Kubernetes is an open source container orchestration engine for automating deployment, scaling, and management of containerized applications. The open source project is hosted by the Cloud Native Computing Foundation.
1.1 - Available Documentation Versions
This website contains documentation for the current version of Kubernetes
and the four previous versions of Kubernetes.
The availability of documentation for a Kubernetes version is separate from whether
that release is currently supported.
Read Support period to learn about
which versions of Kubernetes are officially supported, and for how long.
2 - Getting started
This section lists the different ways to set up and run Kubernetes.
When you install Kubernetes, choose an installation type based on: ease of maintenance, security,
control, available resources, and expertise required to operate and manage a cluster.
You can download Kubernetes to deploy a Kubernetes cluster
on a local machine, into the cloud, or for your own datacenter.
It is recommended to run Kubernetes components as container images wherever
that is possible, and to have Kubernetes manage those components.
Components that run containers - notably, the kubelet - can't be included in this category.
If you don't want to manage a Kubernetes cluster yourself, you could pick a managed service, including
certified platforms.
There are also other standardized and custom solutions across a wide range of cloud and
bare metal environments.
Learning environment
If you're learning Kubernetes, use the tools supported by the Kubernetes community,
or tools in the ecosystem to set up a Kubernetes cluster on a local machine.
See Install tools.
Production environment
When evaluating a solution for a
production environment, consider which aspects of
operating a Kubernetes cluster (or abstractions) you want to manage yourself and which you
prefer to hand off to a provider.
For a cluster you're managing yourself, the officially supported tool
for deploying Kubernetes is kubeadm.
Kubernetes is designed for its control plane to
run on Linux. Within your cluster you can run applications on Linux or other operating systems, including
Windows.
A production-quality Kubernetes cluster requires planning and preparation.
If your Kubernetes cluster is to run critical workloads, it must be configured to be resilient.
This page explains steps you can take to set up a production-ready cluster,
or to promote an existing cluster for production use.
If you're already familiar with production setup and want the links, skip to
What's next.
Production considerations
Typically, a production Kubernetes cluster environment has more requirements than a
personal learning, development, or test environment Kubernetes. A production environment may require
secure access by many users, consistent availability, and the resources to adapt
to changing demands.
As you decide where you want your production Kubernetes environment to live
(on premises or in a cloud) and the amount of management you want to take
on or hand to others, consider how your requirements for a Kubernetes cluster
are influenced by the following issues:
Availability: A single-machine Kubernetes learning environment
has a single point of failure. Creating a highly available cluster means considering:
Separating the control plane from the worker nodes.
Replicating the control plane components on multiple nodes.
Load balancing traffic to the cluster’s API server.
Having enough worker nodes available, or able to quickly become available, as changing workloads warrant it.
Scale: If you expect your production Kubernetes environment to receive a stable amount of
demand, you might be able to set up for the capacity you need and be done. However,
if you expect demand to grow over time or change dramatically based on things like
season or special events, you need to plan how to scale to relieve increased
pressure from more requests to the control plane and worker nodes or scale down to reduce unused
resources.
Security and access management: You have full admin privileges on your own
Kubernetes learning cluster. But shared clusters with important workloads, and
more than one or two users, require a more refined approach to who and what can
access cluster resources. You can use role-based access control
(RBAC) and other
security mechanisms to make sure that users and workloads can get access to the
resources they need, while keeping workloads, and the cluster itself, secure.
You can set limits on the resources that users and workloads can access
by managing policies and
container resources.
Before building a Kubernetes production environment on your own, consider
handing off some or all of this job to
Turnkey Cloud Solutions
providers or other Kubernetes Partners.
Options include:
Serverless: Just run workloads on third-party equipment without managing
a cluster at all. You will be charged for things like CPU usage, memory, and
disk requests.
Managed control plane: Let the provider manage the scale and availability
of the cluster's control plane, as well as handle patches and upgrades.
Managed worker nodes: Configure pools of nodes to meet your needs,
then the provider makes sure those nodes are available and ready to implement
upgrades when needed.
Integration: There are providers that integrate Kubernetes with other
services you may need, such as storage, container registries, authentication
methods, and development tools.
Whether you build a production Kubernetes cluster yourself or work with
partners, review the following sections to evaluate your needs as they relate
to your cluster’s control plane, worker nodes, user access, and
workload resources.
Production cluster setup
In a production-quality Kubernetes cluster, the control plane manages the
cluster from services that can be spread across multiple computers
in different ways. Each worker node, however, represents a single entity that
is configured to run Kubernetes pods.
Production control plane
The simplest Kubernetes cluster has the entire control plane and worker node
services running on the same machine. You can grow that environment by adding
worker nodes, as reflected in the diagram illustrated in
Kubernetes Components.
If the cluster is meant to be available for a short period of time, or can be
discarded if something goes seriously wrong, this might meet your needs.
If you need a more permanent, highly available cluster, however, you should
consider ways of extending the control plane. By design, one-machine control
plane services running on a single machine are not highly available.
If keeping the cluster up and running
and ensuring that it can be repaired if something goes wrong is important,
consider these steps:
Choose deployment tools: You can deploy a control plane using tools such
as kubeadm, kops, and kubespray. See
Installing Kubernetes with deployment tools
to learn tips for production-quality deployments using each of those deployment
methods. Different Container Runtimes
are available to use with your deployments.
Manage certificates: Secure communications between control plane services
are implemented using certificates. Certificates are automatically generated
during deployment or you can generate them using your own certificate authority.
See PKI certificates and requirements for details.
Configure load balancer for apiserver: Configure a load balancer
to distribute external API requests to the apiserver service instances running on different nodes. See
Create an External Load Balancer
for details.
Separate and backup etcd service: The etcd services can either run on the
same machines as other control plane services or run on separate machines, for
extra security and availability. Because etcd stores cluster configuration data,
backing up the etcd database should be done regularly to ensure that you can
repair that database if needed.
See the etcd FAQ for details on configuring and using etcd.
See Operating etcd clusters for Kubernetes
and Set up a High Availability etcd cluster with kubeadm
for details.
Create multiple control plane systems: For high availability, the
control plane should not be limited to a single machine. If the control plane
services are run by an init service (such as systemd), each service should run on at
least three machines. However, running control plane services as pods in
Kubernetes ensures that the replicated number of services that you request
will always be available.
The scheduler should be fault tolerant,
but not highly available. Some deployment tools set up Raft
consensus algorithm to do leader election of Kubernetes services. If the
primary goes away, another service elects itself and take over.
Span multiple zones: If keeping your cluster available at all times is
critical, consider creating a cluster that runs across multiple data centers,
referred to as zones in cloud environments. Groups of zones are referred to as regions.
By spreading a cluster across
multiple zones in the same region, it can improve the chances that your
cluster will continue to function even if one zone becomes unavailable.
See Running in multiple zones for details.
Manage on-going features: If you plan to keep your cluster over time,
there are tasks you need to do to maintain its health and security. For example,
if you installed with kubeadm, there are instructions to help you with
Certificate Management
and Upgrading kubeadm clusters.
See Administer a Cluster
for a longer list of Kubernetes administrative tasks.
Production-quality workloads need to be resilient and anything they rely
on needs to be resilient (such as CoreDNS). Whether you manage your own
control plane or have a cloud provider do it for you, you still need to
consider how you want to manage your worker nodes (also referred to
simply as nodes).
Configure nodes: Nodes can be physical or virtual machines. If you want to
create and manage your own nodes, you can install a supported operating system,
then add and run the appropriate
Node services. Consider:
The demands of your workloads when you set up nodes by having appropriate memory, CPU, and disk speed and storage capacity available.
Whether generic computer systems will do or you have workloads that need GPU processors, Windows nodes, or VM isolation.
Validate nodes: See Valid node setup
for information on how to ensure that a node meets the requirements to join
a Kubernetes cluster.
Add nodes to the cluster: If you are managing your own cluster you can
add nodes by setting up your own machines and either adding them manually or
having them register themselves to the cluster’s apiserver. See the
Nodes section for information on how to set up Kubernetes to add nodes in these ways.
Scale nodes: Have a plan for expanding the capacity your cluster will
eventually need. See Considerations for large clusters
to help determine how many nodes you need, based on the number of pods and
containers you need to run. If you are managing nodes yourself, this can mean
purchasing and installing your own physical equipment.
Autoscale nodes: Most cloud providers support
Cluster Autoscaler
to replace unhealthy nodes or grow and shrink the number of nodes as demand requires. See the
Frequently Asked Questions
for how the autoscaler works and
Deployment
for how it is implemented by different cloud providers. For on-premises, there
are some virtualization platforms that can be scripted to spin up new nodes
based on demand.
Set up node health checks: For important workloads, you want to make sure
that the nodes and pods running on those nodes are healthy. Using the
Node Problem Detector
daemon, you can ensure your nodes are healthy.
Production user management
In production, you may be moving from a model where you or a small group of
people are accessing the cluster to where there may potentially be dozens or
hundreds of people. In a learning environment or platform prototype, you might have a single
administrative account for everything you do. In production, you will want
more accounts with different levels of access to different namespaces.
Taking on a production-quality cluster means deciding how you
want to selectively allow access by other users. In particular, you need to
select strategies for validating the identities of those who try to access your
cluster (authentication) and deciding if they have permissions to do what they
are asking (authorization):
Authentication: The apiserver can authenticate users using client
certificates, bearer tokens, an authenticating proxy, or HTTP basic auth.
You can choose which authentication methods you want to use.
Using plugins, the apiserver can leverage your organization’s existing
authentication methods, such as LDAP or Kerberos. See
Authentication
for a description of these different methods of authenticating Kubernetes users.
Authorization: When you set out to authorize your regular users, you will probably choose
between RBAC and ABAC authorization. See Authorization Overview
to review different modes for authorizing user accounts (as well as service account access to
your cluster):
Role-based access control (RBAC): Lets you
assign access to your cluster by allowing specific sets of permissions to authenticated users.
Permissions can be assigned for a specific namespace (Role) or across the entire cluster
(ClusterRole). Then using RoleBindings and ClusterRoleBindings, those permissions can be attached
to particular users.
Attribute-based access control (ABAC): Lets you
create policies based on resource attributes in the cluster and will allow or deny access
based on those attributes. Each line of a policy file identifies versioning properties (apiVersion
and kind) and a map of spec properties to match the subject (user or group), resource property,
non-resource property (/version or /apis), and readonly. See
Examples for details.
As someone setting up authentication and authorization on your production Kubernetes cluster, here are some things to consider:
Set the authorization mode: When the Kubernetes API server
(kube-apiserver)
starts, the supported authentication modes must be set using the --authorization-mode
flag. For example, that flag in the kube-adminserver.yaml file (in /etc/kubernetes/manifests)
could be set to Node,RBAC. This would allow Node and RBAC authorization for authenticated requests.
Create user certificates and role bindings (RBAC): If you are using RBAC
authorization, users can create a CertificateSigningRequest (CSR) that can be
signed by the cluster CA. Then you can bind Roles and ClusterRoles to each user.
See Certificate Signing Requests
for details.
Create policies that combine attributes (ABAC): If you are using ABAC
authorization, you can assign combinations of attributes to form policies to
authorize selected users or groups to access particular resources (such as a
pod), namespace, or apiGroup. For more information, see
Examples.
Consider Admission Controllers: Additional forms of authorization for
requests that can come in through the API server include
Webhook Token Authentication.
Webhooks and other special authorization types need to be enabled by adding
Admission Controllers
to the API server.
Set limits on workload resources
Demands from production workloads can cause pressure both inside and outside
of the Kubernetes control plane. Consider these items when setting up for the
needs of your cluster's workloads:
Prepare for DNS demand: If you expect workloads to massively scale up,
your DNS service must be ready to scale up as well. See
Autoscale the DNS service in a Cluster.
Create additional service accounts: User accounts determine what users can
do on a cluster, while a service account defines pod access within a particular
namespace. By default, a pod takes on the default service account from its namespace.
See Managing Service Accounts
for information on creating a new service account. For example, you might want to:
Add secrets that a pod could use to pull images from a particular container registry. See
Configure Service Accounts for Pods
for an example.
If you choose to build your own cluster, plan how you want to
handle certificates
and set up high availability for features such as
etcd
and the
API server.
Note: Dockershim has been removed from the Kubernetes project as of release 1.24. Read the Dockershim Removal FAQ for further details.
You need to install a
container runtime
into each node in the cluster so that Pods can run there. This page outlines
what is involved and describes related tasks for setting up nodes.
Kubernetes releases before v1.24 included a direct integration with Docker Engine,
using a component named dockershim. That special direct integration is no longer
part of Kubernetes (this removal was
announced
as part of the v1.20 release).
You can read
Check whether Dockershim removal affects you
to understand how this removal might affect you. To learn about migrating from using dockershim, see
Migrating from dockershim.
If you are running a version of Kubernetes other than v1.25,
check the documentation for that version.
Install and configure prerequisites
The following steps apply common settings for Kubernetes nodes on Linux.
You can skip a particular setting if you're certain you don't need it.
For more information, see Network Plugin Requirements or the documentation for your specific container runtime.
Forwarding IPv4 and letting iptables see bridged traffic
Verify that the br_netfilter module is loaded by running lsmod | grep br_netfilter.
To load it explicitly, run sudo modprobe br_netfilter.
In order for a Linux node's iptables to correctly view bridged traffic, verify that net.bridge.bridge-nf-call-iptables is set to 1 in your sysctl config. For example:
cat <<EOF | sudo tee /etc/modules-load.d/k8s.conf
overlay
br_netfilter
EOFsudo modprobe overlay
sudo modprobe br_netfilter
# sysctl params required by setup, params persist across rebootscat <<EOF | sudo tee /etc/sysctl.d/k8s.conf
net.bridge.bridge-nf-call-iptables = 1
net.bridge.bridge-nf-call-ip6tables = 1
net.ipv4.ip_forward = 1
EOF# Apply sysctl params without rebootsudo sysctl --system
Cgroup drivers
On Linux, control groups
are used to constrain resources that are allocated to processes.
Both kubelet and the
underlying container runtime need to interface with control groups to enforce
resource management for pods and containers and set
resources such as cpu/memory requests and limits. To interface with control
groups, the kubelet and the container runtime need to use a cgroup driver.
It's critical that the kubelet and the container runtime uses the same cgroup
driver and are configured the same.
The cgroupfs driver is the default cgroup driver in the kubelet. When the cgroupfs
driver is used, the kubelet and the container runtime directly interface with
the cgroup filesystem to configure cgroups.
The cgroupfs driver is not recommended when
systemd is the
init system because systemd expects a single cgroup manager on
the system. Additionally, if you use cgroup v2
, use the systemd cgroup driver instead of
cgroupfs.
systemd cgroup driver
When systemd is chosen as the init
system for a Linux distribution, the init process generates and consumes a root control group
(cgroup) and acts as a cgroup manager.
systemd has a tight integration with cgroups and allocates a cgroup per systemd
unit. As a result, if you use systemd as the init system with the cgroupfs
driver, the system gets two different cgroup managers.
Two cgroup managers result in two views of the available and in-use resources in
the system. In some cases, nodes that are configured to use cgroupfs for the
kubelet and container runtime, but use systemd for the rest of the processes become
unstable under resource pressure.
The approach to mitigate this instability is to use systemd as the cgroup driver for
the kubelet and the container runtime when systemd is the selected init system.
To set systemd as the cgroup driver, edit the
KubeletConfiguration
option of cgroupDriver and set it to systemd. For example:
If you configure systemd as the cgroup driver for the kubelet, you must also
configure systemd as the cgroup driver for the container runtime. Refer to
the documentation for your container runtime for instructions. For example:
Changing the cgroup driver of a Node that has joined a cluster is a sensitive operation.
If the kubelet has created Pods using the semantics of one cgroup driver, changing the container
runtime to another cgroup driver can cause errors when trying to re-create the Pod sandbox
for such existing Pods. Restarting the kubelet may not solve such errors.
If you have automation that makes it feasible, replace the node with another using the updated
configuration, or reinstall it using automation.
Migrating to the systemd driver in kubeadm managed clusters
If you wish to migrate to the systemd cgroup driver in existing kubeadm managed clusters,
follow configuring a cgroup driver.
CRI version support
Your container runtime must support at least v1alpha2 of the container runtime interface.
Kubernetes 1.25 defaults to using v1 of the CRI API.
If a container runtime does not support the v1 API, the kubelet falls back to
using the (deprecated) v1alpha2 API instead.
Container runtimes
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
containerd
This section outlines the necessary steps to use containerd as CRI runtime.
Use the following commands to install Containerd on your system:
Follow the instructions for getting started with containerd. Return to this step once you've created a valid configuration file, config.toml.
You can find this file under the path /etc/containerd/config.toml.
You can find this file under the path C:\Program Files\containerd\config.toml.
On Linux the default CRI socket for containerd is /run/containerd/containerd.sock.
On Windows the default CRI endpoint is npipe://./pipe/containerd-containerd.
Configuring the systemd cgroup driver
To use the systemd cgroup driver in /etc/containerd/config.toml with runc, set
The systemd cgroup driver is recommended if you use cgroup v2.
Note:
If you installed containerd from a package (for example, RPM or .deb), you may find
that the CRI integration plugin is disabled by default.
You need CRI support enabled to use containerd with Kubernetes. Make sure that cri
is not included in thedisabled_plugins list within /etc/containerd/config.toml;
if you made changes to that file, also restart containerd.
If you apply this change, make sure to restart containerd:
CRI-O uses the systemd cgroup driver per default, which is likely to work fine
for you. To switch to the cgroupfs cgroup driver, either edit
/etc/crio/crio.conf or place a drop-in configuration in
/etc/crio/crio.conf.d/02-cgroup-manager.conf, for example:
You should also note the changed conmon_cgroup, which has to be set to the value
pod when using CRI-O with cgroupfs. It is generally necessary to keep the
cgroup driver configuration of the kubelet (usually done via kubeadm) and CRI-O
in sync.
For CRI-O, the CRI socket is /var/run/crio/crio.sock by default.
Overriding the sandbox (pause) image
In your CRI-O config you can set the following
config value:
This config option supports live configuration reload to apply this change: systemctl reload crio or by sending
SIGHUP to the crio process.
Docker Engine
Note: These instructions assume that you are using the
cri-dockerd adapter to integrate
Docker Engine with Kubernetes.
On each of your nodes, install Docker for your Linux distribution as per
Install Docker Engine.
Install cri-dockerd, following
the instructions in that source code repository.
For cri-dockerd, the CRI socket is /run/cri-dockerd.sock by default.
Overriding the sandbox (pause) image
The cri-dockerd adapter accepts a command line argument for
specifying which container image to use as the Pod infrastructure container (“pause image”).
The command line argument to use is --pod-infra-container-image.
Mirantis Container Runtime
Mirantis Container Runtime (MCR) is a commercially
available container runtime that was formerly known as Docker Enterprise Edition.
You can use Mirantis Container Runtime with Kubernetes using the open source
cri-dockerd component, included with MCR.
To learn more about how to install Mirantis Container Runtime,
visit MCR Deployment Guide.
Check the systemd unit named cri-docker.socket to find out the path to the CRI
socket.
Overriding the sandbox (pause) image
The cri-dockerd adapter accepts a command line argument for
specifying which container image to use as the Pod infrastructure container (“pause image”).
The command line argument to use is --pod-infra-container-image.
What's next
As well as a container runtime, your cluster will need a working
network plugin.
2.2.2 - Installing Kubernetes with deployment tools
2.2.2.1 - Bootstrapping clusters with kubeadm
2.2.2.1.1 - Installing kubeadm
This page shows how to install the kubeadm toolbox.
For information on how to create a cluster with kubeadm once you have performed this installation process, see the Creating a cluster with kubeadm page.
Before you begin
A compatible Linux host. The Kubernetes project provides generic instructions for Linux distributions based on Debian and Red Hat, and those distributions without a package manager.
2 GB or more of RAM per machine (any less will leave little room for your apps).
2 CPUs or more.
Full network connectivity between all machines in the cluster (public or private network is fine).
Unique hostname, MAC address, and product_uuid for every node. See here for more details.
Certain ports are open on your machines. See here for more details.
Swap disabled. You MUST disable swap in order for the kubelet to work properly.
Verify the MAC address and product_uuid are unique for every node
You can get the MAC address of the network interfaces using the command ip link or ifconfig -a
The product_uuid can be checked by using the command sudo cat /sys/class/dmi/id/product_uuid
It is very likely that hardware devices will have unique addresses, although some virtual machines may have
identical values. Kubernetes uses these values to uniquely identify the nodes in the cluster.
If these values are not unique to each node, the installation process
may fail.
Check network adapters
If you have more than one network adapter, and your Kubernetes components are not reachable on the default
route, we recommend you add IP route(s) so Kubernetes cluster addresses go via the appropriate adapter.
Check required ports
These
required ports
need to be open in order for Kubernetes components to communicate with each other. You can use tools like netcat to check if a port is open. For example:
nc 127.0.0.1 6443
The pod network plugin you use may also require certain ports to be
open. Since this differs with each pod network plugin, please see the
documentation for the plugins about what port(s) those need.
Note: Docker Engine does not implement the CRI
which is a requirement for a container runtime to work with Kubernetes.
For that reason, an additional service cri-dockerd
has to be installed. cri-dockerd is a project based on the legacy built-in
Docker Engine support that was removed from the kubelet in version 1.24.
The tables below include the known endpoints for supported operating systems:
You will install these packages on all of your machines:
kubeadm: the command to bootstrap the cluster.
kubelet: the component that runs on all of the machines in your cluster
and does things like starting pods and containers.
kubectl: the command line util to talk to your cluster.
kubeadm will not install or manage kubelet or kubectl for you, so you will
need to ensure they match the version of the Kubernetes control plane you want
kubeadm to install for you. If you do not, there is a risk of a version skew occurring that
can lead to unexpected, buggy behaviour. However, one minor version skew between the
kubelet and the control plane is supported, but the kubelet version may never exceed the API
server version. For example, the kubelet running 1.7.0 should be fully compatible with a 1.8.0 API server,
but not vice versa.
Warning: These instructions exclude all Kubernetes packages from any system upgrades.
This is because kubeadm and Kubernetes require
special attention to upgrade.
Setting SELinux in permissive mode by running setenforce 0 and sed ... effectively disables it.
This is required to allow containers to access the host filesystem, which is needed by pod networks for example.
You have to do this until SELinux support is improved in the kubelet.
You can leave SELinux enabled if you know how to configure it but it may require settings that are not supported by kubeadm.
If the baseurl fails because your Red Hat-based distribution cannot interpret basearch, replace \$basearch with your computer's architecture.
Type uname -m to see that value.
For example, the baseurl URL for x86_64 could be: https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64.
Install CNI plugins (required for most pod network):
CRICTL_VERSION="v1.22.0"ARCH="amd64"curl -L "https://github.com/kubernetes-sigs/cri-tools/releases/download/${CRICTL_VERSION}/crictl-${CRICTL_VERSION}-linux-${ARCH}.tar.gz" | sudo tar -C $DOWNLOAD_DIR -xz
Install kubeadm, kubelet, kubectl and add a kubelet systemd service:
RELEASE="$(curl -sSL https://dl.k8s.io/release/stable.txt)"ARCH="amd64"cd$DOWNLOAD_DIRsudo curl -L --remote-name-all https://storage.googleapis.com/kubernetes-release/release/${RELEASE}/bin/linux/${ARCH}/{kubeadm,kubelet,kubectl}sudo chmod +x {kubeadm,kubelet,kubectl}RELEASE_VERSION="v0.4.0"curl -sSL "https://raw.githubusercontent.com/kubernetes/release/${RELEASE_VERSION}/cmd/kubepkg/templates/latest/deb/kubelet/lib/systemd/system/kubelet.service" | sed "s:/usr/bin:${DOWNLOAD_DIR}:g" | sudo tee /etc/systemd/system/kubelet.service
sudo mkdir -p /etc/systemd/system/kubelet.service.d
curl -sSL "https://raw.githubusercontent.com/kubernetes/release/${RELEASE_VERSION}/cmd/kubepkg/templates/latest/deb/kubeadm/10-kubeadm.conf" | sed "s:/usr/bin:${DOWNLOAD_DIR}:g" | sudo tee /etc/systemd/system/kubelet.service.d/10-kubeadm.conf
Enable and start kubelet:
systemctl enable --now kubelet
Note: The Flatcar Container Linux distribution mounts the /usr directory as a read-only filesystem.
Before bootstrapping your cluster, you need to take additional steps to configure a writable directory.
See the Kubeadm Troubleshooting guide to learn how to set up a writable directory.
The kubelet is now restarting every few seconds, as it waits in a crashloop for
kubeadm to tell it what to do.
Configuring a cgroup driver
Both the container runtime and the kubelet have a property called
"cgroup driver", which is important
for the management of cgroups on Linux machines.
Warning:
Matching the container runtime and kubelet cgroup drivers is required or otherwise the kubelet process will fail.
As with any program, you might run into an error installing or running kubeadm.
This page lists some common failure scenarios and have provided steps that can help you understand and fix the problem.
If your problem is not listed below, please follow the following steps:
If no issue exists, please open one and follow the issue template.
If you are unsure about how kubeadm works, you can ask on Slack in #kubeadm,
or open a question on StackOverflow. Please include
relevant tags like #kubernetes and #kubeadm so folks can help you.
Not possible to join a v1.18 Node to a v1.17 cluster due to missing RBAC
In v1.18 kubeadm added prevention for joining a Node in the cluster if a Node with the same name already exists.
This required adding RBAC for the bootstrap-token user to be able to GET a Node object.
However this causes an issue where kubeadm join from v1.18 cannot join a cluster created by kubeadm v1.17.
To workaround the issue you have two options:
Execute kubeadm init phase bootstrap-token on a control-plane node using kubeadm v1.18.
Note that this enables the rest of the bootstrap-token permissions as well.
or
Apply the following RBAC manually using kubectl apply -f ...:
ebtables or some similar executable not found during installation
If you see the following warnings while running kubeadm init
[preflight] WARNING: ebtables not found in system path
[preflight] WARNING: ethtool not found in system path
Then you may be missing ebtables, ethtool or a similar executable on your node. You can install them with the following commands:
For Ubuntu/Debian users, run apt install ebtables ethtool.
For CentOS/Fedora users, run yum install ebtables ethtool.
kubeadm blocks waiting for control plane during installation
If you notice that kubeadm init hangs after printing out the following line:
[apiclient] Created API client, waiting for the control plane to become ready
This may be caused by a number of problems. The most common are:
network connection problems. Check that your machine has full network connectivity before continuing.
the cgroup driver of the container runtime differs from that of the kubelet. To understand how to
configure it properly see Configuring a cgroup driver.
control plane containers are crashlooping or hanging. You can check this by running docker ps
and investigating each container by running docker logs. For other container runtime see
Debugging Kubernetes nodes with crictl.
kubeadm blocks when removing managed containers
The following could happen if the container runtime halts and does not remove
any Kubernetes-managed containers:
sudo kubeadm reset
[preflight] Running pre-flight checks
[reset] Stopping the kubelet service
[reset] Unmounting mounted directories in "/var/lib/kubelet"
[reset] Removing kubernetes-managed containers
(block)
A possible solution is to restart the container runtime and then re-run kubeadm reset.
You can also use crictl to debug the state of the container runtime. See
Debugging Kubernetes nodes with crictl.
Pods in RunContainerError, CrashLoopBackOff or Error state
Right after kubeadm init there should not be any pods in these states.
If there are pods in one of these states right afterkubeadm init, please open an
issue in the kubeadm repo. coredns (or kube-dns) should be in the Pending state
until you have deployed the network add-on.
If you see Pods in the RunContainerError, CrashLoopBackOff or Error state
after deploying the network add-on and nothing happens to coredns (or kube-dns),
it's very likely that the Pod Network add-on that you installed is somehow broken.
You might have to grant it more RBAC privileges or use a newer version. Please file
an issue in the Pod Network providers' issue tracker and get the issue triaged there.
coredns is stuck in the Pending state
This is expected and part of the design. kubeadm is network provider-agnostic, so the admin
should install the pod network add-on
of choice. You have to install a Pod Network
before CoreDNS may be deployed fully. Hence the Pending state before the network is set up.
HostPort services do not work
The HostPort and HostIP functionality is available depending on your Pod Network
provider. Please contact the author of the Pod Network add-on to find out whether
HostPort and HostIP functionality are available.
Calico, Canal, and Flannel CNI providers are verified to support HostPort.
If your network provider does not support the portmap CNI plugin, you may need to use the NodePort feature of
services or use HostNetwork=true.
Pods are not accessible via their Service IP
Many network add-ons do not yet enable hairpin mode
which allows pods to access themselves via their Service IP. This is an issue related to
CNI. Please contact the network
add-on provider to get the latest status of their support for hairpin mode.
If you are using VirtualBox (directly or via Vagrant), you will need to
ensure that hostname -i returns a routable IP address. By default the first
interface is connected to a non-routable host-only network. A work around
is to modify /etc/hosts, see this Vagrantfile
for an example.
TLS certificate errors
The following error indicates a possible certificate mismatch.
# kubectl get pods
Unable to connect to the server: x509: certificate signed by unknown authority (possibly because of "crypto/rsa: verification error" while trying to verify candidate authority certificate "kubernetes")
Verify that the $HOME/.kube/config file contains a valid certificate, and
regenerate a certificate if necessary. The certificates in a kubeconfig file
are base64 encoded. The base64 --decode command can be used to decode the certificate
and openssl x509 -text -noout can be used for viewing the certificate information.
Unset the KUBECONFIG environment variable using:
unset KUBECONFIG
Or set it to the default KUBECONFIG location:
exportKUBECONFIG=/etc/kubernetes/admin.conf
Another workaround is to overwrite the existing kubeconfig for the "admin" user:
By default, kubeadm configures a kubelet with automatic rotation of client certificates by using the /var/lib/kubelet/pki/kubelet-client-current.pem symlink specified in /etc/kubernetes/kubelet.conf.
If this rotation process fails you might see errors such as x509: certificate has expired or is not yet valid
in kube-apiserver logs. To fix the issue you must follow these steps:
Backup and delete /etc/kubernetes/kubelet.conf and /var/lib/kubelet/pki/kubelet-client* from the failed node.
From a working control plane node in the cluster that has /etc/kubernetes/pki/ca.key execute
kubeadm kubeconfig user --org system:nodes --client-name system:node:$NODE > kubelet.conf.
$NODE must be set to the name of the existing failed node in the cluster.
Modify the resulted kubelet.conf manually to adjust the cluster name and server endpoint,
or pass kubeconfig user --config (it accepts InitConfiguration). If your cluster does not have
the ca.key you must sign the embedded certificates in the kubelet.conf externally.
Copy this resulted kubelet.conf to /etc/kubernetes/kubelet.conf on the failed node.
Restart the kubelet (systemctl restart kubelet) on the failed node and wait for
/var/lib/kubelet/pki/kubelet-client-current.pem to be recreated.
Manually edit the kubelet.conf to point to the rotated kubelet client certificates, by replacing
client-certificate-data and client-key-data with:
Default NIC When using flannel as the pod network in Vagrant
The following error might indicate that something was wrong in the pod network:
Error from server (NotFound): the server could not find the requested resource
If you're using flannel as the pod network inside Vagrant, then you will have to specify the default interface name for flannel.
Vagrant typically assigns two interfaces to all VMs. The first, for which all hosts are assigned the IP address 10.0.2.15, is for external traffic that gets NATed.
This may lead to problems with flannel, which defaults to the first interface on a host. This leads to all hosts thinking they have the same public IP address. To prevent this, pass the --iface eth1 flag to flannel so that the second interface is chosen.
Non-public IP used for containers
In some situations kubectl logs and kubectl run commands may return with the following errors in an otherwise functional cluster:
Error from server: Get https://10.19.0.41:10250/containerLogs/default/mysql-ddc65b868-glc5m/mysql: dial tcp 10.19.0.41:10250: getsockopt: no route to host
This may be due to Kubernetes using an IP that can not communicate with other IPs on the seemingly same subnet, possibly by policy of the machine provider.
DigitalOcean assigns a public IP to eth0 as well as a private one to be used internally as anchor for their floating IP feature, yet kubelet will pick the latter as the node's InternalIP instead of the public one.
Use ip addr show to check for this scenario instead of ifconfig because ifconfig will not display the offending alias IP address. Alternatively an API endpoint specific to DigitalOcean allows to query for the anchor IP from the droplet:
The workaround is to tell kubelet which IP to use using --node-ip.
When using DigitalOcean, it can be the public one (assigned to eth0) or
the private one (assigned to eth1) should you want to use the optional
private network. The kubeletExtraArgs section of the kubeadm
NodeRegistrationOptions structure
can be used for this.
Then restart kubelet:
systemctl daemon-reload
systemctl restart kubelet
coredns pods have CrashLoopBackOff or Error state
If you have nodes that are running SELinux with an older version of Docker you might experience a scenario
where the coredns pods are not starting. To solve that you can try one of the following options:
Modify the coredns deployment to set allowPrivilegeEscalation to true:
kubectl -n kube-system get deployment coredns -o yaml | \
sed 's/allowPrivilegeEscalation: false/allowPrivilegeEscalation: true/g' | \
kubectl apply -f -
Another cause for CoreDNS to have CrashLoopBackOff is when a CoreDNS Pod deployed in Kubernetes detects a loop. A number of workarounds
are available to avoid Kubernetes trying to restart the CoreDNS Pod every time CoreDNS detects the loop and exits.
Warning: Disabling SELinux or setting allowPrivilegeEscalation to true can compromise
the security of your cluster.
etcd pods restart continually
If you encounter the following error:
rpc error: code = 2 desc = oci runtime error: exec failed: container_linux.go:247: starting container process caused "process_linux.go:110: decoding init error from pipe caused \"read parent: connection reset by peer\""
this issue appears if you run CentOS 7 with Docker 1.13.1.84.
This version of Docker can prevent the kubelet from executing into the etcd container.
To work around the issue, choose one of these options:
Roll back to an earlier version of Docker, such as 1.13.1-75
Not possible to pass a comma separated list of values to arguments inside a --component-extra-args flag
kubeadm init flags such as --component-extra-args allow you to pass custom arguments to a control-plane
component like the kube-apiserver. However, this mechanism is limited due to the underlying type used for parsing
the values (mapStringString).
If you decide to pass an argument that supports multiple, comma-separated values such as
--apiserver-extra-args "enable-admission-plugins=LimitRanger,NamespaceExists" this flag will fail with
flag: malformed pair, expect string=string. This happens because the list of arguments for
--apiserver-extra-args expects key=value pairs and in this case NamespacesExists is considered
as a key that is missing a value.
Alternatively, you can try separating the key=value pairs like so:
--apiserver-extra-args "enable-admission-plugins=LimitRanger,enable-admission-plugins=NamespaceExists"
but this will result in the key enable-admission-plugins only having the value of NamespaceExists.
kube-proxy scheduled before node is initialized by cloud-controller-manager
In cloud provider scenarios, kube-proxy can end up being scheduled on new worker nodes before
the cloud-controller-manager has initialized the node addresses. This causes kube-proxy to fail
to pick up the node's IP address properly and has knock-on effects to the proxy function managing
load balancers.
The following error can be seen in kube-proxy Pods:
server.go:610] Failed to retrieve node IP: host IP unknown; known addresses: []
proxier.go:340] invalid nodeIP, initializing kube-proxy with 127.0.0.1 as nodeIP
A known solution is to patch the kube-proxy DaemonSet to allow scheduling it on control-plane
nodes regardless of their conditions, keeping it off of other nodes until their initial guarding
conditions abate:
On Linux distributions such as Fedora CoreOS or Flatcar Container Linux, the directory /usr is mounted as a read-only filesystem.
For flex-volume support,
Kubernetes components like the kubelet and kube-controller-manager use the default path of
/usr/libexec/kubernetes/kubelet-plugins/volume/exec/, yet the flex-volume directory must be writeable
for the feature to work.
(Note: FlexVolume was deprecated in the Kubernetes v1.23 release)
To workaround this issue you can configure the flex-volume directory using the kubeadm
configuration file.
On the primary control-plane Node (created using kubeadm init) pass the following
file using --config:
Alternatively, you can modify /etc/fstab to make the /usr mount writeable, but please
be advised that this is modifying a design principle of the Linux distribution.
kubeadm upgrade plan prints out context deadline exceeded error message
This error message is shown when upgrading a Kubernetes cluster with kubeadm in the case of running an external etcd. This is not a critical bug and happens because older versions of kubeadm perform a version check on the external etcd cluster. You can proceed with kubeadm upgrade apply ....
This issue is fixed as of version 1.19.
kubeadm reset unmounts /var/lib/kubelet
If /var/lib/kubelet is being mounted, performing a kubeadm reset will effectively unmount it.
To workaround the issue, re-mount the /var/lib/kubelet directory after performing the kubeadm reset operation.
This is a regression introduced in kubeadm 1.15. The issue is fixed in 1.20.
Cannot use the metrics-server securely in a kubeadm cluster
In a kubeadm cluster, the metrics-server
can be used insecurely by passing the --kubelet-insecure-tls to it. This is not recommended for production clusters.
If you want to use TLS between the metrics-server and the kubelet there is a problem,
since kubeadm deploys a self-signed serving certificate for the kubelet. This can cause the following errors
on the side of the metrics-server:
x509: certificate signed by unknown authority
x509: certificate is valid for IP-foo not IP-bar
Using kubeadm, you can create a minimum viable Kubernetes cluster that conforms to best practices.
In fact, you can use kubeadm to set up a cluster that will pass the
Kubernetes Conformance tests.
kubeadm also supports other cluster lifecycle functions, such as
bootstrap tokens and cluster upgrades.
The kubeadm tool is good if you need:
A simple way for you to try out Kubernetes, possibly for the first time.
A way for existing users to automate setting up a cluster and test their application.
A building block in other ecosystem and/or installer tools with a larger
scope.
You can install and use kubeadm on various machines: your laptop, a set
of cloud servers, a Raspberry Pi, and more. Whether you're deploying into the
cloud or on-premises, you can integrate kubeadm into provisioning systems such
as Ansible or Terraform.
Before you begin
To follow this guide, you need:
One or more machines running a deb/rpm-compatible Linux OS; for example: Ubuntu or CentOS.
2 GiB or more of RAM per machine--any less leaves little room for your
apps.
At least 2 CPUs on the machine that you use as a control-plane node.
Full network connectivity among all machines in the cluster. You can use either a
public or a private network.
You also need to use a version of kubeadm that can deploy the version
of Kubernetes that you want to use in your new cluster.
Kubernetes' version and version skew support policy
applies to kubeadm as well as to Kubernetes overall.
Check that policy to learn about what versions of Kubernetes and kubeadm
are supported. This page is written for Kubernetes v1.25.
The kubeadm tool's overall feature state is General Availability (GA). Some sub-features are
still under active development. The implementation of creating the cluster may change
slightly as the tool evolves, but the overall implementation should be pretty stable.
Note: Any commands under kubeadm alpha are, by definition, supported on an alpha level.
Objectives
Install a single control-plane Kubernetes cluster
Install a Pod network on the cluster so that your Pods can
talk to each other
If you have already installed kubeadm, run
apt-get update && apt-get upgrade or
yum update to get the latest version of kubeadm.
When you upgrade, the kubelet restarts every few seconds as it waits in a crashloop for
kubeadm to tell it what to do. This crashloop is expected and normal.
After you initialize your control-plane, the kubelet runs normally.
Preparing the required container images
This step is optional and only applies in case you wish kubeadm init and kubeadm join
to not download the default container images which are hosted at registry.k8s.io.
Kubeadm has commands that can help you pre-pull the required images
when creating a cluster without an internet connection on its nodes.
See Running kubeadm without an internet connection
for more details.
Kubeadm allows you to use a custom image repository for the required images.
See Using custom images
for more details.
Initializing your control-plane node
The control-plane node is the machine where the control plane components run, including
etcd (the cluster database) and the
API Server
(which the kubectl command line tool
communicates with).
(Recommended) If you have plans to upgrade this single control-plane kubeadm cluster
to high availability you should specify the --control-plane-endpoint to set the shared endpoint
for all control-plane nodes. Such an endpoint can be either a DNS name or an IP address of a load-balancer.
Choose a Pod network add-on, and verify whether it requires any arguments to
be passed to kubeadm init. Depending on which
third-party provider you choose, you might need to set the --pod-network-cidr to
a provider-specific value. See Installing a Pod network add-on.
(Optional) kubeadm tries to detect the container runtime by using a list of well
known endpoints. To use different container runtime or if there are more than one installed
on the provisioned node, specify the --cri-socket argument to kubeadm. See
Installing a runtime.
(Optional) Unless otherwise specified, kubeadm uses the network interface associated
with the default gateway to set the advertise address for this particular control-plane node's API server.
To use a different network interface, specify the --apiserver-advertise-address=<ip-address> argument
to kubeadm init. To deploy an IPv6 Kubernetes cluster using IPv6 addressing, you
must specify an IPv6 address, for example --apiserver-advertise-address=fd00::101
To initialize the control-plane node run:
kubeadm init <args>
Considerations about apiserver-advertise-address and ControlPlaneEndpoint
While --apiserver-advertise-address can be used to set the advertise address for this particular
control-plane node's API server, --control-plane-endpoint can be used to set the shared endpoint
for all control-plane nodes.
--control-plane-endpoint allows both IP addresses and DNS names that can map to IP addresses.
Please contact your network administrator to evaluate possible solutions with respect to such mapping.
Here is an example mapping:
192.168.0.102 cluster-endpoint
Where 192.168.0.102 is the IP address of this node and cluster-endpoint is a custom DNS name that maps to this IP.
This will allow you to pass --control-plane-endpoint=cluster-endpoint to kubeadm init and pass the same DNS name to
kubeadm join. Later you can modify cluster-endpoint to point to the address of your load-balancer in an
high availability scenario.
Turning a single control plane cluster created without --control-plane-endpoint into a highly available cluster
is not supported by kubeadm.
To customize control plane components, including optional IPv6 assignment to liveness probe
for control plane components and etcd server, provide extra arguments to each component as documented in
custom arguments.
If you join a node with a different architecture to your cluster, make sure that your deployed DaemonSets
have container image support for this architecture.
kubeadm init first runs a series of prechecks to ensure that the machine
is ready to run Kubernetes. These prechecks expose warnings and exit on errors. kubeadm init
then downloads and installs the cluster control plane components. This may take several minutes.
After it finishes you should see:
Your Kubernetes control-plane has initialized successfully!
To start using your cluster, you need to run the following as a regular user:
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
You should now deploy a Pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
/docs/concepts/cluster-administration/addons/
You can now join any number of machines by running the following on each node
as root:
kubeadm join <control-plane-host>:<control-plane-port> --token <token> --discovery-token-ca-cert-hash sha256:<hash>
To make kubectl work for your non-root user, run these commands, which are
also part of the kubeadm init output:
Alternatively, if you are the root user, you can run:
exportKUBECONFIG=/etc/kubernetes/admin.conf
Warning: Kubeadm signs the certificate in the admin.conf to have Subject: O = system:masters, CN = kubernetes-admin.
system:masters is a break-glass, super user group that bypasses the authorization layer (e.g. RBAC).
Do not share the admin.conf file with anyone and instead grant users custom permissions by generating
them a kubeconfig file using the kubeadm kubeconfig user command. For more details see
Generating kubeconfig files for additional users.
Make a record of the kubeadm join command that kubeadm init outputs. You
need this command to join nodes to your cluster.
The token is used for mutual authentication between the control-plane node and the joining
nodes. The token included here is secret. Keep it safe, because anyone with this
token can add authenticated nodes to your cluster. These tokens can be listed,
created, and deleted with the kubeadm token command. See the
kubeadm reference guide.
Installing a Pod network add-on
Caution:
This section contains important information about networking setup and
deployment order.
Read all of this advice carefully before proceeding.
You must deploy a
Container Network Interface
(CNI) based Pod network add-on so that your Pods can communicate with each other.
Cluster DNS (CoreDNS) will not start up before a network is installed.
Take care that your Pod network must not overlap with any of the host
networks: you are likely to see problems if there is any overlap.
(If you find a collision between your network plugin's preferred Pod
network and some of your host networks, you should think of a suitable
CIDR block to use instead, then use that during kubeadm init with
--pod-network-cidr and as a replacement in your network plugin's YAML).
By default, kubeadm sets up your cluster to use and enforce use of
RBAC (role based access
control).
Make sure that your Pod network plugin supports RBAC, and so do any manifests
that you use to deploy it.
If you want to use IPv6--either dual-stack, or single-stack IPv6 only
networking--for your cluster, make sure that your Pod network plugin
supports IPv6.
IPv6 support was added to CNI in v0.6.0.
Note: Kubeadm should be CNI agnostic and the validation of CNI providers is out of the scope of our current e2e testing.
If you find an issue related to a CNI plugin you should log a ticket in its respective issue
tracker instead of the kubeadm or kubernetes issue trackers.
Several external projects provide Kubernetes Pod networks using CNI, some of which also
support Network Policy.
You can install a Pod network add-on with the following command on the
control-plane node or a node that has the kubeconfig credentials:
kubectl apply -f <add-on.yaml>
You can install only one Pod network per cluster.
Once a Pod network has been installed, you can confirm that it is working by
checking that the CoreDNS Pod is Running in the output of kubectl get pods --all-namespaces.
And once the CoreDNS Pod is up and running, you can continue by joining your nodes.
If your network is not working or CoreDNS is not in the Running state, check out the
troubleshooting guide
for kubeadm.
Managed node labels
By default, kubeadm enables the NodeRestriction
admission controller that restricts what labels can be self-applied by kubelets on node registration.
The admission controller documentation covers what labels are permitted to be used with the kubelet --node-labels option.
The node-role.kubernetes.io/control-plane label is such a restricted label and kubeadm manually applies it using
a privileged client after a node has been created. To do that manually you can do the same by using kubectl label
and ensure it is using a privileged kubeconfig such as the kubeadm managed /etc/kubernetes/admin.conf.
Control plane node isolation
By default, your cluster will not schedule Pods on the control plane nodes for security
reasons. If you want to be able to schedule Pods on the control plane nodes,
for example for a single machine Kubernetes cluster, run:
This will remove the node-role.kubernetes.io/control-plane:NoSchedule taint
from any nodes that have it, including the control plane nodes, meaning that the
scheduler will then be able to schedule Pods everywhere.
Joining your nodes
The nodes are where your workloads (containers and Pods, etc) run. To add new nodes to your cluster do the following for each machine:
If you do not have the token, you can get it by running the following command on the control-plane node:
kubeadm token list
The output is similar to this:
TOKEN TTL EXPIRES USAGES DESCRIPTION EXTRA GROUPS
8ewj1p.9r9hcjoqgajrj4gi 23h 2018-06-12T02:51:28Z authentication, The default bootstrap system:
signing token generated by bootstrappers:
'kubeadm init'. kubeadm:
default-node-token
By default, tokens expire after 24 hours. If you are joining a node to the cluster after the current token has expired,
you can create a new token by running the following command on the control-plane node:
kubeadm token create
The output is similar to this:
5didvk.d09sbcov8ph2amjw
If you don't have the value of --discovery-token-ca-cert-hash, you can get it by running the
following command chain on the control-plane node:
Note: To specify an IPv6 tuple for <control-plane-host>:<control-plane-port>, IPv6 address must be enclosed in square brackets, for example: [fd00::101]:2073.
The output should look something like:
[preflight] Running pre-flight checks
... (log output of join workflow) ...
Node join complete:
* Certificate signing request sent to control-plane and response
received.
* Kubelet informed of new secure connection details.
Run 'kubectl get nodes' on control-plane to see this machine join.
A few seconds later, you should notice this node in the output from kubectl get nodes when run on the control-plane node.
Note: As the cluster nodes are usually initialized sequentially, the CoreDNS Pods are likely to all run
on the first control-plane node. To provide higher availability, please rebalance the CoreDNS Pods
with kubectl -n kube-system rollout restart deployment coredns after at least one new node is joined.
(Optional) Controlling your cluster from machines other than the control-plane node
In order to get a kubectl on some other computer (e.g. laptop) to talk to your
cluster, you need to copy the administrator kubeconfig file from your control-plane node
to your workstation like this:
scp root@<control-plane-host>:/etc/kubernetes/admin.conf .
kubectl --kubeconfig ./admin.conf get nodes
Note:
The example above assumes SSH access is enabled for root. If that is not the
case, you can copy the admin.conf file to be accessible by some other user
and scp using that other user instead.
The admin.conf file gives the user superuser privileges over the cluster.
This file should be used sparingly. For normal users, it's recommended to
generate an unique credential to which you grant privileges. You can do
this with the kubeadm alpha kubeconfig user --client-name <CN>
command. That command will print out a KubeConfig file to STDOUT which you
should save to a file and distribute to your user. After that, grant
privileges by using kubectl create (cluster)rolebinding.
(Optional) Proxying API Server to localhost
If you want to connect to the API Server from outside the cluster you can use
kubectl proxy:
You can now access the API Server locally at http://localhost:8001/api/v1
Clean up
If you used disposable servers for your cluster, for testing, you can
switch those off and do no further clean up. You can use
kubectl config delete-cluster to delete your local references to the
cluster.
However, if you want to deprovision your cluster more cleanly, you should
first drain the node
and make sure that the node is empty, then deconfigure the node.
Remove the node
Talking to the control-plane node with the appropriate credentials, run:
See the Cluster Networking page for a bigger list
of Pod network add-ons.
See the list of add-ons to
explore other add-ons, including tools for logging, monitoring, network policy, visualization &
control of your Kubernetes cluster.
Configure how your cluster handles logs for cluster events and from
applications running in Pods.
See Logging Architecture for
an overview of what is involved.
While kubeadm allows version skew against some components that it manages, it is recommended that you
match the kubeadm version with the versions of the control plane components, kube-proxy and kubelet.
kubeadm's skew against the Kubernetes version
kubeadm can be used with Kubernetes components that are the same version as kubeadm
or one version older. The Kubernetes version can be specified to kubeadm by using the
--kubernetes-version flag of kubeadm init or the
ClusterConfiguration.kubernetesVersion
field when using --config. This option will control the versions
of kube-apiserver, kube-controller-manager, kube-scheduler and kube-proxy.
Example:
kubeadm is at 1.25
kubernetesVersion must be at 1.25 or 1.24
kubeadm's skew against the kubelet
Similarly to the Kubernetes version, kubeadm can be used with a kubelet version that is the same
version as kubeadm or one version older.
Example:
kubeadm is at 1.25
kubelet on the host must be at 1.25 or 1.24
kubeadm's skew against kubeadm
There are certain limitations on how kubeadm commands can operate on existing nodes or whole clusters
managed by kubeadm.
If new nodes are joined to the cluster, the kubeadm binary used for kubeadm join must match
the last version of kubeadm used to either create the cluster with kubeadm init or to upgrade
the same node with kubeadm upgrade. Similar rules apply to the rest of the kubeadm commands
with the exception of kubeadm upgrade.
Example for kubeadm join:
kubeadm version 1.25 was used to create a cluster with kubeadm init
Joining nodes must use a kubeadm binary that is at version 1.25
Nodes that are being upgraded must use a version of kubeadm that is the same MINOR
version or one MINOR version newer than the version of kubeadm used for managing the
node.
Example for kubeadm upgrade:
kubeadm version 1.24 was used to create or upgrade the node
The version of kubeadm used for upgrading the node must be at 1.24
or 1.25
To learn more about the version skew between the different Kubernetes component see
the Version Skew Policy.
Limitations
Cluster resilience
The cluster created here has a single control-plane node, with a single etcd database
running on it. This means that if the control-plane node fails, your cluster may lose
data and may need to be recreated from scratch.
Workarounds:
Regularly back up etcd. The
etcd data directory configured by kubeadm is at /var/lib/etcd on the control-plane node.
kubeadm deb/rpm packages and binaries are built for amd64, arm (32-bit), arm64, ppc64le, and s390x
following the multi-platform
proposal.
Multiplatform container images for the control plane and addons are also supported since v1.12.
Only some of the network providers offer solutions for all platforms. Please consult the list of
network providers above or the documentation from each provider to figure out whether the provider
supports your chosen platform.
Troubleshooting
If you are running into difficulties with kubeadm, please consult our
troubleshooting docs.
2.2.2.1.4 - Customizing components with the kubeadm API
This page covers how to customize the components that kubeadm deploys. For control plane components
you can use flags in the ClusterConfiguration structure or patches per-node. For the kubelet
and kube-proxy you can use KubeletConfiguration and KubeProxyConfiguration, accordingly.
All of these options are possible via the kubeadm configuration API.
For more details on each field in the configuration you can navigate to our
API reference pages.
Note: Customizing the CoreDNS deployment of kubeadm is currently not supported. You must manually
patch the kube-system/corednsConfigMap
and recreate the CoreDNS Pods after that. Alternatively,
you can skip the default CoreDNS deployment and deploy your own variant.
For more details on that see Using init phases with kubeadm.
Customizing the control plane with flags in ClusterConfiguration
The kubeadm ClusterConfiguration object exposes a way for users to override the default
flags passed to control plane components such as the APIServer, ControllerManager, Scheduler and Etcd.
The components are defined using the following structures:
apiServer
controllerManager
scheduler
etcd
These structures contain a common extraArgs field, that consists of key: value pairs.
To override a flag for a control plane component:
Add the appropriate extraArgs to your configuration.
Add flags to the extraArgs field.
Run kubeadm init with --config <YOUR CONFIG YAML>.
Note: You can generate a ClusterConfiguration object with default values by running kubeadm config print init-defaults
and saving the output to a file of your choice.
Note: The ClusterConfiguration object is currently global in kubeadm clusters. This means that any flags that you add,
will apply to all instances of the same component on different nodes. To apply individual configuration per component
on different nodes you can use patches.
Note: Duplicate flags (keys), or passing the same flag --foo multiple times, is currently not supported.
To workaround that you must use patches.
Kubeadm allows you to pass a directory with patch files to InitConfiguration and JoinConfiguration
on individual nodes. These patches can be used as the last customization step before component configuration
is written to disk.
You can pass this file to kubeadm init with --config <YOUR CONFIG YAML>:
The directory must contain files named target[suffix][+patchtype].extension.
For example, kube-apiserver0+merge.yaml or just etcd.json.
target can be one of kube-apiserver, kube-controller-manager, kube-scheduler, etcd
and kubeletconfiguration.
patchtype can be one of strategic, merge or json and these must match the patching formats
supported by kubectl.
The default patchtype is strategic.
extension must be either json or yaml.
suffix is an optional string that can be used to determine which patches are applied first
alpha-numerically.
Note: If you are using kubeadm upgrade to upgrade your kubeadm nodes you must again provide the same
patches, so that the customization is preserved after upgrade. To do that you can use the --patches
flag, which must point to the same directory. kubeadm upgrade currently does not support a configuration
API structure that can be used for the same purpose.
Customizing the kubelet
To customize the kubelet you can add a KubeletConfiguration
next to the ClusterConfiguration or InitConfiguration separated by --- within the same configuration file.
This file can then be passed to kubeadm init and kubeadm will apply the same base KubeletConfiguration
to all nodes in the cluster.
Alternatively, you can use kubelet flags as overrides by passing them in the
nodeRegistration.kubeletExtraArgs field supported by both InitConfiguration and JoinConfiguration.
Some kubelet flags are deprecated, so check their status in the
kubelet reference documentation before using them.
Note: kubeadm deploys kube-proxy as a DaemonSet, which means
that the KubeProxyConfiguration would apply to all instances of kube-proxy in the cluster.
2.2.2.1.5 - Options for Highly Available Topology
This page explains the two options for configuring the topology of your highly available (HA) Kubernetes clusters.
You can set up an HA cluster:
With stacked control plane nodes, where etcd nodes are colocated with control plane nodes
With external etcd nodes, where etcd runs on separate nodes from the control plane
You should carefully consider the advantages and disadvantages of each topology before setting up an HA cluster.
Note: kubeadm bootstraps the etcd cluster statically. Read the etcd Clustering Guide
for more details.
Stacked etcd topology
A stacked HA cluster is a topology where the distributed
data storage cluster provided by etcd is stacked on top of the cluster formed by the nodes managed by
kubeadm that run control plane components.
Each control plane node runs an instance of the kube-apiserver, kube-scheduler, and kube-controller-manager.
The kube-apiserver is exposed to worker nodes using a load balancer.
Each control plane node creates a local etcd member and this etcd member communicates only with
the kube-apiserver of this node. The same applies to the local kube-controller-manager
and kube-scheduler instances.
This topology couples the control planes and etcd members on the same nodes. It is simpler to set up than a cluster
with external etcd nodes, and simpler to manage for replication.
However, a stacked cluster runs the risk of failed coupling. If one node goes down, both an etcd member and a control
plane instance are lost, and redundancy is compromised. You can mitigate this risk by adding more control plane nodes.
You should therefore run a minimum of three stacked control plane nodes for an HA cluster.
This is the default topology in kubeadm. A local etcd member is created automatically
on control plane nodes when using kubeadm init and kubeadm join --control-plane.
External etcd topology
An HA cluster with external etcd is a topology where the distributed data storage cluster provided by etcd is external to the cluster formed by the nodes that run control plane components.
Like the stacked etcd topology, each control plane node in an external etcd topology runs an instance of the kube-apiserver, kube-scheduler, and kube-controller-manager. And the kube-apiserver is exposed to worker nodes using a load balancer. However, etcd members run on separate hosts, and each etcd host communicates with the kube-apiserver of each control plane node.
This topology decouples the control plane and etcd member. It therefore provides an HA setup where
losing a control plane instance or an etcd member has less impact and does not affect
the cluster redundancy as much as the stacked HA topology.
However, this topology requires twice the number of hosts as the stacked HA topology.
A minimum of three hosts for control plane nodes and three hosts for etcd nodes are required for an HA cluster with this topology.
2.2.2.1.6 - Creating Highly Available Clusters with kubeadm
This page explains two different approaches to setting up a highly available Kubernetes
cluster using kubeadm:
With stacked control plane nodes. This approach requires less infrastructure. The etcd members
and control plane nodes are co-located.
With an external etcd cluster. This approach requires more infrastructure. The
control plane nodes and etcd members are separated.
Before proceeding, you should carefully consider which approach best meets the needs of your applications
and environment. Options for Highly Available topology outlines the advantages and disadvantages of each.
If you encounter issues with setting up the HA cluster, please report these
in the kubeadm issue tracker.
Caution: This page does not address running your cluster on a cloud provider. In a cloud
environment, neither approach documented here works with Service objects of type
LoadBalancer, or with dynamic PersistentVolumes.
Before you begin
The prerequisites depend on which topology you have selected for your cluster's
control plane:
Three or more machines that meet kubeadm's minimum requirements for
the control-plane nodes. Having an odd number of control plane nodes can help
with leader selection in the case of machine or zone failure.
Three or more machines that meet kubeadm's minimum requirements for
the control-plane nodes. Having an odd number of control plane nodes can help
with leader selection in the case of machine or zone failure.
including a container runtime, already set up and working
Full network connectivity between all machines in the cluster (public or
private network)
Superuser privileges on all machines using sudo
You can use a different tool; this guide uses sudo in the examples.
SSH access from one device to all nodes in the system
kubeadm and kubelet already installed on all machines.
And you also need:
Three or more additional machines, that will become etcd cluster members.
Having an odd number of members in the etcd cluster is a requirement for achieving
optimal voting quorum.
These machines again need to have kubeadm and kubelet installed.
These machines also require a container runtime, that is already set up and working.
Each host should have access read and fetch images from the Kubernetes container image registry, registry.k8s.io.
If you want to deploy a highly-available cluster where the hosts do not have access to pull images, this is possible. You must ensure by some other means that the correct container images are already available on the relevant hosts.
Command line interface
To manage Kubernetes once your cluster is set up, you should
install kubectl on your PC. It is also useful
to install the kubectl tool on each control plane node, as this can be
helpful for troubleshooting.
First steps for both methods
Create load balancer for kube-apiserver
Note: There are many configurations for load balancers. The following example is only one
option. Your cluster requirements may need a different configuration.
Create a kube-apiserver load balancer with a name that resolves to DNS.
In a cloud environment you should place your control plane nodes behind a TCP
forwarding load balancer. This load balancer distributes traffic to all
healthy control plane nodes in its target list. The health check for
an apiserver is a TCP check on the port the kube-apiserver listens on
(default value :6443).
It is not recommended to use an IP address directly in a cloud environment.
The load balancer must be able to communicate with all control plane nodes
on the apiserver port. It must also allow incoming traffic on its
listening port.
Make sure the address of the load balancer always matches
the address of kubeadm's ControlPlaneEndpoint.
Add the first control plane node to the load balancer, and test the
connection:
nc -v <LOAD_BALANCER_IP> <PORT>
A connection refused error is expected because the API server is not yet
running. A timeout, however, means the load balancer cannot communicate
with the control plane node. If a timeout occurs, reconfigure the load
balancer to communicate with the control plane node.
Add the remaining control plane nodes to the load balancer target group.
You can use the --kubernetes-version flag to set the Kubernetes version to use.
It is recommended that the versions of kubeadm, kubelet, kubectl and Kubernetes match.
The --control-plane-endpoint flag should be set to the address or DNS and port of the load balancer.
The --upload-certs flag is used to upload the certificates that should be shared
across all the control-plane instances to the cluster. If instead, you prefer to copy certs across
control-plane nodes manually or using automation tools, please remove this flag and refer to Manual
certificate distribution section below.
Note: The kubeadm init flags --config and --certificate-key cannot be mixed, therefore if you want
to use the kubeadm configuration
you must add the certificateKey field in the appropriate config locations
(under InitConfiguration and JoinConfiguration: controlPlane).
Note: Some CNI network plugins require additional configuration, for example specifying the pod IP CIDR, while others do not.
See the CNI network documentation.
To add a pod CIDR pass the flag --pod-network-cidr, or if you are using a kubeadm configuration file
set the podSubnet field under the networking object of ClusterConfiguration.
The output looks similar to:
...
You can now join any number of control-plane node by running the following command on each as a root:
kubeadm join 192.168.0.200:6443 --token 9vr73a.a8uxyaju799qwdjv --discovery-token-ca-cert-hash sha256:7c2e69131a36ae2a042a339b33381c6d0d43887e2de83720eff5359e26aec866 --control-plane --certificate-key f8902e114ef118304e561c3ecd4d0b543adc226b7a07f675f56564185ffe0c07
Please note that the certificate-key gives access to cluster sensitive data, keep it secret!
As a safeguard, uploaded-certs will be deleted in two hours; If necessary, you can use kubeadm init phase upload-certs to reload certs afterward.
Then you can join any number of worker nodes by running the following on each as root:
kubeadm join 192.168.0.200:6443 --token 9vr73a.a8uxyaju799qwdjv --discovery-token-ca-cert-hash sha256:7c2e69131a36ae2a042a339b33381c6d0d43887e2de83720eff5359e26aec866
Copy this output to a text file. You will need it later to join control plane and worker nodes to
the cluster.
When --upload-certs is used with kubeadm init, the certificates of the primary control plane
are encrypted and uploaded in the kubeadm-certs Secret.
To re-upload the certificates and generate a new decryption key, use the following command on a
control plane
node that is already joined to the cluster:
You can also specify a custom --certificate-key during init that can later be used by join.
To generate such a key you can use the following command:
kubeadm certs certificate-key
Note: The kubeadm-certs Secret and decryption key expire after two hours.
Caution: As stated in the command output, the certificate key gives access to cluster sensitive data, keep it secret!
Apply the CNI plugin of your choice:
Follow these instructions
to install the CNI provider. Make sure the configuration corresponds to the Pod CIDR specified in the
kubeadm configuration file (if applicable).
Note: You must pick a network plugin that suits your use case and deploy it before you move on to next step.
If you don't do this, you will not be able to launch your cluster properly.
Type the following and watch the pods of the control plane components get started:
kubectl get pod -n kube-system -w
Steps for the rest of the control plane nodes
For each additional control plane node you should:
Execute the join command that was previously given to you by the kubeadm init output on the first node.
It should look something like this:
The --control-plane flag tells kubeadm join to create a new control plane.
The --certificate-key ... will cause the control plane certificates to be downloaded
from the kubeadm-certs Secret in the cluster and be decrypted using the given key.
You can join multiple control-plane nodes in parallel.
External etcd nodes
Setting up a cluster with external etcd nodes is similar to the procedure used for stacked etcd
with the exception that you should setup etcd first, and you should pass the etcd information
in the kubeadm config file.
Set up the etcd cluster
Follow these instructions to set up the etcd cluster.
Replace the value of CONTROL_PLANE with the user@host of the first control-plane node.
Set up the first control plane node
Create a file called kubeadm-config.yaml with the following contents:
---apiVersion:kubeadm.k8s.io/v1beta3kind:ClusterConfigurationkubernetesVersion:stablecontrolPlaneEndpoint:"LOAD_BALANCER_DNS:LOAD_BALANCER_PORT"# change this (see below)etcd:external:endpoints:- https://ETCD_0_IP:2379# change ETCD_0_IP appropriately- https://ETCD_1_IP:2379# change ETCD_1_IP appropriately- https://ETCD_2_IP:2379# change ETCD_2_IP appropriatelycaFile:/etc/kubernetes/pki/etcd/ca.crtcertFile:/etc/kubernetes/pki/apiserver-etcd-client.crtkeyFile:/etc/kubernetes/pki/apiserver-etcd-client.key
Note: The difference between stacked etcd and external etcd here is that the external etcd setup requires
a configuration file with the etcd endpoints under the external object for etcd.
In the case of the stacked etcd topology, this is managed automatically.
Replace the following variables in the config template with the appropriate values for your cluster:
LOAD_BALANCER_DNS
LOAD_BALANCER_PORT
ETCD_0_IP
ETCD_1_IP
ETCD_2_IP
The following steps are similar to the stacked etcd setup:
Run sudo kubeadm init --config kubeadm-config.yaml --upload-certs on this node.
Write the output join commands that are returned to a text file for later use.
Apply the CNI plugin of your choice.
Note: You must pick a network plugin that suits your use case and deploy it before you move on to next step.
If you don't do this, you will not be able to launch your cluster properly.
Steps for the rest of the control plane nodes
The steps are the same as for the stacked etcd setup:
Make sure the first control plane node is fully initialized.
Join each control plane node with the join command you saved to a text file. It's recommended
to join the control plane nodes one at a time.
Don't forget that the decryption key from --certificate-key expires after two hours, by default.
Common tasks after bootstrapping control plane
Install workers
Worker nodes can be joined to the cluster with the command you stored previously
as the output from the kubeadm init command:
If you choose to not use kubeadm init with the --upload-certs flag this means that
you are going to have to manually copy the certificates from the primary control plane node to the
joining control plane nodes.
There are many ways to do this. The following example uses ssh and scp:
SSH is required if you want to control all nodes from a single machine.
Enable ssh-agent on your main device that has access to all other nodes in
the system:
eval $(ssh-agent)
Add your SSH identity to the session:
ssh-add ~/.ssh/path_to_private_key
SSH between nodes to check that the connection is working correctly.
When you SSH to any node, add the -A flag. This flag allows the node that you
have logged into via SSH to access the SSH agent on your PC. Consider alternative
methods if you do not fully trust the security of your user session on the node.
ssh -A 10.0.0.7
When using sudo on any node, make sure to preserve the environment so SSH
forwarding works:
sudo -E -s
After configuring SSH on all the nodes you should run the following script on the first
control plane node after running kubeadm init. This script will copy the certificates from
the first control plane node to the other control plane nodes:
In the following example, replace CONTROL_PLANE_IPS with the IP addresses of the
other control plane nodes.
USER=ubuntu # customizableCONTROL_PLANE_IPS="10.0.0.7 10.0.0.8"for host in ${CONTROL_PLANE_IPS}; do scp /etc/kubernetes/pki/ca.crt "${USER}"@$host:
scp /etc/kubernetes/pki/ca.key "${USER}"@$host:
scp /etc/kubernetes/pki/sa.key "${USER}"@$host:
scp /etc/kubernetes/pki/sa.pub "${USER}"@$host:
scp /etc/kubernetes/pki/front-proxy-ca.crt "${USER}"@$host:
scp /etc/kubernetes/pki/front-proxy-ca.key "${USER}"@$host:
scp /etc/kubernetes/pki/etcd/ca.crt "${USER}"@$host:etcd-ca.crt
# Skip the next line if you are using external etcd scp /etc/kubernetes/pki/etcd/ca.key "${USER}"@$host:etcd-ca.key
done
Caution: Copy only the certificates in the above list. kubeadm will take care of generating the rest of the certificates
with the required SANs for the joining control-plane instances. If you copy all the certificates by mistake,
the creation of additional nodes could fail due to a lack of required SANs.
Then on each joining control plane node you have to run the following script before running kubeadm join.
This script will move the previously copied certificates from the home directory to /etc/kubernetes/pki:
USER=ubuntu # customizablemkdir -p /etc/kubernetes/pki/etcd
mv /home/${USER}/ca.crt /etc/kubernetes/pki/
mv /home/${USER}/ca.key /etc/kubernetes/pki/
mv /home/${USER}/sa.pub /etc/kubernetes/pki/
mv /home/${USER}/sa.key /etc/kubernetes/pki/
mv /home/${USER}/front-proxy-ca.crt /etc/kubernetes/pki/
mv /home/${USER}/front-proxy-ca.key /etc/kubernetes/pki/
mv /home/${USER}/etcd-ca.crt /etc/kubernetes/pki/etcd/ca.crt
# Skip the next line if you are using external etcdmv /home/${USER}/etcd-ca.key /etc/kubernetes/pki/etcd/ca.key
2.2.2.1.7 - Set up a High Availability etcd Cluster with kubeadm
Note: While kubeadm is being used as the management tool for external etcd nodes
in this guide, please note that kubeadm does not plan to support certificate rotation
or upgrades for such nodes. The long term plan is to empower the tool
etcdadm to manage these
aspects.
By default, kubeadm runs a local etcd instance on each control plane node.
It is also possible to treat the etcd cluster as external and provision
etcd instances on separate hosts. The differences between the two approaches are covered in the
Options for Highly Available topology page.
This task walks through the process of creating a high availability external
etcd cluster of three members that can be used by kubeadm during cluster creation.
Before you begin
Three hosts that can talk to each other over TCP ports 2379 and 2380. This
document assumes these default ports. However, they are configurable through
the kubeadm config file.
Each host must have systemd and a bash compatible shell installed.
Each host should have access to the Kubernetes container image registry (registry.k8s.io) or list/pull the required etcd image using
kubeadm config images list/pull. This guide will setup etcd instances as
static pods managed by a kubelet.
Some infrastructure to copy files between hosts. For example ssh and scp
can satisfy this requirement.
Setting up the cluster
The general approach is to generate all certs on one node and only distribute
the necessary files to the other nodes.
Note: kubeadm contains all the necessary cryptographic machinery to generate
the certificates described below; no other cryptographic tooling is required for
this example.
Note: The examples below use IPv4 addresses but you can also configure kubeadm, the kubelet and etcd
to use IPv6 addresses. Dual-stack is supported by some Kubernetes options, but not by etcd. For more details
on Kubernetes dual-stack support see Dual-stack support with kubeadm.
Configure the kubelet to be a service manager for etcd.
Note: You must do this on every host where etcd should be running.
Since etcd was created first, you must override the service priority by creating a new unit file
that has higher precedence than the kubeadm-provided kubelet unit file.
cat << EOF > /etc/systemd/system/kubelet.service.d/20-etcd-service-manager.conf
[Service]
ExecStart=
# Replace "systemd" with the cgroup driver of your container runtime. The default value in the kubelet is "cgroupfs".
# Replace the value of "--container-runtime-endpoint" for a different container runtime if needed.
ExecStart=/usr/bin/kubelet --address=127.0.0.1 --pod-manifest-path=/etc/kubernetes/manifests --cgroup-driver=systemd --container-runtime=remote --container-runtime-endpoint=unix:///var/run/containerd/containerd.sock
Restart=always
EOFsystemctl daemon-reload
systemctl restart kubelet
Check the kubelet status to ensure it is running.
systemctl status kubelet
Create configuration files for kubeadm.
Generate one kubeadm configuration file for each host that will have an etcd
member running on it using the following script.
# Update HOST0, HOST1 and HOST2 with the IPs of your hostsexportHOST0=10.0.0.6
exportHOST1=10.0.0.7
exportHOST2=10.0.0.8
# Update NAME0, NAME1 and NAME2 with the hostnames of your hostsexportNAME0="infra0"exportNAME1="infra1"exportNAME2="infra2"# Create temp directories to store files that will end up on other hosts.mkdir -p /tmp/${HOST0}/ /tmp/${HOST1}/ /tmp/${HOST2}/
HOSTS=(${HOST0}${HOST1}${HOST2})NAMES=(${NAME0}${NAME1}${NAME2})for i in "${!HOSTS[@]}"; doHOST=${HOSTS[$i]}NAME=${NAMES[$i]}cat << EOF > /tmp/${HOST}/kubeadmcfg.yaml
---
apiVersion: "kubeadm.k8s.io/v1beta3"
kind: InitConfiguration
nodeRegistration:
name: ${NAME}
localAPIEndpoint:
advertiseAddress: ${HOST}
---
apiVersion: "kubeadm.k8s.io/v1beta3"
kind: ClusterConfiguration
etcd:
local:
serverCertSANs:
- "${HOST}"
peerCertSANs:
- "${HOST}"
extraArgs:
initial-cluster: ${NAMES[0]}=https://${HOSTS[0]}:2380,${NAMES[1]}=https://${HOSTS[1]}:2380,${NAMES[2]}=https://${HOSTS[2]}:2380
initial-cluster-state: new
name: ${NAME}
listen-peer-urls: https://${HOST}:2380
listen-client-urls: https://${HOST}:2379
advertise-client-urls: https://${HOST}:2379
initial-advertise-peer-urls: https://${HOST}:2380
EOFdone
Generate the certificate authority
If you already have a CA then the only action that is copying the CA's crt and
key file to /etc/kubernetes/pki/etcd/ca.crt and
/etc/kubernetes/pki/etcd/ca.key. After those files have been copied,
proceed to the next step, "Create certificates for each member".
If you do not already have a CA then run this command on $HOST0 (where you
generated the configuration files for kubeadm).
kubeadm init phase certs etcd-ca
This creates two files
/etc/kubernetes/pki/etcd/ca.crt
/etc/kubernetes/pki/etcd/ca.key
Create certificates for each member
kubeadm init phase certs etcd-server --config=/tmp/${HOST2}/kubeadmcfg.yaml
kubeadm init phase certs etcd-peer --config=/tmp/${HOST2}/kubeadmcfg.yaml
kubeadm init phase certs etcd-healthcheck-client --config=/tmp/${HOST2}/kubeadmcfg.yaml
kubeadm init phase certs apiserver-etcd-client --config=/tmp/${HOST2}/kubeadmcfg.yaml
cp -R /etc/kubernetes/pki /tmp/${HOST2}/
# cleanup non-reusable certificatesfind /etc/kubernetes/pki -not -name ca.crt -not -name ca.key -type f -delete
kubeadm init phase certs etcd-server --config=/tmp/${HOST1}/kubeadmcfg.yaml
kubeadm init phase certs etcd-peer --config=/tmp/${HOST1}/kubeadmcfg.yaml
kubeadm init phase certs etcd-healthcheck-client --config=/tmp/${HOST1}/kubeadmcfg.yaml
kubeadm init phase certs apiserver-etcd-client --config=/tmp/${HOST1}/kubeadmcfg.yaml
cp -R /etc/kubernetes/pki /tmp/${HOST1}/
find /etc/kubernetes/pki -not -name ca.crt -not -name ca.key -type f -delete
kubeadm init phase certs etcd-server --config=/tmp/${HOST0}/kubeadmcfg.yaml
kubeadm init phase certs etcd-peer --config=/tmp/${HOST0}/kubeadmcfg.yaml
kubeadm init phase certs etcd-healthcheck-client --config=/tmp/${HOST0}/kubeadmcfg.yaml
kubeadm init phase certs apiserver-etcd-client --config=/tmp/${HOST0}/kubeadmcfg.yaml
# No need to move the certs because they are for HOST0# clean up certs that should not be copied off this hostfind /tmp/${HOST2} -name ca.key -type f -delete
find /tmp/${HOST1} -name ca.key -type f -delete
Copy certificates and kubeadm configs
The certificates have been generated and now they must be moved to their
respective hosts.
Now that the certificates and configs are in place it's time to create the
manifests. On each host run the kubeadm command to generate a static manifest
for etcd.
root@HOST0 $ kubeadm init phase etcd local --config=/tmp/${HOST0}/kubeadmcfg.yaml
root@HOST1 $ kubeadm init phase etcd local --config=$HOME/kubeadmcfg.yaml
root@HOST2 $ kubeadm init phase etcd local --config=$HOME/kubeadmcfg.yaml
Set ${ETCD_TAG} to the version tag of your etcd image. For example 3.4.3-0. To see the etcd image and tag that kubeadm uses execute kubeadm config images list --kubernetes-version ${K8S_VERSION}, where ${K8S_VERSION} is for example v1.17.0
Set ${HOST0}to the IP address of the host you are testing.
What's next
Once you have a working 3 member etcd cluster, you can continue setting up a
highly available control plane using the external etcd method with
kubeadm.
2.2.2.1.8 - Configuring each kubelet in your cluster using kubeadm
Note: Dockershim has been removed from the Kubernetes project as of release 1.24. Read the Dockershim Removal FAQ for further details.
FEATURE STATE:Kubernetes v1.11 [stable]
The lifecycle of the kubeadm CLI tool is decoupled from the
kubelet, which is a daemon that runs
on each node within the Kubernetes cluster. The kubeadm CLI tool is executed by the user when Kubernetes is
initialized or upgraded, whereas the kubelet is always running in the background.
Since the kubelet is a daemon, it needs to be maintained by some kind of an init
system or service manager. When the kubelet is installed using DEBs or RPMs,
systemd is configured to manage the kubelet. You can use a different service
manager instead, but you need to configure it manually.
Some kubelet configuration details need to be the same across all kubelets involved in the cluster, while
other configuration aspects need to be set on a per-kubelet basis to accommodate the different
characteristics of a given machine (such as OS, storage, and networking). You can manage the configuration
of your kubelets manually, but kubeadm now provides a KubeletConfiguration API type for
managing your kubelet configurations centrally.
Kubelet configuration patterns
The following sections describe patterns to kubelet configuration that are simplified by
using kubeadm, rather than managing the kubelet configuration for each Node manually.
Propagating cluster-level configuration to each kubelet
You can provide the kubelet with default values to be used by kubeadm init and kubeadm join
commands. Interesting examples include using a different container runtime or setting the default subnet
used by services.
If you want your services to use the subnet 10.96.0.0/12 as the default for services, you can pass
the --service-cidr parameter to kubeadm:
kubeadm init --service-cidr 10.96.0.0/12
Virtual IPs for services are now allocated from this subnet. You also need to set the DNS address used
by the kubelet, using the --cluster-dns flag. This setting needs to be the same for every kubelet
on every manager and Node in the cluster. The kubelet provides a versioned, structured API object
that can configure most parameters in the kubelet and push out this configuration to each running
kubelet in the cluster. This object is called
KubeletConfiguration.
The KubeletConfiguration allows the user to specify flags such as the cluster DNS IP addresses expressed as
a list of values to a camelCased key, illustrated by the following example:
For more details on the KubeletConfiguration have a look at this section.
Providing instance-specific configuration details
Some hosts require specific kubelet configurations due to differences in hardware, operating system,
networking, or other host-specific parameters. The following list provides a few examples.
The path to the DNS resolution file, as specified by the --resolv-conf kubelet
configuration flag, may differ among operating systems, or depending on whether you are using
systemd-resolved. If this path is wrong, DNS resolution will fail on the Node whose kubelet
is configured incorrectly.
The Node API object .metadata.name is set to the machine's hostname by default,
unless you are using a cloud provider. You can use the --hostname-override flag to override the
default behavior if you need to specify a Node name different from the machine's hostname.
Currently, the kubelet cannot automatically detect the cgroup driver used by the container runtime,
but the value of --cgroup-driver must match the cgroup driver used by the container runtime to ensure
the health of the kubelet.
To specify the container runtime you must set its endpoint with the
--container-runtime-endpoint=<path> flag.
It is possible to configure the kubelet that kubeadm will start if a custom
KubeletConfiguration
API object is passed with a configuration file like so kubeadm ... --config some-config-file.yaml.
By calling kubeadm config print init-defaults --component-configs KubeletConfiguration you can
see all the default values for this structure.
It is also possible to apply instance-specific patches over the base KubeletConfiguration.
Have a look at Customizing the kubelet
for more details.
Workflow when using kubeadm init
When you call kubeadm init, the kubelet configuration is marshalled to disk
at /var/lib/kubelet/config.yaml, and also uploaded to a kubelet-config ConfigMap in the kube-system
namespace of the cluster. A kubelet configuration file is also written to /etc/kubernetes/kubelet.conf
with the baseline cluster-wide configuration for all kubelets in the cluster. This configuration file
points to the client certificates that allow the kubelet to communicate with the API server. This
addresses the need to
propagate cluster-level configuration to each kubelet.
To address the second pattern of
providing instance-specific configuration details,
kubeadm writes an environment file to /var/lib/kubelet/kubeadm-flags.env, which contains a list of
flags to pass to the kubelet when it starts. The flags are presented in the file like this:
In addition to the flags used when starting the kubelet, the file also contains dynamic
parameters such as the cgroup driver and whether to use a different container runtime socket
(--cri-socket).
After marshalling these two files to disk, kubeadm attempts to run the following two
commands, if you are using systemd:
If the reload and restart are successful, the normal kubeadm init workflow continues.
Workflow when using kubeadm join
When you run kubeadm join, kubeadm uses the Bootstrap Token credential to perform
a TLS bootstrap, which fetches the credential needed to download the
kubelet-config ConfigMap and writes it to /var/lib/kubelet/config.yaml. The dynamic
environment file is generated in exactly the same way as kubeadm init.
Next, kubeadm runs the following two commands to load the new configuration into the kubelet:
After the kubelet loads the new configuration, kubeadm writes the
/etc/kubernetes/bootstrap-kubelet.conf KubeConfig file, which contains a CA certificate and Bootstrap
Token. These are used by the kubelet to perform the TLS Bootstrap and obtain a unique
credential, which is stored in /etc/kubernetes/kubelet.conf.
When the /etc/kubernetes/kubelet.conf file is written, the kubelet has finished performing the TLS Bootstrap.
Kubeadm deletes the /etc/kubernetes/bootstrap-kubelet.conf file after completing the TLS Bootstrap.
The kubelet drop-in file for systemd
kubeadm ships with configuration for how systemd should run the kubelet.
Note that the kubeadm CLI command never touches this drop-in file.
Note: The contents below are just an example. If you don't want to use a package manager
follow the guide outlined in the Without a package manager)
section.
[Service]
Environment="KUBELET_KUBECONFIG_ARGS=--bootstrap-kubeconfig=/etc/kubernetes/bootstrap-kubelet.conf --kubeconfig=/etc/kubernetes/kubelet.conf"
Environment="KUBELET_CONFIG_ARGS=--config=/var/lib/kubelet/config.yaml"
# This is a file that "kubeadm init" and "kubeadm join" generate at runtime, populating
the KUBELET_KUBEADM_ARGS variable dynamically
EnvironmentFile=-/var/lib/kubelet/kubeadm-flags.env
# This is a file that the user can use for overrides of the kubelet args as a last resort. Preferably,
# the user should use the .NodeRegistration.KubeletExtraArgs object in the configuration files instead.
# KUBELET_EXTRA_ARGS should be sourced from this file.
EnvironmentFile=-/etc/default/kubelet
ExecStart=
ExecStart=/usr/bin/kubelet $KUBELET_KUBECONFIG_ARGS $KUBELET_CONFIG_ARGS $KUBELET_KUBEADM_ARGS $KUBELET_EXTRA_ARGS
This file specifies the default locations for all of the files managed by kubeadm for the kubelet.
The KubeConfig file to use for the TLS Bootstrap is /etc/kubernetes/bootstrap-kubelet.conf,
but it is only used if /etc/kubernetes/kubelet.conf does not exist.
The KubeConfig file with the unique kubelet identity is /etc/kubernetes/kubelet.conf.
The file containing the kubelet's ComponentConfig is /var/lib/kubelet/config.yaml.
The dynamic environment file that contains KUBELET_KUBEADM_ARGS is sourced from /var/lib/kubelet/kubeadm-flags.env.
The file that can contain user-specified flag overrides with KUBELET_EXTRA_ARGS is sourced from
/etc/default/kubelet (for DEBs), or /etc/sysconfig/kubelet (for RPMs). KUBELET_EXTRA_ARGS
is last in the flag chain and has the highest priority in the event of conflicting settings.
Kubernetes binaries and package contents
The DEB and RPM packages shipped with the Kubernetes releases are:
Package name
Description
kubeadm
Installs the /usr/bin/kubeadm CLI tool and the kubelet drop-in file for the kubelet.
Your Kubernetes cluster includes dual-stack networking, which means that cluster networking lets you use either address family. In a cluster, the control plane can assign both an IPv4 address and an IPv6 address to a single Pod or a Service.
For each server that you want to use as a node, make sure it allows IPv6 forwarding. On Linux, you can set this by running run sysctl -w net.ipv6.conf.all.forwarding=1 as the root user on each server.
You need to have an IPv4 and and IPv6 address range to use. Cluster operators typically
use private address ranges for IPv4. For IPv6, a cluster operator typically chooses a global
unicast address block from within 2000::/3, using a range that is assigned to the operator.
You don't have to route the cluster's IP address ranges to the public internet.
The size of the IP address allocations should be suitable for the number of Pods and
Services that you are planning to run.
Note: If you are upgrading an existing cluster with the kubeadm upgrade command,
kubeadm does not support making modifications to the pod IP address range
(“cluster CIDR”) nor to the cluster's Service address range (“Service CIDR”).
Create a dual-stack cluster
To create a dual-stack cluster with kubeadm init you can pass command line arguments
similar to the following example:
# These address ranges are exampleskubeadm init --pod-network-cidr=10.244.0.0/16,2001:db8:42:0::/56 --service-cidr=10.96.0.0/16,2001:db8:42:1::/112
To make things clearer, here is an example kubeadm
configuration filekubeadm-config.yaml for the primary dual-stack control plane node.
advertiseAddress in InitConfiguration specifies the IP address that the API Server will advertise it is listening on. The value of advertiseAddress equals the --apiserver-advertise-address flag of kubeadm init
Run kubeadm to initiate the dual-stack control plane node:
kubeadm init --config=kubeadm-config.yaml
The kube-controller-manager flags --node-cidr-mask-size-ipv4|--node-cidr-mask-size-ipv6 are set with default values. See configure IPv4/IPv6 dual stack.
Note: The --apiserver-advertise-address flag does not support dual-stack.
Join a node to dual-stack cluster
Before joining a node, make sure that the node has IPv6 routable network interface and allows IPv6 forwarding.
Here is an example kubeadm configuration filekubeadm-config.yaml for joining a worker node to the cluster.
apiVersion:kubeadm.k8s.io/v1beta3kind:JoinConfigurationdiscovery:bootstrapToken:apiServerEndpoint:10.100.0.1:6443token:"clvldh.vjjwg16ucnhp94qr"caCertHashes:- "sha256:a4863cde706cfc580a439f842cc65d5ef112b7b2be31628513a9881cf0d9fe0e"# change auth info above to match the actual token and CA certificate hash for your clusternodeRegistration:kubeletExtraArgs:node-ip:10.100.0.3,fd00:1:2:3::3
Also, here is an example kubeadm configuration filekubeadm-config.yaml for joining another control plane node to the cluster.
apiVersion:kubeadm.k8s.io/v1beta3kind:JoinConfigurationcontrolPlane:localAPIEndpoint:advertiseAddress:"10.100.0.2"bindPort:6443discovery:bootstrapToken:apiServerEndpoint:10.100.0.1:6443token:"clvldh.vjjwg16ucnhp94qr"caCertHashes:- "sha256:a4863cde706cfc580a439f842cc65d5ef112b7b2be31628513a9881cf0d9fe0e"# change auth info above to match the actual token and CA certificate hash for your clusternodeRegistration:kubeletExtraArgs:node-ip:10.100.0.4,fd00:1:2:3::4
advertiseAddress in JoinConfiguration.controlPlane specifies the IP address that the API Server will advertise it is listening on. The value of advertiseAddress equals the --apiserver-advertise-address flag of kubeadm join.
kubeadm join --config=kubeadm-config.yaml
Create a single-stack cluster
Note: Dual-stack support doesn't mean that you need to use dual-stack addressing.
You can deploy a single-stack cluster that has the dual-stack networking feature enabled.
To make things more clear, here is an example kubeadm
configuration filekubeadm-config.yaml for the single-stack control plane node.
kops uses DNS for discovery, both inside the cluster and outside, so that you can reach the kubernetes API server
from clients.
kops has a strong opinion on the cluster name: it should be a valid DNS name. By doing so you will
no longer get your clusters confused, you can share clusters with your colleagues unambiguously,
and you can reach them without relying on remembering an IP address.
You can, and probably should, use subdomains to divide your clusters. As our example we will use
useast1.dev.example.com. The API server endpoint will then be api.useast1.dev.example.com.
A Route53 hosted zone can serve subdomains. Your hosted zone could be useast1.dev.example.com,
but also dev.example.com or even example.com. kops works with any of these, so typically
you choose for organization reasons (e.g. you are allowed to create records under dev.example.com,
but not under example.com).
Let's assume you're using dev.example.com as your hosted zone. You create that hosted zone using
the normal process, or
with a command such as aws route53 create-hosted-zone --name dev.example.com --caller-reference 1.
You must then set up your NS records in the parent domain, so that records in the domain will resolve. Here,
you would create NS records in example.com for dev. If it is a root domain name you would configure the NS
records at your domain registrar (e.g. example.com would need to be configured where you bought example.com).
Verify your route53 domain setup (it is the #1 cause of problems!). You can double-check that
your cluster is configured correctly if you have the dig tool by running:
dig NS dev.example.com
You should see the 4 NS records that Route53 assigned your hosted zone.
(3/5) Create an S3 bucket to store your clusters state
kops lets you manage your clusters even after installation. To do this, it must keep track of the clusters
that you have created, along with their configuration, the keys they are using etc. This information is stored
in an S3 bucket. S3 permissions are used to control access to the bucket.
Multiple clusters can use the same S3 bucket, and you can share an S3 bucket between your colleagues that
administer the same clusters - this is much easier than passing around kubecfg files. But anyone with access
to the S3 bucket will have administrative access to all your clusters, so you don't want to share it beyond
the operations team.
So typically you have one S3 bucket for each ops team (and often the name will correspond
to the name of the hosted zone above!)
In our example, we chose dev.example.com as our hosted zone, so let's pick clusters.dev.example.com as
the S3 bucket name.
Export AWS_PROFILE (if you need to select a profile for the AWS CLI to work)
Create the S3 bucket using aws s3 mb s3://clusters.dev.example.com
You can export KOPS_STATE_STORE=s3://clusters.dev.example.com and then kops will use this location by default.
We suggest putting this in your bash profile or similar.
(4/5) Build your cluster configuration
Run kops create cluster to create your cluster configuration:
kops will create the configuration for your cluster. Note that it only creates the configuration, it does
not actually create the cloud resources - you'll do that in the next step with a kops update cluster. This
give you an opportunity to review the configuration or change it.
It prints commands you can use to explore further:
List your clusters with: kops get cluster
Edit this cluster with: kops edit cluster useast1.dev.example.com
Edit your node instance group: kops edit ig --name=useast1.dev.example.com nodes
Edit your master instance group: kops edit ig --name=useast1.dev.example.com master-us-east-1c
If this is your first time using kops, do spend a few minutes to try those out! An instance group is a
set of instances, which will be registered as kubernetes nodes. On AWS this is implemented via auto-scaling-groups.
You can have several instance groups, for example if you wanted nodes that are a mix of spot and on-demand instances, or
GPU and non-GPU instances.
(5/5) Create the cluster in AWS
Run "kops update cluster" to create your cluster in AWS:
kops update cluster useast1.dev.example.com --yes
That takes a few seconds to run, but then your cluster will likely take a few minutes to actually be ready.
kops update cluster will be the tool you'll use whenever you change the configuration of your cluster; it
applies the changes you have made to the configuration to your cluster - reconfiguring AWS or kubernetes as needed.
For example, after you kops edit ig nodes, then kops update cluster --yes to apply your configuration, and
sometimes you will also have to kops rolling-update cluster to roll out the configuration immediately.
Without --yes, kops update cluster will show you a preview of what it is going to do. This is handy
for production clusters!
Explore other add-ons
See the list of add-ons to explore other add-ons, including tools for logging, monitoring, network policy, visualization, and control of your Kubernetes cluster.
Cleanup
To delete your cluster: kops delete cluster useast1.dev.example.com --yes
Contribute to kops by addressing or raising an issue GitHub Issues
2.2.2.3 - Installing Kubernetes with Kubespray
This quickstart helps to install a Kubernetes cluster hosted on GCE, Azure, OpenStack, AWS, vSphere, Equinix Metal (formerly Packet), Oracle Cloud Infrastructure (Experimental) or Baremetal with Kubespray.
Kubespray is a composition of Ansible playbooks, inventory, provisioning tools, and domain knowledge for generic OS/Kubernetes clusters configuration management tasks.
Kubespray provides:
Highly available cluster.
Composable (Choice of the network plugin for instance).
Provision servers with the following requirements:
Minimum required version of Kubernetes is v1.22
Ansible v2.11+, Jinja 2.11+ and python-netaddr is installed on the machine that will run Ansible commands
The target servers must have access to the Internet in order to pull docker images. Otherwise, additional configuration is required See (Offline Environment)
The target servers are configured to allow IPv4 forwarding.
If using IPv6 for pods and services, the target servers are configured to allow IPv6 forwarding.
The firewalls are not managed, you'll need to implement your own rules the way you used to. in order to avoid any issue during deployment you should disable your firewall.
If kubespray is run from non-root user account, correct privilege escalation method should be configured in the target servers. Then the ansible_become flag or command parameters --become or -b should be specified.
Kubespray provides the following utilities to help provision your environment:
Terraform scripts for the following cloud providers:
Kubespray customizations can be made to a variable file. If you are getting started with Kubespray, consider using the Kubespray defaults to deploy your cluster and explore Kubernetes.
Large deployments (100+ nodes) may require specific adjustments for best results.
(5/5) Verify the deployment
Kubespray provides a way to verify inter-pod connectivity and DNS resolve with Netchecker. Netchecker ensures the netchecker-agents pods can resolve DNS requests and ping each over within the default namespace. Those pods mimic similar behavior as the rest of the workloads and serve as cluster health indicators.
Cluster operations
Kubespray provides additional playbooks to manage your cluster: scale and upgrade.
Scale your cluster
You can add worker nodes from your cluster by running the scale playbook. For more information, see "Adding nodes".
You can remove worker nodes from your cluster by running the remove-node playbook. For more information, see "Remove nodes".
Upgrade your cluster
You can upgrade your cluster by running the upgrade-cluster playbook. For more information, see "Upgrades".
Cleanup
You can reset your nodes and wipe out all components installed with Kubespray via the reset playbook.
Caution: When running the reset playbook, be sure not to accidentally target your production cluster!
Feedback
Slack Channel: #kubespray (You can get your invite here).
This page provides a list of Kubernetes certified solution providers. From each
provider page, you can learn how to install and setup production
ready clusters.
2.3 - Best practices
2.3.1 - Considerations for large clusters
A cluster is a set of nodes (physical
or virtual machines) running Kubernetes agents, managed by the
control plane.
Kubernetes v1.25 supports clusters with up to 5000 nodes. More specifically,
Kubernetes is designed to accommodate configurations that meet all of the following criteria:
No more than 110 pods per node
No more than 5000 nodes
No more than 150000 total pods
No more than 300000 total containers
You can scale your cluster by adding or removing nodes. The way you do this depends
on how your cluster is deployed.
Cloud provider resource quotas
To avoid running into cloud provider quota issues, when creating a cluster with many nodes,
consider:
Requesting a quota increase for cloud resources such as:
Computer instances
CPUs
Storage volumes
In-use IP addresses
Packet filtering rule sets
Number of load balancers
Network subnets
Log streams
Gating the cluster scaling actions to bring up new nodes in batches, with a pause
between batches, because some cloud providers rate limit the creation of new instances.
Control plane components
For a large cluster, you need a control plane with sufficient compute and other
resources.
Typically you would run one or two control plane instances per failure zone,
scaling those instances vertically first and then scaling horizontally after reaching
the point of falling returns to (vertical) scale.
You should run at least one instance per failure zone to provide fault-tolerance. Kubernetes
nodes do not automatically steer traffic towards control-plane endpoints that are in the
same failure zone; however, your cloud provider might have its own mechanisms to do this.
For example, using a managed load balancer, you configure the load balancer to send traffic
that originates from the kubelet and Pods in failure zone A, and direct that traffic only
to the control plane hosts that are also in zone A. If a single control-plane host or
endpoint failure zone A goes offline, that means that all the control-plane traffic for
nodes in zone A is now being sent between zones. Running multiple control plane hosts in
each zone makes that outcome less likely.
etcd storage
To improve performance of large clusters, you can store Event objects in a separate
dedicated etcd instance.
When creating a cluster, you can (using custom tooling):
start and configure additional etcd instance
configure the API server to use it for storing events
Kubernetes resource limits
help to minimize the impact of memory leaks and other ways that pods and containers can
impact on other components. These resource limits apply to
addon resources just as they apply to application workloads.
For example, you can set CPU and memory limits for a logging component:
Addons' default limits are typically based on data collected from experience running
each addon on small or medium Kubernetes clusters. When running on large
clusters, addons often consume more of some resources than their default limits.
If a large cluster is deployed without adjusting these values, the addon(s)
may continuously get killed because they keep hitting the memory limit.
Alternatively, the addon may run but with poor performance due to CPU time
slice restrictions.
To avoid running into cluster addon resource issues, when creating a cluster with
many nodes, consider the following:
Some addons scale vertically - there is one replica of the addon for the cluster
or serving a whole failure zone. For these addons, increase requests and limits
as you scale out your cluster.
Many addons scale horizontally - you add capacity by running more pods - but with
a very large cluster you may also need to raise CPU or memory limits slightly.
The VerticalPodAutoscaler can run in recommender mode to provide suggested
figures for requests and limits.
Some addons run as one copy per node, controlled by a DaemonSet: for example, a node-level log aggregator. Similar to
the case with horizontally-scaled addons, you may also need to raise CPU or memory
limits slightly.
What's next
VerticalPodAutoscaler is a custom resource that you can deploy into your cluster
to help you manage resource requests and limits for pods.
Visit Vertical Pod Autoscaler
to learn more about VerticalPodAutoscaler and how you can use it to scale cluster
components, including cluster-critical addons.
The cluster autoscaler
integrates with a number of cloud providers to help you run the right number of
nodes for the level of resource demand in your cluster.
The addon resizer
helps you in resizing the addons automatically as your cluster's scale changes.
2.3.2 - Running in multiple zones
This page describes running Kubernetes across multiple zones.
Background
Kubernetes is designed so that a single Kubernetes cluster can run
across multiple failure zones, typically where these zones fit within
a logical grouping called a region. Major cloud providers define a region
as a set of failure zones (also called availability zones) that provide
a consistent set of features: within a region, each zone offers the same
APIs and services.
Typical cloud architectures aim to minimize the chance that a failure in
one zone also impairs services in another zone.
Control plane behavior
All control plane components
support running as a pool of interchangeable resources, replicated per
component.
When you deploy a cluster control plane, place replicas of
control plane components across multiple failure zones. If availability is
an important concern, select at least three failure zones and replicate
each individual control plane component (API server, scheduler, etcd,
cluster controller manager) across at least three failure zones.
If you are running a cloud controller manager then you should
also replicate this across all the failure zones you selected.
Note: Kubernetes does not provide cross-zone resilience for the API server
endpoints. You can use various techniques to improve availability for
the cluster API server, including DNS round-robin, SRV records, or
a third-party load balancing solution with health checking.
Node behavior
Kubernetes automatically spreads the Pods for
workload resources (such as Deployment
or StatefulSet)
across different nodes in a cluster. This spreading helps
reduce the impact of failures.
When nodes start up, the kubelet on each node automatically adds
labels to the Node object
that represents that specific kubelet in the Kubernetes API.
These labels can include
zone information.
If your cluster spans multiple zones or regions, you can use node labels
in conjunction with
Pod topology spread constraints
to control how Pods are spread across your cluster among fault domains:
regions, zones, and even specific nodes.
These hints enable the
scheduler to place
Pods for better expected availability, reducing the risk that a correlated
failure affects your whole workload.
For example, you can set a constraint to make sure that the
3 replicas of a StatefulSet are all running in different zones to each
other, whenever that is feasible. You can define this declaratively
without explicitly defining which availability zones are in use for
each workload.
Distributing nodes across zones
Kubernetes' core does not create nodes for you; you need to do that yourself,
or use a tool such as the Cluster API to
manage nodes on your behalf.
Using tools such as the Cluster API you can define sets of machines to run as
worker nodes for your cluster across multiple failure domains, and rules to
automatically heal the cluster in case of whole-zone service disruption.
Manual zone assignment for Pods
You can apply node selector constraints
to Pods that you create, as well as to Pod templates in workload resources
such as Deployment, StatefulSet, or Job.
Storage access for zones
When persistent volumes are created, the PersistentVolumeLabeladmission controller
automatically adds zone labels to any PersistentVolumes that are linked to a specific
zone. The scheduler then ensures,
through its NoVolumeZoneConflict predicate, that pods which claim a given PersistentVolume
are only placed into the same zone as that volume.
You can specify a StorageClass
for PersistentVolumeClaims that specifies the failure domains (zones) that the
storage in that class may use.
To learn about configuring a StorageClass that is aware of failure domains or zones,
see Allowed topologies.
Networking
By itself, Kubernetes does not include zone-aware networking. You can use a
network plugin
to configure cluster networking, and that network solution might have zone-specific
elements. For example, if your cloud provider supports Services with
type=LoadBalancer, the load balancer might only send traffic to Pods running in the
same zone as the load balancer element processing a given connection.
Check your cloud provider's documentation for details.
For custom or on-premises deployments, similar considerations apply.
Service and
Ingress behavior, including handling
of different failure zones, does vary depending on exactly how your cluster is set up.
Fault recovery
When you set up your cluster, you might also need to consider whether and how
your setup can restore service if all the failure zones in a region go
off-line at the same time. For example, do you rely on there being at least
one node able to run Pods in a zone?
Make sure that any cluster-critical repair work does not rely
on there being at least one healthy node in your cluster. For example: if all nodes
are unhealthy, you might need to run a repair Job with a special
toleration so that the repair
can complete enough to bring at least one node into service.
Kubernetes doesn't come with an answer for this challenge; however, it's
something to consider.
What's next
To learn how the scheduler places Pods in a cluster, honoring the configured constraints,
visit Scheduling and Eviction.
2.3.3 - Validate node setup
Node Conformance Test
Node conformance test is a containerized test framework that provides a system
verification and functionality test for a node. The test validates whether the
node meets the minimum requirements for Kubernetes; a node that passes the test
is qualified to join a Kubernetes cluster.
Node Prerequisite
To run node conformance test, a node must satisfy the same prerequisites as a
standard Kubernetes node. At a minimum, the node should have the following
daemons installed:
Container Runtime (Docker)
Kubelet
Running Node Conformance Test
To run the node conformance test, perform the following steps:
Work out the value of the --kubeconfig option for the kubelet; for example:
--kubeconfig=/var/lib/kubelet/config.yaml.
Because the test framework starts a local control plane to test the kubelet,
use http://localhost:8080 as the URL of the API server.
There are some other kubelet command line parameters you may want to use:
--cloud-provider: If you are using --cloud-provider=gce, you should
remove the flag to run the test.
Run the node conformance test with command:
# $CONFIG_DIR is the pod manifest path of your Kubelet.# $LOG_DIR is the test output path.sudo docker run -it --rm --privileged --net=host \
-v /:/rootfs -v $CONFIG_DIR:$CONFIG_DIR -v $LOG_DIR:/var/result \
registry.k8s.io/node-test:0.2
Running Node Conformance Test for Other Architectures
Kubernetes also provides node conformance test docker images for other
architectures:
Arch
Image
amd64
node-test-amd64
arm
node-test-arm
arm64
node-test-arm64
Running Selected Test
To run specific tests, overwrite the environment variable FOCUS with the
regular expression of tests you want to run.
sudo docker run -it --rm --privileged --net=host \
-v /:/rootfs:ro -v $CONFIG_DIR:$CONFIG_DIR -v $LOG_DIR:/var/result \
-e FOCUS=MirrorPod \ # Only run MirrorPod test registry.k8s.io/node-test:0.2
To skip specific tests, overwrite the environment variable SKIP with the
regular expression of tests you want to skip.
sudo docker run -it --rm --privileged --net=host \
-v /:/rootfs:ro -v $CONFIG_DIR:$CONFIG_DIR -v $LOG_DIR:/var/result \
-e SKIP=MirrorPod \ # Run all conformance tests but skip MirrorPod test registry.k8s.io/node-test:0.2
Node conformance test is a containerized version of node e2e test.
By default, it runs all conformance tests.
Theoretically, you can run any node e2e test if you configure the container and
mount required volumes properly. But it is strongly recommended to only run conformance
test, because it requires much more complex configuration to run non-conformance test.
Caveats
The test leaves some docker images on the node, including the node conformance
test image and images of containers used in the functionality
test.
The test leaves dead containers on the node. These containers are created
during the functionality test.
2.3.4 - Enforcing Pod Security Standards
This page provides an overview of best practices when it comes to enforcing
Pod Security Standards.
Using the built-in Pod Security Admission Controller
Namespaces that lack any configuration at all should be considered significant gaps in your cluster
security model. We recommend taking the time to analyze the types of workloads occurring in each
namespace, and by referencing the Pod Security Standards, decide on an appropriate level for
each of them. Unlabeled namespaces should only indicate that they've yet to be evaluated.
In the scenario that all workloads in all namespaces have the same security requirements,
we provide an example
that illustrates how the PodSecurity labels can be applied in bulk.
Embrace the principle of least privilege
In an ideal world, every pod in every namespace would meet the requirements of the restricted
policy. However, this is not possible nor practical, as some workloads will require elevated
privileges for legitimate reasons.
Namespaces allowing privileged workloads should establish and enforce appropriate access controls.
For workloads running in those permissive namespaces, maintain documentation about their unique
security requirements. If at all possible, consider how those requirements could be further
constrained.
Adopt a multi-mode strategy
The audit and warn modes of the Pod Security Standards admission controller make it easy to
collect important security insights about your pods without breaking existing workloads.
It is good practice to enable these modes for all namespaces, setting them to the desired level
and version you would eventually like to enforce. The warnings and audit annotations generated in
this phase can guide you toward that state. If you expect workload authors to make changes to fit
within the desired level, enable the warn mode. If you expect to use audit logs to monitor/drive
changes to fit within the desired level, enable the audit mode.
When you have the enforce mode set to your desired value, these modes can still be useful in a
few different ways:
By setting warn to the same level as enforce, clients will receive warnings when attempting
to create Pods (or resources that have Pod templates) that do not pass validation. This will help
them update those resources to become compliant.
In Namespaces that pin enforce to a specific non-latest version, setting the audit and warn
modes to the same level as enforce, but to the latest version, gives visibility into settings
that were allowed by previous versions but are not allowed per current best practices.
Third-party alternatives
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Other alternatives for enforcing security profiles are being developed in the Kubernetes
ecosystem:
The decision to go with a built-in solution (e.g. PodSecurity admission controller) versus a
third-party tool is entirely dependent on your own situation. When evaluating any solution,
trust of your supply chain is crucial. Ultimately, using any of the aforementioned approaches
will be better than doing nothing.
2.3.5 - PKI certificates and requirements
Kubernetes requires PKI certificates for authentication over TLS.
If you install Kubernetes with kubeadm, the certificates that your cluster requires are automatically generated.
You can also generate your own certificates -- for example, to keep your private keys more secure by not storing them on the API server.
This page explains the certificates that your cluster requires.
How certificates are used by your cluster
Kubernetes requires PKI for the following operations:
Client certificates for the kubelet to authenticate to the API server
Client certificates for administrators of the cluster to authenticate to the API server
Client certificates for the API server to talk to the kubelets
Client certificate for the API server to talk to etcd
Client certificate/kubeconfig for the controller manager to talk to the API server
Client certificate/kubeconfig for the scheduler to talk to the API server.
Client and server certificates for the front-proxy
Note:front-proxy certificates are required only if you run kube-proxy to support an extension API server.
etcd also implements mutual TLS to authenticate clients and peers.
Where certificates are stored
If you install Kubernetes with kubeadm, most certificates are stored in /etc/kubernetes/pki. All paths in this documentation are relative to that directory, with the exception of user account certificates which kubeadm places in /etc/kubernetes.
Configure certificates manually
If you don't want kubeadm to generate the required certificates, you can create them using a single root CA or by providing all certificates. See Certificates for details on creating your own certificate authority.
See Certificate Management with kubeadm for more on managing certificates.
Single root CA
You can create a single root CA, controlled by an administrator. This root CA can then create multiple intermediate CAs, and delegate all further creation to Kubernetes itself.
On top of the above CAs, it is also necessary to get a public/private key pair for service account management, sa.key and sa.pub.
The following example illustrates the CA key and certificate files shown in the previous table:
If you don't wish to copy the CA private keys to your cluster, you can generate all certificates yourself.
Required certificates:
Default CN
Parent CA
O (in Subject)
kind
hosts (SAN)
kube-etcd
etcd-ca
server, client
<hostname>, <Host_IP>, localhost, 127.0.0.1
kube-etcd-peer
etcd-ca
server, client
<hostname>, <Host_IP>, localhost, 127.0.0.1
kube-etcd-healthcheck-client
etcd-ca
client
kube-apiserver-etcd-client
etcd-ca
system:masters
client
kube-apiserver
kubernetes-ca
server
<hostname>, <Host_IP>, <advertise_IP>, [1]
kube-apiserver-kubelet-client
kubernetes-ca
system:masters
client
front-proxy-client
kubernetes-front-proxy-ca
client
[1]: any other IP or DNS name you contact your cluster on (as used by kubeadm
the load balancer stable IP and/or DNS name, kubernetes, kubernetes.default, kubernetes.default.svc,
kubernetes.default.svc.cluster, kubernetes.default.svc.cluster.local)
where kind maps to one or more of the x509 key usage types:
kind
Key usage
server
digital signature, key encipherment, server auth
client
digital signature, key encipherment, client auth
Note: Hosts/SAN listed above are the recommended ones for getting a working cluster; if required by a specific setup, it is possible to add additional SANs on all the server certificates.
Note:
For kubeadm users only:
The scenario where you are copying to your cluster CA certificates without private keys is referred as external CA in the kubeadm documentation.
If you are comparing the above list with a kubeadm generated PKI, please be aware that kube-etcd, kube-etcd-peer and kube-etcd-healthcheck-client certificates
are not generated in case of external etcd.
Certificate paths
Certificates should be placed in a recommended path (as used by kubeadm).
Paths should be specified using the given argument regardless of location.
Same considerations apply for the service account key pair:
private key path
public key path
command
argument
sa.key
kube-controller-manager
--service-account-private-key-file
sa.pub
kube-apiserver
--service-account-key-file
The following example illustrates the file paths from the previous tables you need to provide if you are generating all of your own keys and certificates:
You must manually configure these administrator account and service accounts:
filename
credential name
Default CN
O (in Subject)
admin.conf
default-admin
kubernetes-admin
system:masters
kubelet.conf
default-auth
system:node:<nodeName> (see note)
system:nodes
controller-manager.conf
default-controller-manager
system:kube-controller-manager
scheduler.conf
default-scheduler
system:kube-scheduler
Note: The value of <nodeName> for kubelet.confmust match precisely the value of the node name provided by the kubelet as it registers with the apiserver. For further details, read the Node Authorization.
For each config, generate an x509 cert/key pair with the given CN and O.
The Concepts section helps you learn about the parts of the Kubernetes system and the abstractions Kubernetes uses to represent your cluster, and helps you obtain a deeper understanding of how Kubernetes works.
3.1 - Overview
Get a high-level outline of Kubernetes and the components it is built from.
3.1.1 - What is Kubernetes?
Kubernetes is a portable, extensible, open source platform for managing containerized workloads and services, that facilitates both declarative configuration and automation. It has a large, rapidly growing ecosystem. Kubernetes services, support, and tools are widely available.
This page is an overview of Kubernetes.
Kubernetes is a portable, extensible, open source platform for managing containerized workloads and services, that facilitates both declarative configuration and automation. It has a large, rapidly growing ecosystem. Kubernetes services, support, and tools are widely available.
The name Kubernetes originates from Greek, meaning helmsman or pilot. K8s as an abbreviation results from counting the eight letters between the "K" and the "s". Google open-sourced the Kubernetes project in 2014. Kubernetes combines over 15 years of Google's experience running production workloads at scale with best-of-breed ideas and practices from the community.
Going back in time
Let's take a look at why Kubernetes is so useful by going back in time.
Traditional deployment era:
Early on, organizations ran applications on physical servers. There was no way to define resource boundaries for applications in a physical server, and this caused resource allocation issues. For example, if multiple applications run on a physical server, there can be instances where one application would take up most of the resources, and as a result, the other applications would underperform. A solution for this would be to run each application on a different physical server. But this did not scale as resources were underutilized, and it was expensive for organizations to maintain many physical servers.
Virtualized deployment era: As a solution, virtualization was introduced. It allows you to run multiple Virtual Machines (VMs) on a single physical server's CPU. Virtualization allows applications to be isolated between VMs and provides a level of security as the information of one application cannot be freely accessed by another application.
Virtualization allows better utilization of resources in a physical server and allows better scalability because an application can be added or updated easily, reduces hardware costs, and much more. With virtualization you can present a set of physical resources as a cluster of disposable virtual machines.
Each VM is a full machine running all the components, including its own operating system, on top of the virtualized hardware.
Container deployment era: Containers are similar to VMs, but they have relaxed isolation properties to share the Operating System (OS) among the applications. Therefore, containers are considered lightweight. Similar to a VM, a container has its own filesystem, share of CPU, memory, process space, and more. As they are decoupled from the underlying infrastructure, they are portable across clouds and OS distributions.
Containers have become popular because they provide extra benefits, such as:
Agile application creation and deployment: increased ease and efficiency of container image creation compared to VM image use.
Continuous development, integration, and deployment: provides for reliable and frequent container image build and deployment with quick and efficient rollbacks (due to image immutability).
Dev and Ops separation of concerns: create application container images at build/release time rather than deployment time, thereby decoupling applications from infrastructure.
Observability: not only surfaces OS-level information and metrics, but also application health and other signals.
Environmental consistency across development, testing, and production: Runs the same on a laptop as it does in the cloud.
Cloud and OS distribution portability: Runs on Ubuntu, RHEL, CoreOS, on-premises, on major public clouds, and anywhere else.
Application-centric management: Raises the level of abstraction from running an OS on virtual hardware to running an application on an OS using logical resources.
Loosely coupled, distributed, elastic, liberated micro-services: applications are broken into smaller, independent pieces and can be deployed and managed dynamically – not a monolithic stack running on one big single-purpose machine.
Resource utilization: high efficiency and density.
Why you need Kubernetes and what it can do
Containers are a good way to bundle and run your applications. In a production environment, you need to manage the containers that run the applications and ensure that there is no downtime. For example, if a container goes down, another container needs to start. Wouldn't it be easier if this behavior was handled by a system?
That's how Kubernetes comes to the rescue! Kubernetes provides you with a framework to run distributed systems resiliently. It takes care of scaling and failover for your application, provides deployment patterns, and more. For example, Kubernetes can easily manage a canary deployment for your system.
Kubernetes provides you with:
Service discovery and load balancing
Kubernetes can expose a container using the DNS name or using their own IP address. If traffic to a container is high, Kubernetes is able to load balance and distribute the network traffic so that the deployment is stable.
Storage orchestration
Kubernetes allows you to automatically mount a storage system of your choice, such as local storages, public cloud providers, and more.
Automated rollouts and rollbacks
You can describe the desired state for your deployed containers using Kubernetes, and it can change the actual state to the desired state at a controlled rate. For example, you can automate Kubernetes to create new containers for your deployment, remove existing containers and adopt all their resources to the new container.
Automatic bin packing
You provide Kubernetes with a cluster of nodes that it can use to run containerized tasks. You tell Kubernetes how much CPU and memory (RAM) each container needs. Kubernetes can fit containers onto your nodes to make the best use of your resources.
Self-healing
Kubernetes restarts containers that fail, replaces containers, kills containers that don't respond to your user-defined health check, and doesn't advertise them to clients until they are ready to serve.
Secret and configuration management
Kubernetes lets you store and manage sensitive information, such as passwords, OAuth tokens, and SSH keys. You can deploy and update secrets and application configuration without rebuilding your container images, and without exposing secrets in your stack configuration.
What Kubernetes is not
Kubernetes is not a traditional, all-inclusive PaaS (Platform as a Service) system. Since Kubernetes operates at the container level rather than at the hardware level, it provides some generally applicable features common to PaaS offerings, such as deployment, scaling, load balancing, and lets users integrate their logging, monitoring, and alerting solutions. However, Kubernetes is not monolithic, and these default solutions are optional and pluggable. Kubernetes provides the building blocks for building developer platforms, but preserves user choice and flexibility where it is important.
Kubernetes:
Does not limit the types of applications supported. Kubernetes aims to support an extremely diverse variety of workloads, including stateless, stateful, and data-processing workloads. If an application can run in a container, it should run great on Kubernetes.
Does not deploy source code and does not build your application. Continuous Integration, Delivery, and Deployment (CI/CD) workflows are determined by organization cultures and preferences as well as technical requirements.
Does not provide application-level services, such as middleware (for example, message buses), data-processing frameworks (for example, Spark), databases (for example, MySQL), caches, nor cluster storage systems (for example, Ceph) as built-in services. Such components can run on Kubernetes, and/or can be accessed by applications running on Kubernetes through portable mechanisms, such as the Open Service Broker.
Does not dictate logging, monitoring, or alerting solutions. It provides some integrations as proof of concept, and mechanisms to collect and export metrics.
Does not provide nor mandate a configuration language/system (for example, Jsonnet). It provides a declarative API that may be targeted by arbitrary forms of declarative specifications.
Does not provide nor adopt any comprehensive machine configuration, maintenance, management, or self-healing systems.
Additionally, Kubernetes is not a mere orchestration system. In fact, it eliminates the need for orchestration. The technical definition of orchestration is execution of a defined workflow: first do A, then B, then C. In contrast, Kubernetes comprises a set of independent, composable control processes that continuously drive the current state towards the provided desired state. It shouldn't matter how you get from A to C. Centralized control is also not required. This results in a system that is easier to use and more powerful, robust, resilient, and extensible.
A Kubernetes cluster consists of the components that represent the control plane and a set of machines called nodes.
When you deploy Kubernetes, you get a cluster.
A Kubernetes cluster consists of a set of worker machines, called nodes,
that run containerized applications. Every cluster has at least one worker node.
The worker node(s) host the Pods that are
the components of the application workload. The
control plane manages the worker
nodes and the Pods in the cluster. In production environments, the control plane usually
runs across multiple computers and a cluster usually runs multiple nodes, providing
fault-tolerance and high availability.
This document outlines the various components you need to have for
a complete and working Kubernetes cluster.
The components of a Kubernetes cluster
Control Plane Components
The control plane's components make global decisions about the cluster (for example, scheduling), as well as detecting and responding to cluster events (for example, starting up a new pod when a deployment's replicas field is unsatisfied).
Control plane components can be run on any machine in the cluster. However,
for simplicity, set up scripts typically start all control plane components on
the same machine, and do not run user containers on this machine. See
Creating Highly Available clusters with kubeadm
for an example control plane setup that runs across multiple machines.
kube-apiserver
The API server is a component of the Kubernetes
control plane that exposes the Kubernetes API.
The API server is the front end for the Kubernetes control plane.
The main implementation of a Kubernetes API server is kube-apiserver.
kube-apiserver is designed to scale horizontally—that is, it scales by deploying more instances.
You can run several instances of kube-apiserver and balance traffic between those instances.
etcd
Consistent and highly-available key value store used as Kubernetes' backing store for all cluster data.
If your Kubernetes cluster uses etcd as its backing store, make sure you have a
back up plan
for those data.
You can find in-depth information about etcd in the official documentation.
kube-scheduler
Control plane component that watches for newly created
Pods with no assigned
node, and selects a node for them
to run on.
Factors taken into account for scheduling decisions include:
individual and collective resource requirements, hardware/software/policy
constraints, affinity and anti-affinity specifications, data locality,
inter-workload interference, and deadlines.
kube-controller-manager
Control plane component that runs controller processes.
Logically, each controller is a separate process, but to reduce complexity, they are all compiled into a single binary and run in a single process.
Some types of these controllers are:
Node controller: Responsible for noticing and responding when nodes go down.
Job controller: Watches for Job objects that represent one-off tasks, then creates
Pods to run those tasks to completion.
Endpoints controller: Populates the Endpoints object (that is, joins Services & Pods).
Service Account & Token controllers: Create default accounts and API access tokens for new namespaces.
cloud-controller-manager
A Kubernetes control plane component
that embeds cloud-specific control logic. The cloud controller manager lets you link your
cluster into your cloud provider's API, and separates out the components that interact
with that cloud platform from components that only interact with your cluster.
The cloud-controller-manager only runs controllers that are specific to your cloud provider.
If you are running Kubernetes on your own premises, or in a learning environment inside your
own PC, the cluster does not have a cloud controller manager.
As with the kube-controller-manager, the cloud-controller-manager combines several logically
independent control loops into a single binary that you run as a single process. You can
scale horizontally (run more than one copy) to improve performance or to help tolerate failures.
The following controllers can have cloud provider dependencies:
Node controller: For checking the cloud provider to determine if a node has been deleted in the cloud after it stops responding
Route controller: For setting up routes in the underlying cloud infrastructure
Service controller: For creating, updating and deleting cloud provider load balancers
Node Components
Node components run on every node, maintaining running pods and providing the Kubernetes runtime environment.
kubelet
An agent that runs on each node in the cluster. It makes sure that containers are running in a Pod.
The kubelet takes a set of PodSpecs that are provided through various mechanisms and ensures that the containers described in those PodSpecs are running and healthy. The kubelet doesn't manage containers which were not created by Kubernetes.
kube-proxy
kube-proxy is a network proxy that runs on each
node in your cluster,
implementing part of the Kubernetes
Service concept.
kube-proxy
maintains network rules on nodes. These network rules allow network
communication to your Pods from network sessions inside or outside of
your cluster.
kube-proxy uses the operating system packet filtering layer if there is one
and it's available. Otherwise, kube-proxy forwards the traffic itself.
Container runtime
The container runtime is the software that is responsible for running containers.
Addons use Kubernetes resources (DaemonSet,
Deployment, etc)
to implement cluster features. Because these are providing cluster-level features, namespaced resources
for addons belong within the kube-system namespace.
Selected addons are described below; for an extended list of available addons, please
see Addons.
DNS
While the other addons are not strictly required, all Kubernetes clusters should have cluster DNS, as many examples rely on it.
Cluster DNS is a DNS server, in addition to the other DNS server(s) in your environment, which serves DNS records for Kubernetes services.
Containers started by Kubernetes automatically include this DNS server in their DNS searches.
Web UI (Dashboard)
Dashboard is a general purpose, web-based UI for Kubernetes clusters. It allows users to manage and troubleshoot applications running in the cluster, as well as the cluster itself.
Container Resource Monitoring
Container Resource Monitoring records generic time-series metrics
about containers in a central database, and provides a UI for browsing that data.
Cluster-level Logging
A cluster-level logging mechanism is responsible for
saving container logs to a central log store with search/browsing interface.
The Kubernetes API lets you query and manipulate the state of objects in Kubernetes. The core of Kubernetes' control plane is the API server and the HTTP API that it exposes. Users, the different parts of your cluster, and external components all communicate with one another through the API server.
The core of Kubernetes' control plane
is the API server. The API server
exposes an HTTP API that lets end users, different parts of your cluster, and
external components communicate with one another.
The Kubernetes API lets you query and manipulate the state of API objects in Kubernetes
(for example: Pods, Namespaces, ConfigMaps, and Events).
Most operations can be performed through the
kubectl command-line interface or other
command-line tools, such as
kubeadm, which in turn use the
API. However, you can also access the API directly using REST calls.
Consider using one of the client libraries
if you are writing an application using the Kubernetes API.
OpenAPI specification
Complete API details are documented using OpenAPI.
OpenAPI V2
The Kubernetes API server serves an aggregated OpenAPI v2 spec via the
/openapi/v2 endpoint. You can request the response format using
request headers as follows:
Valid request header values for OpenAPI v2 queries
Kubernetes implements an alternative Protobuf based serialization format that
is primarily intended for intra-cluster communication. For more information
about this format, see the Kubernetes Protobuf serialization design proposal and the
Interface Definition Language (IDL) files for each schema located in the Go
packages that define the API objects.
OpenAPI V3
FEATURE STATE:Kubernetes v1.24 [beta]
Kubernetes v1.25 offers beta support for publishing its APIs as OpenAPI v3; this is a
beta feature that is enabled by default.
You can disable the beta feature by turning off the
feature gate named OpenAPIV3
for the kube-apiserver component.
A discovery endpoint /openapi/v3 is provided to see a list of all
group/versions available. This endpoint only returns JSON. These group/versions
are provided in the following format:
The relative URLs are pointing to immutable OpenAPI descriptions, in
order to improve client-side caching. The proper HTTP caching headers
are also set by the API server for that purpose (Expires to 1 year in
the future, and Cache-Control to immutable). When an obsolete URL is
used, the API server returns a redirect to the newest URL.
The Kubernetes API server publishes an OpenAPI v3 spec per Kubernetes
group version at the /openapi/v3/apis/<group>/<version>?hash=<hash>
endpoint.
Refer to the table below for accepted request headers.
Valid request header values for OpenAPI v3 queries
Kubernetes stores the serialized state of objects by writing them into
etcd.
API groups and versioning
To make it easier to eliminate fields or restructure resource representations,
Kubernetes supports multiple API versions, each at a different API path, such
as /api/v1 or /apis/rbac.authorization.k8s.io/v1alpha1.
Versioning is done at the API level rather than at the resource or field level
to ensure that the API presents a clear, consistent view of system resources
and behavior, and to enable controlling access to end-of-life and/or
experimental APIs.
To make it easier to evolve and to extend its API, Kubernetes implements
API groups that can be
enabled or disabled.
API resources are distinguished by their API group, resource type, namespace
(for namespaced resources), and name. The API server handles the conversion between
API versions transparently: all the different versions are actually representations
of the same persisted data. The API server may serve the same underlying data
through multiple API versions.
For example, suppose there are two API versions, v1 and v1beta1, for the same
resource. If you originally created an object using the v1beta1 version of its
API, you can later read, update, or delete that object
using either the v1beta1 or the v1 API version.
API changes
Any system that is successful needs to grow and change as new use cases emerge or existing ones change.
Therefore, Kubernetes has designed the Kubernetes API to continuously change and grow.
The Kubernetes project aims to not break compatibility with existing clients, and to maintain that
compatibility for a length of time so that other projects have an opportunity to adapt.
In general, new API resources and new resource fields can be added often and frequently.
Elimination of resources or fields requires following the
API deprecation policy.
Kubernetes makes a strong commitment to maintain compatibility for official Kubernetes APIs
once they reach general availability (GA), typically at API version v1. Additionally,
Kubernetes keeps compatibility even for beta API versions wherever feasible:
if you adopt a beta API you can continue to interact with your cluster using that API,
even after the feature goes stable.
Note: Although Kubernetes also aims to maintain compatibility for alpha APIs versions, in some
circumstances this is not possible. If you use any alpha API versions, check the release notes
for Kubernetes when upgrading your cluster, in case the API did change.
Learn about API endpoints, resource types and samples by reading
API Reference.
Learn about what constitutes a compatible change, and how to change the API, from
API changes.
3.1.4 - Working with Kubernetes Objects
Kubernetes objects are persistent entities in the Kubernetes system. Kubernetes uses these entities to represent the state of your cluster. Learn about the Kubernetes object model and how to work with these objects.
3.1.4.1 - Understanding Kubernetes Objects
This page explains how Kubernetes objects are represented in the Kubernetes API, and how you can
express them in .yaml format.
Understanding Kubernetes objects
Kubernetes objects are persistent entities in the Kubernetes system. Kubernetes uses these
entities to represent the state of your cluster. Specifically, they can describe:
What containerized applications are running (and on which nodes)
The resources available to those applications
The policies around how those applications behave, such as restart policies, upgrades, and fault-tolerance
A Kubernetes object is a "record of intent"--once you create the object, the Kubernetes system
will constantly work to ensure that object exists. By creating an object, you're effectively
telling the Kubernetes system what you want your cluster's workload to look like; this is your
cluster's desired state.
To work with Kubernetes objects--whether to create, modify, or delete them--you'll need to use the
Kubernetes API. When you use the kubectl command-line
interface, for example, the CLI makes the necessary Kubernetes API calls for you. You can also use
the Kubernetes API directly in your own programs using one of the
Client Libraries.
Object Spec and Status
Almost every Kubernetes object includes two nested object fields that govern
the object's configuration: the object spec and the object status.
For objects that have a spec, you have to set this when you create the object,
providing a description of the characteristics you want the resource to have:
its desired state.
The status describes the current state of the object, supplied and updated
by the Kubernetes system and its components. The Kubernetes
control plane continually
and actively manages every object's actual state to match the desired state you
supplied.
For example: in Kubernetes, a Deployment is an object that can represent an
application running on your cluster. When you create the Deployment, you
might set the Deployment spec to specify that you want three replicas of
the application to be running. The Kubernetes system reads the Deployment
spec and starts three instances of your desired application--updating
the status to match your spec. If any of those instances should fail
(a status change), the Kubernetes system responds to the difference
between spec and status by making a correction--in this case, starting
a replacement instance.
When you create an object in Kubernetes, you must provide the object spec that describes its
desired state, as well as some basic information about the object (such as a name). When you use
the Kubernetes API to create the object (either directly or via kubectl), that API request must
include that information as JSON in the request body. Most often, you provide the information to
kubectl in a .yaml file.kubectl converts the information to JSON when making the API
request.
Here's an example .yaml file that shows the required fields and object spec for a Kubernetes Deployment:
apiVersion:apps/v1kind:Deploymentmetadata:name:nginx-deploymentspec:selector:matchLabels:app:nginxreplicas:2# tells deployment to run 2 pods matching the templatetemplate:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80
One way to create a Deployment using a .yaml file like the one above is to use the
kubectl apply command
in the kubectl command-line interface, passing the .yaml file as an argument. Here's an example:
In the .yaml file for the Kubernetes object you want to create, you'll need to set values for the following fields:
apiVersion - Which version of the Kubernetes API you're using to create this object
kind - What kind of object you want to create
metadata - Data that helps uniquely identify the object, including a name string, UID, and optional namespace
spec - What state you desire for the object
The precise format of the object spec is different for every Kubernetes object, and contains
nested fields specific to that object. The Kubernetes API Reference
can help you find the spec format for all of the objects you can create using Kubernetes.
For example, see the spec field
for the Pod API reference.
For each Pod, the .spec field specifies the pod and its desired state (such as the container image name for
each container within that pod).
Another example of an object specification is the
spec field
for the StatefulSet API. For StatefulSet, the .spec field specifies the StatefulSet and
its desired state.
Within the .spec of a StatefulSet is a template
for Pod objects. That template describes Pods that the StatefulSet controller will create in order to
satisfy the StatefulSet specification.
Different kinds of object can also have different .status; again, the API reference pages
detail the structure of that .status field, and its content for each different type of object.
What's next
Learn about the most important basic Kubernetes objects, such as Pod.
The kubectl command-line tool supports several different ways to create and manage
Kubernetes objects. This document provides an overview of the different
approaches. Read the Kubectl book for
details of managing objects by Kubectl.
Management techniques
Warning: A Kubernetes object should be managed using only one technique. Mixing
and matching techniques for the same object results in undefined behavior.
Management technique
Operates on
Recommended environment
Supported writers
Learning curve
Imperative commands
Live objects
Development projects
1+
Lowest
Imperative object configuration
Individual files
Production projects
1
Moderate
Declarative object configuration
Directories of files
Production projects
1+
Highest
Imperative commands
When using imperative commands, a user operates directly on live objects
in a cluster. The user provides operations to
the kubectl command as arguments or flags.
This is the recommended way to get started or to run a one-off task in
a cluster. Because this technique operates directly on live
objects, it provides no history of previous configurations.
Examples
Run an instance of the nginx container by creating a Deployment object:
kubectl create deployment nginx --image nginx
Trade-offs
Advantages compared to object configuration:
Commands are expressed as a single action word.
Commands require only a single step to make changes to the cluster.
Disadvantages compared to object configuration:
Commands do not integrate with change review processes.
Commands do not provide an audit trail associated with changes.
Commands do not provide a source of records except for what is live.
Commands do not provide a template for creating new objects.
Imperative object configuration
In imperative object configuration, the kubectl command specifies the
operation (create, replace, etc.), optional flags and at least one file
name. The file specified must contain a full definition of the object
in YAML or JSON format.
See the API reference
for more details on object definitions.
Warning: The imperative replace command replaces the existing
spec with the newly provided one, dropping all changes to the object missing from
the configuration file. This approach should not be used with resource
types whose specs are updated independently of the configuration file.
Services of type LoadBalancer, for example, have their externalIPs field updated
independently from the configuration by the cluster.
Examples
Create the objects defined in a configuration file:
kubectl create -f nginx.yaml
Delete the objects defined in two configuration files:
kubectl delete -f nginx.yaml -f redis.yaml
Update the objects defined in a configuration file by overwriting
the live configuration:
kubectl replace -f nginx.yaml
Trade-offs
Advantages compared to imperative commands:
Object configuration can be stored in a source control system such as Git.
Object configuration can integrate with processes such as reviewing changes before push and audit trails.
Object configuration provides a template for creating new objects.
Disadvantages compared to imperative commands:
Object configuration requires basic understanding of the object schema.
Object configuration requires the additional step of writing a YAML file.
Advantages compared to declarative object configuration:
Imperative object configuration behavior is simpler and easier to understand.
As of Kubernetes version 1.5, imperative object configuration is more mature.
Disadvantages compared to declarative object configuration:
Imperative object configuration works best on files, not directories.
Updates to live objects must be reflected in configuration files, or they will be lost during the next replacement.
Declarative object configuration
When using declarative object configuration, a user operates on object
configuration files stored locally, however the user does not define the
operations to be taken on the files. Create, update, and delete operations
are automatically detected per-object by kubectl. This enables working on
directories, where different operations might be needed for different objects.
Note: Declarative object configuration retains changes made by other
writers, even if the changes are not merged back to the object configuration file.
This is possible by using the patch API operation to write only
observed differences, instead of using the replace
API operation to replace the entire object configuration.
Examples
Process all object configuration files in the configs directory, and create or
patch the live objects. You can first diff to see what changes are going to be
made, and then apply:
Advantages compared to imperative object configuration:
Changes made directly to live objects are retained, even if they are not merged back into the configuration files.
Declarative object configuration has better support for operating on directories and automatically detecting operation types (create, patch, delete) per-object.
Disadvantages compared to imperative object configuration:
Declarative object configuration is harder to debug and understand results when they are unexpected.
Partial updates using diffs create complex merge and patch operations.
Each object in your cluster has a Name that is unique for that type of resource.
Every Kubernetes object also has a UID that is unique across your whole cluster.
For example, you can only have one Pod named myapp-1234 within the same namespace, but you can have one Pod and one Deployment that are each named myapp-1234.
For non-unique user-provided attributes, Kubernetes provides labels and annotations.
Names
A client-provided string that refers to an object in a resource URL, such as /api/v1/pods/some-name.
Only one object of a given kind can have a given name at a time. However, if you delete the object, you can make a new object with the same name.
Note: In cases when objects represent a physical entity, like a Node representing a physical host, when the host is re-created under the same name without deleting and re-creating the Node, Kubernetes treats the new host as the old one, which may lead to inconsistencies.
Below are four types of commonly used name constraints for resources.
DNS Subdomain Names
Most resource types require a name that can be used as a DNS subdomain name
as defined in RFC 1123.
This means the name must:
contain no more than 253 characters
contain only lowercase alphanumeric characters, '-' or '.'
start with an alphanumeric character
end with an alphanumeric character
RFC 1123 Label Names
Some resource types require their names to follow the DNS
label standard as defined in RFC 1123.
This means the name must:
contain at most 63 characters
contain only lowercase alphanumeric characters or '-'
start with an alphanumeric character
end with an alphanumeric character
RFC 1035 Label Names
Some resource types require their names to follow the DNS
label standard as defined in RFC 1035.
This means the name must:
contain at most 63 characters
contain only lowercase alphanumeric characters or '-'
start with an alphabetic character
end with an alphanumeric character
Path Segment Names
Some resource types require their names to be able to be safely encoded as a
path segment. In other words, the name may not be "." or ".." and the name may
not contain "/" or "%".
Here's an example manifest for a Pod named nginx-demo.
Note: Some resource types have additional restrictions on their names.
UIDs
A Kubernetes systems-generated string to uniquely identify objects.
Every object created over the whole lifetime of a Kubernetes cluster has a distinct UID. It is intended to distinguish between historical occurrences of similar entities.
Kubernetes UIDs are universally unique identifiers (also known as UUIDs).
UUIDs are standardized as ISO/IEC 9834-8 and as ITU-T X.667.
In Kubernetes, namespaces provides a mechanism for isolating groups of resources within a single cluster. Names of resources need to be unique within a namespace, but not across namespaces. Namespace-based scoping is applicable only for namespaced objects (e.g. Deployments, Services, etc) and not for cluster-wide objects (e.g. StorageClass, Nodes, PersistentVolumes, etc).
When to Use Multiple Namespaces
Namespaces are intended for use in environments with many users spread across multiple
teams, or projects. For clusters with a few to tens of users, you should not
need to create or think about namespaces at all. Start using namespaces when you
need the features they provide.
Namespaces provide a scope for names. Names of resources need to be unique within a namespace,
but not across namespaces. Namespaces cannot be nested inside one another and each Kubernetes
resource can only be in one namespace.
Namespaces are a way to divide cluster resources between multiple users (via resource quota).
It is not necessary to use multiple namespaces to separate slightly different
resources, such as different versions of the same software: use
labels to distinguish
resources within the same namespace.
Note: Avoid creating namespaces with the prefix kube-, since it is reserved for Kubernetes system namespaces.
Viewing namespaces
You can list the current namespaces in a cluster using:
kubectl get namespace
NAME STATUS AGE
default Active 1d
kube-node-lease Active 1d
kube-public Active 1d
kube-system Active 1d
Kubernetes starts with four initial namespaces:
default The default namespace for objects with no other namespace
kube-system The namespace for objects created by the Kubernetes system
kube-public This namespace is created automatically and is readable by all users (including those not authenticated). This namespace is mostly reserved for cluster usage, in case that some resources should be visible and readable publicly throughout the whole cluster. The public aspect of this namespace is only a convention, not a requirement.
kube-node-lease This namespace holds Lease
objects associated with each node. Node leases allow the kubelet to send
heartbeats so that the control plane
can detect node failure.
Setting the namespace for a request
To set the namespace for a current request, use the --namespace flag.
For example:
kubectl run nginx --image=nginx --namespace=<insert-namespace-name-here>
kubectl get pods --namespace=<insert-namespace-name-here>
Setting the namespace preference
You can permanently save the namespace for all subsequent kubectl commands in that
context.
When you create a Service,
it creates a corresponding DNS entry.
This entry is of the form <service-name>.<namespace-name>.svc.cluster.local, which means
that if a container only uses <service-name>, it will resolve to the service which
is local to a namespace. This is useful for using the same configuration across
multiple namespaces such as Development, Staging and Production. If you want to reach
across namespaces, you need to use the fully qualified domain name (FQDN).
By creating namespaces with the same name as public top-level
domains, Services in these
namespaces can have short DNS names that overlap with public DNS records.
Workloads from any namespace performing a DNS lookup without a trailing dot will
be redirected to those services, taking precedence over public DNS.
To mitigate this, limit privileges for creating namespaces to trusted users. If
required, you could additionally configure third-party security controls, such
as admission
webhooks,
to block creating any namespace with the name of public
TLDs.
Not All Objects are in a Namespace
Most Kubernetes resources (e.g. pods, services, replication controllers, and others) are
in some namespaces. However namespace resources are not themselves in a namespace.
And low-level resources, such as
nodes and
persistentVolumes, are not in any namespace.
To see which Kubernetes resources are and aren't in a namespace:
# In a namespacekubectl api-resources --namespaced=true# Not in a namespacekubectl api-resources --namespaced=false
Automatic labelling
FEATURE STATE:Kubernetes 1.21 [beta]
The Kubernetes control plane sets an immutable labelkubernetes.io/metadata.name on all namespaces, provided that the NamespaceDefaultLabelNamefeature gate is enabled.
The value of the label is the namespace name.
Labels are key/value pairs that are attached to objects, such as pods.
Labels are intended to be used to specify identifying attributes of objects that are meaningful and relevant to users, but do not directly imply semantics to the core system.
Labels can be used to organize and to select subsets of objects. Labels can be attached to objects at creation time and subsequently added and modified at any time.
Each object can have a set of key/value labels defined. Each Key must be unique for a given object.
Labels allow for efficient queries and watches and are ideal for use in UIs
and CLIs. Non-identifying information should be recorded using
annotations.
Motivation
Labels enable users to map their own organizational structures onto system objects in a loosely coupled fashion, without requiring clients to store these mappings.
Service deployments and batch processing pipelines are often multi-dimensional entities (e.g., multiple partitions or deployments, multiple release tracks, multiple tiers, multiple micro-services per tier). Management often requires cross-cutting operations, which breaks encapsulation of strictly hierarchical representations, especially rigid hierarchies determined by the infrastructure rather than by users.
These are examples of commonly used labels; you are free to develop your own conventions. Keep in mind that label Key must be unique for a given object.
Syntax and character set
Labels are key/value pairs. Valid label keys have two segments: an optional prefix and name, separated by a slash (/). The name segment is required and must be 63 characters or less, beginning and ending with an alphanumeric character ([a-z0-9A-Z]) with dashes (-), underscores (_), dots (.), and alphanumerics between. The prefix is optional. If specified, the prefix must be a DNS subdomain: a series of DNS labels separated by dots (.), not longer than 253 characters in total, followed by a slash (/).
If the prefix is omitted, the label Key is presumed to be private to the user. Automated system components (e.g. kube-scheduler, kube-controller-manager, kube-apiserver, kubectl, or other third-party automation) which add labels to end-user objects must specify a prefix.
The kubernetes.io/ and k8s.io/ prefixes are reserved for Kubernetes core components.
Valid label value:
must be 63 characters or less (can be empty),
unless empty, must begin and end with an alphanumeric character ([a-z0-9A-Z]),
could contain dashes (-), underscores (_), dots (.), and alphanumerics between.
For example, here's the configuration file for a Pod that has two labels environment: production and app: nginx :
Unlike names and UIDs, labels do not provide uniqueness. In general, we expect many objects to carry the same label(s).
Via a label selector, the client/user can identify a set of objects. The label selector is the core grouping primitive in Kubernetes.
The API currently supports two types of selectors: equality-based and set-based.
A label selector can be made of multiple requirements which are comma-separated. In the case of multiple requirements, all must be satisfied so the comma separator acts as a logical AND (&&) operator.
The semantics of empty or non-specified selectors are dependent on the context,
and API types that use selectors should document the validity and meaning of
them.
Note: For some API types, such as ReplicaSets, the label selectors of two instances must not overlap within a namespace, or the controller can see that as conflicting instructions and fail to determine how many replicas should be present.
Caution: For both equality-based and set-based conditions there is no logical OR (||) operator. Ensure your filter statements are structured accordingly.
Equality-based requirement
Equality- or inequality-based requirements allow filtering by label keys and values. Matching objects must satisfy all of the specified label constraints, though they may have additional labels as well.
Three kinds of operators are admitted =,==,!=. The first two represent equality (and are synonyms), while the latter represents inequality. For example:
environment = production
tier != frontend
The former selects all resources with key equal to environment and value equal to production.
The latter selects all resources with key equal to tier and value distinct from frontend, and all resources with no labels with the tier key.
One could filter for resources in production excluding frontend using the comma operator: environment=production,tier!=frontend
One usage scenario for equality-based label requirement is for Pods to specify
node selection criteria. For example, the sample Pod below selects nodes with
the label "accelerator=nvidia-tesla-p100".
Set-based label requirements allow filtering keys according to a set of values. Three kinds of operators are supported: in,notin and exists (only the key identifier). For example:
environment in (production, qa)
tier notin (frontend, backend)
partition
!partition
The first example selects all resources with key equal to environment and value equal to production or qa.
The second example selects all resources with key equal to tier and values other than frontend and backend, and all resources with no labels with the tier key.
The third example selects all resources including a label with key partition; no values are checked.
The fourth example selects all resources without a label with key partition; no values are checked.
Similarly the comma separator acts as an AND operator. So filtering resources with a partition key (no matter the value) and with environment different than qa can be achieved using partition,environment notin (qa).
The set-based label selector is a general form of equality since environment=production is equivalent to environment in (production); similarly for != and notin.
Set-based requirements can be mixed with equality-based requirements. For example: partition in (customerA, customerB),environment!=qa.
API
LIST and WATCH filtering
LIST and WATCH operations may specify label selectors to filter the sets of objects returned using a query parameter. Both requirements are permitted (presented here as they would appear in a URL query string):
Both label selector styles can be used to list or watch resources via a REST client. For example, targeting apiserver with kubectl and using equality-based one may write:
kubectl get pods -l environment=production,tier=frontend
or using set-based requirements:
kubectl get pods -l 'environment in (production),tier in (frontend)'
As already mentioned set-based requirements are more expressive. For instance, they can implement the OR operator on values:
kubectl get pods -l 'environment in (production, qa)'
or restricting negative matching via exists operator:
kubectl get pods -l 'environment,environment notin (frontend)'
Set references in API objects
Some Kubernetes objects, such as services
and replicationcontrollers,
also use label selectors to specify sets of other resources, such as
pods.
Service and ReplicationController
The set of pods that a service targets is defined with a label selector. Similarly, the population of pods that a replicationcontroller should manage is also defined with a label selector.
Labels selectors for both objects are defined in json or yaml files using maps, and only equality-based requirement selectors are supported:
"selector": {
"component" : "redis",
}
or
selector:component:redis
this selector (respectively in json or yaml format) is equivalent to component=redis or component in (redis).
matchLabels is a map of {key,value} pairs. A single {key,value} in the matchLabels map is equivalent to an element of matchExpressions, whose key field is "key", the operator is "In", and the values array contains only "value". matchExpressions is a list of pod selector requirements. Valid operators include In, NotIn, Exists, and DoesNotExist. The values set must be non-empty in the case of In and NotIn. All of the requirements, from both matchLabels and matchExpressions are ANDed together -- they must all be satisfied in order to match.
Selecting sets of nodes
One use case for selecting over labels is to constrain the set of nodes onto which a pod can schedule.
See the documentation on node selection for more information.
3.1.4.6 - Annotations
You can use Kubernetes annotations to attach arbitrary non-identifying metadata
to objects. Clients such as tools and libraries can retrieve this metadata.
Attaching metadata to objects
You can use either labels or annotations to attach metadata to Kubernetes
objects. Labels can be used to select objects and to find
collections of objects that satisfy certain conditions. In contrast, annotations
are not used to identify and select objects. The metadata
in an annotation can be small or large, structured or unstructured, and can
include characters not permitted by labels.
Note: The keys and the values in the map must be strings. In other words, you cannot use
numeric, boolean, list or other types for either the keys or the values.
Here are some examples of information that could be recorded in annotations:
Fields managed by a declarative configuration layer. Attaching these fields
as annotations distinguishes them from default values set by clients or
servers, and from auto-generated fields and fields set by
auto-sizing or auto-scaling systems.
Build, release, or image information like timestamps, release IDs, git branch,
PR numbers, image hashes, and registry address.
Pointers to logging, monitoring, analytics, or audit repositories.
Client library or tool information that can be used for debugging purposes:
for example, name, version, and build information.
User or tool/system provenance information, such as URLs of related objects
from other ecosystem components.
Lightweight rollout tool metadata: for example, config or checkpoints.
Phone or pager numbers of persons responsible, or directory entries that
specify where that information can be found, such as a team web site.
Directives from the end-user to the implementations to modify behavior or
engage non-standard features.
Instead of using annotations, you could store this type of information in an
external database or directory, but that would make it much harder to produce
shared client libraries and tools for deployment, management, introspection,
and the like.
Syntax and character set
Annotations are key/value pairs. Valid annotation keys have two segments: an optional prefix and name, separated by a slash (/). The name segment is required and must be 63 characters or less, beginning and ending with an alphanumeric character ([a-z0-9A-Z]) with dashes (-), underscores (_), dots (.), and alphanumerics between. The prefix is optional. If specified, the prefix must be a DNS subdomain: a series of DNS labels separated by dots (.), not longer than 253 characters in total, followed by a slash (/).
If the prefix is omitted, the annotation Key is presumed to be private to the user. Automated system components (e.g. kube-scheduler, kube-controller-manager, kube-apiserver, kubectl, or other third-party automation) which add annotations to end-user objects must specify a prefix.
The kubernetes.io/ and k8s.io/ prefixes are reserved for Kubernetes core components.
For example, here's the configuration file for a Pod that has the annotation imageregistry: https://hub.docker.com/ :
Field selectors let you select Kubernetes resources based on the value of one or more resource fields. Here are some examples of field selector queries:
metadata.name=my-service
metadata.namespace!=default
status.phase=Pending
This kubectl command selects all Pods for which the value of the status.phase field is Running:
kubectl get pods --field-selector status.phase=Running
Note: Field selectors are essentially resource filters. By default, no selectors/filters are applied, meaning that all resources of the specified type are selected. This makes the kubectl queries kubectl get pods and kubectl get pods --field-selector "" equivalent.
Supported fields
Supported field selectors vary by Kubernetes resource type. All resource types support the metadata.name and metadata.namespace fields. Using unsupported field selectors produces an error. For example:
kubectl get ingress --field-selector foo.bar=baz
Error from server (BadRequest): Unable to find "ingresses" that match label selector "", field selector "foo.bar=baz": "foo.bar" is not a known field selector: only "metadata.name", "metadata.namespace"
Supported operators
You can use the =, ==, and != operators with field selectors (= and == mean the same thing). This kubectl command, for example, selects all Kubernetes Services that aren't in the default namespace:
kubectl get services --all-namespaces --field-selector metadata.namespace!=default
Chained selectors
As with label and other selectors, field selectors can be chained together as a comma-separated list. This kubectl command selects all Pods for which the status.phase does not equal Running and the spec.restartPolicy field equals Always:
kubectl get pods --field-selector=status.phase!=Running,spec.restartPolicy=Always
Multiple resource types
You can use field selectors across multiple resource types. This kubectl command selects all Statefulsets and Services that are not in the default namespace:
kubectl get statefulsets,services --all-namespaces --field-selector metadata.namespace!=default
3.1.4.8 - Finalizers
Finalizers are namespaced keys that tell Kubernetes to wait until specific
conditions are met before it fully deletes resources marked for deletion.
Finalizers alert controllers
to clean up resources the deleted object owned.
When you tell Kubernetes to delete an object that has finalizers specified for
it, the Kubernetes API marks the object for deletion by populating .metadata.deletionTimestamp,
and returns a 202 status code (HTTP "Accepted"). The target object remains in a terminating state while the
control plane, or other components, take the actions defined by the finalizers.
After these actions are complete, the controller removes the relevant finalizers
from the target object. When the metadata.finalizers field is empty,
Kubernetes considers the deletion complete and deletes the object.
You can use finalizers to control garbage collection
of resources. For example, you can define a finalizer to clean up related resources or
infrastructure before the controller deletes the target resource.
You can use finalizers to control garbage collection
of resources by alerting controllers to perform specific cleanup tasks before
deleting the target resource.
Finalizers don't usually specify the code to execute. Instead, they are
typically lists of keys on a specific resource similar to annotations.
Kubernetes specifies some finalizers automatically, but you can also specify
your own.
How finalizers work
When you create a resource using a manifest file, you can specify finalizers in
the metadata.finalizers field. When you attempt to delete the resource, the
API server handling the delete request notices the values in the finalizers field
and does the following:
Modifies the object to add a metadata.deletionTimestamp field with the
time you started the deletion.
Prevents the object from being removed until its metadata.finalizers field is empty.
Returns a 202 status code (HTTP "Accepted")
The controller managing that finalizer notices the update to the object setting the
metadata.deletionTimestamp, indicating deletion of the object has been requested.
The controller then attempts to satisfy the requirements of the finalizers
specified for that resource. Each time a finalizer condition is satisfied, the
controller removes that key from the resource's finalizers field. When the
finalizers field is emptied, an object with a deletionTimestamp field set
is automatically deleted. You can also use finalizers to prevent deletion of unmanaged resources.
A common example of a finalizer is kubernetes.io/pv-protection, which prevents
accidental deletion of PersistentVolume objects. When a PersistentVolume
object is in use by a Pod, Kubernetes adds the pv-protection finalizer. If you
try to delete the PersistentVolume, it enters a Terminating status, but the
controller can't delete it because the finalizer exists. When the Pod stops
using the PersistentVolume, Kubernetes clears the pv-protection finalizer,
and the controller deletes the volume.
Owner references, labels, and finalizers
Like labels,
owner references
describe the relationships between objects in Kubernetes, but are used for a
different purpose. When a
controller manages objects
like Pods, it uses labels to track changes to groups of related objects. For
example, when a Job creates one or
more Pods, the Job controller applies labels to those pods and tracks changes to
any Pods in the cluster with the same label.
The Job controller also adds owner references to those Pods, pointing at the
Job that created the Pods. If you delete the Job while these Pods are running,
Kubernetes uses the owner references (not labels) to determine which Pods in the
cluster need cleanup.
Kubernetes also processes finalizers when it identifies owner references on a
resource targeted for deletion.
In some situations, finalizers can block the deletion of dependent objects,
which can cause the targeted owner object to remain for
longer than expected without being fully deleted. In these situations, you
should check finalizers and owner references on the target owner and dependent
objects to troubleshoot the cause.
Note: In cases where objects are stuck in a deleting state, avoid manually
removing finalizers to allow deletion to continue. Finalizers are usually added
to resources for a reason, so forcefully removing them can lead to issues in
your cluster. This should only be done when the purpose of the finalizer is
understood and is accomplished in another way (for example, manually cleaning
up some dependent object).
In Kubernetes, some objects are owners of other objects. For example, a
ReplicaSet is the owner of a set of Pods. These owned objects are dependents
of their owner.
Ownership is different from the labels and selectors
mechanism that some resources also use. For example, consider a Service that
creates EndpointSlice objects. The Service uses labels to allow the control plane to
determine which EndpointSlice objects are used for that Service. In addition
to the labels, each EndpointSlice that is managed on behalf of a Service has
an owner reference. Owner references help different parts of Kubernetes avoid
interfering with objects they don’t control.
Owner references in object specifications
Dependent objects have a metadata.ownerReferences field that references their
owner object. A valid owner reference consists of the object name and a UID
within the same namespace as the dependent object. Kubernetes sets the value of
this field automatically for objects that are dependents of other objects like
ReplicaSets, DaemonSets, Deployments, Jobs and CronJobs, and ReplicationControllers.
You can also configure these relationships manually by changing the value of
this field. However, you usually don't need to and can allow Kubernetes to
automatically manage the relationships.
Dependent objects also have an ownerReferences.blockOwnerDeletion field that
takes a boolean value and controls whether specific dependents can block garbage
collection from deleting their owner object. Kubernetes automatically sets this
field to true if a controller
(for example, the Deployment controller) sets the value of the
metadata.ownerReferences field. You can also set the value of the
blockOwnerDeletion field manually to control which dependents block garbage
collection.
A Kubernetes admission controller controls user access to change this field for
dependent resources, based on the delete permissions of the owner. This control
prevents unauthorized users from delaying owner object deletion.
Note:
Cross-namespace owner references are disallowed by design.
Namespaced dependents can specify cluster-scoped or namespaced owners.
A namespaced owner must exist in the same namespace as the dependent.
If it does not, the owner reference is treated as absent, and the dependent
is subject to deletion once all owners are verified absent.
Cluster-scoped dependents can only specify cluster-scoped owners.
In v1.20+, if a cluster-scoped dependent specifies a namespaced kind as an owner,
it is treated as having an unresolvable owner reference, and is not able to be garbage collected.
In v1.20+, if the garbage collector detects an invalid cross-namespace ownerReference,
or a cluster-scoped dependent with an ownerReference referencing a namespaced kind, a warning Event
with a reason of OwnerRefInvalidNamespace and an involvedObject of the invalid dependent is reported.
You can check for that kind of Event by running
kubectl get events -A --field-selector=reason=OwnerRefInvalidNamespace.
Ownership and finalizers
When you tell Kubernetes to delete a resource, the API server allows the
managing controller to process any finalizer rules
for the resource. Finalizers
prevent accidental deletion of resources your cluster may still need to function
correctly. For example, if you try to delete a PersistentVolume that is still
in use by a Pod, the deletion does not happen immediately because the
PersistentVolume has the kubernetes.io/pv-protection finalizer on it.
Instead, the volume remains in the Terminating status until Kubernetes clears
the finalizer, which only happens after the PersistentVolume is no longer
bound to a Pod.
Kubernetes also adds finalizers to an owner resource when you use either
foreground or orphan cascading deletion.
In foreground deletion, it adds the foreground finalizer so that the
controller must delete dependent resources that also have
ownerReferences.blockOwnerDeletion=true before it deletes the owner. If you
specify an orphan deletion policy, Kubernetes adds the orphan finalizer so
that the controller ignores dependent resources after it deletes the owner
object.
You can visualize and manage Kubernetes objects with more tools than kubectl and
the dashboard. A common set of labels allows tools to work interoperably, describing
objects in a common manner that all tools can understand.
In addition to supporting tooling, the recommended labels describe applications
in a way that can be queried.
The metadata is organized around the concept of an application. Kubernetes is not
a platform as a service (PaaS) and doesn't have or enforce a formal notion of an application.
Instead, applications are informal and described with metadata. The definition of
what an application contains is loose.
Note: These are recommended labels. They make it easier to manage applications
but aren't required for any core tooling.
Shared labels and annotations share a common prefix: app.kubernetes.io. Labels
without a prefix are private to users. The shared prefix ensures that shared labels
do not interfere with custom user labels.
Labels
In order to take full advantage of using these labels, they should be applied
on every resource object.
Key
Description
Example
Type
app.kubernetes.io/name
The name of the application
mysql
string
app.kubernetes.io/instance
A unique name identifying the instance of an application
mysql-abcxzy
string
app.kubernetes.io/version
The current version of the application (e.g., a semantic version, revision hash, etc.)
5.7.21
string
app.kubernetes.io/component
The component within the architecture
database
string
app.kubernetes.io/part-of
The name of a higher level application this one is part of
wordpress
string
app.kubernetes.io/managed-by
The tool being used to manage the operation of an application
helm
string
app.kubernetes.io/created-by
The controller/user who created this resource
controller-manager
string
To illustrate these labels in action, consider the following StatefulSet object:
# This is an excerptapiVersion:apps/v1kind:StatefulSetmetadata:labels:app.kubernetes.io/name:mysqlapp.kubernetes.io/instance:mysql-abcxzyapp.kubernetes.io/version:"5.7.21"app.kubernetes.io/component:databaseapp.kubernetes.io/part-of:wordpressapp.kubernetes.io/managed-by:helmapp.kubernetes.io/created-by:controller-manager
Applications And Instances Of Applications
An application can be installed one or more times into a Kubernetes cluster and,
in some cases, the same namespace. For example, WordPress can be installed more
than once where different websites are different installations of WordPress.
The name of an application and the instance name are recorded separately. For
example, WordPress has a app.kubernetes.io/name of wordpress while it has
an instance name, represented as app.kubernetes.io/instance with a value of
wordpress-abcxzy. This enables the application and instance of the application
to be identifiable. Every instance of an application must have a unique name.
Examples
To illustrate different ways to use these labels the following examples have varying complexity.
A Simple Stateless Service
Consider the case for a simple stateless service deployed using Deployment and Service objects. The following two snippets represent how the labels could be used in their simplest form.
The Deployment is used to oversee the pods running the application itself.
Consider a slightly more complicated application: a web application (WordPress)
using a database (MySQL), installed using Helm. The following snippets illustrate
the start of objects used to deploy this application.
The start to the following Deployment is used for WordPress:
With the MySQL StatefulSet and Service you'll notice information about both MySQL and WordPress, the broader application, are included.
3.2 - Cluster Architecture
The architectural concepts behind Kubernetes.
3.2.1 - Nodes
Kubernetes runs your workload by placing containers into Pods to run on Nodes.
A node may be a virtual or physical machine, depending on the cluster. Each node
is managed by the
control plane
and contains the services necessary to run
Pods.
Typically you have several nodes in a cluster; in a learning or resource-limited
environment, you might have only one node.
There are two main ways to have Nodes added to the API server:
The kubelet on a node self-registers to the control plane
You (or another human user) manually add a Node object
After you create a Node object,
or the kubelet on a node self-registers, the control plane checks whether the new Node object is
valid. For example, if you try to create a Node from the following JSON manifest:
Kubernetes creates a Node object internally (the representation). Kubernetes checks
that a kubelet has registered to the API server that matches the metadata.name
field of the Node. If the node is healthy (i.e. all necessary services are running),
then it is eligible to run a Pod. Otherwise, that node is ignored for any cluster activity
until it becomes healthy.
Note:
Kubernetes keeps the object for the invalid Node and continues checking to see whether
it becomes healthy.
You, or a controller, must explicitly
delete the Node object to stop that health checking.
The name identifies a Node. Two Nodes
cannot have the same name at the same time. Kubernetes also assumes that a resource with the same
name is the same object. In case of a Node, it is implicitly assumed that an instance using the
same name will have the same state (e.g. network settings, root disk contents)
and attributes like node labels. This may lead to
inconsistencies if an instance was modified without changing its name. If the Node needs to be
replaced or updated significantly, the existing Node object needs to be removed from API server
first and re-added after the update.
Self-registration of Nodes
When the kubelet flag --register-node is true (the default), the kubelet will attempt to
register itself with the API server. This is the preferred pattern, used by most distros.
For self-registration, the kubelet is started with the following options:
--kubeconfig - Path to credentials to authenticate itself to the API server.
--cloud-provider - How to talk to a cloud provider
to read metadata about itself.
--register-node - Automatically register with the API server.
--register-with-taints - Register the node with the given list of
taints (comma separated <key>=<value>:<effect>).
As mentioned in the Node name uniqueness section,
when Node configuration needs to be updated, it is a good practice to re-register
the node with the API server. For example, if the kubelet being restarted with
the new set of --node-labels, but the same Node name is used, the change will
not take an effect, as labels are being set on the Node registration.
Pods already scheduled on the Node may misbehave or cause issues if the Node
configuration will be changed on kubelet restart. For example, already running
Pod may be tainted against the new labels assigned to the Node, while other
Pods, that are incompatible with that Pod will be scheduled based on this new
label. Node re-registration ensures all Pods will be drained and properly
re-scheduled.
Manual Node administration
You can create and modify Node objects using
kubectl.
When you want to create Node objects manually, set the kubelet flag --register-node=false.
You can modify Node objects regardless of the setting of --register-node.
For example, you can set labels on an existing Node or mark it unschedulable.
You can use labels on Nodes in conjunction with node selectors on Pods to control
scheduling. For example, you can constrain a Pod to only be eligible to run on
a subset of the available nodes.
Marking a node as unschedulable prevents the scheduler from placing new pods onto
that Node but does not affect existing Pods on the Node. This is useful as a
preparatory step before a node reboot or other maintenance.
Note: Pods that are part of a DaemonSet tolerate
being run on an unschedulable Node. DaemonSets typically provide node-local services
that should run on the Node even if it is being drained of workload applications.
Node status
A Node's status contains the following information:
You can use kubectl to view a Node's status and other details:
kubectl describe node <insert-node-name-here>
Each section of the output is described below.
Addresses
The usage of these fields varies depending on your cloud provider or bare metal configuration.
HostName: The hostname as reported by the node's kernel. Can be overridden via the kubelet
--hostname-override parameter.
ExternalIP: Typically the IP address of the node that is externally routable (available from
outside the cluster).
InternalIP: Typically the IP address of the node that is routable only within the cluster.
Conditions
The conditions field describes the status of all Running nodes. Examples of conditions include:
Node conditions, and a description of when each condition applies.
Node Condition
Description
Ready
True if the node is healthy and ready to accept pods, False if the node is not healthy and is not accepting pods, and Unknown if the node controller has not heard from the node in the last node-monitor-grace-period (default is 40 seconds)
DiskPressure
True if pressure exists on the disk size—that is, if the disk capacity is low; otherwise False
MemoryPressure
True if pressure exists on the node memory—that is, if the node memory is low; otherwise False
PIDPressure
True if pressure exists on the processes—that is, if there are too many processes on the node; otherwise False
NetworkUnavailable
True if the network for the node is not correctly configured, otherwise False
Note: If you use command-line tools to print details of a cordoned Node, the Condition includes
SchedulingDisabled. SchedulingDisabled is not a Condition in the Kubernetes API; instead,
cordoned nodes are marked Unschedulable in their spec.
In the Kubernetes API, a node's condition is represented as part of the .status
of the Node resource. For example, the following JSON structure describes a healthy node:
If the status of the Ready condition remains Unknown or False for longer
than the pod-eviction-timeout (an argument passed to the
kube-controller-manager), then the node controller triggers
API-initiated eviction
for all Pods assigned to that node. The default eviction timeout duration is
five minutes.
In some cases when the node is unreachable, the API server is unable to communicate
with the kubelet on the node. The decision to delete the pods cannot be communicated to
the kubelet until communication with the API server is re-established. In the meantime,
the pods that are scheduled for deletion may continue to run on the partitioned node.
The node controller does not force delete pods until it is confirmed that they have stopped
running in the cluster. You can see the pods that might be running on an unreachable node as
being in the Terminating or Unknown state. In cases where Kubernetes cannot deduce from the
underlying infrastructure if a node has permanently left a cluster, the cluster administrator
may need to delete the node object by hand. Deleting the node object from Kubernetes causes
all the Pod objects running on the node to be deleted from the API server and frees up their
names.
When problems occur on nodes, the Kubernetes control plane automatically creates
taints that match the conditions
affecting the node.
The scheduler takes the Node's taints into consideration when assigning a Pod to a Node.
Pods can also have tolerations that let
them run on a Node even though it has a specific taint.
Describes the resources available on the node: CPU, memory, and the maximum
number of pods that can be scheduled onto the node.
The fields in the capacity block indicate the total amount of resources that a
Node has. The allocatable block indicates the amount of resources on a
Node that is available to be consumed by normal Pods.
You may read more about capacity and allocatable resources while learning how
to reserve compute resources
on a Node.
Info
Describes general information about the node, such as kernel version, Kubernetes
version (kubelet and kube-proxy version), container runtime details, and which
operating system the node uses.
The kubelet gathers this information from the node and publishes it into
the Kubernetes API.
Heartbeats
Heartbeats, sent by Kubernetes nodes, help your cluster determine the
availability of each node, and to take action when failures are detected.
For nodes there are two forms of heartbeats:
updates to the .status of a Node
Lease objects
within the kube-node-leasenamespace.
Each Node has an associated Lease object.
Compared to updates to .status of a Node, a Lease is a lightweight resource.
Using Leases for heartbeats reduces the performance impact of these updates
for large clusters.
The kubelet is responsible for creating and updating the .status of Nodes,
and for updating their related Leases.
The kubelet updates the node's .status either when there is change in status
or if there has been no update for a configured interval. The default interval
for .status updates to Nodes is 5 minutes, which is much longer than the 40
second default timeout for unreachable nodes.
The kubelet creates and then updates its Lease object every 10 seconds
(the default update interval). Lease updates occur independently from
updates to the Node's .status. If the Lease update fails, the kubelet retries,
using exponential backoff that starts at 200 milliseconds and capped at 7 seconds.
Node controller
The node controller is a
Kubernetes control plane component that manages various aspects of nodes.
The node controller has multiple roles in a node's life. The first is assigning a
CIDR block to the node when it is registered (if CIDR assignment is turned on).
The second is keeping the node controller's internal list of nodes up to date with
the cloud provider's list of available machines. When running in a cloud
environment and whenever a node is unhealthy, the node controller asks the cloud
provider if the VM for that node is still available. If not, the node
controller deletes the node from its list of nodes.
The third is monitoring the nodes' health. The node controller is
responsible for:
In the case that a node becomes unreachable, updating the Ready condition
in the Node's .status field. In this case the node controller sets the
Ready condition to Unknown.
If a node remains unreachable: triggering
API-initiated eviction
for all of the Pods on the unreachable node. By default, the node controller
waits 5 minutes between marking the node as Unknown and submitting
the first eviction request.
By default, the node controller checks the state of each node every 5 seconds.
This period can be configured using the --node-monitor-period flag on the
kube-controller-manager component.
Rate limits on eviction
In most cases, the node controller limits the eviction rate to
--node-eviction-rate (default 0.1) per second, meaning it won't evict pods
from more than 1 node per 10 seconds.
The node eviction behavior changes when a node in a given availability zone
becomes unhealthy. The node controller checks what percentage of nodes in the zone
are unhealthy (the Ready condition is Unknown or False) at
the same time:
If the fraction of unhealthy nodes is at least --unhealthy-zone-threshold
(default 0.55), then the eviction rate is reduced.
If the cluster is small (i.e. has less than or equal to
--large-cluster-size-threshold nodes - default 50), then evictions are stopped.
Otherwise, the eviction rate is reduced to --secondary-node-eviction-rate
(default 0.01) per second.
The reason these policies are implemented per availability zone is because one
availability zone might become partitioned from the control plane while the others remain
connected. If your cluster does not span multiple cloud provider availability zones,
then the eviction mechanism does not take per-zone unavailability into account.
A key reason for spreading your nodes across availability zones is so that the
workload can be shifted to healthy zones when one entire zone goes down.
Therefore, if all nodes in a zone are unhealthy, then the node controller evicts at
the normal rate of --node-eviction-rate. The corner case is when all zones are
completely unhealthy (none of the nodes in the cluster are healthy). In such a
case, the node controller assumes that there is some problem with connectivity
between the control plane and the nodes, and doesn't perform any evictions.
(If there has been an outage and some nodes reappear, the node controller does
evict pods from the remaining nodes that are unhealthy or unreachable).
The node controller is also responsible for evicting pods running on nodes with
NoExecute taints, unless those pods tolerate that taint.
The node controller also adds taints
corresponding to node problems like node unreachable or not ready. This means
that the scheduler won't place Pods onto unhealthy nodes.
Resource capacity tracking
Node objects track information about the Node's resource capacity: for example, the amount
of memory available and the number of CPUs.
Nodes that self register report their capacity during
registration. If you manually add a Node, then
you need to set the node's capacity information when you add it.
The Kubernetes scheduler ensures that
there are enough resources for all the Pods on a Node. The scheduler checks that the sum
of the requests of containers on the node is no greater than the node's capacity.
That sum of requests includes all containers managed by the kubelet, but excludes any
containers started directly by the container runtime, and also excludes any
processes running outside of the kubelet's control.
The kubelet attempts to detect node system shutdown and terminates pods running on the node.
Kubelet ensures that pods follow the normal
pod termination process
during the node shutdown.
The Graceful node shutdown feature depends on systemd since it takes advantage of
systemd inhibitor locks to
delay the node shutdown with a given duration.
Graceful node shutdown is controlled with the GracefulNodeShutdownfeature gate which is
enabled by default in 1.21.
Note that by default, both configuration options described below,
shutdownGracePeriod and shutdownGracePeriodCriticalPods are set to zero,
thus not activating the graceful node shutdown functionality.
To activate the feature, the two kubelet config settings should be configured appropriately and
set to non-zero values.
During a graceful shutdown, kubelet terminates pods in two phases:
Graceful node shutdown feature is configured with two
KubeletConfiguration options:
shutdownGracePeriod:
Specifies the total duration that the node should delay the shutdown by. This is the total
grace period for pod termination for both regular and
critical pods.
shutdownGracePeriodCriticalPods:
Specifies the duration used to terminate
critical pods
during a node shutdown. This value should be less than shutdownGracePeriod.
For example, if shutdownGracePeriod=30s, and
shutdownGracePeriodCriticalPods=10s, kubelet will delay the node shutdown by
30 seconds. During the shutdown, the first 20 (30-10) seconds would be reserved
for gracefully terminating normal pods, and the last 10 seconds would be
reserved for terminating critical pods.
Note:
When pods were evicted during the graceful node shutdown, they are marked as shutdown.
Running kubectl get pods shows the status of the the evicted pods as Terminated.
And kubectl describe pod indicates that the pod was evicted because of node shutdown:
Reason: Terminated
Message: Pod was terminated in response to imminent node shutdown.
Non Graceful node shutdown
FEATURE STATE:Kubernetes v1.24 [alpha]
A node shutdown action may not be detected by kubelet's Node Shutdown Manager,
either because the command does not trigger the inhibitor locks mechanism used by
kubelet or because of a user error, i.e., the ShutdownGracePeriod and
ShutdownGracePeriodCriticalPods are not configured properly. Please refer to above
section Graceful Node Shutdown for more details.
When a node is shutdown but not detected by kubelet's Node Shutdown Manager, the pods
that are part of a StatefulSet will be stuck in terminating status on
the shutdown node and cannot move to a new running node. This is because kubelet on
the shutdown node is not available to delete the pods so the StatefulSet cannot
create a new pod with the same name. If there are volumes used by the pods, the
VolumeAttachments will not be deleted from the original shutdown node so the volumes
used by these pods cannot be attached to a new running node. As a result, the
application running on the StatefulSet cannot function properly. If the original
shutdown node comes up, the pods will be deleted by kubelet and new pods will be
created on a different running node. If the original shutdown node does not come up,
these pods will be stuck in terminating status on the shutdown node forever.
To mitigate the above situation, a user can manually add the taint node kubernetes.io/out-of-service with either NoExecute or NoSchedule effect to
a Node marking it out-of-service.
If the NodeOutOfServiceVolumeDetachfeature gate
is enabled on kube-controller-manager, and a Node is marked out-of-service with this taint, the
pods on the node will be forcefully deleted if there are no matching tolerations on it and volume
detach operations for the pods terminating on the node will happen immediately. This allows the
Pods on the out-of-service node to recover quickly on a different node.
During a non-graceful shutdown, Pods are terminated in the two phases:
Force delete the Pods that do not have matching out-of-service tolerations.
Immediately perform detach volume operation for such pods.
Note:
Before adding the taint node.kubernetes.io/out-of-service , it should be verified
that the node is already in shutdown or power off state (not in the middle of
restarting).
The user is required to manually remove the out-of-service taint after the pods are
moved to a new node and the user has checked that the shutdown node has been
recovered since the user was the one who originally added the taint.
Pod Priority based graceful node shutdown
FEATURE STATE:Kubernetes v1.23 [alpha]
To provide more flexibility during graceful node shutdown around the ordering
of pods during shutdown, graceful node shutdown honors the PriorityClass for
Pods, provided that you enabled this feature in your cluster. The feature
allows cluster administers to explicitly define the ordering of pods
during graceful node shutdown based on
priority classes.
The Graceful Node Shutdown feature, as described
above, shuts down pods in two phases, non-critical pods, followed by critical
pods. If additional flexibility is needed to explicitly define the ordering of
pods during shutdown in a more granular way, pod priority based graceful
shutdown can be used.
When graceful node shutdown honors pod priorities, this makes it possible to do
graceful node shutdown in multiple phases, each phase shutting down a
particular priority class of pods. The kubelet can be configured with the exact
phases and shutdown time per phase.
Assuming the following custom pod
priority classes
in a cluster,
Pod priority class name
Pod priority class value
custom-class-a
100000
custom-class-b
10000
custom-class-c
1000
regular/unset
0
Within the kubelet configuration
the settings for shutdownGracePeriodByPodPriority could look like:
Pod priority class value
Shutdown period
100000
10 seconds
10000
180 seconds
1000
120 seconds
0
60 seconds
The corresponding kubelet config YAML configuration would be:
The above table implies that any pod with priority value >= 100000 will get
just 10 seconds to stop, any pod with value >= 10000 and < 100000 will get 180
seconds to stop, any pod with value >= 1000 and < 10000 will get 120 seconds to stop.
Finally, all other pods will get 60 seconds to stop.
One doesn't have to specify values corresponding to all of the classes. For
example, you could instead use these settings:
Pod priority class value
Shutdown period
100000
300 seconds
1000
120 seconds
0
60 seconds
In the above case, the pods with custom-class-b will go into the same bucket
as custom-class-c for shutdown.
If there are no pods in a particular range, then the kubelet does not wait
for pods in that priority range. Instead, the kubelet immediately skips to the
next priority class value range.
If this feature is enabled and no configuration is provided, then no ordering
action will be taken.
Using this feature requires enabling the GracefulNodeShutdownBasedOnPodPriorityfeature gate
, and setting ShutdownGracePeriodByPodPriority in the
kubelet config
to the desired configuration containing the pod priority class values and
their respective shutdown periods.
Note: The ability to take Pod priority into account during graceful node shutdown was introduced
as an Alpha feature in Kubernetes v1.23. In Kubernetes 1.25
the feature is Beta and is enabled by default.
Metrics graceful_shutdown_start_time_seconds and graceful_shutdown_end_time_seconds
are emitted under the kubelet subsystem to monitor node shutdowns.
Swap memory management
FEATURE STATE:Kubernetes v1.22 [alpha]
Prior to Kubernetes 1.22, nodes did not support the use of swap memory, and a
kubelet would by default fail to start if swap was detected on a node. In 1.22
onwards, swap memory support can be enabled on a per-node basis.
To enable swap on a node, the NodeSwap feature gate must be enabled on
the kubelet, and the --fail-swap-on command line flag or failSwapOnconfiguration setting
must be set to false.
Warning: When the memory swap feature is turned on, Kubernetes data such as the content
of Secret objects that were written to tmpfs now could be swapped to disk.
A user can also optionally configure memorySwap.swapBehavior in order to
specify how a node will use swap memory. For example,
memorySwap:swapBehavior:LimitedSwap
The available configuration options for swapBehavior are:
LimitedSwap: Kubernetes workloads are limited in how much swap they can
use. Workloads on the node not managed by Kubernetes can still swap.
UnlimitedSwap: Kubernetes workloads can use as much swap memory as they
request, up to the system limit.
If configuration for memorySwap is not specified and the feature gate is
enabled, by default the kubelet will apply the same behaviour as the
LimitedSwap setting.
The behaviour of the LimitedSwap setting depends if the node is running with
v1 or v2 of control groups (also known as "cgroups"):
cgroupsv1: Kubernetes workloads can use any combination of memory and
swap, up to the pod's memory limit, if set.
cgroupsv2: Kubernetes workloads cannot use swap memory.
For more information, and to assist with testing and provide feedback, please
see KEP-2400 and its
design proposal.
3.2.2 - Communication between Nodes and the Control Plane
This document catalogs the communication paths between the API server and the Kubernetes cluster.
The intent is to allow users to customize their installation to harden the network configuration
such that the cluster can be run on an untrusted network (or on fully public IPs on a cloud
provider).
Node to Control Plane
Kubernetes has a "hub-and-spoke" API pattern. All API usage from nodes (or the pods they run)
terminates at the API server. None of the other control plane components are designed to expose
remote services. The API server is configured to listen for remote connections on a secure HTTPS
port (typically 443) with one or more forms of client
authentication enabled.
One or more forms of authorization should be
enabled, especially if anonymous requests
or service account tokens
are allowed.
Nodes should be provisioned with the public root certificate for the cluster such that they can
connect securely to the API server along with valid client credentials. A good approach is that the
client credentials provided to the kubelet are in the form of a client certificate. See
kubelet TLS bootstrapping
for automated provisioning of kubelet client certificates.
Pods that wish to connect to the API server can do so securely by leveraging a service account so
that Kubernetes will automatically inject the public root certificate and a valid bearer token
into the pod when it is instantiated.
The kubernetes service (in default namespace) is configured with a virtual IP address that is
redirected (via kube-proxy) to the HTTPS endpoint on the API server.
The control plane components also communicate with the API server over the secure port.
As a result, the default operating mode for connections from the nodes and pods running on the
nodes to the control plane is secured by default and can run over untrusted and/or public
networks.
Control plane to node
There are two primary communication paths from the control plane (the API server) to the nodes.
The first is from the API server to the kubelet process which runs on each node in the cluster.
The second is from the API server to any node, pod, or service through the API server's proxy
functionality.
API server to kubelet
The connections from the API server to the kubelet are used for:
Fetching logs for pods.
Attaching (usually through kubectl) to running pods.
Providing the kubelet's port-forwarding functionality.
These connections terminate at the kubelet's HTTPS endpoint. By default, the API server does not
verify the kubelet's serving certificate, which makes the connection subject to man-in-the-middle
attacks and unsafe to run over untrusted and/or public networks.
To verify this connection, use the --kubelet-certificate-authority flag to provide the API
server with a root certificate bundle to use to verify the kubelet's serving certificate.
If that is not possible, use SSH tunneling between the API server and kubelet if
required to avoid connecting over an
untrusted or public network.
The connections from the API server to a node, pod, or service default to plain HTTP connections
and are therefore neither authenticated nor encrypted. They can be run over a secure HTTPS
connection by prefixing https: to the node, pod, or service name in the API URL, but they will
not validate the certificate provided by the HTTPS endpoint nor provide client credentials. So
while the connection will be encrypted, it will not provide any guarantees of integrity. These
connections are not currently safe to run over untrusted or public networks.
SSH tunnels
Kubernetes supports SSH tunnels to protect the control plane to nodes communication paths. In this
configuration, the API server initiates an SSH tunnel to each node in the cluster (connecting to
the SSH server listening on port 22) and passes all traffic destined for a kubelet, node, pod, or
service through the tunnel.
This tunnel ensures that the traffic is not exposed outside of the network in which the nodes are
running.
Note: SSH tunnels are currently deprecated, so you shouldn't opt to use them unless you know what you
are doing. The Konnectivity service is a replacement for this
communication channel.
Konnectivity service
FEATURE STATE:Kubernetes v1.18 [beta]
As a replacement to the SSH tunnels, the Konnectivity service provides TCP level proxy for the
control plane to cluster communication. The Konnectivity service consists of two parts: the
Konnectivity server in the control plane network and the Konnectivity agents in the nodes network.
The Konnectivity agents initiate connections to the Konnectivity server and maintain the network
connections.
After enabling the Konnectivity service, all control plane to nodes traffic goes through these
connections.
In robotics and automation, a control loop is
a non-terminating loop that regulates the state of a system.
Here is one example of a control loop: a thermostat in a room.
When you set the temperature, that's telling the thermostat
about your desired state. The actual room temperature is the
current state. The thermostat acts to bring the current state
closer to the desired state, by turning equipment on or off.
In Kubernetes, controllers are control loops that watch the state of your
cluster, then make or request
changes where needed.
Each controller tries to move the current cluster state closer to the desired
state.
Controller pattern
A controller tracks at least one Kubernetes resource type.
These objects
have a spec field that represents the desired state. The
controller(s) for that resource are responsible for making the current
state come closer to that desired state.
The controller might carry the action out itself; more commonly, in Kubernetes,
a controller will send messages to the
API server that have
useful side effects. You'll see examples of this below.
Control via API server
The Job controller is an example of a
Kubernetes built-in controller. Built-in controllers manage state by
interacting with the cluster API server.
Job is a Kubernetes resource that runs a
Pod, or perhaps several Pods, to carry out
a task and then stop.
(Once scheduled, Pod objects become part of the
desired state for a kubelet).
When the Job controller sees a new task it makes sure that, somewhere
in your cluster, the kubelets on a set of Nodes are running the right
number of Pods to get the work done.
The Job controller does not run any Pods or containers
itself. Instead, the Job controller tells the API server to create or remove
Pods.
Other components in the
control plane
act on the new information (there are new Pods to schedule and run),
and eventually the work is done.
After you create a new Job, the desired state is for that Job to be completed.
The Job controller makes the current state for that Job be nearer to your
desired state: creating Pods that do the work you wanted for that Job, so that
the Job is closer to completion.
Controllers also update the objects that configure them.
For example: once the work is done for a Job, the Job controller
updates that Job object to mark it Finished.
(This is a bit like how some thermostats turn a light off to
indicate that your room is now at the temperature you set).
Direct control
By contrast with Job, some controllers need to make changes to
things outside of your cluster.
For example, if you use a control loop to make sure there
are enough Nodes
in your cluster, then that controller needs something outside the
current cluster to set up new Nodes when needed.
Controllers that interact with external state find their desired state from
the API server, then communicate directly with an external system to bring
the current state closer in line.
(There actually is a controller
that horizontally scales the nodes in your cluster.)
The important point here is that the controller makes some change to bring about
your desired state, and then reports current state back to your cluster's API server.
Other control loops can observe that reported data and take their own actions.
In the thermostat example, if the room is very cold then a different controller
might also turn on a frost protection heater. With Kubernetes clusters, the control
plane indirectly works with IP address management tools, storage services,
cloud provider APIs, and other services by
extending Kubernetes to implement that.
Desired versus current state
Kubernetes takes a cloud-native view of systems, and is able to handle
constant change.
Your cluster could be changing at any point as work happens and
control loops automatically fix failures. This means that,
potentially, your cluster never reaches a stable state.
As long as the controllers for your cluster are running and able to make
useful changes, it doesn't matter if the overall state is stable or not.
Design
As a tenet of its design, Kubernetes uses lots of controllers that each manage
a particular aspect of cluster state. Most commonly, a particular control loop
(controller) uses one kind of resource as its desired state, and has a different
kind of resource that it manages to make that desired state happen. For example,
a controller for Jobs tracks Job objects (to discover new work) and Pod objects
(to run the Jobs, and then to see when the work is finished). In this case
something else creates the Jobs, whereas the Job controller creates Pods.
It's useful to have simple controllers rather than one, monolithic set of control
loops that are interlinked. Controllers can fail, so Kubernetes is designed to
allow for that.
Note:
There can be several controllers that create or update the same kind of object.
Behind the scenes, Kubernetes controllers make sure that they only pay attention
to the resources linked to their controlling resource.
For example, you can have Deployments and Jobs; these both create Pods.
The Job controller does not delete the Pods that your Deployment created,
because there is information (labels)
the controllers can use to tell those Pods apart.
Ways of running controllers
Kubernetes comes with a set of built-in controllers that run inside
the kube-controller-manager. These
built-in controllers provide important core behaviors.
The Deployment controller and Job controller are examples of controllers that
come as part of Kubernetes itself ("built-in" controllers).
Kubernetes lets you run a resilient control plane, so that if any of the built-in
controllers were to fail, another part of the control plane will take over the work.
You can find controllers that run outside the control plane, to extend Kubernetes.
Or, if you want, you can write a new controller yourself.
You can run your own controller as a set of Pods,
or externally to Kubernetes. What fits best will depend on what that particular
controller does.
If you want to write your own controller, see
Extension Patterns
in Extending Kubernetes.
3.2.4 - Cloud Controller Manager
FEATURE STATE:Kubernetes v1.11 [beta]
Cloud infrastructure technologies let you run Kubernetes on public, private, and hybrid clouds.
Kubernetes believes in automated, API-driven infrastructure without tight coupling between
components.
The cloud-controller-manager is a Kubernetes control plane component
that embeds cloud-specific control logic. The cloud controller manager lets you link your
cluster into your cloud provider's API, and separates out the components that interact
with that cloud platform from components that only interact with your cluster.
By decoupling the interoperability logic between Kubernetes and the underlying cloud
infrastructure, the cloud-controller-manager component enables cloud providers to release
features at a different pace compared to the main Kubernetes project.
The cloud-controller-manager is structured using a plugin
mechanism that allows different cloud providers to integrate their platforms with Kubernetes.
Design
The cloud controller manager runs in the control plane as a replicated set of processes
(usually, these are containers in Pods). Each cloud-controller-manager implements
multiple controllers in a single
process.
Note: You can also run the cloud controller manager as a Kubernetes
addon rather than as part
of the control plane.
Cloud controller manager functions
The controllers inside the cloud controller manager include:
Node controller
The node controller is responsible for updating Node objects
when new servers are created in your cloud infrastructure. The node controller obtains information about the
hosts running inside your tenancy with the cloud provider. The node controller performs the following functions:
Update a Node object with the corresponding server's unique identifier obtained from the cloud provider API.
Annotating and labelling the Node object with cloud-specific information, such as the region the node
is deployed into and the resources (CPU, memory, etc) that it has available.
Obtain the node's hostname and network addresses.
Verifying the node's health. In case a node becomes unresponsive, this controller checks with
your cloud provider's API to see if the server has been deactivated / deleted / terminated.
If the node has been deleted from the cloud, the controller deletes the Node object from your Kubernetes
cluster.
Some cloud provider implementations split this into a node controller and a separate node
lifecycle controller.
Route controller
The route controller is responsible for configuring routes in the cloud
appropriately so that containers on different nodes in your Kubernetes
cluster can communicate with each other.
Depending on the cloud provider, the route controller might also allocate blocks
of IP addresses for the Pod network.
Service controller
Services integrate with cloud
infrastructure components such as managed load balancers, IP addresses, network
packet filtering, and target health checking. The service controller interacts with your
cloud provider's APIs to set up load balancers and other infrastructure components
when you declare a Service resource that requires them.
Authorization
This section breaks down the access that the cloud controller manager requires
on various API objects, in order to perform its operations.
Node controller
The Node controller only works with Node objects. It requires full access
to read and modify Node objects.
v1/Node:
Get
List
Create
Update
Patch
Watch
Delete
Route controller
The route controller listens to Node object creation and configures
routes appropriately. It requires Get access to Node objects.
v1/Node:
Get
Service controller
The service controller listens to Service object Create, Update and Delete events and then configures Endpoints for those Services appropriately.
To access Services, it requires List, and Watch access. To update Services, it requires Patch and Update access.
To set up Endpoints resources for the Services, it requires access to Create, List, Get, Watch, and Update.
v1/Service:
List
Get
Watch
Patch
Update
Others
The implementation of the core of the cloud controller manager requires access to create Event objects, and to ensure secure operation, it requires access to create ServiceAccounts.
v1/Event:
Create
Patch
Update
v1/ServiceAccount:
Create
The RBAC ClusterRole for the cloud
controller manager looks like:
Want to know how to implement your own cloud controller manager, or extend an existing project?
The cloud controller manager uses Go interfaces to allow implementations from any cloud to be plugged in. Specifically, it uses the CloudProvider interface defined in cloud.go from kubernetes/cloud-provider.
The implementation of the shared controllers highlighted in this document (Node, Route, and Service), and some scaffolding along with the shared cloudprovider interface, is part of the Kubernetes core. Implementations specific to cloud providers are outside the core of Kubernetes and implement the CloudProvider interface.
On Linux, control groups
constrain resources that are allocated to processes.
The kubelet and the
underlying container runtime need to interface with cgroups to enforce
resource mangement for pods and containers which
includes cpu/memory requests and limits for containerized workloads.
There are two versions of cgroups in Linux: cgroup v1 and cgroup v2. cgroup v2 is
the new generation of the cgroup API.
What is cgroup v2?
FEATURE STATE:Kubernetes v1.25 [stable]
cgroup v2 is the next version of the Linux cgroup API. cgroup v2 provides a
unified control system with enhanced resource management
capabilities.
cgroup v2 offers several improvements over cgroup v1, such as the following:
Enhanced resource allocation management and isolation across multiple resources
Unified accounting for different types of memory allocations (network memory, kernel memory, etc)
Accounting for non-immediate resource changes such as page cache write backs
Some Kubernetes features exclusively use cgroup v2 for enhanced resource
management and isolation. For example, the
MemoryQoS feature improves memory QoS
and relies on cgroup v2 primitives.
Using cgroup v2
The recommended way to use cgroup v2 is to use a Linux distribution that
enables and uses cgroup v2 by default.
You can also enable cgroup v2 manually on your Linux distribution by modifying
the kernel cmdline boot arguments. If your distribution uses GRUB,
systemd.unified_cgroup_hierarchy=1 should be added in GRUB_CMDLINE_LINUX
under /etc/default/grub, followed by sudo update-grub. However, the
recommended approach is to use a distribution that already enables cgroup v2 by
default.
Migrating to cgroup v2
To migrate to cgroup v2, ensure that you meet the requirements, then upgrade
to a kernel version that enables cgroup v2 by default.
The kubelet automatically detects that the OS is running on cgroup v2 and
performs accordingly with no additional configuration required.
There should not be any noticeable difference in the user experience when
switching to cgroup v2, unless users are accessing the cgroup file system
directly, either on the node or from within the containers.
cgroup v2 uses a different API than cgroup v1, so if there are any
applications that directly access the cgroup file system, they need to be
updated to newer versions that support cgroup v2. For example:
Some third-party monitoring and security agents may depend on the cgroup filesystem.
Update these agents to versions that support cgroup v2.
If you run cAdvisor as a stand-alone
DaemonSet for monitoring pods and containers, update it to v0.43.0 or later.
If you use JDK, prefer to use JDK 11.0.16 and later or JDK 15 and later, which fully support cgroup v2.
Identify the cgroup version on Linux Nodes
The cgroup version depends on on the Linux distribution being used and the
default cgroup version configured on the OS. To check which cgroup version your
distribution uses, run the stat -fc %T /sys/fs/cgroup/ command on
the node:
The CRI is a plugin interface which enables the kubelet to use a wide variety of
container runtimes, without having a need to recompile the cluster components.
You need a working
container runtime on
each Node in your cluster, so that the
kubelet can launch
Pods and their containers.
The Container Runtime Interface (CRI) is the main protocol for the communication between the kubelet and Container Runtime.
The kubelet acts as a client when connecting to the container runtime via gRPC.
The runtime and image service endpoints have to be available in the container
runtime, which can be configured separately within the kubelet by using the
--image-service-endpoint and --container-runtime-endpointcommand line
flags
For Kubernetes v1.25, the kubelet prefers to use CRI v1.
If a container runtime does not support v1 of the CRI, then the kubelet tries to
negotiate any older supported version.
The v1.25 kubelet can also negotiate CRI v1alpha2, but
this version is considered as deprecated.
If the kubelet cannot negotiate a supported CRI version, the kubelet gives up
and doesn't register as a node.
Upgrading
When upgrading Kubernetes, then the kubelet tries to automatically select the
latest CRI version on restart of the component. If that fails, then the fallback
will take place as mentioned above. If a gRPC re-dial was required because the
container runtime has been upgraded, then the container runtime must also
support the initially selected version or the redial is expected to fail. This
requires a restart of the kubelet.
Garbage collection is a collective term for the various mechanisms Kubernetes uses to clean up
cluster resources. This
allows the clean up of resources like the following:
Many objects in Kubernetes link to each other through owner references.
Owner references tell the control plane which objects are dependent on others.
Kubernetes uses owner references to give the control plane, and other API
clients, the opportunity to clean up related resources before deleting an
object. In most cases, Kubernetes manages owner references automatically.
Ownership is different from the labels and selectors
mechanism that some resources also use. For example, consider a
Service that creates
EndpointSlice objects. The Service uses labels to allow the control plane to
determine which EndpointSlice objects are used for that Service. In addition
to the labels, each EndpointSlice that is managed on behalf of a Service has
an owner reference. Owner references help different parts of Kubernetes avoid
interfering with objects they don’t control.
Note:
Cross-namespace owner references are disallowed by design.
Namespaced dependents can specify cluster-scoped or namespaced owners.
A namespaced owner must exist in the same namespace as the dependent.
If it does not, the owner reference is treated as absent, and the dependent
is subject to deletion once all owners are verified absent.
Cluster-scoped dependents can only specify cluster-scoped owners.
In v1.20+, if a cluster-scoped dependent specifies a namespaced kind as an owner,
it is treated as having an unresolvable owner reference, and is not able to be garbage collected.
In v1.20+, if the garbage collector detects an invalid cross-namespace ownerReference,
or a cluster-scoped dependent with an ownerReference referencing a namespaced kind, a warning Event
with a reason of OwnerRefInvalidNamespace and an involvedObject of the invalid dependent is reported.
You can check for that kind of Event by running
kubectl get events -A --field-selector=reason=OwnerRefInvalidNamespace.
Cascading deletion
Kubernetes checks for and deletes objects that no longer have owner
references, like the pods left behind when you delete a ReplicaSet. When you
delete an object, you can control whether Kubernetes deletes the object's
dependents automatically, in a process called cascading deletion. There are
two types of cascading deletion, as follows:
Foreground cascading deletion
Background cascading deletion
You can also control how and when garbage collection deletes resources that have
owner references using Kubernetes finalizers.
Foreground cascading deletion
In foreground cascading deletion, the owner object you're deleting first enters
a deletion in progress state. In this state, the following happens to the
owner object:
The Kubernetes API server sets the object's metadata.deletionTimestamp
field to the time the object was marked for deletion.
The Kubernetes API server also sets the metadata.finalizers field to
foregroundDeletion.
The object remains visible through the Kubernetes API until the deletion
process is complete.
After the owner object enters the deletion in progress state, the controller
deletes the dependents. After deleting all the dependent objects, the controller
deletes the owner object. At this point, the object is no longer visible in the
Kubernetes API.
During foreground cascading deletion, the only dependents that block owner
deletion are those that have the ownerReference.blockOwnerDeletion=true field.
See Use foreground cascading deletion
to learn more.
Background cascading deletion
In background cascading deletion, the Kubernetes API server deletes the owner
object immediately and the controller cleans up the dependent objects in
the background. By default, Kubernetes uses background cascading deletion unless
you manually use foreground deletion or choose to orphan the dependent objects.
When Kubernetes deletes an owner object, the dependents left behind are called
orphan objects. By default, Kubernetes deletes dependent objects. To learn how
to override this behaviour, see Delete owner objects and orphan dependents.
Garbage collection of unused containers and images
The kubelet performs garbage
collection on unused images every five minutes and on unused containers every
minute. You should avoid using external garbage collection tools, as these can
break the kubelet behavior and remove containers that should exist.
To configure options for unused container and image garbage collection, tune the
kubelet using a configuration file
and change the parameters related to garbage collection using the
KubeletConfiguration
resource type.
Container image lifecycle
Kubernetes manages the lifecycle of all images through its image manager,
which is part of the kubelet, with the cooperation of
cadvisor. The kubelet
considers the following disk usage limits when making garbage collection
decisions:
HighThresholdPercent
LowThresholdPercent
Disk usage above the configured HighThresholdPercent value triggers garbage
collection, which deletes images in order based on the last time they were used,
starting with the oldest first. The kubelet deletes images
until disk usage reaches the LowThresholdPercent value.
Container garbage collection
The kubelet garbage collects unused containers based on the following variables,
which you can define:
MinAge: the minimum age at which the kubelet can garbage collect a
container. Disable by setting to 0.
MaxPerPodContainer: the maximum number of dead containers each Pod pair
can have. Disable by setting to less than 0.
MaxContainers: the maximum number of dead containers the cluster can have.
Disable by setting to less than 0.
In addition to these variables, the kubelet garbage collects unidentified and
deleted containers, typically starting with the oldest first.
MaxPerPodContainer and MaxContainers may potentially conflict with each other
in situations where retaining the maximum number of containers per Pod
(MaxPerPodContainer) would go outside the allowable total of global dead
containers (MaxContainers). In this situation, the kubelet adjusts
MaxPerPodContainer to address the conflict. A worst-case scenario would be to
downgrade MaxPerPodContainer to 1 and evict the oldest containers.
Additionally, containers owned by pods that have been deleted are removed once
they are older than MinAge.
Note: The kubelet only garbage collects the containers it manages.
Configuring garbage collection
You can tune garbage collection of resources by configuring options specific to
the controllers managing those resources. The following pages show you how to
configure garbage collection:
Learn about the TTL controller (beta) that cleans up finished Jobs.
3.3 - Containers
Technology for packaging an application along with its runtime dependencies.
Each container that you run is repeatable; the standardization from having
dependencies included means that you get the same behavior wherever you
run it.
Containers decouple applications from underlying host infrastructure.
This makes deployment easier in different cloud or OS environments.
Container images
A container image is a ready-to-run
software package, containing everything needed to run an application:
the code and any runtime it requires, application and system libraries,
and default values for any essential settings.
By design, a container is immutable: you cannot change the code of a
container that is already running. If you have a containerized application
and want to make changes, you need to build a new image that includes
the change, then recreate the container to start from the updated image.
Container runtimes
The container runtime is the software that is responsible for running containers.
A container image represents binary data that encapsulates an application and all its
software dependencies. Container images are executable software bundles that can run
standalone and that make very well defined assumptions about their runtime environment.
You typically create a container image of your application and push it to a registry
before referring to it in a
Pod
This page provides an outline of the container image concept.
Image names
Container images are usually given a name such as pause, example/mycontainer, or kube-apiserver.
Images can also include a registry hostname; for example: fictional.registry.example/imagename,
and possibly a port number as well; for example: fictional.registry.example:10443/imagename.
If you don't specify a registry hostname, Kubernetes assumes that you mean the Docker public registry.
After the image name part you can add a tag (in the same way you would when using with commands like docker or podman).
Tags let you identify different versions of the same series of images.
Image tags consist of lowercase and uppercase letters, digits, underscores (_),
periods (.), and dashes (-).
There are additional rules about where you can place the separator
characters (_, -, and .) inside an image tag.
If you don't specify a tag, Kubernetes assumes you mean the tag latest.
Updating images
When you first create a Deployment,
StatefulSet, Pod, or other
object that includes a Pod template, then by default the pull policy of all
containers in that pod will be set to IfNotPresent if it is not explicitly
specified. This policy causes the
kubelet to skip pulling an
image if it already exists.
Image pull policy
The imagePullPolicy for a container and the tag of the image affect when the
kubelet attempts to pull (download) the specified image.
Here's a list of the values you can set for imagePullPolicy and the effects
these values have:
IfNotPresent
the image is pulled only if it is not already present locally.
Always
every time the kubelet launches a container, the kubelet queries the container
image registry to resolve the name to an image
digest. If the kubelet has a
container image with that exact digest cached locally, the kubelet uses its cached
image; otherwise, the kubelet pulls the image with the resolved digest,
and uses that image to launch the container.
Never
the kubelet does not try fetching the image. If the image is somehow already present
locally, the kubelet attempts to start the container; otherwise, startup fails.
See pre-pulled images for more details.
The caching semantics of the underlying image provider make even
imagePullPolicy: Always efficient, as long as the registry is reliably accessible.
Your container runtime can notice that the image layers already exist on the node
so that they don't need to be downloaded again.
Note:
You should avoid using the :latest tag when deploying containers in production as
it is harder to track which version of the image is running and more difficult to
roll back properly.
Instead, specify a meaningful tag such as v1.42.0.
To make sure the Pod always uses the same version of a container image, you can specify
the image's digest;
replace <image-name>:<tag> with <image-name>@<digest>
(for example, image@sha256:45b23dee08af5e43a7fea6c4cf9c25ccf269ee113168c19722f87876677c5cb2).
When using image tags, if the image registry were to change the code that the tag on that image represents, you might end up with a mix of Pods running the old and new code. An image digest uniquely identifies a specific version of the image, so Kubernetes runs the same code every time it starts a container with that image name and digest specified. Specifying an image by digest fixes the code that you run so that a change at the registry cannot lead to that mix of versions.
There are third-party admission controllers
that mutate Pods (and pod templates) when they are created, so that the
running workload is defined based on an image digest rather than a tag.
That might be useful if you want to make sure that all your workload is
running the same code no matter what tag changes happen at the registry.
Default image pull policy
When you (or a controller) submit a new Pod to the API server, your cluster sets the
imagePullPolicy field when specific conditions are met:
if you omit the imagePullPolicy field, and the tag for the container image is
:latest, imagePullPolicy is automatically set to Always;
if you omit the imagePullPolicy field, and you don't specify the tag for the
container image, imagePullPolicy is automatically set to Always;
if you omit the imagePullPolicy field, and you specify the tag for the
container image that isn't :latest, the imagePullPolicy is automatically set to
IfNotPresent.
Note:
The value of imagePullPolicy of the container is always set when the object is
first created, and is not updated if the image's tag later changes.
For example, if you create a Deployment with an image whose tag is not:latest, and later update that Deployment's image to a :latest tag, the
imagePullPolicy field will not change to Always. You must manually change
the pull policy of any object after its initial creation.
Required image pull
If you would like to always force a pull, you can do one of the following:
Set the imagePullPolicy of the container to Always.
Omit the imagePullPolicy and use :latest as the tag for the image to use;
Kubernetes will set the policy to Always when you submit the Pod.
Omit the imagePullPolicy and the tag for the image to use;
Kubernetes will set the policy to Always when you submit the Pod.
When a kubelet starts creating containers for a Pod using a container runtime,
it might be possible the container is in Waiting
state because of ImagePullBackOff.
The status ImagePullBackOff means that a container could not start because Kubernetes
could not pull a container image (for reasons such as invalid image name, or pulling
from a private registry without imagePullSecret). The BackOff part indicates
that Kubernetes will keep trying to pull the image, with an increasing back-off delay.
Kubernetes raises the delay between each attempt until it reaches a compiled-in limit,
which is 300 seconds (5 minutes).
Multi-architecture images with image indexes
As well as providing binary images, a container registry can also serve a container image index. An image index can point to multiple image manifests for architecture-specific versions of a container. The idea is that you can have a name for an image (for example: pause, example/mycontainer, kube-apiserver) and allow different systems to fetch the right binary image for the machine architecture they are using.
Kubernetes itself typically names container images with a suffix -$(ARCH). For backward compatibility, please generate the older images with suffixes. The idea is to generate say pause image which has the manifest for all the arch(es) and say pause-amd64 which is backwards compatible for older configurations or YAML files which may have hard coded the images with suffixes.
Using a private registry
Private registries may require keys to read images from them.
Credentials can be provided in several ways:
Configuring Nodes to Authenticate to a Private Registry
all pods can read any configured private registries
requires node configuration by cluster administrator
Pre-pulled Images
all pods can use any images cached on a node
requires root access to all nodes to set up
Specifying ImagePullSecrets on a Pod
only pods which provide own keys can access the private registry
Vendor-specific or local extensions
if you're using a custom node configuration, you (or your cloud
provider) can implement your mechanism for authenticating the node
to the container registry.
These options are explained in more detail below.
Configuring nodes to authenticate to a private registry
Specific instructions for setting credentials depends on the container runtime and registry you chose to use. You should refer to your solution's documentation for the most accurate information.
For an example of configuring a private container image registry, see the
Pull an Image from a Private Registry
task. That example uses a private registry in Docker Hub.
Interpretation of config.json
The interpretation of config.json varies between the original Docker
implementation and the Kubernetes interpretation. In Docker, the auths keys
can only specify root URLs, whereas Kubernetes allows glob URLs as well as
prefix-matched paths. This means that a config.json like this is valid:
The root URL (*my-registry.io) is matched by using the following syntax:
pattern:
{ term }
term:
'*' matches any sequence of non-Separator characters
'?' matches any single non-Separator character
'[' [ '^' ] { character-range } ']'
character class (must be non-empty)
c matches character c (c != '*', '?', '\\', '[')
'\\' c matches character c
character-range:
c matches character c (c != '\\', '-', ']')
'\\' c matches character c
lo '-' hi matches character c for lo <= c <= hi
Image pull operations would now pass the credentials to the CRI container
runtime for every valid pattern. For example the following container image names
would match successfully:
my-registry.io/images
my-registry.io/images/my-image
my-registry.io/images/another-image
sub.my-registry.io/images/my-image
a.sub.my-registry.io/images/my-image
The kubelet performs image pulls sequentially for every found credential. This
means, that multiple entries in config.json are possible, too:
If now a container specifies an image my-registry.io/images/subpath/my-image
to be pulled, then the kubelet will try to download them from both
authentication sources if one of them fails.
Pre-pulled images
Note: This approach is suitable if you can control node configuration. It
will not work reliably if your cloud provider manages nodes and replaces
them automatically.
By default, the kubelet tries to pull each image from the specified registry.
However, if the imagePullPolicy property of the container is set to IfNotPresent or Never,
then a local image is used (preferentially or exclusively, respectively).
If you want to rely on pre-pulled images as a substitute for registry authentication,
you must ensure all nodes in the cluster have the same pre-pulled images.
This can be used to preload certain images for speed or as an alternative to authenticating to a private registry.
All pods will have read access to any pre-pulled images.
Specifying imagePullSecrets on a Pod
Note: This is the recommended approach to run containers based on images
in private registries.
Kubernetes supports specifying container image registry keys on a Pod.
imagePullSecrets must all be in the same namespace as the Pod. The referenced
Secrets must be of type kubernetes.io/dockercfg or kubernetes.io/dockerconfigjson.
Creating a Secret with a Docker config
You need to know the username, registry password and client email address for authenticating
to the registry, as well as its hostname.
Run the following command, substituting the appropriate uppercase values:
This is particularly useful if you are using multiple private container
registries, as kubectl create secret docker-registry creates a Secret that
only works with a single private registry.
Note: Pods can only reference image pull secrets in their own namespace,
so this process needs to be done one time per namespace.
Referring to an imagePullSecrets on a Pod
Now, you can create pods which reference that secret by adding an imagePullSecrets
section to a Pod definition. Each item in the imagePullSecrets array can only
reference a Secret in the same namespace.
This page describes the resources available to Containers in the Container environment.
Container environment
The Kubernetes Container environment provides several important resources to Containers:
A filesystem, which is a combination of an image and one or more volumes.
Information about the Container itself.
Information about other objects in the cluster.
Container information
The hostname of a Container is the name of the Pod in which the Container is running.
It is available through the hostname command or the
gethostname
function call in libc.
The Pod name and namespace are available as environment variables through the
downward API.
User defined environment variables from the Pod definition are also available to the Container,
as are any environment variables specified statically in the container image.
Cluster information
A list of all services that were running when a Container was created is available to that Container as environment variables.
This list is limited to services within the same namespace as the new Container's Pod and Kubernetes control plane services.
For a service named foo that maps to a Container named bar,
the following variables are defined:
FOO_SERVICE_HOST=<the host the service is running on>
FOO_SERVICE_PORT=<the port the service is running on>
Services have dedicated IP addresses and are available to the Container via DNS,
if DNS addon is enabled.
This page describes the RuntimeClass resource and runtime selection mechanism.
RuntimeClass is a feature for selecting the container runtime configuration. The container runtime
configuration is used to run a Pod's containers.
Motivation
You can set a different RuntimeClass between different Pods to provide a balance of
performance versus security. For example, if part of your workload deserves a high
level of information security assurance, you might choose to schedule those Pods so
that they run in a container runtime that uses hardware virtualization. You'd then
benefit from the extra isolation of the alternative runtime, at the expense of some
additional overhead.
You can also use RuntimeClass to run different Pods with the same container runtime
but with different settings.
Setup
Configure the CRI implementation on nodes (runtime dependent)
Create the corresponding RuntimeClass resources
1. Configure the CRI implementation on nodes
The configurations available through RuntimeClass are Container Runtime Interface (CRI)
implementation dependent. See the corresponding documentation (below) for your
CRI implementation for how to configure.
Note: RuntimeClass assumes a homogeneous node configuration across the cluster by default (which means
that all nodes are configured the same way with respect to container runtimes). To support
heterogeneous node configurations, see Scheduling below.
The configurations have a corresponding handler name, referenced by the RuntimeClass. The
handler must be a valid DNS label name.
2. Create the corresponding RuntimeClass resources
The configurations setup in step 1 should each have an associated handler name, which identifies
the configuration. For each handler, create a corresponding RuntimeClass object.
The RuntimeClass resource currently only has 2 significant fields: the RuntimeClass name
(metadata.name) and the handler (handler). The object definition looks like this:
# RuntimeClass is defined in the node.k8s.io API groupapiVersion:node.k8s.io/v1kind:RuntimeClassmetadata:# The name the RuntimeClass will be referenced by.# RuntimeClass is a non-namespaced resource.name:myclass # The name of the corresponding CRI configurationhandler:myconfiguration
Note: It is recommended that RuntimeClass write operations (create/update/patch/delete) be
restricted to the cluster administrator. This is typically the default. See
Authorization Overview for more details.
Usage
Once RuntimeClasses are configured for the cluster, you can specify a
runtimeClassName in the Pod spec to use it. For example:
This will instruct the kubelet to use the named RuntimeClass to run this pod. If the named
RuntimeClass does not exist, or the CRI cannot run the corresponding handler, the pod will enter the
Failed terminal phase. Look for a
corresponding event for an
error message.
If no runtimeClassName is specified, the default RuntimeHandler will be used, which is equivalent
to the behavior when the RuntimeClass feature is disabled.
CRI Configuration
For more details on setting up CRI runtimes, see CRI installation.
Runtime handlers are configured through containerd's configuration at
/etc/containerd/config.toml. Valid handlers are configured under the runtimes section:
By specifying the scheduling field for a RuntimeClass, you can set constraints to
ensure that Pods running with this RuntimeClass are scheduled to nodes that support it.
If scheduling is not set, this RuntimeClass is assumed to be supported by all nodes.
To ensure pods land on nodes supporting a specific RuntimeClass, that set of nodes should have a
common label which is then selected by the runtimeclass.scheduling.nodeSelector field. The
RuntimeClass's nodeSelector is merged with the pod's nodeSelector in admission, effectively taking
the intersection of the set of nodes selected by each. If there is a conflict, the pod will be
rejected.
If the supported nodes are tainted to prevent other RuntimeClass pods from running on the node, you
can add tolerations to the RuntimeClass. As with the nodeSelector, the tolerations are merged
with the pod's tolerations in admission, effectively taking the union of the set of nodes tolerated
by each.
To learn more about configuring the node selector and tolerations, see
Assigning Pods to Nodes.
Pod Overhead
FEATURE STATE:Kubernetes v1.24 [stable]
You can specify overhead resources that are associated with running a Pod. Declaring overhead allows
the cluster (including the scheduler) to account for it when making decisions about Pods and resources.
Pod overhead is defined in RuntimeClass through the overhead field. Through the use of this field,
you can specify the overhead of running pods utilizing this RuntimeClass and ensure these overheads
are accounted for in Kubernetes.
This page describes how kubelet managed Containers can use the Container lifecycle hook framework
to run code triggered by events during their management lifecycle.
Overview
Analogous to many programming language frameworks that have component lifecycle hooks, such as Angular,
Kubernetes provides Containers with lifecycle hooks.
The hooks enable Containers to be aware of events in their management lifecycle
and run code implemented in a handler when the corresponding lifecycle hook is executed.
Container hooks
There are two hooks that are exposed to Containers:
PostStart
This hook is executed immediately after a container is created.
However, there is no guarantee that the hook will execute before the container ENTRYPOINT.
No parameters are passed to the handler.
PreStop
This hook is called immediately before a container is terminated due to an API request or management
event such as a liveness/startup probe failure, preemption, resource contention and others. A call
to the PreStop hook fails if the container is already in a terminated or completed state and the
hook must complete before the TERM signal to stop the container can be sent. The Pod's termination
grace period countdown begins before the PreStop hook is executed, so regardless of the outcome of
the handler, the container will eventually terminate within the Pod's termination grace period. No
parameters are passed to the handler.
A more detailed description of the termination behavior can be found in
Termination of Pods.
Hook handler implementations
Containers can access a hook by implementing and registering a handler for that hook.
There are two types of hook handlers that can be implemented for Containers:
Exec - Executes a specific command, such as pre-stop.sh, inside the cgroups and namespaces of the Container.
Resources consumed by the command are counted against the Container.
HTTP - Executes an HTTP request against a specific endpoint on the Container.
Hook handler execution
When a Container lifecycle management hook is called,
the Kubernetes management system executes the handler according to the hook action,
httpGet and tcpSocket are executed by the kubelet process, and exec is executed in the container.
Hook handler calls are synchronous within the context of the Pod containing the Container.
This means that for a PostStart hook,
the Container ENTRYPOINT and hook fire asynchronously.
However, if the hook takes too long to run or hangs,
the Container cannot reach a running state.
PreStop hooks are not executed asynchronously from the signal to stop the Container; the hook must
complete its execution before the TERM signal can be sent. If a PreStop hook hangs during
execution, the Pod's phase will be Terminating and remain there until the Pod is killed after its
terminationGracePeriodSeconds expires. This grace period applies to the total time it takes for
both the PreStop hook to execute and for the Container to stop normally. If, for example,
terminationGracePeriodSeconds is 60, and the hook takes 55 seconds to complete, and the Container
takes 10 seconds to stop normally after receiving the signal, then the Container will be killed
before it can stop normally, since terminationGracePeriodSeconds is less than the total time
(55+10) it takes for these two things to happen.
If either a PostStart or PreStop hook fails,
it kills the Container.
Users should make their hook handlers as lightweight as possible.
There are cases, however, when long running commands make sense,
such as when saving state prior to stopping a Container.
Hook delivery guarantees
Hook delivery is intended to be at least once,
which means that a hook may be called multiple times for any given event,
such as for PostStart or PreStop.
It is up to the hook implementation to handle this correctly.
Generally, only single deliveries are made.
If, for example, an HTTP hook receiver is down and is unable to take traffic,
there is no attempt to resend.
In some rare cases, however, double delivery may occur.
For instance, if a kubelet restarts in the middle of sending a hook,
the hook might be resent after the kubelet comes back up.
Debugging Hook handlers
The logs for a Hook handler are not exposed in Pod events.
If a handler fails for some reason, it broadcasts an event.
For PostStart, this is the FailedPostStartHook event,
and for PreStop, this is the FailedPreStopHook event.
To generate a failed FailedPostStartHook event yourself, modify the lifecycle-events.yaml file to change the postStart command to "badcommand" and apply it.
Here is some example output of the resulting events you see from running kubectl describe pod lifecycle-demo:
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 7s default-scheduler Successfully assigned default/lifecycle-demo to ip-XXX-XXX-XX-XX.us-east-2...
Normal Pulled 6s kubelet Successfully pulled image "nginx" in 229.604315ms
Normal Pulling 4s (x2 over 6s) kubelet Pulling image "nginx"
Normal Created 4s (x2 over 5s) kubelet Created container lifecycle-demo-container
Normal Started 4s (x2 over 5s) kubelet Started container lifecycle-demo-container
Warning FailedPostStartHook 4s (x2 over 5s) kubelet Exec lifecycle hook ([badcommand]) for Container "lifecycle-demo-container" in Pod "lifecycle-demo_default(30229739-9651-4e5a-9a32-a8f1688862db)" failed - error: command 'badcommand' exited with 126: , message: "OCI runtime exec failed: exec failed: container_linux.go:380: starting container process caused: exec: \"badcommand\": executable file not found in $PATH: unknown\r\n"
Normal Killing 4s (x2 over 5s) kubelet FailedPostStartHook
Normal Pulled 4s kubelet Successfully pulled image "nginx" in 215.66395ms
Warning BackOff 2s (x2 over 3s) kubelet Back-off restarting failed container
Windows applications constitute a large portion of the services and applications that
run in many organizations. Windows containers
provide a way to encapsulate processes and package dependencies, making it easier
to use DevOps practices and follow cloud native patterns for Windows applications.
Organizations with investments in Windows-based applications and Linux-based
applications don't have to look for separate orchestrators to manage their workloads,
leading to increased operational efficiencies across their deployments, regardless
of operating system.
Windows nodes in Kubernetes
To enable the orchestration of Windows containers in Kubernetes, include Windows nodes
in your existing Linux cluster. Scheduling Windows containers in
Pods on Kubernetes is similar to
scheduling Linux-based containers.
In order to run Windows containers, your Kubernetes cluster must include
multiple operating systems.
While you can only run the control plane on Linux,
you can deploy worker nodes running either Windows or Linux.
Windows nodes are
supported provided that the operating system is
Windows Server 2019.
This document uses the term Windows containers to mean Windows containers with
process isolation. Kubernetes does not support running Windows containers with
Hyper-V isolation.
Compatibility and limitations
Some node features are only available if you use a specific
container runtime; others are not available on Windows nodes,
including:
HugePages: not supported for Windows containers
Privileged containers: not supported for Windows containers.
HostProcess Containers offer similar functionality.
TerminationGracePeriod: requires containerD
Not all features of shared namespaces are supported. See API compatibility
for more details.
From an API and kubectl perspective, Windows containers behave in much the same
way as Linux-based containers. However, there are some notable differences in key
functionality which are outlined in this section.
Comparison with Linux
Key Kubernetes elements work the same way in Windows as they do in Linux. This
section refers to several key workload abstractions and how they map to Windows.
A Pod is the basic building block of Kubernetes–the smallest and simplest unit in
the Kubernetes object model that you create or deploy. You may not deploy Windows and
Linux containers in the same Pod. All containers in a Pod are scheduled onto a single
Node where each Node represents a specific platform and architecture. The following
Pod capabilities, properties and events are supported with Windows containers:
Single or multiple containers per Pod with process isolation and volume sharing
Pod status fields
Readiness, liveness, and startup probes
postStart & preStop container lifecycle hooks
ConfigMap, Secrets: as environment variables or volumes
emptyDir volumes
Named pipe host mounts
Resource limits
OS field:
The .spec.os.name field should be set to windows to indicate that the current Pod uses Windows containers.
Note: Starting from 1.25, the IdentifyPodOS feature gate is in GA stage and defaults to be enabled.
If you set the .spec.os.name field to windows,
you must not set the following fields in the .spec of that Pod:
In the above list, wildcards (*) indicate all elements in a list.
For example, spec.containers[*].securityContext refers to the SecurityContext object
for all containers. If any of these fields is specified, the Pod will
not be admitted by the API server.
Pods, workload resources, and Services are critical elements to managing Windows
workloads on Kubernetes. However, on their own they are not enough to enable
the proper lifecycle management of Windows workloads in a dynamic cloud native
environment.
Some kubelet command line options behave differently on Windows, as described below:
The --windows-priorityclass lets you set the scheduling priority of the kubelet process
(see CPU resource management)
The --kube-reserved, --system-reserved , and --eviction-hard flags update
NodeAllocatable
Eviction by using --enforce-node-allocable is not implemented
Eviction by using --eviction-hard and --eviction-soft are not implemented
When running on a Windows node the kubelet does not have memory or CPU
restrictions. --kube-reserved and --system-reserved only subtract from NodeAllocatable
and do not guarantee resource provided for workloads.
See Resource Management for Windows nodes
for more information.
The MemoryPressure Condition is not implemented
The kubelet does not take OOM eviction actions
API compatibility
There are subtle differences in the way the Kubernetes APIs work for Windows due to the OS
and container runtime. Some workload properties were designed for Linux, and fail to run on Windows.
At a high level, these OS concepts are different:
Identity - Linux uses userID (UID) and groupID (GID) which
are represented as integer types. User and group names
are not canonical - they are just an alias in /etc/groups
or /etc/passwd back to UID+GID. Windows uses a larger binary
security identifier (SID)
which is stored in the Windows Security Access Manager (SAM) database. This
database is not shared between the host and containers, or between containers.
File permissions - Windows uses an access control list based on (SIDs), whereas
POSIX systems such as Linux use a bitmask based on object permissions and UID+GID,
plus optional access control lists.
File paths - the convention on Windows is to use \ instead of /. The Go IO
libraries typically accept both and just make it work, but when you're setting a
path or command line that's interpreted inside a container, \ may be needed.
Signals - Windows interactive apps handle termination differently, and can
implement one or more of these:
A UI thread handles well-defined messages including WM_CLOSE.
Console apps handle Ctrl-C or Ctrl-break using a Control Handler.
Services register a Service Control Handler function that can accept
SERVICE_CONTROL_STOP control codes.
Container exit codes follow the same convention where 0 is success, and nonzero is failure.
The specific error codes may differ across Windows and Linux. However, exit codes
passed from the Kubernetes components (kubelet, kube-proxy) are unchanged.
Field compatibility for container specifications
The following list documents differences between how Pod container specifications
work between Windows and Linux:
Huge pages are not implemented in the Windows container
runtime, and are not available. They require asserting a user
privilege
that's not configurable for containers.
requests.cpu and requests.memory - requests are subtracted
from node available resources, so they can be used to avoid overprovisioning a
node. However, they cannot be used to guarantee resources in an overprovisioned
node. They should be applied to all containers as a best practice if the operator
wants to avoid overprovisioning entirely.
securityContext.allowPrivilegeEscalation -
not possible on Windows; none of the capabilities are hooked up
securityContext.capabilities -
POSIX capabilities are not implemented on Windows
securityContext.privileged -
Windows doesn't support privileged containers
securityContext.procMount -
Windows doesn't have a /proc filesystem
securityContext.readOnlyRootFilesystem -
not possible on Windows; write access is required for registry & system
processes to run inside the container
securityContext.runAsGroup -
not possible on Windows as there is no GID support
securityContext.runAsNonRoot -
this setting will prevent containers from running as ContainerAdministrator
which is the closest equivalent to a root user on Windows.
securityContext.runAsUser -
use runAsUserName
instead
securityContext.seLinuxOptions -
not possible on Windows as SELinux is Linux-specific
terminationMessagePath -
this has some limitations in that Windows doesn't support mapping single files. The
default value is /dev/termination-log, which does work because it does not
exist on Windows by default.
Field compatibility for Pod specifications
The following list documents differences between how Pod specifications work between Windows and Linux:
hostIPC and hostpid - host namespace sharing is not possible on Windows
hostNetwork - There is no Windows OS support to share the host network
dnsPolicy - setting the Pod dnsPolicy to ClusterFirstWithHostNet is
not supported on Windows because host networking is not provided. Pods always
run with a container network.
podSecurityContext (see below)
shareProcessNamespace - this is a beta feature, and depends on Linux namespaces
which are not implemented on Windows. Windows cannot share process namespaces or
the container's root filesystem. Only the network can be shared.
terminationGracePeriodSeconds - this is not fully implemented in Docker on Windows,
see the GitHub issue.
The behavior today is that the ENTRYPOINT process is sent CTRL_SHUTDOWN_EVENT,
then Windows waits 5 seconds by default, and finally shuts down
all processes using the normal Windows shutdown behavior. The 5
second default is actually in the Windows registry
inside the container,
so it can be overridden when the container is built.
volumeDevices - this is a beta feature, and is not implemented on Windows.
Windows cannot attach raw block devices to pods.
volumes
If you define an emptyDir volume, you cannot set its volume source to memory.
You cannot enable mountPropagation for volume mounts as this is not
supported on Windows.
The node problem detector (see
Monitor Node Health)
has preliminary support for Windows.
For more information, visit the project's GitHub page.
Pause container
In a Kubernetes Pod, an infrastructure or “pause” container is first created
to host the container. In Linux, the cgroups and namespaces that make up a pod
need a process to maintain their continued existence; the pause process provides
this. Containers that belong to the same pod, including infrastructure and worker
containers, share a common network endpoint (same IPv4 and / or IPv6 address, same
network port spaces). Kubernetes uses pause containers to allow for worker containers
crashing or restarting without losing any of the networking configuration.
Kubernetes maintains a multi-architecture image that includes support for Windows.
For Kubernetes v1.25 the recommended pause image is registry.k8s.io/pause:3.6.
The source code
is available on GitHub.
Microsoft maintains a different multi-architecture image, with Linux and Windows
amd64 support, that you can find as mcr.microsoft.com/oss/kubernetes/pause:3.6.
This image is built from the same source as the Kubernetes maintained image but
all of the Windows binaries are authenticode signed by Microsoft.
The Kubernetes project recommends using the Microsoft maintained image if you are
deploying to a production or production-like environment that requires signed
binaries.
Container runtimes
You need to install a
container runtime
into each node in the cluster so that Pods can run there.
The following container runtimes work with Windows:
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
ContainerD
FEATURE STATE:Kubernetes v1.20 [stable]
You can use ContainerD 1.4.0+
as the container runtime for Kubernetes nodes that run Windows.
On Windows nodes, strict compatibility rules apply where the host OS version must
match the container base image OS version. Only Windows containers with a container
operating system of Windows Server 2019 are fully supported.
For Kubernetes v1.25, operating system compatibility for Windows nodes (and Pods)
is as follows:
Your main source of help for troubleshooting your Kubernetes cluster should start
with the Troubleshooting
page.
Some additional, Windows-specific troubleshooting help is included
in this section. Logs are an important element of troubleshooting
issues in Kubernetes. Make sure to include them any time you seek
troubleshooting assistance from other contributors. Follow the
instructions in the
SIG Windows contributing guide on gathering logs.
Reporting issues and feature requests
If you have what looks like a bug, or you would like to
make a feature request, please follow the SIG Windows contributing guide to create a new issue.
You should first search the list of issues in case it was
reported previously and comment with your experience on the issue and add additional
logs. SIG Windows channel on the Kubernetes Slack is also a great avenue to get some initial support and
troubleshooting ideas prior to creating a ticket.
Deployment tools
The kubeadm tool helps you to deploy a Kubernetes cluster, providing the control
plane to manage the cluster it, and nodes to run your workloads.
Adding Windows nodes
explains how to deploy Windows nodes to your cluster using kubeadm.
The Kubernetes cluster API project also provides means to automate deployment of Windows nodes.
Windows distribution channels
For a detailed explanation of Windows distribution channels see the
Microsoft documentation.
Information on the different Windows Server servicing channels
including their support models can be found at
Windows Server servicing channels.
3.4.2 - Guide for scheduling Windows containers in Kubernetes
Windows applications constitute a large portion of the services and applications that run in many organizations.
This guide walks you through the steps to configure and deploy Windows containers in Kubernetes.
Objectives
Configure an example deployment to run Windows containers on the Windows node
Highlight Windows specific funcationality in Kubernetes
It is important to note that creating and deploying services and workloads on Kubernetes
behaves in much the same way for Linux and Windows containers.
Kubectl commands to interface with the cluster are identical.
The example in the section below is provided to jumpstart your experience with Windows containers.
Getting Started: Deploying a Windows container
The example YAML file below deploys a simple webserver application running inside a Windows container.
Create a service spec named win-webserver.yaml with the contents below:
Note: Port mapping is also supported, but for simplicity this example exposes
port 80 of the container directly to the Service.
Check that all nodes are healthy:
kubectl get nodes
Deploy the service and watch for pod updates:
kubectl apply -f win-webserver.yaml
kubectl get pods -o wide -w
When the service is deployed correctly both Pods are marked as Ready. To exit the watch command, press Ctrl+C.
Check that the deployment succeeded. To verify:
Two pods listed from the Linux control plane node, use kubectl get pods
Node-to-pod communication across the network, curl port 80 of your pod IPs from the Linux control plane node
to check for a web server response
Pod-to-pod communication, ping between pods (and across hosts, if you have more than one Windows node)
using docker exec or kubectl exec
Service-to-pod communication, curl the virtual service IP (seen under kubectl get services)
from the Linux control plane node and from individual pods
Service discovery, curl the service name with the Kubernetes default DNS suffix
Inbound connectivity, curl the NodePort from the Linux control plane node or machines outside of the cluster
Outbound connectivity, curl external IPs from inside the pod using kubectl exec
Note: Windows container hosts are not able to access the IP of services scheduled on them due to current platform limitations of the Windows networking stack.
Only Windows pods are able to access service IPs.
Observability
Capturing logs from workloads
Logs are an important element of observability; they enable users to gain insights
into the operational aspect of workloads and are a key ingredient to troubleshooting issues.
Because Windows containers and workloads inside Windows containers behave differently from Linux containers,
users had a hard time collecting logs, limiting operational visibility.
Windows workloads for example are usually configured to log to ETW (Event Tracing for Windows)
or push entries to the application event log.
LogMonitor, an open source tool by Microsoft,
is the recommended way to monitor configured log sources inside a Windows container.
LogMonitor supports monitoring event logs, ETW providers, and custom application logs,
piping them to STDOUT for consumption by kubectl logs <pod>.
Follow the instructions in the LogMonitor GitHub page to copy its binaries and configuration files
to all your containers and add the necessary entrypoints for LogMonitor to push your logs to STDOUT.
Configuring container user
Using configurable Container usernames
Windows containers can be configured to run their entrypoints and processes
with different usernames than the image defaults.
Learn more about it here.
Managing Workload Identity with Group Managed Service Accounts
Windows container workloads can be configured to use Group Managed Service Accounts (GMSA).
Group Managed Service Accounts are a specific type of Active Directory account that provide automatic password management,
simplified service principal name (SPN) management, and the ability to delegate the management to other administrators across multiple servers.
Containers configured with a GMSA can access external Active Directory Domain resources while carrying the identity configured with the GMSA.
Learn more about configuring and using GMSA for Windows containers here.
Taints and Tolerations
Users need to use some combination of taints and node selectors in order to
schedule Linux and Windows workloads to their respective OS-specific nodes.
The recommended approach is outlined below,
with one of its main goals being that this approach should not break compatibility for existing Linux workloads.
Starting from 1.25, you can (and should) set .spec.os.name for each Pod, to indicate the operating system
that the containers in that Pod are designed for. For Pods that run Linux containers, set
.spec.os.name to linux. For Pods that run Windows containers, set .spec.os.name
to windows.
Note: Starting from 1.25, the IdentifyPodOS feature is in GA stage and defaults to be enabled.
The scheduler does not use the value of .spec.os.name when assigning Pods to nodes. You should
use normal Kubernetes mechanisms for
assigning pods to nodes
to ensure that the control plane for your cluster places pods onto nodes that are running the
appropriate operating system.
The .spec.os.name value has no effect on the scheduling of the Windows pods,
so taints and tolerations and node selectors are still required
to ensure that the Windows pods land onto appropriate Windows nodes.
Ensuring OS-specific workloads land on the appropriate container host
Users can ensure Windows containers can be scheduled on the appropriate host using Taints and Tolerations.
All Kubernetes nodes today have the following default labels:
kubernetes.io/os = [windows|linux]
kubernetes.io/arch = [amd64|arm64|...]
If a Pod specification does not specify a nodeSelector like "kubernetes.io/os": windows,
it is possible the Pod can be scheduled on any host, Windows or Linux.
This can be problematic since a Windows container can only run on Windows and a Linux container can only run on Linux.
The best practice is to use a nodeSelector.
However, we understand that in many cases users have a pre-existing large number of deployments for Linux containers,
as well as an ecosystem of off-the-shelf configurations, such as community Helm charts, and programmatic Pod generation cases, such as with Operators.
In those situations, you may be hesitant to make the configuration change to add nodeSelectors.
The alternative is to use Taints. Because the kubelet can set Taints during registration,
it could easily be modified to automatically add a taint when running on Windows only.
For example: --register-with-taints='os=windows:NoSchedule'
By adding a taint to all Windows nodes, nothing will be scheduled on them (that includes existing Linux Pods).
In order for a Windows Pod to be scheduled on a Windows node,
it would need both the nodeSelector and the appropriate matching toleration to choose Windows.
Handling multiple Windows versions in the same cluster
The Windows Server version used by each pod must match that of the node. If you want to use multiple Windows
Server versions in the same cluster, then you should set additional node labels and nodeSelectors.
Kubernetes 1.17 automatically adds a new label node.kubernetes.io/windows-build to simplify this.
If you're running an older version, then it's recommended to add this label manually to Windows nodes.
This label reflects the Windows major, minor, and build number that need to match for compatibility.
Here are values used today for each Windows Server version.
Product Name
Build Number(s)
Windows Server 2019
10.0.17763
Windows Server, Version 20H2
10.0.19042
Windows Server 2022
10.0.20348
Simplifying with RuntimeClass
RuntimeClass can be used to simplify the process of using taints and tolerations.
A cluster administrator can create a RuntimeClass object which is used to encapsulate these taints and tolerations.
Save this file to runtimeClasses.yml. It includes the appropriate nodeSelector
for the Windows OS, architecture, and version.
Understand Pods, the smallest deployable compute object in Kubernetes, and the higher-level abstractions that help you to run them.
A workload is an application running on Kubernetes.
Whether your workload is a single component or several that work together, on Kubernetes you run
it inside a set of pods.
In Kubernetes, a Pod represents a set of running
containers on your cluster.
Kubernetes pods have a defined lifecycle.
For example, once a pod is running in your cluster then a critical fault on the
node where that pod is running means that
all the pods on that node fail. Kubernetes treats that level of failure as final: you
would need to create a new Pod to recover, even if the node later becomes healthy.
However, to make life considerably easier, you don't need to manage each Pod directly.
Instead, you can use workload resources that manage a set of pods on your behalf.
These resources configure controllers
that make sure the right number of the right kind of pod are running, to match the state
you specified.
Kubernetes provides several built-in workload resources:
Deployment and ReplicaSet
(replacing the legacy resource
ReplicationController).
Deployment is a good fit for managing a stateless application workload on your cluster,
where any Pod in the Deployment is interchangeable and can be replaced if needed.
StatefulSet lets you
run one or more related Pods that do track state somehow. For example, if your workload
records data persistently, you can run a StatefulSet that matches each Pod with a
PersistentVolume. Your code, running in the
Pods for that StatefulSet, can replicate data to other Pods in the same StatefulSet
to improve overall resilience.
DaemonSet defines Pods that provide
node-local facilities. These might be fundamental to the operation of your cluster, such
as a networking helper tool, or be part of an
add-on.
Every time you add a node to your cluster that matches the specification in a DaemonSet,
the control plane schedules a Pod for that DaemonSet onto the new node.
Job and
CronJob
define tasks that run to completion and then stop. Jobs represent one-off tasks, whereas
CronJobs recur according to a schedule.
In the wider Kubernetes ecosystem, you can find third-party workload resources that provide
additional behaviors. Using a
custom resource definition,
you can add in a third-party workload resource if you want a specific behavior that's not part
of Kubernetes' core. For example, if you wanted to run a group of Pods for your application but
stop work unless all the Pods are available (perhaps for some high-throughput distributed task),
then you can implement or install an extension that does provide that feature.
What's next
As well as reading about each resource, you can learn about specific tasks that relate to them:
Once your application is running, you might want to make it available on the internet as
a Service or, for web application only,
using an Ingress.
3.5.1 - Pods
Pods are the smallest deployable units of computing that you can create and manage in Kubernetes.
A Pod (as in a pod of whales or pea pod) is a group of one or more
containers, with shared storage and network resources, and a specification for how to run the containers. A Pod's contents are always co-located and
co-scheduled, and run in a shared context. A Pod models an
application-specific "logical host": it contains one or more application
containers which are relatively tightly coupled.
In non-cloud contexts, applications executed on the same physical or virtual machine are analogous to cloud applications executed on the same logical host.
As well as application containers, a Pod can contain
init containers that run
during Pod startup. You can also inject
ephemeral containers
for debugging if your cluster offers this.
What is a Pod?
Note: While Kubernetes supports more
container runtimes
than just Docker, Docker is the most commonly known
runtime, and it helps to describe Pods using some terminology from Docker.
The shared context of a Pod is a set of Linux namespaces, cgroups, and
potentially other facets of isolation - the same things that isolate a Docker
container. Within a Pod's context, the individual applications may have
further sub-isolations applied.
In terms of Docker concepts, a Pod is similar to a group of Docker containers
with shared namespaces and shared filesystem volumes.
Using Pods
The following is an example of a Pod which consists of a container running the image nginx:1.14.2.
Pods are generally not created directly and are created using workload resources.
See Working with Pods for more information on how Pods are used
with workload resources.
Workload resources for managing pods
Usually you don't need to create Pods directly, even singleton Pods. Instead, create them using workload resources such as Deployment or Job.
If your Pods need to track state, consider the
StatefulSet resource.
Pods in a Kubernetes cluster are used in two main ways:
Pods that run a single container. The "one-container-per-Pod" model is the
most common Kubernetes use case; in this case, you can think of a Pod as a
wrapper around a single container; Kubernetes manages Pods rather than managing
the containers directly.
Pods that run multiple containers that need to work together. A Pod can
encapsulate an application composed of multiple co-located containers that are
tightly coupled and need to share resources. These co-located containers
form a single cohesive unit of service—for example, one container serving data
stored in a shared volume to the public, while a separate sidecar container
refreshes or updates those files.
The Pod wraps these containers, storage resources, and an ephemeral network
identity together as a single unit.
Note: Grouping multiple co-located and co-managed containers in a single Pod is a
relatively advanced use case. You should use this pattern only in specific
instances in which your containers are tightly coupled.
Each Pod is meant to run a single instance of a given application. If you want to
scale your application horizontally (to provide more overall resources by running
more instances), you should use multiple Pods, one for each instance. In
Kubernetes, this is typically referred to as replication.
Replicated Pods are usually created and managed as a group by a workload resource
and its controller.
See Pods and controllers for more information on how
Kubernetes uses workload resources, and their controllers, to implement application
scaling and auto-healing.
How Pods manage multiple containers
Pods are designed to support multiple cooperating processes (as containers) that form
a cohesive unit of service. The containers in a Pod are automatically co-located and
co-scheduled on the same physical or virtual machine in the cluster. The containers
can share resources and dependencies, communicate with one another, and coordinate
when and how they are terminated.
For example, you might have a container that
acts as a web server for files in a shared volume, and a separate "sidecar" container
that updates those files from a remote source, as in the following diagram:
Some Pods have init containers as well as app containers. Init containers run and complete before the app containers are started.
Pods natively provide two kinds of shared resources for their constituent containers:
networking and storage.
Working with Pods
You'll rarely create individual Pods directly in Kubernetes—even singleton Pods. This
is because Pods are designed as relatively ephemeral, disposable entities. When
a Pod gets created (directly by you, or indirectly by a
controller), the new Pod is
scheduled to run on a Node in your cluster.
The Pod remains on that node until the Pod finishes execution, the Pod object is deleted,
the Pod is evicted for lack of resources, or the node fails.
Note: Restarting a container in a Pod should not be confused with restarting a Pod. A Pod
is not a process, but an environment for running container(s). A Pod persists until
it is deleted.
When you create the manifest for a Pod object, make sure the name specified is a valid
DNS subdomain name.
Pod OS
FEATURE STATE:Kubernetes v1.25 [stable]
You should set the .spec.os.name field to either windows or linux to indicate the OS on
which you want the pod to run. These two are the only operating systems supported for now by
Kubernetes. In future, this list may be expanded.
In Kubernetes v1.25, the value you set for this field has no
effect on scheduling of the pods.
Setting the .spec.os.name helps to identify the pod OS
authoratitively and is used for validation. The kubelet refuses to run a Pod where you have
specified a Pod OS, if this isn't the same as the operating system for the node where
that kubelet is running.
The Pod security standards also use this
field to avoid enforcing policies that aren't relevant to that operating system.
Pods and controllers
You can use workload resources to create and manage multiple Pods for you. A controller
for the resource handles replication and rollout and automatic healing in case of
Pod failure. For example, if a Node fails, a controller notices that Pods on that
Node have stopped working and creates a replacement Pod. The scheduler places the
replacement Pod onto a healthy Node.
Here are some examples of workload resources that manage one or more Pods:
Controllers for workload resources create Pods
from a pod template and manage those Pods on your behalf.
PodTemplates are specifications for creating Pods, and are included in workload resources such as
Deployments,
Jobs, and
DaemonSets.
Each controller for a workload resource uses the PodTemplate inside the workload
object to make actual Pods. The PodTemplate is part of the desired state of whatever
workload resource you used to run your app.
The sample below is a manifest for a simple Job with a template that starts one
container. The container in that Pod prints a message then pauses.
apiVersion:batch/v1kind:Jobmetadata:name:hellospec:template:# This is the pod templatespec:containers:- name:helloimage:busybox:1.28command:['sh','-c','echo "Hello, Kubernetes!" && sleep 3600']restartPolicy:OnFailure# The pod template ends here
Modifying the pod template or switching to a new pod template has no direct effect
on the Pods that already exist. If you change the pod template for a workload
resource, that resource needs to create replacement Pods that use the updated template.
For example, the StatefulSet controller ensures that the running Pods match the current
pod template for each StatefulSet object. If you edit the StatefulSet to change its pod
template, the StatefulSet starts to create new Pods based on the updated template.
Eventually, all of the old Pods are replaced with new Pods, and the update is complete.
Each workload resource implements its own rules for handling changes to the Pod template.
If you want to read more about StatefulSet specifically, read
Update strategy in the StatefulSet Basics tutorial.
On Nodes, the kubelet does not
directly observe or manage any of the details around pod templates and updates; those
details are abstracted away. That abstraction and separation of concerns simplifies
system semantics, and makes it feasible to extend the cluster's behavior without
changing existing code.
Pod update and replacement
As mentioned in the previous section, when the Pod template for a workload
resource is changed, the controller creates new Pods based on the updated
template instead of updating or patching the existing Pods.
Kubernetes doesn't prevent you from managing Pods directly. It is possible to
update some fields of a running Pod, in place. However, Pod update operations
like
patch, and
replace
have some limitations:
Most of the metadata about a Pod is immutable. For example, you cannot
change the namespace, name, uid, or creationTimestamp fields;
the generation field is unique. It only accepts updates that increment the
field's current value.
If the metadata.deletionTimestamp is set, no new entry can be added to the
metadata.finalizers list.
Pod updates may not change fields other than spec.containers[*].image,
spec.initContainers[*].image, spec.activeDeadlineSeconds or
spec.tolerations. For spec.tolerations, you can only add new entries.
When updating the spec.activeDeadlineSeconds field, two types of updates
are allowed:
setting the unassigned field to a positive number;
updating the field from a positive number to a smaller, non-negative
number.
Resource sharing and communication
Pods enable data sharing and communication among their constituent
containers.
Storage in Pods
A Pod can specify a set of shared storage
volumes. All containers
in the Pod can access the shared volumes, allowing those containers to
share data. Volumes also allow persistent data in a Pod to survive
in case one of the containers within needs to be restarted. See
Storage for more information on how
Kubernetes implements shared storage and makes it available to Pods.
Pod networking
Each Pod is assigned a unique IP address for each address family. Every
container in a Pod shares the network namespace, including the IP address and
network ports. Inside a Pod (and only then), the containers that belong to the Pod
can communicate with one another using localhost. When containers in a Pod communicate
with entities outside the Pod,
they must coordinate how they use the shared network resources (such as ports).
Within a Pod, containers share an IP address and port space, and
can find each other via localhost. The containers in a Pod can also communicate
with each other using standard inter-process communications like SystemV semaphores
or POSIX shared memory. Containers in different Pods have distinct IP addresses
and can not communicate by OS-level IPC without special configuration.
Containers that want to interact with a container running in a different Pod can
use IP networking to communicate.
Containers within the Pod see the system hostname as being the same as the configured
name for the Pod. There's more about this in the networking
section.
Privileged mode for containers
In Linux, any container in a Pod can enable privileged mode using the privileged (Linux) flag on the security context of the container spec. This is useful for containers that want to use operating system administrative capabilities such as manipulating the network stack or accessing hardware devices.
If your cluster has the WindowsHostProcessContainers feature enabled, you can create a Windows HostProcess pod by setting the windowsOptions.hostProcess flag on the security context of the pod spec. All containers in these pods must run as Windows HostProcess containers. HostProcess pods run directly on the host and can also be used to perform administrative tasks as is done with Linux privileged containers.
Note: Your container runtime must support the concept of a privileged container for this setting to be relevant.
Static Pods
Static Pods are managed directly by the kubelet daemon on a specific node,
without the API server
observing them.
Whereas most Pods are managed by the control plane (for example, a
Deployment), for static
Pods, the kubelet directly supervises each static Pod (and restarts it if it fails).
Static Pods are always bound to one Kubelet on a specific node.
The main use for static Pods is to run a self-hosted control plane: in other words,
using the kubelet to supervise the individual control plane components.
The kubelet automatically tries to create a mirror Pod
on the Kubernetes API server for each static Pod.
This means that the Pods running on a node are visible on the API server,
but cannot be controlled from there.
To understand the context for why Kubernetes wraps a common Pod API in other resources (such as StatefulSets or Deployments), you can read about the prior art, including:
This page describes the lifecycle of a Pod. Pods follow a defined lifecycle, starting
in the Pendingphase, moving through Running if at least one
of its primary containers starts OK, and then through either the Succeeded or
Failed phases depending on whether any container in the Pod terminated in failure.
Whilst a Pod is running, the kubelet is able to restart containers to handle some
kind of faults. Within a Pod, Kubernetes tracks different container
states and determines what action to take to make the Pod
healthy again.
In the Kubernetes API, Pods have both a specification and an actual status. The
status for a Pod object consists of a set of Pod conditions.
You can also inject custom readiness information into the
condition data for a Pod, if that is useful to your application.
Pods are only scheduled once in their lifetime.
Once a Pod is scheduled (assigned) to a Node, the Pod runs on that Node until it stops
or is terminated.
Pod lifetime
Like individual application containers, Pods are considered to be relatively
ephemeral (rather than durable) entities. Pods are created, assigned a unique
ID (UID), and scheduled
to nodes where they remain until termination (according to restart policy) or
deletion.
If a Node dies, the Pods scheduled to that node
are scheduled for deletion after a timeout period.
Pods do not, by themselves, self-heal. If a Pod is scheduled to a
node that then fails, the Pod is deleted; likewise, a Pod won't
survive an eviction due to a lack of resources or Node maintenance. Kubernetes uses a
higher-level abstraction, called a
controller, that handles the work of
managing the relatively disposable Pod instances.
A given Pod (as defined by a UID) is never "rescheduled" to a different node; instead,
that Pod can be replaced by a new, near-identical Pod, with even the same name if
desired, but with a different UID.
When something is said to have the same lifetime as a Pod, such as a
volume,
that means that the thing exists as long as that specific Pod (with that exact UID)
exists. If that Pod is deleted for any reason, and even if an identical replacement
is created, the related thing (a volume, in this example) is also destroyed and
created anew.
Pod diagram
A multi-container Pod that contains a file puller and a
web server that uses a persistent volume for shared storage between the containers.
Pod phase
A Pod's status field is a
PodStatus
object, which has a phase field.
The phase of a Pod is a simple, high-level summary of where the Pod is in its
lifecycle. The phase is not intended to be a comprehensive rollup of observations
of container or Pod state, nor is it intended to be a comprehensive state machine.
The number and meanings of Pod phase values are tightly guarded.
Other than what is documented here, nothing should be assumed about Pods that
have a given phase value.
Here are the possible values for phase:
Value
Description
Pending
The Pod has been accepted by the Kubernetes cluster, but one or more of the containers has not been set up and made ready to run. This includes time a Pod spends waiting to be scheduled as well as the time spent downloading container images over the network.
Running
The Pod has been bound to a node, and all of the containers have been created. At least one container is still running, or is in the process of starting or restarting.
Succeeded
All containers in the Pod have terminated in success, and will not be restarted.
Failed
All containers in the Pod have terminated, and at least one container has terminated in failure. That is, the container either exited with non-zero status or was terminated by the system.
Unknown
For some reason the state of the Pod could not be obtained. This phase typically occurs due to an error in communicating with the node where the Pod should be running.
Note: When a Pod is being deleted, it is shown as Terminating by some kubectl commands.
This Terminating status is not one of the Pod phases.
A Pod is granted a term to terminate gracefully, which defaults to 30 seconds.
You can use the flag --force to terminate a Pod by force.
If a node dies or is disconnected from the rest of the cluster, Kubernetes
applies a policy for setting the phase of all Pods on the lost node to Failed.
Container states
As well as the phase of the Pod overall, Kubernetes tracks the state of
each container inside a Pod. You can use
container lifecycle hooks to
trigger events to run at certain points in a container's lifecycle.
Once the scheduler
assigns a Pod to a Node, the kubelet starts creating containers for that Pod
using a container runtime.
There are three possible container states: Waiting, Running, and Terminated.
To check the state of a Pod's containers, you can use
kubectl describe pod <name-of-pod>. The output shows the state for each container
within that Pod.
Each state has a specific meaning:
Waiting
If a container is not in either the Running or Terminated state, it is Waiting.
A container in the Waiting state is still running the operations it requires in
order to complete start up: for example, pulling the container image from a container
image registry, or applying Secret
data.
When you use kubectl to query a Pod with a container that is Waiting, you also see
a Reason field to summarize why the container is in that state.
Running
The Running status indicates that a container is executing without issues. If there
was a postStart hook configured, it has already executed and finished. When you use
kubectl to query a Pod with a container that is Running, you also see information
about when the container entered the Running state.
Terminated
A container in the Terminated state began execution and then either ran to
completion or failed for some reason. When you use kubectl to query a Pod with
a container that is Terminated, you see a reason, an exit code, and the start and
finish time for that container's period of execution.
If a container has a preStop hook configured, this hook runs before the container enters
the Terminated state.
Container restart policy
The spec of a Pod has a restartPolicy field with possible values Always, OnFailure,
and Never. The default value is Always.
The restartPolicy applies to all containers in the Pod. restartPolicy only
refers to restarts of the containers by the kubelet on the same node. After containers
in a Pod exit, the kubelet restarts them with an exponential back-off delay (10s, 20s,
40s, …), that is capped at five minutes. Once a container has executed for 10 minutes
without any problems, the kubelet resets the restart backoff timer for that container.
Pod conditions
A Pod has a PodStatus, which has an array of
PodConditions
through which the Pod has or has not passed. Kubelet manages the following
PodConditions:
PodScheduled: the Pod has been scheduled to a node.
PodHasNetwork: (alpha feature; must be enabled explicitly) the
Pod sandbox has been successfully created and networking configured.
ContainersReady: all containers in the Pod are ready.
Initialized: all init containers
have completed successfully.
Ready: the Pod is able to serve requests and should be added to the load
balancing pools of all matching Services.
Field name
Description
type
Name of this Pod condition.
status
Indicates whether that condition is applicable, with possible values "True", "False", or "Unknown".
lastProbeTime
Timestamp of when the Pod condition was last probed.
lastTransitionTime
Timestamp for when the Pod last transitioned from one status to another.
reason
Machine-readable, UpperCamelCase text indicating the reason for the condition's last transition.
message
Human-readable message indicating details about the last status transition.
Pod readiness
FEATURE STATE:Kubernetes v1.14 [stable]
Your application can inject extra feedback or signals into PodStatus:
Pod readiness. To use this, set readinessGates in the Pod's spec to
specify a list of additional conditions that the kubelet evaluates for Pod readiness.
Readiness gates are determined by the current state of status.condition
fields for the Pod. If Kubernetes cannot find such a condition in the
status.conditions field of a Pod, the status of the condition
is defaulted to "False".
Here is an example:
kind:Pod...spec:readinessGates:- conditionType:"www.example.com/feature-1"status:conditions:- type:Ready # a built in PodConditionstatus:"False"lastProbeTime:nulllastTransitionTime:2018-01-01T00:00:00Z- type:"www.example.com/feature-1"# an extra PodConditionstatus:"False"lastProbeTime:nulllastTransitionTime:2018-01-01T00:00:00ZcontainerStatuses:- containerID:docker://abcd...ready:true...
The Pod conditions you add must have names that meet the Kubernetes label key format.
Status for Pod readiness
The kubectl patch command does not support patching object status.
To set these status.conditions for the pod, applications and
operators should use
the PATCH action.
You can use a Kubernetes client library to
write code that sets custom Pod conditions for Pod readiness.
For a Pod that uses custom conditions, that Pod is evaluated to be ready only
when both the following statements apply:
All containers in the Pod are ready.
All conditions specified in readinessGates are True.
When a Pod's containers are Ready but at least one custom condition is missing or
False, the kubelet sets the Pod's condition to ContainersReady.
Pod network readiness
FEATURE STATE:Kubernetes v1.25 [alpha]
After a Pod gets scheduled on a node, it needs to be admitted by the Kubelet and
have any volumes mounted. Once these phases are complete, the Kubelet works with
a container runtime (using Container runtime interface (CRI)) to set up a
runtime sandbox and configure networking for the Pod. If the
PodHasNetworkConditionfeature gate is enabled,
Kubelet reports whether a pod has reached this initialization milestone through
the PodHasNetwork condition in the status.conditions field of a Pod.
The PodHasNetwork condition is set to False by the Kubelet when it detects a
Pod does not have a runtime sandbox with networking configured. This occurs in
the following scenarios:
Early in the lifecycle of the Pod, when the kubelet has not yet begun to set up a sandbox for the Pod using the container runtime.
Later in the lifecycle of the Pod, when the Pod sandbox has been destroyed due
to either:
the node rebooting, without the Pod getting evicted
for container runtimes that use virtual machines for isolation, the Pod
sandbox virtual machine rebooting, which then requires creating a new sandbox and fresh container network configuration.
The PodHasNetwork condition is set to True by the kubelet after the
successful completion of sandbox creation and network configuration for the Pod
by the runtime plugin. The kubelet can start pulling container images and create
containers after PodHasNetwork condition has been set to True.
For a Pod with init containers, the kubelet sets the Initialized condition to
True after the init containers have successfully completed (which happens
after successful sandbox creation and network configuration by the runtime
plugin). For a Pod without init containers, the kubelet sets the Initialized
condition to True before sandbox creation and network configuration starts.
Container probes
A probe is a diagnostic
performed periodically by the
kubelet
on a container. To perform a diagnostic,
the kubelet either executes code within the container, or makes
a network request.
Check mechanisms
There are four different ways to check a container using a probe.
Each probe must define exactly one of these four mechanisms:
exec
Executes a specified command inside the container. The diagnostic
is considered successful if the command exits with a status code of 0.
grpc
Performs a remote procedure call using gRPC.
The target should implement
gRPC health checks.
The diagnostic is considered successful if the status
of the response is SERVING.
gRPC probes are an alpha feature and are only available if you
enable the GRPCContainerProbefeature gate.
httpGet
Performs an HTTP GET request against the Pod's IP
address on a specified port and path. The diagnostic is
considered successful if the response has a status code
greater than or equal to 200 and less than 400.
tcpSocket
Performs a TCP check against the Pod's IP address on
a specified port. The diagnostic is considered successful if
the port is open. If the remote system (the container) closes
the connection immediately after it opens, this counts as healthy.
Probe outcome
Each probe has one of three results:
Success
The container passed the diagnostic.
Failure
The container failed the diagnostic.
Unknown
The diagnostic failed (no action should be taken, and the kubelet
will make further checks).
Types of probe
The kubelet can optionally perform and react to three kinds of probes on running
containers:
livenessProbe
Indicates whether the container is running. If
the liveness probe fails, the kubelet kills the container, and the container
is subjected to its restart policy. If a container does not
provide a liveness probe, the default state is Success.
readinessProbe
Indicates whether the container is ready to respond to requests.
If the readiness probe fails, the endpoints controller removes the Pod's IP
address from the endpoints of all Services that match the Pod. The default
state of readiness before the initial delay is Failure. If a container does
not provide a readiness probe, the default state is Success.
startupProbe
Indicates whether the application within the container is started.
All other probes are disabled if a startup probe is provided, until it succeeds.
If the startup probe fails, the kubelet kills the container, and the container
is subjected to its restart policy. If a container does not
provide a startup probe, the default state is Success.
If the process in your container is able to crash on its own whenever it
encounters an issue or becomes unhealthy, you do not necessarily need a liveness
probe; the kubelet will automatically perform the correct action in accordance
with the Pod's restartPolicy.
If you'd like your container to be killed and restarted if a probe fails, then
specify a liveness probe, and specify a restartPolicy of Always or OnFailure.
When should you use a readiness probe?
FEATURE STATE:Kubernetes v1.0 [stable]
If you'd like to start sending traffic to a Pod only when a probe succeeds,
specify a readiness probe. In this case, the readiness probe might be the same
as the liveness probe, but the existence of the readiness probe in the spec means
that the Pod will start without receiving any traffic and only start receiving
traffic after the probe starts succeeding.
If you want your container to be able to take itself down for maintenance, you
can specify a readiness probe that checks an endpoint specific to readiness that
is different from the liveness probe.
If your app has a strict dependency on back-end services, you can implement both
a liveness and a readiness probe. The liveness probe passes when the app itself
is healthy, but the readiness probe additionally checks that each required
back-end service is available. This helps you avoid directing traffic to Pods
that can only respond with error messages.
If your container needs to work on loading large data, configuration files, or
migrations during startup, you can use a
startup probe. However, if you want to
detect the difference between an app that has failed and an app that is still
processing its startup data, you might prefer a readiness probe.
Note: If you want to be able to drain requests when the Pod is deleted, you do not
necessarily need a readiness probe; on deletion, the Pod automatically puts itself
into an unready state regardless of whether the readiness probe exists.
The Pod remains in the unready state while it waits for the containers in the Pod
to stop.
When should you use a startup probe?
FEATURE STATE:Kubernetes v1.20 [stable]
Startup probes are useful for Pods that have containers that take a long time to
come into service. Rather than set a long liveness interval, you can configure
a separate configuration for probing the container as it starts up, allowing
a time longer than the liveness interval would allow.
If your container usually starts in more than
initialDelaySeconds + failureThreshold × periodSeconds, you should specify a
startup probe that checks the same endpoint as the liveness probe. The default for
periodSeconds is 10s. You should then set its failureThreshold high enough to
allow the container to start, without changing the default values of the liveness
probe. This helps to protect against deadlocks.
Termination of Pods
Because Pods represent processes running on nodes in the cluster, it is important to
allow those processes to gracefully terminate when they are no longer needed (rather
than being abruptly stopped with a KILL signal and having no chance to clean up).
The design aim is for you to be able to request deletion and know when processes
terminate, but also be able to ensure that deletes eventually complete.
When you request deletion of a Pod, the cluster records and tracks the intended grace period
before the Pod is allowed to be forcefully killed. With that forceful shutdown tracking in
place, the kubelet attempts graceful
shutdown.
Typically, the container runtime sends a TERM signal to the main process in each
container. Many container runtimes respect the STOPSIGNAL value defined in the container
image and send this instead of TERM.
Once the grace period has expired, the KILL signal is sent to any remaining
processes, and the Pod is then deleted from the
API Server. If the kubelet or the
container runtime's management service is restarted while waiting for processes to terminate, the
cluster retries from the start including the full original grace period.
An example flow:
You use the kubectl tool to manually delete a specific Pod, with the default grace period
(30 seconds).
The Pod in the API server is updated with the time beyond which the Pod is considered "dead"
along with the grace period.
If you use kubectl describe to check on the Pod you're deleting, that Pod shows up as
"Terminating".
On the node where the Pod is running: as soon as the kubelet sees that a Pod has been marked
as terminating (a graceful shutdown duration has been set), the kubelet begins the local Pod
shutdown process.
If one of the Pod's containers has defined a preStophook, the kubelet
runs that hook inside of the container. If the preStop hook is still running after the
grace period expires, the kubelet requests a small, one-off grace period extension of 2
seconds.
Note: If the preStop hook needs longer to complete than the default grace period allows,
you must modify terminationGracePeriodSeconds to suit this.
The kubelet triggers the container runtime to send a TERM signal to process 1 inside each
container.
Note: The containers in the Pod receive the TERM signal at different times and in an arbitrary
order. If the order of shutdowns matters, consider using a preStop hook to synchronize.
At the same time as the kubelet is starting graceful shutdown, the control plane removes that
shutting-down Pod from Endpoints (and, if enabled, EndpointSlice) objects where these represent
a Service with a configured
selector.
ReplicaSets and other workload resources
no longer treat the shutting-down Pod as a valid, in-service replica. Pods that shut down slowly
cannot continue to serve traffic as load balancers (like the service proxy) remove the Pod from
the list of endpoints as soon as the termination grace period begins.
When the grace period expires, the kubelet triggers forcible shutdown. The container runtime sends
SIGKILL to any processes still running in any container in the Pod.
The kubelet also cleans up a hidden pause container if that container runtime uses one.
The kubelet triggers forcible removal of Pod object from the API server, by setting grace period
to 0 (immediate deletion).
The API server deletes the Pod's API object, which is then no longer visible from any client.
Forced Pod termination
Caution: Forced deletions can be potentially disruptive for some workloads and their Pods.
By default, all deletes are graceful within 30 seconds. The kubectl delete command supports
the --grace-period=<seconds> option which allows you to override the default and specify your
own value.
Setting the grace period to 0 forcibly and immediately deletes the Pod from the API
server. If the pod was still running on a node, that forcible deletion triggers the kubelet to
begin immediate cleanup.
Note: You must specify an additional flag --force along with --grace-period=0 in order to perform force deletions.
When a force deletion is performed, the API server does not wait for confirmation
from the kubelet that the Pod has been terminated on the node it was running on. It
removes the Pod in the API immediately so a new Pod can be created with the same
name. On the node, Pods that are set to terminate immediately will still be given
a small grace period before being force killed.
If you need to force-delete Pods that are part of a StatefulSet, refer to the task
documentation for
deleting Pods from a StatefulSet.
Garbage collection of terminated Pods
For failed Pods, the API objects remain in the cluster's API until a human or
controller process
explicitly removes them.
The control plane cleans up terminated Pods (with a phase of Succeeded or
Failed), when the number of Pods exceeds the configured threshold
(determined by terminated-pod-gc-threshold in the kube-controller-manager).
This avoids a resource leak as Pods are created and terminated over time.
For detailed information about Pod and container status in the API, see
the API reference documentation covering
.status for Pod.
3.5.1.2 - Init Containers
This page provides an overview of init containers: specialized containers that run
before app containers in a Pod.
Init containers can contain utilities or setup scripts not present in an app image.
You can specify init containers in the Pod specification alongside the containers
array (which describes app containers).
Understanding init containers
A Pod can have multiple containers
running apps within it, but it can also have one or more init containers, which are run
before the app containers are started.
Init containers are exactly like regular containers, except:
Init containers always run to completion.
Each init container must complete successfully before the next one starts.
If a Pod's init container fails, the kubelet repeatedly restarts that init container until it succeeds.
However, if the Pod has a restartPolicy of Never, and an init container fails during startup of that Pod, Kubernetes treats the overall Pod as failed.
To specify an init container for a Pod, add the initContainers field into
the Pod specification,
as an array of container items (similar to the app containers field and its contents).
See Container in the
API reference for more details.
The status of the init containers is returned in .status.initContainerStatuses
field as an array of the container statuses (similar to the .status.containerStatuses
field).
Differences from regular containers
Init containers support all the fields and features of app containers,
including resource limits, volumes, and security settings. However, the
resource requests and limits for an init container are handled differently,
as documented in Resources.
Also, init containers do not support lifecycle, livenessProbe, readinessProbe, or
startupProbe because they must run to completion before the Pod can be ready.
If you specify multiple init containers for a Pod, kubelet runs each init
container sequentially. Each init container must succeed before the next can run.
When all of the init containers have run to completion, kubelet initializes
the application containers for the Pod and runs them as usual.
Using init containers
Because init containers have separate images from app containers, they
have some advantages for start-up related code:
Init containers can contain utilities or custom code for setup that are not present in an app
image. For example, there is no need to make an image FROM another image just to use a tool like
sed, awk, python, or dig during setup.
The application image builder and deployer roles can work independently without
the need to jointly build a single app image.
Init containers can run with a different view of the filesystem than app containers in the
same Pod. Consequently, they can be given access to
Secrets that app containers cannot access.
Because init containers run to completion before any app containers start, init containers offer
a mechanism to block or delay app container startup until a set of preconditions are met. Once
preconditions are met, all of the app containers in a Pod can start in parallel.
Init containers can securely run utilities or custom code that would otherwise make an app
container image less secure. By keeping unnecessary tools separate you can limit the attack
surface of your app container image.
Examples
Here are some ideas for how to use init containers:
Wait for a Service to
be created, using a shell one-line command like:
for i in {1..100}; do sleep 1; if dig myservice; thenexit 0; fi; done; exit1
Register this Pod with a remote server from the downward API with a command like:
curl -X POST http://$MANAGEMENT_SERVICE_HOST:$MANAGEMENT_SERVICE_PORT/register -d 'instance=$(<POD_NAME>)&ip=$(<POD_IP>)'
Wait for some time before starting the app container with a command like
Place values into a configuration file and run a template tool to dynamically
generate a configuration file for the main app container. For example,
place the POD_IP value in a configuration and generate the main app
configuration file using Jinja.
Init containers in use
This example defines a simple Pod that has two init containers.
The first waits for myservice, and the second waits for mydb. Once both
init containers complete, the Pod runs the app container from its spec section.
apiVersion:v1kind:Podmetadata:name:myapp-podlabels:app.kubernetes.io/name:MyAppspec:containers:- name:myapp-containerimage:busybox:1.28command:['sh','-c','echo The app is running! && sleep 3600']initContainers:- name:init-myserviceimage:busybox:1.28command:['sh','-c',"until nslookup myservice.$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace).svc.cluster.local; do echo waiting for myservice; sleep 2; done"]- name:init-mydbimage:busybox:1.28command:['sh','-c',"until nslookup mydb.$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace).svc.cluster.local; do echo waiting for mydb; sleep 2; done"]
You can start this Pod by running:
kubectl apply -f myapp.yaml
The output is similar to this:
pod/myapp-pod created
And check on its status with:
kubectl get -f myapp.yaml
The output is similar to this:
NAME READY STATUS RESTARTS AGE
myapp-pod 0/1 Init:0/2 0 6m
or for more details:
kubectl describe -f myapp.yaml
The output is similar to this:
Name: myapp-pod
Namespace: default
[...]
Labels: app.kubernetes.io/name=MyApp
Status: Pending
[...]
Init Containers:
init-myservice:
[...]
State: Running
[...]
init-mydb:
[...]
State: Waiting
Reason: PodInitializing
Ready: False
[...]
Containers:
myapp-container:
[...]
State: Waiting
Reason: PodInitializing
Ready: False
[...]
Events:
FirstSeen LastSeen Count From SubObjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
16s 16s 1 {default-scheduler } Normal Scheduled Successfully assigned myapp-pod to 172.17.4.201
16s 16s 1 {kubelet 172.17.4.201} spec.initContainers{init-myservice} Normal Pulling pulling image "busybox"
13s 13s 1 {kubelet 172.17.4.201} spec.initContainers{init-myservice} Normal Pulled Successfully pulled image "busybox"
13s 13s 1 {kubelet 172.17.4.201} spec.initContainers{init-myservice} Normal Created Created container with docker id 5ced34a04634; Security:[seccomp=unconfined]
13s 13s 1 {kubelet 172.17.4.201} spec.initContainers{init-myservice} Normal Started Started container with docker id 5ced34a04634
To see logs for the init containers in this Pod, run:
kubectl logs myapp-pod -c init-myservice # Inspect the first init containerkubectl logs myapp-pod -c init-mydb # Inspect the second init container
At this point, those init containers will be waiting to discover Services named
mydb and myservice.
Here's a configuration you can use to make those Services appear:
You'll then see that those init containers complete, and that the myapp-pod
Pod moves into the Running state:
kubectl get -f myapp.yaml
The output is similar to this:
NAME READY STATUS RESTARTS AGE
myapp-pod 1/1 Running 0 9m
This simple example should provide some inspiration for you to create your own
init containers. What's next contains a link to a more detailed example.
Detailed behavior
During Pod startup, the kubelet delays running init containers until the networking
and storage are ready. Then the kubelet runs the Pod's init containers in the order
they appear in the Pod's spec.
Each init container must exit successfully before
the next container starts. If a container fails to start due to the runtime or
exits with failure, it is retried according to the Pod restartPolicy. However,
if the Pod restartPolicy is set to Always, the init containers use
restartPolicy OnFailure.
A Pod cannot be Ready until all init containers have succeeded. The ports on an
init container are not aggregated under a Service. A Pod that is initializing
is in the Pending state but should have a condition Initialized set to false.
If the Pod restarts, or is restarted, all init containers
must execute again.
Changes to the init container spec are limited to the container image field.
Altering an init container image field is equivalent to restarting the Pod.
Because init containers can be restarted, retried, or re-executed, init container
code should be idempotent. In particular, code that writes to files on EmptyDirs
should be prepared for the possibility that an output file already exists.
Init containers have all of the fields of an app container. However, Kubernetes
prohibits readinessProbe from being used because init containers cannot
define readiness distinct from completion. This is enforced during validation.
Use activeDeadlineSeconds on the Pod to prevent init containers from failing forever.
The active deadline includes init containers.
However it is recommended to use activeDeadlineSeconds only if teams deploy their application
as a Job, because activeDeadlineSeconds has an effect even after initContainer finished.
The Pod which is already running correctly would be killed by activeDeadlineSeconds if you set.
The name of each app and init container in a Pod must be unique; a
validation error is thrown for any container sharing a name with another.
Resources
Given the ordering and execution for init containers, the following rules
for resource usage apply:
The highest of any particular resource request or limit defined on all init
containers is the effective init request/limit. If any resource has no
resource limit specified this is considered as the highest limit.
The Pod's effective request/limit for a resource is the higher of:
the sum of all app containers request/limit for a resource
the effective init request/limit for a resource
Scheduling is done based on effective requests/limits, which means
init containers can reserve resources for initialization that are not used
during the life of the Pod.
The QoS (quality of service) tier of the Pod's effective QoS tier is the
QoS tier for init containers and app containers alike.
Quota and limits are applied based on the effective Pod request and
limit.
Pod level control groups (cgroups) are based on the effective Pod request and
limit, the same as the scheduler.
Pod restart reasons
A Pod can restart, causing re-execution of init containers, for the following
reasons:
The Pod infrastructure container is restarted. This is uncommon and would
have to be done by someone with root access to nodes.
All containers in a Pod are terminated while restartPolicy is set to Always,
forcing a restart, and the init container completion record has been lost due
to garbage collection.
The Pod will not be restarted when the init container image is changed, or the
init container completion record has been lost due to garbage collection. This
applies for Kubernetes v1.20 and later. If you are using an earlier version of
Kubernetes, consult the documentation for the version you are using.
This guide is for application owners who want to build
highly available applications, and thus need to understand
what types of disruptions can happen to Pods.
It is also for cluster administrators who want to perform automated
cluster actions, like upgrading and autoscaling clusters.
Voluntary and involuntary disruptions
Pods do not disappear until someone (a person or a controller) destroys them, or
there is an unavoidable hardware or system software error.
We call these unavoidable cases involuntary disruptions to
an application. Examples are:
a hardware failure of the physical machine backing the node
cluster administrator deletes VM (instance) by mistake
cloud provider or hypervisor failure makes VM disappear
a kernel panic
the node disappears from the cluster due to cluster network partition
Except for the out-of-resources condition, all these conditions
should be familiar to most users; they are not specific
to Kubernetes.
We call other cases voluntary disruptions. These include both
actions initiated by the application owner and those initiated by a Cluster
Administrator. Typical application owner actions include:
deleting the deployment or other controller that manages the pod
updating a deployment's pod template causing a restart
Draining a node from a cluster to scale the cluster down (learn about
Cluster Autoscaling
).
Removing a pod from a node to permit something else to fit on that node.
These actions might be taken directly by the cluster administrator, or by automation
run by the cluster administrator, or by your cluster hosting provider.
Ask your cluster administrator or consult your cloud provider or distribution documentation
to determine if any sources of voluntary disruptions are enabled for your cluster.
If none are enabled, you can skip creating Pod Disruption Budgets.
Caution: Not all voluntary disruptions are constrained by Pod Disruption Budgets. For example,
deleting deployments or pods bypasses Pod Disruption Budgets.
Dealing with disruptions
Here are some ways to mitigate involuntary disruptions:
Replicate your application if you need higher availability. (Learn about running replicated
stateless
and stateful applications.)
For even higher availability when running replicated applications,
spread applications across racks (using
anti-affinity)
or across zones (if using a
multi-zone cluster.)
The frequency of voluntary disruptions varies. On a basic Kubernetes cluster, there are
no automated voluntary disruptions (only user-triggered ones). However, your cluster administrator or hosting provider
may run some additional services which cause voluntary disruptions. For example,
rolling out node software updates can cause voluntary disruptions. Also, some implementations
of cluster (node) autoscaling may cause voluntary disruptions to defragment and compact nodes.
Your cluster administrator or hosting provider should have documented what level of voluntary
disruptions, if any, to expect. Certain configuration options, such as
using PriorityClasses
in your pod spec can also cause voluntary (and involuntary) disruptions.
Pod disruption budgets
FEATURE STATE:Kubernetes v1.21 [stable]
Kubernetes offers features to help you run highly available applications even when you
introduce frequent voluntary disruptions.
As an application owner, you can create a PodDisruptionBudget (PDB) for each application.
A PDB limits the number of Pods of a replicated application that are down simultaneously from
voluntary disruptions. For example, a quorum-based application would
like to ensure that the number of replicas running is never brought below the
number needed for a quorum. A web front end might want to
ensure that the number of replicas serving load never falls below a certain
percentage of the total.
Cluster managers and hosting providers should use tools which
respect PodDisruptionBudgets by calling the Eviction API
instead of directly deleting pods or deployments.
For example, the kubectl drain subcommand lets you mark a node as going out of
service. When you run kubectl drain, the tool tries to evict all of the Pods on
the Node you're taking out of service. The eviction request that kubectl submits on
your behalf may be temporarily rejected, so the tool periodically retries all failed
requests until all Pods on the target node are terminated, or until a configurable timeout
is reached.
A PDB specifies the number of replicas that an application can tolerate having, relative to how
many it is intended to have. For example, a Deployment which has a .spec.replicas: 5 is
supposed to have 5 pods at any given time. If its PDB allows for there to be 4 at a time,
then the Eviction API will allow voluntary disruption of one (but not two) pods at a time.
The group of pods that comprise the application is specified using a label selector, the same
as the one used by the application's controller (deployment, stateful-set, etc).
The "intended" number of pods is computed from the .spec.replicas of the workload resource
that is managing those pods. The control plane discovers the owning workload resource by
examining the .metadata.ownerReferences of the Pod.
Involuntary disruptions cannot be prevented by PDBs; however they
do count against the budget.
Pods which are deleted or unavailable due to a rolling upgrade to an application do count
against the disruption budget, but workload resources (such as Deployment and StatefulSet)
are not limited by PDBs when doing rolling upgrades. Instead, the handling of failures
during application updates is configured in the spec for the specific workload resource.
When a pod is evicted using the eviction API, it is gracefully
terminated, honoring the
terminationGracePeriodSeconds setting in its PodSpec.
PodDisruptionBudget example
Consider a cluster with 3 nodes, node-1 through node-3.
The cluster is running several applications. One of them has 3 replicas initially called
pod-a, pod-b, and pod-c. Another, unrelated pod without a PDB, called pod-x, is also shown.
Initially, the pods are laid out as follows:
node-1
node-2
node-3
pod-a available
pod-b available
pod-c available
pod-x available
All 3 pods are part of a deployment, and they collectively have a PDB which requires
there be at least 2 of the 3 pods to be available at all times.
For example, assume the cluster administrator wants to reboot into a new kernel version to fix a bug in the kernel.
The cluster administrator first tries to drain node-1 using the kubectl drain command.
That tool tries to evict pod-a and pod-x. This succeeds immediately.
Both pods go into the terminating state at the same time.
This puts the cluster in this state:
node-1 draining
node-2
node-3
pod-a terminating
pod-b available
pod-c available
pod-x terminating
The deployment notices that one of the pods is terminating, so it creates a replacement
called pod-d. Since node-1 is cordoned, it lands on another node. Something has
also created pod-y as a replacement for pod-x.
(Note: for a StatefulSet, pod-a, which would be called something like pod-0, would need
to terminate completely before its replacement, which is also called pod-0 but has a
different UID, could be created. Otherwise, the example applies to a StatefulSet as well.)
Now the cluster is in this state:
node-1 draining
node-2
node-3
pod-a terminating
pod-b available
pod-c available
pod-x terminating
pod-d starting
pod-y
At some point, the pods terminate, and the cluster looks like this:
node-1 drained
node-2
node-3
pod-b available
pod-c available
pod-d starting
pod-y
At this point, if an impatient cluster administrator tries to drain node-2 or
node-3, the drain command will block, because there are only 2 available
pods for the deployment, and its PDB requires at least 2. After some time passes, pod-d becomes available.
The cluster state now looks like this:
node-1 drained
node-2
node-3
pod-b available
pod-c available
pod-d available
pod-y
Now, the cluster administrator tries to drain node-2.
The drain command will try to evict the two pods in some order, say
pod-b first and then pod-d. It will succeed at evicting pod-b.
But, when it tries to evict pod-d, it will be refused because that would leave only
one pod available for the deployment.
The deployment creates a replacement for pod-b called pod-e.
Because there are not enough resources in the cluster to schedule
pod-e the drain will again block. The cluster may end up in this
state:
node-1 drained
node-2
node-3
no node
pod-b terminating
pod-c available
pod-e pending
pod-d available
pod-y
At this point, the cluster administrator needs to
add a node back to the cluster to proceed with the upgrade.
You can see how Kubernetes varies the rate at which disruptions
can happen, according to:
how many replicas an application needs
how long it takes to gracefully shutdown an instance
how long it takes a new instance to start up
the type of controller
the cluster's resource capacity
Pod disruption conditions
FEATURE STATE:Kubernetes v1.25 [alpha]
Note: In order to use this behavior, you must enable the PodDisruptionsConditionfeature gate
in your cluster.
When enabled, a dedicated Pod DisruptionTargetcondition is added to indicate
that the Pod is about to be deleted due to a disruption.
The reason field of the condition additionally
indicates one of the following reasons for the Pod termination:
PreemptionByKubeScheduler
Pod is due to be preempted by a scheduler in order to accommodate a new Pod with a higher priority. For more information, see Pod priority preemption.
DeletionByTaintManager
Pod is due to be deleted by Taint Manager (which is part of the node lifecycle controller within kube-controller-manager) due to a NoExecute taint that the Pod does not tolerate; see taint-based evictions.
Pod, that is bound to a no longer existing Node, is due to be deleted by Pod garbage collection.
Note: A Pod disruption might be interrupted. The control plane might re-attempt to
continue the disruption of the same Pod, but it is not guaranteed. As a result,
the DisruptionTarget condition might be added to a Pod, but that Pod might then not actually be
deleted. In such a situation, after some time, the
Pod disruption condition will be cleared.
When using a Job (or CronJob), you may want to use these Pod disruption conditions as part of your Job's
Pod failure policy.
Separating Cluster Owner and Application Owner Roles
Often, it is useful to think of the Cluster Manager
and Application Owner as separate roles with limited knowledge
of each other. This separation of responsibilities
may make sense in these scenarios:
when there are many application teams sharing a Kubernetes cluster, and
there is natural specialization of roles
when third-party tools or services are used to automate cluster management
Pod Disruption Budgets support this separation of roles by providing an
interface between the roles.
If you do not have such a separation of responsibilities in your organization,
you may not need to use Pod Disruption Budgets.
How to perform Disruptive Actions on your Cluster
If you are a Cluster Administrator, and you need to perform a disruptive action on all
the nodes in your cluster, such as a node or system software upgrade, here are some options:
Accept downtime during the upgrade.
Failover to another complete replica cluster.
No downtime, but may be costly both for the duplicated nodes
and for human effort to orchestrate the switchover.
Write disruption tolerant applications and use PDBs.
No downtime.
Minimal resource duplication.
Allows more automation of cluster administration.
Writing disruption-tolerant applications is tricky, but the work to tolerate voluntary
disruptions largely overlaps with work to support autoscaling and tolerating
involuntary disruptions.
Learn about updating a deployment
including steps to maintain its availability during the rollout.
3.5.1.4 - Ephemeral Containers
FEATURE STATE:Kubernetes v1.25 [stable]
This page provides an overview of ephemeral containers: a special type of container
that runs temporarily in an existing Pod to
accomplish user-initiated actions such as troubleshooting. You use ephemeral
containers to inspect services rather than to build applications.
Understanding ephemeral containers
Pods are the fundamental building
block of Kubernetes applications. Since Pods are intended to be disposable and
replaceable, you cannot add a container to a Pod once it has been created.
Instead, you usually delete and replace Pods in a controlled fashion using
deployments.
Sometimes it's necessary to inspect the state of an existing Pod, however, for
example to troubleshoot a hard-to-reproduce bug. In these cases you can run
an ephemeral container in an existing Pod to inspect its state and run
arbitrary commands.
What is an ephemeral container?
Ephemeral containers differ from other containers in that they lack guarantees
for resources or execution, and they will never be automatically restarted, so
they are not appropriate for building applications. Ephemeral containers are
described using the same ContainerSpec as regular containers, but many fields
are incompatible and disallowed for ephemeral containers.
Ephemeral containers may not have ports, so fields such as ports,
livenessProbe, readinessProbe are disallowed.
Pod resource allocations are immutable, so setting resources is disallowed.
Ephemeral containers are created using a special ephemeralcontainers handler
in the API rather than by adding them directly to pod.spec, so it's not
possible to add an ephemeral container using kubectl edit.
Like regular containers, you may not change or remove an ephemeral container
after you have added it to a Pod.
Uses for ephemeral containers
Ephemeral containers are useful for interactive troubleshooting when kubectl exec is insufficient because a container has crashed or a container image
doesn't include debugging utilities.
In particular, distroless images
enable you to deploy minimal container images that reduce attack surface
and exposure to bugs and vulnerabilities. Since distroless images do not include a
shell or any debugging utilities, it's difficult to troubleshoot distroless
images using kubectl exec alone.
When using ephemeral containers, it's helpful to enable process namespace
sharing so
you can view processes in other containers.
This page explains how user namespaces are used in Kubernetes pods. A user
namespace allows to isolate the user running inside the container from the one
in the host.
A process running as root in a container can run as a different (non-root) user
in the host; in other words, the process has full privileges for operations
inside the user namespace, but is unprivileged for operations outside the
namespace.
You can use this feature to reduce the damage a compromised container can do to
the host or other pods in the same node. There are several security
vulnerabilities rated either HIGH or CRITICAL that were not
exploitable when user namespaces is active. It is expected user namespace will
mitigate some future vulnerabilities too.
Before you begin
🛇 This item links to a third party project or product that is not part of Kubernetes itself. More information
This is a Linux only feature. In addition, support is needed in the
container runtime
to use this feature with Kubernetes stateless pods:
CRI-O: v1.25 has support for user namespaces.
containerd: support is planned for the 1.7 release. See containerd
issue #7063 for more details.
User namespaces is a Linux feature that allows to map users in the container to
different users in the host. Furthermore, the capabilities granted to a pod in
a user namespace are valid only in the namespace and void outside of it.
A pod can opt-in to use user nameapces by setting the pod.spec.hostUsers field
to false.
The kubelet will pick host UIDs/GIDs a pod is mapped to, and will do so in a way
to guarantee that no two stateless pods on the same node use the same mapping.
The runAsUser, runAsGroup, fsGroup, etc. fields in the pod.spec always
refer to the user inside the container.
The valid UIDs/GIDs when this feature is enabled is the range 0-65535. This
applies to files and processes (runAsUser, runAsGroup, etc.).
Files using a UID/GID outside this range will be seen as belonging to the
overflow ID, usually 65534 (configured in /proc/sys/kernel/overflowuid and
/proc/sys/kernel/overflowgid). However, it is not possible to modify those
files, even by running as the 65534 user/group.
Most applications that need to run as root but don't access other host
namespaces or resources, should continue to run fine without any changes needed
if user namespaces is activated.
Understanding user namespaces for stateless pods
Several container runtimes with their default configuration (like Docker Engine,
containerd, CRI-O) use Linux namespaces for isolation. Other technologies exist
and can be used with those runtimes too (e.g. Kata Containers uses VMs instead of
Linux namespaces). This page is applicable for container runtimes using Linux
namespaces for isolation.
When creating a pod, by default, several new namespaces are used for isolation:
a network namespace to isolate the network of the container, a PID namespace to
isolate the view of processes, etc. If a user namespace is used, this will
isolate the users in the container from the users in the node.
This means containers can run as root and be mapped to a non-root user on the
host. Inside the container the process will think it is running as root (and
therefore tools like apt, yum, etc. work fine), while in reality the process
doesn't have privileges on the host. You can verify this, for example, if you
check the user the container process is running ps from the host. The user
ps shows is not the same as the user you see if you execute inside the
container the command id.
This abstraction limits what can happen, for example, if the container manages
to escape to the host. Given that the container is running as a non-privileged
user on the host, it is limited what it can do to the host.
Furthermore, as users on each pod will be mapped to different non-overlapping
users in the host, it is limited what they can do to other pods too.
Capabilities granted to a pod are also limited to the pod user namespace and
mostly invalid out of it, some are even completely void. Here are two examples:
CAP_SYS_MODULE does not have any effect if granted to a pod using user
namespaces, the pod isn't able to load kernel modules.
CAP_SYS_ADMIN is limited to the pod's user namespace and invalid outside
of it.
Without using a user namespace a container running as root, in the case of a
container breakout, has root privileges on the node. And if some capability were
granted to the container, the capabilities are valid on the host too. None of
this is true when we use user namespaces.
If you want to know more details about what changes when user namespaces are in
use, see man 7 user_namespaces.
Set up a node to support user namespaces
It is recommended that the host's files and host's processes use UIDs/GIDs in
the range of 0-65535.
The kubelet will assign UIDs/GIDs higher than that to pods. Therefore, to
guarantee as much isolation as possible, the UIDs/GIDs used by the host's files
and host's processes should be in the range 0-65535.
Note that this recommendation is important to mitigate the impact of CVEs like
CVE-2021-25741, where a pod can potentially read arbitrary
files in the hosts. If the UIDs/GIDs of the pod and the host don't overlap, it
is limited what a pod would be able to do: the pod UID/GID won't match the
host's file owner/group.
Limitations
When using a user namespace for the pod, it is disallowed to use other host
namespaces. In particular, if you set hostUsers: false then you are not
allowed to set any of:
hostNetwork: true
hostIPC: true
hostPID: true
The pod is allowed to use no volumes at all or, if using volumes, only these
volume types are allowed:
configmap
secret
projected
downwardAPI
emptyDir
To guarantee that the pod can read the files of such volumes, volumes are
created as if you specified .spec.securityContext.fsGroup as 0 for the Pod.
If it is specified to a different value, this other value will of course be
honored instead.
As a by-product of this, folders and files for these volumes will have
permissions for the group, even if defaultMode or mode to specific items of
the volumes were specified without permissions to groups. For example, it is not
possible to mount these volumes in a way that its files have permissions only
for the owner.
3.5.1.6 - Downward API
There are two ways to expose Pod and container fields to a running container: environment variables, and as files that are populated by a special volume type. Together, these two ways of exposing Pod and container fields are called the downward API.
It is sometimes useful for a container to have information about itself, without
being overly coupled to Kubernetes. The downward API allows containers to consume
information about themselves or the cluster without using the Kubernetes client
or API server.
An example is an existing application that assumes a particular well-known
environment variable holds a unique identifier. One possibility is to wrap the
application, but that is tedious and error-prone, and it violates the goal of low
coupling. A better option would be to use the Pod's name as an identifier, and
inject the Pod's name into the well-known environment variable.
In Kubernetes, there are two ways to expose Pod and container fields to a running container:
Together, these two ways of exposing Pod and container fields are called the
downward API.
Available fields
Only some Kubernetes API fields are available through the downward API. This
section lists which fields you can make available.
You can pass information from available Pod-level fields using fieldRef.
At the API level, the spec for a Pod always defines at least one
Container.
You can pass information from available Container-level fields using
resourceFieldRef.
Information available via fieldRef
For most Pod-level fields, you can provide them to a container either as
an environment variable or using a downwardAPI volume. The fields available
via either mechanism are:
the primary IP address of the node to which the Pod is assigned
status.podIP
the pod's primary IP address (usually, its IPv4 address)
In addition, the following information is available through
a downwardAPI volume fieldRef, but not as environment variables:
metadata.labels
all of the pod's labels, formatted as label-key="escaped-label-value" with one label per line
metadata.annotations
all of the pod's annotations, formatted as annotation-key="escaped-annotation-value" with one annotation per line
Information available via resourceFieldRef
These container-level fields allow you to provide information about
requests and limits
for resources such as CPU and memory.
resource: limits.cpu
A container's CPU limit
resource: requests.cpu
A container's CPU request
resource: limits.memory
A container's memory limit
resource: requests.memory
A container's memory request
resource: limits.hugepages-*
A container's hugepages limit (provided that the DownwardAPIHugePagesfeature gate is enabled)
resource: requests.hugepages-*
A container's hugepages request (provided that the DownwardAPIHugePagesfeature gate is enabled)
resource: limits.ephemeral-storage
A container's ephemeral-storage limit
resource: requests.ephemeral-storage
A container's ephemeral-storage request
Fallback information for resource limits
If CPU and memory limits are not specified for a container, and you use the
downward API to try to expose that information, then the
kubelet defaults to exposing the maximum allocatable value for CPU and memory
based on the node allocatable
calculation.
A Deployment provides declarative updates for Pods and
ReplicaSets.
You describe a desired state in a Deployment, and the Deployment Controller changes the actual state to the desired state at a controlled rate. You can define Deployments to create new ReplicaSets, or to remove existing Deployments and adopt all their resources with new Deployments.
Note: Do not manage ReplicaSets owned by a Deployment. Consider opening an issue in the main Kubernetes repository if your use case is not covered below.
Use Case
The following are typical use cases for Deployments:
Declare the new state of the Pods by updating the PodTemplateSpec of the Deployment. A new ReplicaSet is created and the Deployment manages moving the Pods from the old ReplicaSet to the new one at a controlled rate. Each new ReplicaSet updates the revision of the Deployment.
A Deployment named nginx-deployment is created, indicated by the .metadata.name field.
The Deployment creates three replicated Pods, indicated by the .spec.replicas field.
The .spec.selector field defines how the Deployment finds which Pods to manage.
In this case, you select a label that is defined in the Pod template (app: nginx).
However, more sophisticated selection rules are possible,
as long as the Pod template itself satisfies the rule.
Note: The .spec.selector.matchLabels field is a map of {key,value} pairs.
A single {key,value} in the matchLabels map is equivalent to an element of matchExpressions,
whose key field is "key", the operator is "In", and the values array contains only "value".
All of the requirements, from both matchLabels and matchExpressions, must be satisfied in order to match.
The template field contains the following sub-fields:
The Pods are labeled app: nginxusing the .metadata.labels field.
The Pod template's specification, or .template.spec field, indicates that
the Pods run one container, nginx, which runs the nginxDocker Hub image at version 1.14.2.
Create one container and name it nginx using the .spec.template.spec.containers[0].name field.
Before you begin, make sure your Kubernetes cluster is up and running.
Follow the steps given below to create the above Deployment:
Create the Deployment by running the following command:
Run kubectl get deployments to check if the Deployment was created.
If the Deployment is still being created, the output is similar to the following:
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deployment 0/3 0 0 1s
When you inspect the Deployments in your cluster, the following fields are displayed:
NAME lists the names of the Deployments in the namespace.
READY displays how many replicas of the application are available to your users. It follows the pattern ready/desired.
UP-TO-DATE displays the number of replicas that have been updated to achieve the desired state.
AVAILABLE displays how many replicas of the application are available to your users.
AGE displays the amount of time that the application has been running.
Notice how the number of desired replicas is 3 according to .spec.replicas field.
To see the Deployment rollout status, run kubectl rollout status deployment/nginx-deployment.
The output is similar to:
Waiting for rollout to finish: 2 out of 3 new replicas have been updated...
deployment "nginx-deployment" successfully rolled out
Run the kubectl get deployments again a few seconds later.
The output is similar to this:
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deployment 3/3 3 3 18s
Notice that the Deployment has created all three replicas, and all replicas are up-to-date (they contain the latest Pod template) and available.
To see the ReplicaSet (rs) created by the Deployment, run kubectl get rs. The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-deployment-75675f5897 3 3 3 18s
ReplicaSet output shows the following fields:
NAME lists the names of the ReplicaSets in the namespace.
DESIRED displays the desired number of replicas of the application, which you define when you create the Deployment. This is the desired state.
CURRENT displays how many replicas are currently running.
READY displays how many replicas of the application are available to your users.
AGE displays the amount of time that the application has been running.
Notice that the name of the ReplicaSet is always formatted as [DEPLOYMENT-NAME]-[RANDOM-STRING].
The random string is randomly generated and uses the pod-template-hash as a seed.
To see the labels automatically generated for each Pod, run kubectl get pods --show-labels.
The output is similar to:
NAME READY STATUS RESTARTS AGE LABELS
nginx-deployment-75675f5897-7ci7o 1/1 Running 0 18s app=nginx,pod-template-hash=3123191453
nginx-deployment-75675f5897-kzszj 1/1 Running 0 18s app=nginx,pod-template-hash=3123191453
nginx-deployment-75675f5897-qqcnn 1/1 Running 0 18s app=nginx,pod-template-hash=3123191453
The created ReplicaSet ensures that there are three nginx Pods.
Note:
You must specify an appropriate selector and Pod template labels in a Deployment
(in this case, app: nginx).
Do not overlap labels or selectors with other controllers (including other Deployments and StatefulSets). Kubernetes doesn't stop you from overlapping, and if multiple controllers have overlapping selectors those controllers might conflict and behave unexpectedly.
Pod-template-hash label
Caution: Do not change this label.
The pod-template-hash label is added by the Deployment controller to every ReplicaSet that a Deployment creates or adopts.
This label ensures that child ReplicaSets of a Deployment do not overlap. It is generated by hashing the PodTemplate of the ReplicaSet and using the resulting hash as the label value that is added to the ReplicaSet selector, Pod template labels,
and in any existing Pods that the ReplicaSet might have.
Updating a Deployment
Note: A Deployment's rollout is triggered if and only if the Deployment's Pod template (that is, .spec.template)
is changed, for example if the labels or container images of the template are updated. Other updates, such as scaling the Deployment, do not trigger a rollout.
Follow the steps given below to update your Deployment:
Let's update the nginx Pods to use the nginx:1.16.1 image instead of the nginx:1.14.2 image.
kubectl set image deployment.v1.apps/nginx-deployment nginx=nginx:1.16.1
or use the following command:
kubectl set image deployment/nginx-deployment nginx=nginx:1.16.1
The output is similar to:
deployment.apps/nginx-deployment image updated
Alternatively, you can edit the Deployment and change .spec.template.spec.containers[0].image from nginx:1.14.2 to nginx:1.16.1:
kubectl edit deployment/nginx-deployment
The output is similar to:
deployment.apps/nginx-deployment edited
To see the rollout status, run:
kubectl rollout status deployment/nginx-deployment
The output is similar to this:
Waiting for rollout to finish: 2 out of 3 new replicas have been updated...
or
deployment "nginx-deployment" successfully rolled out
Get more details on your updated Deployment:
After the rollout succeeds, you can view the Deployment by running kubectl get deployments.
The output is similar to this:
NAME READY UP-TO-DATE AVAILABLE AGEnginx-deployment 3/3 3 3 36s
Run kubectl get rs to see that the Deployment updated the Pods by creating a new ReplicaSet and scaling it
up to 3 replicas, as well as scaling down the old ReplicaSet to 0 replicas.
kubectl get rs
The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-deployment-1564180365 3 3 3 6s
nginx-deployment-2035384211 0 0 0 36s
Running get pods should now show only the new Pods:
kubectl get pods
The output is similar to this:
NAME READY STATUS RESTARTS AGE
nginx-deployment-1564180365-khku8 1/1 Running 0 14s
nginx-deployment-1564180365-nacti 1/1 Running 0 14s
nginx-deployment-1564180365-z9gth 1/1 Running 0 14s
Next time you want to update these Pods, you only need to update the Deployment's Pod template again.
Deployment ensures that only a certain number of Pods are down while they are being updated. By default,
it ensures that at least 75% of the desired number of Pods are up (25% max unavailable).
Deployment also ensures that only a certain number of Pods are created above the desired number of Pods.
By default, it ensures that at most 125% of the desired number of Pods are up (25% max surge).
For example, if you look at the above Deployment closely, you will see that it first creates a new Pod,
then deletes an old Pod, and creates another new one. It does not kill old Pods until a sufficient number of
new Pods have come up, and does not create new Pods until a sufficient number of old Pods have been killed.
It makes sure that at least 3 Pods are available and that at max 4 Pods in total are available. In case of
a Deployment with 4 replicas, the number of Pods would be between 3 and 5.
Get details of your Deployment:
kubectl describe deployments
The output is similar to this:
Name: nginx-deployment
Namespace: default
CreationTimestamp: Thu, 30 Nov 2017 10:56:25 +0000
Labels: app=nginx
Annotations: deployment.kubernetes.io/revision=2
Selector: app=nginx
Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=nginx
Containers:
nginx:
Image: nginx:1.16.1
Port: 80/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets: <none>
NewReplicaSet: nginx-deployment-1564180365 (3/3 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 2m deployment-controller Scaled up replica set nginx-deployment-2035384211 to 3
Normal ScalingReplicaSet 24s deployment-controller Scaled up replica set nginx-deployment-1564180365 to 1
Normal ScalingReplicaSet 22s deployment-controller Scaled down replica set nginx-deployment-2035384211 to 2
Normal ScalingReplicaSet 22s deployment-controller Scaled up replica set nginx-deployment-1564180365 to 2
Normal ScalingReplicaSet 19s deployment-controller Scaled down replica set nginx-deployment-2035384211 to 1
Normal ScalingReplicaSet 19s deployment-controller Scaled up replica set nginx-deployment-1564180365 to 3
Normal ScalingReplicaSet 14s deployment-controller Scaled down replica set nginx-deployment-2035384211 to 0
Here you see that when you first created the Deployment, it created a ReplicaSet (nginx-deployment-2035384211)
and scaled it up to 3 replicas directly. When you updated the Deployment, it created a new ReplicaSet
(nginx-deployment-1564180365) and scaled it up to 1 and waited for it to come up. Then it scaled down the old ReplicaSet
to 2 and scaled up the new ReplicaSet to 2 so that at least 3 Pods were available and at most 4 Pods were created at all times.
It then continued scaling up and down the new and the old ReplicaSet, with the same rolling update strategy.
Finally, you'll have 3 available replicas in the new ReplicaSet, and the old ReplicaSet is scaled down to 0.
Note: Kubernetes doesn't count terminating Pods when calculating the number of availableReplicas, which must be between
replicas - maxUnavailable and replicas + maxSurge. As a result, you might notice that there are more Pods than
expected during a rollout, and that the total resources consumed by the Deployment is more than replicas + maxSurge
until the terminationGracePeriodSeconds of the terminating Pods expires.
Rollover (aka multiple updates in-flight)
Each time a new Deployment is observed by the Deployment controller, a ReplicaSet is created to bring up
the desired Pods. If the Deployment is updated, the existing ReplicaSet that controls Pods whose labels
match .spec.selector but whose template does not match .spec.template are scaled down. Eventually, the new
ReplicaSet is scaled to .spec.replicas and all old ReplicaSets is scaled to 0.
If you update a Deployment while an existing rollout is in progress, the Deployment creates a new ReplicaSet
as per the update and start scaling that up, and rolls over the ReplicaSet that it was scaling up previously
-- it will add it to its list of old ReplicaSets and start scaling it down.
For example, suppose you create a Deployment to create 5 replicas of nginx:1.14.2,
but then update the Deployment to create 5 replicas of nginx:1.16.1, when only 3
replicas of nginx:1.14.2 had been created. In that case, the Deployment immediately starts
killing the 3 nginx:1.14.2 Pods that it had created, and starts creating
nginx:1.16.1 Pods. It does not wait for the 5 replicas of nginx:1.14.2 to be created
before changing course.
Label selector updates
It is generally discouraged to make label selector updates and it is suggested to plan your selectors up front.
In any case, if you need to perform a label selector update, exercise great caution and make sure you have grasped
all of the implications.
Note: In API version apps/v1, a Deployment's label selector is immutable after it gets created.
Selector additions require the Pod template labels in the Deployment spec to be updated with the new label too,
otherwise a validation error is returned. This change is a non-overlapping one, meaning that the new selector does
not select ReplicaSets and Pods created with the old selector, resulting in orphaning all old ReplicaSets and
creating a new ReplicaSet.
Selector updates changes the existing value in a selector key -- result in the same behavior as additions.
Selector removals removes an existing key from the Deployment selector -- do not require any changes in the
Pod template labels. Existing ReplicaSets are not orphaned, and a new ReplicaSet is not created, but note that the
removed label still exists in any existing Pods and ReplicaSets.
Rolling Back a Deployment
Sometimes, you may want to rollback a Deployment; for example, when the Deployment is not stable, such as crash looping.
By default, all of the Deployment's rollout history is kept in the system so that you can rollback anytime you want
(you can change that by modifying revision history limit).
Note: A Deployment's revision is created when a Deployment's rollout is triggered. This means that the
new revision is created if and only if the Deployment's Pod template (.spec.template) is changed,
for example if you update the labels or container images of the template. Other updates, such as scaling the Deployment,
do not create a Deployment revision, so that you can facilitate simultaneous manual- or auto-scaling.
This means that when you roll back to an earlier revision, only the Deployment's Pod template part is
rolled back.
Suppose that you made a typo while updating the Deployment, by putting the image name as nginx:1.161 instead of nginx:1.16.1:
kubectl set image deployment/nginx-deployment nginx=nginx:1.161
The output is similar to this:
deployment.apps/nginx-deployment image updated
The rollout gets stuck. You can verify it by checking the rollout status:
kubectl rollout status deployment/nginx-deployment
The output is similar to this:
Waiting for rollout to finish: 1 out of 3 new replicas have been updated...
Press Ctrl-C to stop the above rollout status watch. For more information on stuck rollouts,
read more here.
You see that the number of old replicas (nginx-deployment-1564180365 and nginx-deployment-2035384211) is 2, and new replicas (nginx-deployment-3066724191) is 1.
kubectl get rs
The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-deployment-1564180365 3 3 3 25s
nginx-deployment-2035384211 0 0 0 36s
nginx-deployment-3066724191 1 1 0 6s
Looking at the Pods created, you see that 1 Pod created by new ReplicaSet is stuck in an image pull loop.
kubectl get pods
The output is similar to this:
NAME READY STATUS RESTARTS AGE
nginx-deployment-1564180365-70iae 1/1 Running 0 25s
nginx-deployment-1564180365-jbqqo 1/1 Running 0 25s
nginx-deployment-1564180365-hysrc 1/1 Running 0 25s
nginx-deployment-3066724191-08mng 0/1 ImagePullBackOff 0 6s
Note: The Deployment controller stops the bad rollout automatically, and stops scaling up the new ReplicaSet. This depends on the rollingUpdate parameters (maxUnavailable specifically) that you have specified. Kubernetes by default sets the value to 25%.
Get the description of the Deployment:
kubectl describe deployment
The output is similar to this:
Name: nginx-deployment
Namespace: default
CreationTimestamp: Tue, 15 Mar 2016 14:48:04 -0700
Labels: app=nginx
Selector: app=nginx
Replicas: 3 desired | 1 updated | 4 total | 3 available | 1 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=nginx
Containers:
nginx:
Image: nginx:1.161
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True ReplicaSetUpdated
OldReplicaSets: nginx-deployment-1564180365 (3/3 replicas created)
NewReplicaSet: nginx-deployment-3066724191 (1/1 replicas created)
Events:
FirstSeen LastSeen Count From SubObjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
1m 1m 1 {deployment-controller } Normal ScalingReplicaSet Scaled up replica set nginx-deployment-2035384211 to 3
22s 22s 1 {deployment-controller } Normal ScalingReplicaSet Scaled up replica set nginx-deployment-1564180365 to 1
22s 22s 1 {deployment-controller } Normal ScalingReplicaSet Scaled down replica set nginx-deployment-2035384211 to 2
22s 22s 1 {deployment-controller } Normal ScalingReplicaSet Scaled up replica set nginx-deployment-1564180365 to 2
21s 21s 1 {deployment-controller } Normal ScalingReplicaSet Scaled down replica set nginx-deployment-2035384211 to 1
21s 21s 1 {deployment-controller } Normal ScalingReplicaSet Scaled up replica set nginx-deployment-1564180365 to 3
13s 13s 1 {deployment-controller } Normal ScalingReplicaSet Scaled down replica set nginx-deployment-2035384211 to 0
13s 13s 1 {deployment-controller } Normal ScalingReplicaSet Scaled up replica set nginx-deployment-3066724191 to 1
To fix this, you need to rollback to a previous revision of Deployment that is stable.
Checking Rollout History of a Deployment
Follow the steps given below to check the rollout history:
First, check the revisions of this Deployment:
kubectl rollout history deployment/nginx-deployment
The output is similar to this:
deployments "nginx-deployment"
REVISION CHANGE-CAUSE
1 kubectl apply --filename=https://k8s.io/examples/controllers/nginx-deployment.yaml
2 kubectl set image deployment/nginx-deployment nginx=nginx:1.16.1
3 kubectl set image deployment/nginx-deployment nginx=nginx:1.161
CHANGE-CAUSE is copied from the Deployment annotation kubernetes.io/change-cause to its revisions upon creation. You can specify theCHANGE-CAUSE message by:
Annotating the Deployment with kubectl annotate deployment/nginx-deployment kubernetes.io/change-cause="image updated to 1.16.1"
Manually editing the manifest of the resource.
To see the details of each revision, run:
kubectl rollout history deployment/nginx-deployment --revision=2
For more details about rollout related commands, read kubectl rollout.
The Deployment is now rolled back to a previous stable revision. As you can see, a DeploymentRollback event
for rolling back to revision 2 is generated from Deployment controller.
Check if the rollback was successful and the Deployment is running as expected, run:
kubectl get deployment nginx-deployment
The output is similar to this:
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-deployment 3/3 3 3 30m
Get the description of the Deployment:
kubectl describe deployment nginx-deployment
The output is similar to this:
Name: nginx-deployment
Namespace: default
CreationTimestamp: Sun, 02 Sep 2018 18:17:55 -0500
Labels: app=nginx
Annotations: deployment.kubernetes.io/revision=4
kubernetes.io/change-cause=kubectl set image deployment/nginx-deployment nginx=nginx:1.16.1
Selector: app=nginx
Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: app=nginx
Containers:
nginx:
Image: nginx:1.16.1
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets: <none>
NewReplicaSet: nginx-deployment-c4747d96c (3/3 replicas created)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 12m deployment-controller Scaled up replica set nginx-deployment-75675f5897 to 3
Normal ScalingReplicaSet 11m deployment-controller Scaled up replica set nginx-deployment-c4747d96c to 1
Normal ScalingReplicaSet 11m deployment-controller Scaled down replica set nginx-deployment-75675f5897 to 2
Normal ScalingReplicaSet 11m deployment-controller Scaled up replica set nginx-deployment-c4747d96c to 2
Normal ScalingReplicaSet 11m deployment-controller Scaled down replica set nginx-deployment-75675f5897 to 1
Normal ScalingReplicaSet 11m deployment-controller Scaled up replica set nginx-deployment-c4747d96c to 3
Normal ScalingReplicaSet 11m deployment-controller Scaled down replica set nginx-deployment-75675f5897 to 0
Normal ScalingReplicaSet 11m deployment-controller Scaled up replica set nginx-deployment-595696685f to 1
Normal DeploymentRollback 15s deployment-controller Rolled back deployment "nginx-deployment" to revision 2
Normal ScalingReplicaSet 15s deployment-controller Scaled down replica set nginx-deployment-595696685f to 0
Scaling a Deployment
You can scale a Deployment by using the following command:
Assuming horizontal Pod autoscaling is enabled
in your cluster, you can set up an autoscaler for your Deployment and choose the minimum and maximum number of
Pods you want to run based on the CPU utilization of your existing Pods.
RollingUpdate Deployments support running multiple versions of an application at the same time. When you
or an autoscaler scales a RollingUpdate Deployment that is in the middle of a rollout (either in progress
or paused), the Deployment controller balances the additional replicas in the existing active
ReplicaSets (ReplicaSets with Pods) in order to mitigate risk. This is called proportional scaling.
For example, you are running a Deployment with 10 replicas, maxSurge=3, and maxUnavailable=2.
Ensure that the 10 replicas in your Deployment are running.
kubectl get deploy
The output is similar to this:
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
nginx-deployment 10 10 10 10 50s
You update to a new image which happens to be unresolvable from inside the cluster.
kubectl set image deployment/nginx-deployment nginx=nginx:sometag
The output is similar to this:
deployment.apps/nginx-deployment image updated
The image update starts a new rollout with ReplicaSet nginx-deployment-1989198191, but it's blocked due to the
maxUnavailable requirement that you mentioned above. Check out the rollout status:
kubectl get rs
The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-deployment-1989198191 5 5 0 9s
nginx-deployment-618515232 8 8 8 1m
Then a new scaling request for the Deployment comes along. The autoscaler increments the Deployment replicas
to 15. The Deployment controller needs to decide where to add these new 5 replicas. If you weren't using
proportional scaling, all 5 of them would be added in the new ReplicaSet. With proportional scaling, you
spread the additional replicas across all ReplicaSets. Bigger proportions go to the ReplicaSets with the
most replicas and lower proportions go to ReplicaSets with less replicas. Any leftovers are added to the
ReplicaSet with the most replicas. ReplicaSets with zero replicas are not scaled up.
In our example above, 3 replicas are added to the old ReplicaSet and 2 replicas are added to the
new ReplicaSet. The rollout process should eventually move all replicas to the new ReplicaSet, assuming
the new replicas become healthy. To confirm this, run:
kubectl get deploy
The output is similar to this:
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
nginx-deployment 15 18 7 8 7m
The rollout status confirms how the replicas were added to each ReplicaSet.
kubectl get rs
The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-deployment-1989198191 7 7 0 7m
nginx-deployment-618515232 11 11 11 7m
Pausing and Resuming a rollout of a Deployment
When you update a Deployment, or plan to, you can pause rollouts
for that Deployment before you trigger one or more updates. When
you're ready to apply those changes, you resume rollouts for the
Deployment. This approach allows you to
apply multiple fixes in between pausing and resuming without triggering unnecessary rollouts.
For example, with a Deployment that was created:
Get the Deployment details:
kubectl get deploy
The output is similar to this:
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
nginx 3 3 3 3 1m
Get the rollout status:
kubectl get rs
The output is similar to this:
NAME DESIRED CURRENT READY AGE
nginx-2142116321 3 3 3 1m
Pause by running the following command:
kubectl rollout pause deployment/nginx-deployment
The output is similar to this:
deployment.apps/nginx-deployment paused
Then update the image of the Deployment:
kubectl set image deployment/nginx-deployment nginx=nginx:1.16.1
The output is similar to this:
deployment.apps/nginx-deployment image updated
Notice that no new rollout started:
kubectl rollout history deployment/nginx-deployment
The initial state of the Deployment prior to pausing its rollout will continue its function, but new updates to
the Deployment will not have any effect as long as the Deployment rollout is paused.
Eventually, resume the Deployment rollout and observe a new ReplicaSet coming up with all the new updates:
NAME DESIRED CURRENT READY AGE
nginx-2142116321 0 0 0 2m
nginx-3926361531 3 3 3 28s
Note: You cannot rollback a paused Deployment until you resume it.
Deployment status
A Deployment enters various states during its lifecycle. It can be progressing while
rolling out a new ReplicaSet, it can be complete, or it can fail to progress.
Progressing Deployment
Kubernetes marks a Deployment as progressing when one of the following tasks is performed:
The Deployment creates a new ReplicaSet.
The Deployment is scaling up its newest ReplicaSet.
The Deployment is scaling down its older ReplicaSet(s).
New Pods become ready or available (ready for at least MinReadySeconds).
When the rollout becomes “progressing”, the Deployment controller adds a condition with the following
attributes to the Deployment's .status.conditions:
You can monitor the progress for a Deployment by using kubectl rollout status.
Complete Deployment
Kubernetes marks a Deployment as complete when it has the following characteristics:
All of the replicas associated with the Deployment have been updated to the latest version you've specified, meaning any
updates you've requested have been completed.
All of the replicas associated with the Deployment are available.
No old replicas for the Deployment are running.
When the rollout becomes “complete”, the Deployment controller sets a condition with the following
attributes to the Deployment's .status.conditions:
type: Progressing
status: "True"
reason: NewReplicaSetAvailable
This Progressing condition will retain a status value of "True" until a new rollout
is initiated. The condition holds even when availability of replicas changes (which
does instead affect the Available condition).
You can check if a Deployment has completed by using kubectl rollout status. If the rollout completed
successfully, kubectl rollout status returns a zero exit code.
kubectl rollout status deployment/nginx-deployment
The output is similar to this:
Waiting for rollout to finish: 2 of 3 updated replicas are available...
deployment "nginx-deployment" successfully rolled out
and the exit status from kubectl rollout is 0 (success):
echo$?
0
Failed Deployment
Your Deployment may get stuck trying to deploy its newest ReplicaSet without ever completing. This can occur
due to some of the following factors:
Insufficient quota
Readiness probe failures
Image pull errors
Insufficient permissions
Limit ranges
Application runtime misconfiguration
One way you can detect this condition is to specify a deadline parameter in your Deployment spec:
(.spec.progressDeadlineSeconds). .spec.progressDeadlineSeconds denotes the
number of seconds the Deployment controller waits before indicating (in the Deployment status) that the
Deployment progress has stalled.
The following kubectl command sets the spec with progressDeadlineSeconds to make the controller report
lack of progress of a rollout for a Deployment after 10 minutes:
Once the deadline has been exceeded, the Deployment controller adds a DeploymentCondition with the following
attributes to the Deployment's .status.conditions:
type: Progressing
status: "False"
reason: ProgressDeadlineExceeded
This condition can also fail early and is then set to status value of "False" due to reasons as ReplicaSetCreateError.
Also, the deadline is not taken into account anymore once the Deployment rollout completes.
Note: Kubernetes takes no action on a stalled Deployment other than to report a status condition with
reason: ProgressDeadlineExceeded. Higher level orchestrators can take advantage of it and act accordingly, for
example, rollback the Deployment to its previous version.
Note: If you pause a Deployment rollout, Kubernetes does not check progress against your specified deadline.
You can safely pause a Deployment rollout in the middle of a rollout and resume without triggering
the condition for exceeding the deadline.
You may experience transient errors with your Deployments, either due to a low timeout that you have set or
due to any other kind of error that can be treated as transient. For example, let's suppose you have
insufficient quota. If you describe the Deployment you will notice the following section:
kubectl describe deployment nginx-deployment
The output is similar to this:
<...>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True ReplicaSetUpdated
ReplicaFailure True FailedCreate
<...>
If you run kubectl get deployment nginx-deployment -o yaml, the Deployment status is similar to this:
Eventually, once the Deployment progress deadline is exceeded, Kubernetes updates the status and the
reason for the Progressing condition:
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing False ProgressDeadlineExceeded
ReplicaFailure True FailedCreate
You can address an issue of insufficient quota by scaling down your Deployment, by scaling down other
controllers you may be running, or by increasing quota in your namespace. If you satisfy the quota
conditions and the Deployment controller then completes the Deployment rollout, you'll see the
Deployment's status update with a successful condition (status: "True" and reason: NewReplicaSetAvailable).
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
type: Available with status: "True" means that your Deployment has minimum availability. Minimum availability is dictated
by the parameters specified in the deployment strategy. type: Progressing with status: "True" means that your Deployment
is either in the middle of a rollout and it is progressing or that it has successfully completed its progress and the minimum
required new replicas are available (see the Reason of the condition for the particulars - in our case
reason: NewReplicaSetAvailable means that the Deployment is complete).
You can check if a Deployment has failed to progress by using kubectl rollout status. kubectl rollout status
returns a non-zero exit code if the Deployment has exceeded the progression deadline.
kubectl rollout status deployment/nginx-deployment
The output is similar to this:
Waiting for rollout to finish: 2 out of 3 new replicas have been updated...
error: deployment "nginx" exceeded its progress deadline
and the exit status from kubectl rollout is 1 (indicating an error):
echo$?
1
Operating on a failed deployment
All actions that apply to a complete Deployment also apply to a failed Deployment. You can scale it up/down, roll back
to a previous revision, or even pause it if you need to apply multiple tweaks in the Deployment Pod template.
Clean up Policy
You can set .spec.revisionHistoryLimit field in a Deployment to specify how many old ReplicaSets for
this Deployment you want to retain. The rest will be garbage-collected in the background. By default,
it is 10.
Note: Explicitly setting this field to 0, will result in cleaning up all the history of your Deployment
thus that Deployment will not be able to roll back.
Canary Deployment
If you want to roll out releases to a subset of users or servers using the Deployment, you
can create multiple Deployments, one for each release, following the canary pattern described in
managing resources.
Writing a Deployment Spec
As with all other Kubernetes configs, a Deployment needs .apiVersion, .kind, and .metadata fields.
For general information about working with config files, see
deploying applications,
configuring containers, and using kubectl to manage resources documents.
The name of a Deployment object must be a valid
DNS subdomain name.
The .spec.template and .spec.selector are the only required fields of the .spec.
The .spec.template is a Pod template. It has exactly the same schema as a Pod, except it is nested and does not have an apiVersion or kind.
In addition to required fields for a Pod, a Pod template in a Deployment must specify appropriate
labels and an appropriate restart policy. For labels, make sure not to overlap with other controllers. See selector.
.spec.replicas is an optional field that specifies the number of desired Pods. It defaults to 1.
Should you manually scale a Deployment, example via kubectl scale deployment deployment --replicas=X, and then you update that Deployment based on a manifest
(for example: by running kubectl apply -f deployment.yaml),
then applying that manifest overwrites the manual scaling that you previously did.
If a HorizontalPodAutoscaler (or any
similar API for horizontal scaling) is managing scaling for a Deployment, don't set .spec.replicas.
Instead, allow the Kubernetes
control plane to manage the
.spec.replicas field automatically.
Selector
.spec.selector is a required field that specifies a label selector
for the Pods targeted by this Deployment.
.spec.selector must match .spec.template.metadata.labels, or it will be rejected by the API.
In API version apps/v1, .spec.selector and .metadata.labels do not default to .spec.template.metadata.labels if not set. So they must be set explicitly. Also note that .spec.selector is immutable after creation of the Deployment in apps/v1.
A Deployment may terminate Pods whose labels match the selector if their template is different
from .spec.template or if the total number of such Pods exceeds .spec.replicas. It brings up new
Pods with .spec.template if the number of Pods is less than the desired number.
Note: You should not create other Pods whose labels match this selector, either directly, by creating
another Deployment, or by creating another controller such as a ReplicaSet or a ReplicationController. If you
do so, the first Deployment thinks that it created these other Pods. Kubernetes does not stop you from doing this.
If you have multiple controllers that have overlapping selectors, the controllers will fight with each
other and won't behave correctly.
Strategy
.spec.strategy specifies the strategy used to replace old Pods by new ones.
.spec.strategy.type can be "Recreate" or "RollingUpdate". "RollingUpdate" is
the default value.
Recreate Deployment
All existing Pods are killed before new ones are created when .spec.strategy.type==Recreate.
Note: This will only guarantee Pod termination previous to creation for upgrades. If you upgrade a Deployment, all Pods
of the old revision will be terminated immediately. Successful removal is awaited before any Pod of the new
revision is created. If you manually delete a Pod, the lifecycle is controlled by the ReplicaSet and the
replacement will be created immediately (even if the old Pod is still in a Terminating state). If you need an
"at most" guarantee for your Pods, you should consider using a
StatefulSet.
Rolling Update Deployment
The Deployment updates Pods in a rolling update
fashion when .spec.strategy.type==RollingUpdate. You can specify maxUnavailable and maxSurge to control
the rolling update process.
Max Unavailable
.spec.strategy.rollingUpdate.maxUnavailable is an optional field that specifies the maximum number
of Pods that can be unavailable during the update process. The value can be an absolute number (for example, 5)
or a percentage of desired Pods (for example, 10%). The absolute number is calculated from percentage by
rounding down. The value cannot be 0 if .spec.strategy.rollingUpdate.maxSurge is 0. The default value is 25%.
For example, when this value is set to 30%, the old ReplicaSet can be scaled down to 70% of desired
Pods immediately when the rolling update starts. Once new Pods are ready, old ReplicaSet can be scaled
down further, followed by scaling up the new ReplicaSet, ensuring that the total number of Pods available
at all times during the update is at least 70% of the desired Pods.
Max Surge
.spec.strategy.rollingUpdate.maxSurge is an optional field that specifies the maximum number of Pods
that can be created over the desired number of Pods. The value can be an absolute number (for example, 5) or a
percentage of desired Pods (for example, 10%). The value cannot be 0 if MaxUnavailable is 0. The absolute number
is calculated from the percentage by rounding up. The default value is 25%.
For example, when this value is set to 30%, the new ReplicaSet can be scaled up immediately when the
rolling update starts, such that the total number of old and new Pods does not exceed 130% of desired
Pods. Once old Pods have been killed, the new ReplicaSet can be scaled up further, ensuring that the
total number of Pods running at any time during the update is at most 130% of desired Pods.
Progress Deadline Seconds
.spec.progressDeadlineSeconds is an optional field that specifies the number of seconds you want
to wait for your Deployment to progress before the system reports back that the Deployment has
failed progressing - surfaced as a condition with type: Progressing, status: "False".
and reason: ProgressDeadlineExceeded in the status of the resource. The Deployment controller will keep
retrying the Deployment. This defaults to 600. In the future, once automatic rollback will be implemented, the Deployment
controller will roll back a Deployment as soon as it observes such a condition.
If specified, this field needs to be greater than .spec.minReadySeconds.
Min Ready Seconds
.spec.minReadySeconds is an optional field that specifies the minimum number of seconds for which a newly
created Pod should be ready without any of its containers crashing, for it to be considered available.
This defaults to 0 (the Pod will be considered available as soon as it is ready). To learn more about when
a Pod is considered ready, see Container Probes.
Revision History Limit
A Deployment's revision history is stored in the ReplicaSets it controls.
.spec.revisionHistoryLimit is an optional field that specifies the number of old ReplicaSets to retain
to allow rollback. These old ReplicaSets consume resources in etcd and crowd the output of kubectl get rs. The configuration of each Deployment revision is stored in its ReplicaSets; therefore, once an old ReplicaSet is deleted, you lose the ability to rollback to that revision of Deployment. By default, 10 old ReplicaSets will be kept, however its ideal value depends on the frequency and stability of new Deployments.
More specifically, setting this field to zero means that all old ReplicaSets with 0 replicas will be cleaned up.
In this case, a new Deployment rollout cannot be undone, since its revision history is cleaned up.
Paused
.spec.paused is an optional boolean field for pausing and resuming a Deployment. The only difference between
a paused Deployment and one that is not paused, is that any changes into the PodTemplateSpec of the paused
Deployment will not trigger new rollouts as long as it is paused. A Deployment is not paused by default when
it is created.
Deployment is a top-level resource in the Kubernetes REST API.
Read the
Deployment
object definition to understand the API for deployments.
Read about PodDisruptionBudget and how
you can use it to manage application availability during disruptions.
3.5.2.2 - ReplicaSet
A ReplicaSet's purpose is to maintain a stable set of replica Pods running at any given time. As such, it is often
used to guarantee the availability of a specified number of identical Pods.
How a ReplicaSet works
A ReplicaSet is defined with fields, including a selector that specifies how to identify Pods it can acquire, a number
of replicas indicating how many Pods it should be maintaining, and a pod template specifying the data of new Pods
it should create to meet the number of replicas criteria. A ReplicaSet then fulfills its purpose by creating
and deleting Pods as needed to reach the desired number. When a ReplicaSet needs to create new Pods, it uses its Pod
template.
A ReplicaSet is linked to its Pods via the Pods' metadata.ownerReferences
field, which specifies what resource the current object is owned by. All Pods acquired by a ReplicaSet have their owning
ReplicaSet's identifying information within their ownerReferences field. It's through this link that the ReplicaSet
knows of the state of the Pods it is maintaining and plans accordingly.
A ReplicaSet identifies new Pods to acquire by using its selector. If there is a Pod that has no
OwnerReference or the OwnerReference is not a Controller and it
matches a ReplicaSet's selector, it will be immediately acquired by said ReplicaSet.
When to use a ReplicaSet
A ReplicaSet ensures that a specified number of pod replicas are running at any given
time. However, a Deployment is a higher-level concept that manages ReplicaSets and
provides declarative updates to Pods along with a lot of other useful features.
Therefore, we recommend using Deployments instead of directly using ReplicaSets, unless
you require custom update orchestration or don't require updates at all.
This actually means that you may never need to manipulate ReplicaSet objects:
use a Deployment instead, and define your application in the spec section.
apiVersion:apps/v1kind:ReplicaSetmetadata:name:frontendlabels:app:guestbooktier:frontendspec:# modify replicas according to your casereplicas:3selector:matchLabels:tier:frontendtemplate:metadata:labels:tier:frontendspec:containers:- name:php-redisimage:gcr.io/google_samples/gb-frontend:v3
Saving this manifest into frontend.yaml and submitting it to a Kubernetes cluster will
create the defined ReplicaSet and the Pods that it manages.
While you can create bare Pods with no problems, it is strongly recommended to make sure that the bare Pods do not have
labels which match the selector of one of your ReplicaSets. The reason for this is because a ReplicaSet is not limited
to owning Pods specified by its template-- it can acquire other Pods in the manner specified in the previous sections.
Take the previous frontend ReplicaSet example, and the Pods specified in the following manifest:
As those Pods do not have a Controller (or any object) as their owner reference and match the selector of the frontend
ReplicaSet, they will immediately be acquired by it.
Suppose you create the Pods after the frontend ReplicaSet has been deployed and has set up its initial Pod replicas to
fulfill its replica count requirement:
You shall see that the ReplicaSet has acquired the Pods and has only created new ones according to its spec until the
number of its new Pods and the original matches its desired count. As fetching the Pods:
kubectl get pods
Will reveal in its output:
NAME READY STATUS RESTARTS AGE
frontend-hmmj2 1/1 Running 0 9s
pod1 1/1 Running 0 36s
pod2 1/1 Running 0 36s
In this manner, a ReplicaSet can own a non-homogenous set of Pods
Writing a ReplicaSet manifest
As with all other Kubernetes API objects, a ReplicaSet needs the apiVersion, kind, and metadata fields.
For ReplicaSets, the kind is always a ReplicaSet.
The .spec.template is a pod template which is also
required to have labels in place. In our frontend.yaml example we had one label: tier: frontend.
Be careful not to overlap with the selectors of other controllers, lest they try to adopt this Pod.
For the template's restart policy field,
.spec.template.spec.restartPolicy, the only allowed value is Always, which is the default.
Pod Selector
The .spec.selector field is a label selector. As discussed
earlier these are the labels used to identify potential Pods to acquire. In our
frontend.yaml example, the selector was:
matchLabels:tier:frontend
In the ReplicaSet, .spec.template.metadata.labels must match spec.selector, or it will
be rejected by the API.
Note: For 2 ReplicaSets specifying the same .spec.selector but different
.spec.template.metadata.labels and .spec.template.spec fields, each ReplicaSet ignores the
Pods created by the other ReplicaSet.
Replicas
You can specify how many Pods should run concurrently by setting .spec.replicas. The ReplicaSet will create/delete
its Pods to match this number.
If you do not specify .spec.replicas, then it defaults to 1.
Working with ReplicaSets
Deleting a ReplicaSet and its Pods
To delete a ReplicaSet and all of its Pods, use
kubectl delete. The
Garbage collector automatically deletes all of
the dependent Pods by default.
When using the REST API or the client-go library, you must set propagationPolicy to
Background or Foreground in the -d option. For example:
You can delete a ReplicaSet without affecting any of its Pods using
kubectl delete
with the --cascade=orphan option.
When using the REST API or the client-go library, you must set propagationPolicy to Orphan.
For example:
Once the original is deleted, you can create a new ReplicaSet to replace it. As long
as the old and new .spec.selector are the same, then the new one will adopt the old Pods.
However, it will not make any effort to make existing Pods match a new, different pod template.
To update Pods to a new spec in a controlled way, use a
Deployment, as
ReplicaSets do not support a rolling update directly.
Isolating Pods from a ReplicaSet
You can remove Pods from a ReplicaSet by changing their labels. This technique may be used to remove Pods
from service for debugging, data recovery, etc. Pods that are removed in this way will be replaced automatically (
assuming that the number of replicas is not also changed).
Scaling a ReplicaSet
A ReplicaSet can be easily scaled up or down by simply updating the .spec.replicas field. The ReplicaSet controller
ensures that a desired number of Pods with a matching label selector are available and operational.
When scaling down, the ReplicaSet controller chooses which pods to delete by sorting the available pods to
prioritize scaling down pods based on the following general algorithm:
Pending (and unschedulable) pods are scaled down first
If controller.kubernetes.io/pod-deletion-cost annotation is set, then
the pod with the lower value will come first.
Pods on nodes with more replicas come before pods on nodes with fewer replicas.
If the pods' creation times differ, the pod that was created more recently
comes before the older pod (the creation times are bucketed on an integer log scale
when the LogarithmicScaleDownfeature gate is enabled)
If all of the above match, then selection is random.
The annotation should be set on the pod, the range is [-2147483647, 2147483647]. It represents the cost of
deleting a pod compared to other pods belonging to the same ReplicaSet. Pods with lower deletion
cost are preferred to be deleted before pods with higher deletion cost.
The implicit value for this annotation for pods that don't set it is 0; negative values are permitted.
Invalid values will be rejected by the API server.
This feature is beta and enabled by default. You can disable it using the
feature gatePodDeletionCost in both kube-apiserver and kube-controller-manager.
Note:
This is honored on a best-effort basis, so it does not offer any guarantees on pod deletion order.
Users should avoid updating the annotation frequently, such as updating it based on a metric value,
because doing so will generate a significant number of pod updates on the apiserver.
Example Use Case
The different pods of an application could have different utilization levels. On scale down, the application
may prefer to remove the pods with lower utilization. To avoid frequently updating the pods, the application
should update controller.kubernetes.io/pod-deletion-cost once before issuing a scale down (setting the
annotation to a value proportional to pod utilization level). This works if the application itself controls
the down scaling; for example, the driver pod of a Spark deployment.
ReplicaSet as a Horizontal Pod Autoscaler Target
A ReplicaSet can also be a target for
Horizontal Pod Autoscalers (HPA). That is,
a ReplicaSet can be auto-scaled by an HPA. Here is an example HPA targeting
the ReplicaSet we created in the previous example.
Saving this manifest into hpa-rs.yaml and submitting it to a Kubernetes cluster should
create the defined HPA that autoscales the target ReplicaSet depending on the CPU usage
of the replicated Pods.
Deployment is an object which can own ReplicaSets and update
them and their Pods via declarative, server-side rolling updates.
While ReplicaSets can be used independently, today they're mainly used by Deployments as a mechanism to orchestrate Pod
creation, deletion and updates. When you use Deployments you don't have to worry about managing the ReplicaSets that
they create. Deployments own and manage their ReplicaSets.
As such, it is recommended to use Deployments when you want ReplicaSets.
Bare Pods
Unlike the case where a user directly created Pods, a ReplicaSet replaces Pods that are deleted or
terminated for any reason, such as in the case of node failure or disruptive node maintenance,
such as a kernel upgrade. For this reason, we recommend that you use a ReplicaSet even if your
application requires only a single Pod. Think of it similarly to a process supervisor, only it
supervises multiple Pods across multiple nodes instead of individual processes on a single node. A
ReplicaSet delegates local container restarts to some agent on the node such as Kubelet.
Job
Use a Job instead of a ReplicaSet for Pods that are
expected to terminate on their own (that is, batch jobs).
DaemonSet
Use a DaemonSet instead of a ReplicaSet for Pods that provide a
machine-level function, such as machine monitoring or machine logging. These Pods have a lifetime that is tied
to a machine lifetime: the Pod needs to be running on the machine before other Pods start, and are
safe to terminate when the machine is otherwise ready to be rebooted/shutdown.
ReplicationController
ReplicaSets are the successors to ReplicationControllers.
The two serve the same purpose, and behave similarly, except that a ReplicationController does not support set-based
selector requirements as described in the labels user guide.
As such, ReplicaSets are preferred over ReplicationControllers
ReplicaSet is a top-level resource in the Kubernetes REST API.
Read the
ReplicaSet
object definition to understand the API for replica sets.
Read about PodDisruptionBudget and how
you can use it to manage application availability during disruptions.
3.5.2.3 - StatefulSets
StatefulSet is the workload API object used to manage stateful applications.
Manages the deployment and scaling of a set of Pods, and provides guarantees about the ordering and uniqueness of these Pods.
Like a Deployment, a StatefulSet manages Pods that are based on an identical container spec. Unlike a Deployment, a StatefulSet maintains a sticky identity for each of their Pods. These pods are created from the same spec, but are not interchangeable: each has a persistent identifier that it maintains across any rescheduling.
If you want to use storage volumes to provide persistence for your workload, you can use a StatefulSet as part of the solution. Although individual Pods in a StatefulSet are susceptible to failure, the persistent Pod identifiers make it easier to match existing volumes to the new Pods that replace any that have failed.
Using StatefulSets
StatefulSets are valuable for applications that require one or more of the
following.
Stable, unique network identifiers.
Stable, persistent storage.
Ordered, graceful deployment and scaling.
Ordered, automated rolling updates.
In the above, stable is synonymous with persistence across Pod (re)scheduling.
If an application doesn't require any stable identifiers or ordered deployment,
deletion, or scaling, you should deploy your application using a workload object
that provides a set of stateless replicas.
Deployment or
ReplicaSet may be better suited to your stateless needs.
Limitations
The storage for a given Pod must either be provisioned by a
PersistentVolume Provisioner
based on the requested storage class, or pre-provisioned by an admin.
Deleting and/or scaling a StatefulSet down will not delete the volumes associated with the
StatefulSet. This is done to ensure data safety, which is generally more valuable than an
automatic purge of all related StatefulSet resources.
StatefulSets currently require a Headless Service
to be responsible for the network identity of the Pods. You are responsible for creating this
Service.
StatefulSets do not provide any guarantees on the termination of pods when a StatefulSet is
deleted. To achieve ordered and graceful termination of the pods in the StatefulSet, it is
possible to scale the StatefulSet down to 0 prior to deletion.
The example below demonstrates the components of a StatefulSet.
apiVersion:v1kind:Servicemetadata:name:nginxlabels:app:nginxspec:ports:- port:80name:webclusterIP:Noneselector:app:nginx---apiVersion:apps/v1kind:StatefulSetmetadata:name:webspec:selector:matchLabels:app:nginx# has to match .spec.template.metadata.labelsserviceName:"nginx"replicas:3# by default is 1minReadySeconds:10# by default is 0template:metadata:labels:app:nginx# has to match .spec.selector.matchLabelsspec:terminationGracePeriodSeconds:10containers:- name:nginximage:registry.k8s.io/nginx-slim:0.8ports:- containerPort:80name:webvolumeMounts:- name:wwwmountPath:/usr/share/nginx/htmlvolumeClaimTemplates:- metadata:name:wwwspec:accessModes:["ReadWriteOnce"]storageClassName:"my-storage-class"resources:requests:storage:1Gi
In the above example:
A Headless Service, named nginx, is used to control the network domain.
The StatefulSet, named web, has a Spec that indicates that 3 replicas of the nginx container will be launched in unique Pods.
The volumeClaimTemplates will provide stable storage using
PersistentVolumes provisioned by a
PersistentVolume Provisioner.
You must set the .spec.selector field of a StatefulSet to match the labels of its
.spec.template.metadata.labels. Failing to specify a matching Pod Selector will result in a
validation error during StatefulSet creation.
Volume Claim Templates
You can set the .spec.volumeClaimTemplates which can provide stable storage using
PersistentVolumes provisioned by a PersistentVolume
Provisioner.
Minimum ready seconds
FEATURE STATE:Kubernetes v1.25 [stable]
.spec.minReadySeconds is an optional field that specifies the minimum number of seconds for which a newly
created Pod should be running and ready without any of its containers crashing, for it to be considered available.
This is used to check progression of a rollout when using a Rolling Update strategy.
This field defaults to 0 (the Pod will be considered available as soon as it is ready). To learn more about when
a Pod is considered ready, see Container Probes.
Pod Identity
StatefulSet Pods have a unique identity that is comprised of an ordinal, a
stable network identity, and stable storage. The identity sticks to the Pod,
regardless of which node it's (re)scheduled on.
Ordinal Index
For a StatefulSet with N replicas, each Pod in the StatefulSet will be
assigned an integer ordinal, from 0 up through N-1, that is unique over the Set.
Stable Network ID
Each Pod in a StatefulSet derives its hostname from the name of the StatefulSet
and the ordinal of the Pod. The pattern for the constructed hostname
is $(statefulset name)-$(ordinal). The example above will create three Pods
named web-0,web-1,web-2.
A StatefulSet can use a Headless Service
to control the domain of its Pods. The domain managed by this Service takes the form:
$(service name).$(namespace).svc.cluster.local, where "cluster.local" is the
cluster domain.
As each Pod is created, it gets a matching DNS subdomain, taking the form:
$(podname).$(governing service domain), where the governing service is defined
by the serviceName field on the StatefulSet.
Depending on how DNS is configured in your cluster, you may not be able to look up the DNS
name for a newly-run Pod immediately. This behavior can occur when other clients in the
cluster have already sent queries for the hostname of the Pod before it was created.
Negative caching (normal in DNS) means that the results of previous failed lookups are
remembered and reused, even after the Pod is running, for at least a few seconds.
If you need to discover Pods promptly after they are created, you have a few options:
Query the Kubernetes API directly (for example, using a watch) rather than relying on DNS lookups.
Decrease the time of caching in your Kubernetes DNS provider (typically this means editing the
config map for CoreDNS, which currently caches for 30 seconds).
As mentioned in the limitations section, you are responsible for
creating the Headless Service
responsible for the network identity of the pods.
Here are some examples of choices for Cluster Domain, Service name,
StatefulSet name, and how that affects the DNS names for the StatefulSet's Pods.
For each VolumeClaimTemplate entry defined in a StatefulSet, each Pod receives one
PersistentVolumeClaim. In the nginx example above, each Pod receives a single PersistentVolume
with a StorageClass of my-storage-class and 1 Gib of provisioned storage. If no StorageClass
is specified, then the default StorageClass will be used. When a Pod is (re)scheduled
onto a node, its volumeMounts mount the PersistentVolumes associated with its
PersistentVolume Claims. Note that, the PersistentVolumes associated with the
Pods' PersistentVolume Claims are not deleted when the Pods, or StatefulSet are deleted.
This must be done manually.
Pod Name Label
When the StatefulSet Controller creates a Pod,
it adds a label, statefulset.kubernetes.io/pod-name, that is set to the name of
the Pod. This label allows you to attach a Service to a specific Pod in
the StatefulSet.
Deployment and Scaling Guarantees
For a StatefulSet with N replicas, when Pods are being deployed, they are created sequentially, in order from {0..N-1}.
When Pods are being deleted, they are terminated in reverse order, from {N-1..0}.
Before a scaling operation is applied to a Pod, all of its predecessors must be Running and Ready.
Before a Pod is terminated, all of its successors must be completely shutdown.
The StatefulSet should not specify a pod.Spec.TerminationGracePeriodSeconds of 0. This practice
is unsafe and strongly discouraged. For further explanation, please refer to
force deleting StatefulSet Pods.
When the nginx example above is created, three Pods will be deployed in the order
web-0, web-1, web-2. web-1 will not be deployed before web-0 is
Running and Ready, and web-2 will not be deployed until
web-1 is Running and Ready. If web-0 should fail, after web-1 is Running and Ready, but before
web-2 is launched, web-2 will not be launched until web-0 is successfully relaunched and
becomes Running and Ready.
If a user were to scale the deployed example by patching the StatefulSet such that
replicas=1, web-2 would be terminated first. web-1 would not be terminated until web-2
is fully shutdown and deleted. If web-0 were to fail after web-2 has been terminated and
is completely shutdown, but prior to web-1's termination, web-1 would not be terminated
until web-0 is Running and Ready.
Pod Management Policies
StatefulSet allows you to relax its ordering guarantees while
preserving its uniqueness and identity guarantees via its .spec.podManagementPolicy field.
OrderedReady Pod Management
OrderedReady pod management is the default for StatefulSets. It implements the behavior
described above.
Parallel Pod Management
Parallel pod management tells the StatefulSet controller to launch or
terminate all Pods in parallel, and to not wait for Pods to become Running
and Ready or completely terminated prior to launching or terminating another
Pod. This option only affects the behavior for scaling operations. Updates are not
affected.
Update strategies
A StatefulSet's .spec.updateStrategy field allows you to configure
and disable automated rolling updates for containers, labels, resource request/limits, and
annotations for the Pods in a StatefulSet. There are two possible values:
OnDelete
When a StatefulSet's .spec.updateStrategy.type is set to OnDelete,
the StatefulSet controller will not automatically update the Pods in a
StatefulSet. Users must manually delete Pods to cause the controller to
create new Pods that reflect modifications made to a StatefulSet's .spec.template.
RollingUpdate
The RollingUpdate update strategy implements automated, rolling update for the Pods in a
StatefulSet. This is the default update strategy.
Rolling Updates
When a StatefulSet's .spec.updateStrategy.type is set to RollingUpdate, the
StatefulSet controller will delete and recreate each Pod in the StatefulSet. It will proceed
in the same order as Pod termination (from the largest ordinal to the smallest), updating
each Pod one at a time.
The Kubernetes control plane waits until an updated Pod is Running and Ready prior
to updating its predecessor. If you have set .spec.minReadySeconds (see
Minimum Ready Seconds), the control plane additionally waits that
amount of time after the Pod turns ready, before moving on.
Partitioned rolling updates
The RollingUpdate update strategy can be partitioned, by specifying a
.spec.updateStrategy.rollingUpdate.partition. If a partition is specified, all Pods with an
ordinal that is greater than or equal to the partition will be updated when the StatefulSet's
.spec.template is updated. All Pods with an ordinal that is less than the partition will not
be updated, and, even if they are deleted, they will be recreated at the previous version. If a
StatefulSet's .spec.updateStrategy.rollingUpdate.partition is greater than its .spec.replicas,
updates to its .spec.template will not be propagated to its Pods.
In most cases you will not need to use a partition, but they are useful if you want to stage an
update, roll out a canary, or perform a phased roll out.
Maximum unavailable Pods
FEATURE STATE:Kubernetes v1.24 [alpha]
You can control the maximum number of Pods that can be unavailable during an update
by specifying the .spec.updateStrategy.rollingUpdate.maxUnavailable field.
The value can be an absolute number (for example, 5) or a percentage of desired
Pods (for example, 10%). Absolute number is calculated from the percentage value
by rounding it up. This field cannot be 0. The default setting is 1.
This field applies to all Pods in the range 0 to replicas - 1. If there is any
unavailable Pod in the range 0 to replicas - 1, it will be counted towards
maxUnavailable.
Note: The maxUnavailable field is in Alpha stage and it is honored only by API servers
that are running with the MaxUnavailableStatefulSetfeature gate
enabled.
Forced rollback
When using Rolling Updates with the default
Pod Management Policy (OrderedReady),
it's possible to get into a broken state that requires manual intervention to repair.
If you update the Pod template to a configuration that never becomes Running and
Ready (for example, due to a bad binary or application-level configuration error),
StatefulSet will stop the rollout and wait.
In this state, it's not enough to revert the Pod template to a good configuration.
Due to a known issue,
StatefulSet will continue to wait for the broken Pod to become Ready
(which never happens) before it will attempt to revert it back to the working
configuration.
After reverting the template, you must also delete any Pods that StatefulSet had
already attempted to run with the bad configuration.
StatefulSet will then begin to recreate the Pods using the reverted template.
PersistentVolumeClaim retention
FEATURE STATE:Kubernetes v1.23 [alpha]
The optional .spec.persistentVolumeClaimRetentionPolicy field controls if
and how PVCs are deleted during the lifecycle of a StatefulSet. You must enable the
StatefulSetAutoDeletePVCfeature gate
to use this field. Once enabled, there are two policies you can configure for each
StatefulSet:
whenDeleted
configures the volume retention behavior that applies when the StatefulSet is deleted
whenScaled
configures the volume retention behavior that applies when the replica count of
the StatefulSet is reduced; for example, when scaling down the set.
For each policy that you can configure, you can set the value to either Delete or Retain.
Delete
The PVCs created from the StatefulSet volumeClaimTemplate are deleted for each Pod
affected by the policy. With the whenDeleted policy all PVCs from the
volumeClaimTemplate are deleted after their Pods have been deleted. With the
whenScaled policy, only PVCs corresponding to Pod replicas being scaled down are
deleted, after their Pods have been deleted.
Retain (default)
PVCs from the volumeClaimTemplate are not affected when their Pod is
deleted. This is the behavior before this new feature.
Bear in mind that these policies only apply when Pods are being removed due to the
StatefulSet being deleted or scaled down. For example, if a Pod associated with a StatefulSet
fails due to node failure, and the control plane creates a replacement Pod, the StatefulSet
retains the existing PVC. The existing volume is unaffected, and the cluster will attach it to
the node where the new Pod is about to launch.
The default for policies is Retain, matching the StatefulSet behavior before this new feature.
The StatefulSet controller adds
owner references
to its PVCs, which are then deleted by the garbage collector after the Pod is terminated. This enables the Pod to
cleanly unmount all volumes before the PVCs are deleted (and before the backing PV and
volume are deleted, depending on the retain policy). When you set the whenDeleted
policy to Delete, an owner reference to the StatefulSet instance is placed on all PVCs
associated with that StatefulSet.
The whenScaled policy must delete PVCs only when a Pod is scaled down, and not when a
Pod is deleted for another reason. When reconciling, the StatefulSet controller compares
its desired replica count to the actual Pods present on the cluster. Any StatefulSet Pod
whose id greater than the replica count is condemned and marked for deletion. If the
whenScaled policy is Delete, the condemned Pods are first set as owners to the
associated StatefulSet template PVCs, before the Pod is deleted. This causes the PVCs
to be garbage collected after only the condemned Pods have terminated.
This means that if the controller crashes and restarts, no Pod will be deleted before its
owner reference has been updated appropriate to the policy. If a condemned Pod is
force-deleted while the controller is down, the owner reference may or may not have been
set up, depending on when the controller crashed. It may take several reconcile loops to
update the owner references, so some condemned Pods may have set up owner references and
other may not. For this reason we recommend waiting for the controller to come back up,
which will verify owner references before terminating Pods. If that is not possible, the
operator should verify the owner references on PVCs to ensure the expected objects are
deleted when Pods are force-deleted.
Replicas
.spec.replicas is an optional field that specifies the number of desired Pods. It defaults to 1.
Should you manually scale a deployment, example via kubectl scale statefulset statefulset --replicas=X, and then you update that StatefulSet
based on a manifest (for example: by running kubectl apply -f statefulset.yaml), then applying that manifest overwrites the manual scaling
that you previously did.
If a HorizontalPodAutoscaler
(or any similar API for horizontal scaling) is managing scaling for a
Statefulset, don't set .spec.replicas. Instead, allow the Kubernetes
control plane to manage
the .spec.replicas field automatically.
StatefulSet is a top-level resource in the Kubernetes REST API.
Read the
StatefulSet
object definition to understand the API for stateful sets.
Read about PodDisruptionBudget and how
you can use it to manage application availability during disruptions.
3.5.2.4 - DaemonSet
A DaemonSet ensures that all (or some) Nodes run a copy of a Pod. As nodes are added to the
cluster, Pods are added to them. As nodes are removed from the cluster, those Pods are garbage
collected. Deleting a DaemonSet will clean up the Pods it created.
Some typical uses of a DaemonSet are:
running a cluster storage daemon on every node
running a logs collection daemon on every node
running a node monitoring daemon on every node
In a simple case, one DaemonSet, covering all nodes, would be used for each type of daemon.
A more complex setup might use multiple DaemonSets for a single type of daemon, but with
different flags and/or different memory and cpu requests for different hardware types.
Writing a DaemonSet Spec
Create a DaemonSet
You can describe a DaemonSet in a YAML file. For example, the daemonset.yaml file below
describes a DaemonSet that runs the fluentd-elasticsearch Docker image:
apiVersion:apps/v1kind:DaemonSetmetadata:name:fluentd-elasticsearchnamespace:kube-systemlabels:k8s-app:fluentd-loggingspec:selector:matchLabels:name:fluentd-elasticsearchtemplate:metadata:labels:name:fluentd-elasticsearchspec:tolerations:# these tolerations are to have the daemonset runnable on control plane nodes# remove them if your control plane nodes should not run pods- key:node-role.kubernetes.io/control-planeoperator:Existseffect:NoSchedule- key:node-role.kubernetes.io/masteroperator:Existseffect:NoSchedulecontainers:- name:fluentd-elasticsearchimage:quay.io/fluentd_elasticsearch/fluentd:v2.5.2resources:limits:memory:200Mirequests:cpu:100mmemory:200MivolumeMounts:- name:varlogmountPath:/var/log- name:varlibdockercontainersmountPath:/var/lib/docker/containersreadOnly:trueterminationGracePeriodSeconds:30volumes:- name:varloghostPath:path:/var/log- name:varlibdockercontainershostPath:path:/var/lib/docker/containers
The .spec.template is one of the required fields in .spec.
The .spec.template is a pod template.
It has exactly the same schema as a Pod,
except it is nested and does not have an apiVersion or kind.
In addition to required fields for a Pod, a Pod template in a DaemonSet has to specify appropriate
labels (see pod selector).
A Pod Template in a DaemonSet must have a RestartPolicy
equal to Always, or be unspecified, which defaults to Always.
Pod Selector
The .spec.selector field is a pod selector. It works the same as the .spec.selector of
a Job.
You must specify a pod selector that matches the labels of the
.spec.template.
Also, once a DaemonSet is created,
its .spec.selector can not be mutated. Mutating the pod selector can lead to the
unintentional orphaning of Pods, and it was found to be confusing to users.
The .spec.selector is an object consisting of two fields:
matchExpressions - allows to build more sophisticated selectors by specifying key,
list of values and an operator that relates the key and values.
When the two are specified the result is ANDed.
The .spec.selector must match the .spec.template.metadata.labels.
Config with these two not matching will be rejected by the API.
Running Pods on select Nodes
If you specify a .spec.template.spec.nodeSelector, then the DaemonSet controller will
create Pods on nodes which match that node selector.
Likewise if you specify a .spec.template.spec.affinity,
then DaemonSet controller will create Pods on nodes which match that
node affinity.
If you do not specify either, then the DaemonSet controller will create Pods on all nodes.
How Daemon Pods are scheduled
Scheduled by default scheduler
FEATURE STATE:Kubernetes 1.17 [stable]
A DaemonSet ensures that all eligible nodes run a copy of a Pod. Normally, the
node that a Pod runs on is selected by the Kubernetes scheduler. However,
DaemonSet pods are created and scheduled by the DaemonSet controller instead.
That introduces the following issues:
Inconsistent Pod behavior: Normal Pods waiting to be scheduled are created
and in Pending state, but DaemonSet pods are not created in Pending
state. This is confusing to the user.
Pod preemption
is handled by default scheduler. When preemption is enabled, the DaemonSet controller
will make scheduling decisions without considering pod priority and preemption.
ScheduleDaemonSetPods allows you to schedule DaemonSets using the default
scheduler instead of the DaemonSet controller, by adding the NodeAffinity term
to the DaemonSet pods, instead of the .spec.nodeName term. The default
scheduler is then used to bind the pod to the target host. If node affinity of
the DaemonSet pod already exists, it is replaced (the original node affinity was
taken into account before selecting the target host). The DaemonSet controller only
performs these operations when creating or modifying DaemonSet pods, and no
changes are made to the spec.template of the DaemonSet.
In addition, node.kubernetes.io/unschedulable:NoSchedule toleration is added
automatically to DaemonSet Pods. The default scheduler ignores
unschedulable Nodes when scheduling DaemonSet Pods.
Taints and Tolerations
Although Daemon Pods respect
taints and tolerations,
the following tolerations are added to DaemonSet Pods automatically according to
the related features.
Toleration Key
Effect
Version
Description
node.kubernetes.io/not-ready
NoExecute
1.13+
DaemonSet pods will not be evicted when there are node problems such as a network partition.
node.kubernetes.io/unreachable
NoExecute
1.13+
DaemonSet pods will not be evicted when there are node problems such as a network partition.
node.kubernetes.io/disk-pressure
NoSchedule
1.8+
DaemonSet pods tolerate disk-pressure attributes by default scheduler.
node.kubernetes.io/memory-pressure
NoSchedule
1.8+
DaemonSet pods tolerate memory-pressure attributes by default scheduler.
node.kubernetes.io/unschedulable
NoSchedule
1.12+
DaemonSet pods tolerate unschedulable attributes by default scheduler.
node.kubernetes.io/network-unavailable
NoSchedule
1.12+
DaemonSet pods, who uses host network, tolerate network-unavailable attributes by default scheduler.
Communicating with Daemon Pods
Some possible patterns for communicating with Pods in a DaemonSet are:
Push: Pods in the DaemonSet are configured to send updates to another service, such
as a stats database. They do not have clients.
NodeIP and Known Port: Pods in the DaemonSet can use a hostPort, so that the pods
are reachable via the node IPs.
Clients know the list of node IPs somehow, and know the port by convention.
DNS: Create a headless service
with the same pod selector, and then discover DaemonSets using the endpoints
resource or retrieve multiple A records from DNS.
Service: Create a service with the same Pod selector, and use the service to reach a
daemon on a random node. (No way to reach specific node.)
Updating a DaemonSet
If node labels are changed, the DaemonSet will promptly add Pods to newly matching nodes and delete
Pods from newly not-matching nodes.
You can modify the Pods that a DaemonSet creates. However, Pods do not allow all
fields to be updated. Also, the DaemonSet controller will use the original template the next
time a node (even with the same name) is created.
You can delete a DaemonSet. If you specify --cascade=orphan with kubectl, then the Pods
will be left on the nodes. If you subsequently create a new DaemonSet with the same selector,
the new DaemonSet adopts the existing Pods. If any Pods need replacing the DaemonSet replaces
them according to its updateStrategy.
It is certainly possible to run daemon processes by directly starting them on a node (e.g. using
init, upstartd, or systemd). This is perfectly fine. However, there are several advantages to
running such processes via a DaemonSet:
Ability to monitor and manage logs for daemons in the same way as applications.
Same config language and tools (e.g. Pod templates, kubectl) for daemons and applications.
Running daemons in containers with resource limits increases isolation between daemons from app
containers. However, this can also be accomplished by running the daemons in a container but not in a Pod
(e.g. start directly via Docker).
Bare Pods
It is possible to create Pods directly which specify a particular node to run on. However,
a DaemonSet replaces Pods that are deleted or terminated for any reason, such as in the case of
node failure or disruptive node maintenance, such as a kernel upgrade. For this reason, you should
use a DaemonSet rather than creating individual Pods.
Static Pods
It is possible to create Pods by writing a file to a certain directory watched by Kubelet. These
are called static pods.
Unlike DaemonSet, static Pods cannot be managed with kubectl
or other Kubernetes API clients. Static Pods do not depend on the apiserver, making them useful
in cluster bootstrapping cases. Also, static Pods may be deprecated in the future.
Deployments
DaemonSets are similar to Deployments in that
they both create Pods, and those Pods have processes which are not expected to terminate (e.g. web servers,
storage servers).
Use a Deployment for stateless services, like frontends, where scaling up and down the
number of replicas and rolling out updates are more important than controlling exactly which host
the Pod runs on. Use a DaemonSet when it is important that a copy of a Pod always run on
all or certain hosts, if the DaemonSet provides node-level functionality that allows other Pods to run correctly on that particular node.
For example, network plugins often include a component that runs as a DaemonSet. The DaemonSet component makes sure that the node where it's running has working cluster networking.
DaemonSet is a top-level resource in the Kubernetes REST API.
Read the
DaemonSet
object definition to understand the API for daemon sets.
3.5.2.5 - Jobs
A Job creates one or more Pods and will continue to retry execution of the Pods until a specified number of them successfully terminate.
As pods successfully complete, the Job tracks the successful completions. When a specified number
of successful completions is reached, the task (ie, Job) is complete. Deleting a Job will clean up
the Pods it created. Suspending a Job will delete its active Pods until the Job
is resumed again.
A simple case is to create one Job object in order to reliably run one Pod to completion.
The Job object will start a new Pod if the first Pod fails or is deleted (for example
due to a node hardware failure or a node reboot).
You can also use a Job to run multiple Pods in parallel.
If you want to run a Job (either a single task, or several in parallel) on a schedule,
see CronJob.
Running an example Job
Here is an example Job config. It computes π to 2000 places and prints it out.
It takes around 10s to complete.
To view completed Pods of a Job, use kubectl get pods.
To list all the Pods that belong to a Job in a machine readable form, you can use a command like this:
pods=$(kubectl get pods --selector=job-name=pi --output=jsonpath='{.items[*].metadata.name}')echo$pods
The output is similar to this:
pi-5rwd7
Here, the selector is the same as the selector for the Job. The --output=jsonpath option specifies an expression
with the name from each Pod in the returned list.
The .spec.template is the only required field of the .spec.
The .spec.template is a pod template. It has exactly the same schema as a Pod, except it is nested and does not have an apiVersion or kind.
In addition to required fields for a Pod, a pod template in a Job must specify appropriate
labels (see pod selector) and an appropriate restart policy.
Only a RestartPolicy equal to Never or OnFailure is allowed.
Pod selector
The .spec.selector field is optional. In almost all cases you should not specify it.
See section specifying your own pod selector.
Parallel execution for Jobs
There are three main types of task suitable to run as a Job:
Non-parallel Jobs
normally, only one Pod is started, unless the Pod fails.
the Job is complete as soon as its Pod terminates successfully.
Parallel Jobs with a fixed completion count:
specify a non-zero positive value for .spec.completions.
the Job represents the overall task, and is complete when there are .spec.completions successful Pods.
when using .spec.completionMode="Indexed", each Pod gets a different index in the range 0 to .spec.completions-1.
Parallel Jobs with a work queue:
do not specify .spec.completions, default to .spec.parallelism.
the Pods must coordinate amongst themselves or an external service to determine what each should work on. For example, a Pod might fetch a batch of up to N items from the work queue.
each Pod is independently capable of determining whether or not all its peers are done, and thus that the entire Job is done.
when any Pod from the Job terminates with success, no new Pods are created.
once at least one Pod has terminated with success and all Pods are terminated, then the Job is completed with success.
once any Pod has exited with success, no other Pod should still be doing any work for this task or writing any output. They should all be in the process of exiting.
For a non-parallel Job, you can leave both .spec.completions and .spec.parallelism unset. When both are
unset, both are defaulted to 1.
For a fixed completion count Job, you should set .spec.completions to the number of completions needed.
You can set .spec.parallelism, or leave it unset and it will default to 1.
For a work queue Job, you must leave .spec.completions unset, and set .spec.parallelism to
a non-negative integer.
For more information about how to make use of the different types of job, see the job patterns section.
Controlling parallelism
The requested parallelism (.spec.parallelism) can be set to any non-negative value.
If it is unspecified, it defaults to 1.
If it is specified as 0, then the Job is effectively paused until it is increased.
Actual parallelism (number of pods running at any instant) may be more or less than requested
parallelism, for a variety of reasons:
For fixed completion count Jobs, the actual number of pods running in parallel will not exceed the number of
remaining completions. Higher values of .spec.parallelism are effectively ignored.
For work queue Jobs, no new Pods are started after any Pod has succeeded -- remaining Pods are allowed to complete, however.
If the Job controller failed to create Pods for any reason (lack of ResourceQuota, lack of permission, etc.),
then there may be fewer pods than requested.
The Job controller may throttle new Pod creation due to excessive previous pod failures in the same Job.
When a Pod is gracefully shut down, it takes time to stop.
Completion mode
FEATURE STATE:Kubernetes v1.24 [stable]
Jobs with fixed completion count - that is, jobs that have non null
.spec.completions - can have a completion mode that is specified in .spec.completionMode:
NonIndexed (default): the Job is considered complete when there have been
.spec.completions successfully completed Pods. In other words, each Pod
completion is homologous to each other. Note that Jobs that have null
.spec.completions are implicitly NonIndexed.
Indexed: the Pods of a Job get an associated completion index from 0 to
.spec.completions-1. The index is available through three mechanisms:
The Pod annotation batch.kubernetes.io/job-completion-index.
As part of the Pod hostname, following the pattern $(job-name)-$(index).
When you use an Indexed Job in combination with a
Service, Pods within the Job can use
the deterministic hostnames to address each other via DNS.
From the containerized task, in the environment variable JOB_COMPLETION_INDEX.
The Job is considered complete when there is one successfully completed Pod
for each index. For more information about how to use this mode, see
Indexed Job for Parallel Processing with Static Work Assignment.
Note that, although rare, more than one Pod could be started for the same
index, but only one of them will count towards the completion count.
Handling Pod and container failures
A container in a Pod may fail for a number of reasons, such as because the process in it exited with
a non-zero exit code, or the container was killed for exceeding a memory limit, etc. If this
happens, and the .spec.template.spec.restartPolicy = "OnFailure", then the Pod stays
on the node, but the container is re-run. Therefore, your program needs to handle the case when it is
restarted locally, or else specify .spec.template.spec.restartPolicy = "Never".
See pod lifecycle for more information on restartPolicy.
An entire Pod can also fail, for a number of reasons, such as when the pod is kicked off the node
(node is upgraded, rebooted, deleted, etc.), or if a container of the Pod fails and the
.spec.template.spec.restartPolicy = "Never". When a Pod fails, then the Job controller
starts a new Pod. This means that your application needs to handle the case when it is restarted in a new
pod. In particular, it needs to handle temporary files, locks, incomplete output and the like
caused by previous runs.
Note that even if you specify .spec.parallelism = 1 and .spec.completions = 1 and
.spec.template.spec.restartPolicy = "Never", the same program may
sometimes be started twice.
If you do specify .spec.parallelism and .spec.completions both greater than 1, then there may be
multiple pods running at once. Therefore, your pods must also be tolerant of concurrency.
Pod backoff failure policy
There are situations where you want to fail a Job after some amount of retries
due to a logical error in configuration etc.
To do so, set .spec.backoffLimit to specify the number of retries before
considering a Job as failed. The back-off limit is set by default to 6. Failed
Pods associated with the Job are recreated by the Job controller with an
exponential back-off delay (10s, 20s, 40s ...) capped at six minutes.
The number of retries is calculated in two ways:
The number of Pods with .status.phase = "Failed".
When using restartPolicy = "OnFailure", the number of retries in all the
containers of Pods with .status.phase equal to Pending or Running.
If either of the calculations reaches the .spec.backoffLimit, the Job is
considered failed.
When the JobTrackingWithFinalizers feature is
disabled, the number of failed Pods is only based on Pods that are still present
in the API.
Note: If your job has restartPolicy = "OnFailure", keep in mind that your Pod running the Job
will be terminated once the job backoff limit has been reached. This can make debugging the Job's executable more difficult. We suggest setting
restartPolicy = "Never" when debugging the Job or using a logging system to ensure output
from failed Jobs is not lost inadvertently.
Job termination and cleanup
When a Job completes, no more Pods are created, but the Pods are usually not deleted either.
Keeping them around
allows you to still view the logs of completed pods to check for errors, warnings, or other diagnostic output.
The job object also remains after it is completed so that you can view its status. It is up to the user to delete
old jobs after noting their status. Delete the job with kubectl (e.g. kubectl delete jobs/pi or kubectl delete -f ./job.yaml). When you delete the job using kubectl, all the pods it created are deleted too.
By default, a Job will run uninterrupted unless a Pod fails (restartPolicy=Never) or a Container exits in error (restartPolicy=OnFailure), at which point the Job defers to the
.spec.backoffLimit described above. Once .spec.backoffLimit has been reached the Job will be marked as failed and any running Pods will be terminated.
Another way to terminate a Job is by setting an active deadline.
Do this by setting the .spec.activeDeadlineSeconds field of the Job to a number of seconds.
The activeDeadlineSeconds applies to the duration of the job, no matter how many Pods are created.
Once a Job reaches activeDeadlineSeconds, all of its running Pods are terminated and the Job status will become type: Failed with reason: DeadlineExceeded.
Note that a Job's .spec.activeDeadlineSeconds takes precedence over its .spec.backoffLimit. Therefore, a Job that is retrying one or more failed Pods will not deploy additional Pods once it reaches the time limit specified by activeDeadlineSeconds, even if the backoffLimit is not yet reached.
Note that both the Job spec and the Pod template spec within the Job have an activeDeadlineSeconds field. Ensure that you set this field at the proper level.
Keep in mind that the restartPolicy applies to the Pod, and not to the Job itself: there is no automatic Job restart once the Job status is type: Failed.
That is, the Job termination mechanisms activated with .spec.activeDeadlineSeconds and .spec.backoffLimit result in a permanent Job failure that requires manual intervention to resolve.
Clean up finished jobs automatically
Finished Jobs are usually no longer needed in the system. Keeping them around in
the system will put pressure on the API server. If the Jobs are managed directly
by a higher level controller, such as
CronJobs, the Jobs can be
cleaned up by CronJobs based on the specified capacity-based cleanup policy.
TTL mechanism for finished Jobs
FEATURE STATE:Kubernetes v1.23 [stable]
Another way to clean up finished Jobs (either Complete or Failed)
automatically is to use a TTL mechanism provided by a
TTL controller for
finished resources, by specifying the .spec.ttlSecondsAfterFinished field of
the Job.
When the TTL controller cleans up the Job, it will delete the Job cascadingly,
i.e. delete its dependent objects, such as Pods, together with the Job. Note
that when the Job is deleted, its lifecycle guarantees, such as finalizers, will
be honored.
The Job pi-with-ttl will be eligible to be automatically deleted, 100
seconds after it finishes.
If the field is set to 0, the Job will be eligible to be automatically deleted
immediately after it finishes. If the field is unset, this Job won't be cleaned
up by the TTL controller after it finishes.
Note:
It is recommended to set ttlSecondsAfterFinished field because unmanaged jobs
(Jobs that you created directly, and not indirectly through other workload APIs
such as CronJob) have a default deletion
policy of orphanDependents causing Pods created by an unmanaged Job to be left around
after that Job is fully deleted.
Even though the control plane eventually
garbage collects
the Pods from a deleted Job after they either fail or complete, sometimes those
lingering pods may cause cluster performance degradation or in worst case cause the
cluster to go offline due to this degradation.
You can use LimitRanges and
ResourceQuotas to place a
cap on the amount of resources that a particular namespace can
consume.
Job patterns
The Job object can be used to support reliable parallel execution of Pods. The Job object is not
designed to support closely-communicating parallel processes, as commonly found in scientific
computing. It does support parallel processing of a set of independent but related work items.
These might be emails to be sent, frames to be rendered, files to be transcoded, ranges of keys in a
NoSQL database to scan, and so on.
In a complex system, there may be multiple different sets of work items. Here we are just
considering one set of work items that the user wants to manage together — a batch job.
There are several different patterns for parallel computation, each with strengths and weaknesses.
The tradeoffs are:
One Job object for each work item, vs. a single Job object for all work items. The latter is
better for large numbers of work items. The former creates some overhead for the user and for the
system to manage large numbers of Job objects.
Number of pods created equals number of work items, vs. each Pod can process multiple work items.
The former typically requires less modification to existing code and containers. The latter
is better for large numbers of work items, for similar reasons to the previous bullet.
Several approaches use a work queue. This requires running a queue service,
and modifications to the existing program or container to make it use the work queue.
Other approaches are easier to adapt to an existing containerised application.
The tradeoffs are summarized here, with columns 2 to 4 corresponding to the above tradeoffs.
The pattern names are also links to examples and more detailed description.
When you specify completions with .spec.completions, each Pod created by the Job controller
has an identical spec. This means that
all pods for a task will have the same command line and the same
image, the same volumes, and (almost) the same environment variables. These patterns
are different ways to arrange for pods to work on different things.
This table shows the required settings for .spec.parallelism and .spec.completions for each of the patterns.
Here, W is the number of work items.
When a Job is created, the Job controller will immediately begin creating Pods
to satisfy the Job's requirements and will continue to do so until the Job is
complete. However, you may want to temporarily suspend a Job's execution and
resume it later, or start Jobs in suspended state and have a custom controller
decide later when to start them.
To suspend a Job, you can update the .spec.suspend field of
the Job to true; later, when you want to resume it again, update it to false.
Creating a Job with .spec.suspend set to true will create it in the suspended
state.
When a Job is resumed from suspension, its .status.startTime field will be
reset to the current time. This means that the .spec.activeDeadlineSeconds
timer will be stopped and reset when a Job is suspended and resumed.
When you suspend a Job, any running Pods that don't have a status of Completed will be terminated.
with a SIGTERM signal. The Pod's graceful termination period will be honored and
your Pod must handle this signal in this period. This may involve saving
progress for later or undoing changes. Pods terminated this way will not count
towards the Job's completions count.
An example Job definition in the suspended state can be like so:
The Job's status can be used to determine if a Job is suspended or has been
suspended in the past:
kubectl get jobs/myjob -o yaml
apiVersion:batch/v1kind:Job# .metadata and .spec omittedstatus:conditions:- lastProbeTime:"2021-02-05T13:14:33Z"lastTransitionTime:"2021-02-05T13:14:33Z"status:"True"type:SuspendedstartTime:"2021-02-05T13:13:48Z"
The Job condition of type "Suspended" with status "True" means the Job is
suspended; the lastTransitionTime field can be used to determine how long the
Job has been suspended for. If the status of that condition is "False", then the
Job was previously suspended and is now running. If such a condition does not
exist in the Job's status, the Job has never been stopped.
Events are also created when the Job is suspended and resumed:
kubectl describe jobs/myjob
Name: myjob
...
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 12m job-controller Created pod: myjob-hlrpl
Normal SuccessfulDelete 11m job-controller Deleted pod: myjob-hlrpl
Normal Suspended 11m job-controller Job suspended
Normal SuccessfulCreate 3s job-controller Created pod: myjob-jvb44
Normal Resumed 3s job-controller Job resumed
The last four events, particularly the "Suspended" and "Resumed" events, are
directly a result of toggling the .spec.suspend field. In the time between
these two events, we see that no Pods were created, but Pod creation restarted
as soon as the Job was resumed.
Mutable Scheduling Directives
FEATURE STATE:Kubernetes v1.23 [beta]
Note: In order to use this behavior, you must enable the JobMutableNodeSchedulingDirectivesfeature gate
on the API server.
It is enabled by default.
In most cases a parallel job will want the pods to run with constraints,
like all in the same zone, or all either on GPU model x or y but not a mix of both.
The suspend field is the first step towards achieving those semantics. Suspend allows a
custom queue controller to decide when a job should start; However, once a job is unsuspended,
a custom queue controller has no influence on where the pods of a job will actually land.
This feature allows updating a Job's scheduling directives before it starts, which gives custom queue
controllers the ability to influence pod placement while at the same time offloading actual
pod-to-node assignment to kube-scheduler. This is allowed only for suspended Jobs that have never
been unsuspended before.
The fields in a Job's pod template that can be updated are node affinity, node selector,
tolerations, labels and annotations.
Specifying your own Pod selector
Normally, when you create a Job object, you do not specify .spec.selector.
The system defaulting logic adds this field when the Job is created.
It picks a selector value that will not overlap with any other jobs.
However, in some cases, you might need to override this automatically set selector.
To do this, you can specify the .spec.selector of the Job.
Be very careful when doing this. If you specify a label selector which is not
unique to the pods of that Job, and which matches unrelated Pods, then pods of the unrelated
job may be deleted, or this Job may count other Pods as completing it, or one or both
Jobs may refuse to create Pods or run to completion. If a non-unique selector is
chosen, then other controllers (e.g. ReplicationController) and their Pods may behave
in unpredictable ways too. Kubernetes will not stop you from making a mistake when
specifying .spec.selector.
Here is an example of a case when you might want to use this feature.
Say Job old is already running. You want existing Pods
to keep running, but you want the rest of the Pods it creates
to use a different pod template and for the Job to have a new name.
You cannot update the Job because these fields are not updatable.
Therefore, you delete Job old but leave its pods
running, using kubectl delete jobs/old --cascade=orphan.
Before deleting it, you make a note of what selector it uses:
Then you create a new Job with name new and you explicitly specify the same selector.
Since the existing Pods have label controller-uid=a8f3d00d-c6d2-11e5-9f87-42010af00002,
they are controlled by Job new as well.
You need to specify manualSelector: true in the new Job since you are not using
the selector that the system normally generates for you automatically.
The new Job itself will have a different uid from a8f3d00d-c6d2-11e5-9f87-42010af00002. Setting
manualSelector: true tells the system that you know what you are doing and to allow this
mismatch.
Pod failure policy
FEATURE STATE:Kubernetes v1.25 [alpha]
Note: You can only configure a Pod failure policy for a Job if you have the
JobPodFailurePolicyfeature gate
enabled in your cluster. Additionally, it is recommended
to enable the PodDisruptionsCondition feature gate in order to be able to detect and handle
Pod disruption conditions in the Pod failure policy (see also:
Pod disruption conditions). Both feature gates are
available in Kubernetes v1.25.
A Pod failure policy, defined with the .spec.podFailurePolicy field, enables
your cluster to handle Pod failures based on the container exit codes and the
Pod conditions.
In some situations, you may want to have a better control when handling Pod
failures than the control provided by the Pod backoff failure policy,
which is based on the Job's .spec.backoffLimit. These are some examples of use cases:
To optimize costs of running workloads by avoiding unnecessary Pod restarts,
you can terminate a Job as soon as one of its Pods fails with an exit code
indicating a software bug.
To guarantee that your Job finishes even if there are disruptions, you can
ignore Pod failures caused by disruptions (such preemption,
API-initiated eviction
or taint-based eviction) so
that they don't count towards the .spec.backoffLimit limit of retries.
You can configure a Pod failure policy, in the .spec.podFailurePolicy field,
to meet the above use cases. This policy can handle Pod failures based on the
container exit codes and the Pod conditions.
Here is a manifest for a Job that defines a podFailurePolicy:
apiVersion:batch/v1kind:Jobmetadata:name:job-pod-failure-policy-examplespec:completions:12parallelism:3template:spec:restartPolicy:Nevercontainers:- name:mainimage:docker.io/library/bash:5command:["bash"]# example command simulating a bug which triggers the FailJob actionargs:- -c- echo "Hello world!" && sleep 5 && exit 42backoffLimit:6podFailurePolicy:rules:- action:FailJobonExitCodes:containerName:main # optionaloperator: In # one of:In, NotInvalues:[42]- action: Ignore # one of:Ignore, FailJob, CountonPodConditions:- type:DisruptionTarget # indicates Pod disruption
In the example above, the first rule of the Pod failure policy specifies that
the Job should be marked failed if the main container fails with the 42 exit
code. The following are the rules for the main container specifically:
an exit code of 0 means that the container succeeded
an exit code of 42 means that the entire Job failed
any other exit code represents that the container failed, and hence the entire
Pod. The Pod will be re-created if the total number of restarts is
below backoffLimit. If the backoffLimit is reached the entire Job failed.
Note: Because the Pod template specifies a restartPolicy: Never,
the kubelet does not restart the main container in that particular Pod.
The second rule of the Pod failure policy, specifying the Ignore action for
failed Pods with condition DisruptionTarget excludes Pod disruptions from
being counted towards the .spec.backoffLimit limit of retries.
Note: If the Job failed, either by the Pod failure policy or Pod backoff
failure policy, and the Job is running multiple Pods, Kubernetes terminates all
the Pods in that Job that are still Pending or Running.
These are some requirements and semantics of the API:
if you want to use a .spec.podFailurePolicy field for a Job, you must
also define that Job's pod template with .spec.restartPolicy set to Never.
the Pod failure policy rules you specify under spec.podFailurePolicy.rules
are evaluated in order. Once a rule matches a Pod failure, the remaining rules
are ignored. When no rule matches the Pod failure, the default
handling applies.
you may want to restrict a rule to a specific container by specifing its name
inspec.podFailurePolicy.rules[*].containerName. When not specified the rule
applies to all containers. When specified, it should match one the container
or initContainer names in the Pod template.
you may specify the action taken when a Pod failure policy is matched by
spec.podFailurePolicy.rules[*].action. Possible values are:
FailJob: use to indicate that the Pod's job should be marked as Failed and
all running Pods should be terminated.
Ignore: use to indicate that the counter towards the .spec.backoffLimit
should not be incremented and a replacement Pod should be created.
Count: use to indicate that the Pod should be handled in the default way.
The counter towards the .spec.backoffLimit should be incremented.
When enabled, the control plane tracks new Jobs using the behavior described
below. Jobs created before the feature was enabled are unaffected. As a user,
the only difference you would see is that the control plane tracking of Job
completion is more accurate.
When this feature isn't enabled, the Job Controller
relies on counting the Pods that exist in the cluster to track the Job status,
that is, to keep the counters for succeeded and failed Pods.
However, Pods can be removed for a number of reasons, including:
The garbage collector that removes orphan Pods when a Node goes down.
The garbage collector that removes finished Pods (in Succeeded or Failed
phase) after a threshold.
Human intervention to delete Pods belonging to a Job.
An external controller (not provided as part of Kubernetes) that removes or
replaces Pods.
If you enable the JobTrackingWithFinalizers feature for your cluster, the
control plane keeps track of the Pods that belong to any Job and notices if any
such Pod is removed from the API server. To do that, the Job controller creates Pods with
the finalizer batch.kubernetes.io/job-tracking. The controller removes the
finalizer only after the Pod has been accounted for in the Job status, allowing
the Pod to be removed by other controllers or users.
The Job controller uses the new algorithm for new Jobs only. Jobs created
before the feature is enabled are unaffected. You can determine if the Job
controller is tracking a Job using Pod finalizers by checking if the Job has the
annotation batch.kubernetes.io/job-tracking. You should not manually add
or remove this annotation from Jobs.
Alternatives
Bare Pods
When the node that a Pod is running on reboots or fails, the pod is terminated
and will not be restarted. However, a Job will create new Pods to replace terminated ones.
For this reason, we recommend that you use a Job rather than a bare Pod, even if your application
requires only a single Pod.
Replication Controller
Jobs are complementary to Replication Controllers.
A Replication Controller manages Pods which are not expected to terminate (e.g. web servers), and a Job
manages Pods that are expected to terminate (e.g. batch tasks).
As discussed in Pod Lifecycle, Job is only appropriate
for pods with RestartPolicy equal to OnFailure or Never.
(Note: If RestartPolicy is not set, the default value is Always.)
Single Job starts controller Pod
Another pattern is for a single Job to create a Pod which then creates other Pods, acting as a sort
of custom controller for those Pods. This allows the most flexibility, but may be somewhat
complicated to get started with and offers less integration with Kubernetes.
One example of this pattern would be a Job which starts a Pod which runs a script that in turn
starts a Spark master controller (see spark example), runs a spark
driver, and then cleans up.
An advantage of this approach is that the overall process gets the completion guarantee of a Job
object, but maintains complete control over what Pods are created and how work is assigned to them.
Job is part of the Kubernetes REST API.
Read the
Job
object definition to understand the API for jobs.
Read about CronJob, which you
can use to define a series of Jobs that will run based on a schedule, similar to
the UNIX tool cron.
Practice how to configure handling of retriable and non-retriable pod failures
using podFailurePolicy, based on the step-by-step examples.
3.5.2.6 - Automatic Clean-up for Finished Jobs
FEATURE STATE:Kubernetes v1.23 [stable]
TTL-after-finished controller provides a
TTL (time to live) mechanism to limit the lifetime of resource objects that
have finished execution. TTL controller only handles
Jobs.
TTL-after-finished Controller
The TTL-after-finished controller is only supported for Jobs. A cluster operator can use this feature to clean
up finished Jobs (either Complete or Failed) automatically by specifying the
.spec.ttlSecondsAfterFinished field of a Job, as in this
example.
The TTL-after-finished controller will assume that a job is eligible to be cleaned up
TTL seconds after the job has finished, in other words, when the TTL has expired. When the
TTL-after-finished controller cleans up a job, it will delete it cascadingly, that is to say it will delete
its dependent objects together with it. Note that when the job is deleted,
its lifecycle guarantees, such as finalizers, will be honored.
The TTL seconds can be set at any time. Here are some examples for setting the
.spec.ttlSecondsAfterFinished field of a Job:
Specify this field in the job manifest, so that a Job can be cleaned up
automatically some time after it finishes.
Set this field of existing, already finished jobs, to adopt this new
feature.
Use a
mutating admission webhook
to set this field dynamically at job creation time. Cluster administrators can
use this to enforce a TTL policy for finished jobs.
Use a
mutating admission webhook
to set this field dynamically after the job has finished, and choose
different TTL values based on job status, labels, etc.
Caveat
Updating TTL Seconds
Note that the TTL period, e.g. .spec.ttlSecondsAfterFinished field of Jobs,
can be modified after the job is created or has finished. However, once the
Job becomes eligible to be deleted (when the TTL has expired), the system won't
guarantee that the Jobs will be kept, even if an update to extend the TTL
returns a successful API response.
Time Skew
Because TTL-after-finished controller uses timestamps stored in the Kubernetes jobs to
determine whether the TTL has expired or not, this feature is sensitive to time
skew in the cluster, which may cause TTL-after-finish controller to clean up job objects
at the wrong time.
Clocks aren't always correct, but the difference should be
very small. Please be aware of this risk when setting a non-zero TTL.
If your control plane runs the kube-controller-manager in Pods or bare
containers, the timezone set for the kube-controller-manager container determines the timezone
that the cron job controller uses.
Caution:
The v1 CronJob API
does not officially support setting timezone as explained above.
Setting variables such as CRON_TZ or TZ is not officially supported by the Kubernetes project.
CRON_TZ or TZ is an implementation detail of the internal library being used
for parsing and calculating the next Job creation time. Any usage of it is not
recommended in a production cluster.
When creating the manifest for a CronJob resource, make sure the name you provide
is a valid DNS subdomain name.
The name must be no longer than 52 characters. This is because the CronJob controller will automatically
append 11 characters to the job name provided and there is a constraint that the
maximum length of a Job name is no more than 63 characters.
CronJob
CronJobs are meant for performing regular scheduled actions such as backups,
report generation, and so on. Each of those tasks should be configured to recur
indefinitely (for example: once a day / week / month); you can define the point
in time within that interval when the job should start.
Example
This example CronJob manifest prints the current time and a hello message every minute:
# ┌───────────── minute (0 - 59)
# │ ┌───────────── hour (0 - 23)
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12)
# │ │ │ │ ┌───────────── day of the week (0 - 6) (Sunday to Saturday;
# │ │ │ │ │ 7 is also Sunday on some systems)
# │ │ │ │ │ OR sun, mon, tue, wed, thu, fri, sat
# │ │ │ │ │
# * * * * *
Entry
Description
Equivalent to
@yearly (or @annually)
Run once a year at midnight of 1 January
0 0 1 1 *
@monthly
Run once a month at midnight of the first day of the month
0 0 1 * *
@weekly
Run once a week at midnight on Sunday morning
0 0 * * 0
@daily (or @midnight)
Run once a day at midnight
0 0 * * *
@hourly
Run once an hour at the beginning of the hour
0 * * * *
For example, the line below states that the task must be started every Friday at midnight, as well as on the 13th of each month at midnight:
0 0 13 * 5
To generate CronJob schedule expressions, you can also use web tools like crontab.guru.
Time zones
For CronJobs with no time zone specified, the kube-controller-manager interprets schedules relative to its local time zone.
FEATURE STATE:Kubernetes v1.25 [beta]
If you enable the CronJobTimeZonefeature gate,
you can specify a time zone for a CronJob (if you don't enable that feature gate, or if you are using a version of
Kubernetes that does not have experimental time zone support, all CronJobs in your cluster have an unspecified
timezone).
When you have the feature enabled, you can set spec.timeZone to the name of a valid time zone name. For example, setting
spec.timeZone: "Etc/UTC" instructs Kubernetes to interpret the schedule relative to Coordinated Universal Time.
A time zone database from the Go standard library is included in the binaries and used as a fallback in case an external database is not available on the system.
CronJob limitations
A cron job creates a job object about once per execution time of its schedule. We say "about" because there
are certain circumstances where two jobs might be created, or no job might be created. We attempt to make these rare,
but do not completely prevent them. Therefore, jobs should be idempotent.
If startingDeadlineSeconds is set to a large value or left unset (the default)
and if concurrencyPolicy is set to Allow, the jobs will always run
at least once.
Caution: If startingDeadlineSeconds is set to a value less than 10 seconds, the CronJob may not be scheduled. This is because the CronJob controller checks things every 10 seconds.
For every CronJob, the CronJob Controller checks how many schedules it missed in the duration from its last scheduled time until now. If there are more than 100 missed schedules, then it does not start the job and logs the error
Cannot determine if job needs to be started. Too many missed start time (> 100). Set or decrease .spec.startingDeadlineSeconds or check clock skew.
It is important to note that if the startingDeadlineSeconds field is set (not nil), the controller counts how many missed jobs occurred from the value of startingDeadlineSeconds until now rather than from the last scheduled time until now. For example, if startingDeadlineSeconds is 200, the controller counts how many missed jobs occurred in the last 200 seconds.
A CronJob is counted as missed if it has failed to be created at its scheduled time. For example, If concurrencyPolicy is set to Forbid and a CronJob was attempted to be scheduled when there was a previous schedule still running, then it would count as missed.
For example, suppose a CronJob is set to schedule a new Job every one minute beginning at 08:30:00, and its
startingDeadlineSeconds field is not set. If the CronJob controller happens to
be down from 08:29:00 to 10:21:00, the job will not start as the number of missed jobs which missed their schedule is greater than 100.
To illustrate this concept further, suppose a CronJob is set to schedule a new Job every one minute beginning at 08:30:00, and its
startingDeadlineSeconds is set to 200 seconds. If the CronJob controller happens to
be down for the same period as the previous example (08:29:00 to 10:21:00,) the Job will still start at 10:22:00. This happens as the controller now checks how many missed schedules happened in the last 200 seconds (ie, 3 missed schedules), rather than from the last scheduled time until now.
The CronJob is only responsible for creating Jobs that match its schedule, and
the Job in turn is responsible for the management of the Pods it represents.
Controller version
Starting with Kubernetes v1.21 the second version of the CronJob controller
is the default implementation. To disable the default CronJob controller
and use the original CronJob controller instead, one pass the CronJobControllerV2feature gate
flag to the kube-controller-manager,
and set this flag to false. For example:
--feature-gates="CronJobControllerV2=false"
What's next
Learn about Pods and
Jobs, two concepts
that CronJobs rely upon.
Read about the format
of CronJob .spec.schedule fields.
CronJob is part of the Kubernetes REST API.
Read the
CronJob
object definition to understand the API for Kubernetes cron jobs.
3.5.2.8 - ReplicationController
Note: A Deployment that configures a ReplicaSet is now the recommended way to set up replication.
A ReplicationController ensures that a specified number of pod replicas are running at any one
time. In other words, a ReplicationController makes sure that a pod or a homogeneous set of pods is
always up and available.
How a ReplicationController Works
If there are too many pods, the ReplicationController terminates the extra pods. If there are too few, the
ReplicationController starts more pods. Unlike manually created pods, the pods maintained by a
ReplicationController are automatically replaced if they fail, are deleted, or are terminated.
For example, your pods are re-created on a node after disruptive maintenance such as a kernel upgrade.
For this reason, you should use a ReplicationController even if your application requires
only a single pod. A ReplicationController is similar to a process supervisor,
but instead of supervising individual processes on a single node, the ReplicationController supervises multiple pods
across multiple nodes.
ReplicationController is often abbreviated to "rc" in discussion, and as a shortcut in
kubectl commands.
A simple case is to create one ReplicationController object to reliably run one instance of
a Pod indefinitely. A more complex use case is to run several identical replicas of a replicated
service, such as web servers.
Running an example ReplicationController
This example ReplicationController config runs three copies of the nginx web server.
To list all the pods that belong to the ReplicationController in a machine readable form, you can use a command like this:
pods=$(kubectl get pods --selector=app=nginx --output=jsonpath={.items..metadata.name})echo$pods
The output is similar to this:
nginx-3ntk0 nginx-4ok8v nginx-qrm3m
Here, the selector is the same as the selector for the ReplicationController (seen in the
kubectl describe output), and in a different form in replication.yaml. The --output=jsonpath option
specifies an expression with the name from each pod in the returned list.
Writing a ReplicationController Spec
As with all other Kubernetes config, a ReplicationController needs apiVersion, kind, and metadata fields.
The name of a ReplicationController object must be a valid
DNS subdomain name.
For general information about working with configuration files, see object management.
The .spec.template is the only required field of the .spec.
The .spec.template is a pod template. It has exactly the same schema as a Pod, except it is nested and does not have an apiVersion or kind.
In addition to required fields for a Pod, a pod template in a ReplicationController must specify appropriate
labels and an appropriate restart policy. For labels, make sure not to overlap with other controllers. See pod selector.
For local container restarts, ReplicationControllers delegate to an agent on the node,
for example the Kubelet or Docker.
Labels on the ReplicationController
The ReplicationController can itself have labels (.metadata.labels). Typically, you
would set these the same as the .spec.template.metadata.labels; if .metadata.labels is not specified
then it defaults to .spec.template.metadata.labels. However, they are allowed to be
different, and the .metadata.labels do not affect the behavior of the ReplicationController.
Pod Selector
The .spec.selector field is a label selector. A ReplicationController
manages all the pods with labels that match the selector. It does not distinguish
between pods that it created or deleted and pods that another person or process created or
deleted. This allows the ReplicationController to be replaced without affecting the running pods.
If specified, the .spec.template.metadata.labels must be equal to the .spec.selector, or it will
be rejected by the API. If .spec.selector is unspecified, it will be defaulted to
.spec.template.metadata.labels.
Also you should not normally create any pods whose labels match this selector, either directly, with
another ReplicationController, or with another controller such as Job. If you do so, the
ReplicationController thinks that it created the other pods. Kubernetes does not stop you
from doing this.
If you do end up with multiple controllers that have overlapping selectors, you
will have to manage the deletion yourself (see below).
Multiple Replicas
You can specify how many pods should run concurrently by setting .spec.replicas to the number
of pods you would like to have running concurrently. The number running at any time may be higher
or lower, such as if the replicas were just increased or decreased, or if a pod is gracefully
shutdown, and a replacement starts early.
If you do not specify .spec.replicas, then it defaults to 1.
Working with ReplicationControllers
Deleting a ReplicationController and its Pods
To delete a ReplicationController and all its pods, use kubectl delete. Kubectl will scale the ReplicationController to zero and wait
for it to delete each pod before deleting the ReplicationController itself. If this kubectl
command is interrupted, it can be restarted.
When using the REST API or client library, you need to do the steps explicitly (scale replicas to
0, wait for pod deletions, then delete the ReplicationController).
Deleting only a ReplicationController
You can delete a ReplicationController without affecting any of its pods.
Using kubectl, specify the --cascade=orphan option to kubectl delete.
When using the REST API or client library, you can delete the ReplicationController object.
Once the original is deleted, you can create a new ReplicationController to replace it. As long
as the old and new .spec.selector are the same, then the new one will adopt the old pods.
However, it will not make any effort to make existing pods match a new, different pod template.
To update pods to a new spec in a controlled way, use a rolling update.
Isolating pods from a ReplicationController
Pods may be removed from a ReplicationController's target set by changing their labels. This technique may be used to remove pods from service for debugging and data recovery. Pods that are removed in this way will be replaced automatically (assuming that the number of replicas is not also changed).
Common usage patterns
Rescheduling
As mentioned above, whether you have 1 pod you want to keep running, or 1000, a ReplicationController will ensure that the specified number of pods exists, even in the event of node failure or pod termination (for example, due to an action by another control agent).
Scaling
The ReplicationController enables scaling the number of replicas up or down, either manually or by an auto-scaling control agent, by updating the replicas field.
Rolling updates
The ReplicationController is designed to facilitate rolling updates to a service by replacing pods one-by-one.
As explained in #1353, the recommended approach is to create a new ReplicationController with 1 replica, scale the new (+1) and old (-1) controllers one by one, and then delete the old controller after it reaches 0 replicas. This predictably updates the set of pods regardless of unexpected failures.
Ideally, the rolling update controller would take application readiness into account, and would ensure that a sufficient number of pods were productively serving at any given time.
The two ReplicationControllers would need to create pods with at least one differentiating label, such as the image tag of the primary container of the pod, since it is typically image updates that motivate rolling updates.
Multiple release tracks
In addition to running multiple releases of an application while a rolling update is in progress, it's common to run multiple releases for an extended period of time, or even continuously, using multiple release tracks. The tracks would be differentiated by labels.
For instance, a service might target all pods with tier in (frontend), environment in (prod). Now say you have 10 replicated pods that make up this tier. But you want to be able to 'canary' a new version of this component. You could set up a ReplicationController with replicas set to 9 for the bulk of the replicas, with labels tier=frontend, environment=prod, track=stable, and another ReplicationController with replicas set to 1 for the canary, with labels tier=frontend, environment=prod, track=canary. Now the service is covering both the canary and non-canary pods. But you can mess with the ReplicationControllers separately to test things out, monitor the results, etc.
Using ReplicationControllers with Services
Multiple ReplicationControllers can sit behind a single service, so that, for example, some traffic
goes to the old version, and some goes to the new version.
A ReplicationController will never terminate on its own, but it isn't expected to be as long-lived as services. Services may be composed of pods controlled by multiple ReplicationControllers, and it is expected that many ReplicationControllers may be created and destroyed over the lifetime of a service (for instance, to perform an update of pods that run the service). Both services themselves and their clients should remain oblivious to the ReplicationControllers that maintain the pods of the services.
Writing programs for Replication
Pods created by a ReplicationController are intended to be fungible and semantically identical, though their configurations may become heterogeneous over time. This is an obvious fit for replicated stateless servers, but ReplicationControllers can also be used to maintain availability of master-elected, sharded, and worker-pool applications. Such applications should use dynamic work assignment mechanisms, such as the RabbitMQ work queues, as opposed to static/one-time customization of the configuration of each pod, which is considered an anti-pattern. Any pod customization performed, such as vertical auto-sizing of resources (for example, cpu or memory), should be performed by another online controller process, not unlike the ReplicationController itself.
Responsibilities of the ReplicationController
The ReplicationController ensures that the desired number of pods matches its label selector and are operational. Currently, only terminated pods are excluded from its count. In the future, readiness and other information available from the system may be taken into account, we may add more controls over the replacement policy, and we plan to emit events that could be used by external clients to implement arbitrarily sophisticated replacement and/or scale-down policies.
The ReplicationController is forever constrained to this narrow responsibility. It itself will not perform readiness nor liveness probes. Rather than performing auto-scaling, it is intended to be controlled by an external auto-scaler (as discussed in #492), which would change its replicas field. We will not add scheduling policies (for example, spreading) to the ReplicationController. Nor should it verify that the pods controlled match the currently specified template, as that would obstruct auto-sizing and other automated processes. Similarly, completion deadlines, ordering dependencies, configuration expansion, and other features belong elsewhere. We even plan to factor out the mechanism for bulk pod creation (#170).
The ReplicationController is intended to be a composable building-block primitive. We expect higher-level APIs and/or tools to be built on top of it and other complementary primitives for user convenience in the future. The "macro" operations currently supported by kubectl (run, scale) are proof-of-concept examples of this. For instance, we could imagine something like Asgard managing ReplicationControllers, auto-scalers, services, scheduling policies, canaries, etc.
API Object
Replication controller is a top-level resource in the Kubernetes REST API. More details about the
API object can be found at:
ReplicationController API object.
Alternatives to ReplicationController
ReplicaSet
ReplicaSet is the next-generation ReplicationController that supports the new set-based label selector.
It's mainly used by Deployment as a mechanism to orchestrate pod creation, deletion and updates.
Note that we recommend using Deployments instead of directly using Replica Sets, unless you require custom update orchestration or don't require updates at all.
Deployment (Recommended)
Deployment is a higher-level API object that updates its underlying Replica Sets and their Pods. Deployments are recommended if you want the rolling update functionality, because they are declarative, server-side, and have additional features.
Bare Pods
Unlike in the case where a user directly created pods, a ReplicationController replaces pods that are deleted or terminated for any reason, such as in the case of node failure or disruptive node maintenance, such as a kernel upgrade. For this reason, we recommend that you use a ReplicationController even if your application requires only a single pod. Think of it similarly to a process supervisor, only it supervises multiple pods across multiple nodes instead of individual processes on a single node. A ReplicationController delegates local container restarts to some agent on the node (for example, Kubelet or Docker).
Job
Use a Job instead of a ReplicationController for pods that are expected to terminate on their own
(that is, batch jobs).
DaemonSet
Use a DaemonSet instead of a ReplicationController for pods that provide a
machine-level function, such as machine monitoring or machine logging. These pods have a lifetime that is tied
to a machine lifetime: the pod needs to be running on the machine before other pods start, and are
safe to terminate when the machine is otherwise ready to be rebooted/shutdown.
Learn about Deployment, the replacement
for ReplicationController.
ReplicationController is part of the Kubernetes REST API.
Read the
ReplicationController
object definition to understand the API for replication controllers.
3.6 - Services, Load Balancing, and Networking
Concepts and resources behind networking in Kubernetes.
The Kubernetes network model
Every Pod in a cluster gets its own unique cluster-wide IP address.
This means you do not need to explicitly create links between Pods and you
almost never need to deal with mapping container ports to host ports.
This creates a clean, backwards-compatible model where Pods can be treated
much like VMs or physical hosts from the perspectives of port allocation,
naming, service discovery, load balancing,
application configuration, and migration.
Kubernetes imposes the following fundamental requirements on any networking
implementation (barring any intentional network segmentation policies):
pods can communicate with all other pods on any other node
without NAT
agents on a node (e.g. system daemons, kubelet) can communicate with all
pods on that node
Note: For those platforms that support Pods running in the host network (e.g.
Linux), when pods are attached to the host network of a node they can still communicate
with all pods on all nodes without NAT.
This model is not only less complex overall, but it is principally compatible
with the desire for Kubernetes to enable low-friction porting of apps from VMs
to containers. If your job previously ran in a VM, your VM had an IP and could
talk to other VMs in your project. This is the same basic model.
Kubernetes IP addresses exist at the Pod scope - containers within a Pod
share their network namespaces - including their IP address and MAC address.
This means that containers within a Pod can all reach each other's ports on
localhost. This also means that containers within a Pod must coordinate port
usage, but this is no different from processes in a VM. This is called the
"IP-per-pod" model.
How this is implemented is a detail of the particular container runtime in use.
It is possible to request ports on the Node itself which forward to your Pod
(called host ports), but this is a very niche operation. How that forwarding is
implemented is also a detail of the container runtime. The Pod itself is
blind to the existence or non-existence of host ports.
An abstract way to expose an application running on a set of Pods as a network service.
With Kubernetes you don't need to modify your application to use an unfamiliar service discovery mechanism.
Kubernetes gives Pods their own IP addresses and a single DNS name for a set of Pods,
and can load-balance across them.
Motivation
Kubernetes Pods are created and destroyed
to match the desired state of your cluster. Pods are nonpermanent resources.
If you use a Deployment to run your app,
it can create and destroy Pods dynamically.
Each Pod gets its own IP address, however in a Deployment, the set of Pods
running in one moment in time could be different from
the set of Pods running that application a moment later.
This leads to a problem: if some set of Pods (call them "backends") provides
functionality to other Pods (call them "frontends") inside your cluster,
how do the frontends find out and keep track of which IP address to connect
to, so that the frontend can use the backend part of the workload?
Enter Services.
Service resources
In Kubernetes, a Service is an abstraction which defines a logical set of Pods
and a policy by which to access them (sometimes this pattern is called
a micro-service). The set of Pods targeted by a Service is usually determined
by a selector.
To learn about other ways to define Service endpoints,
see Services without selectors.
For example, consider a stateless image-processing backend which is running with
3 replicas. Those replicas are fungible—frontends do not care which backend
they use. While the actual Pods that compose the backend set may change, the
frontend clients should not need to be aware of that, nor should they need to keep
track of the set of backends themselves.
The Service abstraction enables this decoupling.
Cloud-native service discovery
If you're able to use Kubernetes APIs for service discovery in your application,
you can query the API server
for Endpoints, that get updated whenever the set of Pods in a Service changes.
For non-native applications, Kubernetes offers ways to place a network port or load
balancer in between your application and the backend Pods.
Defining a Service
A Service in Kubernetes is a REST object, similar to a Pod. Like all of the
REST objects, you can POST a Service definition to the API server to create
a new instance.
The name of a Service object must be a valid
RFC 1035 label name.
For example, suppose you have a set of Pods where each listens on TCP port 9376
and contains a label app.kubernetes.io/name=MyApp:
This specification creates a new Service object named "my-service", which
targets TCP port 9376 on any Pod with the app.kubernetes.io/name=MyApp label.
Kubernetes assigns this Service an IP address (sometimes called the "cluster IP"),
which is used by the Service proxies
(see Virtual IPs and service proxies below).
The controller for the Service selector continuously scans for Pods that
match its selector, and then POSTs any updates to an Endpoint object
also named "my-service".
Note: A Service can map any incoming port to a targetPort. By default and
for convenience, the targetPort is set to the same value as the port
field.
Port definitions in Pods have names, and you can reference these names in the
targetPort attribute of a Service. For example, we can bind the targetPort
of the Service to the Pod port in the following way:
This works even if there is a mixture of Pods in the Service using a single
configured name, with the same network protocol available via different
port numbers. This offers a lot of flexibility for deploying and evolving
your Services. For example, you can change the port numbers that Pods expose
in the next version of your backend software, without breaking clients.
The default protocol for Services is TCP; you can also use any other
supported protocol.
As many Services need to expose more than one port, Kubernetes supports multiple
port definitions on a Service object.
Each port definition can have the same protocol, or a different one.
Services without selectors
Services most commonly abstract access to Kubernetes Pods thanks to the selector,
but when used with a corresponding Endpoints object and without a selector, the Service can abstract other kinds of backends,
including ones that run outside the cluster. For example:
You want to have an external database cluster in production, but in your
test environment you use your own databases.
You want to point your Service to a Service in a different
Namespace or on another cluster.
You are migrating a workload to Kubernetes. While evaluating the approach,
you run only a portion of your backends in Kubernetes.
In any of these scenarios you can define a Service without a Pod selector.
For example:
Because this Service has no selector, the corresponding Endpoints object is not
created automatically. You can manually map the Service to the network address and port
where it's running, by adding an Endpoints object manually:
apiVersion:v1kind:Endpointsmetadata:# the name here should match the name of the Servicename:my-servicesubsets:- addresses:- ip:192.0.2.42ports:- port:9376
When you create an Endpoints
object for a Service, you set the name of the new object to be the same as that
of the Service.
Note:
The endpoint IPs must not be: loopback (127.0.0.0/8 for IPv4, ::1/128 for IPv6), or
link-local (169.254.0.0/16 and 224.0.0.0/24 for IPv4, fe80::/64 for IPv6).
Endpoint IP addresses cannot be the cluster IPs of other Kubernetes Services,
because kube-proxy doesn't support virtual IPs
as a destination.
Accessing a Service without a selector works the same as if it had a selector.
In the example above, traffic is routed to the single endpoint defined in
the YAML: 192.0.2.42:9376 (TCP).
Note: The Kubernetes API server does not allow proxying to endpoints that are not mapped to
pods. Actions such as kubectl proxy <service-name> where the service has no
selector will fail due to this constraint. This prevents the Kubernetes API server
from being used as a proxy to endpoints the caller may not be authorized to access.
An ExternalName Service is a special case of Service that does not have
selectors and uses DNS names instead. For more information, see the
ExternalName section later in this document.
Over Capacity Endpoints
If an Endpoints resource has more than 1000 endpoints then a Kubernetes v1.22 (or later)
cluster annotates that Endpoints with endpoints.kubernetes.io/over-capacity: truncated.
This annotation indicates that the affected Endpoints object is over capacity and that
the endpoints controller has truncated the number of endpoints to 1000.
EndpointSlices
FEATURE STATE:Kubernetes v1.21 [stable]
EndpointSlices are an API resource that can provide a more scalable alternative
to Endpoints. Although conceptually quite similar to Endpoints, EndpointSlices
allow for distributing network endpoints across multiple resources. By default,
an EndpointSlice is considered "full" once it reaches 100 endpoints, at which
point additional EndpointSlices will be created to store any additional
endpoints.
EndpointSlices provide additional attributes and functionality which is
described in detail in EndpointSlices.
Application protocol
FEATURE STATE:Kubernetes v1.20 [stable]
The appProtocol field provides a way to specify an application protocol for
each Service port. The value of this field is mirrored by the corresponding
Endpoints and EndpointSlice objects.
This field follows standard Kubernetes label syntax. Values should either be
IANA standard service names or
domain prefixed names such as mycompany.com/my-custom-protocol.
Virtual IPs and service proxies
Every node in a Kubernetes cluster runs a kube-proxy. kube-proxy is
responsible for implementing a form of virtual IP for Services of type other
than ExternalName.
Why not use round-robin DNS?
A question that pops up every now and then is why Kubernetes relies on
proxying to forward inbound traffic to backends. What about other
approaches? For example, would it be possible to configure DNS records that
have multiple A values (or AAAA for IPv6), and rely on round-robin name
resolution?
There are a few reasons for using proxying for Services:
There is a long history of DNS implementations not respecting record TTLs,
and caching the results of name lookups after they should have expired.
Some apps do DNS lookups only once and cache the results indefinitely.
Even if apps and libraries did proper re-resolution, the low or zero TTLs
on the DNS records could impose a high load on DNS that then becomes
difficult to manage.
Later in this page you can read about various kube-proxy implementations work. Overall,
you should note that, when running kube-proxy, kernel level rules may be
modified (for example, iptables rules might get created), which won't get cleaned up,
in some cases until you reboot. Thus, running kube-proxy is something that should
only be done by an administrator which understands the consequences of having a
low level, privileged network proxying service on a computer. Although the kube-proxy
executable supports a cleanup function, this function is not an official feature and
thus is only available to use as-is.
Configuration
Note that the kube-proxy starts up in different modes, which are determined by its configuration.
The kube-proxy's configuration is done via a ConfigMap, and the ConfigMap for kube-proxy
effectively deprecates the behaviour for almost all of the flags for the kube-proxy.
The ConfigMap for the kube-proxy does not support live reloading of configuration.
The ConfigMap parameters for the kube-proxy cannot all be validated and verified on startup.
For example, if your operating system doesn't allow you to run iptables commands,
the standard kernel kube-proxy implementation will not work.
Likewise, if you have an operating system which doesn't support netsh,
it will not run in Windows userspace mode.
User space proxy mode
In this (legacy) mode, kube-proxy watches the Kubernetes control plane for the addition and
removal of Service and Endpoint objects. For each Service it opens a
port (randomly chosen) on the local node. Any connections to this "proxy port"
are proxied to one of the Service's backend Pods (as reported via
Endpoints). kube-proxy takes the SessionAffinity setting of the Service into
account when deciding which backend Pod to use.
Lastly, the user-space proxy installs iptables rules which capture traffic to
the Service's clusterIP (which is virtual) and port. The rules
redirect that traffic to the proxy port which proxies the backend Pod.
By default, kube-proxy in userspace mode chooses a backend via a round-robin algorithm.
iptables proxy mode
In this mode, kube-proxy watches the Kubernetes control plane for the addition and
removal of Service and Endpoint objects. For each Service, it installs
iptables rules, which capture traffic to the Service's clusterIP and port,
and redirect that traffic to one of the Service's
backend sets. For each Endpoint object, it installs iptables rules which
select a backend Pod.
By default, kube-proxy in iptables mode chooses a backend at random.
Using iptables to handle traffic has a lower system overhead, because traffic
is handled by Linux netfilter without the need to switch between userspace and the
kernel space. This approach is also likely to be more reliable.
If kube-proxy is running in iptables mode and the first Pod that's selected
does not respond, the connection fails. This is different from userspace
mode: in that scenario, kube-proxy would detect that the connection to the first
Pod had failed and would automatically retry with a different backend Pod.
You can use Pod readiness probes
to verify that backend Pods are working OK, so that kube-proxy in iptables mode
only sees backends that test out as healthy. Doing this means you avoid
having traffic sent via kube-proxy to a Pod that's known to have failed.
IPVS proxy mode
FEATURE STATE:Kubernetes v1.11 [stable]
In ipvs mode, kube-proxy watches Kubernetes Services and Endpoints,
calls netlink interface to create IPVS rules accordingly and synchronizes
IPVS rules with Kubernetes Services and Endpoints periodically.
This control loop ensures that IPVS status matches the desired
state.
When accessing a Service, IPVS directs traffic to one of the backend Pods.
The IPVS proxy mode is based on netfilter hook function that is similar to
iptables mode, but uses a hash table as the underlying data structure and works
in the kernel space.
That means kube-proxy in IPVS mode redirects traffic with lower latency than
kube-proxy in iptables mode, with much better performance when synchronising
proxy rules. Compared to the other proxy modes, IPVS mode also supports a
higher throughput of network traffic.
IPVS provides more options for balancing traffic to backend Pods;
these are:
rr: round-robin
lc: least connection (smallest number of open connections)
dh: destination hashing
sh: source hashing
sed: shortest expected delay
nq: never queue
Note:
To run kube-proxy in IPVS mode, you must make IPVS available on
the node before starting kube-proxy.
When kube-proxy starts in IPVS proxy mode, it verifies whether IPVS
kernel modules are available. If the IPVS kernel modules are not detected, then kube-proxy
falls back to running in iptables proxy mode.
In these proxy models, the traffic bound for the Service's IP:Port is
proxied to an appropriate backend without the clients knowing anything
about Kubernetes or Services or Pods.
If you want to make sure that connections from a particular client
are passed to the same Pod each time, you can select the session affinity based
on the client's IP addresses by setting service.spec.sessionAffinity to "ClientIP"
(the default is "None").
You can also set the maximum session sticky time by setting
service.spec.sessionAffinityConfig.clientIP.timeoutSeconds appropriately.
(the default value is 10800, which works out to be 3 hours).
Note: On Windows, setting the maximum session sticky time for Services is not supported.
Multi-Port Services
For some Services, you need to expose more than one port.
Kubernetes lets you configure multiple port definitions on a Service object.
When using multiple ports for a Service, you must give all of your ports names
so that these are unambiguous.
For example:
As with Kubernetes names in general, names for ports
must only contain lowercase alphanumeric characters and -. Port names must
also start and end with an alphanumeric character.
For example, the names 123-abc and web are valid, but 123_abc and -web are not.
Choosing your own IP address
You can specify your own cluster IP address as part of a Service creation
request. To do this, set the .spec.clusterIP field. For example, if you
already have an existing DNS entry that you wish to reuse, or legacy systems
that are configured for a specific IP address and difficult to re-configure.
The IP address that you choose must be a valid IPv4 or IPv6 address from within the
service-cluster-ip-range CIDR range that is configured for the API server.
If you try to create a Service with an invalid clusterIP address value, the API
server will return a 422 HTTP status code to indicate that there's a problem.
Traffic policies
External traffic policy
You can set the spec.externalTrafficPolicy field to control how traffic from external sources is routed.
Valid values are Cluster and Local. Set the field to Cluster to route external traffic to all ready endpoints
and Local to only route to ready node-local endpoints. If the traffic policy is Local and there are no node-local
endpoints, the kube-proxy does not forward any traffic for the relevant Service.
Note:
FEATURE STATE:Kubernetes v1.22 [alpha]
If you enable the ProxyTerminatingEndpointsfeature gate
for the kube-proxy, the kube-proxy checks if the node
has local endpoints and whether or not all the local endpoints are marked as terminating.
If there are local endpoints and all of those are terminating, then the kube-proxy ignores
any external traffic policy of Local. Instead, whilst the node-local endpoints remain as all
terminating, the kube-proxy forwards traffic for that Service to healthy endpoints elsewhere,
as if the external traffic policy were set to Cluster.
This forwarding behavior for terminating endpoints exists to allow external load balancers to
gracefully drain connections that are backed by NodePort Services, even when the health check
node port starts to fail. Otherwise, traffic can be lost between the time a node is still in the node pool of a load
balancer and traffic is being dropped during the termination period of a pod.
Internal traffic policy
FEATURE STATE:Kubernetes v1.22 [beta]
You can set the spec.internalTrafficPolicy field to control how traffic from internal sources is routed.
Valid values are Cluster and Local. Set the field to Cluster to route internal traffic to all ready endpoints
and Local to only route to ready node-local endpoints. If the traffic policy is Local and there are no node-local
endpoints, traffic is dropped by kube-proxy.
Discovering services
Kubernetes supports 2 primary modes of finding a Service - environment
variables and DNS.
Environment variables
When a Pod is run on a Node, the kubelet adds a set of environment variables
for each active Service. It adds {SVCNAME}_SERVICE_HOST and {SVCNAME}_SERVICE_PORT variables,
where the Service name is upper-cased and dashes are converted to underscores.
It also supports variables (see makeLinkVariables)
that are compatible with Docker Engine's
"legacy container links" feature.
For example, the Service redis-master which exposes TCP port 6379 and has been
allocated cluster IP address 10.0.0.11, produces the following environment
variables:
When you have a Pod that needs to access a Service, and you are using
the environment variable method to publish the port and cluster IP to the client
Pods, you must create the Service before the client Pods come into existence.
Otherwise, those client Pods won't have their environment variables populated.
If you only use DNS to discover the cluster IP for a Service, you don't need to
worry about this ordering issue.
DNS
You can (and almost always should) set up a DNS service for your Kubernetes
cluster using an add-on.
A cluster-aware DNS server, such as CoreDNS, watches the Kubernetes API for new
Services and creates a set of DNS records for each one. If DNS has been enabled
throughout your cluster then all Pods should automatically be able to resolve
Services by their DNS name.
For example, if you have a Service called my-service in a Kubernetes
namespace my-ns, the control plane and the DNS Service acting together
create a DNS record for my-service.my-ns. Pods in the my-ns namespace
should be able to find the service by doing a name lookup for my-service
(my-service.my-ns would also work).
Pods in other namespaces must qualify the name as my-service.my-ns. These names
will resolve to the cluster IP assigned for the Service.
Kubernetes also supports DNS SRV (Service) records for named ports. If the
my-service.my-ns Service has a port named http with the protocol set to
TCP, you can do a DNS SRV query for _http._tcp.my-service.my-ns to discover
the port number for http, as well as the IP address.
The Kubernetes DNS server is the only way to access ExternalName Services.
You can find more information about ExternalName resolution in
DNS Pods and Services.
Headless Services
Sometimes you don't need load-balancing and a single Service IP. In
this case, you can create what are termed "headless" Services, by explicitly
specifying "None" for the cluster IP (.spec.clusterIP).
You can use a headless Service to interface with other service discovery mechanisms,
without being tied to Kubernetes' implementation.
For headless Services, a cluster IP is not allocated, kube-proxy does not handle
these Services, and there is no load balancing or proxying done by the platform
for them. How DNS is automatically configured depends on whether the Service has
selectors defined:
With selectors
For headless Services that define selectors, the endpoints controller creates
Endpoints records in the API, and modifies the DNS configuration to return
A records (IP addresses) that point directly to the Pods backing the Service.
Without selectors
For headless Services that do not define selectors, the endpoints controller does
not create Endpoints records. However, the DNS system looks for and configures
either:
A records for any Endpoints that share a name with the Service, for all
other types.
Publishing Services (ServiceTypes)
For some parts of your application (for example, frontends) you may want to expose a
Service onto an external IP address, that's outside of your cluster.
Kubernetes ServiceTypes allow you to specify what kind of Service you want.
The default is ClusterIP.
Type values and their behaviors are:
ClusterIP: Exposes the Service on a cluster-internal IP. Choosing this value
makes the Service only reachable from within the cluster. This is the
default ServiceType.
NodePort: Exposes the Service on each Node's IP at a static port
(the NodePort). A ClusterIP Service, to which the NodePort Service
routes, is automatically created. You'll be able to contact the NodePort Service,
from outside the cluster,
by requesting <NodeIP>:<NodePort>.
LoadBalancer: Exposes the Service externally using a cloud
provider's load balancer. NodePort and ClusterIP Services, to which the external
load balancer routes, are automatically created.
ExternalName: Maps the Service to the contents of the
externalName field (e.g. foo.bar.example.com), by returning a CNAME record
with its value. No proxying of any kind is set up.
Note: You need either kube-dns version 1.7 or CoreDNS version 0.0.8 or higher
to use the ExternalName type.
You can also use Ingress to expose your Service.
Ingress is not a Service type, but it acts as the entry point for your cluster.
It lets you consolidate your routing rules into a single resource as it can expose multiple
services under the same IP address.
Type NodePort
If you set the type field to NodePort, the Kubernetes control plane
allocates a port from a range specified by --service-node-port-range flag (default: 30000-32767).
Each node proxies that port (the same port number on every Node) into your Service.
Your Service reports the allocated port in its .spec.ports[*].nodePort field.
If you want to specify particular IP(s) to proxy the port, you can set the
--nodeport-addresses flag for kube-proxy or the equivalent nodePortAddresses
field of the
kube-proxy configuration file
to particular IP block(s).
This flag takes a comma-delimited list of IP blocks (e.g. 10.0.0.0/8, 192.0.2.0/25)
to specify IP address ranges that kube-proxy should consider as local to this node.
For example, if you start kube-proxy with the --nodeport-addresses=127.0.0.0/8 flag,
kube-proxy only selects the loopback interface for NodePort Services.
The default for --nodeport-addresses is an empty list.
This means that kube-proxy should consider all available network interfaces for NodePort.
(That's also compatible with earlier Kubernetes releases).
If you want a specific port number, you can specify a value in the nodePort
field. The control plane will either allocate you that port or report that
the API transaction failed.
This means that you need to take care of possible port collisions yourself.
You also have to use a valid port number, one that's inside the range configured
for NodePort use.
Using a NodePort gives you the freedom to set up your own load balancing solution,
to configure environments that are not fully supported by Kubernetes, or even
to expose one or more nodes' IPs directly.
Note that this Service is visible as <NodeIP>:spec.ports[*].nodePort
and .spec.clusterIP:spec.ports[*].port.
If the --nodeport-addresses flag for kube-proxy or the equivalent field
in the kube-proxy configuration file is set, <NodeIP> would be filtered node IP(s).
For example:
apiVersion:v1kind:Servicemetadata:name:my-servicespec:type:NodePortselector:app.kubernetes.io/name:MyAppports:# By default and for convenience, the `targetPort` is set to the same value as the `port` field.- port:80targetPort:80# Optional field# By default and for convenience, the Kubernetes control plane will allocate a port from a range (default: 30000-32767)nodePort:30007
Type LoadBalancer
On cloud providers which support external load balancers, setting the type
field to LoadBalancer provisions a load balancer for your Service.
The actual creation of the load balancer happens asynchronously, and
information about the provisioned balancer is published in the Service's
.status.loadBalancer field.
For example:
Traffic from the external load balancer is directed at the backend Pods.
The cloud provider decides how it is load balanced.
Some cloud providers allow you to specify the loadBalancerIP. In those cases, the load-balancer is created
with the user-specified loadBalancerIP. If the loadBalancerIP field is not specified,
the loadBalancer is set up with an ephemeral IP address. If you specify a loadBalancerIP
but your cloud provider does not support the feature, the loadbalancerIP field that you
set is ignored.
Note:
On Azure, if you want to use a user-specified public type loadBalancerIP, you first need
to create a static type public IP address resource. This public IP address resource should
be in the same resource group of the other automatically created resources of the cluster.
For example, MC_myResourceGroup_myAKSCluster_eastus.
By default, for LoadBalancer type of Services, when there is more than one port defined, all
ports must have the same protocol, and the protocol must be one which is supported
by the cloud provider.
The feature gate MixedProtocolLBService (enabled by default for the kube-apiserver as of v1.24) allows the use of
different protocols for LoadBalancer type of Services, when there is more than one port defined.
Note: The set of protocols that can be used for LoadBalancer type of Services is still defined by the cloud provider. If a
cloud provider does not support mixed protocols they will provide only a single protocol.
Disabling load balancer NodePort allocation
FEATURE STATE:Kubernetes v1.24 [stable]
You can optionally disable node port allocation for a Service of type=LoadBalancer, by setting
the field spec.allocateLoadBalancerNodePorts to false. This should only be used for load balancer implementations
that route traffic directly to pods as opposed to using node ports. By default, spec.allocateLoadBalancerNodePorts
is true and type LoadBalancer Services will continue to allocate node ports. If spec.allocateLoadBalancerNodePorts
is set to false on an existing Service with allocated node ports, those node ports will not be de-allocated automatically.
You must explicitly remove the nodePorts entry in every Service port to de-allocate those node ports.
Specifying class of load balancer implementation
FEATURE STATE:Kubernetes v1.24 [stable]
spec.loadBalancerClass enables you to use a load balancer implementation other than the cloud provider default.
By default, spec.loadBalancerClass is nil and a LoadBalancer type of Service uses
the cloud provider's default load balancer implementation if the cluster is configured with
a cloud provider using the --cloud-provider component flag.
If spec.loadBalancerClass is specified, it is assumed that a load balancer
implementation that matches the specified class is watching for Services.
Any default load balancer implementation (for example, the one provided by
the cloud provider) will ignore Services that have this field set.
spec.loadBalancerClass can be set on a Service of type LoadBalancer only.
Once set, it cannot be changed.
The value of spec.loadBalancerClass must be a label-style identifier,
with an optional prefix such as "internal-vip" or "example.com/internal-vip".
Unprefixed names are reserved for end-users.
Internal load balancer
In a mixed environment it is sometimes necessary to route traffic from Services inside the same
(virtual) network address block.
In a split-horizon DNS environment you would need two Services to be able to route both external
and internal traffic to your endpoints.
To set an internal load balancer, add one of the following annotations to your Service
depending on the cloud Service provider you're using.
The first specifies the ARN of the certificate to use. It can be either a
certificate from a third party issuer that was uploaded to IAM or one created
within AWS Certificate Manager.
The second annotation specifies which protocol a Pod speaks. For HTTPS and
SSL, the ELB expects the Pod to authenticate itself over the encrypted
connection, using a certificate.
HTTP and HTTPS selects layer 7 proxying: the ELB terminates
the connection with the user, parses headers, and injects the X-Forwarded-For
header with the user's IP address (Pods only see the IP address of the
ELB at the other end of its connection) when forwarding requests.
TCP and SSL selects layer 4 proxying: the ELB forwards traffic without
modifying the headers.
In a mixed-use environment where some ports are secured and others are left unencrypted,
you can use the following annotations:
In the above example, if the Service contained three ports, 80, 443, and
8443, then 443 and 8443 would use the SSL certificate, but 80 would be proxied HTTP.
From Kubernetes v1.9 onwards you can use
predefined AWS SSL policies
with HTTPS or SSL listeners for your Services.
To see which policies are available for use, you can use the aws command line tool:
You can then specify any one of those policies using the
"service.beta.kubernetes.io/aws-load-balancer-ssl-negotiation-policy"
annotation; for example:
Since version 1.3.0, the use of this annotation applies to all ports proxied by the ELB
and cannot be configured otherwise.
ELB Access Logs on AWS
There are several annotations to manage access logs for ELB Services on AWS.
The annotation service.beta.kubernetes.io/aws-load-balancer-access-log-enabled
controls whether access logs are enabled.
The annotation service.beta.kubernetes.io/aws-load-balancer-access-log-emit-interval
controls the interval in minutes for publishing the access logs. You can specify
an interval of either 5 or 60 minutes.
The annotation service.beta.kubernetes.io/aws-load-balancer-access-log-s3-bucket-name
controls the name of the Amazon S3 bucket where load balancer access logs are
stored.
The annotation service.beta.kubernetes.io/aws-load-balancer-access-log-s3-bucket-prefix
specifies the logical hierarchy you created for your Amazon S3 bucket.
metadata:name:my-serviceannotations:# Specifies whether access logs are enabled for the load balancerservice.beta.kubernetes.io/aws-load-balancer-access-log-enabled:"true"# The interval for publishing the access logs. You can specify an interval of either 5 or 60 (minutes).service.beta.kubernetes.io/aws-load-balancer-access-log-emit-interval:"60"# The name of the Amazon S3 bucket where the access logs are storedservice.beta.kubernetes.io/aws-load-balancer-access-log-s3-bucket-name:"my-bucket"# The logical hierarchy you created for your Amazon S3 bucket, for example `my-bucket-prefix/prod`service.beta.kubernetes.io/aws-load-balancer-access-log-s3-bucket-prefix:"my-bucket-prefix/prod"
Connection Draining on AWS
Connection draining for Classic ELBs can be managed with the annotation
service.beta.kubernetes.io/aws-load-balancer-connection-draining-enabled set
to the value of "true". The annotation
service.beta.kubernetes.io/aws-load-balancer-connection-draining-timeout can
also be used to set maximum time, in seconds, to keep the existing connections open before
deregistering the instances.
There are other annotations to manage Classic Elastic Load Balancers that are described below.
metadata:name:my-serviceannotations:# The time, in seconds, that the connection is allowed to be idle (no data has been sent# over the connection) before it is closed by the load balancerservice.beta.kubernetes.io/aws-load-balancer-connection-idle-timeout:"60"# Specifies whether cross-zone load balancing is enabled for the load balancerservice.beta.kubernetes.io/aws-load-balancer-cross-zone-load-balancing-enabled:"true"# A comma-separated list of key-value pairs which will be recorded as# additional tags in the ELB.service.beta.kubernetes.io/aws-load-balancer-additional-resource-tags:"environment=prod,owner=devops"# The number of successive successful health checks required for a backend to# be considered healthy for traffic. Defaults to 2, must be between 2 and 10service.beta.kubernetes.io/aws-load-balancer-healthcheck-healthy-threshold:""# The number of unsuccessful health checks required for a backend to be# considered unhealthy for traffic. Defaults to 6, must be between 2 and 10service.beta.kubernetes.io/aws-load-balancer-healthcheck-unhealthy-threshold:"3"# The approximate interval, in seconds, between health checks of an# individual instance. Defaults to 10, must be between 5 and 300service.beta.kubernetes.io/aws-load-balancer-healthcheck-interval:"20"# The amount of time, in seconds, during which no response means a failed# health check. This value must be less than the service.beta.kubernetes.io/aws-load-balancer-healthcheck-interval# value. Defaults to 5, must be between 2 and 60service.beta.kubernetes.io/aws-load-balancer-healthcheck-timeout:"5"# A list of existing security groups to be configured on the ELB created. Unlike the annotation# service.beta.kubernetes.io/aws-load-balancer-extra-security-groups, this replaces all other# security groups previously assigned to the ELB and also overrides the creation# of a uniquely generated security group for this ELB.# The first security group ID on this list is used as a source to permit incoming traffic to# target worker nodes (service traffic and health checks).# If multiple ELBs are configured with the same security group ID, only a single permit line# will be added to the worker node security groups, that means if you delete any# of those ELBs it will remove the single permit line and block access for all ELBs that shared the same security group ID.# This can cause a cross-service outage if not used properlyservice.beta.kubernetes.io/aws-load-balancer-security-groups:"sg-53fae93f"# A list of additional security groups to be added to the created ELB, this leaves the uniquely# generated security group in place, this ensures that every ELB# has a unique security group ID and a matching permit line to allow traffic to the target worker nodes# (service traffic and health checks).# Security groups defined here can be shared between services.service.beta.kubernetes.io/aws-load-balancer-extra-security-groups:"sg-53fae93f,sg-42efd82e"# A comma separated list of key-value pairs which are used# to select the target nodes for the load balancerservice.beta.kubernetes.io/aws-load-balancer-target-node-labels:"ingress-gw,gw-name=public-api"
Network Load Balancer support on AWS
FEATURE STATE:Kubernetes v1.15 [beta]
To use a Network Load Balancer on AWS, use the annotation service.beta.kubernetes.io/aws-load-balancer-type with the value set to nlb.
Note: NLB only works with certain instance classes; see the
AWS documentation
on Elastic Load Balancing for a list of supported instance types.
Unlike Classic Elastic Load Balancers, Network Load Balancers (NLBs) forward the
client's IP address through to the node. If a Service's .spec.externalTrafficPolicy
is set to Cluster, the client's IP address is not propagated to the end
Pods.
By setting .spec.externalTrafficPolicy to Local, the client IP addresses is
propagated to the end Pods, but this could result in uneven distribution of
traffic. Nodes without any Pods for a particular LoadBalancer Service will fail
the NLB Target Group's health check on the auto-assigned
.spec.healthCheckNodePort and not receive any traffic.
In order to achieve even traffic, either use a DaemonSet or specify a
pod anti-affinity
to not locate on the same node.
In order for client traffic to reach instances behind an NLB, the Node security
groups are modified with the following IP rules:
Rule
Protocol
Port(s)
IpRange(s)
IpRange Description
Health Check
TCP
NodePort(s) (.spec.healthCheckNodePort for .spec.externalTrafficPolicy = Local)
Subnet CIDR
kubernetes.io/rule/nlb/health=<loadBalancerName>
Client Traffic
TCP
NodePort(s)
.spec.loadBalancerSourceRanges (defaults to 0.0.0.0/0)
kubernetes.io/rule/nlb/client=<loadBalancerName>
MTU Discovery
ICMP
3,4
.spec.loadBalancerSourceRanges (defaults to 0.0.0.0/0)
kubernetes.io/rule/nlb/mtu=<loadBalancerName>
In order to limit which client IP's can access the Network Load Balancer,
specify loadBalancerSourceRanges.
spec:loadBalancerSourceRanges:- "143.231.0.0/16"
Note: If .spec.loadBalancerSourceRanges is not set, Kubernetes
allows traffic from 0.0.0.0/0 to the Node Security Group(s). If nodes have
public IP addresses, be aware that non-NLB traffic can also reach all instances
in those modified security groups.
Other CLB annotations on Tencent Kubernetes Engine (TKE)
There are other annotations for managing Cloud Load Balancers on TKE as shown below.
metadata:name:my-serviceannotations:# Bind Loadbalancers with specified nodesservice.kubernetes.io/qcloud-loadbalancer-backends-label:key in (value1, value2)# ID of an existing load balancerservice.kubernetes.io/tke-existed-lbid：lb-6swtxxxx# Custom parameters for the load balancer (LB), does not support modification of LB type yetservice.kubernetes.io/service.extensiveParameters:""# Custom parameters for the LB listenerservice.kubernetes.io/service.listenerParameters:""# Specifies the type of Load balancer;# valid values: classic (Classic Cloud Load Balancer) or application (Application Cloud Load Balancer)service.kubernetes.io/loadbalance-type:xxxxx# Specifies the public network bandwidth billing method;# valid values: TRAFFIC_POSTPAID_BY_HOUR(bill-by-traffic) and BANDWIDTH_POSTPAID_BY_HOUR (bill-by-bandwidth).service.kubernetes.io/qcloud-loadbalancer-internet-charge-type:xxxxxx# Specifies the bandwidth value (value range: [1,2000] Mbps).service.kubernetes.io/qcloud-loadbalancer-internet-max-bandwidth-out:"10"# When this annotation is set，the loadbalancers will only register nodes# with pod running on it, otherwise all nodes will be registered.service.kubernetes.io/local-svc-only-bind-node-with-pod:true
Type ExternalName
Services of type ExternalName map a Service to a DNS name, not to a typical selector such as
my-service or cassandra. You specify these Services with the spec.externalName parameter.
This Service definition, for example, maps
the my-service Service in the prod namespace to my.database.example.com:
Note: ExternalName accepts an IPv4 address string, but as a DNS name comprised of digits, not as an IP address.
ExternalNames that resemble IPv4 addresses are not resolved by CoreDNS or ingress-nginx because ExternalName
is intended to specify a canonical DNS name. To hardcode an IP address, consider using
headless Services.
When looking up the host my-service.prod.svc.cluster.local, the cluster DNS Service
returns a CNAME record with the value my.database.example.com. Accessing
my-service works in the same way as other Services but with the crucial
difference that redirection happens at the DNS level rather than via proxying or
forwarding. Should you later decide to move your database into your cluster, you
can start its Pods, add appropriate selectors or endpoints, and change the
Service's type.
Warning:
You may have trouble using ExternalName for some common protocols, including HTTP and HTTPS.
If you use ExternalName then the hostname used by clients inside your cluster is different from
the name that the ExternalName references.
For protocols that use hostnames this difference may lead to errors or unexpected responses.
HTTP requests will have a Host: header that the origin server does not recognize;
TLS servers will not be able to provide a certificate matching the hostname that the client connected to.
If there are external IPs that route to one or more cluster nodes, Kubernetes Services can be exposed on those
externalIPs. Traffic that ingresses into the cluster with the external IP (as destination IP), on the Service port,
will be routed to one of the Service endpoints. externalIPs are not managed by Kubernetes and are the responsibility
of the cluster administrator.
In the Service spec, externalIPs can be specified along with any of the ServiceTypes.
In the example below, "my-service" can be accessed by clients on "80.11.12.10:80" (externalIP:port)
Using the userspace proxy for VIPs works at small to medium scale, but will
not scale to very large clusters with thousands of Services. The
original design proposal for portals
has more details on this.
Using the userspace proxy obscures the source IP address of a packet accessing
a Service.
This makes some kinds of network filtering (firewalling) impossible. The iptables
proxy mode does not
obscure in-cluster source IPs, but it does still impact clients coming through
a load balancer or node-port.
The Type field is designed as nested functionality - each level adds to the
previous. This is not strictly required on all cloud providers (e.g. Google Compute Engine does
not need to allocate a NodePort to make LoadBalancer work, but AWS does)
but the current API requires it.
Virtual IP implementation
The previous information should be sufficient for many people who want to
use Services. However, there is a lot going on behind the scenes that may be
worth understanding.
Avoiding collisions
One of the primary philosophies of Kubernetes is that you should not be
exposed to situations that could cause your actions to fail through no fault
of your own. For the design of the Service resource, this means not making
you choose your own port number if that choice might collide with
someone else's choice. That is an isolation failure.
In order to allow you to choose a port number for your Services, we must
ensure that no two Services can collide. Kubernetes does that by allocating each
Service its own IP address from within the service-cluster-ip-range
CIDR range that is configured for the API server.
To ensure each Service receives a unique IP, an internal allocator atomically
updates a global allocation map in etcd
prior to creating each Service. The map object must exist in the registry for
Services to get IP address assignments, otherwise creations will
fail with a message indicating an IP address could not be allocated.
In the control plane, a background controller is responsible for creating that
map (needed to support migrating from older versions of Kubernetes that used
in-memory locking). Kubernetes also uses controllers to check for invalid
assignments (eg due to administrator intervention) and for cleaning up allocated
IP addresses that are no longer used by any Services.
IP address ranges for type: ClusterIP Services
FEATURE STATE:Kubernetes v1.25 [beta]
However, there is a problem with this ClusterIP allocation strategy, because a user
can also choose their own address for the service.
This could result in a conflict if the internal allocator selects the same IP address
for another Service.
The ServiceIPStaticSubrangefeature gate is enabled by default in v1.25
and later, using an allocation strategy that divides the ClusterIP range into two bands, based on
the size of the configured service-cluster-ip-range by using the following formula
min(max(16, cidrSize / 16), 256), described as never less than 16 or more than 256,
with a graduated step function between them. Dynamic IP allocations will be preferentially
chosen from the upper band, reducing risks of conflicts with the IPs
assigned from the lower band.
This allows users to use the lower band of the service-cluster-ip-range for their
Services with static IPs assigned with a very low risk of running into conflicts.
Service IP addresses
Unlike Pod IP addresses, which actually route to a fixed destination,
Service IPs are not actually answered by a single host. Instead, kube-proxy
uses iptables (packet processing logic in Linux) to define virtual IP addresses
which are transparently redirected as needed. When clients connect to the
VIP, their traffic is automatically transported to an appropriate endpoint.
The environment variables and DNS for Services are actually populated in
terms of the Service's virtual IP address (and port).
kube-proxy supports three proxy modes—userspace, iptables and IPVS—which
each operate slightly differently.
Userspace
As an example, consider the image processing application described above.
When the backend Service is created, the Kubernetes master assigns a virtual
IP address, for example 10.0.0.1. Assuming the Service port is 1234, the
Service is observed by all of the kube-proxy instances in the cluster.
When a proxy sees a new Service, it opens a new random port, establishes an
iptables redirect from the virtual IP address to this new port, and starts accepting
connections on it.
When a client connects to the Service's virtual IP address, the iptables
rule kicks in, and redirects the packets to the proxy's own port.
The "Service proxy" chooses a backend, and starts proxying traffic from the client to the backend.
This means that Service owners can choose any port they want without risk of
collision. Clients can connect to an IP and port, without being aware
of which Pods they are actually accessing.
iptables
Again, consider the image processing application described above.
When the backend Service is created, the Kubernetes control plane assigns a virtual
IP address, for example 10.0.0.1. Assuming the Service port is 1234, the
Service is observed by all of the kube-proxy instances in the cluster.
When a proxy sees a new Service, it installs a series of iptables rules which
redirect from the virtual IP address to per-Service rules. The per-Service
rules link to per-Endpoint rules which redirect traffic (using destination NAT)
to the backends.
When a client connects to the Service's virtual IP address the iptables rule kicks in.
A backend is chosen (either based on session affinity or randomly) and packets are
redirected to the backend. Unlike the userspace proxy, packets are never
copied to userspace, the kube-proxy does not have to be running for the virtual
IP address to work, and Nodes see traffic arriving from the unaltered client IP
address.
This same basic flow executes when traffic comes in through a node-port or
through a load-balancer, though in those cases the client IP does get altered.
IPVS
iptables operations slow down dramatically in large scale cluster e.g 10,000 Services.
IPVS is designed for load balancing and based on in-kernel hash tables.
So you can achieve performance consistency in large number of Services from IPVS-based kube-proxy.
Meanwhile, IPVS-based kube-proxy has more sophisticated load balancing algorithms
(least conns, locality, weighted, persistence).
API Object
Service is a top-level resource in the Kubernetes REST API. You can find more details
about the Service API object.
Supported protocols
TCP
You can use TCP for any kind of Service, and it's the default network protocol.
UDP
You can use UDP for most Services. For type=LoadBalancer Services, UDP support
depends on the cloud provider offering this facility.
SCTP
FEATURE STATE:Kubernetes v1.20 [stable]
When using a network plugin that supports SCTP traffic, you can use SCTP for
most Services. For type=LoadBalancer Services, SCTP support depends on the cloud
provider offering this facility. (Most do not).
Warnings
Support for multihomed SCTP associations
Warning:
The support of multihomed SCTP associations requires that the CNI plugin can support the
assignment of multiple interfaces and IP addresses to a Pod.
NAT for multihomed SCTP associations requires special logic in the corresponding kernel modules.
Windows
Note: SCTP is not supported on Windows based nodes.
Userspace kube-proxy
Warning: The kube-proxy does not support the management of SCTP associations when it is in userspace mode.
HTTP
If your cloud provider supports it, you can use a Service in LoadBalancer mode
to set up external HTTP / HTTPS reverse proxying, forwarded to the Endpoints
of the Service.
Note: You can also use Ingress in place of Service
to expose HTTP/HTTPS Services.
PROXY protocol
If your cloud provider supports it,
you can use a Service in LoadBalancer mode to configure a load balancer outside
of Kubernetes itself, that will forward connections prefixed with
PROXY protocol.
The load balancer will send an initial series of octets describing the
incoming connection, similar to this example
3.6.2 - Topology-aware traffic routing with topology keys
FEATURE STATE:Kubernetes v1.21 [deprecated]
Note: This feature, specifically the alpha topologyKeys API, is deprecated since
Kubernetes v1.21.
Topology Aware Hints,
introduced in Kubernetes v1.21, provide similar functionality.
Service Topology enables a service to route traffic based upon the Node
topology of the cluster. For example, a service can specify that traffic be
preferentially routed to endpoints that are on the same Node as the client, or
in the same availability zone.
Topology-aware traffic routing
By default, traffic sent to a ClusterIP or NodePort Service may be routed to
any backend address for the Service. Kubernetes 1.7 made it possible to
route "external" traffic to the Pods running on the same Node that received the
traffic. For ClusterIP Services, the equivalent same-node preference for
routing wasn't possible; nor could you configure your cluster to favor routing
to endpoints within the same zone.
By setting topologyKeys on a Service, you're able to define a policy for routing
traffic based upon the Node labels for the originating and destination Nodes.
The label matching between the source and destination lets you, as a cluster
operator, designate sets of Nodes that are "closer" and "farther" from one another.
You can define labels to represent whatever metric makes sense for your own
requirements.
In public clouds, for example, you might prefer to keep network traffic within the
same zone, because interzonal traffic has a cost associated with it (and intrazonal
traffic typically does not). Other common needs include being able to route traffic
to a local Pod managed by a DaemonSet, or directing traffic to Nodes connected to the
same top-of-rack switch for the lowest latency.
Using Service Topology
If your cluster has the ServiceTopologyfeature gate enabled, you can control Service traffic
routing by specifying the topologyKeys field on the Service spec. This field
is a preference-order list of Node labels which will be used to sort endpoints
when accessing this Service. Traffic will be directed to a Node whose value for
the first label matches the originating Node's value for that label. If there is
no backend for the Service on a matching Node, then the second label will be
considered, and so forth, until no labels remain.
If no match is found, the traffic will be rejected, as if there were no
backends for the Service at all. That is, endpoints are chosen based on the first
topology key with available backends. If this field is specified and all entries
have no backends that match the topology of the client, the service has no
backends for that client and connections should fail. The special value "*" may
be used to mean "any topology". This catch-all value, if used, only makes sense
as the last value in the list.
If topologyKeys is not specified or empty, no topology constraints will be applied.
Consider a cluster with Nodes that are labeled with their hostname, zone name,
and region name. Then you can set the topologyKeys values of a service to direct
traffic as follows.
Only to endpoints on the same node, failing if no endpoint exists on the node:
["kubernetes.io/hostname"].
Preferentially to endpoints on the same node, falling back to endpoints in the
same zone, followed by the same region, and failing otherwise: ["kubernetes.io/hostname", "topology.kubernetes.io/zone", "topology.kubernetes.io/region"].
This may be useful, for example, in cases where data locality is critical.
Preferentially to the same zone, but fallback on any available endpoint if
none are available within this zone:
["topology.kubernetes.io/zone", "*"].
Constraints
Service topology is not compatible with externalTrafficPolicy=Local, and
therefore a Service cannot use both of these features. It is possible to use
both features in the same cluster on different Services, only not on the same
Service.
Valid topology keys are currently limited to kubernetes.io/hostname,
topology.kubernetes.io/zone, and topology.kubernetes.io/region, but will
be generalized to other node labels in the future.
Topology keys must be valid label keys and at most 16 keys may be specified.
The catch-all value, "*", must be the last value in the topology keys, if
it is used.
Examples
The following are common examples of using the Service Topology feature.
Only Node Local Endpoints
A Service that only routes to node local endpoints. If no endpoints exist on the node, traffic is dropped:
Kubernetes creates DNS records for Services and Pods. You can contact
Services with consistent DNS names instead of IP addresses.
Introduction
Kubernetes DNS schedules a DNS Pod and Service on the cluster, and configures
the kubelets to tell individual containers to use the DNS Service's IP to
resolve DNS names.
Every Service defined in the cluster (including the DNS server itself) is
assigned a DNS name. By default, a client Pod's DNS search list includes the
Pod's own namespace and the cluster's default domain.
Namespaces of Services
A DNS query may return different results based on the namespace of the Pod making
it. DNS queries that don't specify a namespace are limited to the Pod's
namespace. Access Services in other namespaces by specifying it in the DNS query.
For example, consider a Pod in a test namespace. A data Service is in
the prod namespace.
A query for data returns no results, because it uses the Pod's test namespace.
A query for data.prod returns the intended result, because it specifies the
namespace.
DNS queries may be expanded using the Pod's /etc/resolv.conf. Kubelet
sets this file for each Pod. For example, a query for just data may be
expanded to data.test.svc.cluster.local. The values of the search option
are used to expand queries. To learn more about DNS queries, see
the resolv.conf manual page.
In summary, a Pod in the test namespace can successfully resolve either
data.prod or data.prod.svc.cluster.local.
DNS Records
What objects get DNS records?
Services
Pods
The following sections detail the supported DNS record types and layout that is
supported. Any other layout or names or queries that happen to work are
considered implementation details and are subject to change without warning.
For more up-to-date specification, see
Kubernetes DNS-Based Service Discovery.
Services
A/AAAA records
"Normal" (not headless) Services are assigned a DNS A or AAAA record,
depending on the IP family of the Service, for a name of the form
my-svc.my-namespace.svc.cluster-domain.example. This resolves to the cluster IP
of the Service.
"Headless" (without a cluster IP) Services are also assigned a DNS A or AAAA record,
depending on the IP family of the Service, for a name of the form
my-svc.my-namespace.svc.cluster-domain.example. Unlike normal
Services, this resolves to the set of IPs of the Pods selected by the Service.
Clients are expected to consume the set or else use standard round-robin
selection from the set.
SRV records
SRV Records are created for named ports that are part of normal or Headless
Services.
For each named port, the SRV record would have the form
_my-port-name._my-port-protocol.my-svc.my-namespace.svc.cluster-domain.example.
For a regular Service, this resolves to the port number and the domain name:
my-svc.my-namespace.svc.cluster-domain.example.
For a headless Service, this resolves to multiple answers, one for each Pod
that is backing the Service, and contains the port number and the domain name of the Pod
of the form auto-generated-name.my-svc.my-namespace.svc.cluster-domain.example.
Pods
A/AAAA records
In general a Pod has the following DNS resolution:
For example, if a Pod in the default namespace has the IP address 172.17.0.3,
and the domain name for your cluster is cluster.local, then the Pod has a DNS name:
172-17-0-3.default.pod.cluster.local.
Any Pods exposed by a Service have the following DNS resolution available:
Currently when a Pod is created, its hostname is the Pod's metadata.name value.
The Pod spec has an optional hostname field, which can be used to specify the
Pod's hostname. When specified, it takes precedence over the Pod's name to be
the hostname of the Pod. For example, given a Pod with hostname set to
"my-host", the Pod will have its hostname set to "my-host".
The Pod spec also has an optional subdomain field which can be used to specify
its subdomain. For example, a Pod with hostname set to "foo", and subdomain
set to "bar", in namespace "my-namespace", will have the fully qualified
domain name (FQDN) "foo.bar.my-namespace.svc.cluster-domain.example".
Example:
apiVersion:v1kind:Servicemetadata:name:default-subdomainspec:selector:name:busyboxclusterIP:Noneports:- name:foo# Actually, no port is needed.port:1234targetPort:1234---apiVersion:v1kind:Podmetadata:name:busybox1labels:name:busyboxspec:hostname:busybox-1subdomain:default-subdomaincontainers:- image:busybox:1.28command:- sleep- "3600"name:busybox---apiVersion:v1kind:Podmetadata:name:busybox2labels:name:busyboxspec:hostname:busybox-2subdomain:default-subdomaincontainers:- image:busybox:1.28command:- sleep- "3600"name:busybox
If there exists a headless Service in the same namespace as the Pod and with
the same name as the subdomain, the cluster's DNS Server also returns an A or AAAA
record for the Pod's fully qualified hostname.
For example, given a Pod with the hostname set to "busybox-1" and the subdomain set to
"default-subdomain", and a headless Service named "default-subdomain" in
the same namespace, the Pod will see its own FQDN as
"busybox-1.default-subdomain.my-namespace.svc.cluster-domain.example". DNS serves an
A or AAAA record at that name, pointing to the Pod's IP. Both Pods "busybox1" and
"busybox2" can have their distinct A or AAAA records.
The Endpoints object can specify the hostname for any endpoint addresses,
along with its IP.
Note: Because A or AAAA records are not created for Pod names, hostname is required for the Pod's A or AAAA
record to be created. A Pod with no hostname but with subdomain will only create the
A or AAAA record for the headless Service (default-subdomain.my-namespace.svc.cluster-domain.example),
pointing to the Pod's IP address. Also, Pod needs to become ready in order to have a
record unless publishNotReadyAddresses=True is set on the Service.
Pod's setHostnameAsFQDN field
FEATURE STATE:Kubernetes v1.22 [stable]
When a Pod is configured to have fully qualified domain name (FQDN), its hostname is the short hostname. For example, if you have a Pod with the fully qualified domain name busybox-1.default-subdomain.my-namespace.svc.cluster-domain.example, then by default the hostname command inside that Pod returns busybox-1 and the hostname --fqdn command returns the FQDN.
When you set setHostnameAsFQDN: true in the Pod spec, the kubelet writes the Pod's FQDN into the hostname for that Pod's namespace. In this case, both hostname and hostname --fqdn return the Pod's FQDN.
Note:
In Linux, the hostname field of the kernel (the nodename field of struct utsname) is limited to 64 characters.
If a Pod enables this feature and its FQDN is longer than 64 character, it will fail to start. The Pod will remain in Pending status (ContainerCreating as seen by kubectl) generating error events, such as Failed to construct FQDN from Pod hostname and cluster domain, FQDN long-FQDN is too long (64 characters is the max, 70 characters requested). One way of improving user experience for this scenario is to create an admission webhook controller to control FQDN size when users create top level objects, for example, Deployment.
Pod's DNS Policy
DNS policies can be set on a per-Pod basis. Currently Kubernetes supports the
following Pod-specific DNS policies. These policies are specified in the
dnsPolicy field of a Pod Spec.
"Default": The Pod inherits the name resolution configuration from the node
that the Pods run on.
See related discussion
for more details.
"ClusterFirst": Any DNS query that does not match the configured cluster
domain suffix, such as "www.kubernetes.io", is forwarded to the upstream
nameserver inherited from the node. Cluster administrators may have extra
stub-domain and upstream DNS servers configured.
See related discussion
for details on how DNS queries are handled in those cases.
"ClusterFirstWithHostNet": For Pods running with hostNetwork, you should
explicitly set its DNS policy "ClusterFirstWithHostNet".
Note: This is not supported on Windows. See below for details
"None": It allows a Pod to ignore DNS settings from the Kubernetes
environment. All DNS settings are supposed to be provided using the
dnsConfig field in the Pod Spec.
See Pod's DNS config subsection below.
Note: "Default" is not the default DNS policy. If dnsPolicy is not
explicitly specified, then "ClusterFirst" is used.
The example below shows a Pod with its DNS policy set to
"ClusterFirstWithHostNet" because it has hostNetwork set to true.
Pod's DNS Config allows users more control on the DNS settings for a Pod.
The dnsConfig field is optional and it can work with any dnsPolicy settings.
However, when a Pod's dnsPolicy is set to "None", the dnsConfig field has
to be specified.
Below are the properties a user can specify in the dnsConfig field:
nameservers: a list of IP addresses that will be used as DNS servers for the
Pod. There can be at most 3 IP addresses specified. When the Pod's dnsPolicy
is set to "None", the list must contain at least one IP address, otherwise
this property is optional.
The servers listed will be combined to the base nameservers generated from the
specified DNS policy with duplicate addresses removed.
searches: a list of DNS search domains for hostname lookup in the Pod.
This property is optional. When specified, the provided list will be merged
into the base search domain names generated from the chosen DNS policy.
Duplicate domain names are removed.
Kubernetes allows for at most 6 search domains.
options: an optional list of objects where each object may have a name
property (required) and a value property (optional). The contents in this
property will be merged to the options generated from the specified DNS policy.
Duplicate entries are removed.
The following is an example Pod with custom DNS settings:
By default, for Pod's DNS Config, Kubernetes allows at most 6 search domains and
a list of search domains of up to 256 characters.
If the feature gate ExpandedDNSConfig is enabled for the kube-apiserver and
the kubelet, it is allowed for Kubernetes to have at most 32 search domains and
a list of search domains of up to 2048 characters.
DNS resolution on Windows nodes
ClusterFirstWithHostNet is not supported for Pods that run on Windows nodes.
Windows treats all names with a . as a FQDN and skips FQDN resolution.
On Windows, there are multiple DNS resolvers that can be used. As these come with
slightly different behaviors, using the
Resolve-DNSName
powershell cmdlet for name query resolutions is recommended.
On Linux, you have a DNS suffix list, which is used after resolution of a name as fully
qualified has failed.
On Windows, you can only have 1 DNS suffix, which is the DNS suffix associated with that
Pod's namespace (example: mydns.svc.cluster.local). Windows can resolve FQDNs, Services,
or network name which can be resolved with this single suffix. For example, a Pod spawned
in the default namespace, will have the DNS suffix default.svc.cluster.local.
Inside a Windows Pod, you can resolve both kubernetes.default.svc.cluster.local
and kubernetes, but not the partially qualified names (kubernetes.default or
kubernetes.default.svc).
Now that you have a continuously running, replicated application you can expose it on a network.
Kubernetes assumes that pods can communicate with other pods, regardless of which host they land on. Kubernetes gives every pod its own cluster-private IP address, so you do not need to explicitly create links between pods or map container ports to host ports. This means that containers within a Pod can all reach each other's ports on localhost, and all pods in a cluster can see each other without NAT. The rest of this document elaborates on how you can run reliable services on such a networking model.
This guide uses a simple nginx server to demonstrate proof of concept.
Exposing pods to the cluster
We did this in a previous example, but let's do it once again and focus on the networking perspective.
Create an nginx Pod, and note that it has a container port specification:
You should be able to ssh into any node in your cluster and use a tool such as curl to make queries against both IPs. Note that the containers are not using port 80 on the node, nor are there any special NAT rules to route traffic to the pod. This means you can run multiple nginx pods on the same node all using the same containerPort, and access them from any other pod or node in your cluster using the assigned IP address for the Service. If you want to arrange for a specific port on the host Node to be forwarded to backing Pods, you can - but the networking model should mean that you do not need to do so.
So we have pods running nginx in a flat, cluster wide, address space. In theory, you could talk to these pods directly, but what happens when a node dies? The pods die with it, and the Deployment will create new ones, with different IPs. This is the problem a Service solves.
A Kubernetes Service is an abstraction which defines a logical set of Pods running somewhere in your cluster, that all provide the same functionality. When created, each Service is assigned a unique IP address (also called clusterIP). This address is tied to the lifespan of the Service, and will not change while the Service is alive. Pods can be configured to talk to the Service, and know that communication to the Service will be automatically load-balanced out to some pod that is a member of the Service.
You can create a Service for your 2 nginx replicas with kubectl expose:
kubectl expose deployment/my-nginx
service/my-nginx exposed
This is equivalent to kubectl apply -f the following yaml:
This specification will create a Service which targets TCP port 80 on any Pod
with the run: my-nginx label, and expose it on an abstracted Service port
(targetPort: is the port the container accepts traffic on, port: is the
abstracted Service port, which can be any port other pods use to access the
Service).
View Service
API object to see the list of supported fields in service definition.
Check your Service:
kubectl get svc my-nginx
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-nginx ClusterIP 10.0.162.149 <none> 80/TCP 21s
As mentioned previously, a Service is backed by a group of Pods. These Pods are
exposed through endpoints. The Service's selector will be evaluated continuously
and the results will be POSTed to an Endpoints object also named my-nginx.
When a Pod dies, it is automatically removed from the endpoints, and new Pods
matching the Service's selector will automatically get added to the endpoints.
Check the endpoints, and note that the IPs are the same as the Pods created in
the first step:
NAME ENDPOINTS AGE
my-nginx 10.244.2.5:80,10.244.3.4:80 1m
You should now be able to curl the nginx Service on <CLUSTER-IP>:<PORT> from
any node in your cluster. Note that the Service IP is completely virtual, it
never hits the wire. If you're curious about how this works you can read more
about the service proxy.
Accessing the Service
Kubernetes supports 2 primary modes of finding a Service - environment variables
and DNS. The former works out of the box while the latter requires the
CoreDNS cluster addon.
Note: If the service environment variables are not desired (because possible clashing with expected program ones,
too many variables to process, only using DNS, etc) you can disable this mode by setting the enableServiceLinks
flag to false on the pod spec.
Environment Variables
When a Pod runs on a Node, the kubelet adds a set of environment variables for
each active Service. This introduces an ordering problem. To see why, inspect
the environment of your running nginx Pods (your Pod name will be different):
kubectl exec my-nginx-3800858182-jr4a2 -- printenv | grep SERVICE
Note there's no mention of your Service. This is because you created the replicas
before the Service. Another disadvantage of doing this is that the scheduler might
put both Pods on the same machine, which will take your entire Service down if
it dies. We can do this the right way by killing the 2 Pods and waiting for the
Deployment to recreate them. This time around the Service exists before the
replicas. This will give you scheduler-level Service spreading of your Pods
(provided all your nodes have equal capacity), as well as the right environment
variables:
Kubernetes offers a DNS cluster addon Service that automatically assigns dns names to other Services. You can check if it's running on your cluster:
kubectl get services kube-dns --namespace=kube-system
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kube-dns ClusterIP 10.0.0.10 <none> 53/UDP,53/TCP 8m
The rest of this section will assume you have a Service with a long lived IP
(my-nginx), and a DNS server that has assigned a name to that IP. Here we use the CoreDNS cluster addon (application name kube-dns), so you can talk to the Service from any pod in your cluster using standard methods (e.g. gethostbyname()). If CoreDNS isn't running, you can enable it referring to the CoreDNS README or Installing CoreDNS. Let's run another curl application to test this:
kubectl run curl --image=radial/busyboxplus:curl -i --tty
Waiting for pod default/curl-131556218-9fnch to be running, status is Pending, pod ready: false
Hit enter for command prompt
Till now we have only accessed the nginx server from within the cluster. Before exposing the Service to the internet, you want to make sure the communication channel is secure. For this, you will need:
Self signed certificates for https (unless you already have an identity certificate)
An nginx server configured to use the certificates
A secret that makes the certificates accessible to pods
You can acquire all these from the nginx https example. This requires having go and make tools installed. If you don't want to install those, then follow the manual steps later. In short:
At this point you can reach the nginx server from any node.
kubectl get pods -o yaml | grep -i podip
podIP: 10.244.3.5
node $ curl -k https://10.244.3.5
...
<h1>Welcome to nginx!</h1>
Note how we supplied the -k parameter to curl in the last step, this is because we don't know anything about the pods running nginx at certificate generation time,
so we have to tell curl to ignore the CName mismatch. By creating a Service we linked the CName used in the certificate with the actual DNS name used by pods during Service lookup.
Let's test this from a pod (the same secret is being reused for simplicity, the pod only needs nginx.crt to access the Service):
For some parts of your applications you may want to expose a Service onto an
external IP address. Kubernetes supports two ways of doing this: NodePorts and
LoadBalancers. The Service created in the last section already used NodePort,
so your nginx HTTPS replica is ready to serve traffic on the internet if your
node has a public IP.
Let's now recreate the Service to use a cloud load balancer. Change the Type of my-nginx Service from NodePort to LoadBalancer:
kubectl edit svc my-nginx
kubectl get svc my-nginx
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-nginx LoadBalancer 10.0.162.149 xx.xxx.xxx.xxx 8080:30163/TCP 21s
curl https://<EXTERNAL-IP> -k
...
<title>Welcome to nginx!</title>
The IP address in the EXTERNAL-IP column is the one that is available on the public internet. The CLUSTER-IP is only available inside your
cluster/private cloud network.
Note that on AWS, type LoadBalancer creates an ELB, which uses a (long)
hostname, not an IP. It's too long to fit in the standard kubectl get svc
output, in fact, so you'll need to do kubectl describe service my-nginx to
see it. You'll see something like this:
kubectl describe service my-nginx
...
LoadBalancer Ingress: a320587ffd19711e5a37606cf4a74574-1142138393.us-east-1.elb.amazonaws.com
...
An API object that manages external access to the services in a cluster, typically HTTP.
Ingress may provide load balancing, SSL termination and name-based virtual hosting.
Terminology
For clarity, this guide defines the following terms:
Node: A worker machine in Kubernetes, part of a cluster.
Cluster: A set of Nodes that run containerized applications managed by Kubernetes. For this example, and in most common Kubernetes deployments, nodes in the cluster are not part of the public internet.
Edge router: A router that enforces the firewall policy for your cluster. This could be a gateway managed by a cloud provider or a physical piece of hardware.
Cluster network: A set of links, logical or physical, that facilitate communication within a cluster according to the Kubernetes networking model.
Service: A Kubernetes Service that identifies a set of Pods using label selectors. Unless mentioned otherwise, Services are assumed to have virtual IPs only routable within the cluster network.
What is Ingress?
Ingress exposes HTTP and HTTPS routes from outside the cluster to
services within the cluster.
Traffic routing is controlled by rules defined on the Ingress resource.
Here is a simple example where an Ingress sends all its traffic to one Service:
Figure. Ingress
An Ingress may be configured to give Services externally-reachable URLs, load balance traffic, terminate SSL / TLS, and offer name-based virtual hosting. An Ingress controller is responsible for fulfilling the Ingress, usually with a load balancer, though it may also configure your edge router or additional frontends to help handle the traffic.
An Ingress does not expose arbitrary ports or protocols. Exposing services other than HTTP and HTTPS to the internet typically
uses a service of type Service.Type=NodePort or
Service.Type=LoadBalancer.
Prerequisites
You must have an Ingress controller to satisfy an Ingress. Only creating an Ingress resource has no effect.
An Ingress needs apiVersion, kind, metadata and spec fields.
The name of an Ingress object must be a valid
DNS subdomain name.
For general information about working with config files, see deploying applications, configuring containers, managing resources.
Ingress frequently uses annotations to configure some options depending on the Ingress controller, an example of which
is the rewrite-target annotation.
Different Ingress controllers support different annotations. Review the documentation for
your choice of Ingress controller to learn which annotations are supported.
The Ingress spec
has all the information needed to configure a load balancer or proxy server. Most importantly, it
contains a list of rules matched against all incoming requests. Ingress resource only supports rules
for directing HTTP(S) traffic.
There are some ingress controllers, that work without the definition of a
default IngressClass. For example, the Ingress-NGINX controller can be
configured with a flag--watch-ingress-without-class. It is recommended though, to specify the
default IngressClass as shown below.
Ingress rules
Each HTTP rule contains the following information:
An optional host. In this example, no host is specified, so the rule applies to all inbound
HTTP traffic through the IP address specified. If a host is provided (for example,
foo.bar.com), the rules apply to that host.
A list of paths (for example, /testpath), each of which has an associated
backend defined with a service.name and a service.port.name or
service.port.number. Both the host and path must match the content of an
incoming request before the load balancer directs traffic to the referenced
Service.
A backend is a combination of Service and port names as described in the
Service doc or a custom resource backend by way of a CRD. HTTP (and HTTPS) requests to the
Ingress that matches the host and path of the rule are sent to the listed backend.
A defaultBackend is often configured in an Ingress controller to service any requests that do not
match a path in the spec.
DefaultBackend
An Ingress with no rules sends all traffic to a single default backend and .spec.defaultBackend
is the backend that should handle requests in that case.
The defaultBackend is conventionally a configuration option of the
Ingress controller and
is not specified in your Ingress resources.
If no .spec.rules are specified, .spec.defaultBackend must be specified.
If defaultBackend is not set, the handling of requests that do not match any of the rules will be up to the
ingress controller (consult the documentation for your ingress controller to find out how it handles this case).
If none of the hosts or paths match the HTTP request in the Ingress objects, the traffic is
routed to your default backend.
Resource backends
A Resource backend is an ObjectRef to another Kubernetes resource within the
same namespace as the Ingress object. A Resource is a mutually exclusive
setting with Service, and will fail validation if both are specified. A common
usage for a Resource backend is to ingress data to an object storage backend
with static assets.
Each path in an Ingress is required to have a corresponding path type. Paths
that do not include an explicit pathType will fail validation. There are three
supported path types:
ImplementationSpecific: With this path type, matching is up to the
IngressClass. Implementations can treat this as a separate pathType or treat
it identically to Prefix or Exact path types.
Exact: Matches the URL path exactly and with case sensitivity.
Prefix: Matches based on a URL path prefix split by /. Matching is case
sensitive and done on a path element by element basis. A path element refers
to the list of labels in the path split by the / separator. A request is a
match for path p if every p is an element-wise prefix of p of the
request path.
Note: If the last element of the path is a substring of the last
element in request path, it is not a match (for example: /foo/bar
matches/foo/bar/baz, but does not match /foo/barbaz).
Examples
Kind
Path(s)
Request path(s)
Matches?
Prefix
/
(all paths)
Yes
Exact
/foo
/foo
Yes
Exact
/foo
/bar
No
Exact
/foo
/foo/
No
Exact
/foo/
/foo
No
Prefix
/foo
/foo, /foo/
Yes
Prefix
/foo/
/foo, /foo/
Yes
Prefix
/aaa/bb
/aaa/bbb
No
Prefix
/aaa/bbb
/aaa/bbb
Yes
Prefix
/aaa/bbb/
/aaa/bbb
Yes, ignores trailing slash
Prefix
/aaa/bbb
/aaa/bbb/
Yes, matches trailing slash
Prefix
/aaa/bbb
/aaa/bbb/ccc
Yes, matches subpath
Prefix
/aaa/bbb
/aaa/bbbxyz
No, does not match string prefix
Prefix
/, /aaa
/aaa/ccc
Yes, matches /aaa prefix
Prefix
/, /aaa, /aaa/bbb
/aaa/bbb
Yes, matches /aaa/bbb prefix
Prefix
/, /aaa, /aaa/bbb
/ccc
Yes, matches / prefix
Prefix
/aaa
/ccc
No, uses default backend
Mixed
/foo (Prefix), /foo (Exact)
/foo
Yes, prefers Exact
Multiple matches
In some cases, multiple paths within an Ingress will match a request. In those
cases precedence will be given first to the longest matching path. If two paths
are still equally matched, precedence will be given to paths with an exact path
type over prefix path type.
Hostname wildcards
Hosts can be precise matches (for example “foo.bar.com”) or a wildcard (for
example “*.foo.com”). Precise matches require that the HTTP host header
matches the host field. Wildcard matches require the HTTP host header is
equal to the suffix of the wildcard rule.
Ingresses can be implemented by different controllers, often with different
configuration. Each Ingress should specify a class, a reference to an
IngressClass resource that contains additional configuration including the name
of the controller that should implement the class.
The default scope for IngressClass parameters is cluster-wide.
If you set the .spec.parameters field and don't set
.spec.parameters.scope, or if you set .spec.parameters.scope to
Cluster, then the IngressClass refers to a cluster-scoped resource.
The kind (in combination the apiGroup) of the parameters
refers to a cluster-scoped API (possibly a custom resource), and
the name of the parameters identifies a specific cluster scoped
resource for that API.
For example:
---apiVersion:networking.k8s.io/v1kind:IngressClassmetadata:name:external-lb-1spec:controller:example.com/ingress-controllerparameters:# The parameters for this IngressClass are specified in a# ClusterIngressParameter (API group k8s.example.net) named# "external-config-1". This definition tells Kubernetes to# look for a cluster-scoped parameter resource.scope:ClusterapiGroup:k8s.example.netkind:ClusterIngressParametername:external-config-1
FEATURE STATE:Kubernetes v1.23 [stable]
If you set the .spec.parameters field and set
.spec.parameters.scope to Namespace, then the IngressClass refers
to a namespaced-scoped resource. You must also set the namespace
field within .spec.parameters to the namespace that contains
the parameters you want to use.
The kind (in combination the apiGroup) of the parameters
refers to a namespaced API (for example: ConfigMap), and
the name of the parameters identifies a specific resource
in the namespace you specified in namespace.
Namespace-scoped parameters help the cluster operator delegate control over the
configuration (for example: load balancer settings, API gateway definition)
that is used for a workload. If you used a cluster-scoped parameter then either:
the cluster operator team needs to approve a different team's changes every
time there's a new configuration change being applied.
the cluster operator must define specific access controls, such as
RBAC roles and bindings, that let
the application team make changes to the cluster-scoped parameters resource.
The IngressClass API itself is always cluster-scoped.
Here is an example of an IngressClass that refers to parameters that are
namespaced:
---apiVersion:networking.k8s.io/v1kind:IngressClassmetadata:name:external-lb-2spec:controller:example.com/ingress-controllerparameters:# The parameters for this IngressClass are specified in an# IngressParameter (API group k8s.example.com) named "external-config",# that's in the "external-configuration" namespace.scope:NamespaceapiGroup:k8s.example.comkind:IngressParameternamespace:external-configurationname:external-config
Deprecated annotation
Before the IngressClass resource and ingressClassName field were added in
Kubernetes 1.18, Ingress classes were specified with a
kubernetes.io/ingress.class annotation on the Ingress. This annotation was
never formally defined, but was widely supported by Ingress controllers.
The newer ingressClassName field on Ingresses is a replacement for that
annotation, but is not a direct equivalent. While the annotation was generally
used to reference the name of the Ingress controller that should implement the
Ingress, the field is a reference to an IngressClass resource that contains
additional Ingress configuration, including the name of the Ingress controller.
Default IngressClass
You can mark a particular IngressClass as default for your cluster. Setting the
ingressclass.kubernetes.io/is-default-class annotation to true on an
IngressClass resource will ensure that new Ingresses without an
ingressClassName field specified will be assigned this default IngressClass.
Caution: If you have more than one IngressClass marked as the default for your cluster,
the admission controller prevents creating new Ingress objects that don't have
an ingressClassName specified. You can resolve this by ensuring that at most 1
IngressClass is marked as default in your cluster.
There are some ingress controllers, that work without the definition of a
default IngressClass. For example, the Ingress-NGINX controller can be
configured with a flag--watch-ingress-without-class. It is recommended though, to specify the
default IngressClass:
There are existing Kubernetes concepts that allow you to expose a single Service
(see alternatives). You can also do this with an Ingress by specifying a
default backend with no rules.
If you create it using kubectl apply -f you should be able to view the state
of the Ingress you added:
kubectl get ingress test-ingress
NAME CLASS HOSTS ADDRESS PORTS AGE
test-ingress external-lb * 203.0.113.123 80 59s
Where 203.0.113.123 is the IP allocated by the Ingress controller to satisfy
this Ingress.
Note: Ingress controllers and load balancers may take a minute or two to allocate an IP address.
Until that time, you often see the address listed as <pending>.
Simple fanout
A fanout configuration routes traffic from a single IP address to more than one Service,
based on the HTTP URI being requested. An Ingress allows you to keep the number of load balancers
down to a minimum. For example, a setup like:
When you create the Ingress with kubectl apply -f:
kubectl describe ingress simple-fanout-example
Name: simple-fanout-example
Namespace: default
Address: 178.91.123.132
Default backend: default-http-backend:80 (10.8.2.3:8080)
Rules:
Host Path Backends
---- ---- --------
foo.bar.com
/foo service1:4200 (10.8.0.90:4200)
/bar service2:8080 (10.8.0.91:8080)
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ADD 22s loadbalancer-controller default/test
The Ingress controller provisions an implementation-specific load balancer
that satisfies the Ingress, as long as the Services (service1, service2) exist.
When it has done so, you can see the address of the load balancer at the
Address field.
Note: Depending on the Ingress controller
you are using, you may need to create a default-http-backend
Service.
Name based virtual hosting
Name-based virtual hosts support routing HTTP traffic to multiple host names at the same IP address.
Figure. Ingress Name Based Virtual hosting
The following Ingress tells the backing load balancer to route requests based on
the Host header.
If you create an Ingress resource without any hosts defined in the rules, then any
web traffic to the IP address of your Ingress controller can be matched without a name based
virtual host being required.
For example, the following Ingress routes traffic
requested for first.bar.com to service1, second.bar.com to service2, and any traffic whose request host header doesn't match first.bar.com and second.bar.com to service3.
You can secure an Ingress by specifying a Secret
that contains a TLS private key and certificate. The Ingress resource only
supports a single TLS port, 443, and assumes TLS termination at the ingress point
(traffic to the Service and its Pods is in plaintext).
If the TLS configuration section in an Ingress specifies different hosts, they are
multiplexed on the same port according to the hostname specified through the
SNI TLS extension (provided the Ingress controller supports SNI). The TLS secret
must contain keys named tls.crt and tls.key that contain the certificate
and private key to use for TLS. For example:
Referencing this secret in an Ingress tells the Ingress controller to
secure the channel from the client to the load balancer using TLS. You need to make
sure the TLS secret you created came from a certificate that contains a Common
Name (CN), also known as a Fully Qualified Domain Name (FQDN) for https-example.foo.com.
Note: Keep in mind that TLS will not work on the default rule because the
certificates would have to be issued for all the possible sub-domains. Therefore,
hosts in the tls section need to explicitly match the host in the rules
section.
Note: There is a gap between TLS features supported by various Ingress
controllers. Please refer to documentation on
nginx,
GCE, or any other
platform specific Ingress controller to understand how TLS works in your environment.
Load balancing
An Ingress controller is bootstrapped with some load balancing policy settings
that it applies to all Ingress, such as the load balancing algorithm, backend
weight scheme, and others. More advanced load balancing concepts
(e.g. persistent sessions, dynamic weights) are not yet exposed through the
Ingress. You can instead get these features through the load balancer used for
a Service.
It's also worth noting that even though health checks are not exposed directly
through the Ingress, there exist parallel concepts in Kubernetes such as
readiness probes
that allow you to achieve the same end result. Please review the controller
specific documentation to see how they handle health checks (for example:
nginx, or
GCE).
Updating an Ingress
To update an existing Ingress to add a new Host, you can update it by editing the resource:
kubectl describe ingress test
Name: test
Namespace: default
Address: 178.91.123.132
Default backend: default-http-backend:80 (10.8.2.3:8080)
Rules:
Host Path Backends
---- ---- --------
foo.bar.com
/foo service1:80 (10.8.0.90:80)
Annotations:
nginx.ingress.kubernetes.io/rewrite-target: /
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ADD 35s loadbalancer-controller default/test
kubectl edit ingress test
This pops up an editor with the existing configuration in YAML format.
Modify it to include the new Host:
After you save your changes, kubectl updates the resource in the API server, which tells the
Ingress controller to reconfigure the load balancer.
Verify this:
kubectl describe ingress test
Name: test
Namespace: default
Address: 178.91.123.132
Default backend: default-http-backend:80 (10.8.2.3:8080)
Rules:
Host Path Backends
---- ---- --------
foo.bar.com
/foo service1:80 (10.8.0.90:80)
bar.baz.com
/foo service2:80 (10.8.0.91:80)
Annotations:
nginx.ingress.kubernetes.io/rewrite-target: /
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ADD 45s loadbalancer-controller default/test
You can achieve the same outcome by invoking kubectl replace -f on a modified Ingress YAML file.
Failing across availability zones
Techniques for spreading traffic across failure domains differ between cloud providers.
Please check the documentation of the relevant Ingress controller for details.
Alternatives
You can expose a Service in multiple ways that don't directly involve the Ingress resource:
In order for the Ingress resource to work, the cluster must have an ingress controller running.
Unlike other types of controllers which run as part of the kube-controller-manager binary, Ingress controllers
are not started automatically with a cluster. Use this page to choose the ingress controller implementation
that best fits your cluster.
Kubernetes as a project supports and maintains AWS, GCE, and
nginx ingress controllers.
Additional controllers
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Skipper HTTP router and reverse proxy for service composition, including use cases like Kubernetes Ingress, designed as a library to build your custom proxy.
Tyk Operator extends Ingress with Custom Resources to bring API Management capabilities to Ingress. Tyk Operator works with the Open Source Tyk Gateway & Tyk Cloud control plane.
You may deploy any number of ingress controllers using ingress class
within a cluster. Note the .metadata.name of your ingress class resource. When you create an ingress you would need that name to specify the ingressClassName field on your Ingress object (refer to IngressSpec v1 reference. ingressClassName is a replacement of the older annotation method.
If you do not specify an IngressClass for an Ingress, and your cluster has exactly one IngressClass marked as default, then Kubernetes applies the cluster's default IngressClass to the Ingress.
You mark an IngressClass as default by setting the ingressclass.kubernetes.io/is-default-class annotation on that IngressClass, with the string value "true".
Ideally, all ingress controllers should fulfill this specification, but the various ingress
controllers operate slightly differently.
Note: Make sure you review your ingress controller's documentation to understand the caveats of choosing it.
EndpointSlices provide a simple way to track network endpoints within a
Kubernetes cluster. They offer a more scalable and extensible alternative to
Endpoints.
Motivation
The Endpoints API has provided a simple and straightforward way of
tracking network endpoints in Kubernetes. Unfortunately as Kubernetes clusters
and Services have grown to handle and
send more traffic to more backend Pods, limitations of that original API became
more visible.
Most notably, those included challenges with scaling to larger numbers of
network endpoints.
Since all network endpoints for a Service were stored in a single Endpoints
resource, those resources could get quite large. That affected the performance
of Kubernetes components (notably the master control plane) and resulted in
significant amounts of network traffic and processing when Endpoints changed.
EndpointSlices help you mitigate those issues as well as provide an extensible
platform for additional features such as topological routing.
EndpointSlice resources
In Kubernetes, an EndpointSlice contains references to a set of network
endpoints. The control plane automatically creates EndpointSlices
for any Kubernetes Service that has a selector specified. These EndpointSlices include
references to all the Pods that match the Service selector. EndpointSlices group
network endpoints together by unique combinations of protocol, port number, and
Service name.
The name of a EndpointSlice object must be a valid
DNS subdomain name.
As an example, here's a sample EndpointSlice resource for the example
Kubernetes Service.
By default, the control plane creates and manages EndpointSlices to have no
more than 100 endpoints each. You can configure this with the
--max-endpoints-per-slicekube-controller-manager
flag, up to a maximum of 1000.
EndpointSlices can act as the source of truth for
kube-proxy when it comes to
how to route internal traffic. When enabled, they should provide a performance
improvement for services with large numbers of endpoints.
Address types
EndpointSlices support three address types:
IPv4
IPv6
FQDN (Fully Qualified Domain Name)
Conditions
The EndpointSlice API stores conditions about endpoints that may be useful for consumers.
The three conditions are ready, serving, and terminating.
Ready
ready is a condition that maps to a Pod's Ready condition. A running Pod with the Ready
condition set to True should have this EndpointSlice condition also set to true. For
compatibility reasons, ready is NEVER true when a Pod is terminating. Consumers should refer
to the serving condition to inspect the readiness of terminating Pods. The only exception to
this rule is for Services with spec.publishNotReadyAddresses set to true. Endpoints for these
Services will always have the ready condition set to true.
Serving
FEATURE STATE:Kubernetes v1.22 [beta]
serving is identical to the ready condition, except it does not account for terminating states.
Consumers of the EndpointSlice API should check this condition if they care about pod readiness while
the pod is also terminating.
Note: Although serving is almost identical to ready, it was added to prevent break the existing meaning
of ready. It may be unexpected for existing clients if ready could be true for terminating
endpoints, since historically terminating endpoints were never included in the Endpoints or
EndpointSlice API to begin with. For this reason, ready is alwaysfalse for terminating
endpoints, and a new condition serving was added in v1.20 so that clients can track readiness
for terminating pods independent of the existing semantics for ready.
Terminating
FEATURE STATE:Kubernetes v1.22 [beta]
Terminating is a condition that indicates whether an endpoint is terminating.
For pods, this is any pod that has a deletion timestamp set.
Topology information
Each endpoint within an EndpointSlice can contain relevant topology information.
The topology information includes the location of the endpoint and information
about the corresponding Node and zone. These are available in the following
per endpoint fields on EndpointSlices:
nodeName - The name of the Node this endpoint is on.
zone - The zone this endpoint is in.
Note:
In the v1 API, the per endpoint topology was effectively removed in favor of
the dedicated fields nodeName and zone.
Setting arbitrary topology fields on the endpoint field of an EndpointSlice
resource has been deprecated and is not supported in the v1 API.
Instead, the v1 API supports setting individual nodeName and zone fields.
These fields are automatically translated between API versions. For example, the
value of the "topology.kubernetes.io/zone" key in the topology field in
the v1beta1 API is accessible as the zone field in the v1 API.
Management
Most often, the control plane (specifically, the endpoint slice
controller) creates and
manages EndpointSlice objects. There are a variety of other use cases for
EndpointSlices, such as service mesh implementations, that could result in other
entities or controllers managing additional sets of EndpointSlices.
To ensure that multiple entities can manage EndpointSlices without interfering
with each other, Kubernetes defines the
labelendpointslice.kubernetes.io/managed-by, which indicates the entity managing
an EndpointSlice.
The endpoint slice controller sets endpointslice-controller.k8s.io as the value
for this label on all EndpointSlices it manages. Other entities managing
EndpointSlices should also set a unique value for this label.
Ownership
In most use cases, EndpointSlices are owned by the Service that the endpoint
slice object tracks endpoints for. This ownership is indicated by an owner
reference on each EndpointSlice as well as a kubernetes.io/service-name
label that enables simple lookups of all EndpointSlices belonging to a Service.
EndpointSlice mirroring
In some cases, applications create custom Endpoints resources. To ensure that
these applications do not need to concurrently write to both Endpoints and
EndpointSlice resources, the cluster's control plane mirrors most Endpoints
resources to corresponding EndpointSlices.
The control plane mirrors Endpoints resources unless:
the Endpoints resource has a endpointslice.kubernetes.io/skip-mirror label
set to true.
the Endpoints resource has a control-plane.alpha.kubernetes.io/leader
annotation.
the corresponding Service resource does not exist.
the corresponding Service resource has a non-nil selector.
Individual Endpoints resources may translate into multiple EndpointSlices. This
will occur if an Endpoints resource has multiple subsets or includes endpoints
with multiple IP families (IPv4 and IPv6). A maximum of 1000 addresses per
subset will be mirrored to EndpointSlices.
Distribution of EndpointSlices
Each EndpointSlice has a set of ports that applies to all endpoints within the
resource. When named ports are used for a Service, Pods may end up with
different target port numbers for the same named port, requiring different
EndpointSlices. This is similar to the logic behind how subsets are grouped
with Endpoints.
The control plane tries to fill EndpointSlices as full as possible, but does not
actively rebalance them. The logic is fairly straightforward:
Iterate through existing EndpointSlices, remove endpoints that are no longer
desired and update matching endpoints that have changed.
Iterate through EndpointSlices that have been modified in the first step and
fill them up with any new endpoints needed.
If there's still new endpoints left to add, try to fit them into a previously
unchanged slice and/or create new ones.
Importantly, the third step prioritizes limiting EndpointSlice updates over a
perfectly full distribution of EndpointSlices. As an example, if there are 10
new endpoints to add and 2 EndpointSlices with room for 5 more endpoints each,
this approach will create a new EndpointSlice instead of filling up the 2
existing EndpointSlices. In other words, a single EndpointSlice creation is
preferrable to multiple EndpointSlice updates.
With kube-proxy running on each Node and watching EndpointSlices, every change
to an EndpointSlice becomes relatively expensive since it will be transmitted to
every Node in the cluster. This approach is intended to limit the number of
changes that need to be sent to every Node, even if it may result with multiple
EndpointSlices that are not full.
In practice, this less than ideal distribution should be rare. Most changes
processed by the EndpointSlice controller will be small enough to fit in an
existing EndpointSlice, and if not, a new EndpointSlice is likely going to be
necessary soon anyway. Rolling updates of Deployments also provide a natural
repacking of EndpointSlices with all Pods and their corresponding endpoints
getting replaced.
Duplicate endpoints
Due to the nature of EndpointSlice changes, endpoints may be represented in more
than one EndpointSlice at the same time. This naturally occurs as changes to
different EndpointSlice objects can arrive at the Kubernetes client watch/cache
at different times. Implementations using EndpointSlice must be able to have the
endpoint appear in more than one slice. A reference implementation of how to
perform endpoint deduplication can be found in the EndpointSliceCache
implementation in kube-proxy.
Service Internal Traffic Policy enables internal traffic restrictions to only route
internal traffic to endpoints within the node the traffic originated from. The
"internal" traffic here refers to traffic originated from Pods in the current
cluster. This can help to reduce costs and improve performance.
Using Service Internal Traffic Policy
The ServiceInternalTrafficPolicyfeature gate
is a Beta feature and enabled by default.
When the feature is enabled, you can enable the internal-only traffic policy for a
Services, by setting its
.spec.internalTrafficPolicy to Local.
This tells kube-proxy to only use node local endpoints for cluster internal traffic.
Note: For pods on nodes with no endpoints for a given Service, the Service
behaves as if it has zero endpoints (for Pods on this node) even if the service
does have endpoints on other nodes.
The following example shows what a Service looks like when you set
.spec.internalTrafficPolicy to Local:
The kube-proxy filters the endpoints it routes to based on the
spec.internalTrafficPolicy setting. When it's set to Local, only node local
endpoints are considered. When it's Cluster or missing, all endpoints are
considered.
When the feature gateServiceInternalTrafficPolicy is enabled, spec.internalTrafficPolicy defaults to "Cluster".
Topology Aware Hints enable topology aware routing by including suggestions
for how clients should consume endpoints. This approach adds metadata to enable
consumers of EndpointSlice and / or Endpoints objects, so that traffic to
those network endpoints can be routed closer to where it originated.
For example, you can route traffic within a locality to reduce
costs, or to improve network performance.
Note: The "topology-aware hints" feature is at Beta stage and it is NOT enabled
by default. To try out this feature, you have to enable the TopologyAwareHintsfeature gate.
Motivation
Kubernetes clusters are increasingly deployed in multi-zone environments.
Topology Aware Hints provides a mechanism to help keep traffic within the zone
it originated from. This concept is commonly referred to as "Topology Aware
Routing". When calculating the endpoints for a Service,
the EndpointSlice controller considers the topology (region and zone) of each endpoint
and populates the hints field to allocate it to a zone.
Cluster components such as the kube-proxy
can then consume those hints, and use them to influence how the traffic is routed
(favoring topologically closer endpoints).
Using Topology Aware Hints
You can activate Topology Aware Hints for a Service by setting the
service.kubernetes.io/topology-aware-hints annotation to auto. This tells
the EndpointSlice controller to set topology hints if it is deemed safe.
Importantly, this does not guarantee that hints will always be set.
How it works
The functionality enabling this feature is split into two components: The
EndpointSlice controller and the kube-proxy. This section provides a high level overview
of how each component implements this feature.
EndpointSlice controller
The EndpointSlice controller is responsible for setting hints on EndpointSlices
when this feature is enabled. The controller allocates a proportional amount of
endpoints to each zone. This proportion is based on the
allocatable
CPU cores for nodes running in that zone. For example, if one zone had 2 CPU
cores and another zone only had 1 CPU core, the controller would allocated twice
as many endpoints to the zone with 2 CPU cores.
The following example shows what an EndpointSlice looks like when hints have
been populated:
The kube-proxy component filters the endpoints it routes to based on the hints set by
the EndpointSlice controller. In most cases, this means that the kube-proxy is able
to route traffic to endpoints in the same zone. Sometimes the controller allocates endpoints
from a different zone to ensure more even distribution of endpoints between zones.
This would result in some traffic being routed to other zones.
Safeguards
The Kubernetes control plane and the kube-proxy on each node apply some
safeguard rules before using Topology Aware Hints. If these don't check out,
the kube-proxy selects endpoints from anywhere in your cluster, regardless of the
zone.
Insufficient number of endpoints: If there are less endpoints than zones
in a cluster, the controller will not assign any hints.
Impossible to achieve balanced allocation: In some cases, it will be
impossible to achieve a balanced allocation of endpoints among zones. For
example, if zone-a is twice as large as zone-b, but there are only 2
endpoints, an endpoint allocated to zone-a may receive twice as much traffic
as zone-b. The controller does not assign hints if it can't get this "expected
overload" value below an acceptable threshold for each zone. Importantly this
is not based on real-time feedback. It is still possible for individual
endpoints to become overloaded.
One or more Nodes has insufficient information: If any node does not have
a topology.kubernetes.io/zone label or is not reporting a value for
allocatable CPU, the control plane does not set any topology-aware endpoint
hints and so kube-proxy does not filter endpoints by zone.
One or more endpoints does not have a zone hint: When this happens,
the kube-proxy assumes that a transition from or to Topology Aware Hints is
underway. Filtering endpoints for a Service in this state would be dangerous
so the kube-proxy falls back to using all endpoints.
A zone is not represented in hints: If the kube-proxy is unable to find
at least one endpoint with a hint targeting the zone it is running in, it falls
to using endpoints from all zones. This is most likely to happen as you add
a new zone into your existing cluster.
Constraints
Topology Aware Hints are not used when either externalTrafficPolicy or
internalTrafficPolicy is set to Local on a Service. It is possible to use
both features in the same cluster on different Services, just not on the same
Service.
This approach will not work well for Services that have a large proportion of
traffic originating from a subset of zones. Instead this assumes that incoming
traffic will be roughly proportional to the capacity of the Nodes in each
zone.
The EndpointSlice controller ignores unready nodes as it calculates the
proportions of each zone. This could have unintended consequences if a large
portion of nodes are unready.
The EndpointSlice controller does not take into account tolerations when deploying calculating the
proportions of each zone. If the Pods backing a Service are limited to a
subset of Nodes in the cluster, this will not be taken into account.
This may not work well with autoscaling. For example, if a lot of traffic is
originating from a single zone, only the endpoints allocated to that zone will
be handling that traffic. That could result in Horizontal Pod Autoscaler
either not picking up on this event, or newly added pods starting in a
different zone.
If you want to control traffic flow at the IP address or port level (OSI layer 3 or 4), then you might consider using Kubernetes NetworkPolicies for particular applications in your cluster. NetworkPolicies are an application-centric construct which allow you to specify how a pod is allowed to communicate with various network "entities" (we use the word "entity" here to avoid overloading the more common terms such as "endpoints" and "services", which have specific Kubernetes connotations) over the network. NetworkPolicies apply to a connection with a pod on one or both ends, and are not relevant to other connections.
The entities that a Pod can communicate with are identified through a combination of the following 3 identifiers:
Other pods that are allowed (exception: a pod cannot block access to itself)
Namespaces that are allowed
IP blocks (exception: traffic to and from the node where a Pod is running is always allowed, regardless of the IP address of the Pod or the node)
When defining a pod- or namespace- based NetworkPolicy, you use a selector to specify what traffic is allowed to and from the Pod(s) that match the selector.
Meanwhile, when IP based NetworkPolicies are created, we define policies based on IP blocks (CIDR ranges).
Prerequisites
Network policies are implemented by the network plugin. To use network policies, you must be using a networking solution which supports NetworkPolicy. Creating a NetworkPolicy resource without a controller that implements it will have no effect.
The Two Sorts of Pod Isolation
There are two sorts of isolation for a pod: isolation for egress, and isolation for ingress. They concern what connections may be established. "Isolation" here is not absolute, rather it means "some restrictions apply". The alternative, "non-isolated for $direction", means that no restrictions apply in the stated direction. The two sorts of isolation (or not) are declared independently, and are both relevant for a connection from one pod to another.
By default, a pod is non-isolated for egress; all outbound connections are allowed. A pod is isolated for egress if there is any NetworkPolicy that both selects the pod and has "Egress" in its policyTypes; we say that such a policy applies to the pod for egress. When a pod is isolated for egress, the only allowed connections from the pod are those allowed by the egress list of some NetworkPolicy that applies to the pod for egress. The effects of those egress lists combine additively.
By default, a pod is non-isolated for ingress; all inbound connections are allowed. A pod is isolated for ingress if there is any NetworkPolicy that both selects the pod and has "Ingress" in its policyTypes; we say that such a policy applies to the pod for ingress. When a pod is isolated for ingress, the only allowed connections into the pod are those from the pod's node and those allowed by the ingress list of some NetworkPolicy that applies to the pod for ingress. The effects of those ingress lists combine additively.
Network policies do not conflict; they are additive. If any policy or policies apply to a given pod for a given direction, the connections allowed in that direction from that pod is the union of what the applicable policies allow. Thus, order of evaluation does not affect the policy result.
For a connection from a source pod to a destination pod to be allowed, both the egress policy on the source pod and the ingress policy on the destination pod need to allow the connection. If either side does not allow the connection, it will not happen.
The NetworkPolicy resource
See the NetworkPolicy reference for a full definition of the resource.
Note: POSTing this to the API server for your cluster will have no effect unless your chosen networking solution supports network policy.
Mandatory Fields: As with all other Kubernetes config, a NetworkPolicy
needs apiVersion, kind, and metadata fields. For general information
about working with config files, see
Configure a Pod to Use a ConfigMap,
and Object Management.
spec: NetworkPolicy spec has all the information needed to define a particular network policy in the given namespace.
podSelector: Each NetworkPolicy includes a podSelector which selects the grouping of pods to which the policy applies. The example policy selects pods with the label "role=db". An empty podSelector selects all pods in the namespace.
policyTypes: Each NetworkPolicy includes a policyTypes list which may include either Ingress, Egress, or both. The policyTypes field indicates whether or not the given policy applies to ingress traffic to selected pod, egress traffic from selected pods, or both. If no policyTypes are specified on a NetworkPolicy then by default Ingress will always be set and Egress will be set if the NetworkPolicy has any egress rules.
ingress: Each NetworkPolicy may include a list of allowed ingress rules. Each rule allows traffic which matches both the from and ports sections. The example policy contains a single rule, which matches traffic on a single port, from one of three sources, the first specified via an ipBlock, the second via a namespaceSelector and the third via a podSelector.
egress: Each NetworkPolicy may include a list of allowed egress rules. Each rule allows traffic which matches both the to and ports sections. The example policy contains a single rule, which matches traffic on a single port to any destination in 10.0.0.0/24.
So, the example NetworkPolicy:
isolates "role=db" pods in the "default" namespace for both ingress and egress traffic (if they weren't already isolated)
(Ingress rules) allows connections to all pods in the "default" namespace with the label "role=db" on TCP port 6379 from:
any pod in the "default" namespace with the label "role=frontend"
any pod in a namespace with the label "project=myproject"
IP addresses in the ranges 172.17.0.0–172.17.0.255 and 172.17.2.0–172.17.255.255 (ie, all of 172.17.0.0/16 except 172.17.1.0/24)
(Egress rules) allows connections from any pod in the "default" namespace with the label "role=db" to CIDR 10.0.0.0/24 on TCP port 5978
There are four kinds of selectors that can be specified in an ingressfrom section or egressto section:
podSelector: This selects particular Pods in the same namespace as the NetworkPolicy which should be allowed as ingress sources or egress destinations.
namespaceSelector: This selects particular namespaces for which all Pods should be allowed as ingress sources or egress destinations.
namespaceSelectorandpodSelector: A single to/from entry that specifies both namespaceSelector and podSelector selects particular Pods within particular namespaces. Be careful to use correct YAML syntax; this policy:
contains two elements in the from array, and allows connections from Pods in the local Namespace with the label role=client, or from any Pod in any namespace with the label user=alice.
When in doubt, use kubectl describe to see how Kubernetes has interpreted the policy.
ipBlock: This selects particular IP CIDR ranges to allow as ingress sources or egress destinations. These should be cluster-external IPs, since Pod IPs are ephemeral and unpredictable.
Cluster ingress and egress mechanisms often require rewriting the source or destination IP
of packets. In cases where this happens, it is not defined whether this happens before or
after NetworkPolicy processing, and the behavior may be different for different
combinations of network plugin, cloud provider, Service implementation, etc.
In the case of ingress, this means that in some cases you may be able to filter incoming
packets based on the actual original source IP, while in other cases, the "source IP" that
the NetworkPolicy acts on may be the IP of a LoadBalancer or of the Pod's node, etc.
For egress, this means that connections from pods to Service IPs that get rewritten to
cluster-external IPs may or may not be subject to ipBlock-based policies.
Default policies
By default, if no policies exist in a namespace, then all ingress and egress traffic is allowed to and from pods in that namespace. The following examples let you change the default behavior
in that namespace.
Default deny all ingress traffic
You can create a "default" ingress isolation policy for a namespace by creating a NetworkPolicy that selects all pods but does not allow any ingress traffic to those pods.
This ensures that even pods that aren't selected by any other NetworkPolicy will still be isolated for ingress. This policy does not affect isolation for egress from any pod.
Allow all ingress traffic
If you want to allow all incoming connections to all pods in a namespace, you can create a policy that explicitly allows that.
With this policy in place, no additional policy or policies can cause any incoming connection to those pods to be denied. This policy has no effect on isolation for egress from any pod.
Default deny all egress traffic
You can create a "default" egress isolation policy for a namespace by creating a NetworkPolicy that selects all pods but does not allow any egress traffic from those pods.
This ensures that even pods that aren't selected by any other NetworkPolicy will not be allowed egress traffic. This policy does not
change the ingress isolation behavior of any pod.
Allow all egress traffic
If you want to allow all connections from all pods in a namespace, you can create a policy that explicitly allows all outgoing connections from pods in that namespace.
With this policy in place, no additional policy or policies can cause any outgoing connection from those pods to be denied. This policy has no effect on isolation for ingress to any pod.
Default deny all ingress and all egress traffic
You can create a "default" policy for a namespace which prevents all ingress AND egress traffic by creating the following NetworkPolicy in that namespace.
This ensures that even pods that aren't selected by any other NetworkPolicy will not be allowed ingress or egress traffic.
SCTP support
FEATURE STATE:Kubernetes v1.20 [stable]
As a stable feature, this is enabled by default. To disable SCTP at a cluster level, you (or your cluster administrator) will need to disable the SCTPSupportfeature gate for the API server with --feature-gates=SCTPSupport=false,….
When the feature gate is enabled, you can set the protocol field of a NetworkPolicy to SCTP.
Note: You must be using a CNI plugin that supports SCTP protocol NetworkPolicies.
Targeting a range of ports
FEATURE STATE:Kubernetes v1.25 [stable]
When writing a NetworkPolicy, you can target a range of ports instead of a single port.
This is achievable with the usage of the endPort field, as the following example:
The above rule allows any Pod with label role=db on the namespace default to communicate
with any IP within the range 10.0.0.0/24 over TCP, provided that the target
port is between the range 32000 and 32768.
The following restrictions apply when using this field:
The endPort field must be equal to or greater than the port field.
endPort can only be defined if port is also defined.
Both ports must be numeric.
Note: Your cluster must be using a CNI plugin that
supports the endPort field in NetworkPolicy specifications.
If your network plugin
does not support the endPort field and you specify a NetworkPolicy with that,
the policy will be applied only for the single port field.
Targeting a Namespace by its name
FEATURE STATE:Kubernetes 1.22 [stable]
The Kubernetes control plane sets an immutable label kubernetes.io/metadata.name on all
namespaces, provided that the NamespaceDefaultLabelNamefeature gate is enabled.
The value of the label is the namespace name.
While NetworkPolicy cannot target a namespace by its name with some object field, you can use the
standardized label to target a specific namespace.
What you can't do with network policies (at least, not yet)
As of Kubernetes 1.25, the following functionality does not exist in the NetworkPolicy API, but you might be able to implement workarounds using Operating System components (such as SELinux, OpenVSwitch, IPTables, and so on) or Layer 7 technologies (Ingress controllers, Service Mesh implementations) or admission controllers. In case you are new to network security in Kubernetes, its worth noting that the following User Stories cannot (yet) be implemented using the NetworkPolicy API.
Forcing internal cluster traffic to go through a common gateway (this might be best served with a service mesh or other proxy).
Anything TLS related (use a service mesh or ingress controller for this).
Node specific policies (you can use CIDR notation for these, but you cannot target nodes by their Kubernetes identities specifically).
Targeting of services by name (you can, however, target pods or namespaces by their labels, which is often a viable workaround).
Creation or management of "Policy requests" that are fulfilled by a third party.
Default policies which are applied to all namespaces or pods (there are some third party Kubernetes distributions and projects which can do this).
Advanced policy querying and reachability tooling.
The ability to log network security events (for example connections that are blocked or accepted).
The ability to explicitly deny policies (currently the model for NetworkPolicies are deny by default, with only the ability to add allow rules).
The ability to prevent loopback or incoming host traffic (Pods cannot currently block localhost access, nor do they have the ability to block access from their resident node).
See more recipes for common scenarios enabled by the NetworkPolicy resource.
3.6.11 - IPv4/IPv6 dual-stack
FEATURE STATE:Kubernetes v1.23 [stable]
IPv4/IPv6 dual-stack networking enables the allocation of both IPv4 and IPv6 addresses to
Pods and Services.
IPv4/IPv6 dual-stack networking is enabled by default for your Kubernetes cluster starting in
1.21, allowing the simultaneous assignment of both IPv4 and IPv6 addresses.
Supported Features
IPv4/IPv6 dual-stack on your Kubernetes cluster provides the following features:
Dual-stack Pod networking (a single IPv4 and IPv6 address assignment per Pod)
IPv4 and IPv6 enabled Services
Pod off-cluster egress routing (eg. the Internet) via both IPv4 and IPv6 interfaces
Prerequisites
The following prerequisites are needed in order to utilize IPv4/IPv6 dual-stack Kubernetes clusters:
Kubernetes 1.20 or later
For information about using dual-stack services with earlier
Kubernetes versions, refer to the documentation for that version
of Kubernetes.
Provider support for dual-stack networking (Cloud provider or otherwise must be able to provide
Kubernetes nodes with routable IPv4/IPv6 network interfaces)
--node-cidr-mask-size-ipv4|--node-cidr-mask-size-ipv6 defaults to /24 for IPv4 and /64 for IPv6
kube-proxy:
--cluster-cidr=<IPv4 CIDR>,<IPv6 CIDR>
kubelet:
when there is no --cloud-provider the administrator can pass a comma-separated pair of IP
addresses via --node-ip to manually configure dual-stack .status.addresses for that Node.
If a Pod runs on that node in HostNetwork mode, the Pod reports these IP addresses in its
.status.podIPs field.
All podIPs in a node match the IP family preference defined by the .status.addresses
field for that Node.
Note:
An example of an IPv4 CIDR: 10.244.0.0/16 (though you would supply your own address range)
An example of an IPv6 CIDR: fdXY:IJKL:MNOP:15::/64 (this shows the format but is not a valid
address - see RFC 4193)
Services
You can create Services which can use IPv4, IPv6, or both.
The address family of a Service defaults to the address family of the first service cluster IP
range (configured via the --service-cluster-ip-range flag to the kube-apiserver).
When you define a Service you can optionally configure it as dual stack. To specify the behavior you want, you
set the .spec.ipFamilyPolicy field to one of the following values:
SingleStack: Single-stack service. The control plane allocates a cluster IP for the Service,
using the first configured service cluster IP range.
PreferDualStack:
Allocates IPv4 and IPv6 cluster IPs for the Service.
RequireDualStack: Allocates Service .spec.ClusterIPs from both IPv4 and IPv6 address ranges.
Selects the .spec.ClusterIP from the list of .spec.ClusterIPs based on the address family
of the first element in the .spec.ipFamilies array.
If you would like to define which IP family to use for single stack or define the order of IP
families for dual-stack, you can choose the address families by setting an optional field,
.spec.ipFamilies, on the Service.
Note: The .spec.ipFamilies field is immutable because the .spec.ClusterIP cannot be reallocated on a
Service that already exists. If you want to change .spec.ipFamilies, delete and recreate the
Service.
You can set .spec.ipFamilies to any of the following array values:
["IPv4"]
["IPv6"]
["IPv4","IPv6"] (dual stack)
["IPv6","IPv4"] (dual stack)
The first family you list is used for the legacy .spec.ClusterIP field.
Dual-stack Service configuration scenarios
These examples demonstrate the behavior of various dual-stack Service configuration scenarios.
Dual-stack options on new Services
This Service specification does not explicitly define .spec.ipFamilyPolicy. When you create
this Service, Kubernetes assigns a cluster IP for the Service from the first configured
service-cluster-ip-range and sets the .spec.ipFamilyPolicy to SingleStack. (Services
without selectors and
headless Services with selectors
will behave in this same way.)
This Service specification explicitly defines PreferDualStack in .spec.ipFamilyPolicy. When
you create this Service on a dual-stack cluster, Kubernetes assigns both IPv4 and IPv6
addresses for the service. The control plane updates the .spec for the Service to record the IP
address assignments. The field .spec.ClusterIPs is the primary field, and contains both assigned
IP addresses; .spec.ClusterIP is a secondary field with its value calculated from
.spec.ClusterIPs.
For the .spec.ClusterIP field, the control plane records the IP address that is from the
same address family as the first service cluster IP range.
On a single-stack cluster, the .spec.ClusterIPs and .spec.ClusterIP fields both only list
one address.
On a cluster with dual-stack enabled, specifying RequireDualStack in .spec.ipFamilyPolicy
behaves the same as PreferDualStack.
This Service specification explicitly defines IPv6 and IPv4 in .spec.ipFamilies as well
as defining PreferDualStack in .spec.ipFamilyPolicy. When Kubernetes assigns an IPv6 and
IPv4 address in .spec.ClusterIPs, .spec.ClusterIP is set to the IPv6 address because that is
the first element in the .spec.ClusterIPs array, overriding the default.
These examples demonstrate the default behavior when dual-stack is newly enabled on a cluster
where Services already exist. (Upgrading an existing cluster to 1.21 or beyond will enable
dual-stack.)
When dual-stack is enabled on a cluster, existing Services (whether IPv4 or IPv6) are
configured by the control plane to set .spec.ipFamilyPolicy to SingleStack and set
.spec.ipFamilies to the address family of the existing Service. The existing Service cluster IP
will be stored in .spec.ClusterIPs.
When dual-stack is enabled on a cluster, existing
headless Services with selectors are
configured by the control plane to set .spec.ipFamilyPolicy to SingleStack and set
.spec.ipFamilies to the address family of the first service cluster IP range (configured via the
--service-cluster-ip-range flag to the kube-apiserver) even though .spec.ClusterIP is set to
None.
Switching Services between single-stack and dual-stack
Services can be changed from single-stack to dual-stack and from dual-stack to single-stack.
To change a Service from single-stack to dual-stack, change .spec.ipFamilyPolicy from
SingleStack to PreferDualStack or RequireDualStack as desired. When you change this
Service from single-stack to dual-stack, Kubernetes assigns the missing address family so that the
Service now has IPv4 and IPv6 addresses.
Edit the Service specification updating the .spec.ipFamilyPolicy from SingleStack to PreferDualStack.
Before:
spec:ipFamilyPolicy:SingleStack
After:
spec:ipFamilyPolicy:PreferDualStack
To change a Service from dual-stack to single-stack, change .spec.ipFamilyPolicy from
PreferDualStack or RequireDualStack to SingleStack. When you change this Service from
dual-stack to single-stack, Kubernetes retains only the first element in the .spec.ClusterIPs
array, and sets .spec.ClusterIP to that IP address and sets .spec.ipFamilies to the address
family of .spec.ClusterIPs.
Headless Services without selector
For Headless Services without selectors
and without .spec.ipFamilyPolicy explicitly set, the .spec.ipFamilyPolicy field defaults to
RequireDualStack.
Service type LoadBalancer
To provision a dual-stack load balancer for your Service:
Set the .spec.type field to LoadBalancer
Set .spec.ipFamilyPolicy field to PreferDualStack or RequireDualStack
Note: To use a dual-stack LoadBalancer type Service, your cloud provider must support IPv4 and IPv6
load balancers.
Egress traffic
If you want to enable egress traffic in order to reach off-cluster destinations (eg. the public
Internet) from a Pod that uses non-publicly routable IPv6 addresses, you need to enable the Pod to
use a publicly routed IPv6 address via a mechanism such as transparent proxying or IP
masquerading. The ip-masq-agent project
supports IP masquerading on dual-stack clusters.
Kubernetes on Windows does not support single-stack "IPv6-only" networking. However,
dual-stack IPv4/IPv6 networking for pods and nodes with single-family services
is supported.
You can use IPv4/IPv6 dual-stack networking with l2bridge networks.
Note: Overlay (VXLAN) networks on Windows do not support dual-stack networking.
You can read more about the different network modes for Windows within the
Networking on Windows topic.
Kubernetes supports running nodes on either Linux or Windows. You can mix both kinds of node
within a single cluster.
This page provides an overview to networking specific to the Windows operating system.
Container networking on Windows
Networking for Windows containers is exposed through
CNI plugins.
Windows containers function similarly to virtual machines in regards to
networking. Each container has a virtual network adapter (vNIC) which is connected
to a Hyper-V virtual switch (vSwitch). The Host Networking Service (HNS) and the
Host Compute Service (HCS) work together to create containers and attach container
vNICs to networks. HCS is responsible for the management of containers whereas HNS
is responsible for the management of networking resources such as:
Virtual networks (including creation of vSwitches)
Endpoints / vNICs
Namespaces
Policies including packet encapsulations, load-balancing rules, ACLs, and NAT rules.
The Windows HNS and vSwitch implement namespacing and can
create virtual NICs as needed for a pod or container. However, many configurations such
as DNS, routes, and metrics are stored in the Windows registry database rather than as
files inside /etc, which is how Linux stores those configurations. The Windows registry for the container
is separate from that of the host, so concepts like mapping /etc/resolv.conf from
the host into a container don't have the same effect they would on Linux. These must
be configured using Windows APIs run in the context of that container. Therefore
CNI implementations need to call the HNS instead of relying on file mappings to pass
network details into the pod or container.
Network modes
Windows supports five different networking drivers/modes: L2bridge, L2tunnel,
Overlay (Beta), Transparent, and NAT. In a heterogeneous cluster with Windows and Linux
worker nodes, you need to select a networking solution that is compatible on both
Windows and Linux. The following table lists the out-of-tree plugins are supported on Windows,
with recommendations on when to use each CNI:
Network Driver
Description
Container Packet Modifications
Network Plugins
Network Plugin Characteristics
L2bridge
Containers are attached to an external vSwitch. Containers are attached to the underlay network, although the physical network doesn't need to learn the container MACs because they are rewritten on ingress/egress.
MAC is rewritten to host MAC, IP may be rewritten to host IP using HNS OutboundNAT policy.
win-bridge uses L2bridge network mode, connects containers to the underlay of hosts, offering best performance. Requires user-defined routes (UDR) for inter-node connectivity.
L2Tunnel
This is a special case of l2bridge, but only used on Azure. All packets are sent to the virtualization host where SDN policy is applied.
Azure-CNI allows integration of containers with Azure vNET, and allows them to leverage the set of capabilities that Azure Virtual Network provides. For example, securely connect to Azure services or use Azure NSGs. See azure-cni for some examples
Overlay
Containers are given a vNIC connected to an external vSwitch. Each overlay network gets its own IP subnet, defined by a custom IP prefix.The overlay network driver uses VXLAN encapsulation.
win-overlay should be used when virtual container networks are desired to be isolated from underlay of hosts (e.g. for security reasons). Allows for IPs to be re-used for different overlay networks (which have different VNID tags) if you are restricted on IPs in your datacenter. This option requires KB4489899 on Windows Server 2019.
Requires an external vSwitch. Containers are attached to an external vSwitch which enables intra-pod communication via logical networks (logical switches and routers).
Packet is encapsulated either via GENEVE or STT tunneling to reach pods which are not on the same host. Packets are forwarded or dropped via the tunnel metadata information supplied by the ovn network controller. NAT is done for north-south communication.
Deploy via ansible. Distributed ACLs can be applied via Kubernetes policies. IPAM support. Load-balancing can be achieved without kube-proxy. NATing is done without using iptables/netsh.
NAT (not used in Kubernetes)
Containers are given a vNIC connected to an internal vSwitch. DNS/DHCP is provided using an internal component called WinNAT
This plugin supports delegating to one of the reference CNI plugins (win-overlay,
win-bridge), to work in conjunction with Flannel daemon on Windows (Flanneld) for
automatic node subnet lease assignment and HNS network creation. This plugin reads
in its own configuration file (cni.conf), and aggregates it with the environment
variables from the FlannelD generated subnet.env file. It then delegates to one of
the reference CNI plugins for network plumbing, and sends the correct configuration
containing the node-assigned subnet to the IPAM plugin (for example: host-local).
For Node, Pod, and Service objects, the following network flows are supported for
TCP/UDP traffic:
Pod → Pod (IP)
Pod → Pod (Name)
Pod → Service (Cluster IP)
Pod → Service (PQDN, but only if there are no ".")
Pod → Service (FQDN)
Pod → external (IP)
Pod → external (DNS)
Node → Pod
Pod → Node
IP address management (IPAM)
The following IPAM options are supported on Windows:
A Kubernetes Service is an abstraction
that defines a logical set of Pods and a means to access them over a network.
In a cluster that includes Windows nodes, you can use the following types of Service:
On Windows, you can use the following settings to configure Services and load
balancing behavior:
Windows Service Settings
Feature
Description
Minimum Supported Windows OS build
How to enable
Session affinity
Ensures that connections from a particular client are passed to the same Pod each time.
Windows Server 2022
Set service.spec.sessionAffinity to "ClientIP"
Direct Server Return (DSR)
Load balancing mode where the IP address fixups and the LBNAT occurs at the container vSwitch port directly; service traffic arrives with the source IP set as the originating pod IP.
Windows Server 2019
Set the following flags in kube-proxy: --feature-gates="WinDSR=true" --enable-dsr=true
Preserve-Destination
Skips DNAT of service traffic, thereby preserving the virtual IP of the target service in packets reaching the backend Pod. Also disables node-node forwarding.
Windows Server, version 1903
Set "preserve-destination": "true" in service annotations and enable DSR in kube-proxy.
IPv4/IPv6 dual-stack networking
Native IPv4-to-IPv4 in parallel with IPv6-to-IPv6 communications to, from, and within a cluster
Ensures that source IP of incoming ingress traffic gets preserved. Also disables node-node forwarding.
Windows Server 2019
Set service.spec.externalTrafficPolicy to "Local" and enable DSR in kube-proxy
Warning:
There are known issue with NodePort Services on overlay networking, if the destination node is running Windows Server 2022.
To avoid the issue entirely, you can configure the service with externalTrafficPolicy: Local.
There are known issues with Pod to Pod connectivity on l2bridge network on Windows Server 2022 with KB5005619 or higher installed.
To workaround the issue and restore Pod to Pod connectivity, you can disable the WinDSR feature in kube-proxy.
The following networking functionality is not supported on Windows nodes:
Host networking mode
Local NodePort access from the node itself (works for other nodes or external clients)
More than 64 backend pods (or unique destination addresses) for a single Service
IPv6 communication between Windows pods connected to overlay networks
Local Traffic Policy in non-DSR mode
Outbound communication using the ICMP protocol via the win-overlay, win-bridge, or using the Azure-CNI plugin.
Specifically, the Windows data plane (VFP)
doesn't support ICMP packet transpositions, and this means:
ICMP packets directed to destinations within the same network (such as pod to pod communication via ping)
work as expected;
TCP/UDP packets work as expected;
ICMP packets directed to pass through a remote network (e.g. pod to external internet communication via ping)
cannot be transposed and thus will not be routed back to their source;
Since TCP/UDP packets can still be transposed, you can substitute ping <destination> with
curl <destination> when debugging connectivity with the outside world.
Other limitations:
Windows reference network plugins win-bridge and win-overlay do not implement
CNI spec v0.4.0,
due to a missing CHECK implementation.
The Flannel VXLAN CNI plugin has the following limitations on Windows:
Node-pod connectivity is only possible for local pods with Flannel v0.12.0 (or higher).
Flannel is restricted to using VNI 4096 and UDP port 4789. See the official
Flannel VXLAN
backend docs for more details on these parameters.
3.7 - Storage
Ways to provide both long-term and temporary storage to Pods in your cluster.
3.7.1 - Volumes
On-disk files in a container are ephemeral, which presents some problems for
non-trivial applications when running in containers. One problem
is the loss of files when a container crashes. The kubelet restarts the container
but with a clean state. A second problem occurs when sharing files
between containers running together in a Pod.
The Kubernetes volume abstraction
solves both of these problems.
Familiarity with Pods is suggested.
Background
Docker has a concept of
volumes, though it is
somewhat looser and less managed. A Docker volume is a directory on
disk or in another container. Docker provides volume
drivers, but the functionality is somewhat limited.
Kubernetes supports many types of volumes. A Pod
can use any number of volume types simultaneously.
Ephemeral volume types have a lifetime of a pod, but persistent volumes exist beyond
the lifetime of a pod. When a pod ceases to exist, Kubernetes destroys ephemeral volumes;
however, Kubernetes does not destroy persistent volumes.
For any kind of volume in a given pod, data is preserved across container restarts.
At its core, a volume is a directory, possibly with some data in it, which
is accessible to the containers in a pod. How that directory comes to be, the
medium that backs it, and the contents of it are determined by the particular
volume type used.
To use a volume, specify the volumes to provide for the Pod in .spec.volumes
and declare where to mount those volumes into containers in .spec.containers[*].volumeMounts.
A process in a container sees a filesystem view composed from the initial contents of
the container image, plus volumes
(if defined) mounted inside the container.
The process sees a root filesystem that initially matches the contents of the container
image.
Any writes to within that filesystem hierarchy, if allowed, affect what that process views
when it performs a subsequent filesystem access.
Volumes mount at the specified paths within
the image.
For each container defined within a Pod, you must independently specify where
to mount each volume that the container uses.
Volumes cannot mount within other volumes (but see Using subPath
for a related mechanism). Also, a volume cannot contain a hard link to anything in
a different volume.
Types of Volumes
Kubernetes supports several types of volumes.
awsElasticBlockStore (deprecated)
FEATURE STATE:Kubernetes v1.17 [deprecated]
An awsElasticBlockStore volume mounts an Amazon Web Services (AWS)
EBS volume into your pod. Unlike
emptyDir, which is erased when a pod is removed, the contents of an EBS
volume are persisted and the volume is unmounted. This means that an
EBS volume can be pre-populated with data, and that data can be shared between pods.
Note: You must create an EBS volume by using aws ec2 create-volume or the AWS API before you can use it.
There are some restrictions when using an awsElasticBlockStore volume:
the nodes on which pods are running must be AWS EC2 instances
those instances need to be in the same region and availability zone as the EBS volume
EBS only supports a single EC2 instance mounting a volume
Creating an AWS EBS volume
Before you can use an EBS volume with a pod, you need to create it.
Make sure the zone matches the zone you brought up your cluster in. Check that the size and EBS volume
type are suitable for your use.
AWS EBS configuration example
apiVersion:v1kind:Podmetadata:name:test-ebsspec:containers:- image:registry.k8s.io/test-webservername:test-containervolumeMounts:- mountPath:/test-ebsname:test-volumevolumes:- name:test-volume# This AWS EBS volume must already exist.awsElasticBlockStore:volumeID:"<volume id>"fsType:ext4
If the EBS volume is partitioned, you can supply the optional field partition: "<partition number>" to specify which partition to mount on.
AWS EBS CSI migration
FEATURE STATE:Kubernetes v1.25 [stable]
The CSIMigration feature for awsElasticBlockStore, when enabled, redirects
all plugin operations from the existing in-tree plugin to the ebs.csi.aws.com Container
Storage Interface (CSI) driver. In order to use this feature, the AWS EBS CSI
driver
must be installed on the cluster.
AWS EBS CSI migration complete
FEATURE STATE:Kubernetes v1.17 [alpha]
To disable the awsElasticBlockStore storage plugin from being loaded by the controller manager
and the kubelet, set the InTreePluginAWSUnregister flag to true.
azureDisk (deprecated)
FEATURE STATE:Kubernetes v1.19 [deprecated]
The azureDisk volume type mounts a Microsoft Azure Data Disk into a pod.
The CSIMigration feature for azureDisk, when enabled, redirects all plugin operations
from the existing in-tree plugin to the disk.csi.azure.com Container
Storage Interface (CSI) Driver. In order to use this feature, the
Azure Disk CSI Driver
must be installed on the cluster.
azureDisk CSI migration complete
FEATURE STATE:Kubernetes v1.21 [alpha]
To disable the azureDisk storage plugin from being loaded by the controller manager
and the kubelet, set the InTreePluginAzureDiskUnregister flag to true.
azureFile (deprecated)
FEATURE STATE:Kubernetes v1.21 [deprecated]
The azureFile volume type mounts a Microsoft Azure File volume (SMB 2.1 and 3.0)
into a pod.
The CSIMigration feature for azureFile, when enabled, redirects all plugin operations
from the existing in-tree plugin to the file.csi.azure.com Container
Storage Interface (CSI) Driver. In order to use this feature, the Azure File CSI
Driver
must be installed on the cluster and the CSIMigrationAzureFilefeature gates must be enabled.
Azure File CSI driver does not support using same volume with different fsgroups. If
CSIMigrationAzureFile is enabled, using same volume with different fsgroups won't be supported at all.
azureFile CSI migration complete
FEATURE STATE:Kubernetes v1.21 [alpha]
To disable the azureFile storage plugin from being loaded by the controller manager
and the kubelet, set the InTreePluginAzureFileUnregister flag to true.
cephfs
A cephfs volume allows an existing CephFS volume to be
mounted into your Pod. Unlike emptyDir, which is erased when a pod is
removed, the contents of a cephfs volume are preserved and the volume is merely
unmounted. This means that a cephfs volume can be pre-populated with data, and
that data can be shared between pods. The cephfs volume can be mounted by multiple
writers simultaneously.
Note: You must have your own Ceph server running with the share exported before you can use it.
Note: Kubernetes must be configured with the OpenStack cloud provider.
The cinder volume type is used to mount the OpenStack Cinder volume into your pod.
Cinder volume configuration example
apiVersion:v1kind:Podmetadata:name:test-cinderspec:containers:- image:registry.k8s.io/test-webservername:test-cinder-containervolumeMounts:- mountPath:/test-cindername:test-volumevolumes:- name:test-volume# This OpenStack volume must already exist.cinder:volumeID:"<volume id>"fsType:ext4
OpenStack CSI migration
FEATURE STATE:Kubernetes v1.24 [stable]
The CSIMigration feature for Cinder is enabled by default since Kubernetes 1.21.
It redirects all plugin operations from the existing in-tree plugin to the
cinder.csi.openstack.org Container Storage Interface (CSI) Driver.
OpenStack Cinder CSI Driver
must be installed on the cluster.
To disable the in-tree Cinder plugin from being loaded by the controller manager
and the kubelet, you can enable the InTreePluginOpenStackUnregisterfeature gate.
configMap
A ConfigMap
provides a way to inject configuration data into pods.
The data stored in a ConfigMap can be referenced in a volume of type
configMap and then consumed by containerized applications running in a pod.
When referencing a ConfigMap, you provide the name of the ConfigMap in the
volume. You can customize the path to use for a specific
entry in the ConfigMap. The following configuration shows how to mount
the log-config ConfigMap onto a Pod called configmap-pod:
The log-config ConfigMap is mounted as a volume, and all contents stored in
its log_level entry are mounted into the Pod at path /etc/config/log_level.
Note that this path is derived from the volume's mountPath and the path
keyed with log_level.
Note:
You must create a ConfigMap
before you can use it.
A container using a ConfigMap as a subPath volume mount will not
receive ConfigMap updates.
Text data is exposed as files using the UTF-8 character encoding. For other character encodings, use binaryData.
downwardAPI
A downwardAPI volume makes downward API
data available to applications. Within the volume, you can find the exposed
data as read-only files in plain text format.
Note: A container using the downward API as a subPath volume mount does not
receive updates when field values change.
An emptyDir volume is first created when a Pod is assigned to a node, and
exists as long as that Pod is running on that node. As the name says, the
emptyDir volume is initially empty. All containers in the Pod can read and write the same
files in the emptyDir volume, though that volume can be mounted at the same
or different paths in each container. When a Pod is removed from a node for
any reason, the data in the emptyDir is deleted permanently.
Note: A container crashing does not remove a Pod from a node. The data in an emptyDir volume
is safe across container crashes.
Some uses for an emptyDir are:
scratch space, such as for a disk-based merge sort
checkpointing a long computation for recovery from crashes
holding files that a content-manager container fetches while a webserver
container serves the data
Depending on your environment, emptyDir volumes are stored on whatever medium that backs the
node such as disk or SSD, or network storage. However, if you set the emptyDir.medium field
to "Memory", Kubernetes mounts a tmpfs (RAM-backed filesystem) for you instead.
While tmpfs is very fast, be aware that unlike disks, tmpfs is cleared on
node reboot and any files you write count against your container's
memory limit.
Note: If the SizeMemoryBackedVolumesfeature gate is enabled,
you can specify a size for memory backed volumes. If no size is specified, memory
backed volumes are sized to 50% of the memory on a Linux host.
An fc volume type allows an existing fibre channel block storage volume
to mount in a Pod. You can specify single or multiple target world wide names (WWNs)
using the parameter targetWWNs in your Volume configuration. If multiple WWNs are specified,
targetWWNs expect that those WWNs are from multi-path connections.
Note: You must configure FC SAN Zoning to allocate and mask those LUNs (volumes) to the target WWNs
beforehand so that Kubernetes hosts can access them.
A gcePersistentDisk volume mounts a Google Compute Engine (GCE)
persistent disk (PD) into your Pod.
Unlike emptyDir, which is erased when a pod is removed, the contents of a PD are
preserved and the volume is merely unmounted. This means that a PD can be
pre-populated with data, and that data can be shared between pods.
Note: You must create a PD using gcloud or the GCE API or UI before you can use it.
There are some restrictions when using a gcePersistentDisk:
the nodes on which Pods are running must be GCE VMs
those VMs need to be in the same GCE project and zone as the persistent disk
One feature of GCE persistent disk is concurrent read-only access to a persistent disk.
A gcePersistentDisk volume permits multiple consumers to simultaneously
mount a persistent disk as read-only. This means that you can pre-populate a PD with your dataset
and then serve it in parallel from as many Pods as you need. Unfortunately,
PDs can only be mounted by a single consumer in read-write mode. Simultaneous
writers are not allowed.
Using a GCE persistent disk with a Pod controlled by a ReplicaSet will fail unless
the PD is read-only or the replica count is 0 or 1.
Creating a GCE persistent disk
Before you can use a GCE persistent disk with a Pod, you need to create it.
apiVersion:v1kind:Podmetadata:name:test-pdspec:containers:- image:registry.k8s.io/test-webservername:test-containervolumeMounts:- mountPath:/test-pdname:test-volumevolumes:- name:test-volume# This GCE PD must already exist.gcePersistentDisk:pdName:my-data-diskfsType:ext4
Regional persistent disks
The Regional persistent disks
feature allows the creation of persistent disks that are available in two zones
within the same region. In order to use this feature, the volume must be provisioned
as a PersistentVolume; referencing the volume directly from a pod is not supported.
Manually provisioning a Regional PD PersistentVolume
Dynamic provisioning is possible using a
StorageClass for GCE PD.
Before creating a PersistentVolume, you must create the persistent disk:
apiVersion:v1kind:PersistentVolumemetadata:name:test-volumespec:capacity:storage:400GiaccessModes:- ReadWriteOncegcePersistentDisk:pdName:my-data-diskfsType:ext4nodeAffinity:required:nodeSelectorTerms:- matchExpressions:# failure-domain.beta.kubernetes.io/zone should be used prior to 1.21- key:topology.kubernetes.io/zoneoperator:Invalues:- us-central1-a- us-central1-b
GCE CSI migration
FEATURE STATE:Kubernetes v1.25 [stable]
The CSIMigration feature for GCE PD, when enabled, redirects all plugin operations
from the existing in-tree plugin to the pd.csi.storage.gke.io Container
Storage Interface (CSI) Driver. In order to use this feature, the GCE PD CSI
Driver
must be installed on the cluster.
GCE CSI migration complete
FEATURE STATE:Kubernetes v1.21 [alpha]
To disable the gcePersistentDisk storage plugin from being loaded by the controller manager
and the kubelet, set the InTreePluginGCEUnregister flag to true.
gitRepo (deprecated)
Warning: The gitRepo volume type is deprecated. To provision a container with a git repo, mount an EmptyDir into an InitContainer that clones the repo using git, then mount the EmptyDir into the Pod's container.
A gitRepo volume is an example of a volume plugin. This plugin
mounts an empty directory and clones a git repository into this directory
for your Pod to use.
A glusterfs volume allows a Glusterfs (an open
source networked filesystem) volume to be mounted into your Pod. Unlike
emptyDir, which is erased when a Pod is removed, the contents of a
glusterfs volume are preserved and the volume is merely unmounted. This
means that a glusterfs volume can be pre-populated with data, and that data can
be shared between pods. GlusterFS can be mounted by multiple writers
simultaneously.
Note: You must have your own GlusterFS installation running before you can use it.
HostPath volumes present many security risks, and it is a best practice to avoid the use of
HostPaths when possible. When a HostPath volume must be used, it should be scoped to only the
required file or directory, and mounted as ReadOnly.
If restricting HostPath access to specific directories through AdmissionPolicy, volumeMounts MUST
be required to use readOnly mounts for the policy to be effective.
A hostPath volume mounts a file or directory from the host node's filesystem
into your Pod. This is not something that most Pods will need, but it offers a
powerful escape hatch for some applications.
For example, some uses for a hostPath are:
running a container that needs access to Docker internals; use a hostPath
of /var/lib/docker
running cAdvisor in a container; use a hostPath of /sys
allowing a Pod to specify whether a given hostPath should exist prior to the
Pod running, whether it should be created, and what it should exist as
In addition to the required path property, you can optionally specify a type for a hostPath volume.
The supported values for field type are:
Value
Behavior
Empty string (default) is for backward compatibility, which means that no checks will be performed before mounting the hostPath volume.
DirectoryOrCreate
If nothing exists at the given path, an empty directory will be created there as needed with permission set to 0755, having the same group and ownership with Kubelet.
Directory
A directory must exist at the given path
FileOrCreate
If nothing exists at the given path, an empty file will be created there as needed with permission set to 0644, having the same group and ownership with Kubelet.
File
A file must exist at the given path
Socket
A UNIX socket must exist at the given path
CharDevice
A character device must exist at the given path
BlockDevice
A block device must exist at the given path
Watch out when using this type of volume, because:
HostPaths can expose privileged system credentials (such as for the Kubelet) or privileged APIs
(such as container runtime socket), which can be used for container escape or to attack other
parts of the cluster.
Pods with identical configuration (such as created from a PodTemplate) may
behave differently on different nodes due to different files on the nodes
The files or directories created on the underlying hosts are only writable by root. You
either need to run your process as root in a
privileged Container or modify the file
permissions on the host to be able to write to a hostPath volume
hostPath configuration example
apiVersion:v1kind:Podmetadata:name:test-pdspec:containers:- image:registry.k8s.io/test-webservername:test-containervolumeMounts:- mountPath:/test-pdname:test-volumevolumes:- name:test-volumehostPath:# directory location on hostpath:/data# this field is optionaltype:Directory
Caution: The FileOrCreate mode does not create the parent directory of the file. If the parent directory
of the mounted file does not exist, the pod fails to start. To ensure that this mode works,
you can try to mount directories and files separately, as shown in the
FileOrCreateconfiguration.
hostPath FileOrCreate configuration example
apiVersion:v1kind:Podmetadata:name:test-webserverspec:containers:- name:test-webserverimage:registry.k8s.io/test-webserver:latestvolumeMounts:- mountPath:/var/local/aaaname:mydir- mountPath:/var/local/aaa/1.txtname:myfilevolumes:- name:mydirhostPath:# Ensure the file directory is created.path:/var/local/aaatype:DirectoryOrCreate- name:myfilehostPath:path:/var/local/aaa/1.txttype:FileOrCreate
iscsi
An iscsi volume allows an existing iSCSI (SCSI over IP) volume to be mounted
into your Pod. Unlike emptyDir, which is erased when a Pod is removed, the
contents of an iscsi volume are preserved and the volume is merely
unmounted. This means that an iscsi volume can be pre-populated with data, and
that data can be shared between pods.
Note: You must have your own iSCSI server running with the volume created before you can use it.
A feature of iSCSI is that it can be mounted as read-only by multiple consumers
simultaneously. This means that you can pre-populate a volume with your dataset
and then serve it in parallel from as many Pods as you need. Unfortunately,
iSCSI volumes can only be mounted by a single consumer in read-write mode.
Simultaneous writers are not allowed.
A local volume represents a mounted local storage device such as a disk,
partition or directory.
Local volumes can only be used as a statically created PersistentVolume. Dynamic
provisioning is not supported.
Compared to hostPath volumes, local volumes are used in a durable and
portable manner without manually scheduling pods to nodes. The system is aware
of the volume's node constraints by looking at the node affinity on the PersistentVolume.
However, local volumes are subject to the availability of the underlying
node and are not suitable for all applications. If a node becomes unhealthy,
then the local volume becomes inaccessible by the pod. The pod using this volume
is unable to run. Applications using local volumes must be able to tolerate this
reduced availability, as well as potential data loss, depending on the
durability characteristics of the underlying disk.
The following example shows a PersistentVolume using a local volume and
nodeAffinity:
You must set a PersistentVolume nodeAffinity when using local volumes.
The Kubernetes scheduler uses the PersistentVolume nodeAffinity to schedule
these Pods to the correct node.
PersistentVolume volumeMode can be set to "Block" (instead of the default
value "Filesystem") to expose the local volume as a raw block device.
When using local volumes, it is recommended to create a StorageClass with
volumeBindingMode set to WaitForFirstConsumer. For more details, see the
local StorageClass example.
Delaying volume binding ensures that the PersistentVolumeClaim binding decision
will also be evaluated with any other node constraints the Pod may have,
such as node resource requirements, node selectors, Pod affinity, and Pod anti-affinity.
An external static provisioner can be run separately for improved management of
the local volume lifecycle. Note that this provisioner does not support dynamic
provisioning yet. For an example on how to run an external local provisioner,
see the local volume provisioner user
guide.
Note: The local PersistentVolume requires manual cleanup and deletion by the
user if the external static provisioner is not used to manage the volume
lifecycle.
nfs
An nfs volume allows an existing NFS (Network File System) share to be
mounted into a Pod. Unlike emptyDir, which is erased when a Pod is
removed, the contents of an nfs volume are preserved and the volume is merely
unmounted. This means that an NFS volume can be pre-populated with data, and
that data can be shared between pods. NFS can be mounted by multiple
writers simultaneously.
Note: You must have your own NFS server running with the share exported before you can use it.
A persistentVolumeClaim volume is used to mount a
PersistentVolume into a Pod. PersistentVolumeClaims
are a way for users to "claim" durable storage (such as a GCE PersistentDisk or an
iSCSI volume) without knowing the details of the particular cloud environment.
A portworxVolume is an elastic block storage layer that runs hyperconverged with
Kubernetes. Portworx fingerprints storage
in a server, tiers based on capabilities, and aggregates capacity across multiple servers.
Portworx runs in-guest in virtual machines or on bare metal Linux nodes.
A portworxVolume can be dynamically created through Kubernetes or it can also
be pre-provisioned and referenced inside a Pod.
Here is an example Pod referencing a pre-provisioned Portworx volume:
apiVersion:v1kind:Podmetadata:name:test-portworx-volume-podspec:containers:- image:registry.k8s.io/test-webservername:test-containervolumeMounts:- mountPath:/mntname:pxvolvolumes:- name:pxvol# This Portworx volume must already exist.portworxVolume:volumeID:"pxvol"fsType:"<fs-type>"
Note: Make sure you have an existing PortworxVolume with name pxvol
before using it in the Pod.
The CSIMigration feature for Portworx has been added but disabled by default in Kubernetes 1.23 since it's in alpha state.
It has been beta now since v1.25 but it is still turned off by default.
It redirects all plugin operations from the existing in-tree plugin to the
pxd.portworx.com Container Storage Interface (CSI) Driver.
Portworx CSI Driver
must be installed on the cluster.
To enable the feature, set CSIMigrationPortworx=true in kube-controller-manager and kubelet.
projected
A projected volume maps several existing volume sources into the same
directory. For more details, see projected volumes.
rbd
An rbd volume allows a
Rados Block Device (RBD) volume to mount
into your Pod. Unlike emptyDir, which is erased when a pod is removed, the
contents of an rbd volume are preserved and the volume is unmounted. This
means that a RBD volume can be pre-populated with data, and that data can be
shared between pods.
Note: You must have a Ceph installation running before you can use RBD.
A feature of RBD is that it can be mounted as read-only by multiple consumers
simultaneously. This means that you can pre-populate a volume with your dataset
and then serve it in parallel from as many pods as you need. Unfortunately,
RBD volumes can only be mounted by a single consumer in read-write mode.
Simultaneous writers are not allowed.
The CSIMigration feature for RBD, when enabled, redirects all plugin
operations from the existing in-tree plugin to the rbd.csi.ceph.comCSI driver. In order to use this
feature, the
Ceph CSI driver
must be installed on the cluster and the CSIMigrationRBDfeature gate
must be enabled. (Note that the csiMigrationRBD flag has been removed and
replaced with CSIMigrationRBD in release v1.24)
Note:
As a Kubernetes cluster operator that administers storage, here are the
prerequisites that you must complete before you attempt migration to the
RBD CSI driver:
You must install the Ceph CSI driver (rbd.csi.ceph.com), v3.5.0 or above,
into your Kubernetes cluster.
considering the clusterID field is a required parameter for CSI driver for
its operations, but in-tree StorageClass has monitors field as a required
parameter, a Kubernetes storage admin has to create a clusterID based on the
monitors hash ( ex:#echo -n '<monitors_string>' | md5sum) in the CSI config map and keep the monitors
under this clusterID configuration.
Also, if the value of adminId in the in-tree Storageclass is different from
admin, the adminSecretName mentioned in the in-tree Storageclass has to be
patched with the base64 value of the adminId parameter value, otherwise this
step can be skipped.
secret
A secret volume is used to pass sensitive information, such as passwords, to
Pods. You can store secrets in the Kubernetes API and mount them as files for
use by pods without coupling to Kubernetes directly. secret volumes are
backed by tmpfs (a RAM-backed filesystem) so they are never written to
non-volatile storage.
Note: You must create a Secret in the Kubernetes API before you can use it.
Note: A container using a Secret as a subPath volume mount will not
receive Secret updates.
Note: We recommend to use vSphere CSI out-of-tree driver instead.
A vsphereVolume is used to mount a vSphere VMDK volume into your Pod. The contents
of a volume are preserved when it is unmounted. It supports both VMFS and VSAN datastore.
The CSIMigrationvSphere feature for vsphereVolume is enabled by default as of Kubernetes v1.25.
All plugin operations from the in-tree vspherevolume will be redirected to the csi.vsphere.vmware.comCSI driver unless CSIMigrationvSphere feature gate is disabled.
As of Kubernetes v1.25, vSphere releases less than 7.0u2 are not supported for the
(deprecated) in-tree vSphere storage driver. You must run vSphere 7.0u2 or later
in order to either continue using the deprecated driver, or to migrate to
the replacement CSI driver.
If you are running a version of Kubernetes other than v1.25, consult
the documentation for that version of Kubernetes.
Note:
The following StorageClass parameters from the built-in vsphereVolume plugin are not supported by the vSphere CSI driver:
diskformat
hostfailurestotolerate
forceprovisioning
cachereservation
diskstripes
objectspacereservation
iopslimit
Existing volumes created using these parameters will be migrated to the vSphere CSI driver,
but new volumes created by the vSphere CSI driver will not be honoring these parameters.
vSphere CSI migration complete
FEATURE STATE:Kubernetes v1.19 [beta]
To turn off the vsphereVolume plugin from being loaded by the controller manager and the kubelet, you need to set InTreePluginvSphereUnregister feature flag to true. You must install a csi.vsphere.vmware.comCSI driver on all worker nodes.
Using subPath
Sometimes, it is useful to share one volume for multiple uses in a single pod.
The volumeMounts.subPath property specifies a sub-path inside the referenced volume
instead of its root.
The following example shows how to configure a Pod with a LAMP stack (Linux Apache MySQL PHP)
using a single, shared volume. This sample subPath configuration is not recommended
for production use.
The PHP application's code and assets map to the volume's html folder and
the MySQL database is stored in the volume's mysql folder. For example:
Use the subPathExpr field to construct subPath directory names from
downward API environment variables.
The subPath and subPathExpr properties are mutually exclusive.
In this example, a Pod uses subPathExpr to create a directory pod1 within
the hostPath volume /var/log/pods.
The hostPath volume takes the Pod name from the downwardAPI.
The host directory /var/log/pods/pod1 is mounted at /logs in the container.
apiVersion:v1kind:Podmetadata:name:pod1spec:containers:- name:container1env:- name:POD_NAMEvalueFrom:fieldRef:apiVersion:v1fieldPath:metadata.nameimage:busybox:1.28command:["sh","-c","while [ true ]; do echo 'Hello'; sleep 10; done | tee -a /logs/hello.txt"]volumeMounts:- name:workdir1mountPath:/logs# The variable expansion uses round brackets (not curly brackets).subPathExpr:$(POD_NAME)restartPolicy:Nevervolumes:- name:workdir1hostPath:path:/var/log/pods
Resources
The storage media (such as Disk or SSD) of an emptyDir volume is determined by the
medium of the filesystem holding the kubelet root dir (typically
/var/lib/kubelet). There is no limit on how much space an emptyDir or
hostPath volume can consume, and no isolation between containers or between
pods.
To learn about requesting space using a resource specification, see
how to manage resources.
Out-of-tree volume plugins
The out-of-tree volume plugins include
Container Storage Interface (CSI), and also FlexVolume (which is deprecated). These plugins enable storage vendors to create custom storage plugins
without adding their plugin source code to the Kubernetes repository.
Previously, all volume plugins were "in-tree". The "in-tree" plugins were built, linked, compiled,
and shipped with the core Kubernetes binaries. This meant that adding a new storage system to
Kubernetes (a volume plugin) required checking code into the core Kubernetes code repository.
Both CSI and FlexVolume allow volume plugins to be developed independent of
the Kubernetes code base, and deployed (installed) on Kubernetes clusters as
extensions.
For storage vendors looking to create an out-of-tree volume plugin, please refer
to the volume plugin FAQ.
csi
Container Storage Interface
(CSI) defines a standard interface for container orchestration systems (like
Kubernetes) to expose arbitrary storage systems to their container workloads.
Note: Support for CSI spec versions 0.2 and 0.3 are deprecated in Kubernetes
v1.13 and will be removed in a future release.
Note: CSI drivers may not be compatible across all Kubernetes releases.
Please check the specific CSI driver's documentation for supported
deployments steps for each Kubernetes release and a compatibility matrix.
Once a CSI compatible volume driver is deployed on a Kubernetes cluster, users
may use the csi volume type to attach or mount the volumes exposed by the
CSI driver.
A csi volume can be used in a Pod in three different ways:
The following fields are available to storage administrators to configure a CSI
persistent volume:
driver: A string value that specifies the name of the volume driver to use.
This value must correspond to the value returned in the GetPluginInfoResponse
by the CSI driver as defined in the CSI spec.
It is used by Kubernetes to identify which CSI driver to call out to, and by
CSI driver components to identify which PV objects belong to the CSI driver.
volumeHandle: A string value that uniquely identifies the volume. This value
must correspond to the value returned in the volume.id field of the
CreateVolumeResponse by the CSI driver as defined in the CSI spec.
The value is passed as volume_id on all calls to the CSI volume driver when
referencing the volume.
readOnly: An optional boolean value indicating whether the volume is to be
"ControllerPublished" (attached) as read only. Default is false. This value is
passed to the CSI driver via the readonly field in the
ControllerPublishVolumeRequest.
fsType: If the PV's VolumeMode is Filesystem then this field may be used
to specify the filesystem that should be used to mount the volume. If the
volume has not been formatted and formatting is supported, this value will be
used to format the volume.
This value is passed to the CSI driver via the VolumeCapability field of
ControllerPublishVolumeRequest, NodeStageVolumeRequest, and
NodePublishVolumeRequest.
volumeAttributes: A map of string to string that specifies static properties
of a volume. This map must correspond to the map returned in the
volume.attributes field of the CreateVolumeResponse by the CSI driver as
defined in the CSI spec.
The map is passed to the CSI driver via the volume_context field in the
ControllerPublishVolumeRequest, NodeStageVolumeRequest, and
NodePublishVolumeRequest.
controllerPublishSecretRef: A reference to the secret object containing
sensitive information to pass to the CSI driver to complete the CSI
ControllerPublishVolume and ControllerUnpublishVolume calls. This field is
optional, and may be empty if no secret is required. If the Secret
contains more than one secret, all secrets are passed.
nodeExpandSecretRef: A reference to the secret containing sensitive
information to pass to the CSI driver to complete the CSI
NodeExpandVolume call. This field is optional, and may be empty if no
secret is required. If the object contains more than one secret, all
secrets are passed. When you have configured secret data for node-initiated
volume expansion, the kubelet passes that data via the NodeExpandVolume()
call to the CSI driver. In order to use the nodeExpandSecretRef field, your
cluster should be running Kubernetes version 1.25 or later and you must enable
the feature gate
named CSINodeExpandSecret for each kube-apiserver and for the kubelet on every
node. You must also be using a CSI driver that supports or requires secret data during
node-initiated storage resize operations.
nodePublishSecretRef: A reference to the secret object containing
sensitive information to pass to the CSI driver to complete the CSI
NodePublishVolume call. This field is optional, and may be empty if no
secret is required. If the secret object contains more than one secret, all
secrets are passed.
nodeStageSecretRef: A reference to the secret object containing
sensitive information to pass to the CSI driver to complete the CSI
NodeStageVolume call. This field is optional, and may be empty if no secret
is required. If the Secret contains more than one secret, all secrets
are passed.
CSI raw block volume support
FEATURE STATE:Kubernetes v1.18 [stable]
Vendors with external CSI drivers can implement raw block volume support
in Kubernetes workloads.
You can directly configure CSI volumes within the Pod
specification. Volumes specified in this way are ephemeral and do not
persist across pod restarts. See Ephemeral
Volumes
for more information.
CSI node plugins need to perform various privileged
operations like scanning of disk devices and mounting of file systems. These operations
differ for each host operating system. For Linux worker nodes, containerized CSI node
node plugins are typically deployed as privileged containers. For Windows worker nodes,
privileged operations for containerized CSI node plugins is supported using
csi-proxy, a community-managed,
stand-alone binary that needs to be pre-installed on each Windows node.
For more details, refer to the deployment guide of the CSI plugin you wish to deploy.
Migrating to CSI drivers from in-tree plugins
FEATURE STATE:Kubernetes v1.25 [stable]
The CSIMigration feature directs operations against existing in-tree
plugins to corresponding CSI plugins (which are expected to be installed and configured).
As a result, operators do not have to make any
configuration changes to existing Storage Classes, PersistentVolumes or PersistentVolumeClaims
(referring to in-tree plugins) when transitioning to a CSI driver that supersedes an in-tree plugin.
The operations and features that are supported include:
provisioning/delete, attach/detach, mount/unmount and resizing of volumes.
In-tree plugins that support CSIMigration and have a corresponding CSI driver implemented
are listed in Types of Volumes.
The following in-tree plugins support persistent storage on Windows nodes:
FlexVolume is an out-of-tree plugin interface that uses an exec-based model to interface
with storage drivers. The FlexVolume driver binaries must be installed in a pre-defined
volume plugin path on each node and in some cases the control plane nodes as well.
Pods interact with FlexVolume drivers through the flexVolume in-tree volume plugin.
For more details, see the FlexVolume README document.
The following FlexVolume plugins,
deployed as PowerShell scripts on the host, support Windows nodes:
FlexVolume is deprecated. Using an out-of-tree CSI driver is the recommended way to integrate external storage with Kubernetes.
Maintainers of FlexVolume driver should implement a CSI Driver and help to migrate users of FlexVolume drivers to CSI.
Users of FlexVolume should move their workloads to use the equivalent CSI Driver.
Mount propagation
Mount propagation allows for sharing volumes mounted by a container to
other containers in the same pod, or even to other pods on the same node.
Mount propagation of a volume is controlled by the mountPropagation field
in Container.volumeMounts. Its values are:
None - This volume mount will not receive any subsequent mounts
that are mounted to this volume or any of its subdirectories by the host.
In similar fashion, no mounts created by the container will be visible on
the host. This is the default mode.
HostToContainer - This volume mount will receive all subsequent mounts
that are mounted to this volume or any of its subdirectories.
In other words, if the host mounts anything inside the volume mount, the
container will see it mounted there.
Similarly, if any Pod with Bidirectional mount propagation to the same
volume mounts anything there, the container with HostToContainer mount
propagation will see it.
Bidirectional - This volume mount behaves the same the HostToContainer mount.
In addition, all volume mounts created by the container will be propagated
back to the host and to all containers of all pods that use the same volume.
A typical use case for this mode is a Pod with a FlexVolume or CSI driver or
a Pod that needs to mount something on the host using a hostPath volume.
Warning:Bidirectional mount propagation can be dangerous. It can damage
the host operating system and therefore it is allowed only in privileged
containers. Familiarity with Linux kernel behavior is strongly recommended.
In addition, any volume mounts created by containers in pods must be destroyed
(unmounted) by the containers on termination.
Configuration
Before mount propagation can work properly on some deployments (CoreOS,
RedHat/Centos, Ubuntu) mount share must be configured correctly in
Docker as shown below.
Edit your Docker's systemd service file. Set MountFlags as follows:
MountFlags=shared
Or, remove MountFlags=slave if present. Then restart the Docker daemon:
This document describes persistent volumes in Kubernetes. Familiarity with volumes is suggested.
Introduction
Managing storage is a distinct problem from managing compute instances. The PersistentVolume subsystem provides an API for users and administrators that abstracts details of how storage is provided from how it is consumed. To do this, we introduce two new API resources: PersistentVolume and PersistentVolumeClaim.
A PersistentVolume (PV) is a piece of storage in the cluster that has been provisioned by an administrator or dynamically provisioned using Storage Classes. It is a resource in the cluster just like a node is a cluster resource. PVs are volume plugins like Volumes, but have a lifecycle independent of any individual Pod that uses the PV. This API object captures the details of the implementation of the storage, be that NFS, iSCSI, or a cloud-provider-specific storage system.
A PersistentVolumeClaim (PVC) is a request for storage by a user. It is similar to a Pod. Pods consume node resources and PVCs consume PV resources. Pods can request specific levels of resources (CPU and Memory). Claims can request specific size and access modes (e.g., they can be mounted ReadWriteOnce, ReadOnlyMany or ReadWriteMany, see AccessModes).
While PersistentVolumeClaims allow a user to consume abstract storage resources, it is common that users need PersistentVolumes with varying properties, such as performance, for different problems. Cluster administrators need to be able to offer a variety of PersistentVolumes that differ in more ways than size and access modes, without exposing users to the details of how those volumes are implemented. For these needs, there is the StorageClass resource.
PVs are resources in the cluster. PVCs are requests for those resources and also act as claim checks to the resource. The interaction between PVs and PVCs follows this lifecycle:
Provisioning
There are two ways PVs may be provisioned: statically or dynamically.
Static
A cluster administrator creates a number of PVs. They carry the details of the real storage, which is available for use by cluster users. They exist in the Kubernetes API and are available for consumption.
Dynamic
When none of the static PVs the administrator created match a user's PersistentVolumeClaim,
the cluster may try to dynamically provision a volume specially for the PVC.
This provisioning is based on StorageClasses: the PVC must request a
storage class and
the administrator must have created and configured that class for dynamic
provisioning to occur. Claims that request the class "" effectively disable
dynamic provisioning for themselves.
To enable dynamic storage provisioning based on storage class, the cluster administrator
needs to enable the DefaultStorageClassadmission controller
on the API server. This can be done, for example, by ensuring that DefaultStorageClass is
among the comma-delimited, ordered list of values for the --enable-admission-plugins flag of
the API server component. For more information on API server command-line flags,
check kube-apiserver documentation.
Binding
A user creates, or in the case of dynamic provisioning, has already created, a PersistentVolumeClaim with a specific amount of storage requested and with certain access modes. A control loop in the master watches for new PVCs, finds a matching PV (if possible), and binds them together. If a PV was dynamically provisioned for a new PVC, the loop will always bind that PV to the PVC. Otherwise, the user will always get at least what they asked for, but the volume may be in excess of what was requested. Once bound, PersistentVolumeClaim binds are exclusive, regardless of how they were bound. A PVC to PV binding is a one-to-one mapping, using a ClaimRef which is a bi-directional binding between the PersistentVolume and the PersistentVolumeClaim.
Claims will remain unbound indefinitely if a matching volume does not exist. Claims will be bound as matching volumes become available. For example, a cluster provisioned with many 50Gi PVs would not match a PVC requesting 100Gi. The PVC can be bound when a 100Gi PV is added to the cluster.
Using
Pods use claims as volumes. The cluster inspects the claim to find the bound volume and mounts that volume for a Pod. For volumes that support multiple access modes, the user specifies which mode is desired when using their claim as a volume in a Pod.
Once a user has a claim and that claim is bound, the bound PV belongs to the user for as long as they need it. Users schedule Pods and access their claimed PVs by including a persistentVolumeClaim section in a Pod's volumes block. See Claims As Volumes for more details on this.
Storage Object in Use Protection
The purpose of the Storage Object in Use Protection feature is to ensure that PersistentVolumeClaims (PVCs) in active use by a Pod and PersistentVolume (PVs) that are bound to PVCs are not removed from the system, as this may result in data loss.
Note: PVC is in active use by a Pod when a Pod object exists that is using the PVC.
If a user deletes a PVC in active use by a Pod, the PVC is not removed immediately. PVC removal is postponed until the PVC is no longer actively used by any Pods. Also, if an admin deletes a PV that is bound to a PVC, the PV is not removed immediately. PV removal is postponed until the PV is no longer bound to a PVC.
You can see that a PVC is protected when the PVC's status is Terminating and the Finalizers list includes kubernetes.io/pvc-protection:
When a user is done with their volume, they can delete the PVC objects from the API that allows reclamation of the resource. The reclaim policy for a PersistentVolume tells the cluster what to do with the volume after it has been released of its claim. Currently, volumes can either be Retained, Recycled, or Deleted.
Retain
The Retain reclaim policy allows for manual reclamation of the resource. When the PersistentVolumeClaim is deleted, the PersistentVolume still exists and the volume is considered "released". But it is not yet available for another claim because the previous claimant's data remains on the volume. An administrator can manually reclaim the volume with the following steps.
Delete the PersistentVolume. The associated storage asset in external infrastructure (such as an AWS EBS, GCE PD, Azure Disk, or Cinder volume) still exists after the PV is deleted.
Manually clean up the data on the associated storage asset accordingly.
Manually delete the associated storage asset.
If you want to reuse the same storage asset, create a new PersistentVolume with the same storage asset definition.
Delete
For volume plugins that support the Delete reclaim policy, deletion removes both the PersistentVolume object from Kubernetes, as well as the associated storage asset in the external infrastructure, such as an AWS EBS, GCE PD, Azure Disk, or Cinder volume. Volumes that were dynamically provisioned inherit the reclaim policy of their StorageClass, which defaults to Delete. The administrator should configure the StorageClass according to users' expectations; otherwise, the PV must be edited or patched after it is created. See Change the Reclaim Policy of a PersistentVolume.
Recycle
Warning: The Recycle reclaim policy is deprecated. Instead, the recommended approach is to use dynamic provisioning.
If supported by the underlying volume plugin, the Recycle reclaim policy performs a basic scrub (rm -rf /thevolume/*) on the volume and makes it available again for a new claim.
However, an administrator can configure a custom recycler Pod template using
the Kubernetes controller manager command line arguments as described in the
reference.
The custom recycler Pod template must contain a volumes specification, as
shown in the example below:
However, the particular path specified in the custom recycler Pod template in the volumes part is replaced with the particular path of the volume that is being recycled.
PersistentVolume deletion protection finalizer
FEATURE STATE:Kubernetes v1.23 [alpha]
Finalizers can be added on a PersistentVolume to ensure that PersistentVolumes
having Delete reclaim policy are deleted only after the backing storage are deleted.
The newly introduced finalizers kubernetes.io/pv-controller and external-provisioner.volume.kubernetes.io/finalizer
are only added to dynamically provisioned volumes.
The finalizer kubernetes.io/pv-controller is added to in-tree plugin volumes. The following is an example
When the CSIMigration{provider} feature flag is enabled for a specific in-tree volume plugin,
the kubernetes.io/pv-controller finalizer is replaced by the
external-provisioner.volume.kubernetes.io/finalizer finalizer.
By specifying a PersistentVolume in a PersistentVolumeClaim, you declare a binding between that specific PV and PVC.
If the PersistentVolume exists and has not reserved PersistentVolumeClaims through its claimRef field, then the PersistentVolume and PersistentVolumeClaim will be bound.
The binding happens regardless of some volume matching criteria, including node affinity.
The control plane still checks that storage class, access modes, and requested storage size are valid.
apiVersion:v1kind:PersistentVolumeClaimmetadata:name:foo-pvcnamespace:foospec:storageClassName:""# Empty string must be explicitly set otherwise default StorageClass will be setvolumeName:foo-pv...
This method does not guarantee any binding privileges to the PersistentVolume. If other PersistentVolumeClaims could use the PV that you specify, you first need to reserve that storage volume. Specify the relevant PersistentVolumeClaim in the claimRef field of the PV so that other PVCs can not bind to it.
This is useful if you want to consume PersistentVolumes that have their claimPolicy set
to Retain, including cases where you are reusing an existing PV.
Expanding Persistent Volumes Claims
FEATURE STATE:Kubernetes v1.11 [beta]
Support for expanding PersistentVolumeClaims (PVCs) is enabled by default. You can expand
the following types of volumes:
To request a larger volume for a PVC, edit the PVC object and specify a larger
size. This triggers expansion of the volume that backs the underlying PersistentVolume. A
new PersistentVolume is never created to satisfy the claim. Instead, an existing volume is resized.
Warning: Directly editing the size of a PersistentVolume can prevent an automatic resize of that volume.
If you edit the capacity of a PersistentVolume, and then edit the .spec of a matching
PersistentVolumeClaim to make the size of the PersistentVolumeClaim match the PersistentVolume,
then no storage resize happens.
The Kubernetes control plane will see that the desired state of both resources matches,
conclude that the backing volume size has been manually
increased and that no resize is necessary.
CSI Volume expansion
FEATURE STATE:Kubernetes v1.16 [beta]
Support for expanding CSI volumes is enabled by default but it also requires a specific CSI driver to support volume expansion. Refer to documentation of the specific CSI driver for more information.
Resizing a volume containing a file system
You can only resize volumes containing a file system if the file system is XFS, Ext3, or Ext4.
When a volume contains a file system, the file system is only resized when a new Pod is using
the PersistentVolumeClaim in ReadWrite mode. File system expansion is either done when a Pod is starting up
or when a Pod is running and the underlying file system supports online expansion.
FlexVolumes (deprecated since Kubernetes v1.23) allow resize if the driver is configured with the
RequiresFSResize capability to true. The FlexVolume can be resized on Pod restart.
Resizing an in-use PersistentVolumeClaim
FEATURE STATE:Kubernetes v1.24 [stable]
In this case, you don't need to delete and recreate a Pod or deployment that is using an existing PVC.
Any in-use PVC automatically becomes available to its Pod as soon as its file system has been expanded.
This feature has no effect on PVCs that are not in use by a Pod or deployment. You must create a Pod that
uses the PVC before the expansion can complete.
Similar to other volume types - FlexVolume volumes can also be expanded when in-use by a Pod.
Note: FlexVolume resize is possible only when the underlying driver supports resize.
Note: Expanding EBS volumes is a time-consuming operation. Also, there is a per-volume quota of one modification every 6 hours.
Recovering from Failure when Expanding Volumes
If a user specifies a new size that is too big to be satisfied by underlying storage system, expansion of PVC will be continuously retried until user or cluster administrator takes some action. This can be undesirable and hence Kubernetes provides following methods of recovering from such failures.
If expanding underlying storage fails, the cluster administrator can manually recover the Persistent Volume Claim (PVC) state and cancel the resize requests. Otherwise, the resize requests are continuously retried by the controller without administrator intervention.
Mark the PersistentVolume(PV) that is bound to the PersistentVolumeClaim(PVC) with Retain reclaim policy.
Delete the PVC. Since PV has Retain reclaim policy - we will not lose any data when we recreate the PVC.
Delete the claimRef entry from PV specs, so as new PVC can bind to it. This should make the PV Available.
Re-create the PVC with smaller size than PV and set volumeName field of the PVC to the name of the PV. This should bind new PVC to existing PV.
Don't forget to restore the reclaim policy of the PV.
FEATURE STATE:Kubernetes v1.23 [alpha]
Note: Recovery from failing PVC expansion by users is available as an alpha feature since Kubernetes 1.23. The RecoverVolumeExpansionFailure feature must be enabled for this feature to work. Refer to the feature gate documentation for more information.
If the feature gates RecoverVolumeExpansionFailure is
enabled in your cluster, and expansion has failed for a PVC, you can retry expansion with a
smaller size than the previously requested value. To request a new expansion attempt with a
smaller proposed size, edit .spec.resources for that PVC and choose a value that is less than the
value you previously tried.
This is useful if expansion to a higher value did not succeed because of capacity constraint.
If that has happened, or you suspect that it might have, you can retry expansion by specifying a
size that is within the capacity limits of underlying storage provider. You can monitor status of resize operation by watching .status.resizeStatus and events on the PVC.
Note that,
although you can specify a lower amount of storage than what was requested previously,
the new value must still be higher than .status.capacity.
Kubernetes does not support shrinking a PVC to less than its current size.
Types of Persistent Volumes
PersistentVolume types are implemented as plugins. Kubernetes currently supports the following plugins:
vsphereVolume - vSphere VMDK volume
(deprecated in v1.19)
Older versions of Kubernetes also supported the following in-tree PersistentVolume types:
photonPersistentDisk - Photon controller persistent disk.
(not available starting v1.15)
scaleIO - ScaleIO volume
(not available starting v1.21)
flocker - Flocker storage
(not available starting v1.25)
quobyte - Quobyte volume
(not available starting v1.25)
storageos - StorageOS volume
(not available starting v1.25)
Persistent Volumes
Each PV contains a spec and status, which is the specification and status of the volume.
The name of a PersistentVolume object must be a valid
DNS subdomain name.
Note: Helper programs relating to the volume type may be required for consumption of a PersistentVolume within a cluster. In this example, the PersistentVolume is of type NFS and the helper program /sbin/mount.nfs is required to support the mounting of NFS filesystems.
Capacity
Generally, a PV will have a specific storage capacity. This is set using the PV's capacity attribute. Read the glossary term Quantity to understand the units expected by capacity.
Currently, storage size is the only resource that can be set or requested. Future attributes may include IOPS, throughput, etc.
Volume Mode
FEATURE STATE:Kubernetes v1.18 [stable]
Kubernetes supports two volumeModes of PersistentVolumes: Filesystem and Block.
volumeMode is an optional API parameter.
Filesystem is the default mode used when volumeMode parameter is omitted.
A volume with volumeMode: Filesystem is mounted into Pods into a directory. If the volume
is backed by a block device and the device is empty, Kubernetes creates a filesystem
on the device before mounting it for the first time.
You can set the value of volumeMode to Block to use a volume as a raw block device.
Such volume is presented into a Pod as a block device, without any filesystem on it.
This mode is useful to provide a Pod the fastest possible way to access a volume, without
any filesystem layer between the Pod and the volume. On the other hand, the application
running in the Pod must know how to handle a raw block device.
See Raw Block Volume Support
for an example on how to use a volume with volumeMode: Block in a Pod.
Access Modes
A PersistentVolume can be mounted on a host in any way supported by the resource provider. As shown in the table below, providers will have different capabilities and each PV's access modes are set to the specific modes supported by that particular volume. For example, NFS can support multiple read/write clients, but a specific NFS PV might be exported on the server as read-only. Each PV gets its own set of access modes describing that specific PV's capabilities.
The access modes are:
ReadWriteOnce
the volume can be mounted as read-write by a single node. ReadWriteOnce access mode still can allow multiple pods to access the volume when the pods are running on the same node.
ReadOnlyMany
the volume can be mounted as read-only by many nodes.
ReadWriteMany
the volume can be mounted as read-write by many nodes.
ReadWriteOncePod
the volume can be mounted as read-write by a single Pod. Use ReadWriteOncePod access mode if you want to ensure that only one pod across whole cluster can read that PVC or write to it. This is only supported for CSI volumes and Kubernetes version 1.22+.
Note: Kubernetes uses volume access modes to match PersistentVolumeClaims and PersistentVolumes.
In some cases, the volume access modes also constrain where the PersistentVolume can be mounted.
Volume access modes do not enforce write protection once the storage has been mounted.
Even if the access modes are specified as ReadWriteOnce, ReadOnlyMany, or ReadWriteMany, they don't set any constraints on the volume.
For example, even if a PersistentVolume is created as ReadOnlyMany, it is no guarantee that it will be read-only.
If the access modes are specified as ReadWriteOncePod, the volume is constrained and can be mounted on only a single Pod.
Important! A volume can only be mounted using one access mode at a time, even if it supports many. For example, a GCEPersistentDisk can be mounted as ReadWriteOnce by a single node or ReadOnlyMany by many nodes, but not at the same time.
A PV can have a class, which is specified by setting the
storageClassName attribute to the name of a
StorageClass.
A PV of a particular class can only be bound to PVCs requesting
that class. A PV with no storageClassName has no class and can only be bound
to PVCs that request no particular class.
In the past, the annotation volume.beta.kubernetes.io/storage-class was used instead
of the storageClassName attribute. This annotation is still working; however,
it will become fully deprecated in a future Kubernetes release.
Reclaim Policy
Current reclaim policies are:
Retain -- manual reclamation
Recycle -- basic scrub (rm -rf /thevolume/*)
Delete -- associated storage asset such as AWS EBS, GCE PD, Azure Disk, or OpenStack Cinder volume is deleted
Currently, only NFS and HostPath support recycling. AWS EBS, GCE PD, Azure Disk, and Cinder volumes support deletion.
Mount Options
A Kubernetes administrator can specify additional mount options for when a Persistent Volume is mounted on a node.
Note: Not all Persistent Volume types support mount options.
The following volume types support mount options:
awsElasticBlockStore
azureDisk
azureFile
cephfs
cinder (deprecated in v1.18)
gcePersistentDisk
glusterfs
iscsi
nfs
rbd
vsphereVolume
Mount options are not validated. If a mount option is invalid, the mount fails.
In the past, the annotation volume.beta.kubernetes.io/mount-options was used instead
of the mountOptions attribute. This annotation is still working; however,
it will become fully deprecated in a future Kubernetes release.
Node Affinity
Note: For most volume types, you do not need to set this field. It is automatically populated for AWS EBS, GCE PD and Azure Disk volume block types. You need to explicitly set this for local volumes.
A PV can specify node affinity to define constraints that limit what nodes this volume can be accessed from. Pods that use a PV will only be scheduled to nodes that are selected by the node affinity. To specify node affinity, set nodeAffinity in the .spec of a PV. The PersistentVolume API reference has more details on this field.
Phase
A volume will be in one of the following phases:
Available -- a free resource that is not yet bound to a claim
Bound -- the volume is bound to a claim
Released -- the claim has been deleted, but the resource is not yet reclaimed by the cluster
Failed -- the volume has failed its automatic reclamation
The CLI will show the name of the PVC bound to the PV.
PersistentVolumeClaims
Each PVC contains a spec and status, which is the specification and status of the claim.
The name of a PersistentVolumeClaim object must be a valid
DNS subdomain name.
apiVersion:v1kind:PersistentVolumeClaimmetadata:name:myclaimspec:accessModes:- ReadWriteOncevolumeMode:Filesystemresources:requests:storage:8GistorageClassName:slowselector:matchLabels:release:"stable"matchExpressions:- {key: environment, operator: In, values:[dev]}
Claims use the same convention as volumes to indicate the consumption of the volume as either a filesystem or block device.
Resources
Claims, like Pods, can request specific quantities of a resource. In this case, the request is for storage. The same resource model applies to both volumes and claims.
Selector
Claims can specify a label selector to further filter the set of volumes. Only the volumes whose labels match the selector can be bound to the claim. The selector can consist of two fields:
matchLabels - the volume must have a label with this value
matchExpressions - a list of requirements made by specifying key, list of values, and operator that relates the key and values. Valid operators include In, NotIn, Exists, and DoesNotExist.
All of the requirements, from both matchLabels and matchExpressions, are ANDed together – they must all be satisfied in order to match.
Class
A claim can request a particular class by specifying the name of a
StorageClass
using the attribute storageClassName.
Only PVs of the requested class, ones with the same storageClassName as the PVC, can
be bound to the PVC.
PVCs don't necessarily have to request a class. A PVC with its storageClassName set
equal to "" is always interpreted to be requesting a PV with no class, so it
can only be bound to PVs with no class (no annotation or one set equal to
""). A PVC with no storageClassName is not quite the same and is treated differently
by the cluster, depending on whether the
DefaultStorageClass admission plugin
is turned on.
If the admission plugin is turned on, the administrator may specify a
default StorageClass. All PVCs that have no storageClassName can be bound only to
PVs of that default. Specifying a default StorageClass is done by setting the
annotation storageclass.kubernetes.io/is-default-class equal to true in
a StorageClass object. If the administrator does not specify a default, the
cluster responds to PVC creation as if the admission plugin were turned off. If
more than one default is specified, the admission plugin forbids the creation of
all PVCs.
If the admission plugin is turned off, there is no notion of a default
StorageClass. All PVCs that have storageClassName set to "" can be
bound only to PVs that have storageClassName also set to "".
However, PVCs with missing storageClassName can be updated later once
default StorageClass becomes available. If the PVC gets updated it will no
longer bind to PVs that have storageClassName also set to "".
Depending on installation method, a default StorageClass may be deployed
to a Kubernetes cluster by addon manager during installation.
When a PVC specifies a selector in addition to requesting a StorageClass,
the requirements are ANDed together: only a PV of the requested class and with
the requested labels may be bound to the PVC.
Note: Currently, a PVC with a non-empty selector can't have a PV dynamically provisioned for it.
In the past, the annotation volume.beta.kubernetes.io/storage-class was used instead
of storageClassName attribute. This annotation is still working; however,
it won't be supported in a future Kubernetes release.
Retroactive default StorageClass assignment
FEATURE STATE:Kubernetes v1.25 [alpha]
You can create a PersistentVolumeClaim without specifying a storageClassName for the new PVC, and you can do so even when no default StorageClass exists in your cluster. In this case, the new PVC creates as you defined it, and the storageClassName of that PVC remains unset until default becomes available.
However, if you enable the RetroactiveDefaultStorageClass feature gate then Kubernetes behaves differently: existing PVCs without storageClassName update to use the new default StorageClass.
When a default StorageClass becomes available, the control plane identifies any existing PVCs without storageClassName. For the PVCs that either have an empty value for storageClassName or do not have this key, the control plane then updates those PVCs to set storageClassName to match the new default StorageClass. If you have an existing PVC where the storageClassName is "", and you configure a default StorageClass, then this PVC will not get updated.
In order to keep binding to PVs with storageClassName set to "" (while a default StorageClass is present), you need to set the storageClassName of the associated PVC to "".
This behavior helps administrators change default StorageClass by removing the old one first and then creating or setting another one. This brief window while there is no default causes PVCs without storageClassName created at that time to not have any default, but due to the retroactive default StorageClass assignment this way of changing defaults is safe.
Claims As Volumes
Pods access storage by using the claim as a volume. Claims must exist in the same namespace as the Pod using the claim. The cluster finds the claim in the Pod's namespace and uses it to get the PersistentVolume backing the claim. The volume is then mounted to the host and into the Pod.
PersistentVolumes binds are exclusive, and since PersistentVolumeClaims are namespaced objects, mounting claims with "Many" modes (ROX, RWX) is only possible within one namespace.
PersistentVolumes typed hostPath
A hostPath PersistentVolume uses a file or directory on the Node to emulate network-attached storage.
See an example of hostPath typed volume.
Raw Block Volume Support
FEATURE STATE:Kubernetes v1.18 [stable]
The following volume plugins support raw block volumes, including dynamic provisioning where
applicable:
Note: When adding a raw block device for a Pod, you specify the device path in the container instead of a mount path.
Binding Block Volumes
If a user requests a raw block volume by indicating this using the volumeMode field in the PersistentVolumeClaim spec, the binding rules differ slightly from previous releases that didn't consider this mode as part of the spec.
Listed is a table of possible combinations the user and admin might specify for requesting a raw block device. The table indicates if the volume will be bound or not given the combinations:
Volume binding matrix for statically provisioned volumes:
PV volumeMode
PVC volumeMode
Result
unspecified
unspecified
BIND
unspecified
Block
NO BIND
unspecified
Filesystem
BIND
Block
unspecified
NO BIND
Block
Block
BIND
Block
Filesystem
NO BIND
Filesystem
Filesystem
BIND
Filesystem
Block
NO BIND
Filesystem
unspecified
BIND
Note: Only statically provisioned volumes are supported for alpha release. Administrators should take care to consider these values when working with raw block devices.
Volume Snapshot and Restore Volume from Snapshot Support
FEATURE STATE:Kubernetes v1.20 [stable]
Volume snapshots only support the out-of-tree CSI volume plugins. For details, see Volume Snapshots.
In-tree volume plugins are deprecated. You can read about the deprecated volume plugins in the Volume Plugin FAQ.
Create a PersistentVolumeClaim from a Volume Snapshot
Kubernetes supports custom volume populators.
To use custom volume populators, you must enable the AnyVolumeDataSourcefeature gate for
the kube-apiserver and kube-controller-manager.
Volume populators take advantage of a PVC spec field called dataSourceRef. Unlike the
dataSource field, which can only contain either a reference to another PersistentVolumeClaim
or to a VolumeSnapshot, the dataSourceRef field can contain a reference to any object in the
same namespace, except for core objects other than PVCs. For clusters that have the feature
gate enabled, use of the dataSourceRef is preferred over dataSource.
Data source references
The dataSourceRef field behaves almost the same as the dataSource field. If either one is
specified while the other is not, the API server will give both fields the same value. Neither
field can be changed after creation, and attempting to specify different values for the two
fields will result in a validation error. Therefore the two fields will always have the same
contents.
There are two differences between the dataSourceRef field and the dataSource field that
users should be aware of:
The dataSource field ignores invalid values (as if the field was blank) while the
dataSourceRef field never ignores values and will cause an error if an invalid value is
used. Invalid values are any core object (objects with no apiGroup) except for PVCs.
The dataSourceRef field may contain different types of objects, while the dataSource field
only allows PVCs and VolumeSnapshots.
Users should always use dataSourceRef on clusters that have the feature gate enabled, and
fall back to dataSource on clusters that do not. It is not necessary to look at both fields
under any circumstance. The duplicated values with slightly different semantics exist only for
backwards compatibility. In particular, a mixture of older and newer controllers are able to
interoperate because the fields are the same.
Using volume populators
Volume populators are controllers that can
create non-empty volumes, where the contents of the volume are determined by a Custom Resource.
Users create a populated volume by referring to a Custom Resource using the dataSourceRef field:
Because volume populators are external components, attempts to create a PVC that uses one
can fail if not all the correct components are installed. External controllers should generate
events on the PVC to provide feedback on the status of the creation, including warnings if
the PVC cannot be created due to some missing component.
You can install the alpha volume data source validator
controller into your cluster. That controller generates warning Events on a PVC in the case that no populator
is registered to handle that kind of data source. When a suitable populator is installed for a PVC, it's the
responsibility of that populator controller to report Events that relate to volume creation and issues during
the process.
Writing Portable Configuration
If you're writing configuration templates or examples that run on a wide range of clusters
and need persistent storage, it is recommended that you use the following pattern:
Include PersistentVolumeClaim objects in your bundle of config (alongside
Deployments, ConfigMaps, etc).
Do not include PersistentVolume objects in the config, since the user instantiating
the config may not have permission to create PersistentVolumes.
Give the user the option of providing a storage class name when instantiating
the template.
If the user provides a storage class name, put that value into the
persistentVolumeClaim.storageClassName field.
This will cause the PVC to match the right storage
class if the cluster has StorageClasses enabled by the admin.
If the user does not provide a storage class name, leave the
persistentVolumeClaim.storageClassName field as nil. This will cause a
PV to be automatically provisioned for the user with the default StorageClass
in the cluster. Many cluster environments have a default StorageClass installed,
or administrators can create their own default StorageClass.
In your tooling, watch for PVCs that are not getting bound after some time
and surface this to the user, as this may indicate that the cluster has no
dynamic storage support (in which case the user should create a matching PV)
or the cluster has no storage system (in which case the user cannot deploy
config requiring PVCs).
Each projected volume source is listed in the spec under sources. The
parameters are nearly the same with two exceptions:
For secrets, the secretName field has been changed to name to be consistent
with ConfigMap naming.
The defaultMode can only be specified at the projected level and not for each
volume source. However, as illustrated above, you can explicitly set the mode
for each individual projection.
serviceAccountToken projected volumes
When the TokenRequestProjection feature is enabled, you can inject the token
for the current service account
into a Pod at a specified path. For example:
The example Pod has a projected volume containing the injected service account
token. Containers in this Pod can use that token to access the Kubernetes API
server, authenticating with the identity of the pod's ServiceAccount.
The audience field contains the intended audience of the
token. A recipient of the token must identify itself with an identifier specified
in the audience of the token, and otherwise should reject the token. This field
is optional and it defaults to the identifier of the API server.
The expirationSeconds is the expected duration of validity of the service account
token. It defaults to 1 hour and must be at least 10 minutes (600 seconds). An administrator
can also limit its maximum value by specifying the --service-account-max-token-expiration
option for the API server. The path field specifies a relative path to the mount point
of the projected volume.
Note: A container using a projected volume source as a subPath
volume mount will not receive updates for those volume sources.
SecurityContext interactions
The proposal for file permission handling in projected service account volume enhancement introduced the projected files having the correct owner permissions set.
Linux
In Linux pods that have a projected volume and RunAsUser set in the Pod
SecurityContext,
the projected files have the correct ownership set including container user
ownership.
Windows
In Windows pods that have a projected volume and RunAsUsername set in the
Pod SecurityContext, the ownership is not enforced due to the way user
accounts are managed in Windows. Windows stores and manages local user and group
accounts in a database file called Security Account Manager (SAM). Each
container maintains its own instance of the SAM database, to which the host has
no visibility into while the container is running. Windows containers are
designed to run the user mode portion of the OS in isolation from the host,
hence the maintenance of a virtual SAM database. As a result, the kubelet running
on the host does not have the ability to dynamically configure host file
ownership for virtualized container accounts. It is recommended that if files on
the host machine are to be shared with the container then they should be placed
into their own volume mount outside of C:\.
By default, the projected files will have the following ownership as shown for
an example projected volume file:
This implies all administrator users like ContainerAdministrator will have
read, write and execute access while, non-administrator users will have read and
execute access.
Note:
In general, granting the container access to the host is discouraged as it can
open the door for potential security exploits.
Creating a Windows Pod with RunAsUser in it's SecurityContext will result in
the Pod being stuck at ContainerCreating forever. So it is advised to not use
the Linux only RunAsUser option with Windows Pods.
3.7.4 - Ephemeral Volumes
This document describes ephemeral volumes in Kubernetes. Familiarity
with volumes is suggested, in
particular PersistentVolumeClaim and PersistentVolume.
Some application need additional storage but don't care whether that
data is stored persistently across restarts. For example, caching
services are often limited by memory size and can move infrequently
used data into storage that is slower than memory with little impact
on overall performance.
Other applications expect some read-only input data to be present in
files, like configuration data or secret keys.
Ephemeral volumes are designed for these use cases. Because volumes
follow the Pod's lifetime and get created and deleted along with the
Pod, Pods can be stopped and restarted without being limited to where
some persistent volume is available.
Ephemeral volumes are specified inline in the Pod spec, which
simplifies application deployment and management.
Types of ephemeral volumes
Kubernetes supports several different kinds of ephemeral volumes for
different purposes:
emptyDir: empty at Pod startup,
with storage coming locally from the kubelet base directory (usually
the root disk) or RAM
generic ephemeral volumes, which
can be provided by all storage drivers that also support persistent volumes
emptyDir, configMap, downwardAPI, secret are provided as
local ephemeral
storage.
They are managed by kubelet on each node.
CSI ephemeral volumes must be provided by third-party CSI storage
drivers.
Generic ephemeral volumes can be provided by third-party CSI storage
drivers, but also by any other storage driver that supports dynamic
provisioning. Some CSI drivers are written specifically for CSI
ephemeral volumes and do not support dynamic provisioning: those then
cannot be used for generic ephemeral volumes.
The advantage of using third-party drivers is that they can offer
functionality that Kubernetes itself does not support, for example
storage with different performance characteristics than the disk that
is managed by kubelet, or injecting different data.
CSI ephemeral volumes
FEATURE STATE:Kubernetes v1.25 [stable]
Note: CSI ephemeral volumes are only supported by a subset of CSI drivers.
The Kubernetes CSI Drivers list
shows which drivers support ephemeral volumes.
Conceptually, CSI ephemeral volumes are similar to configMap,
downwardAPI and secret volume types: the storage is managed locally on each
node and is created together with other local resources after a Pod has been
scheduled onto a node. Kubernetes has no concept of rescheduling Pods
anymore at this stage. Volume creation has to be unlikely to fail,
otherwise Pod startup gets stuck. In particular, storage capacity
aware Pod scheduling is not
supported for these volumes. They are currently also not covered by
the storage resource usage limits of a Pod, because that is something
that kubelet can only enforce for storage that it manages itself.
Here's an example manifest for a Pod that uses CSI ephemeral storage:
The volumeAttributes determine what volume is prepared by the
driver. These attributes are specific to each driver and not
standardized. See the documentation of each CSI driver for further
instructions.
CSI driver restrictions
CSI ephemeral volumes allow users to provide volumeAttributes
directly to the CSI driver as part of the Pod spec. A CSI driver
allowing volumeAttributes that are typically restricted to
administrators is NOT suitable for use in an inline ephemeral volume.
For example, parameters that are normally defined in the StorageClass
should not be exposed to users through the use of inline ephemeral volumes.
Cluster administrators who need to restrict the CSI drivers that are
allowed to be used as inline volumes within a Pod spec may do so by:
Removing Ephemeral from volumeLifecycleModes in the CSIDriver spec, which prevents the
driver from being used as an inline ephemeral volume.
Generic ephemeral volumes are similar to emptyDir volumes in the
sense that they provide a per-pod directory for scratch data that is
usually empty after provisioning. But they may also have additional
features:
Storage can be local or network-attached.
Volumes can have a fixed size that Pods are not able to exceed.
Volumes may have some initial data, depending on the driver and
parameters.
The key design idea is that the
parameters for a volume claim
are allowed inside a volume source of the Pod. Labels, annotations and
the whole set of fields for a PersistentVolumeClaim are supported. When such a Pod gets
created, the ephemeral volume controller then creates an actual PersistentVolumeClaim
object in the same namespace as the Pod and ensures that the PersistentVolumeClaim
gets deleted when the Pod gets deleted.
That triggers volume binding and/or provisioning, either immediately if
the StorageClass uses immediate volume binding or when the Pod is
tentatively scheduled onto a node (WaitForFirstConsumer volume
binding mode). The latter is recommended for generic ephemeral volumes
because then the scheduler is free to choose a suitable node for
the Pod. With immediate binding, the scheduler is forced to select a node that has
access to the volume once it is available.
In terms of resource ownership,
a Pod that has generic ephemeral storage is the owner of the PersistentVolumeClaim(s)
that provide that ephemeral storage. When the Pod is deleted,
the Kubernetes garbage collector deletes the PVC, which then usually
triggers deletion of the volume because the default reclaim policy of
storage classes is to delete volumes. You can create quasi-ephemeral local storage
using a StorageClass with a reclaim policy of retain: the storage outlives the Pod,
and in this case you need to ensure that volume clean up happens separately.
While these PVCs exist, they can be used like any other PVC. In
particular, they can be referenced as data source in volume cloning or
snapshotting. The PVC object also holds the current status of the
volume.
PersistentVolumeClaim naming
Naming of the automatically created PVCs is deterministic: the name is
a combination of Pod name and volume name, with a hyphen (-) in the
middle. In the example above, the PVC name will be
my-app-scratch-volume. This deterministic naming makes it easier to
interact with the PVC because one does not have to search for it once
the Pod name and volume name are known.
The deterministic naming also introduces a potential conflict between different
Pods (a Pod "pod-a" with volume "scratch" and another Pod with name
"pod" and volume "a-scratch" both end up with the same PVC name
"pod-a-scratch") and between Pods and manually created PVCs.
Such conflicts are detected: a PVC is only used for an ephemeral
volume if it was created for the Pod. This check is based on the
ownership relationship. An existing PVC is not overwritten or
modified. But this does not resolve the conflict because without the
right PVC, the Pod cannot start.
Caution: Take care when naming Pods and volumes inside the
same namespace, so that these conflicts can't occur.
Security
Enabling the GenericEphemeralVolume feature allows users to create
PVCs indirectly if they can create Pods, even if they do not have
permission to create PVCs directly. Cluster administrators must be
aware of this. If this does not fit their security model, they should
use an admission webhook
that rejects objects like Pods that have a generic ephemeral volume.
The normal namespace quota for PVCs
still applies, so even if users are allowed to use this new mechanism, they cannot use
it to circumvent other policies.
This document describes the concept of a StorageClass in Kubernetes. Familiarity
with volumes and
persistent volumes is suggested.
Introduction
A StorageClass provides a way for administrators to describe the "classes" of
storage they offer. Different classes might map to quality-of-service levels,
or to backup policies, or to arbitrary policies determined by the cluster
administrators. Kubernetes itself is unopinionated about what classes
represent. This concept is sometimes called "profiles" in other storage
systems.
The StorageClass Resource
Each StorageClass contains the fields provisioner, parameters, and
reclaimPolicy, which are used when a PersistentVolume belonging to the
class needs to be dynamically provisioned.
The name of a StorageClass object is significant, and is how users can
request a particular class. Administrators set the name and other parameters
of a class when first creating StorageClass objects, and the objects cannot
be updated once they are created.
Administrators can specify a default StorageClass only for PVCs that don't
request any particular class to bind to: see the
PersistentVolumeClaim section
for details.
You are not restricted to specifying the "internal" provisioners
listed here (whose names are prefixed with "kubernetes.io" and shipped
alongside Kubernetes). You can also run and specify external provisioners,
which are independent programs that follow a specification
defined by Kubernetes. Authors of external provisioners have full discretion
over where their code lives, how the provisioner is shipped, how it needs to be
run, what volume plugin it uses (including Flex), etc. The repository
kubernetes-sigs/sig-storage-lib-external-provisioner
houses a library for writing external provisioners that implements the bulk of
the specification. Some external provisioners are listed under the repository
kubernetes-sigs/sig-storage-lib-external-provisioner.
For example, NFS doesn't provide an internal provisioner, but an external
provisioner can be used. There are also cases when 3rd party storage
vendors provide their own external provisioner.
Reclaim Policy
PersistentVolumes that are dynamically created by a StorageClass will have the
reclaim policy specified in the reclaimPolicy field of the class, which can be
either Delete or Retain. If no reclaimPolicy is specified when a
StorageClass object is created, it will default to Delete.
PersistentVolumes that are created manually and managed via a StorageClass will have
whatever reclaim policy they were assigned at creation.
Allow Volume Expansion
FEATURE STATE:Kubernetes v1.11 [beta]
PersistentVolumes can be configured to be expandable. This feature when set to true,
allows the users to resize the volume by editing the corresponding PVC object.
The following types of volumes support volume expansion, when the underlying
StorageClass has the field allowVolumeExpansion set to true.
Table of Volume types and the version of Kubernetes they require
Volume type
Required Kubernetes version
gcePersistentDisk
1.11
awsElasticBlockStore
1.11
Cinder
1.11
glusterfs
1.11
rbd
1.11
Azure File
1.11
Azure Disk
1.11
Portworx
1.11
FlexVolume
1.13
CSI
1.14 (alpha), 1.16 (beta)
Note: You can only use the volume expansion feature to grow a Volume, not to shrink it.
Mount Options
PersistentVolumes that are dynamically created by a StorageClass will have the
mount options specified in the mountOptions field of the class.
If the volume plugin does not support mount options but mount options are
specified, provisioning will fail. Mount options are not validated on either
the class or PV. If a mount option is invalid, the PV mount fails.
The Immediate mode indicates that volume binding and dynamic
provisioning occurs once the PersistentVolumeClaim is created. For storage
backends that are topology-constrained and not globally accessible from all Nodes
in the cluster, PersistentVolumes will be bound or provisioned without knowledge of the Pod's scheduling
requirements. This may result in unschedulable Pods.
A cluster administrator can address this issue by specifying the WaitForFirstConsumer mode which
will delay the binding and provisioning of a PersistentVolume until a Pod using the PersistentVolumeClaim is created.
PersistentVolumes will be selected or provisioned conforming to the topology that is
specified by the Pod's scheduling constraints. These include, but are not limited to, resource
requirements,
node selectors,
pod affinity and
anti-affinity,
and taints and tolerations.
The following plugins support WaitForFirstConsumer with dynamic provisioning:
CSI volumes are also supported with dynamic provisioning
and pre-created PVs, but you'll need to look at the documentation for a specific CSI driver
to see its supported topology keys and examples.
Note:
If you choose to use WaitForFirstConsumer, do not use nodeName in the Pod spec
to specify node affinity. If nodeName is used in this case, the scheduler will be bypassed and PVC will remain in pending state.
Instead, you can use node selector for hostname in this case as shown below.
When a cluster operator specifies the WaitForFirstConsumer volume binding mode, it is no longer necessary
to restrict provisioning to specific topologies in most situations. However,
if still required, allowedTopologies can be specified.
This example demonstrates how to restrict the topology of provisioned volumes to specific
zones and should be used as a replacement for the zone and zones parameters for the
supported plugins.
Storage Classes have parameters that describe volumes belonging to the storage
class. Different parameters may be accepted depending on the provisioner. For
example, the value io1, for the parameter type, and the parameter
iopsPerGB are specific to EBS. When a parameter is omitted, some default is
used.
There can be at most 512 parameters defined for a StorageClass.
The total length of the parameters object including its keys and values cannot
exceed 256 KiB.
type: io1, gp2, sc1, st1. See
AWS docs
for details. Default: gp2.
zone (Deprecated): AWS zone. If neither zone nor zones is specified, volumes are
generally round-robin-ed across all active zones where Kubernetes cluster
has a node. zone and zones parameters must not be used at the same time.
zones (Deprecated): A comma separated list of AWS zone(s). If neither zone nor zones
is specified, volumes are generally round-robin-ed across all active zones
where Kubernetes cluster has a node. zone and zones parameters must not
be used at the same time.
iopsPerGB: only for io1 volumes. I/O operations per second per GiB. AWS
volume plugin multiplies this with size of requested volume to compute IOPS
of the volume and caps it at 20 000 IOPS (maximum supported by AWS, see
AWS docs.
A string is expected here, i.e. "10", not 10.
fsType: fsType that is supported by kubernetes. Default: "ext4".
encrypted: denotes whether the EBS volume should be encrypted or not.
Valid values are "true" or "false". A string is expected here,
i.e. "true", not true.
kmsKeyId: optional. The full Amazon Resource Name of the key to use when
encrypting the volume. If none is supplied but encrypted is true, a key is
generated by AWS. See AWS docs for valid ARN value.
Note:zone and zones parameters are deprecated and replaced with
allowedTopologies
zone (Deprecated): GCE zone. If neither zone nor zones is specified, volumes are
generally round-robin-ed across all active zones where Kubernetes cluster has
a node. zone and zones parameters must not be used at the same time.
zones (Deprecated): A comma separated list of GCE zone(s). If neither zone nor zones
is specified, volumes are generally round-robin-ed across all active zones
where Kubernetes cluster has a node. zone and zones parameters must not
be used at the same time.
fstype: ext4 or xfs. Default: ext4. The defined filesystem type must be supported by the host operating system.
replication-type: none or regional-pd. Default: none.
If replication-type is set to none, a regular (zonal) PD will be provisioned.
If replication-type is set to regional-pd, a
Regional Persistent Disk
will be provisioned. It's highly recommended to have
volumeBindingMode: WaitForFirstConsumer set, in which case when you create
a Pod that consumes a PersistentVolumeClaim which uses this StorageClass, a
Regional Persistent Disk is provisioned with two zones. One zone is the same
as the zone that the Pod is scheduled in. The other zone is randomly picked
from the zones available to the cluster. Disk zones can be further constrained
using allowedTopologies.
Note:zone and zones parameters are deprecated and replaced with
allowedTopologies
resturl: Gluster REST service/Heketi service url which provision gluster
volumes on demand. The general format should be IPaddress:Port and this is
a mandatory parameter for GlusterFS dynamic provisioner. If Heketi service is
exposed as a routable service in openshift/kubernetes setup, this can have a
format similar to http://heketi-storage-project.cloudapps.mystorage.com
where the fqdn is a resolvable Heketi service url.
restauthenabled : Gluster REST service authentication boolean that enables
authentication to the REST server. If this value is "true", restuser and
restuserkey or secretNamespace + secretName have to be filled. This
option is deprecated, authentication is enabled when any of restuser,
restuserkey, secretName or secretNamespace is specified.
restuser : Gluster REST service/Heketi user who has access to create volumes
in the Gluster Trusted Pool.
restuserkey : Gluster REST service/Heketi user's password which will be used
for authentication to the REST server. This parameter is deprecated in favor
of secretNamespace + secretName.
secretNamespace, secretName : Identification of Secret instance that
contains user password to use when talking to Gluster REST service. These
parameters are optional, empty password will be used when both
secretNamespace and secretName are omitted. The provided secret must have
type "kubernetes.io/glusterfs", for example created in this way:
clusterid: 630372ccdc720a92c681fb928f27b53f is the ID of the cluster
which will be used by Heketi when provisioning the volume. It can also be a
list of clusterids, for example:
"8452344e2becec931ece4e33c4674e4e,42982310de6c63381718ccfa6d8cf397". This
is an optional parameter.
gidMin, gidMax : The minimum and maximum value of GID range for the
StorageClass. A unique value (GID) in this range ( gidMin-gidMax ) will be
used for dynamically provisioned volumes. These are optional values. If not
specified, the volume will be provisioned with a value between 2000-2147483647
which are defaults for gidMin and gidMax respectively.
volumetype : The volume type and its parameters can be configured with this
optional value. If the volume type is not mentioned, it's up to the provisioner
to decide the volume type.
For example:
Replica volume: volumetype: replicate:3 where '3' is replica count.
Disperse/EC volume: volumetype: disperse:4:2 where '4' is data and '2' is the redundancy count.
Distribute volume: volumetype: none
For available volume types and administration options, refer to the
Administration Guide.
When persistent volumes are dynamically provisioned, the Gluster plugin
automatically creates an endpoint and a headless service in the name
gluster-dynamic-<claimname>. The dynamic endpoint and service are automatically
deleted when the persistent volume claim is deleted.
server: Server is the hostname or IP address of the NFS server.
path: Path that is exported by the NFS server.
readOnly: A flag indicating whether the storage will be mounted as read only (default false).
Kubernetes doesn't include an internal NFS provisioner. You need to use an external provisioner to create a StorageClass for NFS.
Here are some examples:
datastore: The user can also specify the datastore in the StorageClass.
The volume will be created on the datastore specified in the StorageClass,
which in this case is VSANDatastore. This field is optional. If the
datastore is not specified, then the volume will be created on the datastore
specified in the vSphere config file used to initialize the vSphere Cloud
Provider.
Storage Policy Management inside kubernetes
Using existing vCenter SPBM policy
One of the most important features of vSphere for Storage Management is
policy based Management. Storage Policy Based Management (SPBM) is a
storage policy framework that provides a single unified control plane
across a broad range of data services and storage solutions. SPBM enables
vSphere administrators to overcome upfront storage provisioning challenges,
such as capacity planning, differentiated service levels and managing
capacity headroom.
The SPBM policies can be specified in the StorageClass using the
storagePolicyName parameter.
Virtual SAN policy support inside Kubernetes
Vsphere Infrastructure (VI) Admins will have the ability to specify custom
Virtual SAN Storage Capabilities during dynamic volume provisioning. You
can now define storage requirements, such as performance and availability,
in the form of storage capabilities during dynamic volume provisioning.
The storage capability requirements are converted into a Virtual SAN
policy which are then pushed down to the Virtual SAN layer when a
persistent volume (virtual disk) is being created. The virtual disk is
distributed across the Virtual SAN datastore to meet the requirements.
monitors: Ceph monitors, comma delimited. This parameter is required.
adminId: Ceph client ID that is capable of creating images in the pool.
Default is "admin".
adminSecretName: Secret Name for adminId. This parameter is required.
The provided secret must have type "kubernetes.io/rbd".
adminSecretNamespace: The namespace for adminSecretName. Default is "default".
pool: Ceph RBD pool. Default is "rbd".
userId: Ceph client ID that is used to map the RBD image. Default is the
same as adminId.
userSecretName: The name of Ceph Secret for userId to map RBD image. It
must exist in the same namespace as PVCs. This parameter is required.
The provided secret must have type "kubernetes.io/rbd", for example created in this
way:
userSecretNamespace: The namespace for userSecretName.
fsType: fsType that is supported by kubernetes. Default: "ext4".
imageFormat: Ceph RBD image format, "1" or "2". Default is "2".
imageFeatures: This parameter is optional and should only be used if you
set imageFormat to "2". Currently supported features are layering only.
Default is "", and no features are turned on.
skuName: Azure storage account Sku tier. Default is empty.
location: Azure storage account location. Default is empty.
storageAccount: Azure storage account name. If a storage account is provided,
it must reside in the same resource group as the cluster, and location is
ignored. If a storage account is not provided, a new storage account will be
created in the same resource group as the cluster.
storageaccounttype: Azure storage account Sku tier. Default is empty.
kind: Possible values are shared, dedicated, and managed (default).
When kind is shared, all unmanaged disks are created in a few shared
storage accounts in the same resource group as the cluster. When kind is
dedicated, a new dedicated storage account will be created for the new
unmanaged disk in the same resource group as the cluster. When kind is
managed, all managed disks are created in the same resource group as
the cluster.
resourceGroup: Specify the resource group in which the Azure disk will be created.
It must be an existing resource group name. If it is unspecified, the disk will be
placed in the same resource group as the current Kubernetes cluster.
Premium VM can attach both Standard_LRS and Premium_LRS disks, while Standard
VM can only attach Standard_LRS disks.
Managed VM can only attach managed disks and unmanaged VM can only attach
unmanaged disks.
skuName: Azure storage account Sku tier. Default is empty.
location: Azure storage account location. Default is empty.
storageAccount: Azure storage account name. Default is empty. If a storage
account is not provided, all storage accounts associated with the resource
group are searched to find one that matches skuName and location. If a
storage account is provided, it must reside in the same resource group as the
cluster, and skuName and location are ignored.
secretNamespace: the namespace of the secret that contains the Azure Storage
Account Name and Key. Default is the same as the Pod.
secretName: the name of the secret that contains the Azure Storage Account Name and
Key. Default is azure-storage-account-<accountName>-secret
readOnly: a flag indicating whether the storage will be mounted as read only.
Defaults to false which means a read/write mount. This setting will impact the
ReadOnly setting in VolumeMounts as well.
During storage provisioning, a secret named by secretName is created for the
mounting credentials. If the cluster has enabled both
RBAC and
Controller Roles,
add the create permission of resource secret for clusterrole
system:controller:persistent-volume-binder.
In a multi-tenancy context, it is strongly recommended to set the value for
secretNamespace explicitly, otherwise the storage account credentials may
be read by other users.
fs: filesystem to be laid out: none/xfs/ext4 (default: ext4).
block_size: block size in Kbytes (default: 32).
repl: number of synchronous replicas to be provided in the form of
replication factor 1..3 (default: 1) A string is expected here i.e.
"1" and not 1.
priority_io: determines whether the volume will be created from higher
performance or a lower priority storage high/medium/low (default: low).
snap_interval: clock/time interval in minutes for when to trigger snapshots.
Snapshots are incremental based on difference with the prior snapshot, 0
disables snaps (default: 0). A string is expected here i.e.
"70" and not 70.
aggregation_level: specifies the number of chunks the volume would be
distributed into, 0 indicates a non-aggregated volume (default: 0). A string
is expected here i.e. "0" and not 0
ephemeral: specifies whether the volume should be cleaned-up after unmount
or should be persistent. emptyDir use case can set this value to true and
persistent volumes use case such as for databases like Cassandra should set
to false, true/false (default false). A string is expected here i.e.
"true" and not true.
Local volumes do not currently support dynamic provisioning, however a StorageClass
should still be created to delay volume binding until Pod scheduling. This is
specified by the WaitForFirstConsumer volume binding mode.
Delaying volume binding allows the scheduler to consider all of a Pod's
scheduling constraints when choosing an appropriate PersistentVolume for a
PersistentVolumeClaim.
3.7.6 - Dynamic Volume Provisioning
Dynamic volume provisioning allows storage volumes to be created on-demand.
Without dynamic provisioning, cluster administrators have to manually make
calls to their cloud or storage provider to create new storage volumes, and
then create PersistentVolume objects
to represent them in Kubernetes. The dynamic provisioning feature eliminates
the need for cluster administrators to pre-provision storage. Instead, it
automatically provisions storage when it is requested by users.
Background
The implementation of dynamic volume provisioning is based on the API object StorageClass
from the API group storage.k8s.io. A cluster administrator can define as many
StorageClass objects as needed, each specifying a volume plugin (aka
provisioner) that provisions a volume and the set of parameters to pass to
that provisioner when provisioning.
A cluster administrator can define and expose multiple flavors of storage (from
the same or different storage systems) within a cluster, each with a custom set
of parameters. This design also ensures that end users don't have to worry
about the complexity and nuances of how storage is provisioned, but still
have the ability to select from multiple storage options.
More information on storage classes can be found
here.
Enabling Dynamic Provisioning
To enable dynamic provisioning, a cluster administrator needs to pre-create
one or more StorageClass objects for users.
StorageClass objects define which provisioner should be used and what parameters
should be passed to that provisioner when dynamic provisioning is invoked.
The name of a StorageClass object must be a valid
DNS subdomain name.
The following manifest creates a storage class "slow" which provisions standard
disk-like persistent disks.
Users request dynamically provisioned storage by including a storage class in
their PersistentVolumeClaim. Before Kubernetes v1.6, this was done via the
volume.beta.kubernetes.io/storage-class annotation. However, this annotation
is deprecated since v1.9. Users now can and should instead use the
storageClassName field of the PersistentVolumeClaim object. The value of
this field must match the name of a StorageClass configured by the
administrator (see below).
To select the "fast" storage class, for example, a user would create the
following PersistentVolumeClaim:
This claim results in an SSD-like Persistent Disk being automatically
provisioned. When the claim is deleted, the volume is destroyed.
Defaulting Behavior
Dynamic provisioning can be enabled on a cluster such that all claims are
dynamically provisioned if no storage class is specified. A cluster administrator
can enable this behavior by:
An administrator can mark a specific StorageClass as default by adding the
storageclass.kubernetes.io/is-default-classannotation to it.
When a default StorageClass exists in a cluster and a user creates a
PersistentVolumeClaim with storageClassName unspecified, the
DefaultStorageClass admission controller automatically adds the
storageClassName field pointing to the default storage class.
Note that there can be at most one default storage class on a cluster, or
a PersistentVolumeClaim without storageClassName explicitly specified cannot
be created.
Topology Awareness
In Multi-Zone clusters, Pods can be spread across
Zones in a Region. Single-Zone storage backends should be provisioned in the Zones where
Pods are scheduled. This can be accomplished by setting the Volume Binding
Mode.
3.7.7 - Volume Snapshots
In Kubernetes, a VolumeSnapshot represents a snapshot of a volume on a storage system. This document assumes that you are already familiar with Kubernetes persistent volumes.
Introduction
Similar to how API resources PersistentVolume and PersistentVolumeClaim are used to provision volumes for users and administrators, VolumeSnapshotContent and VolumeSnapshot API resources are provided to create volume snapshots for users and administrators.
A VolumeSnapshotContent is a snapshot taken from a volume in the cluster that has been provisioned by an administrator. It is a resource in the cluster just like a PersistentVolume is a cluster resource.
A VolumeSnapshot is a request for snapshot of a volume by a user. It is similar to a PersistentVolumeClaim.
VolumeSnapshotClass allows you to specify different attributes belonging to a VolumeSnapshot. These attributes may differ among snapshots taken from the same volume on the storage system and therefore cannot be expressed by using the same StorageClass of a PersistentVolumeClaim.
Volume snapshots provide Kubernetes users with a standardized way to copy a volume's contents at a particular point in time without creating an entirely new volume. This functionality enables, for example, database administrators to backup databases before performing edit or delete modifications.
Users need to be aware of the following when using this feature:
API Objects VolumeSnapshot, VolumeSnapshotContent, and VolumeSnapshotClass are CRDs, not part of the core API.
VolumeSnapshot support is only available for CSI drivers.
As part of the deployment process of VolumeSnapshot, the Kubernetes team provides a snapshot controller to be deployed into the control plane, and a sidecar helper container called csi-snapshotter to be deployed together with the CSI driver. The snapshot controller watches VolumeSnapshot and VolumeSnapshotContent objects and is responsible for the creation and deletion of VolumeSnapshotContent object. The sidecar csi-snapshotter watches VolumeSnapshotContent objects and triggers CreateSnapshot and DeleteSnapshot operations against a CSI endpoint.
There is also a validating webhook server which provides tightened validation on snapshot objects. This should be installed by the Kubernetes distros along with the snapshot controller and CRDs, not CSI drivers. It should be installed in all Kubernetes clusters that has the snapshot feature enabled.
CSI drivers may or may not have implemented the volume snapshot functionality. The CSI drivers that have provided support for volume snapshot will likely use the csi-snapshotter. See CSI Driver documentation for details.
The CRDs and snapshot controller installations are the responsibility of the Kubernetes distribution.
Lifecycle of a volume snapshot and volume snapshot content
VolumeSnapshotContents are resources in the cluster. VolumeSnapshots are requests for those resources. The interaction between VolumeSnapshotContents and VolumeSnapshots follow this lifecycle:
Provisioning Volume Snapshot
There are two ways snapshots may be provisioned: pre-provisioned or dynamically provisioned.
Pre-provisioned
A cluster administrator creates a number of VolumeSnapshotContents. They carry the details of the real volume snapshot on the storage system which is available for use by cluster users. They exist in the Kubernetes API and are available for consumption.
Dynamic
Instead of using a pre-existing snapshot, you can request that a snapshot to be dynamically taken from a PersistentVolumeClaim. The VolumeSnapshotClass specifies storage provider-specific parameters to use when taking a snapshot.
Binding
The snapshot controller handles the binding of a VolumeSnapshot object with an appropriate VolumeSnapshotContent object, in both pre-provisioned and dynamically provisioned scenarios. The binding is a one-to-one mapping.
In the case of pre-provisioned binding, the VolumeSnapshot will remain unbound until the requested VolumeSnapshotContent object is created.
Persistent Volume Claim as Snapshot Source Protection
The purpose of this protection is to ensure that in-use
PersistentVolumeClaim
API objects are not removed from the system while a snapshot is being taken from it (as this may result in data loss).
While a snapshot is being taken of a PersistentVolumeClaim, that PersistentVolumeClaim is in-use. If you delete a PersistentVolumeClaim API object in active use as a snapshot source, the PersistentVolumeClaim object is not removed immediately. Instead, removal of the PersistentVolumeClaim object is postponed until the snapshot is readyToUse or aborted.
Delete
Deletion is triggered by deleting the VolumeSnapshot object, and the DeletionPolicy will be followed. If the DeletionPolicy is Delete, then the underlying storage snapshot will be deleted along with the VolumeSnapshotContent object. If the DeletionPolicy is Retain, then both the underlying snapshot and VolumeSnapshotContent remain.
persistentVolumeClaimName is the name of the PersistentVolumeClaim data source for the snapshot. This field is required for dynamically provisioning a snapshot.
A volume snapshot can request a particular class by specifying the name of a
VolumeSnapshotClass
using the attribute volumeSnapshotClassName. If nothing is set, then the default class is used if available.
For pre-provisioned snapshots, you need to specify a volumeSnapshotContentName as the source for the snapshot as shown in the following example. The volumeSnapshotContentName source field is required for pre-provisioned snapshots.
Each VolumeSnapshotContent contains a spec and status. In dynamic provisioning, the snapshot common controller creates VolumeSnapshotContent objects. Here is an example:
volumeHandle is the unique identifier of the volume created on the storage backend and returned by the CSI driver during the volume creation. This field is required for dynamically provisioning a snapshot. It specifies the volume source of the snapshot.
For pre-provisioned snapshots, you (as cluster administrator) are responsible for creating the VolumeSnapshotContent object as follows.
snapshotHandle is the unique identifier of the volume snapshot created on the storage backend. This field is required for the pre-provisioned snapshots. It specifies the CSI snapshot id on the storage system that this VolumeSnapshotContent represents.
sourceVolumeMode is the mode of the volume whose snapshot is taken. The value
of the sourceVolumeMode field can be either Filesystem or Block. If the
source volume mode is not specified, Kubernetes treats the snapshot as if the
source volume's mode is unknown.
Converting the volume mode of a Snapshot
If the VolumeSnapshots API installed on your cluster supports the sourceVolumeMode
field, then the API has the capability to prevent unauthorized users from converting
the mode of a volume.
To check if your cluster has capability for this feature, run the following command:
$ kubectl get crd volumesnapshotcontent -o yaml
If you want to allow users to create a PersistentVolumeClaim from an existing
VolumeSnapshot, but with a different volume mode than the source, the annotation
snapshot.storage.kubernetes.io/allowVolumeModeChange: "true"needs to be added to
the VolumeSnapshotContent that corresponds to the VolumeSnapshot.
For pre-provisioned snapshots, Spec.SourceVolumeMode needs to be populated
by the cluster administrator.
An example VolumeSnapshotContent resource with this feature enabled would look like:
This document describes the concept of VolumeSnapshotClass in Kubernetes. Familiarity
with volume snapshots and
storage classes is suggested.
Introduction
Just like StorageClass provides a way for administrators to describe the "classes"
of storage they offer when provisioning a volume, VolumeSnapshotClass provides a
way to describe the "classes" of storage when provisioning a volume snapshot.
The VolumeSnapshotClass Resource
Each VolumeSnapshotClass contains the fields driver, deletionPolicy, and parameters,
which are used when a VolumeSnapshot belonging to the class needs to be
dynamically provisioned.
The name of a VolumeSnapshotClass object is significant, and is how users can
request a particular class. Administrators set the name and other parameters
of a class when first creating VolumeSnapshotClass objects, and the objects cannot
be updated once they are created.
Note: Installation of the CRDs is the responsibility of the Kubernetes distribution. Without the required CRDs present, the creation of a VolumeSnapshotClass fails.
Administrators can specify a default VolumeSnapshotClass for VolumeSnapshots
that don't request any particular class to bind to by adding the
snapshot.storage.kubernetes.io/is-default-class: "true" annotation:
Volume snapshot classes have a driver that determines what CSI volume plugin is
used for provisioning VolumeSnapshots. This field must be specified.
DeletionPolicy
Volume snapshot classes have a deletionPolicy. It enables you to configure what happens to a VolumeSnapshotContent when the VolumeSnapshot object it is bound to is to be deleted. The deletionPolicy of a volume snapshot class can either be Retain or Delete. This field must be specified.
If the deletionPolicy is Delete, then the underlying storage snapshot will be deleted along with the VolumeSnapshotContent object. If the deletionPolicy is Retain, then both the underlying snapshot and VolumeSnapshotContent remain.
Parameters
Volume snapshot classes have parameters that describe volume snapshots belonging to
the volume snapshot class. Different parameters may be accepted depending on the
driver.
3.7.9 - CSI Volume Cloning
This document describes the concept of cloning existing CSI Volumes in Kubernetes. Familiarity with Volumes is suggested.
Introduction
The CSI Volume Cloning feature adds support for specifying existing PVCs in the dataSource field to indicate a user would like to clone a Volume.
A Clone is defined as a duplicate of an existing Kubernetes Volume that can be consumed as any standard Volume would be. The only difference is that upon provisioning, rather than creating a "new" empty Volume, the back end device creates an exact duplicate of the specified Volume.
The implementation of cloning, from the perspective of the Kubernetes API, adds the ability to specify an existing PVC as a dataSource during new PVC creation. The source PVC must be bound and available (not in use).
Users need to be aware of the following when using this feature:
Cloning support (VolumePVCDataSource) is only available for CSI drivers.
Cloning support is only available for dynamic provisioners.
CSI drivers may or may not have implemented the volume cloning functionality.
You can only clone a PVC when it exists in the same namespace as the destination PVC (source and destination must be in the same namespace).
Cloning is supported with a different Storage Class.
Destination volume can be the same or a different storage class as the source.
Default storage class can be used and storageClassName omitted in the spec.
Cloning can only be performed between two volumes that use the same VolumeMode setting (if you request a block mode volume, the source MUST also be block mode)
Provisioning
Clones are provisioned like any other PVC with the exception of adding a dataSource that references an existing PVC in the same namespace.
Note: You must specify a capacity value for spec.resources.requests.storage, and the value you specify must be the same or larger than the capacity of the source volume.
The result is a new PVC with the name clone-of-pvc-1 that has the exact same content as the specified source pvc-1.
Usage
Upon availability of the new PVC, the cloned PVC is consumed the same as other PVC. It's also expected at this point that the newly created PVC is an independent object. It can be consumed, cloned, snapshotted, or deleted independently and without consideration for it's original dataSource PVC. This also implies that the source is not linked in any way to the newly created clone, it may also be modified or deleted without affecting the newly created clone.
3.7.10 - Storage Capacity
Storage capacity is limited and may vary depending on the node on
which a pod runs: network-attached storage might not be accessible by
all nodes, or storage is local to a node to begin with.
FEATURE STATE:Kubernetes v1.24 [stable]
This page describes how Kubernetes keeps track of storage capacity and
how the scheduler uses that information to schedule Pods onto nodes
that have access to enough storage capacity for the remaining missing
volumes. Without storage capacity tracking, the scheduler may choose a
node that doesn't have enough capacity to provision a volume and
multiple scheduling retries will be needed.
Before you begin
Kubernetes v1.25 includes cluster-level API support for
storage capacity tracking. To use this you must also be using a CSI driver that
supports capacity tracking. Consult the documentation for the CSI drivers that
you use to find out whether this support is available and, if so, how to use
it. If you are not running Kubernetes v1.25, check the
documentation for that version of Kubernetes.
API
There are two API extensions for this feature:
CSIStorageCapacity objects:
these get produced by a CSI driver in the namespace
where the driver is installed. Each object contains capacity
information for one storage class and defines which nodes have
access to that storage.
the CSIDriver object for the driver has StorageCapacity set to
true.
In that case, the scheduler only considers nodes for the Pod which
have enough storage available to them. This check is very
simplistic and only compares the size of the volume against the
capacity listed in CSIStorageCapacity objects with a topology that
includes the node.
For volumes with Immediate volume binding mode, the storage driver
decides where to create the volume, independently of Pods that will
use the volume. The scheduler then schedules Pods onto nodes where the
volume is available after the volume has been created.
For CSI ephemeral volumes,
scheduling always happens without considering storage capacity. This
is based on the assumption that this volume type is only used by
special CSI drivers which are local to a node and do not need
significant resources there.
Rescheduling
When a node has been selected for a Pod with WaitForFirstConsumer
volumes, that decision is still tentative. The next step is that the
CSI storage driver gets asked to create the volume with a hint that the
volume is supposed to be available on the selected node.
Because Kubernetes might have chosen a node based on out-dated
capacity information, it is possible that the volume cannot really be
created. The node selection is then reset and the Kubernetes scheduler
tries again to find a node for the Pod.
Limitations
Storage capacity tracking increases the chance that scheduling works
on the first try, but cannot guarantee this because the scheduler has
to decide based on potentially out-dated information. Usually, the
same retry mechanism as for scheduling without any storage capacity
information handles scheduling failures.
One situation where scheduling can fail permanently is when a Pod uses
multiple volumes: one volume might have been created already in a
topology segment which then does not have enough capacity left for
another volume. Manual intervention is necessary to recover from this,
for example by increasing capacity or deleting the volume that was
already created.
This page describes the maximum number of volumes that can be attached
to a Node for various cloud providers.
Cloud providers like Google, Amazon, and Microsoft typically have a limit on
how many volumes can be attached to a Node. It is important for Kubernetes to
respect those limits. Otherwise, Pods scheduled on a Node could get stuck
waiting for volumes to attach.
Kubernetes default limits
The Kubernetes scheduler has default limits on the number of volumes
that can be attached to a Node:
You can change these limits by setting the value of the
KUBE_MAX_PD_VOLS environment variable, and then starting the scheduler.
CSI drivers might have a different procedure, see their documentation
on how to customize their limits.
Use caution if you set a limit that is higher than the default limit. Consult
the cloud provider's documentation to make sure that Nodes can actually support
the limit you set.
The limit applies to the entire cluster, so it affects all Nodes.
Dynamic volume limits
FEATURE STATE:Kubernetes v1.17 [stable]
Dynamic volume limits are supported for following volume types.
Amazon EBS
Google Persistent Disk
Azure Disk
CSI
For volumes managed by in-tree volume plugins, Kubernetes automatically determines the Node
type and enforces the appropriate maximum number of volumes for the node. For example:
For Amazon EBS disks on M5,C5,R5,T3 and Z1D instance types, Kubernetes allows only 25
volumes to be attached to a Node. For other instance types on
Amazon Elastic Compute Cloud (EC2),
Kubernetes allows 39 volumes to be attached to a Node.
On Azure, up to 64 disks can be attached to a node, depending on the node type. For more details, refer to Sizes for virtual machines in Azure.
If a CSI storage driver advertises a maximum number of volumes for a Node (using NodeGetInfo), the kube-scheduler honors that limit.
Refer to the CSI specifications for details.
For volumes managed by in-tree plugins that have been migrated to a CSI driver, the maximum number of volumes will be the one reported by the CSI driver.
3.7.12 - Volume Health Monitoring
FEATURE STATE:Kubernetes v1.21 [alpha]
CSI volume health monitoring allows CSI Drivers to detect abnormal volume conditions from the underlying storage systems and report them as events on PVCs or Pods.
Volume health monitoring
Kubernetes volume health monitoring is part of how Kubernetes implements the Container Storage Interface (CSI). Volume health monitoring feature is implemented in two components: an External Health Monitor controller, and the kubelet.
If a CSI Driver supports Volume Health Monitoring feature from the controller side, an event will be reported on the related PersistentVolumeClaim (PVC) when an abnormal volume condition is detected on a CSI volume.
The External Health Monitor controller also watches for node failure events. You can enable node failure monitoring by setting the enable-node-watcher flag to true. When the external health monitor detects a node failure event, the controller reports an Event will be reported on the PVC to indicate that pods using this PVC are on a failed node.
If a CSI Driver supports Volume Health Monitoring feature from the node side, an Event will be reported on every Pod using the PVC when an abnormal volume condition is detected on a CSI volume. In addition, Volume Health information is exposed as Kubelet VolumeStats metrics. A new metric kubelet_volume_stats_health_status_abnormal is added. This metric includes two labels: namespace and persistentvolumeclaim. The count is either 1 or 0. 1 indicates the volume is unhealthy, 0 indicates volume is healthy. For more information, please check KEP.
Note: You need to enable the CSIVolumeHealthfeature gate to use this feature from the node side.
This page provides an storage overview specific to the Windows operating system.
Persistent storage
Windows has a layered filesystem driver to mount container layers and create a copy
filesystem based on NTFS. All file paths in the container are resolved only within
the context of that container.
With Docker, volume mounts can only target a directory in the container, and not
an individual file. This limitation does not apply to containerd.
Volume mounts cannot project files or directories back to the host filesystem.
Read-only filesystems are not supported because write access is always required
for the Windows registry and SAM database. However, read-only volumes are supported.
Volume user-masks and permissions are not available. Because the SAM is not shared
between the host & container, there's no mapping between them. All permissions are
resolved within the context of the container.
As a result, the following storage functionality is not supported on Windows nodes:
Volume subpath mounts: only the entire volume can be mounted in a Windows container
Subpath volume mounting for Secrets
Host mount projection
Read-only root filesystem (mapped volumes still support readOnly)
Block device mapping
Memory as the storage medium (for example, emptyDir.medium set to Memory)
File system features like uid/gid; per-user Linux filesystem permissions
Kubernetes volumes enable complex
applications, with data persistence and Pod volume sharing requirements, to be deployed
on Kubernetes. Management of persistent volumes associated with a specific storage
back-end or protocol includes actions such as provisioning/de-provisioning/resizing
of volumes, attaching/detaching a volume to/from a Kubernetes node and
mounting/dismounting a volume to/from individual containers in a pod that needs to
persist data.
Volume management components are shipped as Kubernetes volume
plugin.
The following broad classes of Kubernetes volume plugins are supported on Windows:
Resources that Kubernetes provides for configuring Pods.
3.8.1 - Configuration Best Practices
This document highlights and consolidates configuration best practices that are introduced throughout the user guide, Getting Started documentation, and examples.
This is a living document. If you think of something that is not on this list but might be useful to others, please don't hesitate to file an issue or submit a PR.
General Configuration Tips
When defining configurations, specify the latest stable API version.
Configuration files should be stored in version control before being pushed to the cluster. This allows you to quickly roll back a configuration change if necessary. It also aids cluster re-creation and restoration.
Write your configuration files using YAML rather than JSON. Though these formats can be used interchangeably in almost all scenarios, YAML tends to be more user-friendly.
Group related objects into a single file whenever it makes sense. One file is often easier to manage than several. See the guestbook-all-in-one.yaml file as an example of this syntax.
Note also that many kubectl commands can be called on a directory. For example, you can call kubectl apply on a directory of config files.
Don't specify default values unnecessarily: simple, minimal configuration will make errors less likely.
Put object descriptions in annotations, to allow better introspection.
"Naked" Pods versus ReplicaSets, Deployments, and Jobs
Don't use naked Pods (that is, Pods not bound to a ReplicaSet or Deployment) if you can avoid it. Naked Pods will not be rescheduled in the event of a node failure.
A Deployment, which both creates a ReplicaSet to ensure that the desired number of Pods is always available, and specifies a strategy to replace Pods (such as RollingUpdate), is almost always preferable to creating Pods directly, except for some explicit restartPolicy: Never scenarios. A Job may also be appropriate.
Services
Create a Service before its corresponding backend workloads (Deployments or ReplicaSets), and before any workloads that need to access it. When Kubernetes starts a container, it provides environment variables pointing to all the Services which were running when the container was started. For example, if a Service named foo exists, all containers will get the following variables in their initial environment:
FOO_SERVICE_HOST=<the host the Service is running on>
FOO_SERVICE_PORT=<the port the Service is running on>
This does imply an ordering requirement - any Service that a Pod wants to access must be created before the Pod itself, or else the environment variables will not be populated. DNS does not have this restriction.
An optional (though strongly recommended) cluster add-on is a DNS server. The
DNS server watches the Kubernetes API for new Services and creates a set of DNS records for each. If DNS has been enabled throughout the cluster then all Pods should be able to do name resolution of Services automatically.
Don't specify a hostPort for a Pod unless it is absolutely necessary. When you bind a Pod to a hostPort, it limits the number of places the Pod can be scheduled, because each <hostIP, hostPort, protocol> combination must be unique. If you don't specify the hostIP and protocol explicitly, Kubernetes will use 0.0.0.0 as the default hostIP and TCP as the default protocol.
If you explicitly need to expose a Pod's port on the node, consider using a NodePort Service before resorting to hostPort.
Avoid using hostNetwork, for the same reasons as hostPort.
Use headless Services (which have a ClusterIP of None) for service discovery when you don't need kube-proxy load balancing.
Using Labels
Define and use labels that identify semantic attributes of your application or Deployment, such as { app.kubernetes.io/name: MyApp, tier: frontend, phase: test, deployment: v3 }. You can use these labels to select the appropriate Pods for other resources; for example, a Service that selects all tier: frontend Pods, or all phase: test components of app.kubernetes.io/name: MyApp. See the guestbook app for examples of this approach.
A Service can be made to span multiple Deployments by omitting release-specific labels from its selector. When you need to update a running service without downtime, use a Deployment.
A desired state of an object is described by a Deployment, and if changes to that spec are applied, the deployment controller changes the actual state to the desired state at a controlled rate.
Use the Kubernetes common labels for common use cases. These standardized labels enrich the metadata in a way that allows tools, including kubectl and dashboard, to work in an interoperable way.
You can manipulate labels for debugging. Because Kubernetes controllers (such as ReplicaSet) and Services match to Pods using selector labels, removing the relevant labels from a Pod will stop it from being considered by a controller or from being served traffic by a Service. If you remove the labels of an existing Pod, its controller will create a new Pod to take its place. This is a useful way to debug a previously "live" Pod in a "quarantine" environment. To interactively remove or add labels, use kubectl label.
Using kubectl
Use kubectl apply -f <directory>. This looks for Kubernetes configuration in all .yaml, .yml, and .json files in <directory> and passes it to apply.
A ConfigMap is an API object used to store non-confidential data in key-value pairs.
Pods can consume ConfigMaps as
environment variables, command-line arguments, or as configuration files in a
volume.
A ConfigMap allows you to decouple environment-specific configuration from your container images, so that your applications are easily portable.
Caution: ConfigMap does not provide secrecy or encryption.
If the data you want to store are confidential, use a
Secret rather than a ConfigMap,
or use additional (third party) tools to keep your data private.
Motivation
Use a ConfigMap for setting configuration data separately from application code.
For example, imagine that you are developing an application that you can run on your
own computer (for development) and in the cloud (to handle real traffic).
You write the code to look in an environment variable named DATABASE_HOST.
Locally, you set that variable to localhost. In the cloud, you set it to
refer to a Kubernetes Service
that exposes the database component to your cluster.
This lets you fetch a container image running in the cloud and
debug the exact same code locally if needed.
A ConfigMap is not designed to hold large chunks of data. The data stored in a
ConfigMap cannot exceed 1 MiB. If you need to store settings that are
larger than this limit, you may want to consider mounting a volume or use a
separate database or file service.
ConfigMap object
A ConfigMap is an API object
that lets you store configuration for other objects to use. Unlike most
Kubernetes objects that have a spec, a ConfigMap has data and binaryData
fields. These fields accept key-value pairs as their values. Both the data
field and the binaryData are optional. The data field is designed to
contain UTF-8 strings while the binaryData field is designed to
contain binary data as base64-encoded strings.
Each key under the data or the binaryData field must consist of
alphanumeric characters, -, _ or .. The keys stored in data must not
overlap with the keys in the binaryData field.
Starting from v1.19, you can add an immutable field to a ConfigMap
definition to create an immutable ConfigMap.
ConfigMaps and Pods
You can write a Pod spec that refers to a ConfigMap and configures the container(s)
in that Pod based on the data in the ConfigMap. The Pod and the ConfigMap must be in
the same namespace.
Note: The spec of a static Pod cannot refer to a ConfigMap
or any other API objects.
Here's an example ConfigMap that has some keys with single values,
and other keys where the value looks like a fragment of a configuration
format.
apiVersion:v1kind:ConfigMapmetadata:name:game-demodata:# property-like keys; each key maps to a simple valueplayer_initial_lives:"3"ui_properties_file_name:"user-interface.properties"# file-like keysgame.properties:| enemy.types=aliens,monsters
player.maximum-lives=5user-interface.properties:| color.good=purple
color.bad=yellow
allow.textmode=true
There are four different ways that you can use a ConfigMap to configure
a container inside a Pod:
Inside a container command and args
Environment variables for a container
Add a file in read-only volume, for the application to read
Write code to run inside the Pod that uses the Kubernetes API to read a ConfigMap
These different methods lend themselves to different ways of modeling
the data being consumed.
For the first three methods, the
kubelet uses the data from
the ConfigMap when it launches container(s) for a Pod.
The fourth method means you have to write code to read the ConfigMap and its data.
However, because you're using the Kubernetes API directly, your application can
subscribe to get updates whenever the ConfigMap changes, and react
when that happens. By accessing the Kubernetes API directly, this
technique also lets you access a ConfigMap in a different namespace.
Here's an example Pod that uses values from game-demo to configure a Pod:
apiVersion:v1kind:Podmetadata:name:configmap-demo-podspec:containers:- name:demoimage:alpinecommand:["sleep","3600"]env:# Define the environment variable- name:PLAYER_INITIAL_LIVES# Notice that the case is different here# from the key name in the ConfigMap.valueFrom:configMapKeyRef:name:game-demo # The ConfigMap this value comes from.key:player_initial_lives# The key to fetch.- name:UI_PROPERTIES_FILE_NAMEvalueFrom:configMapKeyRef:name:game-demokey:ui_properties_file_namevolumeMounts:- name:configmountPath:"/config"readOnly:truevolumes:# You set volumes at the Pod level, then mount them into containers inside that Pod- name:configconfigMap:# Provide the name of the ConfigMap you want to mount.name:game-demo# An array of keys from the ConfigMap to create as filesitems:- key:"game.properties"path:"game.properties"- key:"user-interface.properties"path:"user-interface.properties"
A ConfigMap doesn't differentiate between single line property values and
multi-line file-like values.
What matters is how Pods and other objects consume those values.
For this example, defining a volume and mounting it inside the demo
container as /config creates two files,
/config/game.properties and /config/user-interface.properties,
even though there are four keys in the ConfigMap. This is because the Pod
definition specifies an items array in the volumes section.
If you omit the items array entirely, every key in the ConfigMap becomes
a file with the same name as the key, and you get 4 files.
Using ConfigMaps
ConfigMaps can be mounted as data volumes. ConfigMaps can also be used by other
parts of the system, without being directly exposed to the Pod. For example,
ConfigMaps can hold data that other parts of the system should use for configuration.
The most common way to use ConfigMaps is to configure settings for
containers running in a Pod in the same namespace. You can also use a
ConfigMap separately.
For example, you
might encounter addons
or operators that
adjust their behavior based on a ConfigMap.
Using ConfigMaps as files from a Pod
To consume a ConfigMap in a volume in a Pod:
Create a ConfigMap or use an existing one. Multiple Pods can reference the
same ConfigMap.
Modify your Pod definition to add a volume under .spec.volumes[]. Name
the volume anything, and have a .spec.volumes[].configMap.name field set
to reference your ConfigMap object.
Add a .spec.containers[].volumeMounts[] to each container that needs the
ConfigMap. Specify .spec.containers[].volumeMounts[].readOnly = true and
.spec.containers[].volumeMounts[].mountPath to an unused directory name
where you would like the ConfigMap to appear.
Modify your image or command line so that the program looks for files in
that directory. Each key in the ConfigMap data map becomes the filename
under mountPath.
This is an example of a Pod that mounts a ConfigMap in a volume:
Each ConfigMap you want to use needs to be referred to in .spec.volumes.
If there are multiple containers in the Pod, then each container needs its
own volumeMounts block, but only one .spec.volumes is needed per ConfigMap.
Mounted ConfigMaps are updated automatically
When a ConfigMap currently consumed in a volume is updated, projected keys are eventually updated as well.
The kubelet checks whether the mounted ConfigMap is fresh on every periodic sync.
However, the kubelet uses its local cache for getting the current value of the ConfigMap.
The type of the cache is configurable using the ConfigMapAndSecretChangeDetectionStrategy field in
the KubeletConfiguration struct.
A ConfigMap can be either propagated by watch (default), ttl-based, or by redirecting
all requests directly to the API server.
As a result, the total delay from the moment when the ConfigMap is updated to the moment
when new keys are projected to the Pod can be as long as the kubelet sync period + cache
propagation delay, where the cache propagation delay depends on the chosen cache type
(it equals to watch propagation delay, ttl of cache, or zero correspondingly).
ConfigMaps consumed as environment variables are not updated automatically and require a pod restart.
Note: A container using a ConfigMap as a subPath volume mount will not receive ConfigMap updates.
Immutable ConfigMaps
FEATURE STATE:Kubernetes v1.21 [stable]
The Kubernetes feature Immutable Secrets and ConfigMaps provides an option to set
individual Secrets and ConfigMaps as immutable. For clusters that extensively use ConfigMaps
(at least tens of thousands of unique ConfigMap to Pod mounts), preventing changes to their
data has the following advantages:
protects you from accidental (or unwanted) updates that could cause applications outages
improves performance of your cluster by significantly reducing load on kube-apiserver, by
closing watches for ConfigMaps marked as immutable.
This feature is controlled by the ImmutableEphemeralVolumesfeature gate.
You can create an immutable ConfigMap by setting the immutable field to true.
For example:
Once a ConfigMap is marked as immutable, it is not possible to revert this change
nor to mutate the contents of the data or the binaryData field. You can
only delete and recreate the ConfigMap. Because existing Pods maintain a mount point
to the deleted ConfigMap, it is recommended to recreate these pods.
Read The Twelve-Factor App to understand the motivation for
separating code from configuration.
3.8.3 - Secrets
A Secret is an object that contains a small amount of sensitive data such as
a password, a token, or a key. Such information might otherwise be put in a
Pod specification or in a
container image. Using a
Secret means that you don't need to include confidential data in your
application code.
Because Secrets can be created independently of the Pods that use them, there
is less risk of the Secret (and its data) being exposed during the workflow of
creating, viewing, and editing Pods. Kubernetes, and applications that run in
your cluster, can also take additional precautions with Secrets, such as avoiding
writing secret data to nonvolatile storage.
Secrets are similar to ConfigMaps
but are specifically intended to hold confidential data.
Caution:
Kubernetes Secrets are, by default, stored unencrypted in the API server's underlying data store (etcd). Anyone with API access can retrieve or modify a Secret, and so can anyone with access to etcd.
Additionally, anyone who is authorized to create a Pod in a namespace can use that access to read any Secret in that namespace; this includes indirect access such as the ability to create a Deployment.
In order to safely use Secrets, take at least the following steps:
Enable or configure RBAC rules that
restrict reading and writing the Secret. Be aware that secrets can be obtained
implicitly by anyone with the permission to create a Pod.
Where appropriate, also use mechanisms such as RBAC to limit which principals are allowed
to create new Secrets or replace existing ones.
The Kubernetes control plane also uses Secrets; for example,
bootstrap token Secrets are a mechanism to
help automate node registration.
Alternatives to Secrets
Rather than using a Secret to protect confidential data, you can pick from alternatives.
Here are some of your options:
if your cloud-native component needs to authenticate to another application that you
know is running within the same Kubernetes cluster, you can use a
ServiceAccount
and its tokens to identify your client.
there are third-party tools that you can run, either within or outside your cluster,
that provide secrets management. For example, a service that Pods access over HTTPS,
that reveals a secret if the client correctly authenticates (for example, with a ServiceAccount
token).
for authentication, you can implement a custom signer for X.509 certificates, and use
CertificateSigningRequests
to let that custom signer issue certificates to Pods that need them.
you can use a device plugin
to expose node-local encryption hardware to a specific Pod. For example, you can schedule
trusted Pods onto nodes that provide a Trusted Platform Module, configured out-of-band.
You can also combine two or more of those options, including the option to use Secret objects themselves.
For example: implement (or deploy) an operator
that fetches short-lived session tokens from an external service, and then creates Secrets based
on those short-lived session tokens. Pods running in your cluster can make use of the session tokens,
and operator ensures they are valid. This separation means that you can run Pods that are unaware of
the exact mechanisms for issuing and refreshing those session tokens.
You can specify the data and/or the stringData field when creating a
configuration file for a Secret. The data and the stringData fields are optional.
The values for all keys in the data field have to be base64-encoded strings.
If the conversion to base64 string is not desirable, you can choose to specify
the stringData field instead, which accepts arbitrary strings as values.
The keys of data and stringData must consist of alphanumeric characters,
-, _ or .. All key-value pairs in the stringData field are internally
merged into the data field. If a key appears in both the data and the
stringData field, the value specified in the stringData field takes
precedence.
Size limit
Individual secrets are limited to 1MiB in size. This is to discourage creation
of very large secrets that could exhaust the API server and kubelet memory.
However, creation of many smaller secrets could also exhaust memory. You can
use a resource quota to limit the
number of Secrets (or other resources) in a namespace.
Editing a Secret
You can edit an existing Secret using kubectl:
kubectl edit secrets mysecret
This opens your default editor and allows you to update the base64 encoded Secret
values in the data field; for example:
# Please edit the object below. Lines beginning with a '#' will be ignored,# and an empty file will abort the edit. If an error occurs while saving this file, it will be# reopened with the relevant failures.#apiVersion:v1data:username:YWRtaW4=password:MWYyZDFlMmU2N2Rmkind:Secretmetadata:annotations:kubectl.kubernetes.io/last-applied-configuration:{... }creationTimestamp:2020-01-22T18:41:56Zname:mysecretnamespace:defaultresourceVersion:"164619"uid:cfee02d6-c137-11e5-8d73-42010af00002type:Opaque
That example manifest defines a Secret with two keys in the data field: username and password.
The values are Base64 strings in the manifest; however, when you use the Secret with a Pod
then the kubelet provides the decoded data to the Pod and its containers.
You can package many keys and values into one Secret, or use many Secrets, whichever is convenient.
Using a Secret
Secrets can be mounted as data volumes or exposed as
environment variables
to be used by a container in a Pod. Secrets can also be used by other parts of the
system, without being directly exposed to the Pod. For example, Secrets can hold
credentials that other parts of the system should use to interact with external
systems on your behalf.
Secret volume sources are validated to ensure that the specified object
reference actually points to an object of type Secret. Therefore, a Secret
needs to be created before any Pods that depend on it.
If the Secret cannot be fetched (perhaps because it does not exist, or
due to a temporary lack of connection to the API server) the kubelet
periodically retries running that Pod. The kubelet also reports an Event
for that Pod, including details of the problem fetching the Secret.
Optional Secrets
When you define a container environment variable based on a Secret,
you can mark it as optional. The default is for the Secret to be
required.
None of a Pod's containers will start until all non-optional Secrets are
available.
If a Pod references a specific key in a Secret and that Secret does exist, but
is missing the named key, the Pod fails during startup.
Using Secrets as files from a Pod
If you want to access data from a Secret in a Pod, one way to do that is to
have Kubernetes make the value of that Secret be available as a file inside
the filesystem of one or more of the Pod's containers.
To configure that, you:
Create a secret or use an existing one. Multiple Pods can reference the same secret.
Modify your Pod definition to add a volume under .spec.volumes[]. Name the volume anything, and have a .spec.volumes[].secret.secretName field equal to the name of the Secret object.
Add a .spec.containers[].volumeMounts[] to each container that needs the secret. Specify .spec.containers[].volumeMounts[].readOnly = true and .spec.containers[].volumeMounts[].mountPath to an unused directory name where you would like the secrets to appear.
Modify your image or command line so that the program looks for files in that directory. Each key in the secret data map becomes the filename under mountPath.
This is an example of a Pod that mounts a Secret named mysecret in a volume:
apiVersion:v1kind:Podmetadata:name:mypodspec:containers:- name:mypodimage:redisvolumeMounts:- name:foomountPath:"/etc/foo"readOnly:truevolumes:- name:foosecret:secretName:mysecretoptional:false# default setting; "mysecret" must exist
Each Secret you want to use needs to be referred to in .spec.volumes.
If there are multiple containers in the Pod, then each container needs its
own volumeMounts block, but only one .spec.volumes is needed per Secret.
Note:
Versions of Kubernetes before v1.22 automatically created credentials for accessing
the Kubernetes API. This older mechanism was based on creating token Secrets that
could then be mounted into running Pods.
In more recent versions, including Kubernetes v1.25, API credentials
are obtained directly by using the TokenRequest API,
and are mounted into Pods using a projected volume.
The tokens obtained using this method have bounded lifetimes, and are automatically
invalidated when the Pod they are mounted into is deleted.
You can still manually create
a service account token Secret; for example, if you need a token that never expires.
However, using the TokenRequest
subresource to obtain a token to access the API is recommended instead.
You can use the kubectl create token
command to obtain a token from the TokenRequest API.
Projection of Secret keys to specific paths
You can also control the paths within the volume where Secret keys are projected.
You can use the .spec.volumes[].secret.items field to change the target path of each key:
the username key from mysecret is available to the container at the path
/etc/foo/my-group/my-username instead of at /etc/foo/username.
the password key from that Secret object is not projected.
If .spec.volumes[].secret.items is used, only keys specified in items are projected.
To consume all keys from the Secret, all of them must be listed in the items field.
If you list keys explicitly, then all listed keys must exist in the corresponding Secret.
Otherwise, the volume is not created.
Secret files permissions
You can set the POSIX file access permission bits for a single Secret key.
If you don't specify any permissions, 0644 is used by default.
You can also set a default mode for the entire Secret volume and override per key if needed.
For example, you can specify a default mode like this:
The secret is mounted on /etc/foo; all the files created by the
secret volume mount have permission 0400.
Note: If you're defining a Pod or a Pod template using JSON, beware that the JSON
specification doesn't support octal notation. You can use the decimal value
for the defaultMode (for example, 0400 in octal is 256 in decimal) instead.
If you're writing YAML, you can write the defaultMode in octal.
Consuming Secret values from volumes
Inside the container that mounts a secret volume, the secret keys appear as
files. The secret values are base64 decoded and stored inside these files.
This is the result of commands executed inside the container from the example above:
ls /etc/foo/
The output is similar to:
username
password
cat /etc/foo/username
The output is similar to:
admin
cat /etc/foo/password
The output is similar to:
1f2d1e2e67df
The program in a container is responsible for reading the secret data from these
files, as needed.
Mounted Secrets are updated automatically
When a volume contains data from a Secret, and that Secret is updated, Kubernetes tracks
this and updates the data in the volume, using an eventually-consistent approach.
Note: A container using a Secret as a
subPath volume mount does not receive
automated Secret updates.
The kubelet keeps a cache of the current keys and values for the Secrets that are used in
volumes for pods on that node.
You can configure the way that the kubelet detects changes from the cached values. The configMapAndSecretChangeDetectionStrategy field in
the kubelet configuration controls which strategy the kubelet uses. The default strategy is Watch.
Updates to Secrets can be either propagated by an API watch mechanism (the default), based on
a cache with a defined time-to-live, or polled from the cluster API server on each kubelet
synchronisation loop.
As a result, the total delay from the moment when the Secret is updated to the moment
when new keys are projected to the Pod can be as long as the kubelet sync period + cache
propagation delay, where the cache propagation delay depends on the chosen cache type
(following the same order listed in the previous paragraph, these are:
watch propagation delay, the configured cache TTL, or zero for direct polling).
Create a Secret (or use an existing one). Multiple Pods can reference the same Secret.
Modify your Pod definition in each container that you wish to consume the value of a secret
key to add an environment variable for each secret key you wish to consume. The environment
variable that consumes the secret key should populate the secret's name and key in env[].valueFrom.secretKeyRef.
Modify your image and/or command line so that the program looks for values in the specified
environment variables.
This is an example of a Pod that uses a Secret via environment variables:
apiVersion:v1kind:Podmetadata:name:secret-env-podspec:containers:- name:mycontainerimage:redisenv:- name:SECRET_USERNAMEvalueFrom:secretKeyRef:name:mysecretkey:usernameoptional:false# same as default; "mysecret" must exist# and include a key named "username"- name:SECRET_PASSWORDvalueFrom:secretKeyRef:name:mysecretkey:passwordoptional:false# same as default; "mysecret" must exist# and include a key named "password"restartPolicy:Never
Invalid environment variables
Secrets used to populate environment variables by the envFrom field that have keys
that are considered invalid environment variable names will have those keys
skipped. The Pod is allowed to start.
If you define a Pod with an invalid variable name, the failed Pod startup includes
an event with the reason set to InvalidVariableNames and a message that lists the
skipped invalid keys. The following example shows a Pod that refers to a Secret
named mysecret, where mysecret contains 2 invalid keys: 1badkey and 2alsobad.
kubectl get events
The output is similar to:
LASTSEEN FIRSTSEEN COUNT NAME KIND SUBOBJECT TYPE REASON
0s 0s 1 dapi-test-pod Pod Warning InvalidEnvironmentVariableNames kubelet, 127.0.0.1 Keys [1badkey, 2alsobad] from the EnvFrom secret default/mysecret were skipped since they are considered invalid environment variable names.
Consuming Secret values from environment variables
Inside a container that consumes a Secret using environment variables, the secret keys appear
as normal environment variables. The values of those variables are the base64 decoded values
of the secret data.
This is the result of commands executed inside the container from the example above:
echo"$SECRET_USERNAME"
The output is similar to:
admin
echo"$SECRET_PASSWORD"
The output is similar to:
1f2d1e2e67df
Note: If a container already consumes a Secret in an environment variable,
a Secret update will not be seen by the container unless it is
restarted. There are third party solutions for triggering restarts when
secrets change.
Container image pull secrets
If you want to fetch container images from a private repository, you need a way for
the kubelet on each node to authenticate to that repository. You can configure
image pull secrets to make this possible. These secrets are configured at the Pod
level.
The imagePullSecrets field for a Pod is a list of references to Secrets in the same namespace
as the Pod.
You can use an imagePullSecrets to pass image registry access credentials to
the kubelet. The kubelet uses this information to pull a private image on behalf of your Pod.
See PodSpec in the Pod API reference
for more information about the imagePullSecrets field.
Using imagePullSecrets
The imagePullSecrets field is a list of references to secrets in the same namespace.
You can use an imagePullSecrets to pass a secret that contains a Docker (or other) image registry
password to the kubelet. The kubelet uses this information to pull a private image on behalf of your Pod.
See the PodSpec API for more information about the imagePullSecrets field.
Manually specifying an imagePullSecret
You can learn how to specify imagePullSecrets from the container images
documentation.
Arranging for imagePullSecrets to be automatically attached
You can manually create imagePullSecrets, and reference these from
a ServiceAccount. Any Pods created with that ServiceAccount
or created with that ServiceAccount by default, will get their imagePullSecrets
field set to that of the service account.
See Add ImagePullSecrets to a service account
for a detailed explanation of that process.
Using Secrets with static Pods
You cannot use ConfigMaps or Secrets with
static Pods.
Use envFrom to define all of the Secret's data as container environment variables. The key from the Secret becomes the environment variable name in the Pod.
You can also create a kustomization.yaml with a secretGenerator field containing ssh keys.
Caution:
Think carefully before sending your own SSH keys: other users of the cluster may have access
to the Secret.
You could instead create an SSH private key representing a service identity that you want to be
accessible to all the users with whom you share the Kubernetes cluster, and that you can revoke
if the credentials are compromised.
Now you can create a Pod which references the secret with the SSH key and
consumes it in a volume:
The container is then free to use the secret data to establish an SSH connection.
Use case: Pods with prod / test credentials
This example illustrates a Pod which consumes a secret containing production
credentials and another Pod which consumes a secret with test environment
credentials.
You can create a kustomization.yaml with a secretGenerator field or run
kubectl create secret.
Special characters such as $, \, *, =, and ! will be interpreted by your shell and require escaping.
In most shells, the easiest way to escape the password is to surround it with single quotes (').
For example, if your actual password is S!B\*d$zDsb=, you should execute the command this way:
You can make your data "hidden" by defining a key that begins with a dot.
This key represents a dotfile or "hidden" file. For example, when the following secret
is mounted into a volume, secret-volume:
The volume will contain a single file, called .secret-file, and
the dotfile-test-container will have this file present at the path
/etc/secret-volume/.secret-file.
Note: Files beginning with dot characters are hidden from the output of ls -l;
you must use ls -la to see them when listing directory contents.
Use case: Secret visible to one container in a Pod
Consider a program that needs to handle HTTP requests, do some complex business
logic, and then sign some messages with an HMAC. Because it has complex
application logic, there might be an unnoticed remote file reading exploit in
the server, which could expose the private key to an attacker.
This could be divided into two processes in two containers: a frontend container
which handles user interaction and business logic, but which cannot see the
private key; and a signer container that can see the private key, and responds
to simple signing requests from the frontend (for example, over localhost networking).
With this partitioned approach, an attacker now has to trick the application
server into doing something rather arbitrary, which may be harder than getting
it to read a file.
Types of Secret
When creating a Secret, you can specify its type using the type field of
the Secret
resource, or certain equivalent kubectl command line flags (if available).
The Secret type is used to facilitate programmatic handling of the Secret data.
Kubernetes provides several built-in types for some common usage scenarios.
These types vary in terms of the validations performed and the constraints
Kubernetes imposes on them.
Built-in Type
Usage
Opaque
arbitrary user-defined data
kubernetes.io/service-account-token
ServiceAccount token
kubernetes.io/dockercfg
serialized ~/.dockercfg file
kubernetes.io/dockerconfigjson
serialized ~/.docker/config.json file
kubernetes.io/basic-auth
credentials for basic authentication
kubernetes.io/ssh-auth
credentials for SSH authentication
kubernetes.io/tls
data for a TLS client or server
bootstrap.kubernetes.io/token
bootstrap token data
You can define and use your own Secret type by assigning a non-empty string as the
type value for a Secret object (an empty string is treated as an Opaque type).
Kubernetes doesn't impose any constraints on the type name. However, if you
are using one of the built-in types, you must meet all the requirements defined
for that type.
If you are defining a type of secret that's for public use, follow the convention
and structure the secret type to have your domain name before the name, separated
by a /. For example: cloud-hosting.example.net/cloud-api-credentials.
Opaque secrets
Opaque is the default Secret type if omitted from a Secret configuration file.
When you create a Secret using kubectl, you will use the generic
subcommand to indicate an Opaque Secret type. For example, the following
command creates an empty Secret of type Opaque.
kubectl create secret generic empty-secret
kubectl get secret empty-secret
The output looks like:
NAME TYPE DATA AGE
empty-secret Opaque 0 2m6s
The DATA column shows the number of data items stored in the Secret.
In this case, 0 means you have created an empty Secret.
Service account token Secrets
A kubernetes.io/service-account-token type of Secret is used to store a
token credential that identifies a
service account.
Since 1.22, this type of Secret is no longer used to mount credentials into Pods,
and obtaining tokens via the TokenRequest
API is recommended instead of using service account token Secret objects.
Tokens obtained from the TokenRequest API are more secure than ones stored in Secret objects,
because they have a bounded lifetime and are not readable by other API clients.
You can use the kubectl create token
command to obtain a token from the TokenRequest API.
You should only create a service account token Secret object
if you can't use the TokenRequest API to obtain a token,
and the security exposure of persisting a non-expiring token credential
in a readable API object is acceptable to you.
When using this Secret type, you need to ensure that the
kubernetes.io/service-account.name annotation is set to an existing
service account name. If you are creating both the ServiceAccount and
the Secret objects, you should create the ServiceAccount object first.
After the Secret is created, a Kubernetes controller
fills in some other fields such as the kubernetes.io/service-account.uid annotation, and the
token key in the data field, which is populated with an authentication token.
The following example configuration declares a service account token Secret:
apiVersion:v1kind:Secretmetadata:name:secret-sa-sampleannotations:kubernetes.io/service-account.name:"sa-name"type:kubernetes.io/service-account-tokendata:# You can include additional key value pairs as you do with Opaque Secretsextra:YmFyCg==
After creating the Secret, wait for Kubernetes to populate the token key in the data field.
See the ServiceAccount
documentation for more information on how service accounts work.
You can also check the automountServiceAccountToken field and the
serviceAccountName field of the
Pod
for information on referencing service account credentials from within Pods.
Docker config Secrets
You can use one of the following type values to create a Secret to
store the credentials for accessing a container image registry:
kubernetes.io/dockercfg
kubernetes.io/dockerconfigjson
The kubernetes.io/dockercfg type is reserved to store a serialized
~/.dockercfg which is the legacy format for configuring Docker command line.
When using this Secret type, you have to ensure the Secret data field
contains a .dockercfg key whose value is content of a ~/.dockercfg file
encoded in the base64 format.
The kubernetes.io/dockerconfigjson type is designed for storing a serialized
JSON that follows the same format rules as the ~/.docker/config.json file
which is a new format for ~/.dockercfg.
When using this Secret type, the data field of the Secret object must
contain a .dockerconfigjson key, in which the content for the
~/.docker/config.json file is provided as a base64 encoded string.
Below is an example for a kubernetes.io/dockercfg type of Secret:
Note: If you do not want to perform the base64 encoding, you can choose to use the
stringData field instead.
When you create these types of Secrets using a manifest, the API
server checks whether the expected key exists in the data field, and
it verifies if the value provided can be parsed as a valid JSON. The API
server doesn't validate if the JSON actually is a Docker config file.
When you do not have a Docker config file, or you want to use kubectl
to create a Secret for accessing a container registry, you can do:
That command creates a Secret of type kubernetes.io/dockerconfigjson.
If you dump the .data.dockerconfigjson field from that new Secret and then
decode it from base64:
kubectl get secret secret-tiger-docker -o jsonpath='{.data.*}' | base64 -d
then the output is equivalent to this JSON document (which is also a valid
Docker configuration file):
Note: The auth value there is base64 encoded; it is obscured but not secret.
Anyone who can read that Secret can learn the registry access bearer token.
Basic authentication Secret
The kubernetes.io/basic-auth type is provided for storing credentials needed
for basic authentication. When using this Secret type, the data field of the
Secret must contain one of the following two keys:
username: the user name for authentication
password: the password or token for authentication
Both values for the above two keys are base64 encoded strings. You can, of
course, provide the clear text content using the stringData for Secret
creation.
The following manifest is an example of a basic authentication Secret:
apiVersion:v1kind:Secretmetadata:name:secret-basic-authtype:kubernetes.io/basic-authstringData:username:admin # required field for kubernetes.io/basic-authpassword:t0p-Secret# required field for kubernetes.io/basic-auth
The basic authentication Secret type is provided only for convenience.
You can create an Opaque type for credentials used for basic authentication.
However, using the defined and public Secret type (kubernetes.io/basic-auth) helps other
people to understand the purpose of your Secret, and sets a convention for what key names
to expect.
The Kubernetes API verifies that the required keys are set for a Secret
of this type.
SSH authentication secrets
The builtin type kubernetes.io/ssh-auth is provided for storing data used in
SSH authentication. When using this Secret type, you will have to specify a
ssh-privatekey key-value pair in the data (or stringData) field
as the SSH credential to use.
The following manifest is an example of a Secret used for SSH public/private
key authentication:
apiVersion:v1kind:Secretmetadata:name:secret-ssh-authtype:kubernetes.io/ssh-authdata:# the data is abbreviated in this examplessh-privatekey:|MIIEpQIBAAKCAQEAulqb/Y ...
The SSH authentication Secret type is provided only for user's convenience.
You could instead create an Opaque type Secret for credentials used for SSH authentication.
However, using the defined and public Secret type (kubernetes.io/ssh-auth) helps other
people to understand the purpose of your Secret, and sets a convention for what key names
to expect.
and the API server does verify if the required keys are provided in a Secret
configuration.
Caution: SSH private keys do not establish trusted communication between an SSH client and
host server on their own. A secondary means of establishing trust is needed to
mitigate "man in the middle" attacks, such as a known_hosts file added to a
ConfigMap.
TLS secrets
Kubernetes provides a builtin Secret type kubernetes.io/tls for storing
a certificate and its associated key that are typically used for TLS.
One common use for TLS secrets is to configure encryption in transit for
an Ingress, but you can also use it
with other resources or directly in your workload.
When using this type of Secret, the tls.key and the tls.crt key must be provided
in the data (or stringData) field of the Secret configuration, although the API
server doesn't actually validate the values for each key.
The following YAML contains an example config for a TLS Secret:
apiVersion:v1kind:Secretmetadata:name:secret-tlstype:kubernetes.io/tlsdata:# the data is abbreviated in this exampletls.crt:|MIIC2DCCAcCgAwIBAgIBATANBgkqh ...tls.key:|MIIEpgIBAAKCAQEA7yn3bRHQ5FHMQ ...
The TLS Secret type is provided for user's convenience. You can create an Opaque
for credentials used for TLS server and/or client. However, using the builtin Secret
type helps ensure the consistency of Secret format in your project; the API server
does verify if the required keys are provided in a Secret configuration.
When creating a TLS Secret using kubectl, you can use the tls subcommand
as shown in the following example:
The public/private key pair must exist before hand. The public key certificate
for --cert must be DER format as per
Section 5.1 of RFC 7468,
and must match the given private key for --key (PKCS #8 in DER format;
Section 11 of RFC 7468).
Note:
A kubernetes.io/tls Secret stores the Base64-encoded DER data for keys and
certificates. If you're familiar with PEM format for private keys and for certificates,
the base64 data are the same as that format except that you omit
the initial and the last lines that are used in PEM.
For example, for a certificate, you do not include --------BEGIN CERTIFICATE-----
and -------END CERTIFICATE----.
Bootstrap token Secrets
A bootstrap token Secret can be created by explicitly specifying the Secret
type to bootstrap.kubernetes.io/token. This type of Secret is designed for
tokens used during the node bootstrap process. It stores tokens used to sign
well-known ConfigMaps.
A bootstrap token Secret is usually created in the kube-system namespace and
named in the form bootstrap-token-<token-id> where <token-id> is a 6 character
string of the token ID.
As a Kubernetes manifest, a bootstrap token Secret might look like the
following:
A bootstrap type Secret has the following keys specified under data:
token-id: A random 6 character string as the token identifier. Required.
token-secret: A random 16 character string as the actual token secret. Required.
description: A human-readable string that describes what the token is
used for. Optional.
expiration: An absolute UTC time using RFC3339 specifying when the token
should be expired. Optional.
usage-bootstrap-<usage>: A boolean flag indicating additional usage for
the bootstrap token.
auth-extra-groups: A comma-separated list of group names that will be
authenticated as in addition to the system:bootstrappers group.
The above YAML may look confusing because the values are all in base64 encoded
strings. In fact, you can create an identical Secret using the following YAML:
apiVersion:v1kind:Secretmetadata:# Note how the Secret is namedname:bootstrap-token-5emitj# A bootstrap token Secret usually resides in the kube-system namespacenamespace:kube-systemtype:bootstrap.kubernetes.io/tokenstringData:auth-extra-groups:"system:bootstrappers:kubeadm:default-node-token"expiration:"2020-09-13T04:39:10Z"# This token ID is used in the nametoken-id:"5emitj"token-secret:"kq4gihvszzgn1p0r"# This token can be used for authenticationusage-bootstrap-authentication:"true"# and it can be used for signingusage-bootstrap-signing:"true"
Immutable Secrets
FEATURE STATE:Kubernetes v1.21 [stable]
Kubernetes lets you mark specific Secrets (and ConfigMaps) as immutable.
Preventing changes to the data of an existing Secret has the following benefits:
protects you from accidental (or unwanted) updates that could cause applications outages
(for clusters that extensively use Secrets - at least tens of thousands of unique Secret
to Pod mounts), switching to immutable Secrets improves the performance of your cluster
by significantly reducing load on kube-apiserver. The kubelet does not need to maintain
a [watch] on any Secrets that are marked as immutable.
Marking a Secret as immutable
You can create an immutable Secret by setting the immutable field to true. For example,
You can also update any existing mutable Secret to make it immutable.
Note: Once a Secret or ConfigMap is marked as immutable, it is not possible to revert this change
nor to mutate the contents of the data field. You can only delete and recreate the Secret.
Existing Pods maintain a mount point to the deleted Secret - it is recommended to recreate
these pods.
Information security for Secrets
Although ConfigMap and Secret work similarly, Kubernetes applies some additional
protection for Secret objects.
Secrets often hold values that span a spectrum of importance, many of which can
cause escalations within Kubernetes (e.g. service account tokens) and to
external systems. Even if an individual app can reason about the power of the
Secrets it expects to interact with, other apps within the same namespace can
render those assumptions invalid.
A Secret is only sent to a node if a Pod on that node requires it.
For mounting secrets into Pods, the kubelet stores a copy of the data into a tmpfs
so that the confidential data is not written to durable storage.
Once the Pod that depends on the Secret is deleted, the kubelet deletes its local copy
of the confidential data from the Secret.
There may be several containers in a Pod. By default, containers you define
only have access to the default ServiceAccount and its related Secret.
You must explicitly define environment variables or map a volume into a
container in order to provide access to any other Secret.
There may be Secrets for several Pods on the same node. However, only the
Secrets that a Pod requests are potentially visible within its containers.
Therefore, one Pod does not have access to the Secrets of another Pod.
Warning: Any privileged containers on a node are liable to have access to all Secrets used
on that node.
Security recommendations for developers
Applications still need to protect the value of confidential information after reading it
from an environment variable or volume. For example, your application must avoid logging
the secret data in the clear or transmitting it to an untrusted party.
If you are defining multiple containers in a Pod, and only one of those
containers needs access to a Secret, define the volume mount or environment
variable configuration so that the other containers do not have access to that
Secret.
If you configure a Secret through a manifest,
with the secret data encoded as base64, sharing this file or checking it in to a
source repository means the secret is available to everyone who can read the manifest.
Base64 encoding is not an encryption method, it provides no additional confidentiality
over plain text.
When deploying applications that interact with the Secret API, you should
limit access using
authorization policies such as
RBAC.
In the Kubernetes API, watch and list requests for Secrets within a namespace
are extremely powerful capabilities. Avoid granting this access where feasible, since
listing Secrets allows the clients to inspect the values of every Secret in that
namespace.
Security recommendations for cluster administrators
Caution: A user who can create a Pod that uses a Secret can also see the value of that Secret. Even
if cluster policies do not allow a user to read the Secret directly, the same user could
have access to run a Pod that then exposes the Secret.
Reserve the ability to watch or list all secrets in a cluster (using the Kubernetes
API), so that only the most privileged, system-level components can perform this action.
When deploying applications that interact with the Secret API, you should
limit access using
authorization policies such as
RBAC.
In the API server, objects (including Secrets) are persisted into
etcd; therefore:
only allow cluster administrators to access etcd (this includes read-only access);
enable encryption at rest
for Secret objects, so that the data of these Secrets are not stored in the clear
into etcd;
consider wiping / shredding the durable storage used by etcd once it is
no longer in use;
if there are multiple etcd instances, make sure that etcd is
using SSL/TLS for communication between etcd peers.
3.8.4 - Resource Management for Pods and Containers
When you specify a Pod, you can optionally specify how
much of each resource a container needs.
The most common resources to specify are CPU and memory (RAM); there are others.
When you specify the resource request for containers in a Pod, the
kube-scheduler uses this
information to decide which node to place the Pod on. When you specify a resource limit
for a container, the kubelet enforces those limits so that the running container is not
allowed to use more of that resource than the limit you set. The kubelet also reserves
at least the request amount of that system resource specifically for that container
to use.
Requests and limits
If the node where a Pod is running has enough of a resource available, it's possible (and
allowed) for a container to use more resource than its request for that resource specifies.
However, a container is not allowed to use more than its resource limit.
For example, if you set a memory request of 256 MiB for a container, and that container is in
a Pod scheduled to a Node with 8GiB of memory and no other Pods, then the container can try to use
more RAM.
If you set a memory limit of 4GiB for that container, the kubelet (and
container runtime) enforce the limit.
The runtime prevents the container from using more than the configured resource limit. For example:
when a process in the container tries to consume more than the allowed amount of memory,
the system kernel terminates the process that attempted the allocation, with an out of memory
(OOM) error.
Limits can be implemented either reactively (the system intervenes once it sees a violation)
or by enforcement (the system prevents the container from ever exceeding the limit). Different
runtimes can have different ways to implement the same restrictions.
Note: If you specify a limit for a resource, but do not specify any request, and no admission-time
mechanism has applied a default request for that resource, then Kubernetes copies the limit
you specified and uses it as the requested value for the resource.
Resource types
CPU and memory are each a resource type. A resource type has a base unit.
CPU represents compute processing and is specified in units of Kubernetes CPUs.
Memory is specified in units of bytes.
For Linux workloads, you can specify huge page resources.
Huge pages are a Linux-specific feature where the node kernel allocates blocks of memory
that are much larger than the default page size.
For example, on a system where the default page size is 4KiB, you could specify a limit,
hugepages-2Mi: 80Mi. If the container tries allocating over 40 2MiB huge pages (a
total of 80 MiB), that allocation fails.
Note: You cannot overcommit hugepages-* resources.
This is different from the memory and cpu resources.
CPU and memory are collectively referred to as compute resources, or resources. Compute
resources are measurable quantities that can be requested, allocated, and
consumed. They are distinct from
API resources. API resources, such as Pods and
Services are objects that can be read and modified
through the Kubernetes API server.
Resource requests and limits of Pod and container
For each container, you can specify resource limits and requests,
including the following:
Although you can only specify requests and limits for individual containers,
it is also useful to think about the overall resource requests and limits for
a Pod.
For a particular resource, a Pod resource request/limit is the sum of the
resource requests/limits of that type for each container in the Pod.
Resource units in Kubernetes
CPU resource units
Limits and requests for CPU resources are measured in cpu units.
In Kubernetes, 1 CPU unit is equivalent to 1 physical CPU core,
or 1 virtual core, depending on whether the node is a physical host
or a virtual machine running inside a physical machine.
Fractional requests are allowed. When you define a container with
spec.containers[].resources.requests.cpu set to 0.5, you are requesting half
as much CPU time compared to if you asked for 1.0 CPU.
For CPU resource units, the quantity expression 0.1 is equivalent to the
expression 100m, which can be read as "one hundred millicpu". Some people say
"one hundred millicores", and this is understood to mean the same thing.
CPU resource is always specified as an absolute amount of resource, never as a relative amount. For example,
500m CPU represents the roughly same amount of computing power whether that container
runs on a single-core, dual-core, or 48-core machine.
Note: Kubernetes doesn't allow you to specify CPU resources with a precision finer than
1m. Because of this, it's useful to specify CPU units less than 1.0 or 1000m using
the milliCPU form; for example, 5m rather than 0.005.
Memory resource units
Limits and requests for memory are measured in bytes. You can express memory as
a plain integer or as a fixed-point number using one of these
quantity suffixes:
E, P, T, G, M, k. You can also use the power-of-two equivalents: Ei, Pi, Ti, Gi,
Mi, Ki. For example, the following represent roughly the same value:
128974848, 129e6, 129M, 128974848000m, 123Mi
Pay attention to the case of the suffixes. If you request 400m of memory, this is a request
for 0.4 bytes. Someone who types that probably meant to ask for 400 mebibytes (400Mi)
or 400 megabytes (400M).
Container resources example
The following Pod has two containers. Both containers are defined with a request for
0.25 CPU
and 64MiB (226 bytes) of memory. Each container has a limit of 0.5
CPU and 128MiB of memory. You can say the Pod has a request of 0.5 CPU and 128
MiB of memory, and a limit of 1 CPU and 256MiB of memory.
When you create a Pod, the Kubernetes scheduler selects a node for the Pod to
run on. Each node has a maximum capacity for each of the resource types: the
amount of CPU and memory it can provide for Pods. The scheduler ensures that,
for each resource type, the sum of the resource requests of the scheduled
containers is less than the capacity of the node.
Note that although actual memory
or CPU resource usage on nodes is very low, the scheduler still refuses to place
a Pod on a node if the capacity check fails. This protects against a resource
shortage on a node when resource usage later increases, for example, during a
daily peak in request rate.
How Kubernetes applies resource requests and limits
When the kubelet starts a container as part of a Pod, the kubelet passes that container's
requests and limits for memory and CPU to the container runtime.
On Linux, the container runtime typically configures
kernel cgroups that apply and enforce the
limits you defined.
The CPU limit defines a hard ceiling on how much CPU time that the container can use.
During each scheduling interval (time slice), the Linux kernel checks to see if this
limit is exceeded; if so, the kernel waits before allowing that cgroup to resume execution.
The CPU request typically defines a weighting. If several different containers (cgroups)
want to run on a contended system, workloads with larger CPU requests are allocated more
CPU time than workloads with small requests.
The memory request is mainly used during (Kubernetes) Pod scheduling. On a node that uses
cgroups v2, the container runtime might use the memory request as a hint to set
memory.min and memory.low.
The memory limit defines a memory limit for that cgroup. If the container tries to
allocate more memory than this limit, the Linux kernel out-of-memory subsystem activates
and, typically, intervenes by stopping one of the processes in the container that tried
to allocate memory. If that process is the container's PID 1, and the container is marked
as restartable, Kubernetes restarts the container.
The memory limit for the Pod or container can also apply to pages in memory backed
volumes, such as an emptyDir. The kubelet tracks tmpfs emptyDir volumes as container
memory use, rather than as local ephemeral storage.
If a container exceeds its memory request and the node that it runs on becomes short of
memory overall, it is likely that the Pod the container belongs to will be
evicted.
A container might or might not be allowed to exceed its CPU limit for extended periods of time.
However, container runtimes don't terminate Pods or containers for excessive CPU usage.
To determine whether a container cannot be scheduled or is being killed due to resource limits,
see the Troubleshooting section.
Monitoring compute & memory resource usage
The kubelet reports the resource usage of a Pod as part of the Pod
status.
If optional tools for monitoring
are available in your cluster, then Pod resource usage can be retrieved either
from the Metrics API
directly or from your monitoring tools.
Local ephemeral storage
FEATURE STATE:Kubernetes v1.25 [stable]
Nodes have local ephemeral storage, backed by
locally-attached writeable devices or, sometimes, by RAM.
"Ephemeral" means that there is no long-term guarantee about durability.
Pods use ephemeral local storage for scratch space, caching, and for logs.
The kubelet can provide scratch space to Pods using local ephemeral storage to
mount emptyDirvolumes into containers.
The kubelet also uses this kind of storage to hold
node-level container logs,
container images, and the writable layers of running containers.
Caution: If a node fails, the data in its ephemeral storage can be lost.
Your applications cannot expect any performance SLAs (disk IOPS for example)
from local ephemeral storage.
As a beta feature, Kubernetes lets you track, reserve and limit the amount
of ephemeral local storage a Pod can consume.
Configurations for local ephemeral storage
Kubernetes supports two ways to configure local ephemeral storage on a node:
In this configuration, you place all different kinds of ephemeral local data
(emptyDir volumes, writeable layers, container images, logs) into one filesystem.
The most effective way to configure the kubelet means dedicating this filesystem
to Kubernetes (kubelet) data.
The kubelet also writes
node-level container logs
and treats these similarly to ephemeral local storage.
The kubelet writes logs to files inside its configured log directory (/var/log
by default); and has a base directory for other locally stored data
(/var/lib/kubelet by default).
Typically, both /var/lib/kubelet and /var/log are on the system root filesystem,
and the kubelet is designed with that layout in mind.
Your node can have as many other filesystems, not used for Kubernetes,
as you like.
You have a filesystem on the node that you're using for ephemeral data that
comes from running Pods: logs, and emptyDir volumes. You can use this filesystem
for other data (for example: system logs not related to Kubernetes); it can even
be the root filesystem.
The kubelet also writes
node-level container logs
into the first filesystem, and treats these similarly to ephemeral local storage.
You also use a separate filesystem, backed by a different logical storage device.
In this configuration, the directory where you tell the kubelet to place
container image layers and writeable layers is on this second filesystem.
The first filesystem does not hold any image layers or writeable layers.
Your node can have as many other filesystems, not used for Kubernetes,
as you like.
The kubelet can measure how much local storage it is using. It does this provided
that you have set up the node using one of the supported configurations for local
ephemeral storage.
the LocalStorageCapacityIsolationfeature gate
is enabled (the feature is on by default), and you have set up the node using one
of the supported configurations for local ephemeral storage.
Quotas are faster and more accurate than directory scanning. The
LocalStorageCapacityIsolationFSQuotaMonitoringfeature gate is enabled (the feature is on by default),
If you have a different configuration, then the kubelet does not apply resource
limits for ephemeral local storage.
Note: The kubelet tracks tmpfs emptyDir volumes as container memory use, rather
than as local ephemeral storage.
Setting requests and limits for local ephemeral storage
You can specify ephemeral-storage for managing local ephemeral storage. Each
container of a Pod can specify either or both of the following:
Limits and requests for ephemeral-storage are measured in byte quantities.
You can express storage as a plain integer or as a fixed-point number using one of these suffixes:
E, P, T, G, M, k. You can also use the power-of-two equivalents: Ei, Pi, Ti, Gi,
Mi, Ki. For example, the following quantities all represent roughly the same value:
128974848
129e6
129M
123Mi
Pay attention to the case of the suffixes. If you request 400m of ephemeral-storage, this is a request
for 0.4 bytes. Someone who types that probably meant to ask for 400 mebibytes (400Mi)
or 400 megabytes (400M).
In the following example, the Pod has two containers. Each container has a request of
2GiB of local ephemeral storage. Each container has a limit of 4GiB of local ephemeral
storage. Therefore, the Pod has a request of 4GiB of local ephemeral storage, and
a limit of 8GiB of local ephemeral storage.
How Pods with ephemeral-storage requests are scheduled
When you create a Pod, the Kubernetes scheduler selects a node for the Pod to
run on. Each node has a maximum amount of local ephemeral storage it can provide for Pods.
For more information, see
Node Allocatable.
The scheduler ensures that the sum of the resource requests of the scheduled containers is less than the capacity of the node.
Ephemeral storage consumption management
If the kubelet is managing local ephemeral storage as a resource, then the
kubelet measures storage use in:
emptyDir volumes, except tmpfsemptyDir volumes
directories holding node-level logs
writeable container layers
If a Pod is using more ephemeral storage than you allow it to, the kubelet
sets an eviction signal that triggers Pod eviction.
For container-level isolation, if a container's writable layer and log
usage exceeds its storage limit, the kubelet marks the Pod for eviction.
For pod-level isolation the kubelet works out an overall Pod storage limit by
summing the limits for the containers in that Pod. In this case, if the sum of
the local ephemeral storage usage from all containers and also the Pod's emptyDir
volumes exceeds the overall Pod storage limit, then the kubelet also marks the Pod
for eviction.
Caution:
If the kubelet is not measuring local ephemeral storage, then a Pod
that exceeds its local storage limit will not be evicted for breaching
local storage resource limits.
However, if the filesystem space for writeable container layers, node-level logs,
or emptyDir volumes falls low, the node
taints itself as short on local storage
and this taint triggers eviction for any Pods that don't specifically tolerate the taint.
See the supported configurations
for ephemeral local storage.
The kubelet supports different ways to measure Pod storage use:
The kubelet performs regular, scheduled checks that scan each
emptyDir volume, container log directory, and writeable container layer.
The scan measures how much space is used.
Note:
In this mode, the kubelet does not track open file descriptors
for deleted files.
If you (or a container) create a file inside an emptyDir volume,
something then opens that file, and you delete the file while it is
still open, then the inode for the deleted file stays until you close
that file but the kubelet does not categorize the space as in use.
FEATURE STATE:Kubernetes v1.25 [beta]
Project quotas are an operating-system level feature for managing
storage use on filesystems. With Kubernetes, you can enable project
quotas for monitoring storage use. Make sure that the filesystem
backing the emptyDir volumes, on the node, provides project quota support.
For example, XFS and ext4fs offer project quotas.
Note: Project quotas let you monitor storage use; they do not enforce limits.
Kubernetes uses project IDs starting from 1048576. The IDs in use are
registered in /etc/projects and /etc/projid. If project IDs in
this range are used for other purposes on the system, those project
IDs must be registered in /etc/projects and /etc/projid so that
Kubernetes does not use them.
Quotas are faster and more accurate than directory scanning. When a
directory is assigned to a project, all files created under a
directory are created in that project, and the kernel merely has to
keep track of how many blocks are in use by files in that project.
If a file is created and deleted, but has an open file descriptor,
it continues to consume space. Quota tracking records that space accurately
whereas directory scans overlook the storage used by deleted files.
If you want to use project quotas, you should:
Enable the LocalStorageCapacityIsolationFSQuotaMonitoring=truefeature gate
using the featureGates field in the
kubelet configuration
or the --feature-gates command line flag.
Ensure that the root filesystem (or optional runtime filesystem)
has project quotas enabled. All XFS filesystems support project quotas.
For ext4 filesystems, you need to enable the project quota tracking feature
while the filesystem is not mounted.
# For ext4, with /dev/block-device not mountedsudo tune2fs -O project -Q prjquota /dev/block-device
Ensure that the root filesystem (or optional runtime filesystem) is
mounted with project quotas enabled. For both XFS and ext4fs, the
mount option is named prjquota.
Extended resources
Extended resources are fully-qualified resource names outside the
kubernetes.io domain. They allow cluster operators to advertise and users to
consume the non-Kubernetes-built-in resources.
There are two steps required to use Extended Resources. First, the cluster
operator must advertise an Extended Resource. Second, users must request the
Extended Resource in Pods.
Managing extended resources
Node-level extended resources
Node-level extended resources are tied to nodes.
Device plugin managed resources
See Device
Plugin
for how to advertise device plugin managed resources on each node.
Other resources
To advertise a new node-level extended resource, the cluster operator can
submit a PATCH HTTP request to the API server to specify the available
quantity in the status.capacity for a node in the cluster. After this
operation, the node's status.capacity will include a new resource. The
status.allocatable field is updated automatically with the new resource
asynchronously by the kubelet.
Because the scheduler uses the node's status.allocatable value when
evaluating Pod fitness, the scheduler only takes account of the new value after
that asynchronous update. There may be a short delay between patching the
node capacity with a new resource and the time when the first Pod that requests
the resource can be scheduled on that node.
Example:
Here is an example showing how to use curl to form an HTTP request that
advertises five "example.com/foo" resources on node k8s-node-1 whose master
is k8s-master.
Note: In the preceding request, ~1 is the encoding for the character /
in the patch path. The operation path value in JSON-Patch is interpreted as a
JSON-Pointer. For more details, see
IETF RFC 6901, section 3.
Cluster-level extended resources
Cluster-level extended resources are not tied to nodes. They are usually managed
by scheduler extenders, which handle the resource consumption and resource quota.
You can specify the extended resources that are handled by scheduler extenders
in scheduler configuration
Example:
The following configuration for a scheduler policy indicates that the
cluster-level extended resource "example.com/foo" is handled by the scheduler
extender.
The scheduler sends a Pod to the scheduler extender only if the Pod requests
"example.com/foo".
The ignoredByScheduler field specifies that the scheduler does not check
the "example.com/foo" resource in its PodFitsResources predicate.
Users can consume extended resources in Pod specs like CPU and memory.
The scheduler takes care of the resource accounting so that no more than the
available amount is simultaneously allocated to Pods.
The API server restricts quantities of extended resources to whole numbers.
Examples of valid quantities are 3, 3000m and 3Ki. Examples of
invalid quantities are 0.5 and 1500m.
Note: Extended resources replace Opaque Integer Resources.
Users can use any domain name prefix other than kubernetes.io which is reserved.
To consume an extended resource in a Pod, include the resource name as a key
in the spec.containers[].resources.limits map in the container spec.
Note: Extended resources cannot be overcommitted, so request and limit
must be equal if both are present in a container spec.
A Pod is scheduled only if all of the resource requests are satisfied, including
CPU, memory and any extended resources. The Pod remains in the PENDING state
as long as the resource request cannot be satisfied.
Example:
The Pod below requests 2 CPUs and 1 "example.com/foo" (an extended resource).
Process ID (PID) limits allow for the configuration of a kubelet
to limit the number of PIDs that a given Pod can consume. See
PID Limiting for information.
Troubleshooting
My Pods are pending with event message FailedScheduling
If the scheduler cannot find any node where a Pod can fit, the Pod remains
unscheduled until a place can be found. An
Event is produced
each time the scheduler fails to find a place for the Pod. You can use kubectl
to view the events for a Pod; for example:
kubectl describe pod frontend | grep -A 9999999999 Events
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 23s default-scheduler 0/42 nodes available: insufficient cpu
In the preceding example, the Pod named "frontend" fails to be scheduled due to
insufficient CPU resource on any node. Similar error messages can also suggest
failure due to insufficient memory (PodExceedsFreeMemory). In general, if a Pod
is pending with a message of this type, there are several things to try:
Add more nodes to the cluster.
Terminate unneeded Pods to make room for pending Pods.
Check that the Pod is not larger than all the nodes. For example, if all the
nodes have a capacity of cpu: 1, then a Pod with a request of cpu: 1.1 will
never be scheduled.
Check for node taints. If most of your nodes are tainted, and the new Pod does
not tolerate that taint, the scheduler only considers placements onto the
remaining nodes that don't have that taint.
You can check node capacities and amounts allocated with the
kubectl describe nodes command. For example:
In the preceding output, you can see that if a Pod requests more than 1.120 CPUs
or more than 6.23Gi of memory, that Pod will not fit on the node.
By looking at the “Pods” section, you can see which Pods are taking up space on
the node.
The amount of resources available to Pods is less than the node capacity because
system daemons use a portion of the available resources. Within the Kubernetes API,
each Node has a .status.allocatable field
(see NodeStatus
for details).
The .status.allocatable field describes the amount of resources that are available
to Pods on that node (for example: 15 virtual CPUs and 7538 MiB of memory).
For more information on node allocatable resources in Kubernetes, see
Reserve Compute Resources for System Daemons.
You can configure resource quotas
to limit the total amount of resources that a namespace can consume.
Kubernetes enforces quotas for objects in particular namespace when there is a
ResourceQuota in that namespace.
For example, if you assign specific namespaces to different teams, you
can add ResourceQuotas into those namespaces. Setting resource quotas helps to
prevent one team from using so much of any resource that this over-use affects other teams.
You should also consider what access you grant to that namespace:
full write access to a namespace allows someone with that access to remove any
resource, including a configured ResourceQuota.
My container is terminated
Your container might get terminated because it is resource-starved. To check
whether a container is being killed because it is hitting a resource limit, call
kubectl describe pod on the Pod of interest:
kubectl describe pod simmemleak-hra99
The output is similar to:
Name: simmemleak-hra99
Namespace: default
Image(s): saadali/simmemleak
Node: kubernetes-node-tf0f/10.240.216.66
Labels: name=simmemleak
Status: Running
Reason:
Message:
IP: 10.244.2.75
Containers:
simmemleak:
Image: saadali/simmemleak:latest
Limits:
cpu: 100m
memory: 50Mi
State: Running
Started: Tue, 07 Jul 2019 12:54:41 -0700
Last State: Terminated
Reason: OOMKilled
Exit Code: 137
Started: Fri, 07 Jul 2019 12:54:30 -0700
Finished: Fri, 07 Jul 2019 12:54:33 -0700
Ready: False
Restart Count: 5
Conditions:
Type Status
Ready False
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 42s default-scheduler Successfully assigned simmemleak-hra99 to kubernetes-node-tf0f
Normal Pulled 41s kubelet Container image "saadali/simmemleak:latest" already present on machine
Normal Created 41s kubelet Created container simmemleak
Normal Started 40s kubelet Started container simmemleak
Normal Killing 32s kubelet Killing container with id ead3fb35-5cf5-44ed-9ae1-488115be66c6: Need to kill Pod
In the preceding example, the Restart Count: 5 indicates that the simmemleak
container in the Pod was terminated and restarted five times (so far).
The OOMKilled reason shows that the container tried to use more memory than its limit.
Your next step might be to check the application code for a memory leak. If you
find that the application is behaving how you expect, consider setting a higher
memory limit (and possibly request) for that container.
3.8.5 - Organizing Cluster Access Using kubeconfig Files
Use kubeconfig files to organize information about clusters, users, namespaces, and
authentication mechanisms. The kubectl command-line tool uses kubeconfig files to
find the information it needs to choose a cluster and communicate with the API server
of a cluster.
Note: A file that is used to configure access to clusters is called
a kubeconfig file. This is a generic way of referring to configuration files.
It does not mean that there is a file named kubeconfig.
Warning: Only use kubeconfig files from trusted sources. Using a specially-crafted kubeconfig file could result in malicious code execution or file exposure.
If you must use an untrusted kubeconfig file, inspect it carefully first, much as you would a shell script.
By default, kubectl looks for a file named config in the $HOME/.kube directory.
You can specify other kubeconfig files by setting the KUBECONFIG environment
variable or by setting the
--kubeconfig flag.
Supporting multiple clusters, users, and authentication mechanisms
Suppose you have several clusters, and your users and components authenticate
in a variety of ways. For example:
A running kubelet might authenticate using certificates.
A user might authenticate using tokens.
Administrators might have sets of certificates that they provide to individual users.
With kubeconfig files, you can organize your clusters, users, and namespaces.
You can also define contexts to quickly and easily switch between
clusters and namespaces.
Context
A context element in a kubeconfig file is used to group access parameters
under a convenient name. Each context has three parameters: cluster, namespace, and user.
By default, the kubectl command-line tool uses parameters from
the current context to communicate with the cluster.
To choose the current context:
kubectl config use-context
The KUBECONFIG environment variable
The KUBECONFIG environment variable holds a list of kubeconfig files.
For Linux and Mac, the list is colon-delimited. For Windows, the list
is semicolon-delimited. The KUBECONFIG environment variable is not
required. If the KUBECONFIG environment variable doesn't exist,
kubectl uses the default kubeconfig file, $HOME/.kube/config.
If the KUBECONFIG environment variable does exist, kubectl uses
an effective configuration that is the result of merging the files
listed in the KUBECONFIG environment variable.
Merging kubeconfig files
To see your configuration, enter this command:
kubectl config view
As described previously, the output might be from a single kubeconfig file,
or it might be the result of merging several kubeconfig files.
Here are the rules that kubectl uses when it merges kubeconfig files:
If the --kubeconfig flag is set, use only the specified file. Do not merge.
Only one instance of this flag is allowed.
Otherwise, if the KUBECONFIG environment variable is set, use it as a
list of files that should be merged.
Merge the files listed in the KUBECONFIG environment variable
according to these rules:
Ignore empty filenames.
Produce errors for files with content that cannot be deserialized.
The first file to set a particular value or map key wins.
Never change the value or map key.
Example: Preserve the context of the first file to set current-context.
Example: If two files specify a red-user, use only values from the first file's red-user.
Even if the second file has non-conflicting entries under red-user, discard them.
Otherwise, use the default kubeconfig file, $HOME/.kube/config, with no merging.
Determine the context to use based on the first hit in this chain:
Use the --context command-line flag if it exists.
Use the current-context from the merged kubeconfig files.
An empty context is allowed at this point.
Determine the cluster and user. At this point, there might or might not be a context.
Determine the cluster and user based on the first hit in this chain,
which is run twice: once for user and once for cluster:
Use a command-line flag if it exists: --user or --cluster.
If the context is non-empty, take the user or cluster from the context.
The user and cluster can be empty at this point.
Determine the actual cluster information to use. At this point, there might or
might not be cluster information.
Build each piece of the cluster information based on this chain; the first hit wins:
Use command line flags if they exist: --server, --certificate-authority, --insecure-skip-tls-verify.
If any cluster information attributes exist from the merged kubeconfig files, use them.
If there is no server location, fail.
Determine the actual user information to use. Build user information using the same
rules as cluster information, except allow only one authentication
technique per user:
Use command line flags if they exist: --client-certificate, --client-key, --username, --password, --token.
Use the user fields from the merged kubeconfig files.
If there are two conflicting techniques, fail.
For any information still missing, use default values and potentially
prompt for authentication information.
File references
File and path references in a kubeconfig file are relative to the location of the kubeconfig file.
File references on the command line are relative to the current working directory.
In $HOME/.kube/config, relative paths are stored relatively, and absolute paths
are stored absolutely.
Proxy
You can configure kubectl to use a proxy per cluster using proxy-url in your kubeconfig file, like this:
This page outlines the differences in how resources are managed between Linux and Windows.
On Linux nodes, cgroups are used
as a pod boundary for resource control. Containers are created within that boundary
for network, process and file system isolation. The Linux cgroup APIs can be used to
gather CPU, I/O, and memory use statistics.
In contrast, Windows uses a job object per container with a system namespace filter
to contain all processes in a container and provide logical isolation from the
host.
(Job objects are a Windows process isolation mechanism and are different from
what Kubernetes refers to as a Job).
There is no way to run a Windows container without the namespace filtering in
place. This means that system privileges cannot be asserted in the context of the
host, and thus privileged containers are not available on Windows.
Containers cannot assume an identity from the host because the Security Account Manager
(SAM) is separate.
Memory management
Windows does not have an out-of-memory process killer as Linux does. Windows always
treats all user-mode memory allocations as virtual, and pagefiles are mandatory.
Windows nodes do not overcommit memory for processes. The
net effect is that Windows won't reach out of memory conditions the same way Linux
does, and processes page to disk instead of being subject to out of memory (OOM)
termination. If memory is over-provisioned and all physical memory is exhausted,
then paging can slow down performance.
CPU management
Windows can limit the amount of CPU time allocated for different processes but cannot
guarantee a minimum amount of CPU time.
On Windows, the kubelet supports a command-line flag to set the
scheduling priority of the
kubelet process: --windows-priorityclass. This flag allows the kubelet process to get
more CPU time slices when compared to other processes running on the Windows host.
More information on the allowable values and their meaning is available at
Windows Priority Classes.
To ensure that running Pods do not starve the kubelet of CPU cycles, set this flag to ABOVE_NORMAL_PRIORITY_CLASS or above.
Resource reservation
To account for memory and CPU used by the operating system, the container runtime, and by
Kubernetes host processes such as the kubelet, you can (and should) reserve
memory and CPU resources with the --kube-reserved and/or --system-reserved kubelet flags.
On Windows these values are only used to calculate the node's
allocatable resources.
Caution:
As you deploy workloads, set resource memory and CPU limits on containers.
This also subtracts from NodeAllocatable and helps the cluster-wide scheduler in determining which pods to place on which nodes.
Scheduling pods without limits may over-provision the Windows nodes and in extreme
cases can cause the nodes to become unhealthy.
On Windows, a good practice is to reserve at least 2GiB of memory.
To determine how much CPU to reserve,
identify the maximum pod density for each node and monitor the CPU usage of
the system services running there, then choose a value that meets your workload needs.
3.9 - Security
Concepts for keeping your cloud-native workload secure.
3.9.1 - Overview of Cloud Native Security
A model for thinking about Kubernetes security in the context of Cloud Native security.
This overview defines a model for thinking about Kubernetes security in the context of Cloud Native security.
Warning: This container security model provides suggestions, not proven information security policies.
The 4C's of Cloud Native security
You can think about security in layers. The 4C's of Cloud Native security are Cloud,
Clusters, Containers, and Code.
Note: This layered approach augments the defense in depth
computing approach to security, which is widely regarded as a best practice for securing
software systems.
The 4C's of Cloud Native Security
Each layer of the Cloud Native security model builds upon the next outermost layer.
The Code layer benefits from strong base (Cloud, Cluster, Container) security layers.
You cannot safeguard against poor security standards in the base layers by addressing
security at the Code level.
Cloud
In many ways, the Cloud (or co-located servers, or the corporate datacenter) is the
trusted computing base
of a Kubernetes cluster. If the Cloud layer is vulnerable (or
configured in a vulnerable way) then there is no guarantee that the components built
on top of this base are secure. Each cloud provider makes security recommendations
for running workloads securely in their environment.
Cloud provider security
If you are running a Kubernetes cluster on your own hardware or a different cloud provider,
consult your documentation for security best practices.
Here are links to some of the popular cloud providers' security documentation:
Suggestions for securing your infrastructure in a Kubernetes cluster:
Infrastructure security
Area of Concern for Kubernetes Infrastructure
Recommendation
Network access to API Server (Control plane)
All access to the Kubernetes control plane is not allowed publicly on the internet and is controlled by network access control lists restricted to the set of IP addresses needed to administer the cluster.
Network access to Nodes (nodes)
Nodes should be configured to only accept connections (via network access control lists) from the control plane on the specified ports, and accept connections for services in Kubernetes of type NodePort and LoadBalancer. If possible, these nodes should not be exposed on the public internet entirely.
Kubernetes access to Cloud Provider API
Each cloud provider needs to grant a different set of permissions to the Kubernetes control plane and nodes. It is best to provide the cluster with cloud provider access that follows the principle of least privilege for the resources it needs to administer. The Kops documentation provides information about IAM policies and roles.
Access to etcd
Access to etcd (the datastore of Kubernetes) should be limited to the control plane only. Depending on your configuration, you should attempt to use etcd over TLS. More information can be found in the etcd documentation.
etcd Encryption
Wherever possible it's a good practice to encrypt all storage at rest, and since etcd holds the state of the entire cluster (including Secrets) its disk should especially be encrypted at rest.
Cluster
There are two areas of concern for securing Kubernetes:
Securing the cluster components that are configurable
Securing the applications which run in the cluster
Components of the Cluster
If you want to protect your cluster from accidental or malicious access and adopt
good information practices, read and follow the advice about
securing your cluster.
Components in the cluster (your application)
Depending on the attack surface of your application, you may want to focus on specific
aspects of security. For example: If you are running a service (Service A) that is critical
in a chain of other resources and a separate workload (Service B) which is
vulnerable to a resource exhaustion attack, then the risk of compromising Service A
is high if you do not limit the resources of Service B. The following table lists
areas of security concerns and recommendations for securing workloads running in Kubernetes:
Container security is outside the scope of this guide. Here are general recommendations and
links to explore this topic:
Area of Concern for Containers
Recommendation
Container Vulnerability Scanning and OS Dependency Security
As part of an image build step, you should scan your containers for known vulnerabilities.
Image Signing and Enforcement
Sign container images to maintain a system of trust for the content of your containers.
Disallow privileged users
When constructing containers, consult your documentation for how to create users inside of the containers that have the least level of operating system privilege necessary in order to carry out the goal of the container.
Application code is one of the primary attack surfaces over which you have the most control.
While securing application code is outside of the Kubernetes security topic, here
are recommendations to protect application code:
Code security
Code security
Area of Concern for Code
Recommendation
Access over TLS only
If your code needs to communicate by TCP, perform a TLS handshake with the client ahead of time. With the exception of a few cases, encrypt everything in transit. Going one step further, it's a good idea to encrypt network traffic between services. This can be done through a process known as mutual TLS authentication or mTLS which performs a two sided verification of communication between two certificate holding services.
Limiting port ranges of communication
This recommendation may be a bit self-explanatory, but wherever possible you should only expose the ports on your service that are absolutely essential for communication or metric gathering.
3rd Party Dependency Security
It is a good practice to regularly scan your application's third party libraries for known security vulnerabilities. Each programming language has a tool for performing this check automatically.
Static Code Analysis
Most languages provide a way for a snippet of code to be analyzed for any potentially unsafe coding practices. Whenever possible you should perform checks using automated tooling that can scan codebases for common security errors. Some of the tools can be found at: https://owasp.org/www-community/Source_Code_Analysis_Tools
Dynamic probing attacks
There are a few automated tools that you can run against your service to try some of the well known service attacks. These include SQL injection, CSRF, and XSS. One of the most popular dynamic analysis tools is the OWASP Zed Attack proxy tool.
A detailed look at the different policy levels defined in the Pod Security Standards.
The Pod Security Standards define three different policies to broadly cover the security
spectrum. These policies are cumulative and range from highly-permissive to highly-restrictive.
This guide outlines the requirements of each policy.
Profile
Description
Privileged
Unrestricted policy, providing the widest possible level of permissions. This policy allows for known privilege escalations.
Baseline
Minimally restrictive policy which prevents known privilege escalations. Allows the default (minimally specified) Pod configuration.
Restricted
Heavily restricted policy, following current Pod hardening best practices.
Profile Details
Privileged
The Privileged policy is purposely-open, and entirely unrestricted. This type of policy is
typically aimed at system- and infrastructure-level workloads managed by privileged, trusted users.
The Privileged policy is defined by an absence of restrictions. Allow-by-default
mechanisms (such as gatekeeper) may be Privileged by default. In contrast, for a deny-by-default mechanism (such as Pod
Security Policy) the Privileged policy should disable all restrictions.
Baseline
The Baseline policy is aimed at ease of adoption for common containerized workloads while
preventing known privilege escalations. This policy is targeted at application operators and
developers of non-critical applications. The following listed controls should be
enforced/disallowed:
Note: In this table, wildcards (*) indicate all elements in a list. For example,
spec.containers[*].securityContext refers to the Security Context object for all defined
containers. If any of the listed containers fails to meet the requirements, the entire pod will
fail validation.
Baseline policy specification
Control
Policy
HostProcess
Windows pods offer the ability to run HostProcess containers which enables privileged access to the Windows node. Privileged access to the host is disallowed in the baseline policy.
HostPorts should be disallowed, or at minimum restricted to a known list.
Restricted Fields
spec.containers[*].ports[*].hostPort
spec.initContainers[*].ports[*].hostPort
spec.ephemeralContainers[*].ports[*].hostPort
Allowed Values
Undefined/nil
Known list
0
AppArmor
On supported hosts, the runtime/default AppArmor profile is applied by default. The baseline policy should prevent overriding or disabling the default AppArmor profile, or restrict overrides to an allowed set of profiles.
Sysctls can disable security mechanisms or affect all containers on a host, and should be disallowed except for an allowed "safe" subset. A sysctl is considered safe if it is namespaced in the container or the Pod, and it is isolated from other Pods or processes on the same Node.
Restricted Fields
spec.securityContext.sysctls[*].name
Allowed Values
Undefined/nil
kernel.shm_rmid_forced
net.ipv4.ip_local_port_range
net.ipv4.ip_unprivileged_port_start
net.ipv4.tcp_syncookies
net.ipv4.ping_group_range
Restricted
The Restricted policy is aimed at enforcing current Pod hardening best practices, at the
expense of some compatibility. It is targeted at operators and developers of security-critical
applications, as well as lower-trust users. The following listed controls should be
enforced/disallowed:
Note: In this table, wildcards (*) indicate all elements in a list. For example,
spec.containers[*].securityContext refers to the Security Context object for all defined
containers. If any of the listed containers fails to meet the requirements, the entire pod will
fail validation.
Restricted policy specification
Control
Policy
Everything from the baseline profile.
Volume Types
The restricted policy only permits the following volume types.
Restricted Fields
spec.volumes[*]
Allowed Values
Every item in the spec.volumes[*] list must set one of the following fields to a non-null value:
spec.volumes[*].configMap
spec.volumes[*].csi
spec.volumes[*].downwardAPI
spec.volumes[*].emptyDir
spec.volumes[*].ephemeral
spec.volumes[*].persistentVolumeClaim
spec.volumes[*].projected
spec.volumes[*].secret
Privilege Escalation (v1.8+)
Privilege escalation (such as via set-user-ID or set-group-ID file mode) should not be allowed. This is Linux only policy in v1.25+ (spec.os.name != windows)
Seccomp profile must be explicitly set to one of the allowed values. Both the Unconfined profile and the absence of a profile are prohibited. This is Linux only policy in v1.25+ (spec.os.name != windows)
The container fields may be undefined/nil if the pod-level
spec.securityContext.seccompProfile.type field is set appropriately.
Conversely, the pod-level field may be undefined/nil if _all_ container-
level fields are set.
Capabilities (v1.22+)
Containers must drop ALL capabilities, and are only permitted to add back
the NET_BIND_SERVICE capability. This is Linux only policy in v1.25+ (.spec.os.name != "windows")
Decoupling policy definition from policy instantiation allows for a common understanding and
consistent language of policies across clusters, independent of the underlying enforcement
mechanism.
As mechanisms mature, they will be defined below on a per-policy basis. The methods of enforcement
of individual policies are not defined here.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Other alternatives for enforcing policies are being developed in the Kubernetes ecosystem, such as:
Kubernetes lets you use nodes that run either Linux or Windows. You can mix both kinds of
node in one cluster.
Windows in Kubernetes has some limitations and differentiators from Linux-based
workloads. Specifically, many of the Pod securityContext fields
have no effect on Windows.
Note: Kubelets prior to v1.24 don't enforce the pod OS field, and if a cluster has nodes on versions earlier than v1.24 the restricted policies should be pinned to a version prior to v1.25.
Restricted Pod Security Standard changes
Another important change, made in Kubernetes v1.25 is that the restricted Pod security
has been updated to use the pod.spec.os.name field. Based on the OS name, certain policies that are specific
to a particular OS can be relaxed for the other OS.
OS-specific policy controls
Restrictions on the following controls are only required if .spec.os.name is not windows:
Privilege Escalation
Seccomp
Linux Capabilities
FAQ
Why isn't there a profile between privileged and baseline?
The three profiles defined here have a clear linear progression from most secure (restricted) to least
secure (privileged), and cover a broad set of workloads. Privileges required above the baseline
policy are typically very application specific, so we do not offer a standard profile in this
niche. This is not to say that the privileged profile should always be used in this case, but that
policies in this space need to be defined on a case-by-case basis.
SIG Auth may reconsider this position in the future, should a clear need for other profiles arise.
What's the difference between a security profile and a security context?
Security Contexts configure Pods and
Containers at runtime. Security contexts are defined as part of the Pod and container specifications
in the Pod manifest, and represent parameters to the container runtime.
Security profiles are control plane mechanisms to enforce specific settings in the Security Context,
as well as other related parameters outside the Security Context. As of July 2021,
Pod Security Policies are deprecated in favor of the
built-in Pod Security Admission Controller.
What about sandboxed Pods?
There is not currently an API standard that controls whether a Pod is considered sandboxed or
not. Sandbox Pods may be identified by the use of a sandboxed runtime (such as gVisor or Kata
Containers), but there is no standard definition of what a sandboxed runtime is.
The protections necessary for sandboxed workloads can differ from others. For example, the need to
restrict privileged permissions is lessened when the workload is isolated from the underlying
kernel. This allows for workloads requiring heightened permissions to still be isolated.
Additionally, the protection of sandboxed workloads is highly dependent on the method of
sandboxing. As such, no single recommended profile is recommended for all sandboxed workloads.
3.9.3 - Pod Security Admission
An overview of the Pod Security Admission Controller, which can enforce the Pod Security Standards.
FEATURE STATE:Kubernetes v1.25 [stable]
The Kubernetes Pod Security Standards define
different isolation levels for Pods. These standards let you define how you want to restrict the
behavior of pods in a clear, consistent fashion.
Kubernetes offers a built-in Pod Securityadmission controller to enforce the Pod Security Standards. Pod security restrictions
are applied at the namespace level when pods are
created.
Built-in Pod Security admission enforcement
This page is part of the documentation for Kubernetes v1.25.
If you are running a different version of Kubernetes, consult the documentation for that release.
Pod Security levels
Pod Security admission places requirements on a Pod's Security
Context and other related fields according
to the three levels defined by the Pod Security
Standards: privileged, baseline, and
restricted. Refer to the Pod Security Standards
page for an in-depth look at those requirements.
Pod Security Admission labels for namespaces
Once the feature is enabled or the webhook is installed, you can configure namespaces to define the admission
control mode you want to use for pod security in each namespace. Kubernetes defines a set of
labels that you can set to define which of the
predefined Pod Security Standard levels you want to use for a namespace. The label you select
defines what action the control plane
takes if a potential violation is detected:
Pod Security Admission modes
Mode
Description
enforce
Policy violations will cause the pod to be rejected.
audit
Policy violations will trigger the addition of an audit annotation to the event recorded in the audit log, but are otherwise allowed.
warn
Policy violations will trigger a user-facing warning, but are otherwise allowed.
A namespace can configure any or all modes, or even set a different level for different modes.
For each mode, there are two labels that determine the policy used:
# The per-mode level label indicates which policy level to apply for the mode.## MODE must be one of `enforce`, `audit`, or `warn`.# LEVEL must be one of `privileged`, `baseline`, or `restricted`.pod-security.kubernetes.io/<MODE>:<LEVEL># Optional: per-mode version label that can be used to pin the policy to the# version that shipped with a given Kubernetes minor version (for example v1.25).## MODE must be one of `enforce`, `audit`, or `warn`.# VERSION must be a valid Kubernetes minor version, or `latest`.pod-security.kubernetes.io/<MODE>-version:<VERSION>
Pods are often created indirectly, by creating a workload
object such as a Deployment or Job. The workload object defines a
Pod template and a controller for the
workload resource creates Pods based on that template. To help catch violations early, both the
audit and warning modes are applied to the workload resources. However, enforce mode is not
applied to workload resources, only to the resulting pod objects.
Exemptions
You can define exemptions from pod security enforcement in order to allow the creation of pods that
would have otherwise been prohibited due to the policy associated with a given namespace.
Exemptions can be statically configured in the
Admission Controller configuration.
Exemptions must be explicitly enumerated. Requests meeting exemption criteria are ignored by the
Admission Controller (all enforce, audit and warn behaviors are skipped). Exemption dimensions include:
Usernames: requests from users with an exempt authenticated (or impersonated) username are
ignored.
RuntimeClassNames: pods and workload resources specifying an exempt runtime class name are
ignored.
Namespaces: pods and workload resources in an exempt namespace are ignored.
Caution: Most pods are created by a controller in response to a workload
resource, meaning that exempting an end user will only
exempt them from enforcement when creating pods directly, but not when creating a workload resource.
Controller service accounts (such as system:serviceaccount:kube-system:replicaset-controller)
should generally not be exempted, as doing so would implicitly exempt any user that can create the
corresponding workload resource.
Updates to the following pod fields are exempt from policy checks, meaning that if a pod update
request only changes these fields, it will not be denied even if the pod is in violation of the
current policy level:
Any metadata updates except changes to the seccomp or AppArmor annotations:
PodSecurityPolicy was deprecated
in Kubernetes v1.21, and removed from Kubernetes in v1.25.
Instead of using PodSecurityPolicy, you can enforce similar restrictions on Pods using
either or both:
If you are not running Kubernetes v1.25, check the documentation for
your version of Kubernetes.
3.9.5 - Security For Windows Nodes
This page describes security considerations and best practices specific to the Windows operating system.
Protection for Secret data on nodes
On Windows, data from Secrets are written out in clear text onto the node's local
storage (as compared to using tmpfs / in-memory filesystems on Linux). As a cluster
operator, you should take both of the following additional measures:
Use file ACLs to secure the Secrets' file location.
RunAsUsername
can be specified for Windows Pods or containers to execute the container
processes as specific user. This is roughly equivalent to
RunAsUser.
Windows containers offer two default user accounts, ContainerUser and ContainerAdministrator.
The differences between these two user accounts are covered in
When to use ContainerAdmin and ContainerUser user accounts
within Microsoft's Secure Windows containers documentation.
Local users can be added to container images during the container build process.
Note:
Nano Server based images run as
ContainerUser by default
Server Core based images run as
ContainerAdministrator by default
Linux-specific pod security context mechanisms (such as SELinux, AppArmor, Seccomp, or custom
POSIX capabilities) are not supported on Windows nodes.
Privileged containers are not supported
on Windows.
Instead HostProcess containers
can be used on Windows to perform many of the tasks performed by privileged containers on Linux.
3.9.6 - Controlling Access to the Kubernetes API
This page provides an overview of controlling access to the Kubernetes API.
Users access the Kubernetes API using kubectl,
client libraries, or by making REST requests. Both human users and
Kubernetes service accounts can be
authorized for API access.
When a request reaches the API, it goes through several stages, illustrated in the
following diagram:
Transport security
By default, the Kubernetes API server listens on port 6443 on the first non-localhost network interface, protected by TLS. In a typical production Kubernetes cluster, the API serves on port 443. The port can be changed with the --secure-port, and the listening IP address with the --bind-address flag.
The API server presents a certificate. This certificate may be signed using
a private certificate authority (CA), or based on a public key infrastructure linked
to a generally recognized CA. The certificate and corresponding private key can be set by using the --tls-cert-file and --tls-private-key-file flags.
If your cluster uses a private certificate authority, you need a copy of that CA
certificate configured into your ~/.kube/config on the client, so that you can
trust the connection and be confident it was not intercepted.
Your client can present a TLS client certificate at this stage.
Authentication
Once TLS is established, the HTTP request moves to the Authentication step.
This is shown as step 1 in the diagram.
The cluster creation script or cluster admin configures the API server to run
one or more Authenticator modules.
Authenticators are described in more detail in
Authentication.
The input to the authentication step is the entire HTTP request; however, it typically
examines the headers and/or client certificate.
Authentication modules include client certificates, password, and plain tokens,
bootstrap tokens, and JSON Web Tokens (used for service accounts).
Multiple authentication modules can be specified, in which case each one is tried in sequence,
until one of them succeeds.
If the request cannot be authenticated, it is rejected with HTTP status code 401.
Otherwise, the user is authenticated as a specific username, and the user name
is available to subsequent steps to use in their decisions. Some authenticators
also provide the group memberships of the user, while other authenticators
do not.
While Kubernetes uses usernames for access control decisions and in request logging,
it does not have a User object nor does it store usernames or other information about
users in its API.
Authorization
After the request is authenticated as coming from a specific user, the request must be authorized. This is shown as step 2 in the diagram.
A request must include the username of the requester, the requested action, and the object affected by the action. The request is authorized if an existing policy declares that the user has permissions to complete the requested action.
For example, if Bob has the policy below, then he can read pods only in the namespace projectCaribou:
If Bob makes a request to write (create or update) to the objects in the projectCaribou namespace, his authorization is denied. If Bob makes a request to read (get) objects in a different namespace such as projectFish, then his authorization is denied.
Kubernetes authorization requires that you use common REST attributes to interact with existing organization-wide or cloud-provider-wide access control systems. It is important to use REST formatting because these control systems might interact with other APIs besides the Kubernetes API.
Kubernetes supports multiple authorization modules, such as ABAC mode, RBAC Mode, and Webhook mode. When an administrator creates a cluster, they configure the authorization modules that should be used in the API server. If more than one authorization modules are configured, Kubernetes checks each module, and if any module authorizes the request, then the request can proceed. If all of the modules deny the request, then the request is denied (HTTP status code 403).
To learn more about Kubernetes authorization, including details about creating policies using the supported authorization modules, see Authorization.
Admission control
Admission Control modules are software modules that can modify or reject requests.
In addition to the attributes available to Authorization modules, Admission
Control modules can access the contents of the object that is being created or modified.
Admission controllers act on requests that create, modify, delete, or connect to (proxy) an object.
Admission controllers do not act on requests that merely read objects.
When multiple admission controllers are configured, they are called in order.
This is shown as step 3 in the diagram.
Unlike Authentication and Authorization modules, if any admission controller module
rejects, then the request is immediately rejected.
In addition to rejecting objects, admission controllers can also set complex defaults for
fields.
Once a request passes all admission controllers, it is validated using the validation routines
for the corresponding API object, and then written to the object store (shown as step 4).
Auditing
Kubernetes auditing provides a security-relevant, chronological set of records documenting the sequence of actions in a cluster.
The cluster audits the activities generated by users, by applications that use the Kubernetes API, and by the control plane itself.
how Pods can use
Secrets
to obtain API credentials.
3.9.7 - Role Based Access Control Good Practices
Principles and practices for good RBAC design for cluster operators.
Kubernetes RBAC is a key security control
to ensure that cluster users and workloads have only the access to resources required to
execute their roles. It is important to ensure that, when designing permissions for cluster
users, the cluster administrator understands the areas where privilege escalation could occur,
to reduce the risk of excessive access leading to security incidents.
The good practices laid out here should be read in conjunction with the general
RBAC documentation.
General good practice
Least privilege
Ideally, minimal RBAC rights should be assigned to users and service accounts. Only permissions
explicitly required for their operation should be used. While each cluster will be different,
some general rules that can be applied are :
Assign permissions at the namespace level where possible. Use RoleBindings as opposed to
ClusterRoleBindings to give users rights only within a specific namespace.
Avoid providing wildcard permissions when possible, especially to all resources.
As Kubernetes is an extensible system, providing wildcard access gives rights
not just to all object types that currently exist in the cluster, but also to all future object types
which are created in the future.
Administrators should not use cluster-admin accounts except where specifically needed.
Providing a low privileged account with
impersonation rights
can avoid accidental modification of cluster resources.
Avoid adding users to the system:masters group. Any user who is a member of this group
bypasses all RBAC rights checks and will always have unrestricted superuser access, which cannot be
revoked by removing RoleBindings or ClusterRoleBindings. As an aside, if a cluster is
using an authorization webhook, membership of this group also bypasses that webhook (requests
from users who are members of that group are never sent to the webhook)
Minimize distribution of privileged tokens
Ideally, pods shouldn't be assigned service accounts that have been granted powerful permissions
(for example, any of the rights listed under privilege escalation risks).
In cases where a workload requires powerful permissions, consider the following practices:
Limit the number of nodes running powerful pods. Ensure that any DaemonSets you run
are necessary and are run with least privilege to limit the blast radius of container escapes.
Avoid running powerful pods alongside untrusted or publicly-exposed ones. Consider using
Taints and Toleration,
NodeAffinity, or
PodAntiAffinity
to ensure pods don't run alongside untrusted or less-trusted Pods. Pay especial attention to
situations where less-trustworthy Pods are not meeting the Restricted Pod Security Standard.
Hardening
Kubernetes defaults to providing access which may not be required in every cluster. Reviewing
the RBAC rights provided by default can provide opportunities for security hardening.
In general, changes should not be made to rights provided to system: accounts some options
to harden cluster rights exist:
Review bindings for the system:unauthenticated group and remove them where possible, as this gives
access to anyone who can contact the API server at a network level.
Avoid the default auto-mounting of service account tokens by setting
automountServiceAccountToken: false. For more details, see
using default service account token.
Setting this value for a Pod will overwrite the service account setting, workloads
which require service account tokens can still mount them.
Periodic review
It is vital to periodically review the Kubernetes RBAC settings for redundant entries and
possible privilege escalations.
If an attacker is able to create a user account with the same name as a deleted user,
they can automatically inherit all the rights of the deleted user, especially the
rights assigned to that user.
Kubernetes RBAC - privilege escalation risks
Within Kubernetes RBAC there are a number of privileges which, if granted, can allow a user or a service account
to escalate their privileges in the cluster or affect systems outside the cluster.
This section is intended to provide visibility of the areas where cluster operators
should take care, to ensure that they do not inadvertently allow for more access to clusters than intended.
Listing secrets
It is generally clear that allowing get access on Secrets will allow a user to read their contents.
It is also important to note that list and watch access also effectively allow for users to reveal the Secret contents.
For example, when a List response is returned (for example, via kubectl get secrets -A -o yaml), the response
includes the contents of all Secrets.
Workload creation
Users who are able to create workloads (either Pods, or
workload resources that manage Pods) will
be able to gain access to the underlying node unless restrictions based on the Kubernetes
Pod Security Standards are in place.
Users who can run privileged Pods can use that access to gain node access and potentially to
further elevate their privileges. Where you do not fully trust a user or other principal
with the ability to create suitably secure and isolated Pods, you should enforce either the
Baseline or Restricted Pod Security Standard.
You can use Pod Security admission
or other (third party) mechanisms to implement that enforcement.
You can also use the deprecated PodSecurityPolicy mechanism
to restrict users' abilities to create privileged Pods (N.B. PodSecurityPolicy is scheduled for removal
in version 1.25).
Creating a workload in a namespace also grants indirect access to Secrets in that namespace.
Creating a pod in kube-system or a similarly privileged namespace can grant a user access to
Secrets they would not have through RBAC directly.
Persistent volume creation
As noted in the PodSecurityPolicy
documentation, access to create PersistentVolumes can allow for escalation of access to the underlying host.
Where access to persistent storage is required trusted administrators should create
PersistentVolumes, and constrained users should use PersistentVolumeClaims to access that storage.
Access to proxy subresource of Nodes
Users with access to the proxy sub-resource of node objects have rights to the Kubelet API,
which allows for command execution on every pod on the node(s) to which they have rights.
This access bypasses audit logging and admission control, so care should be taken before
granting rights to this resource.
Escalate verb
Generally, the RBAC system prevents users from creating clusterroles with more rights than the user possesses.
The exception to this is the escalate verb. As noted in the RBAC documentation,
users with this right can effectively escalate their privileges.
Bind verb
Similar to the escalate verb, granting users this right allows for the bypass of Kubernetes
in-built protections against privilege escalation, allowing users to create bindings to
roles with rights they do not already have.
Impersonate verb
This verb allows users to impersonate and gain the rights of other users in the cluster.
Care should be taken when granting it, to ensure that excessive permissions cannot be gained
via one of the impersonated accounts.
CSRs and certificate issuing
The CSR API allows for users with create rights to CSRs and update rights on certificatesigningrequests/approval
where the signer is kubernetes.io/kube-apiserver-client to create new client certificates
which allow users to authenticate to the cluster. Those client certificates can have arbitrary
names including duplicates of Kubernetes system components. This will effectively allow for privilege escalation.
Token request
Users with create rights on serviceaccounts/token can create TokenRequests to issue
tokens for existing service accounts.
Control admission webhooks
Users with control over validatingwebhookconfigurations or mutatingwebhookconfigurations
can control webhooks that can read any object admitted to the cluster, and in the case of
mutating webhooks, also mutate admitted objects.
Kubernetes RBAC - denial of service risks
Object creation denial-of-service
Users who have rights to create objects in a cluster may be able to create sufficient large
objects to create a denial of service condition either based on the size or number of objects, as discussed in
etcd used by Kubernetes is vulnerable to OOM attack. This may be
specifically relevant in multi-tenant clusters if semi-trusted or untrusted users
are allowed limited access to a system.
One option for mitigation of this issue would be to use
resource quotas
to limit the quantity of objects which can be created.
This page provides an overview of available configuration options and best practices for cluster multi-tenancy.
Sharing clusters saves costs and simplifies administration. However, sharing clusters also presents challenges such as security, fairness, and managing noisy neighbors.
Clusters can be shared in many ways. In some cases, different applications may run in the same cluster. In other cases, multiple instances of the same application may run in the same cluster, one for each end user. All these types of sharing are frequently described using the umbrella term multi-tenancy.
While Kubernetes does not have first-class concepts of end users or tenants, it provides several features to help manage different tenancy requirements. These are discussed below.
Use cases
The first step to determining how to share your cluster is understanding your use case, so you can evaluate the patterns and tools available. In general, multi-tenancy in Kubernetes clusters falls into two broad categories, though many variations and hybrids are also possible.
Multiple teams
A common form of multi-tenancy is to share a cluster between multiple teams within an organization, each of whom may operate one or more workloads. These workloads frequently need to communicate with each other, and with other workloads located on the same or different clusters.
In this scenario, members of the teams often have direct access to Kubernetes resources via tools such as kubectl, or indirect access through GitOps controllers or other types of release automation tools. There is often some level of trust between members of different teams, but Kubernetes policies such as RBAC, quotas, and network policies are essential to safely and fairly share clusters.
Multiple customers
The other major form of multi-tenancy frequently involves a Software-as-a-Service (SaaS) vendor running multiple instances of a workload for customers. This business model is so strongly associated with this deployment style that many people call it "SaaS tenancy." However, a better term might be "multi-customer tenancy,” since SaaS vendors may also use other deployment models, and this deployment model can also be used outside of SaaS.
In this scenario, the customers do not have access to the cluster; Kubernetes is invisible from their perspective and is only used by the vendor to manage the workloads. Cost optimization is frequently a critical concern, and Kubernetes policies are used to ensure that the workloads are strongly isolated from each other.
Terminology
Tenants
When discussing multi-tenancy in Kubernetes, there is no single definition for a "tenant". Rather, the definition of a tenant will vary depending on whether multi-team or multi-customer tenancy is being discussed.
In multi-team usage, a tenant is typically a team, where each team typically deploys a small number of workloads that scales with the complexity of the service. However, the definition of "team" may itself be fuzzy, as teams may be organized into higher-level divisions or subdivided into smaller teams.
By contrast, if each team deploys dedicated workloads for each new client, they are using a multi-customer model of tenancy. In this case, a "tenant" is simply a group of users who share a single workload. This may be as large as an entire company, or as small as a single team at that company.
In many cases, the same organization may use both definitions of "tenants" in different contexts. For example, a platform team may offer shared services such as security tools and databases to multiple internal “customers” and a SaaS vendor may also have multiple teams sharing a development cluster. Finally, hybrid architectures are also possible, such as a SaaS provider using a combination of per-customer workloads for sensitive data, combined with multi-tenant shared services.
A cluster showing coexisting tenancy models
Isolation
There are several ways to design and build multi-tenant solutions with Kubernetes. Each of these methods comes with its own set of tradeoffs that impact the isolation level, implementation effort, operational complexity, and cost of service.
A Kubernetes cluster consists of a control plane which runs Kubernetes software, and a data plane consisting of worker nodes where tenant workloads are executed as pods. Tenant isolation can be applied in both the control plane and the data plane based on organizational requirements.
The level of isolation offered is sometimes described using terms like “hard” multi-tenancy, which implies strong isolation, and “soft” multi-tenancy, which implies weaker isolation. In particular, "hard" multi-tenancy is often used to describe cases where the tenants do not trust each other, often from security and resource sharing perspectives (e.g. guarding against attacks such as data exfiltration or DoS). Since data planes typically have much larger attack surfaces, "hard" multi-tenancy often requires extra attention to isolating the data-plane, though control plane isolation also remains critical.
However, the terms "hard" and "soft" can often be confusing, as there is no single definition that will apply to all users. Rather, "hardness" or "softness" is better understood as a broad spectrum, with many different techniques that can be used to maintain different types of isolation in your clusters, based on your requirements.
In more extreme cases, it may be easier or necessary to forgo any cluster-level sharing at all and assign each tenant their dedicated cluster, possibly even running on dedicated hardware if VMs are not considered an adequate security boundary. This may be easier with managed Kubernetes clusters, where the overhead of creating and operating clusters is at least somewhat taken on by a cloud provider. The benefit of stronger tenant isolation must be evaluated against the cost and complexity of managing multiple clusters. The Multi-cluster SIG is responsible for addressing these types of use cases.
The remainder of this page focuses on isolation techniques used for shared Kubernetes clusters. However, even if you are considering dedicated clusters, it may be valuable to review these recommendations, as it will give you the flexibility to shift to shared clusters in the future if your needs or capabilities change.
Control plane isolation
Control plane isolation ensures that different tenants cannot access or affect each others' Kubernetes API resources.
Namespaces
In Kubernetes, a Namespace provides a mechanism for isolating groups of API resources within a single cluster. This isolation has two key dimensions:
Object names within a namespace can overlap with names in other namespaces, similar to files in folders. This allows tenants to name their resources without having to consider what other tenants are doing.
Many Kubernetes security policies are scoped to namespaces. For example, RBAC Roles and Network Policies are namespace-scoped resources. Using RBAC, Users and Service Accounts can be restricted to a namespace.
In a multi-tenant environment, a Namespace helps segment a tenant's workload into a logical and distinct management unit. In fact, a common practice is to isolate every workload in its own namespace, even if multiple workloads are operated by the same tenant. This ensures that each workload has its own identity and can be configured with an appropriate security policy.
The namespace isolation model requires configuration of several other Kubernetes resources, networking plugins, and adherence to security best practices to properly isolate tenant workloads. These considerations are discussed below.
Access controls
The most important type of isolation for the control plane is authorization. If teams or their workloads can access or modify each others' API resources, they can change or disable all other types of policies thereby negating any protection those policies may offer. As a result, it is critical to ensure that each tenant has the appropriate access to only the namespaces they need, and no more. This is known as the "Principle of Least Privilege."
Role-based access control (RBAC) is commonly used to enforce authorization in the Kubernetes control plane, for both users and workloads (service accounts). Roles and role bindings are Kubernetes objects that are used at a namespace level to enforce access control in your application; similar objects exist for authorizing access to cluster-level objects, though these are less useful for multi-tenant clusters.
In a multi-team environment, RBAC must be used to restrict tenants' access to the appropriate namespaces, and ensure that cluster-wide resources can only be accessed or modified by privileged users such as cluster administrators.
If a policy ends up granting a user more permissions than they need, this is likely a signal that the namespace containing the affected resources should be refactored into finer-grained namespaces. Namespace management tools may simplify the management of these finer-grained namespaces by applying common RBAC policies to different namespaces, while still allowing fine-grained policies where necessary.
Quotas
Kubernetes workloads consume node resources, like CPU and memory. In a multi-tenant environment, you can use
Resource Quotas to manage resource usage of tenant workloads.
For the multiple teams use case, where tenants have access to the Kubernetes API, you can use resource quotas
to limit the number of API resources (for example: the number of Pods, or the number of ConfigMaps)
that a tenant can create. Limits on object count ensure fairness and aim to avoid noisy neighbor issues from
affecting other tenants that share a control plane.
Resource quotas are namespaced objects. By mapping tenants to namespaces, cluster admins can use quotas to ensure that a tenant cannot monopolize a cluster's resources or overwhelm its control plane. Namespace management tools simplify the administration of quotas. In addition, while Kubernetes quotas only apply within a single namespace, some namespace management tools allow groups of namespaces to share quotas, giving administrators far more flexibility with less effort than built-in quotas.
Quotas prevent a single tenant from consuming greater than their allocated share of resources hence minimizing the “noisy neighbor” issue, where one tenant negatively impacts the performance of other tenants' workloads.
When you apply a quota to namespace, Kubernetes requires you to also specify resource requests and limits for each container. Limits are the upper bound for the amount of resources that a container can consume. Containers that attempt to consume resources that exceed the configured limits will either be throttled or killed, based on the resource type. When resource requests are set lower than limits, each container is guaranteed the requested amount but there may still be some potential for impact across workloads.
Quotas cannot protect against all kinds of resource sharing, such as network traffic. Node isolation (described below) may be a better solution for this problem.
Data Plane Isolation
Data plane isolation ensures that pods and workloads for different tenants are sufficiently isolated.
Network isolation
By default, all pods in a Kubernetes cluster are allowed to communicate with each other, and all network traffic is unencrypted. This can lead to security vulnerabilities where traffic is accidentally or maliciously sent to an unintended destination, or is intercepted by a workload on a compromised node.
Pod-to-pod communication can be controlled using Network Policies, which restrict communication between pods using namespace labels or IP address ranges. In a multi-tenant environment where strict network isolation between tenants is required, starting with a default policy that denies communication between pods is recommended with another rule that allows all pods to query the DNS server for name resolution. With such a default policy in place, you can begin adding more permissive rules that allow for communication within a namespace. This scheme can be further refined as required. Note that this only applies to pods within a single control plane; pods that belong to different virtual control planes cannot talk to each other via Kubernetes networking.
Namespace management tools may simplify the creation of default or common network policies. In addition, some of these tools allow you to enforce a consistent set of namespace labels across your cluster, ensuring that they are a trusted basis for your policies.
Warning: Network policies require a CNI plugin that supports the implementation of network policies. Otherwise, NetworkPolicy resources will be ignored.
More advanced network isolation may be provided by service meshes, which provide OSI Layer 7 policies based on workload identity, in addition to namespaces. These higher-level policies can make it easier to manage namespace-based multi-tenancy, especially when multiple namespaces are dedicated to a single tenant. They frequently also offer encryption using mutual TLS, protecting your data even in the presence of a compromised node, and work across dedicated or virtual clusters. However, they can be significantly more complex to manage and may not be appropriate for all users.
Storage isolation
Kubernetes offers several types of volumes that can be used as persistent storage for workloads. For security and data-isolation, dynamic volume provisioning is recommended and volume types that use node resources should be avoided.
StorageClasses allow you to describe custom "classes" of storage offered by your cluster, based on quality-of-service levels, backup policies, or custom policies determined by the cluster administrators.
Pods can request storage using a PersistentVolumeClaim. A PersistentVolumeClaim is a namespaced resource, which enables isolating portions of the storage system and dedicating it to tenants within the shared Kubernetes cluster. However, it is important to note that a PersistentVolume is a cluster-wide resource and has a lifecycle independent of workloads and namespaces.
For example, you can configure a separate StorageClass for each tenant and use this to strengthen isolation.
If a StorageClass is shared, you should set a reclaim policy of Delete
to ensure that a PersistentVolume cannot be reused across different namespaces.
Sandboxing containers
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Kubernetes pods are composed of one or more containers that execute on worker nodes. Containers utilize OS-level virtualization and hence offer a weaker isolation boundary than virtual machines that utilize hardware-based virtualization.
In a shared environment, unpatched vulnerabilities in the application and system layers can be exploited by attackers for container breakouts and remote code execution that allow access to host resources. In some applications, like a Content Management System (CMS), customers may be allowed the ability to upload and execute untrusted scripts or code. In either case, mechanisms to further isolate and protect workloads using strong isolation are desirable.
Sandboxing provides a way to isolate workloads running in a shared cluster. It typically involves running each pod in a separate execution environment such as a virtual machine or a userspace kernel. Sandboxing is often recommended when you are running untrusted code, where workloads are assumed to be malicious. Part of the reason this type of isolation is necessary is because containers are processes running on a shared kernel; they mount file systems like /sys and /proc from the underlying host, making them less secure than an application that runs on a virtual machine which has its own kernel. While controls such as seccomp, AppArmor, and SELinux can be used to strengthen the security of containers, it is hard to apply a universal set of rules to all workloads running in a shared cluster. Running workloads in a sandbox environment helps to insulate the host from container escapes, where an attacker exploits a vulnerability to gain access to the host system and all the processes/files running on that host.
Virtual machines and userspace kernels are 2 popular approaches to sandboxing. The following sandboxing implementations are available:
gVisor intercepts syscalls from containers and runs them through a userspace kernel, written in Go, with limited access to the underlying host.
Kata Containers is an OCI compliant runtime that allows you to run containers in a VM. The hardware virtualization available in Kata offers an added layer of security for containers running untrusted code.
Node Isolation
Node isolation is another technique that you can use to isolate tenant workloads from each other. With node isolation, a set of nodes is dedicated to running pods from a particular tenant and co-mingling of tenant pods is prohibited. This configuration reduces the noisy tenant issue, as all pods running on a node will belong to a single tenant. The risk of information disclosure is slightly lower with node isolation because an attacker that manages to escape from a container will only have access to the containers and volumes mounted to that node.
Although workloads from different tenants are running on different nodes, it is important to be aware that the kubelet and (unless using virtual control planes) the API service are still shared services. A skilled attacker could use the permissions assigned to the kubelet or other pods running on the node to move laterally within the cluster and gain access to tenant workloads running on other nodes. If this is a major concern, consider implementing compensating controls such as seccomp, AppArmor or SELinux or explore using sandboxed containers or creating separate clusters for each tenant.
Node isolation is a little easier to reason about from a billing standpoint than sandboxing containers since you can charge back per node rather than per pod. It also has fewer compatibility and performance issues and may be easier to implement than sandboxing containers. For example, nodes for each tenant can be configured with taints so that only pods with the corresponding toleration can run on them. A mutating webhook could then be used to automatically add tolerations and node affinities to pods deployed into tenant namespaces so that they run on a specific set of nodes designated for that tenant.
This section discusses other Kubernetes constructs and patterns that are relevant for multi-tenancy.
API Priority and Fairness
API priority and fairness is a Kubernetes feature that allows you to assign a priority to certain pods running within the cluster. When an application calls the Kubernetes API, the API server evaluates the priority assigned to pod. Calls from pods with higher priority are fulfilled before those with a lower priority. When contention is high, lower priority calls can be queued until the server is less busy or you can reject the requests.
Using API priority and fairness will not be very common in SaaS environments unless you are allowing customers to run applications that interface with the Kubernetes API, e.g. a controller.
Quality-of-Service (QoS)
When you’re running a SaaS application, you may want the ability to offer different Quality-of-Service (QoS) tiers of service to different tenants. For example, you may have freemium service that comes with fewer performance guarantees and features and a for-fee service tier with specific performance guarantees. Fortunately, there are several Kubernetes constructs that can help you accomplish this within a shared cluster, including network QoS, storage classes, and pod priority and preemption. The idea with each of these is to provide tenants with the quality of service that they paid for. Let’s start by looking at networking QoS.
Typically, all pods on a node share a network interface. Without network QoS, some pods may consume an unfair share of the available bandwidth at the expense of other pods. The Kubernetes bandwidth plugin creates an extended resource for networking that allows you to use Kubernetes resources constructs, i.e. requests/limits, to apply rate limits to pods by using Linux tc queues. Be aware that the plugin is considered experimental as per the Network Plugins documentation and should be thoroughly tested before use in production environments.
For storage QoS, you will likely want to create different storage classes or profiles with different performance characteristics. Each storage profile can be associated with a different tier of service that is optimized for different workloads such IO, redundancy, or throughput. Additional logic might be necessary to allow the tenant to associate the appropriate storage profile with their workload.
Finally, there’s pod priority and preemption where you can assign priority values to pods. When scheduling pods, the scheduler will try evicting pods with lower priority when there are insufficient resources to schedule pods that are assigned a higher priority. If you have a use case where tenants have different service tiers in a shared cluster e.g. free and paid, you may want to give higher priority to certain tiers using this feature.
DNS
Kubernetes clusters include a Domain Name System (DNS) service to provide translations from names to IP addresses, for all Services and Pods. By default, the Kubernetes DNS service allows lookups across all namespaces in the cluster.
In multi-tenant environments where tenants can access pods and other Kubernetes resources, or where
stronger isolation is required, it may be necessary to prevent pods from looking up services in other
Namespaces.
You can restrict cross-namespace DNS lookups by configuring security rules for the DNS service.
For example, CoreDNS (the default DNS service for Kubernetes) can leverage Kubernetes metadata
to restrict queries to Pods and Services within a namespace. For more information, read an
example of configuring
this within the CoreDNS documentation.
Operators are Kubernetes controllers that manage applications. Operators can simplify the management of multiple instances of an application, like a database service, which makes them a common building block in the multi-consumer (SaaS) multi-tenancy use case.
Operators used in a multi-tenant environment should follow a stricter set of guidelines. Specifically, the Operator should:
Support creating resources within different tenant namespaces, rather than just in the namespace in which the Operator is deployed.
Ensure that the Pods are configured with resource requests and limits, to ensure scheduling and fairness.
Support configuration of Pods for data-plane isolation techniques such as node isolation and sandboxed containers.
Implementations
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
There are two primary ways to share a Kubernetes cluster for multi-tenancy: using Namespaces (i.e. a Namespace per tenant) or by virtualizing the control plane (i.e. Virtual control plane per tenant).
In both cases, data plane isolation, and management of additional considerations such as API Priority and Fairness, is also recommended.
Namespace isolation is well-supported by Kubernetes, has a negligible resource cost, and provides mechanisms to allow tenants to interact appropriately, such as by allowing service-to-service communication. However, it can be difficult to configure, and doesn't apply to Kubernetes resources that can't be namespaced, such as Custom Resource Definitions, Storage Classes, and Webhooks.
Control plane virtualization allows for isolation of non-namespaced resources at the cost of somewhat higher resource usage and more difficult cross-tenant sharing. It is a good option when namespace isolation is insufficient but dedicated clusters are undesirable, due to the high cost of maintaining them (especially on-prem) or due to their higher overhead and lack of resource sharing. However, even within a virtualized control plane, you will likely see benefits by using namespaces as well.
The two options are discussed in more detail in the following sections:
Namespace per tenant
As previously mentioned, you should consider isolating each workload in its own namespace, even if you are using dedicated clusters or virtualized control planes. This ensures that each workload only has access to its own resources, such as Config Maps and Secrets, and allows you to tailor dedicated security policies for each workload. In addition, it is a best practice to give each namespace names that are unique across your entire fleet (i.e., even if they are in separate clusters), as this gives you the flexibility to switch between dedicated and shared clusters in the future, or to use multi-cluster tooling such as service meshes.
Conversely, there are also advantages to assigning namespaces at the tenant level, not just the workload level, since there are often policies that apply to all workloads owned by a single tenant. However, this raises its own problems. Firstly, this makes it difficult or impossible to customize policies to individual workloads, and secondly, it may be challenging to come up with a single level of "tenancy" that should be given a namespace. For example, an organization may have divisions, teams, and subteams - which should be assigned a namespace?
To solve this, Kubernetes provides the Hierarchical Namespace Controller (HNC), which allows you to organize your namespaces into hierarchies, and share certain policies and resources between them. It also helps you manage namespace labels, namespace lifecycles, and delegated management, and share resource quotas across related namespaces. These capabilities can be useful in both multi-team and multi-customer scenarios.
Other projects that provide similar capabilities and aid in managing namespaced resources are listed below:
Another form of control-plane isolation is to use Kubernetes extensions to provide each tenant a virtual control-plane that enables segmentation of cluster-wide API resources. Data plane isolation techniques can be used with this model to securely manage worker nodes across tenants.
The virtual control plane based multi-tenancy model extends namespace-based multi-tenancy by providing each tenant with dedicated control plane components, and hence complete control over cluster-wide resources and add-on services. Worker nodes are shared across all tenants, and are managed by a Kubernetes cluster that is normally inaccessible to tenants. This cluster is often referred to as a super-cluster (or sometimes as a host-cluster). Since a tenant’s control-plane is not directly associated with underlying compute resources it is referred to as a virtual control plane.
A virtual control plane typically consists of the Kubernetes API server, the controller manager, and the etcd data store. It interacts with the super cluster via a metadata synchronization controller which coordinates changes across tenant control planes and the control plane of the super--cluster.
By using per-tenant dedicated control planes, most of the isolation problems due to sharing one API server among all tenants are solved. Examples include noisy neighbors in the control plane, uncontrollable blast radius of policy misconfigurations, and conflicts between cluster scope objects such as webhooks and CRDs. Hence, the virtual control plane model is particularly suitable for cases where each tenant requires access to a Kubernetes API server and expects the full cluster manageability.
The improved isolation comes at the cost of running and maintaining an individual virtual control plane per tenant. In addition, per-tenant control planes do not solve isolation problems in the data plane, such as node-level noisy neighbors or security threats. These must still be addressed separately.
Policies you can configure that apply to groups of resources.
3.10.1 - Limit Ranges
By default, containers run with unbounded compute resources on a Kubernetes cluster.
With resource quotas, cluster administrators can restrict resource consumption and creation on a namespace basis.
Within a namespace, a Pod or Container can consume as much CPU and memory as defined by the namespace's resource quota. There is a concern that one Pod or Container could monopolize all available resources. A LimitRange is a policy to constrain resource allocations (to Pods or Containers) in a namespace.
A LimitRange provides constraints that can:
Enforce minimum and maximum compute resources usage per Pod or Container in a namespace.
Enforce minimum and maximum storage request per PersistentVolumeClaim in a namespace.
Enforce a ratio between request and limit for a resource in a namespace.
Set default request/limit for compute resources in a namespace and automatically inject them to Containers at runtime.
Enabling LimitRange
LimitRange support has been enabled by default since Kubernetes 1.10.
A LimitRange is enforced in a particular namespace when there is a
LimitRange object in that namespace.
The administrator creates one LimitRange in one namespace.
Users create resources like Pods, Containers, and PersistentVolumeClaims in the namespace.
The LimitRanger admission controller enforces defaults and limits for all Pods and Containers that do not set compute resource requirements and tracks usage to ensure it does not exceed resource minimum, maximum and ratio defined in any LimitRange present in the namespace.
If creating or updating a resource (Pod, Container, PersistentVolumeClaim) that violates a LimitRange constraint, the request to the API server will fail with an HTTP status code 403 FORBIDDEN and a message explaining the constraint that have been violated.
If a LimitRange is activated in a namespace for compute resources like cpu and memory, users must specify
requests or limits for those values. Otherwise, the system may reject Pod creation.
LimitRange validations occurs only at Pod Admission stage, not on Running Pods.
Examples of policies that could be created using limit range are:
In a 2 node cluster with a capacity of 8 GiB RAM and 16 cores, constrain Pods in a namespace to request 100m of CPU with a max limit of 500m for CPU and request 200Mi for Memory with a max limit of 600Mi for Memory.
Define default CPU limit and request to 150m and memory default request to 300Mi for Containers started with no cpu and memory requests in their specs.
In the case where the total limits of the namespace is less than the sum of the limits of the Pods/Containers,
there may be contention for resources. In this case, the Containers or Pods will not be created.
Neither contention nor changes to a LimitRange will affect already created resources.
When several users or teams share a cluster with a fixed number of nodes,
there is a concern that one team could use more than its fair share of resources.
Resource quotas are a tool for administrators to address this concern.
A resource quota, defined by a ResourceQuota object, provides constraints that limit
aggregate resource consumption per namespace. It can limit the quantity of objects that can
be created in a namespace by type, as well as the total amount of compute resources that may
be consumed by resources in that namespace.
Resource quotas work like this:
Different teams work in different namespaces. This can be enforced with RBAC.
The administrator creates one ResourceQuota for each namespace.
Users create resources (pods, services, etc.) in the namespace, and the quota system
tracks usage to ensure it does not exceed hard resource limits defined in a ResourceQuota.
If creating or updating a resource violates a quota constraint, the request will fail with HTTP
status code 403 FORBIDDEN with a message explaining the constraint that would have been violated.
If quota is enabled in a namespace for compute resources like cpu and memory, users must specify
requests or limits for those values; otherwise, the quota system may reject pod creation. Hint: Use
the LimitRanger admission controller to force defaults for pods that make no compute resource requirements.
See the walkthrough
for an example of how to avoid this problem.
Examples of policies that could be created using namespaces and quotas are:
In a cluster with a capacity of 32 GiB RAM, and 16 cores, let team A use 20 GiB and 10 cores,
let B use 10GiB and 4 cores, and hold 2GiB and 2 cores in reserve for future allocation.
Limit the "testing" namespace to using 1 core and 1GiB RAM. Let the "production" namespace
use any amount.
In the case where the total capacity of the cluster is less than the sum of the quotas of the namespaces,
there may be contention for resources. This is handled on a first-come-first-served basis.
Neither contention nor changes to quota will affect already created resources.
Enabling Resource Quota
Resource Quota support is enabled by default for many Kubernetes distributions. It is
enabled when the API server--enable-admission-plugins= flag has ResourceQuota as
one of its arguments.
A resource quota is enforced in a particular namespace when there is a
ResourceQuota in that namespace.
Compute Resource Quota
You can limit the total sum of
compute resources
that can be requested in a given namespace.
The following resource types are supported:
Resource Name
Description
limits.cpu
Across all pods in a non-terminal state, the sum of CPU limits cannot exceed this value.
limits.memory
Across all pods in a non-terminal state, the sum of memory limits cannot exceed this value.
requests.cpu
Across all pods in a non-terminal state, the sum of CPU requests cannot exceed this value.
requests.memory
Across all pods in a non-terminal state, the sum of memory requests cannot exceed this value.
hugepages-<size>
Across all pods in a non-terminal state, the number of huge page requests of the specified size cannot exceed this value.
cpu
Same as requests.cpu
memory
Same as requests.memory
Resource Quota For Extended Resources
In addition to the resources mentioned above, in release 1.10, quota support for
extended resources is added.
As overcommit is not allowed for extended resources, it makes no sense to specify both requests
and limits for the same extended resource in a quota. So for extended resources, only quota items
with prefix requests. is allowed for now.
Take the GPU resource as an example, if the resource name is nvidia.com/gpu, and you want to
limit the total number of GPUs requested in a namespace to 4, you can define a quota as follows:
Across all persistent volume claims associated with the storage-class-name, the total number of persistent volume claims that can exist in the namespace.
For example, if an operator wants to quota storage with gold storage class separate from bronze storage class, the operator can
define a quota as follows:
In release 1.8, quota support for local ephemeral storage is added as an alpha feature:
Resource Name
Description
requests.ephemeral-storage
Across all pods in the namespace, the sum of local ephemeral storage requests cannot exceed this value.
limits.ephemeral-storage
Across all pods in the namespace, the sum of local ephemeral storage limits cannot exceed this value.
ephemeral-storage
Same as requests.ephemeral-storage.
Note: When using a CRI container runtime, container logs will count against the ephemeral storage quota.
This can result in the unexpected eviction of pods that have exhausted their storage quotas.
Refer to Logging Architecture for details.
Object Count Quota
You can set quota for the total number of certain resources of all standard,
namespaced resource types using the following syntax:
count/<resource>.<group> for resources from non-core groups
count/<resource> for resources from the core group
Here is an example set of resources users may want to put under object count quota:
count/persistentvolumeclaims
count/services
count/secrets
count/configmaps
count/replicationcontrollers
count/deployments.apps
count/replicasets.apps
count/statefulsets.apps
count/jobs.batch
count/cronjobs.batch
The same syntax can be used for custom resources.
For example, to create a quota on a widgets custom resource in the example.com API group, use count/widgets.example.com.
When using count/* resource quota, an object is charged against the quota if it exists in server storage.
These types of quotas are useful to protect against exhaustion of storage resources. For example, you may
want to limit the number of Secrets in a server given their large size. Too many Secrets in a cluster can
actually prevent servers and controllers from starting. You can set a quota for Jobs to protect against
a poorly configured CronJob. CronJobs that create too many Jobs in a namespace can lead to a denial of service.
It is also possible to do generic object count quota on a limited set of resources.
The following types are supported:
Resource Name
Description
configmaps
The total number of ConfigMaps that can exist in the namespace.
The total number of Pods in a non-terminal state that can exist in the namespace. A pod is in a terminal state if .status.phase in (Failed, Succeeded) is true.
replicationcontrollers
The total number of ReplicationControllers that can exist in the namespace.
resourcequotas
The total number of ResourceQuotas that can exist in the namespace.
services
The total number of Services that can exist in the namespace.
services.loadbalancers
The total number of Services of type LoadBalancer that can exist in the namespace.
services.nodeports
The total number of Services of type NodePort that can exist in the namespace.
secrets
The total number of Secrets that can exist in the namespace.
For example, pods quota counts and enforces a maximum on the number of pods
created in a single namespace that are not terminal. You might want to set a pods
quota on a namespace to avoid the case where a user creates many small pods and
exhausts the cluster's supply of Pod IPs.
Quota Scopes
Each quota can have an associated set of scopes. A quota will only measure usage for a resource if it matches
the intersection of enumerated scopes.
When a scope is added to the quota, it limits the number of resources it supports to those that pertain to the scope.
Resources specified on the quota outside of the allowed set results in a validation error.
Scope
Description
Terminating
Match pods where .spec.activeDeadlineSeconds >= 0
NotTerminating
Match pods where .spec.activeDeadlineSeconds is nil
BestEffort
Match pods that have best effort quality of service.
NotBestEffort
Match pods that do not have best effort quality of service.
PriorityClass
Match pods that references the specified priority class.
The BestEffort scope restricts a quota to tracking the following resource:
pods
The Terminating, NotTerminating, NotBestEffort and PriorityClass
scopes restrict a quota to tracking the following resources:
pods
cpu
memory
requests.cpu
requests.memory
limits.cpu
limits.memory
Note that you cannot specify both the Terminating and the NotTerminating
scopes in the same quota, and you cannot specify both the BestEffort and
NotBestEffort scopes in the same quota either.
The scopeSelector supports the following values in the operator field:
In
NotIn
Exists
DoesNotExist
When using one of the following values as the scopeName when defining the
scopeSelector, the operator must be Exists.
Terminating
NotTerminating
BestEffort
NotBestEffort
If the operator is In or NotIn, the values field must have at least
one value. For example:
If the operator is Exists or DoesNotExist, the values field must NOT be
specified.
Resource Quota Per PriorityClass
FEATURE STATE:Kubernetes v1.17 [stable]
Pods can be created at a specific priority.
You can control a pod's consumption of system resources based on a pod's priority, by using the scopeSelector
field in the quota spec.
A quota is matched and consumed only if scopeSelector in the quota spec selects the pod.
When quota is scoped for priority class using scopeSelector field, quota object
is restricted to track only following resources:
pods
cpu
memory
ephemeral-storage
limits.cpu
limits.memory
limits.ephemeral-storage
requests.cpu
requests.memory
requests.ephemeral-storage
This example creates a quota object and matches it with pods at specific priorities. The example
works as follows:
Pods in the cluster have one of the three priority classes, "low", "medium", "high".
resourcequota/pods-high created
resourcequota/pods-medium created
resourcequota/pods-low created
Verify that Used quota is 0 using kubectl describe quota.
kubectl describe quota
Name: pods-high
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 0 1k
memory 0 200Gi
pods 0 10
Name: pods-low
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 0 5
memory 0 10Gi
pods 0 10
Name: pods-medium
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 0 10
memory 0 20Gi
pods 0 10
Create a pod with priority "high". Save the following YAML to a
file high-priority-pod.yml.
apiVersion:v1kind:Podmetadata:name:high-priorityspec:containers:- name:high-priorityimage:ubuntucommand:["/bin/sh"]args:["-c","while true; do echo hello; sleep 10;done"]resources:requests:memory:"10Gi"cpu:"500m"limits:memory:"10Gi"cpu:"500m"priorityClassName:high
Apply it with kubectl create.
kubectl create -f ./high-priority-pod.yml
Verify that "Used" stats for "high" priority quota, pods-high, has changed and that
the other two quotas are unchanged.
kubectl describe quota
Name: pods-high
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 500m 1k
memory 10Gi 200Gi
pods 1 10
Name: pods-low
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 0 5
memory 0 10Gi
pods 0 10
Name: pods-medium
Namespace: default
Resource Used Hard
-------- ---- ----
cpu 0 10
memory 0 20Gi
pods 0 10
Cross-namespace Pod Affinity Quota
FEATURE STATE:Kubernetes v1.24 [stable]
Operators can use CrossNamespacePodAffinity quota scope to limit which namespaces are allowed to
have pods with affinity terms that cross namespaces. Specifically, it controls which pods are allowed
to set namespaces or namespaceSelector fields in pod affinity terms.
Preventing users from using cross-namespace affinity terms might be desired since a pod
with anti-affinity constraints can block pods from all other namespaces
from getting scheduled in a failure domain.
Using this scope operators can prevent certain namespaces (foo-ns in the example below)
from having pods that use cross-namespace pod affinity by creating a resource quota object in
that namespace with CrossNamespaceAffinity scope and hard limit of 0:
If operators want to disallow using namespaces and namespaceSelector by default, and
only allow it for specific namespaces, they could configure CrossNamespaceAffinity
as a limited resource by setting the kube-apiserver flag --admission-control-config-file
to the path of the following configuration file:
With the above configuration, pods can use namespaces and namespaceSelector in pod affinity only
if the namespace where they are created have a resource quota object with
CrossNamespaceAffinity scope and a hard limit greater than or equal to the number of pods using those fields.
Requests compared to Limits
When allocating compute resources, each container may specify a request and a limit value for either CPU or memory.
The quota can be configured to quota either value.
If the quota has a value specified for requests.cpu or requests.memory, then it requires that every incoming
container makes an explicit request for those resources. If the quota has a value specified for limits.cpu or limits.memory,
then it requires that every incoming container specifies an explicit limit for those resources.
Viewing and Setting Quotas
Kubectl supports creating, updating, and viewing quotas:
Name: test
Namespace: myspace
Resource Used Hard
-------- ---- ----
count/deployments.apps 1 2
count/pods 2 3
count/replicasets.apps 1 4
count/secrets 1 4
Quota and Cluster Capacity
ResourceQuotas are independent of the cluster capacity. They are
expressed in absolute units. So, if you add nodes to your cluster, this does not
automatically give each namespace the ability to consume more resources.
Sometimes more complex policies may be desired, such as:
Proportionally divide total cluster resources among several teams.
Allow each tenant to grow resource usage as needed, but have a generous
limit to prevent accidental resource exhaustion.
Detect demand from one namespace, add nodes, and increase quota.
Such policies could be implemented using ResourceQuotas as building blocks, by
writing a "controller" that watches the quota usage and adjusts the quota
hard limits of each namespace according to other signals.
Note that resource quota divides up aggregate cluster resources, but it creates no
restrictions around nodes: pods from several namespaces may run on the same node.
Limit Priority Class consumption by default
It may be desired that pods at a particular priority, eg. "cluster-services",
should be allowed in a namespace, if and only if, a matching quota object exists.
With this mechanism, operators are able to restrict usage of certain high
priority classes to a limited number of namespaces and not every namespace
will be able to consume these priority classes by default.
To enforce this, kube-apiserver flag --admission-control-config-file should be
used to pass path to the following configuration file:
Kubernetes allow you to limit the number of process IDs (PIDs) that a
Pod can use.
You can also reserve a number of allocatable PIDs for each node
for use by the operating system and daemons (rather than by Pods).
Process IDs (PIDs) are a fundamental resource on nodes. It is trivial to hit the
task limit without hitting any other resource limits, which can then cause
instability to a host machine.
Cluster administrators require mechanisms to ensure that Pods running in the
cluster cannot induce PID exhaustion that prevents host daemons (such as the
kubelet or
kube-proxy,
and potentially also the container runtime) from running.
In addition, it is important to ensure that PIDs are limited among Pods in order
to ensure they have limited impact on other workloads on the same node.
Note: On certain Linux installations, the operating system sets the PIDs limit to a low default,
such as 32768. Consider raising the value of /proc/sys/kernel/pid_max.
You can configure a kubelet to limit the number of PIDs a given Pod can consume.
For example, if your node's host OS is set to use a maximum of 262144 PIDs and
expect to host less than 250 Pods, one can give each Pod a budget of 1000
PIDs to prevent using up that node's overall number of available PIDs. If the
admin wants to overcommit PIDs similar to CPU or memory, they may do so as well
with some additional risks. Either way, a single Pod will not be able to bring
the whole machine down. This kind of resource limiting helps to prevent simple
fork bombs from affecting operation of an entire cluster.
Per-Pod PID limiting allows administrators to protect one Pod from another, but
does not ensure that all Pods scheduled onto that host are unable to impact the node overall.
Per-Pod limiting also does not protect the node agents themselves from PID exhaustion.
You can also reserve an amount of PIDs for node overhead, separate from the
allocation to Pods. This is similar to how you can reserve CPU, memory, or other
resources for use by the operating system and other facilities outside of Pods
and their containers.
PID limiting is a an important sibling to compute
resource requests
and limits. However, you specify it in a different way: rather than defining a
Pod's resource limit in the .spec for a Pod, you configure the limit as a
setting on the kubelet. Pod-defined PID limits are not currently supported.
Caution: This means that the limit that applies to a Pod may be different depending on
where the Pod is scheduled. To make things simple, it's easiest if all Nodes use
the same PID resource limits and reservations.
Node PID limits
Kubernetes allows you to reserve a number of process IDs for the system use. To
configure the reservation, use the parameter pid=<number> in the
--system-reserved and --kube-reserved command line options to the kubelet.
The value you specified declares that the specified number of process IDs will
be reserved for the system as a whole and for Kubernetes system daemons
respectively.
Note: Before Kubernetes version 1.20, PID resource limiting with Node-level
reservations required enabling the feature
gateSupportNodePidsLimit to work.
Pod PID limits
Kubernetes allows you to limit the number of processes running in a Pod. You
specify this limit at the node level, rather than configuring it as a resource
limit for a particular Pod. Each Node can have a different PID limit.
To configure the limit, you can specify the command line parameter --pod-max-pids
to the kubelet, or set PodPidsLimit in the kubelet
configuration file.
Note: Before Kubernetes version 1.20, PID resource limiting for Pods required enabling
the feature gateSupportPodPidsLimit to work.
PID based eviction
You can configure kubelet to start terminating a Pod when it is misbehaving and consuming abnormal amount of resources.
This feature is called eviction. You can
Configure Out of Resource Handling
for various eviction signals.
Use pid.available eviction signal to configure the threshold for number of PIDs used by Pod.
You can set soft and hard eviction policies.
However, even with the hard eviction policy, if the number of PIDs growing very fast,
node can still get into unstable state by hitting the node PIDs limit.
Eviction signal value is calculated periodically and does NOT enforce the limit.
PID limiting - per Pod and per Node sets the hard limit.
Once the limit is hit, workload will start experiencing failures when trying to get a new PID.
It may or may not lead to rescheduling of a Pod,
depending on how workload reacts on these failures and how liveleness and readiness
probes are configured for the Pod. However, if limits were set correctly,
you can guarantee that other Pods workload and system processes will not run out of PIDs
when one Pod is misbehaving.
In order to support latency-critical and high-throughput workloads, Kubernetes offers a suite of Resource Managers. The managers aim to co-ordinate and optimise node's resources alignment for pods configured with a specific requirement for CPUs, devices, and memory (hugepages) resources.
The main manager, the Topology Manager, is a Kubelet component that co-ordinates the overall resource management process through its policy.
The configuration of individual managers is elaborated in dedicated documents:
In Kubernetes, scheduling refers to making sure that Pods are matched to Nodes so that the kubelet can run them. Preemption is the process of terminating Pods with lower Priority so that Pods with higher Priority can schedule on Nodes. Eviction is the process of proactively terminating one or more Pods on resource-starved Nodes.
In Kubernetes, scheduling refers to making sure that Pods
are matched to Nodes so that the
kubelet can run them. Preemption
is the process of terminating Pods with lower Priority
so that Pods with higher Priority can schedule on Nodes. Eviction is the process
of terminating one or more Pods on Nodes.
Pod disruption is the process by which
Pods on Nodes are terminated either voluntarily or involuntarily.
Voluntary disruptions are started intentionally by application owners or cluster
administrators. Involuntary disruptions are unintentional and can be triggered by
unavoidable issues like Nodes running out of resources, or by accidental deletions.
In Kubernetes, scheduling refers to making sure that Pods
are matched to Nodes so that
Kubelet can run them.
Scheduling overview
A scheduler watches for newly created Pods that have no Node assigned. For
every Pod that the scheduler discovers, the scheduler becomes responsible
for finding the best Node for that Pod to run on. The scheduler reaches
this placement decision taking into account the scheduling principles
described below.
If you want to understand why Pods are placed onto a particular Node,
or if you're planning to implement a custom scheduler yourself, this
page will help you learn about scheduling.
kube-scheduler
kube-scheduler
is the default scheduler for Kubernetes and runs as part of the
control plane.
kube-scheduler is designed so that, if you want and need to, you can
write your own scheduling component and use that instead.
For every newly created pod or other unscheduled pods, kube-scheduler
selects an optimal node for them to run on. However, every container in
pods has different requirements for resources and every pod also has
different requirements. Therefore, existing nodes need to be filtered
according to the specific scheduling requirements.
In a cluster, Nodes that meet the scheduling requirements for a Pod
are called feasible nodes. If none of the nodes are suitable, the pod
remains unscheduled until the scheduler is able to place it.
The scheduler finds feasible Nodes for a Pod and then runs a set of
functions to score the feasible Nodes and picks a Node with the highest
score among the feasible ones to run the Pod. The scheduler then notifies
the API server about this decision in a process called binding.
Factors that need to be taken into account for scheduling decisions include
individual and collective resource requirements, hardware / software /
policy constraints, affinity and anti-affinity specifications, data
locality, inter-workload interference, and so on.
Node selection in kube-scheduler
kube-scheduler selects a node for the pod in a 2-step operation:
Filtering
Scoring
The filtering step finds the set of Nodes where it's feasible to
schedule the Pod. For example, the PodFitsResources filter checks whether a
candidate Node has enough available resource to meet a Pod's specific
resource requests. After this step, the node list contains any suitable
Nodes; often, there will be more than one. If the list is empty, that
Pod isn't (yet) schedulable.
In the scoring step, the scheduler ranks the remaining nodes to choose
the most suitable Pod placement. The scheduler assigns a score to each Node
that survived filtering, basing this score on the active scoring rules.
Finally, kube-scheduler assigns the Pod to the Node with the highest ranking.
If there is more than one node with equal scores, kube-scheduler selects
one of these at random.
There are two supported ways to configure the filtering and scoring behavior
of the scheduler:
Scheduling Policies allow you to configure Predicates for filtering and Priorities for scoring.
Scheduling Profiles allow you to configure Plugins that implement different scheduling stages, including: QueueSort, Filter, Score, Bind, Reserve, Permit, and others. You can also configure the kube-scheduler to run different profiles.
You can constrain a Pod so that it is
restricted to run on particular node(s),
or to prefer to run on particular nodes.
There are several ways to do this and the recommended approaches all use
label selectors to facilitate the selection.
Often, you do not need to set any such constraints; the
scheduler will automatically do a reasonable placement
(for example, spreading your Pods across nodes so as not place Pods on a node with insufficient free resources).
However, there are some circumstances where you may want to control which node
the Pod deploys to, for example, to ensure that a Pod ends up on a node with an SSD attached to it,
or to co-locate Pods from two different services that communicate a lot into the same availability zone.
You can use any of the following methods to choose where Kubernetes schedules
specific Pods:
Note: The value of these labels is cloud provider specific and is not guaranteed to be reliable.
For example, the value of kubernetes.io/hostname may be the same as the node name in some environments
and a different value in other environments.
Node isolation/restriction
Adding labels to nodes allows you to target Pods for scheduling on specific
nodes or groups of nodes. You can use this functionality to ensure that specific
Pods only run on nodes with certain isolation, security, or regulatory
properties.
If you use labels for node isolation, choose label keys that the kubelet
cannot modify. This prevents a compromised node from setting those labels on
itself so that the scheduler schedules workloads onto the compromised node.
The NodeRestriction admission plugin
prevents the kubelet from setting or modifying labels with a
node-restriction.kubernetes.io/ prefix.
To make use of that label prefix for node isolation:
Ensure you are using the Node authorizer and have enabled the NodeRestriction admission plugin.
Add labels with the node-restriction.kubernetes.io/ prefix to your nodes, and use those labels in your node selectors.
For example, example.com.node-restriction.kubernetes.io/fips=true or example.com.node-restriction.kubernetes.io/pci-dss=true.
nodeSelector
nodeSelector is the simplest recommended form of node selection constraint.
You can add the nodeSelector field to your Pod specification and specify the
node labels you want the target node to have.
Kubernetes only schedules the Pod onto nodes that have each of the labels you
specify.
nodeSelector is the simplest way to constrain Pods to nodes with specific
labels. Affinity and anti-affinity expands the types of constraints you can
define. Some of the benefits of affinity and anti-affinity include:
The affinity/anti-affinity language is more expressive. nodeSelector only
selects nodes with all the specified labels. Affinity/anti-affinity gives you
more control over the selection logic.
You can indicate that a rule is soft or preferred, so that the scheduler
still schedules the Pod even if it can't find a matching node.
You can constrain a Pod using labels on other Pods running on the node (or other topological domain),
instead of just node labels, which allows you to define rules for which Pods
can be co-located on a node.
The affinity feature consists of two types of affinity:
Node affinity functions like the nodeSelector field but is more expressive and
allows you to specify soft rules.
Inter-pod affinity/anti-affinity allows you to constrain Pods against labels
on other Pods.
Node affinity
Node affinity is conceptually similar to nodeSelector, allowing you to constrain which nodes your
Pod can be scheduled on based on node labels. There are two types of node
affinity:
requiredDuringSchedulingIgnoredDuringExecution: The scheduler can't
schedule the Pod unless the rule is met. This functions like nodeSelector,
but with a more expressive syntax.
preferredDuringSchedulingIgnoredDuringExecution: The scheduler tries to
find a node that meets the rule. If a matching node is not available, the
scheduler still schedules the Pod.
Note: In the preceding types, IgnoredDuringExecution means that if the node labels
change after Kubernetes schedules the Pod, the Pod continues to run.
You can specify node affinities using the .spec.affinity.nodeAffinity field in
your Pod spec.
The node must have a label with the key topology.kubernetes.io/zone and
the value of that label must be either antarctica-east1 or antarctica-west1.
The node preferably has a label with the key another-node-label-key and
the value another-node-label-value.
You can use the operator field to specify a logical operator for Kubernetes to use when
interpreting the rules. You can use In, NotIn, Exists, DoesNotExist,
Gt and Lt.
NotIn and DoesNotExist allow you to define node anti-affinity behavior.
Alternatively, you can use node taints
to repel Pods from specific nodes.
Note:
If you specify both nodeSelector and nodeAffinity, both must be satisfied
for the Pod to be scheduled onto a node.
If you specify multiple nodeSelectorTerms associated with nodeAffinity
types, then the Pod can be scheduled onto a node if one of the specified nodeSelectorTerms can be
satisfied.
If you specify multiple matchExpressions associated with a single nodeSelectorTerms,
then the Pod can be scheduled onto a node only if all the matchExpressions are
satisfied.
You can specify a weight between 1 and 100 for each instance of the
preferredDuringSchedulingIgnoredDuringExecution affinity type. When the
scheduler finds nodes that meet all the other scheduling requirements of the Pod, the
scheduler iterates through every preferred rule that the node satisfies and adds the
value of the weight for that expression to a sum.
The final sum is added to the score of other priority functions for the node.
Nodes with the highest total score are prioritized when the scheduler makes a
scheduling decision for the Pod.
If there are two possible nodes that match the
preferredDuringSchedulingIgnoredDuringExecution rule, one with the
label-1:key-1 label and another with the label-2:key-2 label, the scheduler
considers the weight of each node and adds the weight to the other scores for
that node, and schedules the Pod onto the node with the highest final score.
Note: If you want Kubernetes to successfully schedule the Pods in this example, you
must have existing nodes with the kubernetes.io/os=linux label.
Node affinity per scheduling profile
FEATURE STATE:Kubernetes v1.20 [beta]
When configuring multiple scheduling profiles, you can associate
a profile with a node affinity, which is useful if a profile only applies to a specific set of nodes.
To do so, add an addedAffinity to the args field of the NodeAffinity plugin
in the scheduler configuration. For example:
The addedAffinity is applied to all Pods that set .spec.schedulerName to foo-scheduler, in addition to the
NodeAffinity specified in the PodSpec.
That is, in order to match the Pod, nodes need to satisfy addedAffinity and
the Pod's .spec.NodeAffinity.
Since the addedAffinity is not visible to end users, its behavior might be
unexpected to them. Use node labels that have a clear correlation to the
scheduler profile name.
Note: The DaemonSet controller, which creates Pods for DaemonSets,
does not support scheduling profiles. When the DaemonSet controller creates
Pods, the default Kubernetes scheduler places those Pods and honors any
nodeAffinity rules in the DaemonSet controller.
Inter-pod affinity and anti-affinity
Inter-pod affinity and anti-affinity allow you to constrain which nodes your
Pods can be scheduled on based on the labels of Pods already running on that
node, instead of the node labels.
Inter-pod affinity and anti-affinity rules take the form "this
Pod should (or, in the case of anti-affinity, should not) run in an X if that X
is already running one or more Pods that meet rule Y", where X is a topology
domain like node, rack, cloud provider zone or region, or similar and Y is the
rule Kubernetes tries to satisfy.
You express these rules (Y) as label selectors
with an optional associated list of namespaces. Pods are namespaced objects in
Kubernetes, so Pod labels also implicitly have namespaces. Any label selectors
for Pod labels should specify the namespaces in which Kubernetes should look for those
labels.
You express the topology domain (X) using a topologyKey, which is the key for
the node label that the system uses to denote the domain. For examples, see
Well-Known Labels, Annotations and Taints.
Note: Inter-pod affinity and anti-affinity require substantial amount of
processing which can slow down scheduling in large clusters significantly. We do
not recommend using them in clusters larger than several hundred nodes.
Note: Pod anti-affinity requires nodes to be consistently labelled, in other words,
every node in the cluster must have an appropriate label matching topologyKey.
If some or all nodes are missing the specified topologyKey label, it can lead
to unintended behavior.
Types of inter-pod affinity and anti-affinity
Similar to node affinity are two types of Pod affinity and
anti-affinity as follows:
requiredDuringSchedulingIgnoredDuringExecution
preferredDuringSchedulingIgnoredDuringExecution
For example, you could use
requiredDuringSchedulingIgnoredDuringExecution affinity to tell the scheduler to
co-locate Pods of two services in the same cloud provider zone because they
communicate with each other a lot. Similarly, you could use
preferredDuringSchedulingIgnoredDuringExecution anti-affinity to spread Pods
from a service across multiple cloud provider zones.
To use inter-pod affinity, use the affinity.podAffinity field in the Pod spec.
For inter-pod anti-affinity, use the affinity.podAntiAffinity field in the Pod
spec.
This example defines one Pod affinity rule and one Pod anti-affinity rule. The
Pod affinity rule uses the "hard"
requiredDuringSchedulingIgnoredDuringExecution, while the anti-affinity rule
uses the "soft" preferredDuringSchedulingIgnoredDuringExecution.
The affinity rule says that the scheduler can only schedule a Pod onto a node if
the node is in the same zone as one or more existing Pods with the label
security=S1. More precisely, the scheduler must place the Pod on a node that has the
topology.kubernetes.io/zone=V label, as long as there is at least one node in
that zone that currently has one or more Pods with the Pod label security=S1.
The anti-affinity rule says that the scheduler should try to avoid scheduling
the Pod onto a node that is in the same zone as one or more Pods with the label
security=S2. More precisely, the scheduler should try to avoid placing the Pod on a node that has the
topology.kubernetes.io/zone=R label if there are other nodes in the
same zone currently running Pods with the Security=S2 Pod label.
To get yourself more familiar with the examples of Pod affinity and anti-affinity,
refer to the design proposal.
You can use the In, NotIn, Exists and DoesNotExist values in the
operator field for Pod affinity and anti-affinity.
In principle, the topologyKey can be any allowed label key with the following
exceptions for performance and security reasons:
For Pod affinity and anti-affinity, an empty topologyKey field is not allowed in both requiredDuringSchedulingIgnoredDuringExecution
and preferredDuringSchedulingIgnoredDuringExecution.
For requiredDuringSchedulingIgnoredDuringExecution Pod anti-affinity rules,
the admission controller LimitPodHardAntiAffinityTopology limits
topologyKey to kubernetes.io/hostname. You can modify or disable the
admission controller if you want to allow custom topologies.
In addition to labelSelector and topologyKey, you can optionally specify a list
of namespaces which the labelSelector should match against using the
namespaces field at the same level as labelSelector and topologyKey.
If omitted or empty, namespaces defaults to the namespace of the Pod where the
affinity/anti-affinity definition appears.
Namespace selector
FEATURE STATE:Kubernetes v1.24 [stable]
You can also select matching namespaces using namespaceSelector, which is a label query over the set of namespaces.
The affinity term is applied to namespaces selected by both namespaceSelector and the namespaces field.
Note that an empty namespaceSelector ({}) matches all namespaces, while a null or empty namespaces list and
null namespaceSelector matches the namespace of the Pod where the rule is defined.
More practical use-cases
Inter-pod affinity and anti-affinity can be even more useful when they are used with higher
level collections such as ReplicaSets, StatefulSets, Deployments, etc. These
rules allow you to configure that a set of workloads should
be co-located in the same defined topology; for example, preferring to place two related
Pods onto the same node.
For example: imagine a three-node cluster. You use the cluster to run a web application
and also an in-memory cache (such as Redis). For this example, also assume that latency between
the web application and the memory cache should be as low as is practical. You could use inter-pod
affinity and anti-affinity to co-locate the web servers with the cache as much as possible.
In the following example Deployment for the Redis cache, the replicas get the label app=store. The
podAntiAffinity rule tells the scheduler to avoid placing multiple replicas
with the app=store label on a single node. This creates each cache in a
separate node.
The following example Deployment for the web servers creates replicas with the label app=web-store.
The Pod affinity rule tells the scheduler to place each replica on a node that has a Pod
with the label app=store. The Pod anti-affinity rule tells the scheduler never to place
multiple app=web-store servers on a single node.
Creating the two preceding Deployments results in the following cluster layout,
where each web server is co-located with a cache, on three separate nodes.
node-1
node-2
node-3
webserver-1
webserver-2
webserver-3
cache-1
cache-2
cache-3
The overall effect is that each cache instance is likely to be accessed by a single client, that
is running on the same node. This approach aims to minimize both skew (imbalanced load) and latency.
You might have other reasons to use Pod anti-affinity.
See the ZooKeeper tutorial
for an example of a StatefulSet configured with anti-affinity for high
availability, using the same technique as this example.
nodeName
nodeName is a more direct form of node selection than affinity or
nodeSelector. nodeName is a field in the Pod spec. If the nodeName field
is not empty, the scheduler ignores the Pod and the kubelet on the named node
tries to place the Pod on that node. Using nodeName overrules using
nodeSelector or affinity and anti-affinity rules.
Some of the limitations of using nodeName to select nodes are:
If the named node does not exist, the Pod will not run, and in
some cases may be automatically deleted.
If the named node does not have the resources to accommodate the
Pod, the Pod will fail and its reason will indicate why,
for example OutOfmemory or OutOfcpu.
Node names in cloud environments are not always predictable or
stable.
Here is an example of a Pod spec using the nodeName field:
You can use topology spread constraints to control how Pods
are spread across your cluster among failure-domains such as regions, zones, nodes, or among any other
topology domains that you define. You might do this to improve performance, expected availability, or
overall utilization.
When you run a Pod on a Node, the Pod itself takes an amount of system resources. These
resources are additional to the resources needed to run the container(s) inside the Pod.
In Kubernetes, Pod Overhead is a way to account for the resources consumed by the Pod
infrastructure on top of the container requests & limits.
In Kubernetes, the Pod's overhead is set at
admission
time according to the overhead associated with the Pod's
RuntimeClass.
A pod's overhead is considered in addition to the sum of container resource requests when
scheduling a Pod. Similarly, the kubelet will include the Pod overhead when sizing the Pod cgroup,
and when carrying out Pod eviction ranking.
Configuring Pod overhead
You need to make sure a RuntimeClass is utilized which defines the overhead field.
Usage example
To work with Pod overhead, you need a RuntimeClass that defines the overhead field. As
an example, you could use the following RuntimeClass definition with a virtualization container
runtime that uses around 120MiB per Pod for the virtual machine and the guest OS:
Workloads which are created which specify the kata-fc RuntimeClass handler will take the memory and
cpu overheads into account for resource quota calculations, node scheduling, as well as Pod cgroup sizing.
Consider running the given example workload, test-pod:
At admission time the RuntimeClass admission controller
updates the workload's PodSpec to include the overhead as described in the RuntimeClass. If the PodSpec already has this field defined,
the Pod will be rejected. In the given example, since only the RuntimeClass name is specified, the admission controller mutates the Pod
to include an overhead.
After the RuntimeClass admission controller has made modifications, you can check the updated
Pod overhead value:
kubectl get pod test-pod -o jsonpath='{.spec.overhead}'
The output is:
map[cpu:250m memory:120Mi]
If a ResourceQuota is defined, the sum of container requests as well as the
overhead field are counted.
When the kube-scheduler is deciding which node should run a new Pod, the scheduler considers that Pod's
overhead as well as the sum of container requests for that Pod. For this example, the scheduler adds the
requests and the overhead, then looks for a node that has 2.25 CPU and 320 MiB of memory available.
Once a Pod is scheduled to a node, the kubelet on that node creates a new cgroup for the Pod. It is within this pod that the underlying
container runtime will create containers.
If the resource has a limit defined for each container (Guaranteed QoS or Burstable QoS with limits defined),
the kubelet will set an upper limit for the pod cgroup associated with that resource (cpu.cfs_quota_us for CPU
and memory.limit_in_bytes memory). This upper limit is based on the sum of the container limits plus the overhead
defined in the PodSpec.
For CPU, if the Pod is Guaranteed or Burstable QoS, the kubelet will set cpu.shares based on the
sum of container requests plus the overhead defined in the PodSpec.
Looking at our example, verify the container requests for the workload:
kubectl get pod test-pod -o jsonpath='{.spec.containers[*].resources.limits}'
The total container requests are 2000m CPU and 200MiB of memory:
The output shows requests for 2250m CPU, and for 320MiB of memory. The requests include Pod overhead:
Namespace Name CPU Requests CPU Limits Memory Requests Memory Limits AGE
--------- ---- ------------ ---------- --------------- ------------- ---
default test-pod 2250m (56%) 2250m (56%) 320Mi (1%) 320Mi (1%) 36m
Verify Pod cgroup limits
Check the Pod's memory cgroups on the node where the workload is running. In the following example,
crictl
is used on the node, which provides a CLI for CRI-compatible container runtimes. This is an
advanced example to show Pod overhead behavior, and it is not expected that users should need to check
cgroups directly on the node.
First, on the particular node, determine the Pod identifier:
# Run this on the node where the Pod is scheduledPOD_ID="$(sudo crictl pods --name test-pod -q)"
From this, you can determine the cgroup path for the Pod:
# Run this on the node where the Pod is scheduledsudo crictl inspectp -o=json $POD_ID | grep cgroupsPath
The resulting cgroup path includes the Pod's pause container. The Pod level cgroup is one directory above.
In this specific case, the pod cgroup path is kubepods/podd7f4b509-cf94-4951-9417-d1087c92a5b2.
Verify the Pod level cgroup setting for memory:
# Run this on the node where the Pod is scheduled.# Also, change the name of the cgroup to match the cgroup allocated for your pod. cat /sys/fs/cgroup/memory/kubepods/podd7f4b509-cf94-4951-9417-d1087c92a5b2/memory.limit_in_bytes
This is 320 MiB, as expected:
335544320
Observability
Some kube_pod_overhead_* metrics are available in kube-state-metrics
to help identify when Pod overhead is being utilized and to help observe stability of workloads
running with a defined overhead.
You can use topology spread constraints to control how
Pods are spread across your cluster
among failure-domains such as regions, zones, nodes, and other user-defined topology
domains. This can help to achieve high availability as well as efficient resource
utilization.
You can set cluster-level constraints as a default,
or configure topology spread constraints for individual workloads.
Motivation
Imagine that you have a cluster of up to twenty nodes, and you want to run a
workload
that automatically scales how many replicas it uses. There could be as few as
two Pods or as many as fifteen.
When there are only two Pods, you'd prefer not to have both of those Pods run on the
same node: you would run the risk that a single node failure takes your workload
offline.
In addition to this basic usage, there are some advanced usage examples that
enable your workloads to benefit on high availability and cluster utilization.
As you scale up and run more Pods, a different concern becomes important. Imagine
that you have three nodes running five Pods each. The nodes have enough capacity
to run that many replicas; however, the clients that interact with this workload
are split across three different datacenters (or infrastructure zones). Now you
have less concern about a single node failure, but you notice that latency is
higher than you'd like, and you are paying for network costs associated with
sending network traffic between the different zones.
You decide that under normal operation you'd prefer to have a similar number of replicas
scheduled into each infrastructure zone,
and you'd like the cluster to self-heal in the case that there is a problem.
Pod topology spread constraints offer you a declarative way to configure that.
topologySpreadConstraints field
The Pod API includes a field, spec.topologySpreadConstraints. The usage of this field looks like
the following:
---apiVersion:v1kind:Podmetadata:name:example-podspec:# Configure a topology spread constrainttopologySpreadConstraints:- maxSkew:<integer>minDomains:<integer># optional; beta since v1.25topologyKey:<string>whenUnsatisfiable:<string>labelSelector:<object>matchLabelKeys:<list># optional; alpha since v1.25nodeAffinityPolicy:[Honor|Ignore]# optional; alpha since v1.25nodeTaintsPolicy:[Honor|Ignore]# optional; alpha since v1.25### other Pod fields go here
You can read more about this field by running kubectl explain Pod.spec.topologySpreadConstraints or
refer to scheduling section of the API reference for Pod.
Spread constraint definition
You can define one or multiple topologySpreadConstraints entries to instruct the
kube-scheduler how to place each incoming Pod in relation to the existing Pods across
your cluster. Those fields are:
maxSkew describes the degree to which Pods may be unevenly distributed. You must
specify this field and the number must be greater than zero. Its semantics differ
according to the value of whenUnsatisfiable:
if you select whenUnsatisfiable: DoNotSchedule, then maxSkew defines the
maximum permitted difference between the number of matching pods in the target
topology and the global minimum
(the minimum number of matching pods in an eligible domain or zero if the number of eligible domains is less than MinDomains).
For example, if you have 3 zones with 2, 2 and 1 matching pods respectively,
MaxSkew is set to 1 then the global minimum is 1.
if you select whenUnsatisfiable: ScheduleAnyway, the scheduler gives higher
precedence to topologies that would help reduce the skew.
minDomains indicates a minimum number of eligible domains. This field is optional.
A domain is a particular instance of a topology. An eligible domain is a domain whose
nodes match the node selector.
Note: The minDomains field is a beta field and enabled by default in 1.25. You can disable it by disabling the
MinDomainsInPodToplogySpreadfeature gate.
The value of minDomains must be greater than 0, when specified.
You can only specify minDomains in conjunction with whenUnsatisfiable: DoNotSchedule.
When the number of eligible domains with match topology keys is less than minDomains,
Pod topology spread treats global minimum as 0, and then the calculation of skew is performed.
The global minimum is the minimum number of matching Pods in an eligible domain,
or zero if the number of eligible domains is less than minDomains.
When the number of eligible domains with matching topology keys equals or is greater than
minDomains, this value has no effect on scheduling.
If you do not specify minDomains, the constraint behaves as if minDomains is 1.
topologyKey is the key of node labels. Nodes that have a label with this key
and identical values are considered to be in the same topology.
We call each instance of a topology (in other words, a <key, value> pair) a domain. The scheduler
will try to put a balanced number of pods into each domain.
Also, we define an eligible domain as a domain whose nodes meet the requirements of
nodeAffinityPolicy and nodeTaintsPolicy.
whenUnsatisfiable indicates how to deal with a Pod if it doesn't satisfy the spread constraint:
DoNotSchedule (default) tells the scheduler not to schedule it.
ScheduleAnyway tells the scheduler to still schedule it while prioritizing nodes that minimize the skew.
labelSelector is used to find matching Pods. Pods
that match this label selector are counted to determine the
number of Pods in their corresponding topology domain.
See Label Selectors
for more details.
matchLabelKeys is a list of pod label keys to select the pods over which
spreading will be calculated. The keys are used to lookup values from the pod labels, those key-value labels are ANDed with labelSelector to select the group of existing pods over which spreading will be calculated for the incoming pod. Keys that don't exist in the pod labels will be ignored. A null or empty list means only match against the labelSelector.
With matchLabelKeys, users don't need to update the pod.spec between different revisions. The controller/operator just needs to set different values to the same label key for different revisions. The scheduler will assume the values automatically based on matchLabelKeys. For example, if users use Deployment, they can use the label keyed with pod-template-hash, which is added automatically by the Deployment controller, to distinguish between different revisions in a single Deployment.
Note: The matchLabelKeys field is an alpha field added in 1.25. You have to enable the
MatchLabelKeysInPodTopologySpreadfeature gate
in order to use it.
nodeAffinityPolicy indicates how we will treat Pod's nodeAffinity/nodeSelector
when calculating pod topology spread skew. Options are:
Honor: only nodes matching nodeAffinity/nodeSelector are included in the calculations.
Ignore: nodeAffinity/nodeSelector are ignored. All nodes are included in the calculations.
If this value is null, the behavior is equivalent to the Honor policy.
Note: The nodeAffinityPolicy is an alpha-level field added in 1.25. You have to enable the
NodeInclusionPolicyInPodTopologySpreadfeature gate
in order to use it.
nodeTaintsPolicy indicates how we will treat node taints when calculating
pod topology spread skew. Options are:
Honor: nodes without taints, along with tainted nodes for which the incoming pod
has a toleration, are included.
Ignore: node taints are ignored. All nodes are included.
If this value is null, the behavior is equivalent to the Ignore policy.
Note: The nodeTaintsPolicy is an alpha-level field added in 1.25. You have to enable the
NodeInclusionPolicyInPodTopologySpreadfeature gate
in order to use it.
When a Pod defines more than one topologySpreadConstraint, those constraints are
combined using a logical AND operation: the kube-scheduler looks for a node for the incoming Pod
that satisfies all the configured constraints.
Node labels
Topology spread constraints rely on node labels to identify the topology
domain(s) that each node is in.
For example, a node might have labels:
region:us-east-1zone:us-east-1a
Note:
For brevity, this example doesn't use the
well-known label keys
topology.kubernetes.io/zone and topology.kubernetes.io/region. However,
those registered label keys are nonetheless recommended rather than the private
(unqualified) label keys region and zone that are used here.
You can't make a reliable assumption about the meaning of a private label key
between different contexts.
Suppose you have a 4-node cluster with the following labels:
NAME STATUS ROLES AGE VERSION LABELS
node1 Ready <none> 4m26s v1.16.0 node=node1,zone=zoneA
node2 Ready <none> 3m58s v1.16.0 node=node2,zone=zoneA
node3 Ready <none> 3m17s v1.16.0 node=node3,zone=zoneB
node4 Ready <none> 2m43s v1.16.0 node=node4,zone=zoneB
Then the cluster is logically viewed as below:
graph TB
subgraph "zoneB"
n3(Node3)
n4(Node4)
end
subgraph "zoneA"
n1(Node1)
n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4 k8s;
class zoneA,zoneB cluster;
Consistency
You should set the same Pod topology spread constraints on all pods in a group.
Usually, if you are using a workload controller such as a Deployment, the pod template
takes care of this for you. If you mix different spread constraints then Kubernetes
follows the API definition of the field; however, the behavior is more likely to become
confusing and troubleshooting is less straightforward.
You need a mechanism to ensure that all the nodes in a topology domain (such as a
cloud provider region) are labelled consistently.
To avoid you needing to manually label nodes, most clusters automatically
populate well-known labels such as topology.kubernetes.io/hostname. Check whether
your cluster supports this.
Topology spread constraint examples
Example: one topology spread constraint
Suppose you have a 4-node cluster where 3 Pods labelled foo: bar are located in
node1, node2 and node3 respectively:
graph BT
subgraph "zoneB"
p3(Pod) --> n3(Node3)
n4(Node4)
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3 k8s;
class zoneA,zoneB cluster;
If you want an incoming Pod to be evenly spread with existing Pods across zones, you
can use a manifest similar to:
From that manifest, topologyKey: zone implies the even distribution will only be applied
to nodes that are labelled zone: <any value> (nodes that don't have a zone label
are skipped). The field whenUnsatisfiable: DoNotSchedule tells the scheduler to let the
incoming Pod stay pending if the scheduler can't find a way to satisfy the constraint.
If the scheduler placed this incoming Pod into zone A, the distribution of Pods would
become [3, 1]. That means the actual skew is then 2 (calculated as 3 - 1), which
violates maxSkew: 1. To satisfy the constraints and context for this example, the
incoming Pod can only be placed onto a node in zone B:
graph BT
subgraph "zoneB"
p3(Pod) --> n3(Node3)
p4(mypod) --> n4(Node4)
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3 k8s;
class p4 plain;
class zoneA,zoneB cluster;
OR
graph BT
subgraph "zoneB"
p3(Pod) --> n3(Node3)
p4(mypod) --> n3
n4(Node4)
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3 k8s;
class p4 plain;
class zoneA,zoneB cluster;
You can tweak the Pod spec to meet various kinds of requirements:
Change maxSkew to a bigger value - such as 2 - so that the incoming Pod can
be placed into zone A as well.
Change topologyKey to node so as to distribute the Pods evenly across nodes
instead of zones. In the above example, if maxSkew remains 1, the incoming
Pod can only be placed onto the node node4.
Change whenUnsatisfiable: DoNotSchedule to whenUnsatisfiable: ScheduleAnyway
to ensure the incoming Pod to be always schedulable (suppose other scheduling APIs
are satisfied). However, it's preferred to be placed into the topology domain which
has fewer matching Pods. (Be aware that this preference is jointly normalized
with other internal scheduling priorities such as resource usage ratio).
Example: multiple topology spread constraints
This builds upon the previous example. Suppose you have a 4-node cluster where 3
existing Pods labeled foo: bar are located on node1, node2 and node3 respectively:
graph BT
subgraph "zoneB"
p3(Pod) --> n3(Node3)
n4(Node4)
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3 k8s;
class p4 plain;
class zoneA,zoneB cluster;
You can combine two topology spread constraints to control the spread of Pods both
by node and by zone:
In this case, to match the first constraint, the incoming Pod can only be placed onto
nodes in zone B; while in terms of the second constraint, the incoming Pod can only be
scheduled to the node node4. The scheduler only considers options that satisfy all
defined constraints, so the only valid placement is onto node node4.
Example: conflicting topology spread constraints
Multiple constraints can lead to conflicts. Suppose you have a 3-node cluster across 2 zones:
graph BT
subgraph "zoneB"
p4(Pod) --> n3(Node3)
p5(Pod) --> n3
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n1
p3(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3,p4,p5 k8s;
class zoneA,zoneB cluster;
If you were to apply
two-constraints.yaml
(the manifest from the previous example)
to this cluster, you would see that the Pod mypod stays in the Pending state.
This happens because: to satisfy the first constraint, the Pod mypod can only
be placed into zone B; while in terms of the second constraint, the Pod mypod
can only schedule to node node2. The intersection of the two constraints returns
an empty set, and the scheduler cannot place the Pod.
To overcome this situation, you can either increase the value of maxSkew or modify
one of the constraints to use whenUnsatisfiable: ScheduleAnyway. Depending on
circumstances, you might also decide to delete an existing Pod manually - for example,
if you are troubleshooting why a bug-fix rollout is not making progress.
Interaction with node affinity and node selectors
The scheduler will skip the non-matching nodes from the skew calculations if the
incoming Pod has spec.nodeSelector or spec.affinity.nodeAffinity defined.
Example: topology spread constraints with node affinity
Suppose you have a 5-node cluster ranging across zones A to C:
graph BT
subgraph "zoneB"
p3(Pod) --> n3(Node3)
n4(Node4)
end
subgraph "zoneA"
p1(Pod) --> n1(Node1)
p2(Pod) --> n2(Node2)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n1,n2,n3,n4,p1,p2,p3 k8s;
class p4 plain;
class zoneA,zoneB cluster;
graph BT
subgraph "zoneC"
n5(Node5)
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class n5 k8s;
class zoneC cluster;
and you know that zone C must be excluded. In this case, you can compose a manifest
as below, so that Pod mypod will be placed into zone B instead of zone C.
Similarly, Kubernetes also respects spec.nodeSelector.
There are some implicit conventions worth noting here:
Only the Pods holding the same namespace as the incoming Pod can be matching candidates.
The scheduler bypasses any nodes that don't have any topologySpreadConstraints[*].topologyKey
present. This implies that:
any Pods located on those bypassed nodes do not impact maxSkew calculation - in the
above example, suppose the node node1 does not have a label "zone", then the 2 Pods will
be disregarded, hence the incoming Pod will be scheduled into zone A.
the incoming Pod has no chances to be scheduled onto this kind of nodes -
in the above example, suppose a node node5 has the mistyped label zone-typo: zoneC
(and no zone label set). After node node5 joins the cluster, it will be bypassed and
Pods for this workload aren't scheduled there.
Be aware of what will happen if the incoming Pod's
topologySpreadConstraints[*].labelSelector doesn't match its own labels. In the
above example, if you remove the incoming Pod's labels, it can still be placed onto
nodes in zone B, since the constraints are still satisfied. However, after that
placement, the degree of imbalance of the cluster remains unchanged - it's still zone A
having 2 Pods labelled as foo: bar, and zone B having 1 Pod labelled as
foo: bar. If this is not what you expect, update the workload's
topologySpreadConstraints[*].labelSelector to match the labels in the pod template.
Cluster-level default constraints
It is possible to set default topology spread constraints for a cluster. Default
topology spread constraints are applied to a Pod if, and only if:
It doesn't define any constraints in its .spec.topologySpreadConstraints.
It belongs to a Service, ReplicaSet, StatefulSet or ReplicationController.
Default constraints can be set as part of the PodTopologySpread plugin
arguments in a scheduling profile.
The constraints are specified with the same API above, except that
labelSelector must be empty. The selectors are calculated from the Services,
ReplicaSets, StatefulSets or ReplicationControllers that the Pod belongs to.
Note: The SelectorSpread plugin
is disabled by default. The Kubernetes project recommends using PodTopologySpread
to achieve similar behavior.
Built-in default constraints
FEATURE STATE:Kubernetes v1.24 [stable]
If you don't configure any cluster-level default constraints for pod topology spreading,
then kube-scheduler acts as if you specified the following default topology constraints:
Also, the legacy SelectorSpread plugin, which provides an equivalent behavior,
is disabled by default.
Note:
The PodTopologySpread plugin does not score the nodes that don't have
the topology keys specified in the spreading constraints. This might result
in a different default behavior compared to the legacy SelectorSpread plugin when
using the default topology constraints.
If your nodes are not expected to have bothkubernetes.io/hostname and
topology.kubernetes.io/zone labels set, define your own constraints
instead of using the Kubernetes defaults.
If you don't want to use the default Pod spreading constraints for your cluster,
you can disable those defaults by setting defaultingType to List and leaving
empty defaultConstraints in the PodTopologySpread plugin configuration:
In Kubernetes, inter-Pod affinity and anti-affinity
control how Pods are scheduled in relation to one another - either more packed
or more scattered.
podAffinity
attracts Pods; you can try to pack any number of Pods into qualifying
topology domain(s)
podAntiAffinity
repels Pods. If you set this to requiredDuringSchedulingIgnoredDuringExecution mode then
only a single Pod can be scheduled into a single topology domain; if you choose
preferredDuringSchedulingIgnoredDuringExecution then you lose the ability to enforce the
constraint.
For finer control, you can specify topology spread constraints to distribute
Pods across different topology domains - to achieve either high availability or
cost-saving. This can also help on rolling update workloads and scaling out
replicas smoothly.
For more context, see the
Motivation
section of the enhancement proposal about Pod topology spread constraints.
Known limitations
There's no guarantee that the constraints remain satisfied when Pods are removed. For
example, scaling down a Deployment may result in imbalanced Pods distribution.
You can use a tool such as the Descheduler
to rebalance the Pods distribution.
Pods matched on tainted nodes are respected.
See Issue 80921.
The scheduler doesn't have prior knowledge of all the zones or other topology
domains that a cluster has. They are determined from the existing nodes in the
cluster. This could lead to a problem in autoscaled clusters, when a node pool (or
node group) is scaled to zero nodes, and you're expecting the cluster to scale up,
because, in this case, those topology domains won't be considered until there is
at least one node in them.
You can work around this by using an cluster autoscaling tool that is aware of
Pod topology spread constraints and is also aware of the overall set of topology
domains.
What's next
The blog article Introducing PodTopologySpread
explains maxSkew in some detail, as well as covering some advanced usage examples.
Read the scheduling section of
the API reference for Pod.
3.11.5 - Taints and Tolerations
Node affinity
is a property of Pods that attracts them to
a set of nodes (either as a preference or a
hard requirement). Taints are the opposite -- they allow a node to repel a set of pods.
Tolerations are applied to pods. Tolerations allow the scheduler to schedule pods with matching
taints. Tolerations allow scheduling but don't guarantee scheduling: the scheduler also
evaluates other parameters
as part of its function.
Taints and tolerations work together to ensure that pods are not scheduled
onto inappropriate nodes. One or more taints are applied to a node; this
marks that the node should not accept any pods that do not tolerate the taints.
Concepts
You add a taint to a node using kubectl taint.
For example,
kubectl taint nodes node1 key1=value1:NoSchedule
places a taint on node node1. The taint has key key1, value value1, and taint effect NoSchedule.
This means that no pod will be able to schedule onto node1 unless it has a matching toleration.
To remove the taint added by the command above, you can run:
kubectl taint nodes node1 key1=value1:NoSchedule-
You specify a toleration for a pod in the PodSpec. Both of the following tolerations "match" the
taint created by the kubectl taint line above, and thus a pod with either toleration would be able
to schedule onto node1:
A toleration "matches" a taint if the keys are the same and the effects are the same, and:
the operator is Exists (in which case no value should be specified), or
the operator is Equal and the values are equal.
Note:
There are two special cases:
An empty key with operator Exists matches all keys, values and effects which means this
will tolerate everything.
An empty effect matches all effects with key key1.
The above example used effect of NoSchedule. Alternatively, you can use effect of PreferNoSchedule.
This is a "preference" or "soft" version of NoSchedule -- the system will try to avoid placing a
pod that does not tolerate the taint on the node, but it is not required. The third kind of effect is
NoExecute, described later.
You can put multiple taints on the same node and multiple tolerations on the same pod.
The way Kubernetes processes multiple taints and tolerations is like a filter: start
with all of a node's taints, then ignore the ones for which the pod has a matching toleration; the
remaining un-ignored taints have the indicated effects on the pod. In particular,
if there is at least one un-ignored taint with effect NoSchedule then Kubernetes will not schedule
the pod onto that node
if there is no un-ignored taint with effect NoSchedule but there is at least one un-ignored taint with
effect PreferNoSchedule then Kubernetes will try to not schedule the pod onto the node
if there is at least one un-ignored taint with effect NoExecute then the pod will be evicted from
the node (if it is already running on the node), and will not be
scheduled onto the node (if it is not yet running on the node).
In this case, the pod will not be able to schedule onto the node, because there is no
toleration matching the third taint. But it will be able to continue running if it is
already running on the node when the taint is added, because the third taint is the only
one of the three that is not tolerated by the pod.
Normally, if a taint with effect NoExecute is added to a node, then any pods that do
not tolerate the taint will be evicted immediately, and pods that do tolerate the
taint will never be evicted. However, a toleration with NoExecute effect can specify
an optional tolerationSeconds field that dictates how long the pod will stay bound
to the node after the taint is added. For example,
means that if this pod is running and a matching taint is added to the node, then
the pod will stay bound to the node for 3600 seconds, and then be evicted. If the
taint is removed before that time, the pod will not be evicted.
Example Use Cases
Taints and tolerations are a flexible way to steer pods away from nodes or evict
pods that shouldn't be running. A few of the use cases are
Dedicated Nodes: If you want to dedicate a set of nodes for exclusive use by
a particular set of users, you can add a taint to those nodes (say,
kubectl taint nodes nodename dedicated=groupName:NoSchedule) and then add a corresponding
toleration to their pods (this would be done most easily by writing a custom
admission controller).
The pods with the tolerations will then be allowed to use the tainted (dedicated) nodes as
well as any other nodes in the cluster. If you want to dedicate the nodes to them and
ensure they only use the dedicated nodes, then you should additionally add a label similar
to the taint to the same set of nodes (e.g. dedicated=groupName), and the admission
controller should additionally add a node affinity to require that the pods can only schedule
onto nodes labeled with dedicated=groupName.
Nodes with Special Hardware: In a cluster where a small subset of nodes have specialized
hardware (for example GPUs), it is desirable to keep pods that don't need the specialized
hardware off of those nodes, thus leaving room for later-arriving pods that do need the
specialized hardware. This can be done by tainting the nodes that have the specialized
hardware (e.g. kubectl taint nodes nodename special=true:NoSchedule or
kubectl taint nodes nodename special=true:PreferNoSchedule) and adding a corresponding
toleration to pods that use the special hardware. As in the dedicated nodes use case,
it is probably easiest to apply the tolerations using a custom
admission controller.
For example, it is recommended to use Extended
Resources
to represent the special hardware, taint your special hardware nodes with the
extended resource name and run the
ExtendedResourceToleration
admission controller. Now, because the nodes are tainted, no pods without the
toleration will schedule on them. But when you submit a pod that requests the
extended resource, the ExtendedResourceToleration admission controller will
automatically add the correct toleration to the pod and that pod will schedule
on the special hardware nodes. This will make sure that these special hardware
nodes are dedicated for pods requesting such hardware and you don't have to
manually add tolerations to your pods.
Taint based Evictions: A per-pod-configurable eviction behavior
when there are node problems, which is described in the next section.
Taint based Evictions
FEATURE STATE:Kubernetes v1.18 [stable]
The NoExecute taint effect, mentioned above, affects pods that are already
running on the node as follows
pods that do not tolerate the taint are evicted immediately
pods that tolerate the taint without specifying tolerationSeconds in
their toleration specification remain bound forever
pods that tolerate the taint with a specified tolerationSeconds remain
bound for the specified amount of time
The node controller automatically taints a Node when certain conditions
are true. The following taints are built in:
node.kubernetes.io/not-ready: Node is not ready. This corresponds to
the NodeCondition Ready being "False".
node.kubernetes.io/unreachable: Node is unreachable from the node
controller. This corresponds to the NodeCondition Ready being "Unknown".
node.kubernetes.io/memory-pressure: Node has memory pressure.
node.kubernetes.io/disk-pressure: Node has disk pressure.
node.kubernetes.io/pid-pressure: Node has PID pressure.
node.kubernetes.io/network-unavailable: Node's network is unavailable.
node.kubernetes.io/unschedulable: Node is unschedulable.
node.cloudprovider.kubernetes.io/uninitialized: When the kubelet is started
with "external" cloud provider, this taint is set on a node to mark it
as unusable. After a controller from the cloud-controller-manager initializes
this node, the kubelet removes this taint.
In case a node is to be evicted, the node controller or the kubelet adds relevant taints
with NoExecute effect. If the fault condition returns to normal the kubelet or node
controller can remove the relevant taint(s).
Note: The control plane limits the rate of adding node new taints to nodes. This rate limiting
manages the number of evictions that are triggered when many nodes become unreachable at
once (for example: if there is a network disruption).
You can specify tolerationSeconds for a Pod to define how long that Pod stays bound
to a failing or unresponsive Node.
For example, you might want to keep an application with a lot of local state
bound to node for a long time in the event of network partition, hoping
that the partition will recover and thus the pod eviction can be avoided.
The toleration you set for that Pod might look like:
Kubernetes automatically adds a toleration for
node.kubernetes.io/not-ready and node.kubernetes.io/unreachable
with tolerationSeconds=300,
unless you, or a controller, set those tolerations explicitly.
These automatically-added tolerations mean that Pods remain bound to
Nodes for 5 minutes after one of these problems is detected.
DaemonSet pods are created with
NoExecute tolerations for the following taints with no tolerationSeconds:
node.kubernetes.io/unreachable
node.kubernetes.io/not-ready
This ensures that DaemonSet pods are never evicted due to these problems.
Taint Nodes by Condition
The control plane, using the node controller,
automatically creates taints with a NoSchedule effect for
node conditions.
The scheduler checks taints, not node conditions, when it makes scheduling
decisions. This ensures that node conditions don't directly affect scheduling.
For example, if the DiskPressure node condition is active, the control plane
adds the node.kubernetes.io/disk-pressure taint and does not schedule new pods
onto the affected node. If the MemoryPressure node condition is active, the
control plane adds the node.kubernetes.io/memory-pressure taint.
You can ignore node conditions for newly created pods by adding the corresponding
Pod tolerations. The control plane also adds the node.kubernetes.io/memory-pressure
toleration on pods that have a QoS class
other than BestEffort. This is because Kubernetes treats pods in the Guaranteed
or Burstable QoS classes (even pods with no memory request set) as if they are
able to cope with memory pressure, while new BestEffort pods are not scheduled
onto the affected node.
The DaemonSet controller automatically adds the following NoSchedule
tolerations to all daemons, to prevent DaemonSets from breaking.
Pods can have priority. Priority indicates the
importance of a Pod relative to other Pods. If a Pod cannot be scheduled, the
scheduler tries to preempt (evict) lower priority Pods to make scheduling of the
pending Pod possible.
Warning:
In a cluster where not all users are trusted, a malicious user could create Pods
at the highest possible priorities, causing other Pods to be evicted/not get
scheduled.
An administrator can use ResourceQuota to prevent users from creating pods at
high priorities.
Create Pods withpriorityClassName set to one of the added
PriorityClasses. Of course you do not need to create the Pods directly;
normally you would add priorityClassName to the Pod template of a
collection object like a Deployment.
Keep reading for more information about these steps.
A PriorityClass is a non-namespaced object that defines a mapping from a
priority class name to the integer value of the priority. The name is specified
in the name field of the PriorityClass object's metadata. The value is
specified in the required value field. The higher the value, the higher the
priority.
The name of a PriorityClass object must be a valid
DNS subdomain name,
and it cannot be prefixed with system-.
A PriorityClass object can have any 32-bit integer value smaller than or equal
to 1 billion. Larger numbers are reserved for critical system Pods that should
not normally be preempted or evicted. A cluster admin should create one
PriorityClass object for each such mapping that they want.
PriorityClass also has two optional fields: globalDefault and description.
The globalDefault field indicates that the value of this PriorityClass should
be used for Pods without a priorityClassName. Only one PriorityClass with
globalDefault set to true can exist in the system. If there is no
PriorityClass with globalDefault set, the priority of Pods with no
priorityClassName is zero.
The description field is an arbitrary string. It is meant to tell users of the
cluster when they should use this PriorityClass.
Notes about PodPriority and existing clusters
If you upgrade an existing cluster without this feature, the priority
of your existing Pods is effectively zero.
Addition of a PriorityClass with globalDefault set to true does not
change the priorities of existing Pods. The value of such a PriorityClass is
used only for Pods created after the PriorityClass is added.
If you delete a PriorityClass, existing Pods that use the name of the
deleted PriorityClass remain unchanged, but you cannot create more Pods that
use the name of the deleted PriorityClass.
Example PriorityClass
apiVersion:scheduling.k8s.io/v1kind:PriorityClassmetadata:name:high-priorityvalue:1000000globalDefault:falsedescription:"This priority class should be used for XYZ service pods only."
Non-preempting PriorityClass
FEATURE STATE:Kubernetes v1.24 [stable]
Pods with preemptionPolicy: Never will be placed in the scheduling queue
ahead of lower-priority pods,
but they cannot preempt other pods.
A non-preempting pod waiting to be scheduled will stay in the scheduling queue,
until sufficient resources are free,
and it can be scheduled.
Non-preempting pods,
like other pods,
are subject to scheduler back-off.
This means that if the scheduler tries these pods and they cannot be scheduled,
they will be retried with lower frequency,
allowing other pods with lower priority to be scheduled before them.
Non-preempting pods may still be preempted by other,
high-priority pods.
preemptionPolicy defaults to PreemptLowerPriority,
which will allow pods of that PriorityClass to preempt lower-priority pods
(as is existing default behavior).
If preemptionPolicy is set to Never,
pods in that PriorityClass will be non-preempting.
An example use case is for data science workloads.
A user may submit a job that they want to be prioritized above other workloads,
but do not wish to discard existing work by preempting running pods.
The high priority job with preemptionPolicy: Never will be scheduled
ahead of other queued pods,
as soon as sufficient cluster resources "naturally" become free.
Example Non-preempting PriorityClass
apiVersion:scheduling.k8s.io/v1kind:PriorityClassmetadata:name:high-priority-nonpreemptingvalue:1000000preemptionPolicy:NeverglobalDefault:falsedescription:"This priority class will not cause other pods to be preempted."
Pod priority
After you have one or more PriorityClasses, you can create Pods that specify one
of those PriorityClass names in their specifications. The priority admission
controller uses the priorityClassName field and populates the integer value of
the priority. If the priority class is not found, the Pod is rejected.
The following YAML is an example of a Pod configuration that uses the
PriorityClass created in the preceding example. The priority admission
controller checks the specification and resolves the priority of the Pod to
1000000.
When Pod priority is enabled, the scheduler orders pending Pods by
their priority and a pending Pod is placed ahead of other pending Pods
with lower priority in the scheduling queue. As a result, the higher
priority Pod may be scheduled sooner than Pods with lower priority if
its scheduling requirements are met. If such Pod cannot be scheduled,
scheduler will continue and tries to schedule other lower priority Pods.
Preemption
When Pods are created, they go to a queue and wait to be scheduled. The
scheduler picks a Pod from the queue and tries to schedule it on a Node. If no
Node is found that satisfies all the specified requirements of the Pod,
preemption logic is triggered for the pending Pod. Let's call the pending Pod P.
Preemption logic tries to find a Node where removal of one or more Pods with
lower priority than P would enable P to be scheduled on that Node. If such a
Node is found, one or more lower priority Pods get evicted from the Node. After
the Pods are gone, P can be scheduled on the Node.
User exposed information
When Pod P preempts one or more Pods on Node N, nominatedNodeName field of Pod
P's status is set to the name of Node N. This field helps scheduler track
resources reserved for Pod P and also gives users information about preemptions
in their clusters.
Please note that Pod P is not necessarily scheduled to the "nominated Node".
The scheduler always tries the "nominated Node" before iterating over any other nodes.
After victim Pods are preempted, they get their graceful termination period. If
another node becomes available while scheduler is waiting for the victim Pods to
terminate, scheduler may use the other node to schedule Pod P. As a result
nominatedNodeName and nodeName of Pod spec are not always the same. Also, if
scheduler preempts Pods on Node N, but then a higher priority Pod than Pod P
arrives, scheduler may give Node N to the new higher priority Pod. In such a
case, scheduler clears nominatedNodeName of Pod P. By doing this, scheduler
makes Pod P eligible to preempt Pods on another Node.
Limitations of preemption
Graceful termination of preemption victims
When Pods are preempted, the victims get their
graceful termination period.
They have that much time to finish their work and exit. If they don't, they are
killed. This graceful termination period creates a time gap between the point
that the scheduler preempts Pods and the time when the pending Pod (P) can be
scheduled on the Node (N). In the meantime, the scheduler keeps scheduling other
pending Pods. As victims exit or get terminated, the scheduler tries to schedule
Pods in the pending queue. Therefore, there is usually a time gap between the
point that scheduler preempts victims and the time that Pod P is scheduled. In
order to minimize this gap, one can set graceful termination period of lower
priority Pods to zero or a small number.
PodDisruptionBudget is supported, but not guaranteed
A PodDisruptionBudget (PDB)
allows application owners to limit the number of Pods of a replicated application
that are down simultaneously from voluntary disruptions. Kubernetes supports
PDB when preempting Pods, but respecting PDB is best effort. The scheduler tries
to find victims whose PDB are not violated by preemption, but if no such victims
are found, preemption will still happen, and lower priority Pods will be removed
despite their PDBs being violated.
Inter-Pod affinity on lower-priority Pods
A Node is considered for preemption only when the answer to this question is
yes: "If all the Pods with lower priority than the pending Pod are removed from
the Node, can the pending Pod be scheduled on the Node?"
Note: Preemption does not necessarily remove all lower-priority
Pods. If the pending Pod can be scheduled by removing fewer than all
lower-priority Pods, then only a portion of the lower-priority Pods are removed.
Even so, the answer to the preceding question must be yes. If the answer is no,
the Node is not considered for preemption.
If a pending Pod has inter-pod affinity
to one or more of the lower-priority Pods on the Node, the inter-Pod affinity
rule cannot be satisfied in the absence of those lower-priority Pods. In this case,
the scheduler does not preempt any Pods on the Node. Instead, it looks for another
Node. The scheduler might find a suitable Node or it might not. There is no
guarantee that the pending Pod can be scheduled.
Our recommended solution for this problem is to create inter-Pod affinity only
towards equal or higher priority Pods.
Cross node preemption
Suppose a Node N is being considered for preemption so that a pending Pod P can
be scheduled on N. P might become feasible on N only if a Pod on another Node is
preempted. Here's an example:
Pod P is being considered for Node N.
Pod Q is running on another Node in the same Zone as Node N.
Pod P has Zone-wide anti-affinity with Pod Q (topologyKey: topology.kubernetes.io/zone).
There are no other cases of anti-affinity between Pod P and other Pods in
the Zone.
In order to schedule Pod P on Node N, Pod Q can be preempted, but scheduler
does not perform cross-node preemption. So, Pod P will be deemed
unschedulable on Node N.
If Pod Q were removed from its Node, the Pod anti-affinity violation would be
gone, and Pod P could possibly be scheduled on Node N.
We may consider adding cross Node preemption in future versions if there is
enough demand and if we find an algorithm with reasonable performance.
Troubleshooting
Pod priority and pre-emption can have unwanted side effects. Here are some
examples of potential problems and ways to deal with them.
Pods are preempted unnecessarily
Preemption removes existing Pods from a cluster under resource pressure to make
room for higher priority pending Pods. If you give high priorities to
certain Pods by mistake, these unintentionally high priority Pods may cause
preemption in your cluster. Pod priority is specified by setting the
priorityClassName field in the Pod's specification. The integer value for
priority is then resolved and populated to the priority field of podSpec.
To address the problem, you can change the priorityClassName for those Pods
to use lower priority classes, or leave that field empty. An empty
priorityClassName is resolved to zero by default.
When a Pod is preempted, there will be events recorded for the preempted Pod.
Preemption should happen only when a cluster does not have enough resources for
a Pod. In such cases, preemption happens only when the priority of the pending
Pod (preemptor) is higher than the victim Pods. Preemption must not happen when
there is no pending Pod, or when the pending Pods have equal or lower priority
than the victims. If preemption happens in such scenarios, please file an issue.
Pods are preempted, but the preemptor is not scheduled
When pods are preempted, they receive their requested graceful termination
period, which is by default 30 seconds. If the victim Pods do not terminate within
this period, they are forcibly terminated. Once all the victims go away, the
preemptor Pod can be scheduled.
While the preemptor Pod is waiting for the victims to go away, a higher priority
Pod may be created that fits on the same Node. In this case, the scheduler will
schedule the higher priority Pod instead of the preemptor.
This is expected behavior: the Pod with the higher priority should take the place
of a Pod with a lower priority.
Higher priority Pods are preempted before lower priority pods
The scheduler tries to find nodes that can run a pending Pod. If no node is
found, the scheduler tries to remove Pods with lower priority from an arbitrary
node in order to make room for the pending pod.
If a node with low priority Pods is not feasible to run the pending Pod, the scheduler
may choose another node with higher priority Pods (compared to the Pods on the
other node) for preemption. The victims must still have lower priority than the
preemptor Pod.
When there are multiple nodes available for preemption, the scheduler tries to
choose the node with a set of Pods with lowest priority. However, if such Pods
have PodDisruptionBudget that would be violated if they are preempted then the
scheduler may choose another node with higher priority Pods.
When multiple nodes exist for preemption and none of the above scenarios apply,
the scheduler chooses a node with the lowest priority.
Interactions between Pod priority and quality of service
Pod priority and QoS class
are two orthogonal features with few interactions and no default restrictions on
setting the priority of a Pod based on its QoS classes. The scheduler's
preemption logic does not consider QoS when choosing preemption targets.
Preemption considers Pod priority and attempts to choose a set of targets with
the lowest priority. Higher-priority Pods are considered for preemption only if
the removal of the lowest priority Pods is not sufficient to allow the scheduler
to schedule the preemptor Pod, or if the lowest priority Pods are protected by
PodDisruptionBudget.
The kubelet uses Priority to determine pod order for node-pressure eviction.
You can use the QoS class to estimate the order in which pods are most likely
to get evicted. The kubelet ranks pods for eviction based on the following factors:
Whether the starved resource usage exceeds requests
kubelet node-pressure eviction does not evict Pods when their
usage does not exceed their requests. If a Pod with lower priority is not
exceeding its requests, it won't be evicted. Another Pod with higher priority
that exceeds its requests may be evicted.
Node-pressure eviction is the process by which the kubelet proactively terminates
pods to reclaim resources on nodes.
The kubelet monitors resources
like CPU, memory, disk space, and filesystem inodes on your cluster's nodes.
When one or more of these resources reach specific consumption levels, the
kubelet can proactively fail one or more pods on the node to reclaim resources
and prevent starvation.
During a node-pressure eviction, the kubelet sets the PodPhase for the
selected pods to Failed. This terminates the pods.
The kubelet does not respect your configured PodDisruptionBudget or the pod's
terminationGracePeriodSeconds. If you use soft eviction thresholds,
the kubelet respects your configured eviction-max-pod-grace-period. If you use
hard eviction thresholds, it uses a 0s grace period for termination.
If the pods are managed by a workload
resource (such as StatefulSet
or Deployment) that
replaces failed pods, the control plane or kube-controller-manager creates new
pods in place of the evicted pods.
Note: The kubelet attempts to reclaim node-level resources
before it terminates end-user pods. For example, it removes unused container
images when disk resources are starved.
The kubelet uses various parameters to make eviction decisions, like the following:
Eviction signals
Eviction thresholds
Monitoring intervals
Eviction signals
Eviction signals are the current state of a particular resource at a specific
point in time. Kubelet uses eviction signals to make eviction decisions by
comparing the signals to eviction thresholds, which are the minimum amount of
the resource that should be available on the node.
In this table, the Description column shows how kubelet gets the value of the
signal. Each signal supports either a percentage or a literal value. Kubelet
calculates the percentage value relative to the total capacity associated with
the signal.
The value for memory.available is derived from the cgroupfs instead of tools
like free -m. This is important because free -m does not work in a
container, and if users use the node allocatable
feature, out of resource decisions
are made local to the end user Pod part of the cgroup hierarchy as well as the
root node. This script
reproduces the same set of steps that the kubelet performs to calculate
memory.available. The kubelet excludes inactive_file (i.e. # of bytes of
file-backed memory on inactive LRU list) from its calculation as it assumes that
memory is reclaimable under pressure.
The kubelet supports the following filesystem partitions:
nodefs: The node's main filesystem, used for local disk volumes, emptyDir,
log storage, and more. For example, nodefs contains /var/lib/kubelet/.
imagefs: An optional filesystem that container runtimes use to store container
images and container writable layers.
Kubelet auto-discovers these filesystems and ignores other filesystems. Kubelet
does not support other configurations.
Some kubelet garbage collection features are deprecated in favor of eviction:
Existing Flag
New Flag
Rationale
--image-gc-high-threshold
--eviction-hard or --eviction-soft
existing eviction signals can trigger image garbage collection
--image-gc-low-threshold
--eviction-minimum-reclaim
eviction reclaims achieve the same behavior
--maximum-dead-containers
-
deprecated once old logs are stored outside of container's context
--maximum-dead-containers-per-container
-
deprecated once old logs are stored outside of container's context
--minimum-container-ttl-duration
-
deprecated once old logs are stored outside of container's context
Eviction thresholds
You can specify custom eviction thresholds for the kubelet to use when it makes
eviction decisions.
Eviction thresholds have the form [eviction-signal][operator][quantity], where:
quantity is the eviction threshold amount, such as 1Gi. The value of quantity
must match the quantity representation used by Kubernetes. You can use either
literal values or percentages (%).
For example, if a node has 10Gi of total memory and you want trigger eviction if
the available memory falls below 1Gi, you can define the eviction threshold as
either memory.available<10% or memory.available<1Gi. You cannot use both.
You can configure soft and hard eviction thresholds.
Soft eviction thresholds
A soft eviction threshold pairs an eviction threshold with a required
administrator-specified grace period. The kubelet does not evict pods until the
grace period is exceeded. The kubelet returns an error on startup if there is no
specified grace period.
You can specify both a soft eviction threshold grace period and a maximum
allowed pod termination grace period for kubelet to use during evictions. If you
specify a maximum allowed grace period and the soft eviction threshold is met,
the kubelet uses the lesser of the two grace periods. If you do not specify a
maximum allowed grace period, the kubelet kills evicted pods immediately without
graceful termination.
You can use the following flags to configure soft eviction thresholds:
eviction-soft: A set of eviction thresholds like memory.available<1.5Gi
that can trigger pod eviction if held over the specified grace period.
eviction-soft-grace-period: A set of eviction grace periods like memory.available=1m30s
that define how long a soft eviction threshold must hold before triggering a Pod eviction.
eviction-max-pod-grace-period: The maximum allowed grace period (in seconds)
to use when terminating pods in response to a soft eviction threshold being met.
Hard eviction thresholds
A hard eviction threshold has no grace period. When a hard eviction threshold is
met, the kubelet kills pods immediately without graceful termination to reclaim
the starved resource.
You can use the eviction-hard flag to configure a set of hard eviction
thresholds like memory.available<1Gi.
The kubelet has the following default hard eviction thresholds:
memory.available<100Mi
nodefs.available<10%
imagefs.available<15%
nodefs.inodesFree<5% (Linux nodes)
Eviction monitoring interval
The kubelet evaluates eviction thresholds based on its configured housekeeping-interval
which defaults to 10s.
Node conditions
The kubelet reports node conditions to reflect that the node is under pressure
because hard or soft eviction threshold is met, independent of configured grace
periods.
The kubelet maps eviction signals to node conditions as follows:
Node Condition
Eviction Signal
Description
MemoryPressure
memory.available
Available memory on the node has satisfied an eviction threshold
DiskPressure
nodefs.available, nodefs.inodesFree, imagefs.available, or imagefs.inodesFree
Available disk space and inodes on either the node's root filesystem or image filesystem has satisfied an eviction threshold
PIDPressure
pid.available
Available processes identifiers on the (Linux) node has fallen below an eviction threshold
The kubelet updates the node conditions based on the configured
--node-status-update-frequency, which defaults to 10s.
Node condition oscillation
In some cases, nodes oscillate above and below soft eviction thresholds without
holding for the defined grace periods. This causes the reported node condition
to constantly switch between true and false, leading to bad eviction decisions.
To protect against oscillation, you can use the eviction-pressure-transition-period
flag, which controls how long the kubelet must wait before transitioning a node
condition to a different state. The transition period has a default value of 5m.
Reclaiming node level resources
The kubelet tries to reclaim node-level resources before it evicts end-user pods.
When a DiskPressure node condition is reported, the kubelet reclaims node-level
resources based on the filesystems on the node.
With imagefs
If the node has a dedicated imagefs filesystem for container runtimes to use,
the kubelet does the following:
If the nodefs filesystem meets the eviction thresholds, the kubelet garbage collects
dead pods and containers.
If the imagefs filesystem meets the eviction thresholds, the kubelet
deletes all unused images.
Without imagefs
If the node only has a nodefs filesystem that meets eviction thresholds,
the kubelet frees up disk space in the following order:
Garbage collect dead pods and containers
Delete unused images
Pod selection for kubelet eviction
If the kubelet's attempts to reclaim node-level resources don't bring the eviction
signal below the threshold, the kubelet begins to evict end-user pods.
The kubelet uses the following parameters to determine the pod eviction order:
As a result, kubelet ranks and evicts pods in the following order:
BestEffort or Burstable pods where the usage exceeds requests. These pods
are evicted based on their Priority and then by how much their usage level
exceeds the request.
Guaranteed pods and Burstable pods where the usage is less than requests
are evicted last, based on their Priority.
Note: The kubelet does not use the pod's QoS class to determine the eviction order.
You can use the QoS class to estimate the most likely pod eviction order when
reclaiming resources like memory. QoS does not apply to EphemeralStorage requests,
so the above scenario will not apply if the node is, for example, under DiskPressure.
Guaranteed pods are guaranteed only when requests and limits are specified for
all the containers and they are equal. These pods will never be evicted because
of another pod's resource consumption. If a system daemon (such as kubelet
and journald) is consuming more resources than were reserved via
system-reserved or kube-reserved allocations, and the node only has
Guaranteed or Burstable pods using less resources than requests left on it,
then the kubelet must choose to evict one of these pods to preserve node stability
and to limit the impact of resource starvation on other pods. In this case, it
will choose to evict pods of lowest Priority first.
When the kubelet evicts pods in response to inode or PID starvation, it uses
the Priority to determine the eviction order, because inodes and PIDs have no
requests.
The kubelet sorts pods differently based on whether the node has a dedicated
imagefs filesystem:
With imagefs
If nodefs is triggering evictions, the kubelet sorts pods based on nodefs
usage (local volumes + logs of all containers).
If imagefs is triggering evictions, the kubelet sorts pods based on the
writable layer usage of all containers.
Without imagefs
If nodefs is triggering evictions, the kubelet sorts pods based on their total
disk usage (local volumes + logs & writable layer of all containers)
Minimum eviction reclaim
In some cases, pod eviction only reclaims a small amount of the starved resource.
This can lead to the kubelet repeatedly hitting the configured eviction thresholds
and triggering multiple evictions.
You can use the --eviction-minimum-reclaim flag or a kubelet config file
to configure a minimum reclaim amount for each resource. When the kubelet notices
that a resource is starved, it continues to reclaim that resource until it
reclaims the quantity you specify.
For example, the following configuration sets minimum reclaim amounts:
In this example, if the nodefs.available signal meets the eviction threshold,
the kubelet reclaims the resource until the signal reaches the threshold of 1Gi,
and then continues to reclaim the minimum amount of 500Mi it until the signal
reaches 1.5Gi.
Similarly, the kubelet reclaims the imagefs resource until the imagefs.available
signal reaches 102Gi.
The default eviction-minimum-reclaim is 0 for all resources.
Node out of memory behavior
If the node experiences an out of memory (OOM) event prior to the kubelet
being able to reclaim memory, the node depends on the oom_killer
to respond.
The kubelet sets an oom_score_adj value for each container based on the QoS for the pod.
Note: The kubelet also sets an oom_score_adj value of -997 for containers in Pods that have
system-node-criticalPriority.
If the kubelet can't reclaim memory before a node experiences OOM, the
oom_killer calculates an oom_score based on the percentage of memory it's
using on the node, and then adds the oom_score_adj to get an effective oom_score
for each container. It then kills the container with the highest score.
This means that containers in low QoS pods that consume a large amount of memory
relative to their scheduling requests are killed first.
Unlike pod eviction, if a container is OOM killed, the kubelet can restart it
based on its RestartPolicy.
Best practices
The following sections describe best practices for eviction configuration.
Schedulable resources and eviction policies
When you configure the kubelet with an eviction policy, you should make sure that
the scheduler will not schedule pods if they will trigger eviction because they
immediately induce memory pressure.
Consider the following scenario:
Node memory capacity: 10Gi
Operator wants to reserve 10% of memory capacity for system daemons (kernel, kubelet, etc.)
Operator wants to evict Pods at 95% memory utilization to reduce incidence of system OOM.
For this to work, the kubelet is launched as follows:
In this configuration, the --system-reserved flag reserves 1.5Gi of memory
for the system, which is 10% of the total memory + the eviction threshold amount.
The node can reach the eviction threshold if a pod is using more than its request,
or if the system is using more than 1Gi of memory, which makes the memory.available
signal fall below 500Mi and triggers the threshold.
DaemonSet
Pod Priority is a major factor in making eviction decisions. If you do not want
the kubelet to evict pods that belong to a DaemonSet, give those pods a high
enough priorityClass in the pod spec. You can also use a lower priorityClass
or the default to only allow DaemonSet pods to run when there are enough
resources.
Known issues
The following sections describe known issues related to out of resource handling.
kubelet may not observe memory pressure right away
By default, the kubelet polls cAdvisor to collect memory usage stats at a
regular interval. If memory usage increases within that window rapidly, the
kubelet may not observe MemoryPressure fast enough, and the OOMKiller
will still be invoked.
You can use the --kernel-memcg-notification flag to enable the memcg
notification API on the kubelet to get notified immediately when a threshold
is crossed.
If you are not trying to achieve extreme utilization, but a sensible measure of
overcommit, a viable workaround for this issue is to use the --kube-reserved
and --system-reserved flags to allocate memory for the system.
active_file memory is not considered as available memory
On Linux, the kernel tracks the number of bytes of file-backed memory on active
LRU list as the active_file statistic. The kubelet treats active_file memory
areas as not reclaimable. For workloads that make intensive use of block-backed
local storage, including ephemeral local storage, kernel-level caches of file
and block data means that many recently accessed cache pages are likely to be
counted as active_file. If enough of these kernel block buffers are on the
active LRU list, the kubelet is liable to observe this as high resource use and
taint the node as experiencing memory pressure - triggering pod eviction.
You can work around that behavior by setting the memory limit and memory request
the same for containers likely to perform intensive I/O activity. You will need
to estimate or measure an optimal memory limit value for that container.
API-initiated eviction is the process by which you use the Eviction API
to create an Eviction object that triggers graceful pod termination.
You can request eviction by calling the Eviction API directly, or programmatically
using a client of the API server, like the kubectl drain command. This
creates an Eviction object, which causes the API server to terminate the Pod.
Using the API to create an Eviction object for a Pod is like performing a
policy-controlled DELETE operation
on the Pod.
Calling the Eviction API
You can use a Kubernetes language client
to access the Kubernetes API and create an Eviction object. To do this, you
POST the attempted operation, similar to the following example:
When you request an eviction using the API, the API server performs admission
checks and responds in one of the following ways:
200 OK: the eviction is allowed, the Eviction subresource is created, and
the Pod is deleted, similar to sending a DELETE request to the Pod URL.
429 Too Many Requests: the eviction is not currently allowed because of the
configured PodDisruptionBudget.
You may be able to attempt the eviction again later. You might also see this
response because of API rate limiting.
500 Internal Server Error: the eviction is not allowed because there is a
misconfiguration, like if multiple PodDisruptionBudgets reference the same Pod.
If the Pod you want to evict isn't part of a workload that has a
PodDisruptionBudget, the API server always returns 200 OK and allows the
eviction.
If the API server allows the eviction, the Pod is deleted as follows:
The Pod resource in the API server is updated with a deletion timestamp,
after which the API server considers the Pod resource to be terminated. The
Pod resource is also marked with the configured grace period.
The kubelet on the node where the local Pod is running notices that the Pod
resource is marked for termination and starts to gracefully shut down the
local Pod.
While the kubelet is shutting the Pod down, the control plane removes the Pod
from Endpoint and
EndpointSlice
objects. As a result, controllers no longer consider the Pod as a valid object.
After the grace period for the Pod expires, the kubelet forcefully terminates
the local Pod.
The kubelet tells the API server to remove the Pod resource.
The API server deletes the Pod resource.
Troubleshooting stuck evictions
In some cases, your applications may enter a broken state, where the Eviction
API will only return 429 or 500 responses until you intervene. This can
happen if, for example, a ReplicaSet creates pods for your application but new
pods do not enter a Ready state. You may also notice this behavior in cases
where the last evicted Pod had a long termination grace period.
If you notice stuck evictions, try one of the following solutions:
Abort or pause the automated operation causing the issue. Investigate the stuck
application before you restart the operation.
Wait a while, then directly delete the Pod from your cluster control plane
instead of using the Eviction API.
In the scheduling-pluginNodeResourcesFit of kube-scheduler, there are two
scoring strategies that support the bin packing of resources: MostAllocated and RequestedToCapacityRatio.
Enabling bin packing using MostAllocated strategy
The MostAllocated strategy scores the nodes based on the utilization of resources, favoring the ones with higher allocation.
For each resource type, you can set a weight to modify its influence in the node score.
To set the MostAllocated strategy for the NodeResourcesFit plugin, use a
scheduler configuration similar to the following:
To learn more about other parameters and their default configuration, see the API documentation for
NodeResourcesFitArgs.
Enabling bin packing using RequestedToCapacityRatio
The RequestedToCapacityRatio strategy allows the users to specify the resources along with weights for
each resource to score nodes based on the request to capacity ratio. This
allows users to bin pack extended resources by using appropriate parameters
to improve the utilization of scarce resources in large clusters. It favors nodes according to a
configured function of the allocated resources. The behavior of the RequestedToCapacityRatio in
the NodeResourcesFit score function can be controlled by the
scoringStrategy field.
Within the scoringStrategy field, you can configure two parameters: requestedToCapacityRatioParam and
resources. The shape in requestedToCapacityRatioParam
parameter allows the user to tune the function as least requested or most
requested based on utilization and score values. The resources parameter
consists of name of the resource to be considered during scoring and weight
specify the weight of each resource.
Below is an example configuration that sets
the bin packing behavior for extended resources intel.com/foo and intel.com/bar
using the requestedToCapacityRatio field.
Referencing the KubeSchedulerConfiguration file with the kube-scheduler
flag --config=/path/to/config/file will pass the configuration to the
scheduler.
To learn more about other parameters and their default configuration, see the API documentation for
NodeResourcesFitArgs.
Tuning the score function
shape is used to specify the behavior of the RequestedToCapacityRatio function.
The above arguments give the node a score of 0 if utilization is 0% and 10 for
utilization 100%, thus enabling bin packing behavior. To enable least
requested the score value must be reversed as follows.
The weight parameter is optional and is set to 1 if not specified. Also, the
weight cannot be set to a negative value.
Node scoring for capacity allocation
This section is intended for those who want to understand the internal details
of this feature.
Below is an example of how the node score is calculated for a given set of values.
The scheduling framework is a pluggable architecture for the Kubernetes scheduler.
It adds a new set of "plugin" APIs to the existing scheduler. Plugins are compiled into the scheduler. The APIs allow most scheduling features to be implemented as plugins, while keeping the
scheduling "core" lightweight and maintainable. Refer to the design proposal of the
scheduling framework for more technical information on the design of the
framework.
Framework workflow
The Scheduling Framework defines a few extension points. Scheduler plugins
register to be invoked at one or more extension points. Some of these plugins
can change the scheduling decisions and some are informational only.
Each attempt to schedule one Pod is split into two phases, the scheduling
cycle and the binding cycle.
Scheduling Cycle & Binding Cycle
The scheduling cycle selects a node for the Pod, and the binding cycle applies
that decision to the cluster. Together, a scheduling cycle and binding cycle are
referred to as a "scheduling context".
Scheduling cycles are run serially, while binding cycles may run concurrently.
A scheduling or binding cycle can be aborted if the Pod is determined to
be unschedulable or if there is an internal error. The Pod will be returned to
the queue and retried.
Extension points
The following picture shows the scheduling context of a Pod and the extension
points that the scheduling framework exposes. In this picture "Filter" is
equivalent to "Predicate" and "Scoring" is equivalent to "Priority function".
One plugin may register at multiple extension points to perform more complex or
stateful tasks.
scheduling framework extension points
QueueSort
These plugins are used to sort Pods in the scheduling queue. A queue sort plugin
essentially provides a Less(Pod1, Pod2) function. Only one queue sort
plugin may be enabled at a time.
PreFilter
These plugins are used to pre-process info about the Pod, or to check certain
conditions that the cluster or the Pod must meet. If a PreFilter plugin returns
an error, the scheduling cycle is aborted.
Filter
These plugins are used to filter out nodes that cannot run the Pod. For each
node, the scheduler will call filter plugins in their configured order. If any
filter plugin marks the node as infeasible, the remaining plugins will not be
called for that node. Nodes may be evaluated concurrently.
PostFilter
These plugins are called after Filter phase, but only when no feasible nodes
were found for the pod. Plugins are called in their configured order. If
any postFilter plugin marks the node as Schedulable, the remaining plugins
will not be called. A typical PostFilter implementation is preemption, which
tries to make the pod schedulable by preempting other Pods.
PreScore
These plugins are used to perform "pre-scoring" work, which generates a sharable
state for Score plugins to use. If a PreScore plugin returns an error, the
scheduling cycle is aborted.
Score
These plugins are used to rank nodes that have passed the filtering phase. The
scheduler will call each scoring plugin for each node. There will be a well
defined range of integers representing the minimum and maximum scores. After the
NormalizeScore phase, the scheduler will combine node
scores from all plugins according to the configured plugin weights.
NormalizeScore
These plugins are used to modify scores before the scheduler computes a final
ranking of Nodes. A plugin that registers for this extension point will be
called with the Score results from the same plugin. This is called
once per plugin per scheduling cycle.
For example, suppose a plugin BlinkingLightScorer ranks Nodes based on how
many blinking lights they have.
funcScoreNode(_ *v1.pod, n *v1.Node) (int, error) {
returngetBlinkingLightCount(n)
}
However, the maximum count of blinking lights may be small compared to
NodeScoreMax. To fix this, BlinkingLightScorer should also register for this
extension point.
If any NormalizeScore plugin returns an error, the scheduling cycle is
aborted.
Note: Plugins wishing to perform "pre-reserve" work should use the
NormalizeScore extension point.
Reserve
A plugin that implements the Reserve extension has two methods, namely Reserve
and Unreserve, that back two informational scheduling phases called Reserve
and Unreserve, respectively. Plugins which maintain runtime state (aka "stateful
plugins") should use these phases to be notified by the scheduler when resources
on a node are being reserved and unreserved for a given Pod.
The Reserve phase happens before the scheduler actually binds a Pod to its
designated node. It exists to prevent race conditions while the scheduler waits
for the bind to succeed. The Reserve method of each Reserve plugin may succeed
or fail; if one Reserve method call fails, subsequent plugins are not executed
and the Reserve phase is considered to have failed. If the Reserve method of
all plugins succeed, the Reserve phase is considered to be successful and the
rest of the scheduling cycle and the binding cycle are executed.
The Unreserve phase is triggered if the Reserve phase or a later phase fails.
When this happens, the Unreserve method of all Reserve plugins will be
executed in the reverse order of Reserve method calls. This phase exists to
clean up the state associated with the reserved Pod.
Caution: The implementation of the Unreserve method in Reserve plugins must be
idempotent and may not fail.
Permit
Permit plugins are invoked at the end of the scheduling cycle for each Pod, to
prevent or delay the binding to the candidate node. A permit plugin can do one of
the three things:
approve
Once all Permit plugins approve a Pod, it is sent for binding.
deny
If any Permit plugin denies a Pod, it is returned to the scheduling queue.
This will trigger the Unreserve phase in Reserve plugins.
wait (with a timeout)
If a Permit plugin returns "wait", then the Pod is kept in an internal "waiting"
Pods list, and the binding cycle of this Pod starts but directly blocks until it
gets approved. If a timeout occurs, wait becomes deny
and the Pod is returned to the scheduling queue, triggering the
Unreserve phase in Reserve plugins.
Note: While any plugin can access the list of "waiting" Pods and approve them
(see FrameworkHandle), we expect only the permit
plugins to approve binding of reserved Pods that are in "waiting" state. Once a Pod
is approved, it is sent to the PreBind phase.
PreBind
These plugins are used to perform any work required before a Pod is bound. For
example, a pre-bind plugin may provision a network volume and mount it on the
target node before allowing the Pod to run there.
If any PreBind plugin returns an error, the Pod is rejected and
returned to the scheduling queue.
Bind
These plugins are used to bind a Pod to a Node. Bind plugins will not be called
until all PreBind plugins have completed. Each bind plugin is called in the
configured order. A bind plugin may choose whether or not to handle the given
Pod. If a bind plugin chooses to handle a Pod, the remaining bind plugins are
skipped.
PostBind
This is an informational extension point. Post-bind plugins are called after a
Pod is successfully bound. This is the end of a binding cycle, and can be used
to clean up associated resources.
Plugin API
There are two steps to the plugin API. First, plugins must register and get
configured, then they use the extension point interfaces. Extension point
interfaces have the following form.
type Plugin interface {
Name() string}
type QueueSortPlugin interface {
Plugin
Less(*v1.pod, *v1.pod) bool}
type PreFilterPlugin interface {
Plugin
PreFilter(context.Context, *framework.CycleState, *v1.pod) error}
// ...
Plugin configuration
You can enable or disable plugins in the scheduler configuration. If you are using
Kubernetes v1.18 or later, most scheduling
plugins are in use and
enabled by default.
In addition to default plugins, you can also implement your own scheduling
plugins and get them configured along with default plugins. You can visit
scheduler-plugins for more details.
If you are using Kubernetes v1.18 or later, you can configure a set of plugins as
a scheduler profile and then define multiple profiles to fit various kinds of workload.
Learn more at multiple profiles.
3.11.11 - Scheduler Performance Tuning
FEATURE STATE:Kubernetes v1.14 [beta]
kube-scheduler
is the Kubernetes default scheduler. It is responsible for placement of Pods
on Nodes in a cluster.
Nodes in a cluster that meet the scheduling requirements of a Pod are
called feasible Nodes for the Pod. The scheduler finds feasible Nodes
for a Pod and then runs a set of functions to score the feasible Nodes,
picking a Node with the highest score among the feasible ones to run
the Pod. The scheduler then notifies the API server about this decision
in a process called Binding.
This page explains performance tuning optimizations that are relevant for
large Kubernetes clusters.
In large clusters, you can tune the scheduler's behaviour balancing
scheduling outcomes between latency (new Pods are placed quickly) and
accuracy (the scheduler rarely makes poor placement decisions).
You configure this tuning setting via kube-scheduler setting
percentageOfNodesToScore. This KubeSchedulerConfiguration setting determines
a threshold for scheduling nodes in your cluster.
Setting the threshold
The percentageOfNodesToScore option accepts whole numeric values between 0
and 100. The value 0 is a special number which indicates that the kube-scheduler
should use its compiled-in default.
If you set percentageOfNodesToScore above 100, kube-scheduler acts as if you
had set a value of 100.
To change the value, edit the
kube-scheduler configuration file
and then restart the scheduler.
In many cases, the configuration file can be found at /etc/kubernetes/config/kube-scheduler.yaml.
After you have made this change, you can run
kubectl get pods -n kube-system | grep kube-scheduler
to verify that the kube-scheduler component is healthy.
Node scoring threshold
To improve scheduling performance, the kube-scheduler can stop looking for
feasible nodes once it has found enough of them. In large clusters, this saves
time compared to a naive approach that would consider every node.
You specify a threshold for how many nodes are enough, as a whole number percentage
of all the nodes in your cluster. The kube-scheduler converts this into an
integer number of nodes. During scheduling, if the kube-scheduler has identified
enough feasible nodes to exceed the configured percentage, the kube-scheduler
stops searching for more feasible nodes and moves on to the
scoring phase.
If you don't specify a threshold, Kubernetes calculates a figure using a
linear formula that yields 50% for a 100-node cluster and yields 10%
for a 5000-node cluster. The lower bound for the automatic value is 5%.
This means that, the kube-scheduler always scores at least 5% of your cluster no
matter how large the cluster is, unless you have explicitly set
percentageOfNodesToScore to be smaller than 5.
If you want the scheduler to score all nodes in your cluster, set
percentageOfNodesToScore to 100.
Example
Below is an example configuration that sets percentageOfNodesToScore to 50%.
percentageOfNodesToScore must be a value between 1 and 100 with the default
value being calculated based on the cluster size. There is also a hardcoded
minimum value of 50 nodes.
Note:
In clusters with less than 50 feasible nodes, the scheduler still
checks all the nodes because there are not enough feasible nodes to stop
the scheduler's search early.
In a small cluster, if you set a low value for percentageOfNodesToScore, your
change will have no or little effect, for a similar reason.
If your cluster has several hundred Nodes or fewer, leave this configuration option
at its default value. Making changes is unlikely to improve the
scheduler's performance significantly.
An important detail to consider when setting this value is that when a smaller
number of nodes in a cluster are checked for feasibility, some nodes are not
sent to be scored for a given Pod. As a result, a Node which could possibly
score a higher value for running the given Pod might not even be passed to the
scoring phase. This would result in a less than ideal placement of the Pod.
You should avoid setting percentageOfNodesToScore very low so that kube-scheduler
does not make frequent, poor Pod placement decisions. Avoid setting the
percentage to anything below 10%, unless the scheduler's throughput is critical
for your application and the score of nodes is not important. In other words, you
prefer to run the Pod on any Node as long as it is feasible.
How the scheduler iterates over Nodes
This section is intended for those who want to understand the internal details
of this feature.
In order to give all the Nodes in a cluster a fair chance of being considered
for running Pods, the scheduler iterates over the nodes in a round robin
fashion. You can imagine that Nodes are in an array. The scheduler starts from
the start of the array and checks feasibility of the nodes until it finds enough
Nodes as specified by percentageOfNodesToScore. For the next Pod, the
scheduler continues from the point in the Node array that it stopped at when
checking feasibility of Nodes for the previous Pod.
If Nodes are in multiple zones, the scheduler iterates over Nodes in various
zones to ensure that Nodes from different zones are considered in the
feasibility checks. As an example, consider six nodes in two zones:
Zone 1: Node 1, Node 2, Node 3, Node 4
Zone 2: Node 5, Node 6
The Scheduler evaluates feasibility of the nodes in this order:
Node 1, Node 5, Node 2, Node 6, Node 3, Node 4
After going over all the Nodes, it goes back to Node 1.
Lower-level detail relevant to creating or administering a Kubernetes cluster.
The cluster administration overview is for anyone creating or administering a Kubernetes cluster.
It assumes some familiarity with core Kubernetes concepts.
Planning a cluster
See the guides in Setup for examples of how to plan, set up, and configure
Kubernetes clusters. The solutions listed in this article are called distros.
Note: Not all distros are actively maintained. Choose distros which have been tested with a recent
version of Kubernetes.
Before choosing a guide, here are some considerations:
Do you want to try out Kubernetes on your computer, or do you want to build a high-availability,
multi-node cluster? Choose distros best suited for your needs.
Will you be using a hosted Kubernetes cluster, such as
Google Kubernetes Engine, or hosting your own cluster?
Will your cluster be on-premises, or in the cloud (IaaS)? Kubernetes does not directly
support hybrid clusters. Instead, you can set up multiple clusters.
If you are configuring Kubernetes on-premises, consider which
networking model fits best.
Will you be running Kubernetes on "bare metal" hardware or on virtual machines (VMs)?
Do you want to run a cluster, or do you expect to do active development of Kubernetes project code?
If the latter, choose an actively-developed distro. Some distros only use binary releases, but
offer a greater variety of choices.
Familiarize yourself with the components needed to run a cluster.
To learn how to generate certificates for your cluster, see Certificates.
3.12.2 - Managing Resources
You've deployed your application and exposed it via a service. Now what? Kubernetes provides a number of tools to help you manage your application deployment, including scaling and updating. Among the features that we will discuss in more depth are configuration files and labels.
Organizing resource configurations
Many applications require multiple resources to be created, such as a Deployment and a Service. Management of multiple resources can be simplified by grouping them together in the same file (separated by --- in YAML). For example:
service/my-nginx-svc created
deployment.apps/my-nginx created
The resources will be created in the order they appear in the file. Therefore, it's best to specify the service first, since that will ensure the scheduler can spread the pods associated with the service as they are created by the controller(s), such as Deployment.
kubectl will read any files with suffixes .yaml, .yml, or .json.
It is a recommended practice to put resources related to the same microservice or application tier into the same file, and to group all of the files associated with your application in the same directory. If the tiers of your application bind to each other using DNS, you can deploy all of the components of your stack together.
A URL can also be specified as a configuration source, which is handy for deploying directly from configuration files checked into GitHub:
Resource creation isn't the only operation that kubectl can perform in bulk. It can also extract resource names from configuration files in order to perform other operations, in particular to delete the same resources you created:
For larger numbers of resources, you'll find it easier to specify the selector (label query) specified using -l or --selector, to filter resources by their labels:
kubectl delete deployment,services -l app=nginx
deployment.apps "my-nginx" deleted
service "my-nginx-svc" deleted
Because kubectl outputs resource names in the same syntax it accepts, you can chain operations using $() or xargs:
kubectl get $(kubectl create -f docs/concepts/cluster-administration/nginx/ -o name | grep service)kubectl create -f docs/concepts/cluster-administration/nginx/ -o name | grep service | xargs -i kubectl get {}
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-nginx-svc LoadBalancer 10.0.0.208 <pending> 80/TCP 0s
With the above commands, we first create resources under examples/application/nginx/ and print the resources created with -o name output format
(print each resource as resource/name). Then we grep only the "service", and then print it with kubectl get.
If you happen to organize your resources across several subdirectories within a particular directory, you can recursively perform the operations on the subdirectories also, by specifying --recursive or -R alongside the --filename,-f flag.
For instance, assume there is a directory project/k8s/development that holds all of the manifests needed for the development environment, organized by resource type:
By default, performing a bulk operation on project/k8s/development will stop at the first level of the directory, not processing any subdirectories. If we had tried to create the resources in this directory using the following command, we would have encountered an error:
kubectl apply -f project/k8s/development
error: you must provide one or more resources by argument or filename (.json|.yaml|.yml|stdin)
Instead, specify the --recursive or -R flag with the --filename,-f flag as such:
namespace/development created
namespace/staging created
configmap/my-config created
deployment.apps/my-deployment created
persistentvolumeclaim/my-pvc created
The examples we've used so far apply at most a single label to any resource. There are many scenarios where multiple labels should be used to distinguish sets from one another.
For instance, different applications would use different values for the app label, but a multi-tier application, such as the guestbook example, would additionally need to distinguish each tier. The frontend could carry the following labels:
labels:app:guestbooktier:frontend
while the Redis master and slave would have different tier labels, and perhaps even an additional role label:
labels:app:guestbooktier:backendrole:master
and
labels:app:guestbooktier:backendrole:slave
The labels allow us to slice and dice our resources along any dimension specified by a label:
kubectl apply -f examples/guestbook/all-in-one/guestbook-all-in-one.yaml
kubectl get pods -Lapp -Ltier -Lrole
NAME READY STATUS RESTARTS AGE
guestbook-redis-slave-2q2yf 1/1 Running 0 3m
guestbook-redis-slave-qgazl 1/1 Running 0 3m
Canary deployments
Another scenario where multiple labels are needed is to distinguish deployments of different releases or configurations of the same component. It is common practice to deploy a canary of a new application release (specified via image tag in the pod template) side by side with the previous release so that the new release can receive live production traffic before fully rolling it out.
For instance, you can use a track label to differentiate different releases.
The primary, stable release would have a track label with value as stable:
and then you can create a new release of the guestbook frontend that carries the track label with different value (i.e. canary), so that two sets of pods would not overlap:
The frontend service would span both sets of replicas by selecting the common subset of their labels (i.e. omitting the track label), so that the traffic will be redirected to both applications:
selector:app:guestbooktier:frontend
You can tweak the number of replicas of the stable and canary releases to determine the ratio of each release that will receive live production traffic (in this case, 3:1).
Once you're confident, you can update the stable track to the new application release and remove the canary one.
Sometimes existing pods and other resources need to be relabeled before creating new resources. This can be done with kubectl label.
For example, if you want to label all your nginx pods as frontend tier, run:
This first filters all pods with the label "app=nginx", and then labels them with the "tier=fe".
To see the pods you labeled, run:
kubectl get pods -l app=nginx -L tier
NAME READY STATUS RESTARTS AGE TIER
my-nginx-2035384211-j5fhi 1/1 Running 0 23m fe
my-nginx-2035384211-u2c7e 1/1 Running 0 23m fe
my-nginx-2035384211-u3t6x 1/1 Running 0 23m fe
This outputs all "app=nginx" pods, with an additional label column of pods' tier (specified with -L or --label-columns).
Sometimes you would want to attach annotations to resources. Annotations are arbitrary non-identifying metadata for retrieval by API clients such as tools, libraries, etc. This can be done with kubectl annotate. For example:
When load on your application grows or shrinks, use kubectl to scale your application. For instance, to decrease the number of nginx replicas from 3 to 1, do:
kubectl scale deployment/my-nginx --replicas=1
deployment.apps/my-nginx scaled
Now you only have one pod managed by the deployment.
kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
my-nginx-2035384211-j5fhi 1/1 Running 0 30m
To have the system automatically choose the number of nginx replicas as needed, ranging from 1 to 3, do:
Sometimes it's necessary to make narrow, non-disruptive updates to resources you've created.
kubectl apply
It is suggested to maintain a set of configuration files in source control
(see configuration as code),
so that they can be maintained and versioned along with the code for the resources they configure.
Then, you can use kubectl apply to push your configuration changes to the cluster.
This command will compare the version of the configuration that you're pushing with the previous version and apply the changes you've made, without overwriting any automated changes to properties you haven't specified.
Note that kubectl apply attaches an annotation to the resource in order to determine the changes to the configuration since the previous invocation. When it's invoked, kubectl apply does a three-way diff between the previous configuration, the provided input and the current configuration of the resource, in order to determine how to modify the resource.
Currently, resources are created without this annotation, so the first invocation of kubectl apply will fall back to a two-way diff between the provided input and the current configuration of the resource. During this first invocation, it cannot detect the deletion of properties set when the resource was created. For this reason, it will not remove them.
All subsequent calls to kubectl apply, and other commands that modify the configuration, such as kubectl replace and kubectl edit, will update the annotation, allowing subsequent calls to kubectl apply to detect and perform deletions using a three-way diff.
kubectl edit
Alternatively, you may also update resources with kubectl edit:
kubectl edit deployment/my-nginx
This is equivalent to first get the resource, edit it in text editor, and then apply the resource with the updated version:
kubectl get deployment my-nginx -o yaml > /tmp/nginx.yaml
vi /tmp/nginx.yaml
# do some edit, and then save the filekubectl apply -f /tmp/nginx.yaml
deployment.apps/my-nginx configured
rm /tmp/nginx.yaml
This allows you to do more significant changes more easily. Note that you can specify the editor with your EDITOR or KUBE_EDITOR environment variables.
For more information, please see kubectl edit document.
In some cases, you may need to update resource fields that cannot be updated once initialized, or you may want to make a recursive change immediately, such as to fix broken pods created by a Deployment. To change such fields, use replace --force, which deletes and re-creates the resource. In this case, you can modify your original configuration file:
Updating your application without a service outage
At some point, you'll eventually need to update your deployed application, typically by specifying a new image or image tag, as in the canary deployment scenario above. kubectl supports several update operations, each of which is applicable to different scenarios.
We'll guide you through how to create and update applications with Deployments.
Let's say you were running version 1.14.2 of nginx:
To update to version 1.16.1, change .spec.template.spec.containers[0].image from nginx:1.14.2 to nginx:1.16.1 using the previous kubectl commands.
kubectl edit deployment/my-nginx
That's it! The Deployment will declaratively update the deployed nginx application progressively behind the scene. It ensures that only a certain number of old replicas may be down while they are being updated, and only a certain number of new replicas may be created above the desired number of pods. To learn more details about it, visit Deployment page.
Networking is a central part of Kubernetes, but it can be challenging to
understand exactly how it is expected to work. There are 4 distinct networking
problems to address:
Highly-coupled container-to-container communications: this is solved by
Pods and localhost communications.
Pod-to-Pod communications: this is the primary focus of this document.
Pod-to-Service communications: this is covered by services.
External-to-Service communications: this is covered by services.
Kubernetes is all about sharing machines between applications. Typically,
sharing machines requires ensuring that two applications do not try to use the
same ports. Coordinating ports across multiple developers is very difficult to
do at scale and exposes users to cluster-level issues outside of their control.
Dynamic port allocation brings a lot of complications to the system - every
application has to take ports as flags, the API servers have to know how to
insert dynamic port numbers into configuration blocks, services have to know
how to find each other, etc. Rather than deal with this, Kubernetes takes a
different approach.
To learn about the Kubernetes networking model, see here.
How to implement the Kubernetes network model
The network model is implemented by the container runtime on each node. The most common container runtimes use Container Network Interface (CNI) plugins to manage their network and security capabilities. Many different CNI plugins exist from many different vendors. Some of these provide only basic features of adding and removing network interfaces, while others provide more sophisticated solutions, such as integration with other container orchestration systems, running multiple CNI plugins, advanced IPAM features etc.
See this page for a non-exhaustive list of networking addons supported by Kubernetes.
What's next
The early design of the networking model and its rationale, and some future
plans are described in more detail in the
networking design document.
3.12.4 - Logging Architecture
Application logs can help you understand what is happening inside your application. The logs are particularly useful for debugging problems and monitoring cluster activity. Most modern applications have some kind of logging mechanism. Likewise, container engines are designed to support logging. The easiest and most adopted logging method for containerized applications is writing to standard output and standard error streams.
However, the native functionality provided by a container engine or runtime is usually not enough for a complete logging solution.
For example, you may want to access your application's logs if a container crashes, a pod gets evicted, or a node dies.
In a cluster, logs should have a separate storage and lifecycle independent of nodes, pods, or containers. This concept is called cluster-level logging.
Cluster-level logging architectures require a separate backend to store, analyze, and query logs. Kubernetes
does not provide a native storage solution for log data. Instead, there are many logging solutions that
integrate with Kubernetes. The following sections describe how to handle and store logs on nodes.
Basic logging in Kubernetes
This example uses a Pod specification with a container
to write text to the standard output stream once per second.
To fetch the logs, use the kubectl logs command, as follows:
kubectl logs counter
The output is:
0: Mon Jan 1 00:00:00 UTC 2001
1: Mon Jan 1 00:00:01 UTC 2001
2: Mon Jan 1 00:00:02 UTC 2001
...
You can use kubectl logs --previous to retrieve logs from a previous instantiation of a container.
If your pod has multiple containers, specify which container's logs you want to access by
appending a container name to the command, with a -c flag, like so:
A container engine handles and redirects any output generated to a containerized application's stdout and stderr streams.
For example, the Docker container engine redirects those two streams to a logging driver, which is configured in Kubernetes to write to a file in JSON format.
Note: The Docker JSON logging driver treats each line as a separate message. When using the Docker logging driver, there is no direct support for multi-line messages. You need to handle multi-line messages at the logging agent level or higher.
By default, if a container restarts, the kubelet keeps one terminated container with its logs. If a pod is evicted from the node, all corresponding containers are also evicted, along with their logs.
An important consideration in node-level logging is implementing log rotation,
so that logs don't consume all available storage on the node. Kubernetes
is not responsible for rotating logs, but rather a deployment tool
should set up a solution to address that.
For example, in Kubernetes clusters, deployed by the kube-up.sh script,
there is a logrotate
tool configured to run each hour. You can also set up a container runtime to
rotate an application's logs automatically.
As an example, you can find detailed information about how kube-up.sh sets
up logging for COS image on GCP in the corresponding
configure-helper script.
When using a CRI container runtime, the kubelet is responsible for rotating the logs and managing the logging directory structure.
The kubelet sends this information to the CRI container runtime and the runtime writes the container logs to the given location.
The two kubelet parameters containerLogMaxSize and containerLogMaxFiles
in kubelet config file
can be used to configure the maximum size for each log file and the maximum number of files allowed for each container respectively.
When you run kubectl logs as in
the basic logging example, the kubelet on the node handles the request and
reads directly from the log file. The kubelet returns the content of the log file.
Note: If an external system has performed the rotation or a CRI container runtime is used,
only the contents of the latest log file will be available through
kubectl logs. For example, if there's a 10MB file, logrotate performs
the rotation and there are two files: one file that is 10MB in size and a second file that is empty.
kubectl logs returns the latest log file which in this example is an empty response.
System component logs
There are two types of system components: those that run in a container and those
that do not run in a container. For example:
The Kubernetes scheduler and kube-proxy run in a container.
The kubelet and container runtime do not run in containers.
On machines with systemd, the kubelet and container runtime write to journald. If
systemd is not present, the kubelet and container runtime write to .log files
in the /var/log directory. System components inside containers always write
to the /var/log directory, bypassing the default logging mechanism.
They use the klog
logging library. You can find the conventions for logging severity for those
components in the development docs on logging.
Similar to the container logs, system component logs in the /var/log
directory should be rotated. In Kubernetes clusters brought up by
the kube-up.sh script, those logs are configured to be rotated by
the logrotate tool daily or once the size exceeds 100MB.
Cluster-level logging architectures
While Kubernetes does not provide a native solution for cluster-level logging, there are several common approaches you can consider. Here are some options:
Use a node-level logging agent that runs on every node.
Include a dedicated sidecar container for logging in an application pod.
Push logs directly to a backend from within an application.
Using a node logging agent
You can implement cluster-level logging by including a node-level logging agent on each node. The logging agent is a dedicated tool that exposes logs or pushes logs to a backend. Commonly, the logging agent is a container that has access to a directory with log files from all of the application containers on that node.
Because the logging agent must run on every node, it is recommended to run the agent
as a DaemonSet.
Node-level logging creates only one agent per node and doesn't require any changes to the applications running on the node.
Containers write to stdout and stderr, but with no agreed format. A node-level agent collects these logs and forwards them for aggregation.
Using a sidecar container with the logging agent
You can use a sidecar container in one of the following ways:
The sidecar container streams application logs to its own stdout.
The sidecar container runs a logging agent, which is configured to pick up logs from an application container.
Streaming sidecar container
By having your sidecar containers write to their own stdout and stderr
streams, you can take advantage of the kubelet and the logging agent that
already run on each node. The sidecar containers read logs from a file, a socket,
or journald. Each sidecar container prints a log to its own stdout or stderr stream.
This approach allows you to separate several log streams from different
parts of your application, some of which can lack support
for writing to stdout or stderr. The logic behind redirecting logs
is minimal, so it's not a significant overhead. Additionally, because
stdout and stderr are handled by the kubelet, you can use built-in tools
like kubectl logs.
For example, a pod runs a single container, and the container
writes to two different log files using two different formats. Here's a
configuration file for the Pod:
apiVersion:v1kind:Podmetadata:name:counterspec:containers:- name:countimage:busybox:1.28args:- /bin/sh- -c- > i=0;
while true;
do
echo "$i: $(date)" >> /var/log/1.log;
echo "$(date) INFO $i" >> /var/log/2.log;
i=$((i+1));
sleep 1;
donevolumeMounts:- name:varlogmountPath:/var/logvolumes:- name:varlogemptyDir:{}
It is not recommended to write log entries with different formats to the same log
stream, even if you managed to redirect both components to the stdout stream of
the container. Instead, you can create two sidecar containers. Each sidecar
container could tail a particular log file from a shared volume and then redirect
the logs to its own stdout stream.
Here's a configuration file for a pod that has two sidecar containers:
Now when you run this pod, you can access each log stream separately by
running the following commands:
kubectl logs counter count-log-1
The output is:
0: Mon Jan 1 00:00:00 UTC 2001
1: Mon Jan 1 00:00:01 UTC 2001
2: Mon Jan 1 00:00:02 UTC 2001
...
kubectl logs counter count-log-2
The output is:
Mon Jan 1 00:00:00 UTC 2001 INFO 0
Mon Jan 1 00:00:01 UTC 2001 INFO 1
Mon Jan 1 00:00:02 UTC 2001 INFO 2
...
The node-level agent installed in your cluster picks up those log streams
automatically without any further configuration. If you like, you can configure
the agent to parse log lines depending on the source container.
Note, that despite low CPU and memory usage (order of a couple of millicores
for cpu and order of several megabytes for memory), writing logs to a file and
then streaming them to stdout can double disk usage. If you have
an application that writes to a single file, it's recommended to set
/dev/stdout as the destination rather than implement the streaming sidecar
container approach.
Sidecar containers can also be used to rotate log files that cannot be
rotated by the application itself. An example of this approach is a small container running logrotate periodically.
However, it's recommended to use stdout and stderr directly and leave rotation
and retention policies to the kubelet.
Sidecar container with a logging agent
If the node-level logging agent is not flexible enough for your situation, you
can create a sidecar container with a separate logging agent that you have
configured specifically to run with your application.
Note: Using a logging agent in a sidecar container can lead
to significant resource consumption. Moreover, you won't be able to access
those logs using kubectl logs because they are not controlled
by the kubelet.
Here are two configuration files that you can use to implement a sidecar container with a logging agent. The first file contains
a ConfigMap to configure fluentd.
apiVersion:v1kind:ConfigMapmetadata:name:fluentd-configdata:fluentd.conf:| <source>
type tail
format none
path /var/log/1.log
pos_file /var/log/1.log.pos
tag count.format1
</source>
<source>
type tail
format none
path /var/log/2.log
pos_file /var/log/2.log.pos
tag count.format2
</source>
<match **>
type google_cloud
</match>
The second file describes a pod that has a sidecar container running fluentd.
The pod mounts a volume where fluentd can pick up its configuration data.
In the sample configurations, you can replace fluentd with any logging agent, reading from any source inside an application container.
Exposing logs directly from the application
Cluster-logging that exposes or pushes logs directly from every application is outside the scope of Kubernetes.
3.12.5 - Metrics For Kubernetes System Components
System component metrics can give a better look into what is happening inside them. Metrics are particularly useful for building dashboards and alerts.
Kubernetes components emit metrics in Prometheus format.
This format is structured plain text, designed so that people and machines can both read it.
Metrics in Kubernetes
In most cases metrics are available on /metrics endpoint of the HTTP server. For components that doesn't expose endpoint by default it can be enabled using --bind-address flag.
In a production environment you may want to configure Prometheus Server or some other metrics scraper
to periodically gather these metrics and make them available in some kind of time series database.
Note that kubelet also exposes metrics in /metrics/cadvisor, /metrics/resource and /metrics/probes endpoints. Those metrics do not have same lifecycle.
If your cluster uses RBAC, reading metrics requires authorization via a user, group or ServiceAccount with a ClusterRole that allows accessing /metrics.
For example:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:prometheusrules:- nonResourceURLs:- "/metrics"verbs:- get
Alpha metrics have no stability guarantees. These metrics can be modified or deleted at any time.
Stable metrics are guaranteed to not change. This means:
A stable metric without a deprecated signature will not be deleted or renamed
A stable metric's type will not be modified
Deprecated metrics are slated for deletion, but are still available for use.
These metrics include an annotation about the version in which they became deprecated.
For example:
Before deprecation
# HELP some_counter this counts things
# TYPE some_counter counter
some_counter 0
After deprecation
# HELP some_counter (Deprecated since 1.15.0) this counts things
# TYPE some_counter counter
some_counter 0
Hidden metrics are no longer published for scraping, but are still available for use. To use a hidden metric, please refer to the Show hidden metrics section.
Deleted metrics are no longer published and cannot be used.
Show hidden metrics
As described above, admins can enable hidden metrics through a command-line flag on a specific binary. This intends to be used as an escape hatch for admins if they missed the migration of the metrics deprecated in the last release.
The flag show-hidden-metrics-for-version takes a version for which you want to show metrics deprecated in that release. The version is expressed as x.y, where x is the major version, y is the minor version. The patch version is not needed even though a metrics can be deprecated in a patch release, the reason for that is the metrics deprecation policy runs against the minor release.
The flag can only take the previous minor version as it's value. All metrics hidden in previous will be emitted if admins set the previous version to show-hidden-metrics-for-version. The too old version is not allowed because this violates the metrics deprecated policy.
Take metric A as an example, here assumed that A is deprecated in 1.n. According to metrics deprecated policy, we can reach the following conclusion:
In release 1.n, the metric is deprecated, and it can be emitted by default.
In release 1.n+1, the metric is hidden by default and it can be emitted by command line show-hidden-metrics-for-version=1.n.
In release 1.n+2, the metric should be removed from the codebase. No escape hatch anymore.
If you're upgrading from release 1.12 to 1.13, but still depend on a metric A deprecated in 1.12, you should set hidden metrics via command line: --show-hidden-metrics=1.12 and remember to remove this metric dependency before upgrading to 1.14
Disable accelerator metrics
The kubelet collects accelerator metrics through cAdvisor. To collect these metrics, for accelerators like NVIDIA GPUs, kubelet held an open handle on the driver. This meant that in order to perform infrastructure changes (for example, updating the driver), a cluster administrator needed to stop the kubelet agent.
The responsibility for collecting accelerator metrics now belongs to the vendor rather than the kubelet. Vendors must provide a container that collects metrics and exposes them to the metrics service (for example, Prometheus).
Controller manager metrics provide important insight into the performance and health of the controller manager.
These metrics include common Go language runtime metrics such as go_routine count and controller specific metrics such as
etcd request latencies or Cloudprovider (AWS, GCE, OpenStack) API latencies that can be used
to gauge the health of a cluster.
Starting from Kubernetes 1.7, detailed Cloudprovider metrics are available for storage operations for GCE, AWS, Vsphere and OpenStack.
These metrics can be used to monitor health of persistent volume operations.
The scheduler exposes optional metrics that reports the requested resources and the desired limits of all running pods. These metrics can be used to build capacity planning dashboards, assess current or historical scheduling limits, quickly identify workloads that cannot schedule due to lack of resources, and compare actual usage to the pod's request.
The kube-scheduler identifies the resource requests and limits configured for each Pod; when either a request or limit is non-zero, the kube-scheduler reports a metrics timeseries. The time series is labelled by:
namespace
pod name
the node where the pod is scheduled or an empty string if not yet scheduled
priority
the assigned scheduler for that pod
the name of the resource (for example, cpu)
the unit of the resource if known (for example, cores)
Once a pod reaches completion (has a restartPolicy of Never or OnFailure and is in the Succeeded or Failed pod phase, or has been deleted and all containers have a terminated state) the series is no longer reported since the scheduler is now free to schedule other pods to run. The two metrics are called kube_pod_resource_request and kube_pod_resource_limit.
The metrics are exposed at the HTTP endpoint /metrics/resources and require the same authorization as the /metrics
endpoint on the scheduler. You must use the --show-hidden-metrics-for-version=1.20 flag to expose these alpha stability metrics.
Disabling metrics
You can explicitly turn off metrics via command line flag --disabled-metrics. This may be desired if, for example, a metric is causing a performance problem. The input is a list of disabled metrics (i.e. --disabled-metrics=metric1,metric2).
Metric cardinality enforcement
Metrics with unbounded dimensions could cause memory issues in the components they instrument. To limit resource use, you can use the --allow-label-value command line option to dynamically configure an allow-list of label values for a metric.
In alpha stage, the flag can only take in a series of mappings as metric label allow-list.
Each mapping is of the format <metric_name>,<label_name>=<allowed_labels> where
<allowed_labels> is a comma-separated list of acceptable label names.
The overall format looks like:
--allow-label-value <metric_name>,<label_name>='<allow_value1>, <allow_value2>...', <metric_name2>,<label_name>='<allow_value1>, <allow_value2>...', ....
Here is an example:
--allow-label-value number_count_metric,odd_number='1,3,5', number_count_metric,even_number='2,4,6', date_gauge_metric,weekend='Saturday,Sunday'
System component logs record events happening in cluster, which can be very useful for debugging.
You can configure log verbosity to see more or less detail.
Logs can be as coarse-grained as showing errors within a component, or as fine-grained as showing step-by-step traces of events (like HTTP access logs, pod state changes, controller actions, or scheduler decisions).
Klog
klog is the Kubernetes logging library. klog
generates log messages for the Kubernetes system components.
Kubernetes is in the process of simplifying logging in its components. The
following klog command line flags are
deprecated
starting with Kubernetes 1.23 and will be removed in a future release:
--add-dir-header
--alsologtostderr
--log-backtrace-at
--log-dir
--log-file
--log-file-max-size
--logtostderr
--one-output
--skip-headers
--skip-log-headers
--stderrthreshold
Output will always be written to stderr, regardless of the output
format. Output redirection is expected to be handled by the component which
invokes a Kubernetes component. This can be a POSIX shell or a tool like
systemd.
In some cases, for example a distroless container or a Windows system service,
those options are not available. Then the
kube-log-runner
binary can be used as wrapper around a Kubernetes component to redirect
output. A prebuilt binary is included in several Kubernetes base images under
its traditional name as /go-runner and as kube-log-runner in server and
node release archives.
This table shows how kube-log-runner invocations correspond to shell redirection:
I1025 00:15:15.525108 1 example.go:79] This is a message
which has a line break.
Structured Logging
FEATURE STATE:Kubernetes v1.23 [beta]
Warning:
Migration to structured log messages is an ongoing process. Not all log messages are structured in this version. When parsing log files, you must also handle unstructured log messages.
Log formatting and value serialization are subject to change.
Structured logging introduces a uniform structure in log messages allowing for programmatic extraction of information. You can store and process structured logs with less effort and cost.
The code which generates a log message determines whether it uses the traditional unstructured klog output
or structured logging.
The default formatting of structured log messages is as text, with a format that
is backward compatible with traditional klog:
I1025 00:15:15.525108 1 controller_utils.go:116] "Pod status updated" pod="kube-system/kubedns" status="ready"
Strings are quoted. Other values are formatted with
%+v, which may cause log messages to
continue on the next line depending on the data.
I1025 00:15:15.525108 1 example.go:116] "Example" data="This is text with a line break\nand \"quotation marks\"." someInt=1 someFloat=0.1 someStruct={StringField: First line,
second line.}
Contextual Logging
FEATURE STATE:Kubernetes v1.24 [alpha]
Contextual logging builds on top of structured logging. It is primarily about
how developers use logging calls: code based on that concept is more flexible
and supports additional use cases as described in the Contextual Logging
KEP.
If developers use additional functions like WithValues or WithName in
their components, then log entries contain additional information that gets
passed into functions by their caller.
Currently this is gated behind the StructuredLogging feature gate and
disabled by default. The infrastructure for this was added in 1.24 without
modifying components. The
component-base/logs/example
command demonstrates how to use the new logging calls and how a component
behaves that supports contextual logging.
$cd$GOPATH/src/k8s.io/kubernetes/staging/src/k8s.io/component-base/logs/example/cmd/
$ go run . --help
...
--feature-gates mapStringBool A set of key=value pairs that describe feature gates for alpha/experimental features. Options are:
AllAlpha=true|false (ALPHA - default=false)
AllBeta=true|false (BETA - default=false)
ContextualLogging=true|false (ALPHA - default=false)
$ go run . --feature-gates ContextualLogging=true...
I0404 18:00:02.916429 451895 logger.go:94] "example/myname: runtime" foo="bar" duration="1m0s"
I0404 18:00:02.916447 451895 logger.go:95] "example: another runtime" foo="bar" duration="1m0s"
The example prefix and foo="bar" were added by the caller of the function
which logs the runtime message and duration="1m0s" value, without having to
modify that function.
With contextual logging disable, WithValues and WithName do nothing and log
calls go through the global klog logger. Therefore this additional information
is not in the log output anymore:
$ go run . --feature-gates ContextualLogging=false...
I0404 18:03:31.171945 452150 logger.go:94] "runtime" duration="1m0s"
I0404 18:03:31.171962 452150 logger.go:95] "another runtime" duration="1m0s"
JSON log format
FEATURE STATE:Kubernetes v1.19 [alpha]
Warning:
JSON output does not support many standard klog flags. For list of unsupported klog flags, see the Command line tool reference.
Not all logs are guaranteed to be written in JSON format (for example, during process start). If you intend to parse logs, make sure you can handle log lines that are not JSON as well.
Field names and JSON serialization are subject to change.
The --logging-format=json flag changes the format of logs from klog native format to JSON format.
Example of JSON log format (pretty printed):
The -v flag controls log verbosity. Increasing the value increases the number of logged events. Decreasing the value decreases the number of logged events.
Increasing verbosity settings logs increasingly less severe events. A verbosity setting of 0 logs only critical events.
Log location
There are two types of system components: those that run in a container and those
that do not run in a container. For example:
The Kubernetes scheduler and kube-proxy run in a container.
On machines with systemd, the kubelet and container runtime write to journald.
Otherwise, they write to .log files in the /var/log directory.
System components inside containers always write to .log files in the /var/log directory,
bypassing the default logging mechanism.
Similar to the container logs, you should rotate system component logs in the /var/log directory.
In Kubernetes clusters created by the kube-up.sh script, log rotation is configured by the logrotate tool.
The logrotate tool rotates logs daily, or once the log size is greater than 100MB.
For a complete guide to collecting traces and using the collector, see
Getting Started with the OpenTelemetry Collector.
However, there are a few things to note that are specific to Kubernetes components.
By default, Kubernetes components export traces using the grpc exporter for OTLP on the
IANA OpenTelemetry port, 4317.
As an example, if the collector is running as a sidecar to a Kubernetes component,
the following receiver configuration will collect spans and log them to standard output:
receivers:otlp:protocols:grpc:exporters:# Replace this exporter with the exporter for your backendlogging:logLevel:debugservice:pipelines:traces:receivers:[otlp]exporters:[logging]
Component traces
kube-apiserver traces
The kube-apiserver generates spans for incoming HTTP requests, and for outgoing requests
to webhooks, etcd, and re-entrant requests. It propagates the
W3C Trace Context with outgoing requests
but does not make use of the trace context attached to incoming requests,
as the kube-apiserver is often a public endpoint.
Enabling tracing in the kube-apiserver
To enable tracing, enable the APIServerTracingfeature gate
on the kube-apiserver. Also, provide the kube-apiserver with a tracing configuration file
with --tracing-config-file=<path-to-config>. This is an example config that records
spans for 1 in 10000 requests, and uses the default OpenTelemetry endpoint:
The kubelet CRI interface and authenticated http servers are instrumented to generate
trace spans. As with the apiserver, the endpoint and sampling rate are configurable.
Trace context propagation is also configured. A parent span's sampling decision is always respected.
A provided tracing configuration sampling rate will apply to spans without a parent.
Enabled without a configured endpoint, the default OpenTelemetry Collector reciever address of "localhost:4317" is set.
Enabling tracing in the kubelet
To enable tracing, enable the KubeletTracingfeature gate
on the kubelet. Also, provide the kubelet with a
tracing configuration.
This is an example snippet of a kubelet config that records spans for 1 in 10000 requests, and uses the default OpenTelemetry endpoint:
Tracing instrumentation is still under active development, and may change
in a variety of ways. This includes span names, attached attributes,
instrumented endpoints, etc. Until this feature graduates to stable,
there are no guarantees of backwards compatibility for tracing instrumentation.
existence and implementation varies from cluster to cluster (e.g. nginx)
sits between all clients and one or more apiservers
acts as load balancer if there are several apiservers.
Cloud Load Balancers on external services:
are provided by some cloud providers (e.g. AWS ELB, Google Cloud Load Balancer)
are created automatically when the Kubernetes service has type LoadBalancer
usually supports UDP/TCP only
SCTP support is up to the load balancer implementation of the cloud provider
implementation varies by cloud provider.
Kubernetes users will typically not need to worry about anything other than the first two types. The cluster admin
will typically ensure that the latter types are set up correctly.
Requesting redirects
Proxies have replaced redirect capabilities. Redirects have been deprecated.
3.12.9 - API Priority and Fairness
FEATURE STATE:Kubernetes v1.20 [beta]
Controlling the behavior of the Kubernetes API server in an overload situation
is a key task for cluster administrators. The kube-apiserver has some controls available
(i.e. the --max-requests-inflight and --max-mutating-requests-inflight
command-line flags) to limit the amount of outstanding work that will be
accepted, preventing a flood of inbound requests from overloading and
potentially crashing the API server, but these flags are not enough to ensure
that the most important requests get through in a period of high traffic.
The API Priority and Fairness feature (APF) is an alternative that improves upon
aforementioned max-inflight limitations. APF classifies
and isolates requests in a more fine-grained way. It also introduces
a limited amount of queuing, so that no requests are rejected in cases
of very brief bursts. Requests are dispatched from queues using a
fair queuing technique so that, for example, a poorly-behaved
controller need not
starve others (even at the same priority level).
This feature is designed to work well with standard controllers, which
use informers and react to failures of API requests with exponential
back-off, and other clients that also work this way.
Caution: Some requests classified as "long-running"—such as remote
command execution or log tailing—are not subject to the API
Priority and Fairness filter. This is also true for the
--max-requests-inflight flag without the API Priority and Fairness
feature enabled. API Priority and Fairness does apply to watch
requests. When API Priority and Fairness is disabled, watch requests
are not subject to the --max-requests-inflight limit.
Enabling/Disabling API Priority and Fairness
The API Priority and Fairness feature is controlled by a feature gate
and is enabled by default. See Feature
Gates
for a general explanation of feature gates and how to enable and
disable them. The name of the feature gate for APF is
"APIPriorityAndFairness". This feature also involves an API Group with: (a) a
v1alpha1 version, disabled by default, and (b) v1beta1 and
v1beta2 versions, enabled by default. You can disable the feature
gate and API group beta versions by adding the following
command-line flags to your kube-apiserver invocation:
kube-apiserver \
--feature-gates=APIPriorityAndFairness=false\
--runtime-config=flowcontrol.apiserver.k8s.io/v1beta1=false,flowcontrol.apiserver.k8s.io/v1beta2=false\
# …and other flags as usual
Alternatively, you can enable the v1alpha1 version of the API group
with --runtime-config=flowcontrol.apiserver.k8s.io/v1alpha1=true.
The command-line flag --enable-priority-and-fairness=false will disable the
API Priority and Fairness feature, even if other flags have enabled it.
Concepts
There are several distinct features involved in the API Priority and Fairness
feature. Incoming requests are classified by attributes of the request using
FlowSchemas, and assigned to priority levels. Priority levels add a degree of
isolation by maintaining separate concurrency limits, so that requests assigned
to different priority levels cannot starve each other. Within a priority level,
a fair-queuing algorithm prevents requests from different flows from starving
each other, and allows for requests to be queued to prevent bursty traffic from
causing failed requests when the average load is acceptably low.
Priority Levels
Without APF enabled, overall concurrency in the API server is limited by the
kube-apiserver flags --max-requests-inflight and
--max-mutating-requests-inflight. With APF enabled, the concurrency limits
defined by these flags are summed and then the sum is divided up among a
configurable set of priority levels. Each incoming request is assigned to a
single priority level, and each priority level will only dispatch as many
concurrent requests as its configuration allows.
The default configuration, for example, includes separate priority levels for
leader-election requests, requests from built-in controllers, and requests from
Pods. This means that an ill-behaved Pod that floods the API server with
requests cannot prevent leader election or actions by the built-in controllers
from succeeding.
Seats Occupied by a Request
The above description of concurrency management is the baseline story.
In it, requests have different durations but are counted equally at
any given moment when comparing against a priority level's concurrency
limit. In the baseline story, each request occupies one unit of
concurrency. The word "seat" is used to mean one unit of concurrency,
inspired by the way each passenger on a train or aircraft takes up one
of the fixed supply of seats.
But some requests take up more than one seat. Some of these are list
requests that the server estimates will return a large number of
objects. These have been found to put an exceptionally heavy burden
on the server, among requests that take a similar amount of time to
run. For this reason, the server estimates the number of objects that
will be returned and considers the request to take a number of seats
that is proportional to that estimated number.
Execution time tweaks for watch requests
API Priority and Fairness manages watch requests, but this involves a
couple more excursions from the baseline behavior. The first concerns
how long a watch request is considered to occupy its seat. Depending
on request parameters, the response to a watch request may or may not
begin with create notifications for all the relevant pre-existing
objects. API Priority and Fairness considers a watch request to be
done with its seat once that initial burst of notifications, if any,
is over.
The normal notifications are sent in a concurrent burst to all
relevant watch response streams whenever the server is notified of an
object create/update/delete. To account for this work, API Priority
and Fairness considers every write request to spend some additional
time occupying seats after the actual writing is done. The server
estimates the number of notifications to be sent and adjusts the write
request's number of seats and seat occupancy time to include this
extra work.
Queuing
Even within a priority level there may be a large number of distinct sources of
traffic. In an overload situation, it is valuable to prevent one stream of
requests from starving others (in particular, in the relatively common case of a
single buggy client flooding the kube-apiserver with requests, that buggy client
would ideally not have much measurable impact on other clients at all). This is
handled by use of a fair-queuing algorithm to process requests that are assigned
the same priority level. Each request is assigned to a flow, identified by the
name of the matching FlowSchema plus a flow distinguisher — which
is either the requesting user, the target resource's namespace, or nothing — and the
system attempts to give approximately equal weight to requests in different
flows of the same priority level.
To enable distinct handling of distinct instances, controllers that have
many instances should authenticate with distinct usernames
After classifying a request into a flow, the API Priority and Fairness
feature then may assign the request to a queue. This assignment uses
a technique known as shuffle sharding, which makes relatively efficient use of
queues to insulate low-intensity flows from high-intensity flows.
The details of the queuing algorithm are tunable for each priority level, and
allow administrators to trade off memory use, fairness (the property that
independent flows will all make progress when total traffic exceeds capacity),
tolerance for bursty traffic, and the added latency induced by queuing.
Exempt requests
Some requests are considered sufficiently important that they are not subject to
any of the limitations imposed by this feature. These exemptions prevent an
improperly-configured flow control configuration from totally disabling an API
server.
Resources
The flow control API involves two kinds of resources.
PriorityLevelConfigurations
define the available isolation classes, the share of the available concurrency
budget that each can handle, and allow for fine-tuning queuing behavior.
FlowSchemas
are used to classify individual inbound requests, matching each to a
single PriorityLevelConfiguration. There is also a v1alpha1 version
of the same API group, and it has the same Kinds with the same syntax and
semantics.
PriorityLevelConfiguration
A PriorityLevelConfiguration represents a single isolation class. Each
PriorityLevelConfiguration has an independent limit on the number of outstanding
requests, and limitations on the number of queued requests.
Concurrency limits for PriorityLevelConfigurations are not specified in absolute
number of requests, but rather in "concurrency shares." The total concurrency
limit for the API Server is distributed among the existing
PriorityLevelConfigurations in proportion with these shares. This allows a
cluster administrator to scale up or down the total amount of traffic to a
server by restarting kube-apiserver with a different value for
--max-requests-inflight (or --max-mutating-requests-inflight), and all
PriorityLevelConfigurations will see their maximum allowed concurrency go up (or
down) by the same fraction.
Caution: With the Priority and Fairness feature enabled, the total concurrency limit for
the server is set to the sum of --max-requests-inflight and
--max-mutating-requests-inflight. There is no longer any distinction made
between mutating and non-mutating requests; if you want to treat them
separately for a given resource, make separate FlowSchemas that match the
mutating and non-mutating verbs respectively.
When the volume of inbound requests assigned to a single
PriorityLevelConfiguration is more than its permitted concurrency level, the
type field of its specification determines what will happen to extra requests.
A type of Reject means that excess traffic will immediately be rejected with
an HTTP 429 (Too Many Requests) error. A type of Queue means that requests
above the threshold will be queued, with the shuffle sharding and fair queuing techniques used
to balance progress between request flows.
The queuing configuration allows tuning the fair queuing algorithm for a
priority level. Details of the algorithm can be read in the
enhancement proposal, but in short:
Increasing queues reduces the rate of collisions between different flows, at
the cost of increased memory usage. A value of 1 here effectively disables the
fair-queuing logic, but still allows requests to be queued.
Increasing queueLengthLimit allows larger bursts of traffic to be
sustained without dropping any requests, at the cost of increased
latency and memory usage.
Changing handSize allows you to adjust the probability of collisions between
different flows and the overall concurrency available to a single flow in an
overload situation.
Note: A larger handSize makes it less likely for two individual flows to collide
(and therefore for one to be able to starve the other), but more likely that
a small number of flows can dominate the apiserver. A larger handSize also
potentially increases the amount of latency that a single high-traffic flow
can cause. The maximum number of queued requests possible from a
single flow is handSize * queueLengthLimit.
Following is a table showing an interesting collection of shuffle
sharding configurations, showing for each the probability that a
given mouse (low-intensity flow) is squished by the elephants (high-intensity flows) for
an illustrative collection of numbers of elephants. See
https://play.golang.org/p/Gi0PLgVHiUg , which computes this table.
Example Shuffle Sharding Configurations
HandSize
Queues
1 elephant
4 elephants
16 elephants
12
32
4.428838398950118e-09
0.11431348830099144
0.9935089607656024
10
32
1.550093439632541e-08
0.0626479840223545
0.9753101519027554
10
64
6.601827268370426e-12
0.00045571320990370776
0.49999929150089345
9
64
3.6310049976037345e-11
0.00045501212304112273
0.4282314876454858
8
64
2.25929199850899e-10
0.0004886697053040446
0.35935114681123076
8
128
6.994461389026097e-13
3.4055790161620863e-06
0.02746173137155063
7
128
1.0579122850901972e-11
6.960839379258192e-06
0.02406157386340147
7
256
7.597695465552631e-14
6.728547142019406e-08
0.0006709661542533682
6
256
2.7134626662687968e-12
2.9516464018476436e-07
0.0008895654642000348
6
512
4.116062922897309e-14
4.982983350480894e-09
2.26025764343413e-05
6
1024
6.337324016514285e-16
8.09060164312957e-11
4.517408062903668e-07
FlowSchema
A FlowSchema matches some inbound requests and assigns them to a
priority level. Every inbound request is tested against every
FlowSchema in turn, starting with those with numerically lowest ---
which we take to be the logically highest --- matchingPrecedence and
working onward. The first match wins.
Caution: Only the first matching FlowSchema for a given request matters. If multiple
FlowSchemas match a single inbound request, it will be assigned based on the one
with the highest matchingPrecedence. If multiple FlowSchemas with equal
matchingPrecedence match the same request, the one with lexicographically
smaller name will win, but it's better not to rely on this, and instead to
ensure that no two FlowSchemas have the same matchingPrecedence.
A FlowSchema matches a given request if at least one of its rules
matches. A rule matches if at least one of its subjectsand at least
one of its resourceRules or nonResourceRules (depending on whether the
incoming request is for a resource or non-resource URL) matches the request.
For the name field in subjects, and the verbs, apiGroups, resources,
namespaces, and nonResourceURLs fields of resource and non-resource rules,
the wildcard * may be specified to match all values for the given field,
effectively removing it from consideration.
A FlowSchema's distinguisherMethod.type determines how requests matching that
schema will be separated into flows. It may be
either ByUser, in which case one requesting user will not be able to starve
other users of capacity, or ByNamespace, in which case requests for resources
in one namespace will not be able to starve requests for resources in other
namespaces of capacity, or it may be blank (or distinguisherMethod may be
omitted entirely), in which case all requests matched by this FlowSchema will be
considered part of a single flow. The correct choice for a given FlowSchema
depends on the resource and your particular environment.
Defaults
Each kube-apiserver maintains two sorts of APF configuration objects:
mandatory and suggested.
Mandatory Configuration Objects
The four mandatory configuration objects reflect fixed built-in
guardrail behavior. This is behavior that the servers have before
those objects exist, and when those objects exist their specs reflect
this behavior. The four mandatory objects are as follows.
The mandatory exempt priority level is used for requests that are
not subject to flow control at all: they will always be dispatched
immediately. The mandatory exempt FlowSchema classifies all
requests from the system:masters group into this priority
level. You may define other FlowSchemas that direct other requests
to this priority level, if appropriate.
The mandatory catch-all priority level is used in combination with
the mandatory catch-all FlowSchema to make sure that every request
gets some kind of classification. Typically you should not rely on
this catch-all configuration, and should create your own catch-all
FlowSchema and PriorityLevelConfiguration (or use the suggested
global-default priority level that is installed by default) as
appropriate. Because it is not expected to be used normally, the
mandatory catch-all priority level has a very small concurrency
share and does not queue requests.
Suggested Configuration Objects
The suggested FlowSchemas and PriorityLevelConfigurations constitute a
reasonable default configuration. You can modify these and/or create
additional configuration objects if you want. If your cluster is
likely to experience heavy load then you should consider what
configuration will work best.
The suggested configuration groups requests into six priority levels:
The node-high priority level is for health updates from nodes.
The system priority level is for non-health requests from the
system:nodes group, i.e. Kubelets, which must be able to contact
the API server in order for workloads to be able to schedule on
them.
The leader-election priority level is for leader election requests from
built-in controllers (in particular, requests for endpoints, configmaps,
or leases coming from the system:kube-controller-manager or
system:kube-scheduler users and service accounts in the kube-system
namespace). These are important to isolate from other traffic because failures
in leader election cause their controllers to fail and restart, which in turn
causes more expensive traffic as the new controllers sync their informers.
The workload-high priority level is for other requests from built-in
controllers.
The workload-low priority level is for requests from any other service
account, which will typically include all requests from controllers running in
Pods.
The global-default priority level handles all other traffic, e.g.
interactive kubectl commands run by nonprivileged users.
The suggested FlowSchemas serve to steer requests into the above
priority levels, and are not enumerated here.
Maintenance of the Mandatory and Suggested Configuration Objects
Each kube-apiserver independently maintains the mandatory and
suggested configuration objects, using initial and periodic behavior.
Thus, in a situation with a mixture of servers of different versions
there may be thrashing as long as different servers have different
opinions of the proper content of these objects.
Each kube-apiserver makes an initial maintenance pass over the
mandatory and suggested configuration objects, and after that does
periodic maintenance (once per minute) of those objects.
For the mandatory configuration objects, maintenance consists of
ensuring that the object exists and, if it does, has the proper spec.
The server refuses to allow a creation or update with a spec that is
inconsistent with the server's guardrail behavior.
Maintenance of suggested configuration objects is designed to allow
their specs to be overridden. Deletion, on the other hand, is not
respected: maintenance will restore the object. If you do not want a
suggested configuration object then you need to keep it around but set
its spec to have minimal consequences. Maintenance of suggested
objects is also designed to support automatic migration when a new
version of the kube-apiserver is rolled out, albeit potentially with
thrashing while there is a mixed population of servers.
Maintenance of a suggested configuration object consists of creating
it --- with the server's suggested spec --- if the object does not
exist. OTOH, if the object already exists, maintenance behavior
depends on whether the kube-apiservers or the users control the
object. In the former case, the server ensures that the object's spec
is what the server suggests; in the latter case, the spec is left
alone.
The question of who controls the object is answered by first looking
for an annotation with key apf.kubernetes.io/autoupdate-spec. If
there is such an annotation and its value is true then the
kube-apiservers control the object. If there is such an annotation
and its value is false then the users control the object. If
neither of those condtions holds then the metadata.generation of the
object is consulted. If that is 1 then the kube-apiservers control
the object. Otherwise the users control the object. These rules were
introduced in release 1.22 and their consideration of
metadata.generation is for the sake of migration from the simpler
earlier behavior. Users who wish to control a suggested configuration
object should set its apf.kubernetes.io/autoupdate-spec annotation
to false.
Maintenance of a mandatory or suggested configuration object also
includes ensuring that it has an apf.kubernetes.io/autoupdate-spec
annotation that accurately reflects whether the kube-apiservers
control the object.
Maintenance also includes deleting objects that are neither mandatory
nor suggested but are annotated
apf.kubernetes.io/autoupdate-spec=true.
Health check concurrency exemption
The suggested configuration gives no special treatment to the health
check requests on kube-apiservers from their local kubelets --- which
tend to use the secured port but supply no credentials. With the
suggested config, these requests get assigned to the global-default
FlowSchema and the corresponding global-default priority level,
where other traffic can crowd them out.
If you add the following additional FlowSchema, this exempts those
requests from rate limiting.
Caution: Making this change also allows any hostile party to then send
health-check requests that match this FlowSchema, at any volume they
like. If you have a web traffic filter or similar external security
mechanism to protect your cluster's API server from general internet
traffic, you can configure rules to block any health check requests
that originate from outside your cluster.
Every HTTP response from an API server with the priority and fairness feature
enabled has two extra headers: X-Kubernetes-PF-FlowSchema-UID and
X-Kubernetes-PF-PriorityLevel-UID, noting the flow schema that matched the request
and the priority level to which it was assigned, respectively. The API objects'
names are not included in these headers in case the requesting user does not
have permission to view them, so when debugging you can use a command like
kubectl get flowschemas -o custom-columns="uid:{metadata.uid},name:{metadata.name}"kubectl get prioritylevelconfigurations -o custom-columns="uid:{metadata.uid},name:{metadata.name}"
to get a mapping of UIDs to names for both FlowSchemas and
PriorityLevelConfigurations.
Observability
Metrics
Note: In versions of Kubernetes before v1.20, the labels flow_schema and
priority_level were inconsistently named flowSchema and priorityLevel,
respectively. If you're running Kubernetes versions v1.19 and earlier, you
should refer to the documentation for your version.
When you enable the API Priority and Fairness feature, the kube-apiserver
exports additional metrics. Monitoring these can help you determine whether your
configuration is inappropriately throttling important traffic, or find
poorly-behaved workloads that may be harming system health.
apiserver_flowcontrol_rejected_requests_total is a counter vector
(cumulative since server start) of requests that were rejected,
broken down by the labels flow_schema (indicating the one that
matched the request), priority_level (indicating the one to which
the request was assigned), and reason. The reason label will be
have one of the following values:
queue-full, indicating that too many requests were already
queued,
concurrency-limit, indicating that the
PriorityLevelConfiguration is configured to reject rather than
queue excess requests, or
time-out, indicating that the request was still in the queue
when its queuing time limit expired.
apiserver_flowcontrol_dispatched_requests_total is a counter
vector (cumulative since server start) of requests that began
executing, broken down by the labels flow_schema (indicating the
one that matched the request) and priority_level (indicating the
one to which the request was assigned).
apiserver_current_inqueue_requests is a gauge vector of recent
high water marks of the number of queued requests, grouped by a
label named request_kind whose value is mutating or readOnly.
These high water marks describe the largest number seen in the one
second window most recently completed. These complement the older
apiserver_current_inflight_requests gauge vector that holds the
last window's high water mark of number of requests actively being
served.
apiserver_flowcontrol_read_vs_write_request_count_samples is a
histogram vector of observations of the then-current number of
requests, broken down by the labels phase (which takes on the
values waiting and executing) and request_kind (which takes on
the values mutating and readOnly). The observations are made
periodically at a high rate. Each observed value is a ratio,
between 0 and 1, of a number of requests divided by the
corresponding limit on the number of requests (queue length limit
for waiting and concurrency limit for executing).
apiserver_flowcontrol_read_vs_write_request_count_watermarks is a
histogram vector of high or low water marks of the number of
requests (divided by the corresponding limit to get a ratio in the
range 0 to 1) broken down by the labels phase (which takes on the
values waiting and executing) and request_kind (which takes on
the values mutating and readOnly); the label mark takes on
values high and low. The water marks are accumulated over
windows bounded by the times when an observation was added to
apiserver_flowcontrol_read_vs_write_request_count_samples. These
water marks show the range of values that occurred between samples.
apiserver_flowcontrol_current_inqueue_requests is a gauge vector
holding the instantaneous number of queued (not executing) requests,
broken down by the labels priority_level and flow_schema.
apiserver_flowcontrol_current_executing_requests is a gauge vector
holding the instantaneous number of executing (not waiting in a
queue) requests, broken down by the labels priority_level and
flow_schema.
apiserver_flowcontrol_request_concurrency_in_use is a gauge vector
holding the instantaneous number of occupied seats, broken down by
the labels priority_level and flow_schema.
apiserver_flowcontrol_priority_level_request_count_samples is a
histogram vector of observations of the then-current number of
requests broken down by the labels phase (which takes on the
values waiting and executing) and priority_level. Each
histogram gets observations taken periodically, up through the last
activity of the relevant sort. The observations are made at a high
rate. Each observed value is a ratio, between 0 and 1, of a number
of requests divided by the corresponding limit on the number of
requests (queue length limit for waiting and concurrency limit for
executing).
apiserver_flowcontrol_priority_level_request_count_watermarks is a
histogram vector of high or low water marks of the number of
requests (divided by the corresponding limit to get a ratio in the
range 0 to 1) broken down by the labels phase (which takes on the
values waiting and executing) and priority_level; the label
mark takes on values high and low. The water marks are
accumulated over windows bounded by the times when an observation
was added to
apiserver_flowcontrol_priority_level_request_count_samples. These
water marks show the range of values that occurred between samples.
apiserver_flowcontrol_priority_level_seat_count_samples is a
histogram vector of observations of the utilization of a priority
level's concurrency limit, broken down by priority_level. This
utilization is the fraction (number of seats occupied) /
(concurrency limit). This metric considers all stages of execution
(both normal and the extra delay at the end of a write to cover for
the corresponding notification work) of all requests except WATCHes;
for those it considers only the initial stage that delivers
notifications of pre-existing objects. Each histogram in the vector
is also labeled with phase: executing (there is no seat limit for
the waiting phase). Each histogram gets observations taken
periodically, up through the last activity of the relevant sort.
The observations
are made at a high rate.
apiserver_flowcontrol_priority_level_seat_count_watermarks is a
histogram vector of high or low water marks of the utilization of a
priority level's concurrency limit, broken down by priority_level
and mark (which takes on values high and low). Each histogram
in the vector is also labeled with phase: executing (there is no
seat limit for the waiting phase). The water marks are accumulated
over windows bounded by the times when an observation was added to
apiserver_flowcontrol_priority_level_seat_count_samples. These
water marks show the range of values that occurred between samples.
apiserver_flowcontrol_request_queue_length_after_enqueue is a
histogram vector of queue lengths for the queues, broken down by
the labels priority_level and flow_schema, as sampled by the
enqueued requests. Each request that gets queued contributes one
sample to its histogram, reporting the length of the queue immediately
after the request was added. Note that this produces different
statistics than an unbiased survey would.
Note: An outlier value in a histogram here means it is likely that a single flow
(i.e., requests by one user or for one namespace, depending on
configuration) is flooding the API server, and being throttled. By contrast,
if one priority level's histogram shows that all queues for that priority
level are longer than those for other priority levels, it may be appropriate
to increase that PriorityLevelConfiguration's concurrency shares.
apiserver_flowcontrol_request_concurrency_limit is a gauge vector
holding the computed concurrency limit (based on the API server's
total concurrency limit and PriorityLevelConfigurations' concurrency
shares), broken down by the label priority_level.
apiserver_flowcontrol_request_wait_duration_seconds is a histogram
vector of how long requests spent queued, broken down by the labels
flow_schema (indicating which one matched the request),
priority_level (indicating the one to which the request was
assigned), and execute (indicating whether the request started
executing).
Note: Since each FlowSchema always assigns requests to a single
PriorityLevelConfiguration, you can add the histograms for all the
FlowSchemas for one priority level to get the effective histogram for
requests assigned to that priority level.
apiserver_flowcontrol_request_execution_seconds is a histogram
vector of how long requests took to actually execute, broken down by
the labels flow_schema (indicating which one matched the request)
and priority_level (indicating the one to which the request was
assigned).
apiserver_flowcontrol_watch_count_samples is a histogram vector of
the number of active WATCH requests relevant to a given write,
broken down by flow_schema and priority_level.
apiserver_flowcontrol_work_estimated_seats is a histogram vector
of the number of estimated seats (maximum of initial and final stage
of execution) associated with requests, broken down by flow_schema
and priority_level.
apiserver_flowcontrol_request_dispatch_no_accommodation_total is a
counter vec of the number of events that in principle could have led
to a request being dispatched but did not, due to lack of available
concurrency, broken down by flow_schema and priority_level. The
relevant sorts of events are arrival of a request and completion of
a request.
Debug endpoints
When you enable the API Priority and Fairness feature, the kube-apiserver
serves the following additional paths at its HTTP[S] ports.
/debug/api_priority_and_fairness/dump_priority_levels - a listing of
all the priority levels and the current state of each. You can fetch like this:
kubectl get --raw /debug/api_priority_and_fairness/dump_priority_levels
For background information on design details for API priority and fairness, see
the enhancement proposal.
You can make suggestions and feature requests via SIG API Machinery
or the feature's slack channel.
3.12.10 - Installing Addons
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Add-ons extend the functionality of Kubernetes.
This page lists some of the available add-ons and links to their respective installation instructions. The list does not try to be exhaustive.
Networking and Network Policy
ACI provides integrated container networking and network security with Cisco ACI.
Antrea operates at Layer 3/4 to provide networking and security services for Kubernetes, leveraging Open vSwitch as the networking data plane.
Calico is a networking and network policy provider. Calico supports a flexible set of networking options so you can choose the most efficient option for your situation, including non-overlay and overlay networks, with or without BGP. Calico uses the same engine to enforce network policy for hosts, pods, and (if using Istio & Envoy) applications at the service mesh layer.
Canal unites Flannel and Calico, providing networking and network policy.
Cilium is a L3 network and network policy plugin that can enforce HTTP/API/L7 policies transparently. Both routing and overlay/encapsulation mode are supported, and it can work on top of other CNI plugins.
CNI-Genie enables Kubernetes to seamlessly connect to a choice of CNI plugins, such as Calico, Canal, Flannel, or Weave.
Contiv provides configurable networking (native L3 using BGP, overlay using vxlan, classic L2, and Cisco-SDN/ACI) for various use cases and a rich policy framework. Contiv project is fully open sourced. The installer provides both kubeadm and non-kubeadm based installation options.
Contrail, based on Tungsten Fabric, is an open source, multi-cloud network virtualization and policy management platform. Contrail and Tungsten Fabric are integrated with orchestration systems such as Kubernetes, OpenShift, OpenStack and Mesos, and provide isolation modes for virtual machines, containers/pods and bare metal workloads.
Flannel is an overlay network provider that can be used with Kubernetes.
Knitter is a plugin to support multiple network interfaces in a Kubernetes pod.
Multus is a Multi plugin for multiple network support in Kubernetes to support all CNI plugins (e.g. Calico, Cilium, Contiv, Flannel), in addition to SRIOV, DPDK, OVS-DPDK and VPP based workloads in Kubernetes.
OVN-Kubernetes is a networking provider for Kubernetes based on OVN (Open Virtual Network), a virtual networking implementation that came out of the Open vSwitch (OVS) project. OVN-Kubernetes provides an overlay based networking implementation for Kubernetes, including an OVS based implementation of load balancing and network policy.
Nodus is an OVN based CNI controller plugin to provide cloud native based Service function chaining(SFC).
NSX-T Container Plug-in (NCP) provides integration between VMware NSX-T and container orchestrators such as Kubernetes, as well as integration between NSX-T and container-based CaaS/PaaS platforms such as Pivotal Container Service (PKS) and OpenShift.
Nuage is an SDN platform that provides policy-based networking between Kubernetes Pods and non-Kubernetes environments with visibility and security monitoring.
Romana is a Layer 3 networking solution for pod networks that also supports the NetworkPolicy API.
Weave Net provides networking and network policy, will carry on working on both sides of a network partition, and does not require an external database.
Service Discovery
CoreDNS is a flexible, extensible DNS server which can be installed as the in-cluster DNS for pods.
Visualization & Control
Dashboard is a dashboard web interface for Kubernetes.
Weave Scope is a tool for graphically visualizing your containers, pods, services etc. Use it in conjunction with a Weave Cloud account or host the UI yourself.
Infrastructure
KubeVirt is an add-on to run virtual machines on Kubernetes. Usually run on bare-metal clusters.
There are several other add-ons documented in the deprecated cluster/addons directory.
Well-maintained ones should be linked to here. PRs welcome!
3.13 - Extending Kubernetes
Different ways to change the behavior of your Kubernetes cluster.
Kubernetes is highly configurable and extensible. As a result, there is rarely a need to fork or
submit patches to the Kubernetes project code.
This guide describes the options for customizing a Kubernetes cluster. It is aimed at
cluster operators who want to understand
how to adapt their Kubernetes cluster to the needs of their work environment. Developers who are
prospective Platform Developers or
Kubernetes Project Contributors will also
find it useful as an introduction to what extension points and patterns exist, and their
trade-offs and limitations.
Overview
Customization approaches can be broadly divided into configuration, which only involves changing
flags, local configuration files, or API resources; and extensions, which involve running
additional programs or services. This document is primarily about extensions.
Configuration
Configuration files and flags are documented in the Reference section of the online
documentation, under each binary:
Flags and configuration files may not always be changeable in a hosted Kubernetes service or a
distribution with managed installation. When they are changeable, they are usually only changeable
by the cluster administrator. Also, they are subject to change in future Kubernetes versions, and
setting them may require restarting processes. For those reasons, they should be used only when
there are no other options.
Built-in Policy APIs, such as ResourceQuota,
PodSecurityPolicies,
NetworkPolicy and Role-based Access Control
(RBAC), are built-in Kubernetes APIs.
APIs are typically used with hosted Kubernetes services and with managed Kubernetes installations.
They are declarative and use the same conventions as other Kubernetes resources like pods,
so new cluster configuration can be repeatable and be managed the same way as applications.
And, where they are stable, they enjoy a
defined support policy like other Kubernetes APIs.
For these reasons, they are preferred over configuration files and flags where suitable.
Extensions
Extensions are software components that extend and deeply integrate with Kubernetes.
They adapt it to support new types and new kinds of hardware.
Many cluster administrators use a hosted or distribution instance of Kubernetes.
These clusters come with extensions pre-installed. As a result, most Kubernetes
users will not need to install extensions and even fewer users will need to author new ones.
Extension Patterns
Kubernetes is designed to be automated by writing client programs. Any
program that reads and/or writes to the Kubernetes API can provide useful
automation. Automation can run on the cluster or off it. By following
the guidance in this doc you can write highly available and robust automation.
Automation generally works with any Kubernetes cluster, including hosted
clusters and managed installations.
There is a specific pattern for writing client programs that work well with
Kubernetes called the Controller pattern. Controllers typically read an
object's .spec, possibly do things, and then update the object's .status.
A controller is a client of Kubernetes. When Kubernetes is the client and calls out to a remote
service, it is called a Webhook. The remote service is called a Webhook Backend. Like
Controllers, Webhooks do add a point of failure.
In the webhook model, Kubernetes makes a network request to a remote service.
In the Binary Plugin model, Kubernetes executes a binary (program).
Binary plugins are used by the kubelet (e.g.
Flex Volume Plugins
and Network Plugins)
and by kubectl.
Below is a diagram showing how the extension points interact with the
Kubernetes control plane.
Extension Points
This diagram shows the extension points in a Kubernetes system.
Users often interact with the Kubernetes API using kubectl.
Kubectl plugins extend the kubectl binary.
They only affect the individual user's local environment, and so cannot enforce site-wide policies.
The API server handles all requests. Several types of extension points in the API server allow
authenticating requests, or blocking them based on their content, editing content, and handling
deletion. These are described in the API Access Extensions section.
The API server serves various kinds of resources. Built-in resource kinds, like pods, are
defined by the Kubernetes project and can't be changed. You can also add resources that you
define, or that other projects have defined, called Custom Resources, as explained in the
Custom Resources section. Custom Resources are often used with API access
extensions.
The Kubernetes scheduler decides which nodes to place pods on. There are several ways to extend
scheduling. These are described in the Scheduler Extensions section.
Much of the behavior of Kubernetes is implemented by programs called Controllers which are
clients of the API server. Controllers are often used in conjunction with Custom Resources.
The kubelet runs on servers, and helps pods appear like virtual servers with their own IPs on
the cluster network. Network Plugins allow for different implementations of
pod networking.
The kubelet also mounts and unmounts volumes for containers. New types of storage can be
supported via Storage Plugins.
If you are unsure where to start, this flowchart can help. Note that some solutions may involve
several types of extensions.
API Extensions
User-Defined Types
Consider adding a Custom Resource to Kubernetes if you want to define new controllers, application
configuration objects or other declarative APIs, and to manage them using Kubernetes tools, such
as kubectl.
Do not use a Custom Resource as data storage for application, user, or monitoring data.
The combination of a custom resource API and a control loop is called the
Operator pattern. The Operator pattern is used to manage
specific, usually stateful, applications. These custom APIs and control loops can also be used to
control other resources, such as storage or policies.
Changing Built-in Resources
When you extend the Kubernetes API by adding custom resources, the added resources always fall
into a new API Groups. You cannot replace or change existing API groups.
Adding an API does not directly let you affect the behavior of existing APIs (e.g. Pods), but API
Access Extensions do.
API Access Extensions
When a request reaches the Kubernetes API Server, it is first Authenticated, then Authorized, then
subject to various types of Admission Control. See
Controlling Access to the Kubernetes API
for more on this flow.
Each of these steps offers extension points.
Kubernetes has several built-in authentication methods that it supports. It can also sit behind an
authenticating proxy, and it can send a token from an Authorization header to a remote service for
verification (a webhook). All of these methods are covered in the
Authentication documentation.
Authentication
Authentication maps headers or certificates
in all requests to a username for the client making the request.
Kubernetes provides several built-in authentication methods, and an
Authentication webhook
method if those don't meet your needs.
Authorization
Authorization determines whether specific
users can read, write, and do other operations on API resources. It works at the level of whole
resources -- it doesn't discriminate based on arbitrary object fields. If the built-in
authorization options don't meet your needs, Authorization webhook
allows calling out to user-provided code to make an authorization decision.
Dynamic Admission Control
After a request is authorized, if it is a write operation, it also goes through
Admission Control steps.
In addition to the built-in steps, there are several extensions:
To make arbitrary admission control decisions, a general
Admission webhook
can be used. Admission Webhooks can reject creations or updates.
Infrastructure Extensions
Storage Plugins
Flex Volumes
allow users to mount volume types without built-in support by having the kubelet call a binary
plugin to mount the volume.
FlexVolume is deprecated since Kubernetes v1.23. The out-of-tree CSI driver is the recommended way
to write volume drivers in Kubernetes. See
Kubernetes Volume Plugin FAQ for Storage Vendors
for more information.
Device Plugins
Device plugins allow a node to discover new Node resources (in addition to the
builtin ones like cpu and memory) via a
Device Plugin.
Network Plugins
Different networking fabrics can be supported via node-level
Network Plugins.
Scheduler Extensions
The scheduler is a special type of controller that watches pods, and assigns
pods to nodes. The default scheduler can be replaced entirely, while
continuing to use other Kubernetes components, or
multiple schedulers
can run at the same time.
This is a significant undertaking, and almost all Kubernetes users find they
do not need to modify the scheduler.
The scheduler also supports a
webhook
that permits a webhook backend (scheduler extension) to filter and prioritize
the nodes chosen for a pod.
Custom resources are extensions of the Kubernetes API. This page discusses when to add a custom
resource to your Kubernetes cluster and when to use a standalone service. It describes the two
methods for adding custom resources and how to choose between them.
Custom resources
A resource is an endpoint in the Kubernetes API that stores a collection of
API objects of a certain kind; for example, the built-in pods resource contains a collection of Pod objects.
A custom resource is an extension of the Kubernetes API that is not necessarily available in a default
Kubernetes installation. It represents a customization of a particular Kubernetes installation. However,
many core Kubernetes functions are now built using custom resources, making Kubernetes more modular.
Custom resources can appear and disappear in a running cluster through dynamic registration,
and cluster admins can update custom resources independently of the cluster itself.
Once a custom resource is installed, users can create and access its objects using
kubectl, just as they do for built-in resources like
Pods.
Custom controllers
On their own, custom resources let you store and retrieve structured data.
When you combine a custom resource with a custom controller, custom resources
provide a true declarative API.
The Kubernetes declarative API
enforces a separation of responsibilities. You declare the desired state of
your resource. The Kubernetes controller keeps the current state of Kubernetes
objects in sync with your declared desired state. This is in contrast to an
imperative API, where you instruct a server what to do.
You can deploy and update a custom controller on a running cluster, independently
of the cluster's lifecycle. Custom controllers can work with any kind of resource,
but they are especially effective when combined with custom resources. The
Operator pattern combines custom
resources and custom controllers. You can use custom controllers to encode domain knowledge
for specific applications into an extension of the Kubernetes API.
Should I add a custom resource to my Kubernetes Cluster?
You want your new types to be readable and writable using kubectl.
kubectl support is not required
You want to view your new types in a Kubernetes UI, such as dashboard, alongside built-in types.
Kubernetes UI support is not required.
You are developing a new API.
You already have a program that serves your API and works well.
You are willing to accept the format restriction that Kubernetes puts on REST resource paths, such as API Groups and Namespaces. (See the API Overview.)
You need to have specific REST paths to be compatible with an already defined REST API.
Your resources are naturally scoped to a cluster or namespaces of a cluster.
Cluster or namespace scoped resources are a poor fit; you need control over the specifics of resource paths.
Your API consists of a relatively small number of relatively small objects (resources).
The objects define configuration of applications or infrastructure.
The objects are updated relatively infrequently.
Humans often need to read and write the objects.
The main operations on the objects are CRUD-y (creating, reading, updating and deleting).
Transactions across objects are not required: the API represents a desired state, not an exact state.
Imperative APIs are not declarative.
Signs that your API might not be declarative include:
The client says "do this", and then gets a synchronous response back when it is done.
The client says "do this", and then gets an operation ID back, and has to check a separate Operation object to determine completion of the request.
You talk about Remote Procedure Calls (RPCs).
Directly storing large amounts of data; for example, > a few kB per object, or > 1000s of objects.
High bandwidth access (10s of requests per second sustained) needed.
Store end-user data (such as images, PII, etc.) or other large-scale data processed by applications.
The natural operations on the objects are not CRUD-y.
The API is not easily modeled as objects.
You chose to represent pending operations with an operation ID or an operation object.
Should I use a configMap or a custom resource?
Use a ConfigMap if any of the following apply:
There is an existing, well-documented config file format, such as a mysql.cnf or pom.xml.
You want to put the entire config file into one key of a configMap.
The main use of the config file is for a program running in a Pod on your cluster to consume the file to configure itself.
Consumers of the file prefer to consume via file in a Pod or environment variable in a pod, rather than the Kubernetes API.
You want to perform rolling updates via Deployment, etc., when the file is updated.
Note: Use a secret for sensitive data, which is similar to a configMap but more secure.
Use a custom resource (CRD or Aggregated API) if most of the following apply:
You want to use Kubernetes client libraries and CLIs to create and update the new resource.
You want top-level support from kubectl; for example, kubectl get my-object object-name.
You want to build new automation that watches for updates on the new object, and then CRUD other objects, or vice versa.
You want to write automation that handles updates to the object.
You want to use Kubernetes API conventions like .spec, .status, and .metadata.
You want the object to be an abstraction over a collection of controlled resources, or a summarization of other resources.
Adding custom resources
Kubernetes provides two ways to add custom resources to your cluster:
CRDs are simple and can be created without any programming.
API Aggregation requires programming, but allows more control over API behaviors like how data is stored and conversion between API versions.
Kubernetes provides these two options to meet the needs of different users, so that neither ease of use nor flexibility is compromised.
Aggregated APIs are subordinate API servers that sit behind the primary API server, which acts as a proxy. This arrangement is called API Aggregation (AA). To users, the Kubernetes API appears extended.
CRDs allow users to create new types of resources without adding another API server. You do not need to understand API Aggregation to use CRDs.
Regardless of how they are installed, the new resources are referred to as Custom Resources to distinguish them from built-in Kubernetes resources (like pods).
CustomResourceDefinitions
The CustomResourceDefinition
API resource allows you to define custom resources.
Defining a CRD object creates a new custom resource with a name and schema that you specify.
The Kubernetes API serves and handles the storage of your custom resource.
The name of a CRD object must be a valid
DNS subdomain name.
This frees you from writing your own API server to handle the custom resource,
but the generic nature of the implementation means you have less flexibility than with
API server aggregation.
Refer to the custom controller example
for an example of how to register a new custom resource, work with instances of your new resource type,
and use a controller to handle events.
API server aggregation
Usually, each resource in the Kubernetes API requires code that handles REST requests and manages persistent storage of objects. The main Kubernetes API server handles built-in resources like pods and services, and can also generically handle custom resources through CRDs.
The aggregation layer allows you to provide specialized
implementations for your custom resources by writing and deploying your own API server.
The main API server delegates requests to your API server for the custom resources that you handle,
making them available to all of its clients.
Choosing a method for adding custom resources
CRDs are easier to use. Aggregated APIs are more flexible. Choose the method that best meets your needs.
Typically, CRDs are a good fit if:
You have a handful of fields
You are using the resource within your company, or as part of a small open-source project (as opposed to a commercial product)
Comparing ease of use
CRDs are easier to create than Aggregated APIs.
CRDs
Aggregated API
Do not require programming. Users can choose any language for a CRD controller.
Requires programming and building binary and image.
No additional service to run; CRDs are handled by API server.
An additional service to create and that could fail.
No ongoing support once the CRD is created. Any bug fixes are picked up as part of normal Kubernetes Master upgrades.
May need to periodically pickup bug fixes from upstream and rebuild and update the Aggregated API server.
No need to handle multiple versions of your API; for example, when you control the client for this resource, you can upgrade it in sync with the API.
You need to handle multiple versions of your API; for example, when developing an extension to share with the world.
Advanced features and flexibility
Aggregated APIs offer more advanced API features and customization of other features; for example, the storage layer.
Feature
Description
CRDs
Aggregated API
Validation
Help users prevent errors and allow you to evolve your API independently of your clients. These features are most useful when there are many clients who can't all update at the same time.
Yes, either via OpenAPI v3.0 validationdefault keyword (GA in 1.17), or via a Mutating Webhook (though this will not be run when reading from etcd for old objects).
Yes
Multi-versioning
Allows serving the same object through two API versions. Can help ease API changes like renaming fields. Less important if you control your client versions.
If you need storage with a different performance mode (for example, a time-series database instead of key-value store) or isolation for security (for example, encryption of sensitive information, etc.)
No
Yes
Custom Business Logic
Perform arbitrary checks or actions when creating, reading, updating or deleting an object
Allows fine-grained access control where user writes the spec section and the controller writes the status section. Allows incrementing object Generation on custom resource data mutation (requires separate spec and status sections in the resource)
Add operations other than CRUD, such as "logs" or "exec".
No
Yes
strategic-merge-patch
The new endpoints support PATCH with Content-Type: application/strategic-merge-patch+json. Useful for updating objects that may be modified both locally, and by the server. For more information, see "Update API Objects in Place Using kubectl patch"
No
Yes
Protocol Buffers
The new resource supports clients that want to use Protocol Buffers
No
Yes
OpenAPI Schema
Is there an OpenAPI (swagger) schema for the types that can be dynamically fetched from the server? Is the user protected from misspelling field names by ensuring only allowed fields are set? Are types enforced (in other words, don't put an int in a string field?)
When you create a custom resource, either via a CRD or an AA, you get many features for your API, compared to implementing it outside the Kubernetes platform:
Feature
What it does
CRUD
The new endpoints support CRUD basic operations via HTTP and kubectl
Watch
The new endpoints support Kubernetes Watch operations via HTTP
Discovery
Clients like kubectl and dashboard automatically offer list, display, and field edit operations on your resources
json-patch
The new endpoints support PATCH with Content-Type: application/json-patch+json
merge-patch
The new endpoints support PATCH with Content-Type: application/merge-patch+json
HTTPS
The new endpoints uses HTTPS
Built-in Authentication
Access to the extension uses the core API server (aggregation layer) for authentication
Built-in Authorization
Access to the extension can reuse the authorization used by the core API server; for example, RBAC.
Finalizers
Block deletion of extension resources until external cleanup happens.
Admission Webhooks
Set default values and validate extension resources during any create/update/delete operation.
UI/CLI Display
Kubectl, dashboard can display extension resources.
Unset versus Empty
Clients can distinguish unset fields from zero-valued fields.
Client Libraries Generation
Kubernetes provides generic client libraries, as well as tools to generate type-specific client libraries.
Labels and annotations
Common metadata across objects that tools know how to edit for core and custom resources.
Preparing to install a custom resource
There are several points to be aware of before adding a custom resource to your cluster.
Third party code and new points of failure
While creating a CRD does not automatically add any new points of failure (for example, by causing third party code to run on your API server), packages (for example, Charts) or other installation bundles often include CRDs as well as a Deployment of third-party code that implements the business logic for a new custom resource.
Installing an Aggregated API server always involves running a new Deployment.
Storage
Custom resources consume storage space in the same way that ConfigMaps do. Creating too many custom resources may overload your API server's storage space.
Aggregated API servers may use the same storage as the main API server, in which case the same warning applies.
Authentication, authorization, and auditing
CRDs always use the same authentication, authorization, and audit logging as the built-in resources of your API server.
If you use RBAC for authorization, most RBAC roles will not grant access to the new resources (except the cluster-admin role or any role created with wildcard rules). You'll need to explicitly grant access to the new resources. CRDs and Aggregated APIs often come bundled with new role definitions for the types they add.
Aggregated API servers may or may not use the same authentication, authorization, and auditing as the primary API server.
Accessing a custom resource
Kubernetes client libraries can be used to access custom resources. Not all client libraries support custom resources. The Go and Python client libraries do.
When you add a custom resource, you can access it using:
kubectl
The kubernetes dynamic client.
A REST client that you write.
A client generated using Kubernetes client generation tools (generating one is an advanced undertaking, but some projects may provide a client along with the CRD or AA).
The aggregation layer allows Kubernetes to be extended with additional APIs, beyond what is offered by the core Kubernetes APIs.
The additional APIs can either be ready-made solutions such as a metrics server, or APIs that you develop yourself.
The aggregation layer is different from Custom Resources, which are a way to make the kube-apiserver recognise new kinds of object.
Aggregation layer
The aggregation layer runs in-process with the kube-apiserver. Until an extension resource is registered, the aggregation layer will do nothing. To register an API, you add an APIService object, which "claims" the URL path in the Kubernetes API. At that point, the aggregation layer will proxy anything sent to that API path (e.g. /apis/myextension.mycompany.io/v1/…) to the registered APIService.
The most common way to implement the APIService is to run an extension API server in Pod(s) that run in your cluster. If you're using the extension API server to manage resources in your cluster, the extension API server (also written as "extension-apiserver") is typically paired with one or more controllers. The apiserver-builder library provides a skeleton for both extension API servers and the associated controller(s).
Response latency
Extension API servers should have low latency networking to and from the kube-apiserver.
Discovery requests are required to round-trip from the kube-apiserver in five seconds or less.
If your extension API server cannot achieve that latency requirement, consider making changes that let you meet it.
3.13.2 - Compute, Storage, and Networking Extensions
3.13.2.1 - Network Plugins
Kubernetes 1.25 supports Container Network Interface
(CNI) plugins for cluster networking. You must use a CNI plugin that is compatible with your cluster and that suits your needs. Different plugins are available (both open- and closed- source) in the wider Kubernetes ecosystem.
You must use a CNI plugin that is compatible with the
v0.4.0 or later
releases of the CNI specification. The Kubernetes project recommends using a plugin that is
compatible with the v1.0.0
CNI specification (plugins can be compatible with multiple spec versions).
Installation
A Container Runtime, in the networking context, is a daemon on a node configured to provide CRI Services for kubelet. In particular, the Container Runtime must be configured to load the CNI plugins required to implement the Kubernetes network model.
Note:
Prior to Kubernetes 1.24, the CNI plugins could also be managed by the kubelet using the cni-bin-dir and network-plugin command-line parameters.
These command-line parameters were removed in Kubernetes 1.24, with management of the CNI no longer in scope for kubelet.
For specific information about how to install and manage a CNI plugin, see the documentation for that plugin or networking provider.
Network Plugin Requirements
For plugin developers and users who regularly build or deploy Kubernetes, the plugin may also need specific configuration to support kube-proxy.
The iptables proxy depends on iptables, and the plugin may need to ensure that container traffic is made available to iptables.
For example, if the plugin connects containers to a Linux bridge, the plugin must set the net/bridge/bridge-nf-call-iptables sysctl to 1 to ensure that the iptables proxy functions correctly.
If the plugin does not use a Linux bridge, but uses something like Open vSwitch or some other mechanism instead, it should ensure container traffic is appropriately routed for the proxy.
By default, if no kubelet network plugin is specified, the noop plugin is used, which sets net/bridge/bridge-nf-call-iptables=1 to ensure simple configurations (like Docker with a bridge) work correctly with the iptables proxy.
Loopback CNI
In addition to the CNI plugin installed on the nodes for implementing the Kubernetes network model, Kubernetes also requires the container runtimes to provide a loopback interface lo, which is used for each sandbox (pod sandboxes, vm sandboxes, ...).
Implementing the loopback interface can be accomplished by re-using the CNI loopback plugin. or by developing your own code to achieve this (see this example from CRI-O).
Support hostPort
The CNI networking plugin supports hostPort. You can use the official portmap
plugin offered by the CNI plugin team or use your own plugin with portMapping functionality.
If you want to enable hostPort support, you must specify portMappings capability in your cni-conf-dir.
For example:
The CNI networking plugin also supports pod ingress and egress traffic shaping. You can use the official bandwidth
plugin offered by the CNI plugin team or use your own plugin with bandwidth control functionality.
If you want to enable traffic shaping support, you must add the bandwidth plugin to your CNI configuration file
(default /etc/cni/net.d) and ensure that the binary is included in your CNI bin dir (default /opt/cni/bin).
Device plugins let you configure your cluster with support for devices or resources that require vendor-specific setup, such as GPUs, NICs, FPGAs, or non-volatile main memory.
Instead of customizing the code for Kubernetes itself, vendors can implement a
device plugin that you deploy either manually or as a DaemonSet.
The targeted devices include GPUs, high-performance NICs, FPGAs, InfiniBand adapters,
and other similar computing resources that may require vendor specific initialization
and setup.
Device plugin registration
The kubelet exports a Registration gRPC service:
service Registration {
rpc Register(RegisterRequest) returns (Empty) {}
}
A device plugin can register itself with the kubelet through this gRPC service.
During the registration, the device plugin needs to send:
The name of its Unix socket.
The Device Plugin API version against which it was built.
The ResourceName it wants to advertise. Here ResourceName needs to follow the
extended resource naming scheme
as vendor-domain/resourcetype.
(For example, an NVIDIA GPU is advertised as nvidia.com/gpu.)
Following a successful registration, the device plugin sends the kubelet the
list of devices it manages, and the kubelet is then in charge of advertising those
resources to the API server as part of the kubelet node status update.
For example, after a device plugin registers hardware-vendor.example/foo with the kubelet
and reports two healthy devices on a node, the node status is updated
to advertise that the node has 2 "Foo" devices installed and available.
Then, users can request devices as part of a Pod specification
(see container).
Requesting extended resources is similar to how you manage requests and limits for
other resources, with the following differences:
Extended resources are only supported as integer resources and cannot be overcommitted.
Devices cannot be shared between containers.
Example
Suppose a Kubernetes cluster is running a device plugin that advertises resource hardware-vendor.example/foo
on certain nodes. Here is an example of a pod requesting this resource to run a demo workload:
---apiVersion:v1kind:Podmetadata:name:demo-podspec:containers:- name:demo-container-1image:registry.k8s.io/pause:2.0resources:limits:hardware-vendor.example/foo:2## This Pod needs 2 of the hardware-vendor.example/foo devices# and can only schedule onto a Node that's able to satisfy# that need.## If the Node has more than 2 of those devices available, the# remainder would be available for other Pods to use.
Device plugin implementation
The general workflow of a device plugin includes the following steps:
Initialization. During this phase, the device plugin performs vendor specific
initialization and setup to make sure the devices are in a ready state.
The plugin starts a gRPC service, with a Unix socket under host path
/var/lib/kubelet/device-plugins/, that implements the following interfaces:
service DevicePlugin {
// GetDevicePluginOptions returns options to be communicated with Device Manager.
rpc GetDevicePluginOptions(Empty) returns (DevicePluginOptions) {}
// ListAndWatch returns a stream of List of Devices
// Whenever a Device state change or a Device disappears, ListAndWatch
// returns the new list
rpc ListAndWatch(Empty) returns (stream ListAndWatchResponse) {}
// Allocate is called during container creation so that the Device
// Plugin can run device specific operations and instruct Kubelet
// of the steps to make the Device available in the container
rpc Allocate(AllocateRequest) returns (AllocateResponse) {}
// GetPreferredAllocation returns a preferred set of devices to allocate
// from a list of available ones. The resulting preferred allocation is not
// guaranteed to be the allocation ultimately performed by the
// devicemanager. It is only designed to help the devicemanager make a more
// informed allocation decision when possible.
rpc GetPreferredAllocation(PreferredAllocationRequest) returns (PreferredAllocationResponse) {}
// PreStartContainer is called, if indicated by Device Plugin during registeration phase,
// before each container start. Device plugin can run device specific operations
// such as resetting the device before making devices available to the container.
rpc PreStartContainer(PreStartContainerRequest) returns (PreStartContainerResponse) {}
}
Note: Plugins are not required to provide useful implementations for
GetPreferredAllocation() or PreStartContainer(). Flags indicating which
(if any) of these calls are available should be set in the DevicePluginOptions
message sent back by a call to GetDevicePluginOptions(). The kubelet will
always call GetDevicePluginOptions() to see which optional functions are
available, before calling any of them directly.
The plugin registers itself with the kubelet through the Unix socket at host
path /var/lib/kubelet/device-plugins/kubelet.sock.
After successfully registering itself, the device plugin runs in serving mode, during which it keeps
monitoring device health and reports back to the kubelet upon any device state changes.
It is also responsible for serving Allocate gRPC requests. During Allocate, the device plugin may
do device-specific preparation; for example, GPU cleanup or QRNG initialization.
If the operations succeed, the device plugin returns an AllocateResponse that contains container
runtime configurations for accessing the allocated devices. The kubelet passes this information
to the container runtime.
Handling kubelet restarts
A device plugin is expected to detect kubelet restarts and re-register itself with the new
kubelet instance. In the current implementation, a new kubelet instance deletes all the existing Unix sockets
under /var/lib/kubelet/device-plugins when it starts. A device plugin can monitor the deletion
of its Unix socket and re-register itself upon such an event.
Device plugin deployment
You can deploy a device plugin as a DaemonSet, as a package for your node's operating system,
or manually.
The canonical directory /var/lib/kubelet/device-plugins requires privileged access,
so a device plugin must run in a privileged security context.
If you're deploying a device plugin as a DaemonSet, /var/lib/kubelet/device-plugins
must be mounted as a Volume
in the plugin's
PodSpec.
If you choose the DaemonSet approach you can rely on Kubernetes to: place the device plugin's
Pod onto Nodes, to restart the daemon Pod after failure, and to help automate upgrades.
API compatibility
Kubernetes device plugin support is in beta. The API may change before stabilization,
in incompatible ways. As a project, Kubernetes recommends that device plugin developers:
Watch for changes in future releases.
Support multiple versions of the device plugin API for backward/forward compatibility.
If you enable the DevicePlugins feature and run device plugins on nodes that need to be upgraded to
a Kubernetes release with a newer device plugin API version, upgrade your device plugins
to support both versions before upgrading these nodes. Taking that approach will
ensure the continuous functioning of the device allocations during the upgrade.
Monitoring device plugin resources
FEATURE STATE:Kubernetes v1.15 [beta]
In order to monitor resources provided by device plugins, monitoring agents need to be able to
discover the set of devices that are in-use on the node and obtain metadata to describe which
container the metric should be associated with. Prometheus metrics
exposed by device monitoring agents should follow the
Kubernetes Instrumentation Guidelines,
identifying containers using pod, namespace, and container prometheus labels.
The kubelet provides a gRPC service to enable discovery of in-use devices, and to provide metadata
for these devices:
// PodResourcesLister is a service provided by the kubelet that provides information about the
// node resources consumed by pods and containers on the node
service PodResourcesLister {
rpc List(ListPodResourcesRequest) returns (ListPodResourcesResponse) {}
rpc GetAllocatableResources(AllocatableResourcesRequest) returns (AllocatableResourcesResponse) {}
}
List gRPC endpoint
The List endpoint provides information on resources of running pods, with details such as the
id of exclusively allocated CPUs, device id as it was reported by device plugins and id of
the NUMA node where these devices are allocated. Also, for NUMA-based machines, it contains the information about memory and hugepages reserved for a container.
// ListPodResourcesResponse is the response returned by List function
message ListPodResourcesResponse {
repeated PodResources pod_resources = 1;
}
// PodResources contains information about the node resources assigned to a pod
message PodResources {
string name = 1;
string namespace = 2;
repeated ContainerResources containers = 3;
}
// ContainerResources contains information about the resources assigned to a container
message ContainerResources {
string name = 1;
repeated ContainerDevices devices = 2;
repeated int64 cpu_ids = 3;
repeated ContainerMemory memory = 4;
}
// ContainerMemory contains information about memory and hugepages assigned to a container
message ContainerMemory {
string memory_type = 1;
uint64 size = 2;
TopologyInfo topology = 3;
}
// Topology describes hardware topology of the resource
message TopologyInfo {
repeated NUMANode nodes = 1;
}
// NUMA representation of NUMA node
message NUMANode {
int64 ID = 1;
}
// ContainerDevices contains information about the devices assigned to a container
message ContainerDevices {
string resource_name = 1;
repeated string device_ids = 2;
TopologyInfo topology = 3;
}
Note:
cpu_ids in the ContainerResources in the List endpoint correspond to exclusive CPUs allocated
to a partilar container. If the goal is to evaluate CPUs that belong to the shared pool, the List
endpoint needs to be used in conjunction with the GetAllocatableResources endpoint as explained
below:
Call GetAllocatableResources to get a list of all the allocatable CPUs
Call GetCpuIds on all ContainerResources in the system
Subtract out all of the CPUs from the GetCpuIds calls from the GetAllocatableResources call
GetAllocatableResources gRPC endpoint
FEATURE STATE:Kubernetes v1.23 [beta]
GetAllocatableResources provides information on resources initially available on the worker node.
It provides more information than kubelet exports to APIServer.
Note:GetAllocatableResources should only be used to evaluate allocatable
resources on a node. If the goal is to evaluate free/unallocated resources it should be used in
conjunction with the List() endpoint. The result obtained by GetAllocatableResources would remain
the same unless the underlying resources exposed to kubelet change. This happens rarely but when
it does (for example: hotplug/hotunplug, device health changes), client is expected to call
GetAlloctableResources endpoint.
However, calling GetAllocatableResources endpoint is not sufficient in case of cpu and/or memory
update and Kubelet needs to be restarted to reflect the correct resource capacity and allocatable.
// AllocatableResourcesResponses contains informations about all the devices known by the kubelet
message AllocatableResourcesResponse {
repeated ContainerDevices devices = 1;
repeated int64 cpu_ids = 2;
repeated ContainerMemory memory = 3;
}
Starting from Kubernetes v1.23, the GetAllocatableResources is enabled by default.
You can disable it by turning off the
KubeletPodResourcesGetAllocatablefeature gate.
Preceding Kubernetes v1.23, to enable this feature kubelet must be started with the following flag:
ContainerDevices do expose the topology information declaring to which NUMA cells the device is affine.
The NUMA cells are identified using a opaque integer ID, which value is consistent to what device
plugins report when they register themselves to the kubelet.
The gRPC service is served over a unix socket at /var/lib/kubelet/pod-resources/kubelet.sock.
Monitoring agents for device plugin resources can be deployed as a daemon, or as a DaemonSet.
The canonical directory /var/lib/kubelet/pod-resources requires privileged access, so monitoring
agents must run in a privileged security context. If a device monitoring agent is running as a
DaemonSet, /var/lib/kubelet/pod-resources must be mounted as a
Volume in the device monitoring agent's
PodSpec.
Support for the PodResourcesLister service requires KubeletPodResourcesfeature gate to be enabled.
It is enabled by default starting with Kubernetes 1.15 and is v1 since Kubernetes 1.20.
Device plugin integration with the Topology Manager
FEATURE STATE:Kubernetes v1.18 [beta]
The Topology Manager is a Kubelet component that allows resources to be co-ordinated in a Topology aligned manner. In order to do this, the Device Plugin API was extended to include a TopologyInfo struct.
Device Plugins that wish to leverage the Topology Manager can send back a populated TopologyInfo struct as part of the device registration, along with the device IDs and the health of the device. The device manager will then use this information to consult with the Topology Manager and make resource assignment decisions.
TopologyInfo supports a nodes field that is either nil (the default) or a list of NUMA nodes. This lets the Device Plugin publish that can span NUMA nodes.
An example TopologyInfo struct populated for a device by a Device Plugin:
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Here are some examples of device plugin implementations:
Operators are software extensions to Kubernetes that make use of
custom resources
to manage applications and their components. Operators follow
Kubernetes principles, notably the control loop.
Motivation
The operator pattern aims to capture the key aim of a human operator who
is managing a service or set of services. Human operators who look after
specific applications and services have deep knowledge of how the system
ought to behave, how to deploy it, and how to react if there are problems.
People who run workloads on Kubernetes often like to use automation to take
care of repeatable tasks. The operator pattern captures how you can write
code to automate a task beyond what Kubernetes itself provides.
Operators in Kubernetes
Kubernetes is designed for automation. Out of the box, you get lots of
built-in automation from the core of Kubernetes. You can use Kubernetes
to automate deploying and running workloads, and you can automate how
Kubernetes does that.
Kubernetes' operator pattern concept lets you extend the cluster's behaviour without modifying the code of Kubernetes itself by linking controllers to one or more custom resources.
Operators are clients of the Kubernetes API that act as controllers for
a Custom Resource.
An example operator
Some of the things that you can use an operator to automate include:
deploying an application on demand
taking and restoring backups of that application's state
handling upgrades of the application code alongside related changes such
as database schemas or extra configuration settings
publishing a Service to applications that don't support Kubernetes APIs to
discover them
simulating failure in all or part of your cluster to test its resilience
choosing a leader for a distributed application without an internal
member election process
What might an operator look like in more detail? Here's an example:
A custom resource named SampleDB, that you can configure into the cluster.
A Deployment that makes sure a Pod is running that contains the
controller part of the operator.
A container image of the operator code.
Controller code that queries the control plane to find out what SampleDB
resources are configured.
The core of the operator is code to tell the API server how to make
reality match the configured resources.
If you add a new SampleDB, the operator sets up PersistentVolumeClaims
to provide durable database storage, a StatefulSet to run SampleDB and
a Job to handle initial configuration.
If you delete it, the operator takes a snapshot, then makes sure that
the StatefulSet and Volumes are also removed.
The operator also manages regular database backups. For each SampleDB
resource, the operator determines when to create a Pod that can connect
to the database and take backups. These Pods would rely on a ConfigMap
and / or a Secret that has database connection details and credentials.
Because the operator aims to provide robust automation for the resource
it manages, there would be additional supporting code. For this example,
code checks to see if the database is running an old version and, if so,
creates Job objects that upgrade it for you.
Deploying operators
The most common way to deploy an operator is to add the
Custom Resource Definition and its associated Controller to your cluster.
The Controller will normally run outside of the
control plane,
much as you would run any containerized application.
For example, you can run the controller in your cluster as a Deployment.
Using an operator
Once you have an operator deployed, you'd use it by adding, modifying or
deleting the kind of resource that the operator uses. Following the above
example, you would set up a Deployment for the operator itself, and then:
kubectl get SampleDB # find configured databaseskubectl edit SampleDB/example-database # manually change some settings
…and that's it! The operator will take care of applying the changes
as well as keeping the existing service in good shape.
Writing your own operator
If there isn't an operator in the ecosystem that implements the behavior you
want, you can code your own.
You also implement an operator (that is, a Controller) using any language / runtime
that can act as a client for the Kubernetes API.
Following are a few libraries and tools you can use to write your own cloud native
operator.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Read CoreOS' original article that introduced the operator pattern (this is an archived version of the original article).
Read an article from Google Cloud about best practices for building operators
4 - Tasks
This section of the Kubernetes documentation contains pages that
show how to do individual tasks. A task page shows how to do a
single thing, typically by giving a short sequence of steps.
The Kubernetes command-line tool, kubectl, allows
you to run commands against Kubernetes clusters.
You can use kubectl to deploy applications, inspect and manage cluster resources,
and view logs. For more information including a complete list of kubectl operations, see the
kubectl reference documentation.
kubectl is installable on a variety of Linux platforms, macOS and Windows.
Find your preferred operating system below.
Like kind, minikube is a tool that lets you run Kubernetes
locally. minikube runs a single-node Kubernetes cluster on your personal
computer (including Windows, macOS and Linux PCs) so that you can try out
Kubernetes, or for daily development work.
You can follow the official
Get Started! guide if your focus is
on getting the tool installed.
You can use the kubeadm tool to create and manage Kubernetes clusters.
It performs the actions necessary to get a minimum viable, secure cluster up and running in a user friendly way.
You must use a kubectl version that is within one minor version difference of your cluster. For example, a v1.25 client can communicate with v1.24, v1.25, and v1.26 control planes.
Using the latest compatible version of kubectl helps avoid unforeseen issues.
Install kubectl on Linux
The following methods exist for installing kubectl on Linux:
If you are on Ubuntu or another Linux distribution that supports the snap package manager, kubectl is available as a snap application.
snap install kubectl --classic
kubectl version --client
If you are on Linux and using Homebrew package manager, kubectl is available for installation.
brew install kubectl
kubectl version --client
Verify kubectl configuration
In order for kubectl to find and access a Kubernetes cluster, it needs a
kubeconfig file,
which is created automatically when you create a cluster using
kube-up.sh
or successfully deploy a Minikube cluster.
By default, kubectl configuration is located at ~/.kube/config.
Check that kubectl is properly configured by getting the cluster state:
kubectl cluster-info
If you see a URL response, kubectl is correctly configured to access your cluster.
If you see a message similar to the following, kubectl is not configured correctly or is not able to connect to a Kubernetes cluster.
The connection to the server <server-name:port> was refused - did you specify the right host or port?
For example, if you are intending to run a Kubernetes cluster on your laptop (locally), you will need a tool like Minikube to be installed first and then re-run the commands stated above.
If kubectl cluster-info returns the url response but you can't access your cluster, to check whether it is configured properly, use:
kubectl cluster-info dump
Optional kubectl configurations and plugins
Enable shell autocompletion
kubectl provides autocompletion support for Bash, Zsh, Fish, and PowerShell, which can save you a lot of typing.
Below are the procedures to set up autocompletion for Bash, Fish, and Zsh.
The kubectl completion script for Bash can be generated with the command kubectl completion bash. Sourcing the completion script in your shell enables kubectl autocompletion.
However, the completion script depends on bash-completion, which means that you have to install this software first (you can test if you have bash-completion already installed by running type _init_completion).
Install bash-completion
bash-completion is provided by many package managers (see here). You can install it with apt-get install bash-completion or yum install bash-completion, etc.
The above commands create /usr/share/bash-completion/bash_completion, which is the main script of bash-completion. Depending on your package manager, you have to manually source this file in your ~/.bashrc file.
To find out, reload your shell and run type _init_completion. If the command succeeds, you're already set, otherwise add the following to your ~/.bashrc file:
source /usr/share/bash-completion/bash_completion
Reload your shell and verify that bash-completion is correctly installed by typing type _init_completion.
Enable kubectl autocompletion
Bash
You now need to ensure that the kubectl completion script gets sourced in all your shell sessions. There are two ways in which you can do this:
kubectl completion bash | sudo tee /etc/bash_completion.d/kubectl > /dev/null
If you have an alias for kubectl, you can extend shell completion to work with that alias:
echo'alias k=kubectl' >>~/.bashrc
echo'complete -o default -F __start_kubectl k' >>~/.bashrc
Note: bash-completion sources all completion scripts in /etc/bash_completion.d.
Both approaches are equivalent. After reloading your shell, kubectl autocompletion should be working.
To enable bash autocompletion in current session of shell, source the ~/.bashrc file:
source ~/.bashrc
The kubectl completion script for Fish can be generated with the command kubectl completion fish. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following line to your ~/.config/fish/config.fish file:
kubectl completion fish | source
After reloading your shell, kubectl autocompletion should be working.
The kubectl completion script for Zsh can be generated with the command kubectl completion zsh. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following to your ~/.zshrc file:
source <(kubectl completion zsh)
If you have an alias for kubectl, kubectl autocompletion will automatically work with it.
After reloading your shell, kubectl autocompletion should be working.
If you get an error like 2: command not found: compdef, then add the following to the beginning of your ~/.zshrc file:
autoload -Uz compinit
compinit
Install kubectl convert plugin
A plugin for Kubernetes command-line tool kubectl, which allows you to convert manifests between different API
versions. This can be particularly helpful to migrate manifests to a non-deprecated api version with newer Kubernetes release.
For more info, visit migrate to non deprecated apis
You must use a kubectl version that is within one minor version difference of your cluster. For example, a v1.25 client can communicate with v1.24, v1.25, and v1.26 control planes.
Using the latest compatible version of kubectl helps avoid unforeseen issues.
Install kubectl on macOS
The following methods exist for installing kubectl on macOS:
Note: Make sure /usr/local/bin is in your PATH environment variable.
Test to ensure the version you installed is up-to-date:
kubectl version --client
Or use this for detailed view of version:
kubectl version --client --output=yaml
Install with Homebrew on macOS
If you are on macOS and using Homebrew package manager, you can install kubectl with Homebrew.
Run the installation command:
brew install kubectl
or
brew install kubernetes-cli
Test to ensure the version you installed is up-to-date:
kubectl version --client
Install with Macports on macOS
If you are on macOS and using Macports package manager, you can install kubectl with Macports.
Run the installation command:
sudo port selfupdate
sudo port install kubectl
Test to ensure the version you installed is up-to-date:
kubectl version --client
Verify kubectl configuration
In order for kubectl to find and access a Kubernetes cluster, it needs a
kubeconfig file,
which is created automatically when you create a cluster using
kube-up.sh
or successfully deploy a Minikube cluster.
By default, kubectl configuration is located at ~/.kube/config.
Check that kubectl is properly configured by getting the cluster state:
kubectl cluster-info
If you see a URL response, kubectl is correctly configured to access your cluster.
If you see a message similar to the following, kubectl is not configured correctly or is not able to connect to a Kubernetes cluster.
The connection to the server <server-name:port> was refused - did you specify the right host or port?
For example, if you are intending to run a Kubernetes cluster on your laptop (locally), you will need a tool like Minikube to be installed first and then re-run the commands stated above.
If kubectl cluster-info returns the url response but you can't access your cluster, to check whether it is configured properly, use:
kubectl cluster-info dump
Optional kubectl configurations and plugins
Enable shell autocompletion
kubectl provides autocompletion support for Bash, Zsh, Fish, and PowerShell which can save you a lot of typing.
Below are the procedures to set up autocompletion for Bash, Fish, and Zsh.
The kubectl completion script for Bash can be generated with kubectl completion bash. Sourcing this script in your shell enables kubectl completion.
However, the kubectl completion script depends on bash-completion which you thus have to previously install.
Warning: There are two versions of bash-completion, v1 and v2. V1 is for Bash 3.2 (which is the default on macOS), and v2 is for Bash 4.1+. The kubectl completion script doesn't work correctly with bash-completion v1 and Bash 3.2. It requires bash-completion v2 and Bash 4.1+. Thus, to be able to correctly use kubectl completion on macOS, you have to install and use Bash 4.1+ (instructions). The following instructions assume that you use Bash 4.1+ (that is, any Bash version of 4.1 or newer).
Upgrade Bash
The instructions here assume you use Bash 4.1+. You can check your Bash's version by running:
echo$BASH_VERSION
If it is too old, you can install/upgrade it using Homebrew:
brew install bash
Reload your shell and verify that the desired version is being used:
echo$BASH_VERSION$SHELL
Homebrew usually installs it at /usr/local/bin/bash.
Install bash-completion
Note: As mentioned, these instructions assume you use Bash 4.1+, which means you will install bash-completion v2 (in contrast to Bash 3.2 and bash-completion v1, in which case kubectl completion won't work).
You can test if you have bash-completion v2 already installed with type _init_completion. If not, you can install it with Homebrew:
brew install bash-completion@2
As stated in the output of this command, add the following to your ~/.bash_profile file:
If you have an alias for kubectl, you can extend shell completion to work with that alias:
echo'alias k=kubectl' >>~/.bash_profile
echo'complete -o default -F __start_kubectl k' >>~/.bash_profile
If you installed kubectl with Homebrew (as explained here), then the kubectl completion script should already be in /usr/local/etc/bash_completion.d/kubectl. In that case, you don't need to do anything.
Note: The Homebrew installation of bash-completion v2 sources all the files in the BASH_COMPLETION_COMPAT_DIR directory, that's why the latter two methods work.
In any case, after reloading your shell, kubectl completion should be working.
The kubectl completion script for Fish can be generated with the command kubectl completion fish. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following line to your ~/.config/fish/config.fish file:
kubectl completion fish | source
After reloading your shell, kubectl autocompletion should be working.
The kubectl completion script for Zsh can be generated with the command kubectl completion zsh. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following to your ~/.zshrc file:
source <(kubectl completion zsh)
If you have an alias for kubectl, kubectl autocompletion will automatically work with it.
After reloading your shell, kubectl autocompletion should be working.
If you get an error like 2: command not found: compdef, then add the following to the beginning of your ~/.zshrc file:
autoload -Uz compinit
compinit
Install kubectl convert plugin
A plugin for Kubernetes command-line tool kubectl, which allows you to convert manifests between different API
versions. This can be particularly helpful to migrate manifests to a non-deprecated api version with newer Kubernetes release.
For more info, visit migrate to non deprecated apis
You must use a kubectl version that is within one minor version difference of your cluster. For example, a v1.25 client can communicate with v1.24, v1.25, and v1.26 control planes.
Using the latest compatible version of kubectl helps avoid unforeseen issues.
Install kubectl on Windows
The following methods exist for installing kubectl on Windows:
Append or prepend the kubectl binary folder to your PATH environment variable.
Test to ensure the version of kubectl is the same as downloaded:
kubectl version --client
Or use this for detailed view of version:
kubectl version --client --output=yaml
Note:Docker Desktop for Windows adds its own version of kubectl to PATH.
If you have installed Docker Desktop before, you may need to place your PATH entry before the one added by the Docker Desktop installer or remove the Docker Desktop's kubectl.
Install on Windows using Chocolatey or Scoop
To install kubectl on Windows you can use either Chocolatey package manager or Scoop command-line installer.
Test to ensure the version you installed is up-to-date:
kubectl version --client
Navigate to your home directory:
# If you're using cmd.exe, run: cd %USERPROFILE%cd ~
Create the .kube directory:
mkdir .kube
Change to the .kube directory you just created:
cd .kube
Configure kubectl to use a remote Kubernetes cluster:
New-Item config -type file
Note: Edit the config file with a text editor of your choice, such as Notepad.
Verify kubectl configuration
In order for kubectl to find and access a Kubernetes cluster, it needs a
kubeconfig file,
which is created automatically when you create a cluster using
kube-up.sh
or successfully deploy a Minikube cluster.
By default, kubectl configuration is located at ~/.kube/config.
Check that kubectl is properly configured by getting the cluster state:
kubectl cluster-info
If you see a URL response, kubectl is correctly configured to access your cluster.
If you see a message similar to the following, kubectl is not configured correctly or is not able to connect to a Kubernetes cluster.
The connection to the server <server-name:port> was refused - did you specify the right host or port?
For example, if you are intending to run a Kubernetes cluster on your laptop (locally), you will need a tool like Minikube to be installed first and then re-run the commands stated above.
If kubectl cluster-info returns the url response but you can't access your cluster, to check whether it is configured properly, use:
kubectl cluster-info dump
Optional kubectl configurations and plugins
Enable shell autocompletion
kubectl provides autocompletion support for Bash, Zsh, Fish, and PowerShell, which can save you a lot of typing.
Below are the procedures to set up autocompletion for PowerShell.
The kubectl completion script for PowerShell can be generated with the command kubectl completion powershell.
To do so in all your shell sessions, add the following line to your $PROFILE file:
This command will regenerate the auto-completion script on every PowerShell start up. You can also add the generated script directly to your $PROFILE file.
To add the generated script to your $PROFILE file, run the following line in your powershell prompt:
kubectl completion powershell >> $PROFILE
After reloading your shell, kubectl autocompletion should be working.
Install kubectl convert plugin
A plugin for Kubernetes command-line tool kubectl, which allows you to convert manifests between different API
versions. This can be particularly helpful to migrate manifests to a non-deprecated api version with newer Kubernetes release.
For more info, visit migrate to non deprecated apis
Snippets to be included in the main kubectl-installs-*.md pages.
4.1.4.1 - bash auto-completion on Linux
Some optional configuration for bash auto-completion on Linux.
Introduction
The kubectl completion script for Bash can be generated with the command kubectl completion bash. Sourcing the completion script in your shell enables kubectl autocompletion.
However, the completion script depends on bash-completion, which means that you have to install this software first (you can test if you have bash-completion already installed by running type _init_completion).
Install bash-completion
bash-completion is provided by many package managers (see here). You can install it with apt-get install bash-completion or yum install bash-completion, etc.
The above commands create /usr/share/bash-completion/bash_completion, which is the main script of bash-completion. Depending on your package manager, you have to manually source this file in your ~/.bashrc file.
To find out, reload your shell and run type _init_completion. If the command succeeds, you're already set, otherwise add the following to your ~/.bashrc file:
source /usr/share/bash-completion/bash_completion
Reload your shell and verify that bash-completion is correctly installed by typing type _init_completion.
Enable kubectl autocompletion
Bash
You now need to ensure that the kubectl completion script gets sourced in all your shell sessions. There are two ways in which you can do this:
kubectl completion bash | sudo tee /etc/bash_completion.d/kubectl > /dev/null
If you have an alias for kubectl, you can extend shell completion to work with that alias:
echo'alias k=kubectl' >>~/.bashrc
echo'complete -o default -F __start_kubectl k' >>~/.bashrc
Note: bash-completion sources all completion scripts in /etc/bash_completion.d.
Both approaches are equivalent. After reloading your shell, kubectl autocompletion should be working.
To enable bash autocompletion in current session of shell, source the ~/.bashrc file:
source ~/.bashrc
4.1.4.2 - bash auto-completion on macOS
Some optional configuration for bash auto-completion on macOS.
Introduction
The kubectl completion script for Bash can be generated with kubectl completion bash. Sourcing this script in your shell enables kubectl completion.
However, the kubectl completion script depends on bash-completion which you thus have to previously install.
Warning: There are two versions of bash-completion, v1 and v2. V1 is for Bash 3.2 (which is the default on macOS), and v2 is for Bash 4.1+. The kubectl completion script doesn't work correctly with bash-completion v1 and Bash 3.2. It requires bash-completion v2 and Bash 4.1+. Thus, to be able to correctly use kubectl completion on macOS, you have to install and use Bash 4.1+ (instructions). The following instructions assume that you use Bash 4.1+ (that is, any Bash version of 4.1 or newer).
Upgrade Bash
The instructions here assume you use Bash 4.1+. You can check your Bash's version by running:
echo$BASH_VERSION
If it is too old, you can install/upgrade it using Homebrew:
brew install bash
Reload your shell and verify that the desired version is being used:
echo$BASH_VERSION$SHELL
Homebrew usually installs it at /usr/local/bin/bash.
Install bash-completion
Note: As mentioned, these instructions assume you use Bash 4.1+, which means you will install bash-completion v2 (in contrast to Bash 3.2 and bash-completion v1, in which case kubectl completion won't work).
You can test if you have bash-completion v2 already installed with type _init_completion. If not, you can install it with Homebrew:
brew install bash-completion@2
As stated in the output of this command, add the following to your ~/.bash_profile file:
If you have an alias for kubectl, you can extend shell completion to work with that alias:
echo'alias k=kubectl' >>~/.bash_profile
echo'complete -o default -F __start_kubectl k' >>~/.bash_profile
If you installed kubectl with Homebrew (as explained here), then the kubectl completion script should already be in /usr/local/etc/bash_completion.d/kubectl. In that case, you don't need to do anything.
Note: The Homebrew installation of bash-completion v2 sources all the files in the BASH_COMPLETION_COMPAT_DIR directory, that's why the latter two methods work.
In any case, after reloading your shell, kubectl completion should be working.
4.1.4.3 - fish auto-completion
Optional configuration to enable fish shell auto-completion.
The kubectl completion script for Fish can be generated with the command kubectl completion fish. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following line to your ~/.config/fish/config.fish file:
kubectl completion fish | source
After reloading your shell, kubectl autocompletion should be working.
4.1.4.4 - kubectl-convert overview
A kubectl plugin that allows you to convert manifests from one version of a Kubernetes API to a different version.
A plugin for Kubernetes command-line tool kubectl, which allows you to convert manifests between different API
versions. This can be particularly helpful to migrate manifests to a non-deprecated api version with newer Kubernetes release.
For more info, visit migrate to non deprecated apis
4.1.4.5 - PowerShell auto-completion
Some optional configuration for powershell auto-completion.
The kubectl completion script for PowerShell can be generated with the command kubectl completion powershell.
To do so in all your shell sessions, add the following line to your $PROFILE file:
This command will regenerate the auto-completion script on every PowerShell start up. You can also add the generated script directly to your $PROFILE file.
To add the generated script to your $PROFILE file, run the following line in your powershell prompt:
kubectl completion powershell >> $PROFILE
After reloading your shell, kubectl autocompletion should be working.
4.1.4.6 - verify kubectl install
How to verify kubectl.
In order for kubectl to find and access a Kubernetes cluster, it needs a
kubeconfig file,
which is created automatically when you create a cluster using
kube-up.sh
or successfully deploy a Minikube cluster.
By default, kubectl configuration is located at ~/.kube/config.
Check that kubectl is properly configured by getting the cluster state:
kubectl cluster-info
If you see a URL response, kubectl is correctly configured to access your cluster.
If you see a message similar to the following, kubectl is not configured correctly or is not able to connect to a Kubernetes cluster.
The connection to the server <server-name:port> was refused - did you specify the right host or port?
For example, if you are intending to run a Kubernetes cluster on your laptop (locally), you will need a tool like Minikube to be installed first and then re-run the commands stated above.
If kubectl cluster-info returns the url response but you can't access your cluster, to check whether it is configured properly, use:
Some optional configuration for zsh auto-completion.
The kubectl completion script for Zsh can be generated with the command kubectl completion zsh. Sourcing the completion script in your shell enables kubectl autocompletion.
To do so in all your shell sessions, add the following to your ~/.zshrc file:
source <(kubectl completion zsh)
If you have an alias for kubectl, kubectl autocompletion will automatically work with it.
After reloading your shell, kubectl autocompletion should be working.
If you get an error like 2: command not found: compdef, then add the following to the beginning of your ~/.zshrc file:
autoload -Uz compinit
compinit
4.2 - Administer a Cluster
Learn common tasks for administering a cluster.
4.2.1 - Administration with kubeadm
4.2.1.1 - Certificate Management with kubeadm
FEATURE STATE:Kubernetes v1.15 [stable]
Client certificates generated by kubeadm expire after 1 year.
This page explains how to manage certificate renewals with kubeadm. It also covers other tasks related
to kubeadm certificate management.
By default, kubeadm generates all the certificates needed for a cluster to run.
You can override this behavior by providing your own certificates.
To do so, you must place them in whatever directory is specified by the
--cert-dir flag or the certificatesDir field of kubeadm's ClusterConfiguration.
By default this is /etc/kubernetes/pki.
If a given certificate and private key pair exists before running kubeadm init,
kubeadm does not overwrite them. This means you can, for example, copy an existing
CA into /etc/kubernetes/pki/ca.crt and /etc/kubernetes/pki/ca.key,
and kubeadm will use this CA for signing the rest of the certificates.
External CA mode
It is also possible to provide only the ca.crt file and not the
ca.key file (this is only available for the root CA file, not other cert pairs).
If all other certificates and kubeconfig files are in place, kubeadm recognizes
this condition and activates the "External CA" mode. kubeadm will proceed without the
CA key on disk.
Instead, run the controller-manager standalone with --controllers=csrsigner and
point to the CA certificate and key.
You can use the check-expiration subcommand to check when certificates expire:
kubeadm certs check-expiration
The output is similar to this:
CERTIFICATE EXPIRES RESIDUAL TIME CERTIFICATE AUTHORITY EXTERNALLY MANAGED
admin.conf Dec 30, 2020 23:36 UTC 364d no
apiserver Dec 30, 2020 23:36 UTC 364d ca no
apiserver-etcd-client Dec 30, 2020 23:36 UTC 364d etcd-ca no
apiserver-kubelet-client Dec 30, 2020 23:36 UTC 364d ca no
controller-manager.conf Dec 30, 2020 23:36 UTC 364d no
etcd-healthcheck-client Dec 30, 2020 23:36 UTC 364d etcd-ca no
etcd-peer Dec 30, 2020 23:36 UTC 364d etcd-ca no
etcd-server Dec 30, 2020 23:36 UTC 364d etcd-ca no
front-proxy-client Dec 30, 2020 23:36 UTC 364d front-proxy-ca no
scheduler.conf Dec 30, 2020 23:36 UTC 364d no
CERTIFICATE AUTHORITY EXPIRES RESIDUAL TIME EXTERNALLY MANAGED
ca Dec 28, 2029 23:36 UTC 9y no
etcd-ca Dec 28, 2029 23:36 UTC 9y no
front-proxy-ca Dec 28, 2029 23:36 UTC 9y no
The command shows expiration/residual time for the client certificates in the /etc/kubernetes/pki folder and for the client certificate embedded in the KUBECONFIG files used by kubeadm (admin.conf, controller-manager.conf and scheduler.conf).
Additionally, kubeadm informs the user if the certificate is externally managed; in this case, the user should take care of managing certificate renewal manually/using other tools.
Warning:kubeadm cannot manage certificates signed by an external CA.
On nodes created with kubeadm init, prior to kubeadm version 1.17, there is a
bug where you manually have to modify the contents of kubelet.conf. After kubeadm init finishes, you should update kubelet.conf to point to the
rotated kubelet client certificates, by replacing client-certificate-data and client-key-data with:
kubeadm renews all the certificates during control plane upgrade.
This feature is designed for addressing the simplest use cases;
if you don't have specific requirements on certificate renewal and perform Kubernetes version upgrades regularly (less than 1 year in between each upgrade), kubeadm will take care of keeping your cluster up to date and reasonably secure.
Note: It is a best practice to upgrade your cluster frequently in order to stay secure.
If you have more complex requirements for certificate renewal, you can opt out from the default behavior by passing --certificate-renewal=false to kubeadm upgrade apply or to kubeadm upgrade node.
Warning: Prior to kubeadm version 1.17 there is a bug
where the default value for --certificate-renewal is false for the kubeadm upgrade node
command. In that case, you should explicitly set --certificate-renewal=true.
Manual certificate renewal
You can renew your certificates manually at any time with the kubeadm certs renew command.
This command performs the renewal using CA (or front-proxy-CA) certificate and key stored in /etc/kubernetes/pki.
After running the command you should restart the control plane Pods. This is required since
dynamic certificate reload is currently not supported for all components and certificates.
Static Pods are managed by the local kubelet
and not by the API Server, thus kubectl cannot be used to delete and restart them.
To restart a static Pod you can temporarily remove its manifest file from /etc/kubernetes/manifests/
and wait for 20 seconds (see the fileCheckFrequency value in KubeletConfiguration struct.
The kubelet will terminate the Pod if it's no longer in the manifest directory.
You can then move the file back and after another fileCheckFrequency period, the kubelet will recreate
the Pod and the certificate renewal for the component can complete.
Warning: If you are running an HA cluster, this command needs to be executed on all the control-plane nodes.
Note:certs renew uses the existing certificates as the authoritative source for attributes (Common Name, Organization, SAN, etc.) instead of the kubeadm-config ConfigMap. It is strongly recommended to keep them both in sync.
kubeadm certs renew provides the following options:
The Kubernetes certificates normally reach their expiration date after one year.
--csr-only can be used to renew certificates with an external CA by generating certificate signing requests (without actually renewing certificates in place); see next paragraph for more information.
It's also possible to renew a single certificate instead of all.
Renew certificates with the Kubernetes certificates API
This section provides more details about how to execute manual certificate renewal using the Kubernetes certificates API.
Caution: These are advanced topics for users who need to integrate their organization's certificate infrastructure into a kubeadm-built cluster. If the default kubeadm configuration satisfies your needs, you should let kubeadm manage certificates instead.
Set up a signer
The Kubernetes Certificate Authority does not work out of the box.
You can configure an external signer such as cert-manager, or you can use the built-in signer.
This section provide more details about how to execute manual certificate renewal using an external CA.
To better integrate with external CAs, kubeadm can also produce certificate signing requests (CSRs).
A CSR represents a request to a CA for a signed certificate for a client.
In kubeadm terms, any certificate that would normally be signed by an on-disk CA can be produced as a CSR instead. A CA, however, cannot be produced as a CSR.
Create certificate signing requests (CSR)
You can create certificate signing requests with kubeadm certs renew --csr-only.
Both the CSR and the accompanying private key are given in the output.
You can pass in a directory with --csr-dir to output the CSRs to the specified location.
If --csr-dir is not specified, the default certificate directory (/etc/kubernetes/pki) is used.
Certificates can be renewed with kubeadm certs renew --csr-only.
As with kubeadm init, an output directory can be specified with the --csr-dir flag.
A CSR contains a certificate's name, domains, and IPs, but it does not specify usages.
It is the responsibility of the CA to specify the correct cert usages
when issuing a certificate.
After a certificate is signed using your preferred method, the certificate and the private key must be copied to the PKI directory (by default /etc/kubernetes/pki).
Certificate authority (CA) rotation
Kubeadm does not support rotation or replacement of CA certificates out of the box.
By default the kubelet serving certificate deployed by kubeadm is self-signed.
This means a connection from external services like the
metrics-server to a
kubelet cannot be secured with TLS.
To configure the kubelets in a new kubeadm cluster to obtain properly signed serving
certificates you must pass the following minimal configuration to kubeadm init:
If you have already created the cluster you must adapt it by doing the following:
Find and edit the kubelet-config-1.25 ConfigMap in the kube-system namespace.
In that ConfigMap, the kubelet key has a
KubeletConfiguration
document as its value. Edit the KubeletConfiguration document to set serverTLSBootstrap: true.
On each node, add the serverTLSBootstrap: true field in /var/lib/kubelet/config.yaml
and restart the kubelet with systemctl restart kubelet
The field serverTLSBootstrap: true will enable the bootstrap of kubelet serving
certificates by requesting them from the certificates.k8s.io API. One known limitation
is that the CSRs (Certificate Signing Requests) for these certificates cannot be automatically
approved by the default signer in the kube-controller-manager -
kubernetes.io/kubelet-serving.
This will require action from the user or a third party controller.
These CSRs can be viewed using:
kubectl get csr
NAME AGE SIGNERNAME REQUESTOR CONDITION
csr-9wvgt 112s kubernetes.io/kubelet-serving system:node:worker-1 Pending
csr-lz97v 1m58s kubernetes.io/kubelet-serving system:node:control-plane-1 Pending
To approve them you can do the following:
kubectl certificate approve <CSR-name>
By default, these serving certificate will expire after one year. Kubeadm sets the
KubeletConfiguration field rotateCertificates to true, which means that close
to expiration a new set of CSRs for the serving certificates will be created and must
be approved to complete the rotation. To understand more see
Certificate Rotation.
If you are looking for a solution for automatic approval of these CSRs it is recommended
that you contact your cloud provider and ask if they have a CSR signer that verifies
the node identity with an out of band mechanism.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Such a controller is not a secure mechanism unless it not only verifies the CommonName
in the CSR but also verifies the requested IPs and domain names. This would prevent
a malicious actor that has access to a kubelet client certificate to create
CSRs requesting serving certificates for any IP or domain name.
Generating kubeconfig files for additional users
During cluster creation, kubeadm signs the certificate in the admin.conf to have
Subject: O = system:masters, CN = kubernetes-admin.
system:masters
is a break-glass, super user group that bypasses the authorization layer (e.g. RBAC).
Sharing the admin.conf with additional users is not recommended!
Instead, you can use the kubeadm kubeconfig user
command to generate kubeconfig files for additional users.
The command accepts a mixture of command line flags and
kubeadm configuration options.
The generated kubeconfig will be written to stdout and can be piped to a file
using kubeadm kubeconfig user ... > somefile.conf.
Example configuration file that can be used with --config:
# example.yamlapiVersion:kubeadm.k8s.io/v1beta3kind:ClusterConfiguration# Will be used as the target "cluster" in the kubeconfigclusterName:"kubernetes"# Will be used as the "server" (IP or DNS name) of this cluster in the kubeconfigcontrolPlaneEndpoint:"some-dns-address:6443"# The cluster CA key and certificate will be loaded from this local directorycertificatesDir:"/etc/kubernetes/pki"
Make sure that these settings match the desired target cluster settings.
To see the settings of an existing cluster use:
kubectl get cm kubeadm-config -n kube-system -o=jsonpath="{.data.ClusterConfiguration}"
The following example will generate a kubeconfig file with credentials valid for 24 hours
for a new user johndoe that is part of the appdevs group:
The Container runtimes page
explains that the systemd driver is recommended for kubeadm based setups instead
of the cgroupfs driver, because kubeadm manages the kubelet as a systemd service.
The page also provides details on how to set up a number of different container runtimes with the
systemd driver by default.
Configuring the kubelet cgroup driver
kubeadm allows you to pass a KubeletConfiguration structure during kubeadm init.
This KubeletConfiguration can include the cgroupDriver field which controls the cgroup
driver of the kubelet.
Note: In v1.22, if the user is not setting the cgroupDriver field under KubeletConfiguration,
kubeadm will default it to systemd.
A minimal example of configuring the field explicitly:
Such a configuration file can then be passed to the kubeadm command:
kubeadm init --config kubeadm-config.yaml
Note:
Kubeadm uses the same KubeletConfiguration for all nodes in the cluster.
The KubeletConfiguration is stored in a ConfigMap
object under the kube-system namespace.
Executing the sub commands init, join and upgrade would result in kubeadm
writing the KubeletConfiguration as a file under /var/lib/kubelet/config.yaml
and passing it to the local node kubelet.
Using the cgroupfs driver
To use cgroupfs and to prevent kubeadm upgrade from modifying the
KubeletConfiguration cgroup driver on existing setups, you must be explicit
about its value. This applies to a case where you do not wish future versions
of kubeadm to apply the systemd driver by default.
If you wish to configure a container runtime to use the cgroupfs driver,
you must refer to the documentation of the container runtime of your choice.
Migrating to the systemd driver
To change the cgroup driver of an existing kubeadm cluster to systemd in-place,
a similar procedure to a kubelet upgrade is required. This must include both
steps outlined below.
Note: Alternatively, it is possible to replace the old nodes in the cluster with new ones
that use the systemd driver. This requires executing only the first step below
before joining the new nodes and ensuring the workloads can safely move to the new
nodes before deleting the old nodes.
Modify the kubelet ConfigMap
Call kubectl edit cm kubelet-config -n kube-system.
Either modify the existing cgroupDriver value or add a new field that looks like this:
cgroupDriver:systemd
This field must be present under the kubelet: section of the ConfigMap.
Update the cgroup driver on all nodes
For each node in the cluster:
Drain the node using kubectl drain <node-name> --ignore-daemonsets
Stop the kubelet using systemctl stop kubelet
Stop the container runtime
Modify the container runtime cgroup driver to systemd
Set cgroupDriver: systemd in /var/lib/kubelet/config.yaml
Execute these steps on nodes one at a time to ensure workloads
have sufficient time to schedule on different nodes.
Once the process is complete ensure that all nodes and workloads are healthy.
4.2.1.3 - Reconfiguring a kubeadm cluster
kubeadm does not support automated ways of reconfiguring components that
were deployed on managed nodes. One way of automating this would be
by using a custom operator.
To modify the components configuration you must manually edit associated cluster
objects and files on disk.
This guide shows the correct sequence of steps that need to be performed
to achieve kubeadm cluster reconfiguration.
Before you begin
You need a cluster that was deployed using kubeadm
Have administrator credentials (/etc/kubernetes/admin.conf) and network connectivity
to a running kube-apiserver in the cluster from a host that has kubectl installed
Have a text editor installed on all hosts
Reconfiguring the cluster
kubeadm writes a set of cluster wide component configuration options in
ConfigMaps and other objects. These objects must be manually edited. The command kubectl edit
can be used for that.
The kubectl edit command will open a text editor where you can edit and save the object directly.
You can use the environment variables KUBECONFIG and KUBE_EDITOR to specify the location of
the kubectl consumed kubeconfig file and preferred text editor.
Note: Upon saving any changes to these cluster objects, components running on nodes may not be
automatically updated. The steps below instruct you on how to perform that manually.
Warning: Component configuration in ConfigMaps is stored as unstructured data (YAML string).
This means that validation will not be performed upon updating the contents of a ConfigMap.
You have to be careful to follow the documented API format for a particular
component configuration and avoid introducing typos and YAML indentation mistakes.
Applying cluster configuration changes
Updating the ClusterConfiguration
During cluster creation and upgrade, kubeadm writes its
ClusterConfiguration
in a ConfigMap called kubeadm-config in the kube-system namespace.
To change a particular option in the ClusterConfiguration you can edit the ConfigMap with this command:
kubectl edit cm -n kube-system kubeadm-config
The configuration is located under the data.ClusterConfiguration key.
Note: The ClusterConfiguration includes a variety of options that affect the configuration of individual
components such as kube-apiserver, kube-scheduler, kube-controller-manager, CoreDNS, etcd and kube-proxy.
Changes to the configuration must be reflected on node components manually.
Reflecting ClusterConfiguration changes on control plane nodes
kubeadm manages the control plane components as static Pod manifests located in
the directory /etc/kubernetes/manifests.
Any changes to the ClusterConfiguration under the apiServer, controllerManager, scheduler or etcd
keys must be reflected in the associated files in the manifests directory on a control plane node.
Such changes may include:
extraArgs - requires updating the list of flags passed to a component container
extraMounts - requires updated the volume mounts for a component container
*SANs - requires writing new certificates with updated Subject Alternative Names.
Before proceeding with these changes, make sure you have backed up the directory /etc/kubernetes/.
The <config-file> contents must match the updated ClusterConfiguration.
The <component-name> value must be the name of the component.
Note: Updating a file in /etc/kubernetes/manifests will tell the kubelet to restart the static Pod for the corresponding component.
Try doing these changes one node at a time to leave the cluster without downtime.
Applying kubelet configuration changes
Updating the KubeletConfiguration
During cluster creation and upgrade, kubeadm writes its
KubeletConfiguration
in a ConfigMap called kubelet-config in the kube-system namespace.
You can edit the ConfigMap with this command:
kubectl edit cm -n kube-system kubelet-config
The configuration is located under the data.kubelet key.
Reflecting the kubelet changes
To reflect the change on kubeadm nodes you must do the following:
Log in to a kubeadm node
Run kubeadm upgrade node phase kubelet-config to download the latest kubelet-config
ConfigMap contents into the local file /var/lib/kubelet/config.conf
Edit the file /var/lib/kubelet/kubeadm-flags.env to apply additional configuration with
flags
Restart the kubelet service with systemctl restart kubelet
Note: Do these changes one node at a time to allow workloads to be rescheduled properly.
Note: During kubeadm upgrade, kubeadm downloads the KubeletConfiguration from the
kubelet-config ConfigMap and overwrite the contents of /var/lib/kubelet/config.conf.
This means that node local configuration must be applied either by flags in
/var/lib/kubelet/kubeadm-flags.env or by manually updating the contents of
/var/lib/kubelet/config.conf after kubeadm upgrade, and then restarting the kubelet.
Applying kube-proxy configuration changes
Updating the KubeProxyConfiguration
During cluster creation and upgrade, kubeadm writes its
KubeProxyConfiguration
in a ConfigMap in the kube-system namespace called kube-proxy.
This ConfigMap is used by the kube-proxy DaemonSet in the kube-system namespace.
To change a particular option in the KubeProxyConfiguration, you can edit the ConfigMap with this command:
kubectl edit cm -n kube-system kube-proxy
The configuration is located under the data.config.conf key.
Reflecting the kube-proxy changes
Once the kube-proxy ConfigMap is updated, you can restart all kube-proxy Pods:
Obtain the Pod names:
kubectl get po -n kube-system | grep kube-proxy
Delete a Pod with:
kubectl delete po -n kube-system <pod-name>
New Pods that use the updated ConfigMap will be created.
Note: Because kubeadm deploys kube-proxy as a DaemonSet, node specific configuration is unsupported.
Applying CoreDNS configuration changes
Updating the CoreDNS Deployment and Service
kubeadm deploys CoreDNS as a Deployment called coredns and with a Service kube-dns,
both in the kube-system namespace.
To update any of the CoreDNS settings, you can edit the Deployment and
Service objects:
Once the CoreDNS changes are applied you can delete the CoreDNS Pods:
Obtain the Pod names:
kubectl get po -n kube-system | grep coredns
Delete a Pod with:
kubectl delete po -n kube-system <pod-name>
New Pods with the updated CoreDNS configuration will be created.
Note: kubeadm does not allow CoreDNS configuration during cluster creation and upgrade.
This means that if you execute kubeadm upgrade apply, your changes to the CoreDNS
objects will be lost and must be reapplied.
Persisting the reconfiguration
During the execution of kubeadm upgrade on a managed node, kubeadm might overwrite configuration
that was applied after the cluster was created (reconfiguration).
Persisting Node object reconfiguration
kubeadm writes Labels, Taints, CRI socket and other information on the Node object for a particular
Kubernetes node. To change any of the contents of this Node object you can use:
kubectl edit no <node-name>
During kubeadm upgrade the contents of such a Node might get overwritten.
If you would like to persist your modifications to the Node object after upgrade,
you can prepare a kubectl patch
and apply it to the Node object:
kubectl patch no <node-name> --patch-file <patch-file>
Persisting control plane component reconfiguration
The main source of control plane configuration is the ClusterConfiguration
object stored in the cluster. To extend the static Pod manifests configuration,
patches can be used.
These patch files must remain as files on the control plane nodes to ensure that
they can be used by the kubeadm upgrade ... --patches <directory>.
If reconfiguration is done to the ClusterConfiguration and static Pod manifests on disk,
the set of node specific patches must be updated accordingly.
Persisting kubelet reconfiguration
Any changes to the KubeletConfiguration stored in /var/lib/kubelet/config.conf will be overwritten on
kubeadm upgrade by downloading the contents of the cluster wide kubelet-config ConfigMap.
To persist kubelet node specific configuration either the file /var/lib/kubelet/config.conf
has to be updated manually post-upgrade or the file /var/lib/kubelet/kubeadm-flags.env can include flags.
The kubelet flags override the associated KubeletConfiguration options, but note that
some of the flags are deprecated.
A kubelet restart will be required after changing /var/lib/kubelet/config.conf or
/var/lib/kubelet/kubeadm-flags.env.
This page explains how to upgrade a Kubernetes cluster created with kubeadm from version
1.24.x to version 1.25.x, and from version
1.25.x to 1.25.y (where y > x). Skipping MINOR versions
when upgrading is unsupported. For more details, please visit Version Skew Policy.
To see information about upgrading clusters created using older versions of kubeadm,
please refer to following pages instead:
The cluster should use a static control plane and etcd pods or external etcd.
Make sure to back up any important components, such as app-level state stored in a database.
kubeadm upgrade does not touch your workloads, only components internal to Kubernetes, but backups are always a best practice.
The instructions below outline when to drain each node during the upgrade process.
If you are performing a minor version upgrade for any kubelet, you must
first drain the node (or nodes) that you are upgrading. In the case of control plane nodes,
they could be running CoreDNS Pods or other critical workloads. For more information see
Draining nodes.
All containers are restarted after upgrade, because the container spec hash value is changed.
To verify that the kubelet service has successfully restarted after the kubelet has been upgraded,
you can execute systemctl status kubelet or view the service logs with journalctl -xeu kubelet.
apt update
apt-cache madison kubeadm
# find the latest 1.25 version in the list
# it should look like 1.25.x-00, where x is the latest patch
yum list --showduplicates kubeadm --disableexcludes=kubernetes
# find the latest 1.25 version in the list
# it should look like 1.25.x-0, where x is the latest patch
Upgrading control plane nodes
The upgrade procedure on control plane nodes should be executed one node at a time.
Pick a control plane node that you wish to upgrade first. It must have the /etc/kubernetes/admin.conf file.
# replace x in 1.25.x-00 with the latest patch version apt-mark unhold kubeadm &&\
apt-get update && apt-get install -y kubeadm=1.25.x-00 &&\
apt-mark hold kubeadm
# replace x in 1.25.x-0 with the latest patch versionyum install -y kubeadm-1.25.x-0 --disableexcludes=kubernetes
Verify that the download works and has the expected version:
kubeadm version
Verify the upgrade plan:
kubeadm upgrade plan
This command checks that your cluster can be upgraded, and fetches the versions you can upgrade to.
It also shows a table with the component config version states.
Note:kubeadm upgrade also automatically renews the certificates that it manages on this node.
To opt-out of certificate renewal the flag --certificate-renewal=false can be used.
For more information see the certificate management guide.
Note: If kubeadm upgrade plan shows any component configs that require manual upgrade, users must provide
a config file with replacement configs to kubeadm upgrade apply via the --config command line flag.
Failing to do so will cause kubeadm upgrade apply to exit with an error and not perform an upgrade.
Choose a version to upgrade to, and run the appropriate command. For example:
# replace x with the patch version you picked for this upgradesudo kubeadm upgrade apply v1.25.x
Once the command finishes you should see:
[upgrade/successful] SUCCESS! Your cluster was upgraded to "v1.25.x". Enjoy!
[upgrade/kubelet] Now that your control plane is upgraded, please proceed with upgrading your kubelets if you haven't already done so.
Manually upgrade your CNI provider plugin.
Your Container Network Interface (CNI) provider may have its own upgrade instructions to follow.
Check the addons page to
find your CNI provider and see whether additional upgrade steps are required.
This step is not required on additional control plane nodes if the CNI provider runs as a DaemonSet.
For the other control plane nodes
Same as the first control plane node but use:
sudo kubeadm upgrade node
instead of:
sudo kubeadm upgrade apply
Also calling kubeadm upgrade plan and upgrading the CNI provider plugin is no longer needed.
Drain the node
Prepare the node for maintenance by marking it unschedulable and evicting the workloads:
# replace <node-to-drain> with the name of your node you are drainingkubectl drain <node-to-drain> --ignore-daemonsets
Bring the node back online by marking it schedulable:
# replace <node-to-drain> with the name of your nodekubectl uncordon <node-to-drain>
Upgrade worker nodes
The upgrade procedure on worker nodes should be executed one node at a time or few nodes at a time,
without compromising the minimum required capacity for running your workloads.
# replace x in 1.25.x-00 with the latest patch versionapt-mark unhold kubeadm &&\
apt-get update && apt-get install -y kubeadm=1.25.x-00 &&\
apt-mark hold kubeadm
# replace x in 1.25.x-0 with the latest patch versionyum install -y kubeadm-1.25.x-0 --disableexcludes=kubernetes
Call "kubeadm upgrade"
For worker nodes this upgrades the local kubelet configuration:
sudo kubeadm upgrade node
Drain the node
Prepare the node for maintenance by marking it unschedulable and evicting the workloads:
# replace <node-to-drain> with the name of your node you are drainingkubectl drain <node-to-drain> --ignore-daemonsets
Bring the node back online by marking it schedulable:
# replace <node-to-drain> with the name of your nodekubectl uncordon <node-to-drain>
Verify the status of the cluster
After the kubelet is upgraded on all nodes verify that all nodes are available again by running
the following command from anywhere kubectl can access the cluster:
kubectl get nodes
The STATUS column should show Ready for all your nodes, and the version number should be updated.
Recovering from a failure state
If kubeadm upgrade fails and does not roll back, for example because of an unexpected shutdown during execution, you can run kubeadm upgrade again.
This command is idempotent and eventually makes sure that the actual state is the desired state you declare.
To recover from a bad state, you can also run kubeadm upgrade apply --force without changing the version that your cluster is running.
During upgrade kubeadm writes the following backup folders under /etc/kubernetes/tmp:
kubeadm-backup-etcd-<date>-<time>
kubeadm-backup-manifests-<date>-<time>
kubeadm-backup-etcd contains a backup of the local etcd member data for this control plane Node.
In case of an etcd upgrade failure and if the automatic rollback does not work, the contents of this folder
can be manually restored in /var/lib/etcd. In case external etcd is used this backup folder will be empty.
kubeadm-backup-manifests contains a backup of the static Pod manifest files for this control plane Node.
In case of a upgrade failure and if the automatic rollback does not work, the contents of this folder can be
manually restored in /etc/kubernetes/manifests. If for some reason there is no difference between a pre-upgrade
and post-upgrade manifest file for a certain component, a backup file for it will not be written.
How it works
kubeadm upgrade apply does the following:
Checks that your cluster is in an upgradeable state:
The API server is reachable
All nodes are in the Ready state
The control plane is healthy
Enforces the version skew policies.
Makes sure the control plane images are available or available to pull to the machine.
Generates replacements and/or uses user supplied overwrites if component configs require version upgrades.
Upgrades the control plane components or rollbacks if any of them fails to come up.
Applies the new CoreDNS and kube-proxy manifests and makes sure that all necessary RBAC rules are created.
Creates new certificate and key files of the API server and backs up old files if they're about to expire in 180 days.
kubeadm upgrade node does the following on additional control plane nodes:
Fetches the kubeadm ClusterConfiguration from the cluster.
Optionally backups the kube-apiserver certificate.
Upgrades the static Pod manifests for the control plane components.
Upgrades the kubelet configuration for this node.
kubeadm upgrade node does the following on worker nodes:
Fetches the kubeadm ClusterConfiguration from the cluster.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Note: If you are running kube-proxy in a HostProcess container within a Pod, and not as a Windows Service, you can upgrade kube-proxy by applying a newer version of your kube-proxy manifests.
Uncordon the node
From a machine with access to the Kubernetes API,
bring the node back online by marking it schedulable:
# replace <node-to-drain> with the name of your nodekubectl uncordon <node-to-drain>
4.2.2 - Migrating from dockershim
This section presents information you need to know when migrating from
dockershim to other container runtimes.
Since the announcement of dockershim deprecation
in Kubernetes 1.20, there were questions on how this will affect various workloads and Kubernetes
installations. Our Dockershim Removal FAQ is there to help you
to understand the problem better.
Dockershim was removed from Kubernetes with the release of v1.24.
If you use Docker Engine via dockershim as your container runtime, and wish to upgrade to v1.24,
it is recommended that you either migrate to another runtime or find an alternative means to obtain Docker Engine support.
Check out container runtimes
section to know your options. Make sure to
report issues you encountered
with the migration. So the issue can be fixed in a timely manner and your cluster would be
ready for dockershim removal.
Your cluster might have more than one kind of node, although this is not a common
configuration.
Check out container runtimes
to understand your options for a container runtime.
There is a
GitHub issue
to track discussion about the deprecation and removal of dockershim.
If you found a defect or other technical concern relating to migrating away from dockershim,
you can report an issue
to the Kubernetes project.
4.2.2.1 - Changing the Container Runtime on a Node from Docker Engine to containerd
This task outlines the steps needed to update your container runtime to containerd from Docker. It
is applicable for cluster operators running Kubernetes 1.23 or earlier. This also covers an
example scenario for migrating from dockershim to containerd. Alternative container runtimes
can be picked from this page.
Before you begin
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Install the containerd.io package from the official Docker repositories.
Instructions for setting up the Docker repository for your respective Linux distribution and
installing the containerd.io package can be found at
Getting started with containerd.
Configure the kubelet to use containerd as its container runtime
Edit the file /var/lib/kubelet/kubeadm-flags.env and add the containerd runtime to the flags.
--container-runtime=remote and
--container-runtime-endpoint=unix:///run/containerd/containerd.sock.
Users using kubeadm should be aware that the kubeadm tool stores the CRI socket for each host as
an annotation in the Node object for that host. To change it you can execute the following command
on a machine that has the kubeadm /etc/kubernetes/admin.conf file.
kubectl edit no <node-name>
This will start a text editor where you can edit the Node object.
To choose a text editor you can set the KUBE_EDITOR environment variable.
Change the value of kubeadm.alpha.kubernetes.io/cri-socket from /var/run/dockershim.sock
to the CRI socket path of your choice (for example unix:///run/containerd/containerd.sock).
Note that new CRI socket paths must be prefixed with unix:// ideally.
Save the changes in the text editor, which will update the Node object.
Restart the kubelet
systemctl start kubelet
Verify that the node is healthy
Run kubectl get nodes -o wide and containerd appears as the runtime for the node we just changed.
Remove Docker Engine
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
The preceding commands don't remove images, containers, volumes, or customized configuration files on your host.
To delete them, follow Docker's instructions to Uninstall Docker Engine.
Caution: Docker's instructions for uninstalling Docker Engine create a risk of deleting containerd. Be careful when executing commands.
4.2.2.2 - Migrate Docker Engine nodes from dockershim to cri-dockerd
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
This page shows you how to migrate your Docker Engine nodes to use cri-dockerd
instead of dockershim. You should follow these steps in these scenarios:
You want to switch away from dockershim and still use Docker Engine to run
containers in Kubernetes.
You want to upgrade to Kubernetes v1.25 and your
existing cluster relies on dockershim, in which case you must migrate
from dockershim and cri-dockerd is one of your options.
To learn more about the removal of dockershim, read the FAQ page.
What is cri-dockerd?
In Kubernetes 1.23 and earlier, you could use Docker Engine with Kubernetes,
relying on a built-in component of Kubernetes named dockershim.
The dockershim component was removed in the Kubernetes 1.24 release; however,
a third-party replacement, cri-dockerd, is available. The cri-dockerd adapter
lets you use Docker Engine through the Container Runtime Interface.
If you want to migrate to cri-dockerd so that you can continue using Docker
Engine as your container runtime, you should do the following for each affected
node:
Install cri-dockerd.
Cordon and drain the node.
Configure the kubelet to use cri-dockerd.
Restart the kubelet.
Verify that the node is healthy.
Test the migration on non-critical nodes first.
You should perform the following steps for each node that you want to migrate
to cri-dockerd.
Cordon the node to stop new Pods scheduling on it:
kubectl cordon <NODE_NAME>
Replace <NODE_NAME> with the name of the node.
Drain the node to safely evict running Pods:
kubectl drain <NODE_NAME> \
--ignore-daemonsets
Configure the kubelet to use cri-dockerd
The following steps apply to clusters set up using the kubeadm tool. If you use
a different tool, you should modify the kubelet using the configuration
instructions for that tool.
Open /var/lib/kubelet/kubeadm-flags.env on each affected node.
Modify the --container-runtime-endpoint flag to
unix:///var/run/cri-dockerd.sock.
The kubeadm tool stores the node's socket as an annotation on the Node object
in the control plane. To modify this socket for each affected node:
Edit the YAML representation of the Node object:
KUBECONFIG=/path/to/admin.conf kubectl edit no <NODE_NAME>
Replace the following:
/path/to/admin.conf: the path to the kubectl configuration file,
admin.conf.
<NODE_NAME>: the name of the node you want to modify.
Change kubeadm.alpha.kubernetes.io/cri-socket from
/var/run/dockershim.sock to unix:///var/run/cri-dockerd.sock.
Save the changes. The Node object is updated on save.
Restart the kubelet
systemctl restart kubelet
Verify that the node is healthy
To check whether the node uses the cri-dockerd endpoint, follow the
instructions in Find out which runtime you use.
The --container-runtime-endpoint flag for the kubelet should be unix:///var/run/cri-dockerd.sock.
4.2.2.3 - Find Out What Container Runtime is Used on a Node
This page outlines steps to find out what container runtime
the nodes in your cluster use.
Depending on the way you run your cluster, the container runtime for the nodes may
have been pre-configured or you need to configure it. If you're using a managed
Kubernetes service, there might be vendor-specific ways to check what container runtime is
configured for the nodes. The method described on this page should work whenever
the execution of kubectl is allowed.
Before you begin
Install and configure kubectl. See Install Tools section for details.
Find out the container runtime used on a Node
Use kubectl to fetch and show node information:
kubectl get nodes -o wide
The output is similar to the following. The column CONTAINER-RUNTIME outputs
the runtime and its version.
For Docker Engine, the output is similar to this:
NAME STATUS VERSION CONTAINER-RUNTIME
node-1 Ready v1.16.15 docker://19.3.1
node-2 Ready v1.16.15 docker://19.3.1
node-3 Ready v1.16.15 docker://19.3.1
If your runtime shows as Docker Engine, you still might not be affected by the
removal of dockershim in Kubernetes v1.24. Check the runtime
endpoint to see if you use dockershim. If you don't use
dockershim, you aren't affected.
For containerd, the output is similar to this:
NAME STATUS VERSION CONTAINER-RUNTIME
node-1 Ready v1.19.6 containerd://1.4.1
node-2 Ready v1.19.6 containerd://1.4.1
node-3 Ready v1.19.6 containerd://1.4.1
Find out more information about container runtimes
on Container Runtimes
page.
Find out what container runtime endpoint you use
The container runtime talks to the kubelet over a Unix socket using the CRI
protocol, which is based on the gRPC
framework. The kubelet acts as a client, and the runtime acts as the server.
In some cases, you might find it useful to know which socket your nodes use. For
example, with the removal of dockershim in Kubernetes v1.24 and later, you might
want to know whether you use Docker Engine with dockershim.
Note: If you currently use Docker Engine in your nodes with cri-dockerd, you aren't
affected by the dockershim removal.
You can check which socket you use by checking the kubelet configuration on your
nodes.
Read the starting commands for the kubelet process:
tr \\0 ' ' < /proc/"$(pgrep kubelet)"/cmdline
If you don't have tr or pgrep, check the command line for the kubelet
process manually.
In the output, look for the --container-runtime flag and the
--container-runtime-endpoint flag.
If your nodes use Kubernetes v1.23 and earlier and these flags aren't
present or if the --container-runtime flag is not remote,
you use the dockershim socket with Docker Engine.
If the --container-runtime-endpoint flag is present, check the socket
name to find out which runtime you use. For example,
unix:///run/containerd/containerd.sock is the containerd endpoint.
If you want to change the Container Runtime on a Node from Docker Engine to containerd,
you can find out more information on migrating from Docker Engine to containerd,
or, if you want to continue using Docker Engine in Kubernetes v1.24 and later, migrate to a
CRI-compatible adapter like cri-dockerd.
To avoid CNI plugin-related errors, verify that you are using or upgrading to a
container runtime that has been tested to work correctly with your version of
Kubernetes.
About the "Incompatible CNI versions" and "Failed to destroy network for sandbox" errors
Service issues exist for pod CNI network setup and tear down in containerd
v1.6.0-v1.6.3 when the CNI plugins have not been upgraded and/or the CNI config
version is not declared in the CNI config files. The containerd team reports, "these issues are resolved in containerd v1.6.4."
With containerd v1.6.0-v1.6.3, if you do not upgrade the CNI plugins and/or
declare the CNI config version, you might encounter the following "Incompatible
CNI versions" or "Failed to destroy network for sandbox" error conditions.
Incompatible CNI versions error
If the version of your CNI plugin does not correctly match the plugin version in
the config because the config version is later than the plugin version, the
containerd log will likely show an error message on startup of a pod similar
to:
If the version of the plugin is missing in the CNI plugin config, the pod may
run. However, stopping the pod generates an error similar to:
ERRO[2022-04-26T00:43:24.518165483Z] StopPodSandbox for "b" failed
error="failed to destroy network for sandbox \"bbc85f891eaf060c5a879e27bba9b6b06450210161dfdecfbb2732959fb6500a\": invalid version \"\": the version is empty"
This error leaves the pod in the not-ready state with a network namespace still
attached. To recover from this problem, edit the CNI config file to add
the missing version information. The next attempt to stop the pod should
be successful.
Updating your CNI plugins and CNI config files
If you're using containerd v1.6.0-v1.6.3 and encountered "Incompatible CNI
versions" or "Failed to destroy network for sandbox" errors, consider updating
your CNI plugins and editing the CNI config files.
Here's an overview of the typical steps for each node:
After stopping your container runtime and kubelet services, perform the
following upgrade operations:
If you're running CNI plugins, upgrade them to the latest version.
If you're using non-CNI plugins, replace them with CNI plugins. Use the
latest version of the plugins.
Update the plugin configuration file to specify or match a version of the
CNI specification that the plugin supports, as shown in the following "An
example containerd configuration
file" section.
For containerd, ensure that you have installed the latest version (v1.0.0
or later) of the CNI loopback plugin.
Upgrade node components (for example, the kubelet) to Kubernetes v1.24
Upgrade to or install the most current version of the container runtime.
Bring the node back into your cluster by restarting your container runtime
and kubelet. Uncordon the node (kubectl uncordon <nodename>).
An example containerd configuration file
The following example shows a configuration for containerd runtime v1.6.x,
which supports a recent version of the CNI specification (v1.0.0).
Please see the documentation from your plugin and networking provider for
further instructions on configuring your system.
On Kubernetes, containerd runtime adds a loopback interface, lo, to pods as a
default behavior. The containerd runtime configures the loopback interface via a
CNI plugin, loopback. The loopback plugin is distributed as part of the
containerd release packages that have the cni designation. containerd
v1.6.0 and later includes a CNI v1.0.0-compatible loopback plugin as well as
other default CNI plugins. The configuration for the loopback plugin is done
internally by containerd, and is set to use CNI v1.0.0. This also means that the
version of the loopback plugin must be v1.0.0 or later when this newer version
containerd is started.
The following bash command generates an example CNI config. Here, the 1.0.0
value for the config version is assigned to the cniVersion field for use when
containerd invokes the CNI bridge plugin.
Update the IP address ranges in the preceding example with ones that are based
on your use case and network addressing plan.
4.2.2.5 - Check whether dockershim removal affects you
The dockershim component of Kubernetes allows to use Docker as a Kubernetes's
container runtime.
Kubernetes' built-in dockershim component was removed in release v1.24.
This page explains how your cluster could be using Docker as a container runtime,
provides details on the role that dockershim plays when in use, and shows steps
you can take to check whether any workloads could be affected by dockershim removal.
Finding if your app has a dependencies on Docker
If you are using Docker for building your application containers, you can still
run these containers on any container runtime. This use of Docker does not count
as a dependency on Docker as a container runtime.
When alternative container runtime is used, executing Docker commands may either
not work or yield unexpected output. This is how you can find whether you have a
dependency on Docker:
Make sure no privileged Pods execute Docker commands (like docker ps),
restart the Docker service (commands such as systemctl restart docker.service),
or modify Docker-specific files such as /etc/docker/daemon.json.
Check for any private registries or image mirror settings in the Docker
configuration file (like /etc/docker/daemon.json). Those typically need to
be reconfigured for another container runtime.
Check that scripts and apps running on nodes outside of your Kubernetes
infrastructure do not execute Docker commands. It might be:
SSH to nodes to troubleshoot;
Node startup scripts;
Monitoring and security agents installed on nodes directly.
Make sure there is no indirect dependencies on dockershim behavior.
This is an edge case and unlikely to affect your application. Some tooling may be configured
to react to Docker-specific behaviors, for example, raise alert on specific metrics or search for
a specific log message as part of troubleshooting instructions.
If you have such tooling configured, test the behavior on test
cluster before migration.
Dependency on Docker explained
A container runtime is software that can
execute the containers that make up a Kubernetes pod. Kubernetes is responsible for orchestration
and scheduling of Pods; on each node, the kubelet
uses the container runtime interface as an abstraction so that you can use any compatible
container runtime.
In its earliest releases, Kubernetes offered compatibility with one container runtime: Docker.
Later in the Kubernetes project's history, cluster operators wanted to adopt additional container runtimes.
The CRI was designed to allow this kind of flexibility - and the kubelet began supporting CRI. However,
because Docker existed before the CRI specification was invented, the Kubernetes project created an
adapter component, dockershim. The dockershim adapter allows the kubelet to interact with Docker as
if Docker were a CRI compatible runtime.
Switching to Containerd as a container runtime eliminates the middleman. All the
same containers can be run by container runtimes like Containerd as before. But
now, since containers schedule directly with the container runtime, they are not visible to Docker.
So any Docker tooling or fancy UI you might have used
before to check on these containers is no longer available.
You cannot get container information using docker ps or docker inspect
commands. As you cannot list containers, you cannot get logs, stop containers,
or execute something inside container using docker exec.
Note: If you're running workloads via Kubernetes, the best way to stop a container is through
the Kubernetes API rather than directly through the container runtime (this advice applies
for all container runtimes, not only Docker).
You can still pull images or build them using docker build command. But images
built or pulled by Docker would not be visible to container runtime and
Kubernetes. They needed to be pushed to some registry to allow them to be used
by Kubernetes.
4.2.2.6 - Migrating telemetry and security agents from dockershim
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Kubernetes' support for direct integration with Docker Engine is deprecated and
has been removed. Most apps do not have a direct dependency on runtime hosting
containers. However, there are still a lot of telemetry and monitoring agents
that have a dependency on Docker to collect containers metadata, logs, and
metrics. This document aggregates information on how to detect these
dependencies as well as links on how to migrate these agents to use generic tools or
alternative runtimes.
Telemetry and security agents
Within a Kubernetes cluster there are a few different ways to run telemetry or
security agents. Some agents have a direct dependency on Docker Engine when
they run as DaemonSets or directly on nodes.
Why do some telemetry agents communicate with Docker Engine?
Historically, Kubernetes was written to work specifically with Docker Engine.
Kubernetes took care of networking and scheduling, relying on Docker Engine for
launching and running containers (within Pods) on a node. Some information that
is relevant to telemetry, such as a pod name, is only available from Kubernetes
components. Other data, such as container metrics, is not the responsibility of
the container runtime. Early telemetry agents needed to query the container
runtime and Kubernetes to report an accurate picture. Over time, Kubernetes
gained the ability to support multiple runtimes, and now supports any runtime
that is compatible with the container runtime interface.
Some telemetry agents rely specifically on Docker Engine tooling. For example, an agent
might run a command such as
docker ps
or docker top to list
containers and processes or docker logs
to receive streamed logs. If nodes in your existing cluster use
Docker Engine, and you switch to a different container runtime,
these commands will not work any longer.
Identify DaemonSets that depend on Docker Engine
If a pod wants to make calls to the dockerd running on the node, the pod must either:
mount the filesystem containing the Docker daemon's privileged socket, as a
volume; or
mount the specific path of the Docker daemon's privileged socket directly, also as a volume.
For example: on COS images, Docker exposes its Unix domain socket at
/var/run/docker.sock This means that the pod spec will include a
hostPath volume mount of /var/run/docker.sock.
Here's a sample shell script to find Pods that have a mount directly mapping the
Docker socket. This script outputs the namespace and name of the pod. You can
remove the grep '/var/run/docker.sock' to review other mounts.
Note: There are alternative ways for a pod to access Docker on the host. For instance, the parent
directory /var/run may be mounted instead of the full path (like in this
example).
The script above only detects the most common uses.
Detecting Docker dependency from node agents
If your cluster nodes are customized and install additional security and
telemetry agents on the node, check with the agent vendor
to verify whether it has any dependency on Docker.
Telemetry and security agent vendors
This section is intended to aggregate information about various telemetry and
security agents that may have a dependency on container runtimes.
We keep the work in progress version of migration instructions for various telemetry and security agent vendors
in Google doc.
Please contact the vendor to get up to date instructions for migrating from dockershim.
Migrate Falco from dockershim
Falco supports any CRI-compatible runtime (containerd is used in the default configuration); the documentation explains all details.
The pod accessing Docker may have name containing:
Check documentation for Prisma Cloud,
under the "Install Prisma Cloud on a CRI (non-Docker) cluster" section.
The pod accessing Docker may be named like:
The SignalFx Smart Agent (deprecated) uses several different monitors for Kubernetes including
kubernetes-cluster, kubelet-stats/kubelet-metrics, and docker-container-stats.
The kubelet-stats monitor was previously deprecated by the vendor, in favor of kubelet-metrics.
The docker-container-stats monitor is the one affected by dockershim removal.
Do not use the docker-container-stats with container runtimes other than Docker Engine.
How to migrate from dockershim-dependent agent:
Remove docker-container-stats from the list of configured monitors.
Note, keeping this monitor enabled with non-dockershim runtime will result in incorrect metrics
being reported when docker is installed on node and no metrics when docker is not installed.
When using client certificate authentication, you can generate certificates
manually through easyrsa, openssl or cfssl.
easyrsa
easyrsa can manually generate certificates for your cluster.
Download, unpack, and initialize the patched version of easyrsa3.
curl -LO https://storage.googleapis.com/kubernetes-release/easy-rsa/easy-rsa.tar.gz
tar xzf easy-rsa.tar.gz
cd easy-rsa-master/easyrsa3
./easyrsa init-pki
Generate a new certificate authority (CA). --batch sets automatic mode;
--req-cn specifies the Common Name (CN) for the CA's new root certificate.
The argument --subject-alt-name sets the possible IPs and DNS names the API server will
be accessed with. The MASTER_CLUSTER_IP is usually the first IP from the service CIDR
that is specified as the --service-cluster-ip-range argument for both the API server and
the controller manager component. The argument --days is used to set the number of days
after which the certificate expires.
The sample below also assumes that you are using cluster.local as the default
DNS domain name.
Create a config file for generating a Certificate Signing Request (CSR).
Be sure to substitute the values marked with angle brackets (e.g. <MASTER_IP>)
with real values before saving this to a file (e.g. csr.conf).
Note that the value for MASTER_CLUSTER_IP is the service cluster IP for the
API server as described in previous subsection.
The sample below also assumes that you are using cluster.local as the default
DNS domain name.
Create a JSON config file for CA certificate signing request (CSR), for example,
ca-csr.json. Be sure to replace the values marked with angle brackets with
real values you want to use.
Generate CA key (ca-key.pem) and certificate (ca.pem):
../cfssl gencert -initca ca-csr.json | ../cfssljson -bare ca
Create a JSON config file for generating keys and certificates for the API
server, for example, server-csr.json. Be sure to replace the values in angle brackets with
real values you want to use. The <MASTER_CLUSTER_IP> is the service cluster
IP for the API server as described in previous subsection.
The sample below also assumes that you are using cluster.local as the default
DNS domain name.
A client node may refuse to recognize a self-signed CA certificate as valid.
For a non-production deployment, or for a deployment that runs behind a company
firewall, you can distribute a self-signed CA certificate to all clients and
refresh the local list for valid certificates.
Updating certificates in /etc/ssl/certs...
1 added, 0 removed; done.
Running hooks in /etc/ca-certificates/update.d....
done.
Certificates API
You can use the certificates.k8s.io API to provision
x509 certificates to use for authentication as documented
in the Managing TLS in a cluster
task page.
4.2.4 - Manage Memory, CPU, and API Resources
4.2.4.1 - Configure Default Memory Requests and Limits for a Namespace
Define a default memory resource limit for a namespace, so that every new Pod in that namespace has a memory resource limit configured.
This page shows how to configure default memory requests and limits for a
namespace.
A Kubernetes cluster can be divided into namespaces. Once you have a namespace that
has a default memory
limit,
and you then try to create a Pod with a container that does not specify its own memory
limit, then the
control plane assigns the default
memory limit to that container.
Kubernetes assigns a default memory request under certain conditions that are explained later in this topic.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Now if you create a Pod in the default-mem-example namespace, and any container
within that Pod does not specify its own values for memory request and memory limit,
then the control plane
applies default values: a memory request of 256MiB and a memory limit of 512MiB.
Here's an example manifest for a Pod that has one container. The container
does not specify a memory request and limit.
kubectl get pod default-mem-demo --output=yaml --namespace=default-mem-example
The output shows that the Pod's container has a memory request of 256 MiB and
a memory limit of 512 MiB. These are the default values specified by the LimitRange.
kubectl get pod default-mem-demo-2 --output=yaml --namespace=default-mem-example
The output shows that the container's memory request is set to match its memory limit.
Notice that the container was not assigned the default memory request value of 256Mi.
kubectl get pod default-mem-demo-3 --output=yaml --namespace=default-mem-example
The output shows that the container's memory request is set to the value specified in the
container's manifest. The container is limited to use no more than 512MiB of
memory, which matches the default memory limit for the namespace.
If your namespace has a memory resource quota
configured,
it is helpful to have a default value in place for memory limit.
Here are three of the restrictions that a resource quota imposes on a namespace:
For every Pod that runs in the namespace, the Pod and each of its containers must have a memory limit.
(If you specify a memory limit for every container in a Pod, Kubernetes can infer the Pod-level memory
limit by adding up the limits for its containers).
Memory limits apply a resource reservation on the node where the Pod in question is scheduled.
The total amount of memory reserved for all Pods in the namespace must not exceed a specified limit.
The total amount of memory actually used by all Pods in the namespace must also not exceed a specified limit.
When you add a LimitRange:
If any Pod in that namespace that includes a container does not specify its own memory limit,
the control plane applies the default memory limit to that container, and the Pod can be
allowed to run in a namespace that is restricted by a memory ResourceQuota.
4.2.4.2 - Configure Default CPU Requests and Limits for a Namespace
Define a default CPU resource limits for a namespace, so that every new Pod in that namespace has a CPU resource limit configured.
This page shows how to configure default CPU requests and limits for a
namespace.
A Kubernetes cluster can be divided into namespaces. If you create a Pod within a
namespace that has a default CPU
limit, and any container in that Pod does not specify
its own CPU limit, then the
control plane assigns the default
CPU limit to that container.
Kubernetes assigns a default CPU
request,
but only under certain conditions that are explained later in this page.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Now if you create a Pod in the default-cpu-example namespace, and any container
in that Pod does not specify its own values for CPU request and CPU limit,
then the control plane applies default values: a CPU request of 0.5 and a default
CPU limit of 1.
Here's a manifest for a Pod that has one container. The container
does not specify a CPU request and limit.
kubectl get pod default-cpu-demo --output=yaml --namespace=default-cpu-example
The output shows that the Pod's only container has a CPU request of 500m cpu
(which you can read as “500 millicpu”), and a CPU limit of 1 cpu.
These are the default values specified by the LimitRange.
kubectl get pod default-cpu-demo-2 --output=yaml --namespace=default-cpu-example
The output shows that the container's CPU request is set to match its CPU limit.
Notice that the container was not assigned the default CPU request value of 0.5 cpu:
resources:
limits:
cpu: "1"
requests:
cpu: "1"
What if you specify a container's request, but not its limit?
Here's an example manifest for a Pod that has one container. The container
specifies a CPU request, but not a limit:
kubectl get pod default-cpu-demo-3 --output=yaml --namespace=default-cpu-example
The output shows that the container's CPU request is set to the value you specified at
the time you created the Pod (in other words: it matches the manifest).
However, the same container's CPU limit is set to 1 cpu, which is the default CPU limit
for that namespace.
resources:
limits:
cpu: "1"
requests:
cpu: 750m
Motivation for default CPU limits and requests
If your namespace has a CPU resource quota
configured,
it is helpful to have a default value in place for CPU limit.
Here are two of the restrictions that a CPU resource quota imposes on a namespace:
For every Pod that runs in the namespace, each of its containers must have a CPU limit.
CPU limits apply a resource reservation on the node where the Pod in question is scheduled.
The total amount of CPU that is reserved for use by all Pods in the namespace must not
exceed a specified limit.
When you add a LimitRange:
If any Pod in that namespace that includes a container does not specify its own CPU limit,
the control plane applies the default CPU limit to that container, and the Pod can be
allowed to run in a namespace that is restricted by a CPU ResourceQuota.
4.2.4.3 - Configure Minimum and Maximum Memory Constraints for a Namespace
Define a range of valid memory resource limits for a namespace, so that every new Pod in that namespace falls within the range you configure.
This page shows how to set minimum and maximum values for memory used by containers
running in a namespace.
You specify minimum and maximum memory values in a
LimitRange
object. If a Pod does not meet the constraints imposed by the LimitRange,
it cannot be created in the namespace.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
kubectl get limitrange mem-min-max-demo-lr --namespace=constraints-mem-example --output=yaml
The output shows the minimum and maximum memory constraints as expected. But
notice that even though you didn't specify default values in the configuration
file for the LimitRange, they were created automatically.
Now whenever you define a Pod within the constraints-mem-example namespace, Kubernetes
performs these steps:
If any container in that Pod does not specify its own memory request and limit,
the control plane assigns the default memory request and limit to that container.
Verify that every container in that Pod requests at least 500 MiB of memory.
Verify that every container in that Pod requests no more than 1024 MiB (1 GiB)
of memory.
Here's a manifest for a Pod that has one container. Within the Pod spec, the sole
container specifies a memory request of 600 MiB and a memory limit of 800 MiB. These satisfy the
minimum and maximum memory constraints imposed by the LimitRange.
Verify that the Pod is running and that its container is healthy:
kubectl get pod constraints-mem-demo --namespace=constraints-mem-example
View detailed information about the Pod:
kubectl get pod constraints-mem-demo --output=yaml --namespace=constraints-mem-example
The output shows that the container within that Pod has a memory request of 600 MiB and
a memory limit of 800 MiB. These satisfy the constraints imposed by the LimitRange for
this namespace:
The output shows that the Pod does not get created, because it defines a container that
requests more memory than is allowed:
Error from server (Forbidden): error when creating "examples/admin/resource/memory-constraints-pod-2.yaml":
pods "constraints-mem-demo-2" is forbidden: maximum memory usage per Container is 1Gi, but limit is 1536Mi.
Attempt to create a Pod that does not meet the minimum memory request
Here's a manifest for a Pod that has one container. That container specifies a
memory request of 100 MiB and a memory limit of 800 MiB.
The output shows that the Pod does not get created, because it defines a container
that requests less memory than the enforced minimum:
Error from server (Forbidden): error when creating "examples/admin/resource/memory-constraints-pod-3.yaml":
pods "constraints-mem-demo-3" is forbidden: minimum memory usage per Container is 500Mi, but request is 100Mi.
Create a Pod that does not specify any memory request or limit
Here's a manifest for a Pod that has one container. The container does not
specify a memory request, and it does not specify a memory limit.
Because your Pod did not define any memory request and limit for that container, the cluster
applied a
default memory request and limit
from the LimitRange.
This means that the definition of that Pod shows those values. You can check it using
kubectl describe:
# Look for the "Requests:" section of the outputkubectl describe pod constraints-mem-demo-4 --namespace=constraints-mem-example
At this point, your Pod might be running or it might not be running. Recall that a prerequisite
for this task is that your Nodes have at least 1 GiB of memory. If each of your Nodes has only
1 GiB of memory, then there is not enough allocatable memory on any Node to accommodate a memory
request of 1 GiB. If you happen to be using Nodes with 2 GiB of memory, then you probably have
enough space to accommodate the 1 GiB request.
Delete your Pod:
kubectl delete pod constraints-mem-demo-4 --namespace=constraints-mem-example
Enforcement of minimum and maximum memory constraints
The maximum and minimum memory constraints imposed on a namespace by a LimitRange are enforced only
when a Pod is created or updated. If you change the LimitRange, it does not affect
Pods that were created previously.
Motivation for minimum and maximum memory constraints
As a cluster administrator, you might want to impose restrictions on the amount of memory that Pods can use.
For example:
Each Node in a cluster has 2 GiB of memory. You do not want to accept any Pod that requests
more than 2 GiB of memory, because no Node in the cluster can support the request.
A cluster is shared by your production and development departments.
You want to allow production workloads to consume up to 8 GiB of memory, but
you want development workloads to be limited to 512 MiB. You create separate namespaces
for production and development, and you apply memory constraints to each namespace.
4.2.4.4 - Configure Minimum and Maximum CPU Constraints for a Namespace
Define a range of valid CPU resource limits for a namespace, so that every new Pod in that namespace falls within the range you configure.
This page shows how to set minimum and maximum values for the CPU resources used by containers
and Pods in a namespace. You specify minimum
and maximum CPU values in a
LimitRange
object. If a Pod does not meet the constraints imposed by the LimitRange, it cannot be created
in the namespace.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
kubectl get limitrange cpu-min-max-demo-lr --output=yaml --namespace=constraints-cpu-example
The output shows the minimum and maximum CPU constraints as expected. But
notice that even though you didn't specify default values in the configuration
file for the LimitRange, they were created automatically.
Now whenever you create a Pod in the constraints-cpu-example namespace (or some other client
of the Kubernetes API creates an equivalent Pod), Kubernetes performs these steps:
If any container in that Pod does not specify its own CPU request and limit, the control plane
assigns the default CPU request and limit to that container.
Verify that every container in that Pod specifies a CPU request that is greater than or equal to 200 millicpu.
Verify that every container in that Pod specifies a CPU limit that is less than or equal to 800 millicpu.
Note: When creating a LimitRange object, you can specify limits on huge-pages
or GPUs as well. However, when both default and defaultRequest are specified
on these resources, the two values must be the same.
Here's a manifest for a Pod that has one container. The container manifest
specifies a CPU request of 500 millicpu and a CPU limit of 800 millicpu. These satisfy the
minimum and maximum CPU constraints imposed by the LimitRange for this namespace.
Verify that the Pod is running and that its container is healthy:
kubectl get pod constraints-cpu-demo --namespace=constraints-cpu-example
View detailed information about the Pod:
kubectl get pod constraints-cpu-demo --output=yaml --namespace=constraints-cpu-example
The output shows that the Pod's only container has a CPU request of 500 millicpu and CPU limit
of 800 millicpu. These satisfy the constraints imposed by the LimitRange.
resources:limits:cpu:800mrequests:cpu:500m
Delete the Pod
kubectl delete pod constraints-cpu-demo --namespace=constraints-cpu-example
Attempt to create a Pod that exceeds the maximum CPU constraint
Here's a manifest for a Pod that has one container. The container specifies a
CPU request of 500 millicpu and a cpu limit of 1.5 cpu.
The output shows that the Pod does not get created, because it defines an unacceptable container.
That container is not acceptable because it specifies a CPU limit that is too large:
Error from server (Forbidden): error when creating "examples/admin/resource/cpu-constraints-pod-2.yaml":
pods "constraints-cpu-demo-2" is forbidden: maximum cpu usage per Container is 800m, but limit is 1500m.
Attempt to create a Pod that does not meet the minimum CPU request
Here's a manifest for a Pod that has one container. The container specifies a
CPU request of 100 millicpu and a CPU limit of 800 millicpu.
The output shows that the Pod does not get created, because it defines an unacceptable container.
That container is not acceptable because it specifies a CPU request that is lower than the
enforced minimum:
Error from server (Forbidden): error when creating "examples/admin/resource/cpu-constraints-pod-3.yaml":
pods "constraints-cpu-demo-3" is forbidden: minimum cpu usage per Container is 200m, but request is 100m.
Create a Pod that does not specify any CPU request or limit
Here's a manifest for a Pod that has one container. The container does not
specify a CPU request, nor does it specify a CPU limit.
kubectl get pod constraints-cpu-demo-4 --namespace=constraints-cpu-example --output=yaml
The output shows that the Pod's single container has a CPU request of 800 millicpu and a
CPU limit of 800 millicpu.
How did that container get those values?
resources:limits:cpu:800mrequests:cpu:800m
Because that container did not specify its own CPU request and limit, the control plane
applied the
default CPU request and limit
from the LimitRange for this namespace.
At this point, your Pod may or may not be running. Recall that a prerequisite for
this task is that your Nodes must have at least 1 CPU available for use. If each of your Nodes has only 1 CPU,
then there might not be enough allocatable CPU on any Node to accommodate a request of 800 millicpu.
If you happen to be using Nodes with 2 CPU, then you probably have enough CPU to accommodate the 800 millicpu request.
Delete your Pod:
kubectl delete pod constraints-cpu-demo-4 --namespace=constraints-cpu-example
Enforcement of minimum and maximum CPU constraints
The maximum and minimum CPU constraints imposed on a namespace by a LimitRange are enforced only
when a Pod is created or updated. If you change the LimitRange, it does not affect
Pods that were created previously.
Motivation for minimum and maximum CPU constraints
As a cluster administrator, you might want to impose restrictions on the CPU resources that Pods can use.
For example:
Each Node in a cluster has 2 CPU. You do not want to accept any Pod that requests
more than 2 CPU, because no Node in the cluster can support the request.
A cluster is shared by your production and development departments.
You want to allow production workloads to consume up to 3 CPU, but you want development workloads to be limited
to 1 CPU. You create separate namespaces for production and development, and you apply CPU constraints to
each namespace.
4.2.4.5 - Configure Memory and CPU Quotas for a Namespace
Define overall memory and CPU resource limits for a namespace.
This page shows how to set quotas for the total amount memory and CPU that
can be used by all Pods running in a namespace.
You specify quotas in a
ResourceQuota
object.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Verify that the Pod is running and that its (only) container is healthy:
kubectl get pod quota-mem-cpu-demo --namespace=quota-mem-cpu-example
Once again, view detailed information about the ResourceQuota:
kubectl get resourcequota mem-cpu-demo --namespace=quota-mem-cpu-example --output=yaml
The output shows the quota along with how much of the quota has been used.
You can see that the memory and CPU requests and limits for your Pod do not
exceed the quota.
In the manifest, you can see that the Pod has a memory request of 700 MiB.
Notice that the sum of the used memory request and this new memory
request exceeds the memory request quota: 600 MiB + 700 MiB > 1 GiB.
The second Pod does not get created. The output shows that creating the second Pod
would cause the memory request total to exceed the memory request quota.
Error from server (Forbidden): error when creating "examples/admin/resource/quota-mem-cpu-pod-2.yaml":
pods "quota-mem-cpu-demo-2" is forbidden: exceeded quota: mem-cpu-demo,
requested: requests.memory=700Mi,used: requests.memory=600Mi, limited: requests.memory=1Gi
Discussion
As you have seen in this exercise, you can use a ResourceQuota to restrict
the memory request total for all Pods running in a namespace.
You can also restrict the totals for memory limit, cpu request, and cpu limit.
Instead of managing total resource use within a namespace, you might want to restrict
individual Pods, or the containers in those Pods. To achieve that kind of limiting, use a
LimitRange.
Restrict how many Pods you can create within a namespace.
This page shows how to set a quota for the total number of Pods that can run
in a Namespace. You specify quotas in a
ResourceQuota
object.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
kubectl get deployment pod-quota-demo --namespace=quota-pod-example --output=yaml
The output shows that even though the Deployment specifies three replicas, only two
Pods were created because of the quota you defined earlier:
spec:...replicas:3...status:availableReplicas:2...lastUpdateTime:2021-04-02T20:57:05Zmessage: 'unable to create pods:pods "pod-quota-demo-1650323038-" is forbidden:exceeded quota: pod-demo, requested: pods=1, used: pods=2, limited:pods=2'
Choice of resource
In this task you have defined a ResourceQuota that limited the total number of Pods, but
you could also limit the total number of other kinds of object. For example, you
might decide to limit how many CronJobs
that can live in a single namespace.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
To get familiar with Cilium easily you can follow the
Cilium Kubernetes Getting Started Guide
to perform a basic DaemonSet installation of Cilium in minikube.
To start minikube, minimal version required is >= v1.5.2, run the with the
following arguments:
minikube version
minikube version: v1.5.2
minikube start --network-plugin=cni
For minikube you can install Cilium using its CLI tool. Cilium will
automatically detect the cluster configuration and will install the appropriate
components for a successful installation:
🔮 Auto-detected Kubernetes kind: minikube
✨ Running "minikube" validation checks
✅ Detected minikube version "1.20.0"
ℹ️ Cilium version not set, using default version "v1.10.0"
🔮 Auto-detected cluster name: minikube
🔮 Auto-detected IPAM mode: cluster-pool
🔮 Auto-detected datapath mode: tunnel
🔑 Generating CA...
2021/05/27 02:54:44 [INFO] generate received request
2021/05/27 02:54:44 [INFO] received CSR
2021/05/27 02:54:44 [INFO] generating key: ecdsa-256
2021/05/27 02:54:44 [INFO] encoded CSR
2021/05/27 02:54:44 [INFO] signed certificate with serial number 48713764918856674401136471229482703021230538642
🔑 Generating certificates for Hubble...
2021/05/27 02:54:44 [INFO] generate received request
2021/05/27 02:54:44 [INFO] received CSR
2021/05/27 02:54:44 [INFO] generating key: ecdsa-256
2021/05/27 02:54:44 [INFO] encoded CSR
2021/05/27 02:54:44 [INFO] signed certificate with serial number 3514109734025784310086389188421560613333279574
🚀 Creating Service accounts...
🚀 Creating Cluster roles...
🚀 Creating ConfigMap...
🚀 Creating Agent DaemonSet...
🚀 Creating Operator Deployment...
⌛ Waiting for Cilium to be installed...
The remainder of the Getting Started Guide explains how to enforce both L3/L4
(i.e., IP address + port) security policies, as well as L7 (e.g., HTTP) security
policies using an example application.
Deploying Cilium for Production Use
For detailed instructions around deploying Cilium for production, see:
Cilium Kubernetes Installation Guide
This documentation includes detailed requirements, instructions and example
production DaemonSet files.
Understanding Cilium components
Deploying a cluster with Cilium adds Pods to the kube-system namespace. To see
this list of Pods run:
kubectl get pods --namespace=kube-system -l k8s-app=cilium
You'll see a list of Pods similar to this:
NAME READY STATUS RESTARTS AGE
cilium-kkdhz 1/1 Running 0 3m23s
...
A cilium Pod runs on each node in your cluster and enforces network policy
on the traffic to/from Pods on that node using Linux BPF.
What's next
Once your cluster is running, you can follow the
Declare Network Policy
to try out Kubernetes NetworkPolicy with Cilium.
Have fun, and if you have questions, contact us using the
Cilium Slack Channel.
4.2.5.4 - Use Kube-router for NetworkPolicy
This page shows how to use Kube-router for NetworkPolicy.
Before you begin
You need to have a Kubernetes cluster running. If you do not already have a cluster, you can create one by using any of the cluster installers like Kops, Bootkube, Kubeadm etc.
Installing Kube-router addon
The Kube-router Addon comes with a Network Policy Controller that watches Kubernetes API server for any NetworkPolicy and pods updated and configures iptables rules and ipsets to allow or block traffic as directed by the policies. Please follow the trying Kube-router with cluster installers guide to install Kube-router addon.
What's next
Once you have installed the Kube-router addon, you can follow the Declare Network Policy to try out Kubernetes NetworkPolicy.
4.2.5.5 - Romana for NetworkPolicy
This page shows how to use Romana for NetworkPolicy.
The Weave Net addon for Kubernetes comes with a
Network Policy Controller
that automatically monitors Kubernetes for any NetworkPolicy annotations on all
namespaces and configures iptables rules to allow or block traffic as directed by the policies.
Test the installation
Verify that the weave works.
Enter the following command:
kubectl get pods -n kube-system -o wide
The output is similar to this:
NAME READY STATUS RESTARTS AGE IP NODE
weave-net-1t1qg 2/2 Running 0 9d 192.168.2.10 worknode3
weave-net-231d7 2/2 Running 1 7d 10.2.0.17 worknodegpu
weave-net-7nmwt 2/2 Running 3 9d 192.168.2.131 masternode
weave-net-pmw8w 2/2 Running 0 9d 192.168.2.216 worknode2
Each Node has a weave Pod, and all Pods are Running and 2/2 READY. (2/2 means that each Pod has weave and weave-npc.)
This page shows how to access clusters using the Kubernetes API.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
When accessing the Kubernetes API for the first time, use the
Kubernetes command-line tool, kubectl.
To access a cluster, you need to know the location of the cluster and have credentials
to access it. Typically, this is automatically set-up when you work through
a Getting started guide,
or someone else set up the cluster and provided you with credentials and a location.
Check the location and credentials that kubectl knows about with this command:
kubectl config view
Many of the examples provide an introduction to using
kubectl. Complete documentation is found in the kubectl manual.
Directly accessing the REST API
kubectl handles locating and authenticating to the API server. If you want to directly access the REST API with an http client like
curl or wget, or a browser, there are multiple ways you can locate and authenticate against the API server:
Run kubectl in proxy mode (recommended). This method is recommended, since it uses the stored apiserver location and verifies the identity of the API server using a self-signed cert. No man-in-the-middle (MITM) attack is possible using this method.
Alternatively, you can provide the location and credentials directly to the http client. This works with client code that is confused by proxies. To protect against man in the middle attacks, you'll need to import a root cert into your browser.
Using the Go or Python client libraries provides accessing kubectl in proxy mode.
Using kubectl proxy
The following command runs kubectl in a mode where it acts as a reverse proxy. It handles
locating the API server and authenticating.
It is possible to avoid using kubectl proxy by passing an authentication token
directly to the API server, like this:
Using grep/cut approach:
# Check all possible clusters, as your .KUBECONFIG may have multiple contexts:kubectl config view -o jsonpath='{"Cluster name\tServer\n"}{range .clusters[*]}{.name}{"\t"}{.cluster.server}{"\n"}{end}'# Select name of cluster you want to interact with from above output:exportCLUSTER_NAME="some_server_name"# Point to the API server referring the cluster nameAPISERVER=$(kubectl config view -o jsonpath="{.clusters[?(@.name==\"$CLUSTER_NAME\")].cluster.server}")# Create a secret to hold a token for the default service accountkubectl apply -f - <<EOF
apiVersion: v1
kind: Secret
metadata:
name: default-token
annotations:
kubernetes.io/service-account.name: default
type: kubernetes.io/service-account-token
EOF# Wait for the token controller to populate the secret with a token:while ! kubectl describe secret default-token | grep -E '^token' >/dev/null; doecho"waiting for token..." >&2 sleep 1done# Get the token valueTOKEN=$(kubectl get secret default-token -o jsonpath='{.data.token}' | base64 --decode)# Explore the API with TOKENcurl -X GET $APISERVER/api --header "Authorization: Bearer $TOKEN" --insecure
The above example uses the --insecure flag. This leaves it subject to MITM
attacks. When kubectl accesses the cluster it uses a stored root certificate
and client certificates to access the server. (These are installed in the
~/.kube directory). Since cluster certificates are typically self-signed, it
may take special configuration to get your http client to use root
certificate.
On some clusters, the API server does not require authentication; it may serve
on localhost, or be protected by a firewall. There is not a standard
for this. Controlling Access to the Kubernetes API
describes how you can configure this as a cluster administrator.
Programmatic access to the API
Kubernetes officially supports client libraries for Go, Python, Java, dotnet, JavaScript, and Haskell. There are other client libraries that are provided and maintained by their authors, not the Kubernetes team. See client libraries for accessing the API from other languages and how they authenticate.
Go client
To get the library, run the following command: go get k8s.io/client-go@kubernetes-<kubernetes-version-number> See https://github.com/kubernetes/client-go/releases to see which versions are supported.
Write an application atop of the client-go clients.
Note: client-go defines its own API objects, so if needed, import API definitions from client-go rather than from the main repository. For example, import "k8s.io/client-go/kubernetes" is correct.
The Go client can use the same kubeconfig file
as the kubectl CLI does to locate and authenticate to the API server. See this example:
package main
import (
"context""fmt""k8s.io/apimachinery/pkg/apis/meta/v1""k8s.io/client-go/kubernetes""k8s.io/client-go/tools/clientcmd")
funcmain() {
// uses the current context in kubeconfig
// path-to-kubeconfig -- for example, /root/.kube/config
config, _ := clientcmd.BuildConfigFromFlags("", "<path-to-kubeconfig>")
// creates the clientset
clientset, _ := kubernetes.NewForConfig(config)
// access the API to list pods
pods, _ := clientset.CoreV1().Pods("").List(context.TODO(), v1.ListOptions{})
fmt.Printf("There are %d pods in the cluster\n", len(pods.Items))
}
The Python client can use the same kubeconfig file
as the kubectl CLI does to locate and authenticate to the API server. See this example:
fromkubernetesimport client, config
config.load_kube_config()
v1=client.CoreV1Api()
print("Listing pods with their IPs:")
ret = v1.list_pod_for_all_namespaces(watch=False)
for i in ret.items:
print("%s\t%s\t%s"% (i.status.pod_ip, i.metadata.namespace, i.metadata.name))
The Java client can use the same kubeconfig file
as the kubectl CLI does to locate and authenticate to the API server. See this example:
packageio.kubernetes.client.examples;importio.kubernetes.client.ApiClient;importio.kubernetes.client.ApiException;importio.kubernetes.client.Configuration;importio.kubernetes.client.apis.CoreV1Api;importio.kubernetes.client.models.V1Pod;importio.kubernetes.client.models.V1PodList;importio.kubernetes.client.util.ClientBuilder;importio.kubernetes.client.util.KubeConfig;importjava.io.FileReader;importjava.io.IOException;/**
* A simple example of how to use the Java API from an application outside a kubernetes cluster
*
* <p>Easiest way to run this: mvn exec:java
* -Dexec.mainClass="io.kubernetes.client.examples.KubeConfigFileClientExample"
*
*/publicclassKubeConfigFileClientExample{publicstaticvoidmain(String[] args)throws IOException, ApiException {// file path to your KubeConfig
String kubeConfigPath ="~/.kube/config";// loading the out-of-cluster config, a kubeconfig from file-system
ApiClient client = ClientBuilder.kubeconfig(KubeConfig.loadKubeConfig(new FileReader(kubeConfigPath))).build();// set the global default api-client to the in-cluster one from above
Configuration.setDefaultApiClient(client);// the CoreV1Api loads default api-client from global configuration.
CoreV1Api api =new CoreV1Api();// invokes the CoreV1Api client
V1PodList list = api.listPodForAllNamespaces(null,null,null,null,null,null,null,null,null); System.out.println("Listing all pods: ");for(V1Pod item : list.getItems()){ System.out.println(item.getMetadata().getName());}}}
This page shows how to specify extended resources for a Node.
Extended resources allow cluster administrators to advertise node-level
resources that would otherwise be unknown to Kubernetes.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Choose one of your Nodes to use for this exercise.
Advertise a new extended resource on one of your Nodes
To advertise a new extended resource on a Node, send an HTTP PATCH request to
the Kubernetes API server. For example, suppose one of your Nodes has four dongles
attached. Here's an example of a PATCH request that advertises four dongle resources
for your Node.
Note that Kubernetes does not need to know what a dongle is or what a dongle is for.
The preceding PATCH request tells Kubernetes that your Node has four things that
you call dongles.
Start a proxy, so that you can easily send requests to the Kubernetes API server:
kubectl proxy
In another command window, send the HTTP PATCH request.
Replace <your-node-name> with the name of your Node:
Note: In the preceding request, ~1 is the encoding for the character / in
the patch path. The operation path value in JSON-Patch is interpreted as a
JSON-Pointer. For more details, see
IETF RFC 6901, section 3.
The output shows that the Node has a capacity of 4 dongles:
Extended resources are similar to memory and CPU resources. For example,
just as a Node has a certain amount of memory and CPU to be shared by all components
running on the Node, it can have a certain number of dongles to be shared
by all components running on the Node. And just as application developers
can create Pods that request a certain amount of memory and CPU, they can
create Pods that request a certain number of dongles.
Extended resources are opaque to Kubernetes; Kubernetes does not
know anything about what they are. Kubernetes knows only that a Node
has a certain number of them. Extended resources must be advertised in integer
amounts. For example, a Node can advertise four dongles, but not 4.5 dongles.
Storage example
Suppose a Node has 800 GiB of a special kind of disk storage. You could
create a name for the special storage, say example.com/special-storage.
Then you could advertise it in chunks of a certain size, say 100 GiB. In that case,
your Node would advertise that it has eight resources of type
example.com/special-storage.
Capacity:...example.com/special-storage:8
If you want to allow arbitrary requests for special storage, you
could advertise special storage in chunks of size 1 byte. In that case, you would advertise
800Gi resources of type example.com/special-storage.
Capacity:...example.com/special-storage:800Gi
Then a Container could request any number of bytes of special storage, up to 800Gi.
Clean up
Here is a PATCH request that removes the dongle advertisement from a Node.
This page shows how to enable and configure autoscaling of the DNS service in
your Kubernetes cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
NAME READY UP-TO-DATE AVAILABLE AGE
...
dns-autoscaler 1/1 1 1 ...
...
If you see "dns-autoscaler" in the output, DNS horizontal autoscaling is
already enabled, and you can skip to
Tuning autoscaling parameters.
Get the name of your DNS Deployment
List the DNS deployments in your cluster in the kube-system namespace:
kubectl get deployment -l k8s-app=kube-dns --namespace=kube-system
The output is similar to this:
NAME READY UP-TO-DATE AVAILABLE AGE
...
coredns 2/2 2 2 ...
...
If you don't see a Deployment for DNS services, you can also look for it by name:
kubectl get deployment --namespace=kube-system
and look for a deployment named coredns or kube-dns.
Your scale target is
Deployment/<your-deployment-name>
where <your-deployment-name> is the name of your DNS Deployment. For example, if
the name of your Deployment for DNS is coredns, your scale target is Deployment/coredns.
Note: CoreDNS is the default DNS service for Kubernetes. CoreDNS sets the label
k8s-app=kube-dns so that it can work in clusters that originally used
kube-dns.
Enable DNS horizontal autoscaling
In this section, you create a new Deployment. The Pods in the Deployment run a
container based on the cluster-proportional-autoscaler-amd64 image.
Create a file named dns-horizontal-autoscaler.yaml with this content:
kind:ServiceAccountapiVersion:v1metadata:name:kube-dns-autoscalernamespace:kube-system---kind:ClusterRoleapiVersion:rbac.authorization.k8s.io/v1metadata:name:system:kube-dns-autoscalerrules:- apiGroups:[""]resources:["nodes"]verbs:["list","watch"]- apiGroups:[""]resources:["replicationcontrollers/scale"]verbs:["get","update"]- apiGroups:["apps"]resources:["deployments/scale","replicasets/scale"]verbs:["get","update"]# Remove the configmaps rule once below issue is fixed:# kubernetes-incubator/cluster-proportional-autoscaler#16- apiGroups:[""]resources:["configmaps"]verbs:["get","create"]---kind:ClusterRoleBindingapiVersion:rbac.authorization.k8s.io/v1metadata:name:system:kube-dns-autoscalersubjects:- kind:ServiceAccountname:kube-dns-autoscalernamespace:kube-systemroleRef:kind:ClusterRolename:system:kube-dns-autoscalerapiGroup:rbac.authorization.k8s.io---apiVersion:apps/v1kind:Deploymentmetadata:name:kube-dns-autoscalernamespace:kube-systemlabels:k8s-app:kube-dns-autoscalerkubernetes.io/cluster-service:"true"spec:selector:matchLabels:k8s-app:kube-dns-autoscalertemplate:metadata:labels:k8s-app:kube-dns-autoscalerspec:priorityClassName:system-cluster-criticalsecurityContext:seccompProfile:type:RuntimeDefaultsupplementalGroups:[65534]fsGroup:65534nodeSelector:kubernetes.io/os:linuxcontainers:- name:autoscalerimage:registry.k8s.io/cpa/cluster-proportional-autoscaler:1.8.4resources:requests:cpu:"20m"memory:"10Mi"command:- /cluster-proportional-autoscaler- --namespace=kube-system- --configmap=kube-dns-autoscaler# Should keep target in sync with cluster/addons/dns/kube-dns.yaml.base- --target=<SCALE_TARGET># When cluster is using large nodes(with more cores), "coresPerReplica" should dominate.# If using small nodes, "nodesPerReplica" should dominate.- --default-params={"linear":{"coresPerReplica":256,"nodesPerReplica":16,"preventSinglePointFailure":true,"includeUnschedulableNodes":true}}- --logtostderr=true- --v=2tolerations:- key:"CriticalAddonsOnly"operator:"Exists"serviceAccountName:kube-dns-autoscaler
In the file, replace <SCALE_TARGET> with your scale target.
Go to the directory that contains your configuration file, and enter this
command to create the Deployment:
Modify the fields according to your needs. The "min" field indicates the
minimal number of DNS backends. The actual number of backends is
calculated using this equation:
Note that the values of both coresPerReplica and nodesPerReplica are
floats.
The idea is that when a cluster is using nodes that have many cores,
coresPerReplica dominates. When a cluster is using nodes that have fewer
cores, nodesPerReplica dominates.
After the manifest file is deleted, the Addon Manager will delete the
dns-autoscaler Deployment.
Understanding how DNS horizontal autoscaling works
The cluster-proportional-autoscaler application is deployed separately from
the DNS service.
An autoscaler Pod runs a client that polls the Kubernetes API server for the
number of nodes and cores in the cluster.
A desired replica count is calculated and applied to the DNS backends based on
the current schedulable nodes and cores and the given scaling parameters.
The scaling parameters and data points are provided via a ConfigMap to the
autoscaler, and it refreshes its parameters table every poll interval to be up
to date with the latest desired scaling parameters.
Changes to the scaling parameters are allowed without rebuilding or restarting
the autoscaler Pod.
The autoscaler provides a controller interface to support two control
patterns: linear and ladder.
This page shows how to change the default Storage Class that is used to
provision volumes for PersistentVolumeClaims that have no special requirements.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Depending on the installation method, your Kubernetes cluster may be deployed with
an existing StorageClass that is marked as default. This default StorageClass
is then used to dynamically provision storage for PersistentVolumeClaims
that do not require any specific storage class. See
PersistentVolumeClaim documentation
for details.
The pre-installed default StorageClass may not fit well with your expected workload;
for example, it might provision storage that is too expensive. If this is the case,
you can either change the default StorageClass or disable it completely to avoid
dynamic provisioning of storage.
Deleting the default StorageClass may not work, as it may be re-created
automatically by the addon manager running in your cluster. Please consult the docs for your installation
for details about addon manager and how to disable individual addons.
Changing the default StorageClass
List the StorageClasses in your cluster:
kubectl get storageclass
The output is similar to this:
NAME PROVISIONER AGE
standard (default) kubernetes.io/gce-pd 1d
gold kubernetes.io/gce-pd 1d
The default StorageClass is marked by (default).
Mark the default StorageClass as non-default:
The default StorageClass has an annotation
storageclass.kubernetes.io/is-default-class set to true. Any other value
or absence of the annotation is interpreted as false.
To mark a StorageClass as non-default, you need to change its value to false:
kubectl patch storageclass standard -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"false"}}}'
where standard is the name of your chosen StorageClass.
Mark a StorageClass as default:
Similar to the previous step, you need to add/set the annotation
storageclass.kubernetes.io/is-default-class=true.
Please note that at most one StorageClass can be marked as default. If two
or more of them are marked as default, a PersistentVolumeClaim without storageClassName explicitly specified cannot be created.
Verify that your chosen StorageClass is default:
kubectl get storageclass
The output is similar to this:
NAME PROVISIONER AGE
standard kubernetes.io/gce-pd 1d
gold (default) kubernetes.io/gce-pd 1d
4.2.10 - Change the Reclaim Policy of a PersistentVolume
This page shows how to change the reclaim policy of a Kubernetes
PersistentVolume.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
PersistentVolumes can have various reclaim policies, including "Retain",
"Recycle", and "Delete". For dynamically provisioned PersistentVolumes,
the default reclaim policy is "Delete". This means that a dynamically provisioned
volume is automatically deleted when a user deletes the corresponding
PersistentVolumeClaim. This automatic behavior might be inappropriate if the volume
contains precious data. In that case, it is more appropriate to use the "Retain"
policy. With the "Retain" policy, if a user deletes a PersistentVolumeClaim,
the corresponding PersistentVolume will not be deleted. Instead, it is moved to the
Released phase, where all of its data can be manually recovered.
where <your-pv-name> is the name of your chosen PersistentVolume.
Note:
On Windows, you must double quote any JSONPath template that contains spaces (not single
quote as shown above for bash). This in turn means that you must use a single quote or escaped
double quote around any literals in the template. For example:
In the preceding output, you can see that the volume bound to claim
default/claim3 has reclaim policy Retain. It will not be automatically
deleted when a user deletes claim default/claim3.
Since cloud providers develop and release at a different pace compared to the Kubernetes project, abstracting the provider-specific code to the cloud-controller-manager binary allows cloud vendors to evolve independently from the core Kubernetes code.
The cloud-controller-manager can be linked to any cloud provider that satisfies cloudprovider.Interface. For backwards compatibility, the cloud-controller-manager provided in the core Kubernetes project uses the same cloud libraries as kube-controller-manager. Cloud providers already supported in Kubernetes core are expected to use the in-tree cloud-controller-manager to transition out of Kubernetes core.
Administration
Requirements
Every cloud has their own set of requirements for running their own cloud provider integration, it should not be too different from the requirements when running kube-controller-manager. As a general rule of thumb you'll need:
cloud authentication/authorization: your cloud may require a token or IAM rules to allow access to their APIs
kubernetes authentication/authorization: cloud-controller-manager may need RBAC rules set to speak to the kubernetes apiserver
high availability: like kube-controller-manager, you may want a high available setup for cloud controller manager using leader election (on by default).
Running cloud-controller-manager
Successfully running cloud-controller-manager requires some changes to your cluster configuration.
kube-apiserver and kube-controller-manager MUST NOT specify the --cloud-provider flag. This ensures that it does not run any cloud specific loops that would be run by cloud controller manager. In the future, this flag will be deprecated and removed.
kubelet must run with --cloud-provider=external. This is to ensure that the kubelet is aware that it must be initialized by the cloud controller manager before it is scheduled any work.
Keep in mind that setting up your cluster to use cloud controller manager will change your cluster behaviour in a few ways:
kubelets specifying --cloud-provider=external will add a taint node.cloudprovider.kubernetes.io/uninitialized with an effect NoSchedule during initialization. This marks the node as needing a second initialization from an external controller before it can be scheduled work. Note that in the event that cloud controller manager is not available, new nodes in the cluster will be left unschedulable. The taint is important since the scheduler may require cloud specific information about nodes such as their region or type (high cpu, gpu, high memory, spot instance, etc).
cloud information about nodes in the cluster will no longer be retrieved using local metadata, but instead all API calls to retrieve node information will go through cloud controller manager. This may mean you can restrict access to your cloud API on the kubelets for better security. For larger clusters you may want to consider if cloud controller manager will hit rate limits since it is now responsible for almost all API calls to your cloud from within the cluster.
The cloud controller manager can implement:
Node controller - responsible for updating kubernetes nodes using cloud APIs and deleting kubernetes nodes that were deleted on your cloud.
Service controller - responsible for loadbalancers on your cloud against services of type LoadBalancer.
Route controller - responsible for setting up network routes on your cloud
any other features you would like to implement if you are running an out-of-tree provider.
Examples
If you are using a cloud that is currently supported in Kubernetes core and would like to adopt cloud controller manager, see the cloud controller manager in kubernetes core.
For cloud controller managers not in Kubernetes core, you can find the respective projects in repositories maintained by cloud vendors or by SIGs.
For providers already in Kubernetes core, you can run the in-tree cloud controller manager as a DaemonSet in your cluster, use the following as a guideline:
# This is an example of how to set up cloud-controller-manager as a Daemonset in your cluster.# It assumes that your masters can run pods and has the role node-role.kubernetes.io/master# Note that this Daemonset will not work straight out of the box for your cloud, this is# meant to be a guideline.---apiVersion:v1kind:ServiceAccountmetadata:name:cloud-controller-managernamespace:kube-system---apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRoleBindingmetadata:name:system:cloud-controller-managerroleRef:apiGroup:rbac.authorization.k8s.iokind:ClusterRolename:cluster-adminsubjects:- kind:ServiceAccountname:cloud-controller-managernamespace:kube-system---apiVersion:apps/v1kind:DaemonSetmetadata:labels:k8s-app:cloud-controller-managername:cloud-controller-managernamespace:kube-systemspec:selector:matchLabels:k8s-app:cloud-controller-managertemplate:metadata:labels:k8s-app:cloud-controller-managerspec:serviceAccountName:cloud-controller-managercontainers:- name:cloud-controller-manager# for in-tree providers we use registry.k8s.io/cloud-controller-manager# this can be replaced with any other image for out-of-tree providersimage:registry.k8s.io/cloud-controller-manager:v1.8.0command:- /usr/local/bin/cloud-controller-manager- --cloud-provider=[YOUR_CLOUD_PROVIDER] # Add your own cloud provider here!- --leader-elect=true- --use-service-account-credentials# these flags will vary for every cloud provider- --allocate-node-cidrs=true- --configure-cloud-routes=true- --cluster-cidr=172.17.0.0/16tolerations:# this is required so CCM can bootstrap itself- key:node.cloudprovider.kubernetes.io/uninitializedvalue:"true"effect:NoSchedule# these tolerations are to have the daemonset runnable on control plane nodes# remove them if your control plane nodes should not run pods- key:node-role.kubernetes.io/control-planeoperator:Existseffect:NoSchedule- key:node-role.kubernetes.io/masteroperator:Existseffect:NoSchedule# this is to restrict CCM to only run on master nodes# the node selector may vary depending on your cluster setupnodeSelector:node-role.kubernetes.io/master:""
Limitations
Running cloud controller manager comes with a few possible limitations. Although these limitations are being addressed in upcoming releases, it's important that you are aware of these limitations for production workloads.
Support for Volumes
Cloud controller manager does not implement any of the volume controllers found in kube-controller-manager as the volume integrations also require coordination with kubelets. As we evolve CSI (container storage interface) and add stronger support for flex volume plugins, necessary support will be added to cloud controller manager so that clouds can fully integrate with volumes. Learn more about out-of-tree CSI volume plugins here.
Scalability
The cloud-controller-manager queries your cloud provider's APIs to retrieve information for all nodes. For very large clusters, consider possible bottlenecks such as resource requirements and API rate limiting.
Chicken and Egg
The goal of the cloud controller manager project is to decouple development of cloud features from the core Kubernetes project. Unfortunately, many aspects of the Kubernetes project has assumptions that cloud provider features are tightly integrated into the project. As a result, adopting this new architecture can create several situations where a request is being made for information from a cloud provider, but the cloud controller manager may not be able to return that information without the original request being complete.
A good example of this is the TLS bootstrapping feature in the Kubelet. TLS bootstrapping assumes that the Kubelet has the ability to ask the cloud provider (or a local metadata service) for all its address types (private, public, etc) but cloud controller manager cannot set a node's address types without being initialized in the first place which requires that the kubelet has TLS certificates to communicate with the apiserver.
As this initiative evolves, changes will be made to address these issues in upcoming releases.
This page shows how to configure quotas for API objects, including
PersistentVolumeClaims and Services. A quota restricts the number of
objects, of a particular type, that can be created in a namespace.
You specify quotas in a
ResourceQuota
object.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
View detailed information about the ResourceQuota:
kubectl get resourcequota object-quota-demo --namespace=quota-object-example --output=yaml
The output shows that in the quota-object-example namespace, there can be at most
one PersistentVolumeClaim, at most two Services of type LoadBalancer, and no Services
of type NodePort.
4.2.13 - Control CPU Management Policies on the Node
FEATURE STATE:Kubernetes v1.12 [beta]
Kubernetes keeps many aspects of how pods execute on nodes abstracted
from the user. This is by design. However, some workloads require
stronger guarantees in terms of latency and/or performance in order to operate
acceptably. The kubelet provides methods to enable more complex workload
placement policies while keeping the abstraction free from explicit placement
directives.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
By default, the kubelet uses CFS quota
to enforce pod CPU limits. When the node runs many CPU-bound pods,
the workload can move to different CPU cores depending on
whether the pod is throttled and which CPU cores are available at
scheduling time. Many workloads are not sensitive to this migration and thus
work fine without any intervention.
However, in workloads where CPU cache affinity and scheduling latency
significantly affect workload performance, the kubelet allows alternative CPU
management policies to determine some placement preferences on the node.
Configuration
The CPU Manager policy is set with the --cpu-manager-policy kubelet
flag or the cpuManagerPolicy field in KubeletConfiguration.
There are two supported policies:
static: allows pods with certain resource characteristics to be
granted increased CPU affinity and exclusivity on the node.
The CPU manager periodically writes resource updates through the CRI in
order to reconcile in-memory CPU assignments with cgroupfs. The reconcile
frequency is set through a new Kubelet configuration value
--cpu-manager-reconcile-period. If not specified, it defaults to the same
duration as --node-status-update-frequency.
The behavior of the static policy can be fine-tuned using the --cpu-manager-policy-options flag.
The flag takes a comma-separated list of key=value policy options.
This feature can be disabled completely using the CPUManagerPolicyOptions feature gate.
The policy options are split into two groups: alpha quality (hidden by default) and beta quality
(visible by default). The groups are guarded respectively by the CPUManagerPolicyAlphaOptions
and CPUManagerPolicyBetaOptions feature gates. Diverging from the Kubernetes standard, these
feature gates guard groups of options, because it would have been too cumbersome to add a feature
gate for each individual option.
Changing the CPU Manager Policy
Since the CPU manger policy can only be applied when kubelet spawns new pods, simply changing from
"none" to "static" won't apply to existing pods. So in order to properly change the CPU manager
policy on a node, perform the following steps:
Remove the old CPU manager state file. The path to this file is
/var/lib/kubelet/cpu_manager_state by default. This clears the state maintained by the
CPUManager so that the cpu-sets set up by the new policy won’t conflict with it.
Edit the kubelet configuration to change the CPU manager policy to the desired value.
Start kubelet.
Repeat this process for every node that needs its CPU manager policy changed. Skipping this
process will result in kubelet crashlooping with the following error:
could not restore state from checkpoint: configured policy "static" differs from state checkpoint policy "none", please drain this node and delete the CPU manager checkpoint file "/var/lib/kubelet/cpu_manager_state" before restarting Kubelet
None policy
The none policy explicitly enables the existing default CPU
affinity scheme, providing no affinity beyond what the OS scheduler does
automatically. Limits on CPU usage for
Guaranteed pods and
Burstable pods
are enforced using CFS quota.
Static policy
The static policy allows containers in Guaranteed pods with integer CPU
requests access to exclusive CPUs on the node. This exclusivity is enforced
using the cpuset cgroup controller.
Note: System services such as the container runtime and the kubelet itself can continue to run on these exclusive CPUs. The exclusivity only extends to other pods.
Note: CPU Manager doesn't support offlining and onlining of
CPUs at runtime. Also, if the set of online CPUs changes on the node,
the node must be drained and CPU manager manually reset by deleting the
state file cpu_manager_state in the kubelet root directory.
This policy manages a shared pool of CPUs that initially contains all CPUs in the
node. The amount of exclusively allocatable CPUs is equal to the total
number of CPUs in the node minus any CPU reservations by the kubelet --kube-reserved or
--system-reserved options. From 1.17, the CPU reservation list can be specified
explicitly by kubelet --reserved-cpus option. The explicit CPU list specified by
--reserved-cpus takes precedence over the CPU reservation specified by
--kube-reserved and --system-reserved. CPUs reserved by these options are taken, in
integer quantity, from the initial shared pool in ascending order by physical
core ID. This shared pool is the set of CPUs on which any containers in
BestEffort and Burstable pods run. Containers in Guaranteed pods with fractional
CPU requests also run on CPUs in the shared pool. Only containers that are
both part of a Guaranteed pod and have integer CPU requests are assigned
exclusive CPUs.
Note: The kubelet requires a CPU reservation greater than zero be made
using either --kube-reserved and/or --system-reserved or --reserved-cpus when
the static policy is enabled. This is because zero CPU reservation would allow the shared
pool to become empty.
As Guaranteed pods whose containers fit the requirements for being statically
assigned are scheduled to the node, CPUs are removed from the shared pool and
placed in the cpuset for the container. CFS quota is not used to bound
the CPU usage of these containers as their usage is bound by the scheduling domain
itself. In others words, the number of CPUs in the container cpuset is equal to the integer
CPU limit specified in the pod spec. This static assignment increases CPU
affinity and decreases context switches due to throttling for the CPU-bound
workload.
Consider the containers in the following pod specs:
spec:containers:- name:nginximage:nginx
This pod runs in the BestEffort QoS class because no resource requests or
limits are specified. It runs in the shared pool.
This pod runs in the Burstable QoS class because resource requests do not
equal limits and the cpu quantity is not specified. It runs in the shared
pool.
This pod runs in the Guaranteed QoS class because requests are equal to limits.
And the container's resource limit for the CPU resource is an integer greater than
or equal to one. The nginx container is granted 2 exclusive CPUs.
This pod runs in the Guaranteed QoS class because requests are equal to limits.
But the container's resource limit for the CPU resource is a fraction. It runs in
the shared pool.
This pod runs in the Guaranteed QoS class because only limits are specified
and requests are set equal to limits when not explicitly specified. And the
container's resource limit for the CPU resource is an integer greater than or
equal to one. The nginx container is granted 2 exclusive CPUs.
Static policy options
You can toggle groups of options on and off based upon their maturity level
using the following feature gates:
CPUManagerPolicyBetaOptions default enabled. Disable to hide beta-level options.
CPUManagerPolicyAlphaOptions default disabled. Enable to show alpha-level options.
You will still have to enable each option using the CPUManagerPolicyOptions kubelet option.
The following policy options exist for the static CPUManager policy:
full-pcpus-only (beta, visible by default)
distribute-cpus-across-numa (alpha, hidden by default)
align-by-socket (alpha, hidden by default)
If the full-pcpus-only policy option is specified, the static policy will always allocate full physical cores.
By default, without this option, the static policy allocates CPUs using a topology-aware best-fit allocation.
On SMT enabled systems, the policy can allocate individual virtual cores, which correspond to hardware threads.
This can lead to different containers sharing the same physical cores; this behaviour in turn contributes
to the noisy neighbours problem.
With the option enabled, the pod will be admitted by the kubelet only if the CPU request of all its containers
can be fulfilled by allocating full physical cores.
If the pod does not pass the admission, it will be put in Failed state with the message SMTAlignmentError.
If the distribute-cpus-across-numapolicy option is specified, the static
policy will evenly distribute CPUs across NUMA nodes in cases where more than
one NUMA node is required to satisfy the allocation.
By default, the CPUManager will pack CPUs onto one NUMA node until it is
filled, with any remaining CPUs simply spilling over to the next NUMA node.
This can cause undesired bottlenecks in parallel code relying on barriers (and
similar synchronization primitives), as this type of code tends to run only as
fast as its slowest worker (which is slowed down by the fact that fewer CPUs
are available on at least one NUMA node).
By distributing CPUs evenly across NUMA nodes, application developers can more
easily ensure that no single worker suffers from NUMA effects more than any
other, improving the overall performance of these types of applications.
If the align-by-socket policy option is specified, CPUs will be considered
aligned at the socket boundary when deciding how to allocate CPUs to a
container. By default, the CPUManager aligns CPU allocations at the NUMA
boundary, which could result in performance degradation if CPUs need to be
pulled from more than one NUMA node to satisfy the allocation. Although it
tries to ensure that all CPUs are allocated from the minimum number of NUMA
nodes, there is no guarantee that those NUMA nodes will be on the same socket.
By directing the CPUManager to explicitly align CPUs at the socket boundary
rather than the NUMA boundary, we are able to avoid such issues. Note, this
policy option is not compatible with TopologyManagersingle-numa-node
policy and does not apply to hardware where the number of sockets is greater
than number of NUMA nodes.
The full-pcpus-only option can be enabled by adding full-pcups-only=true to
the CPUManager policy options.
Likewise, the distribute-cpus-across-numa option can be enabled by adding
distribute-cpus-across-numa=true to the CPUManager policy options.
When both are set, they are "additive" in the sense that CPUs will be
distributed across NUMA nodes in chunks of full-pcpus rather than individual
cores.
The align-by-socket policy option can be enabled by adding align-by-socket=true
to the CPUManager policy options. It is also additive to the full-pcpus-only
and distribute-cpus-across-numa policy options.
4.2.14 - Control Topology Management Policies on a node
FEATURE STATE:Kubernetes v1.18 [beta]
An increasing number of systems leverage a combination of CPUs and hardware accelerators to support latency-critical execution and high-throughput parallel computation. These include workloads in fields such as telecommunications, scientific computing, machine learning, financial services and data analytics. Such hybrid systems comprise a high performance environment.
In order to extract the best performance, optimizations related to CPU isolation, memory and device locality are required. However, in Kubernetes, these optimizations are handled by a disjoint set of components.
Topology Manager is a Kubelet component that aims to coordinate the set of components that are responsible for these optimizations.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.18.
To check the version, enter kubectl version.
How Topology Manager Works
Prior to the introduction of Topology Manager, the CPU and Device Manager in Kubernetes make resource allocation decisions independently of each other.
This can result in undesirable allocations on multiple-socketed systems, performance/latency sensitive applications will suffer due to these undesirable allocations.
Undesirable in this case meaning for example, CPUs and devices being allocated from different NUMA Nodes thus, incurring additional latency.
The Topology Manager is a Kubelet component, which acts as a source of truth so that other Kubelet components can make topology aligned resource allocation choices.
The Topology Manager provides an interface for components, called Hint Providers, to send and receive topology information. Topology Manager has a set of node level policies which are explained below.
The Topology manager receives Topology information from the Hint Providers as a bitmask denoting NUMA Nodes available and a preferred allocation indication. The Topology Manager policies perform a set of operations on the hints provided and converge on the hint determined by the policy to give the optimal result, if an undesirable hint is stored the preferred field for the hint will be set to false. In the current policies preferred is the narrowest preferred mask.
The selected hint is stored as part of the Topology Manager. Depending on the policy configured the pod can be accepted or rejected from the node based on the selected hint.
The hint is then stored in the Topology Manager for use by the Hint Providers when making the resource allocation decisions.
Enable the Topology Manager feature
Support for the Topology Manager requires TopologyManagerfeature gate to be enabled. It is enabled by default starting with Kubernetes 1.18.
Topology Manager Scopes and Policies
The Topology Manager currently:
Aligns Pods of all QoS classes.
Aligns the requested resources that Hint Provider provides topology hints for.
If these conditions are met, the Topology Manager will align the requested resources.
In order to customise how this alignment is carried out, the Topology Manager provides two distinct knobs: scope and policy.
The scope defines the granularity at which you would like resource alignment to be performed (e.g. at the pod or container level). And the policy defines the actual strategy used to carry out the alignment (e.g. best-effort, restricted, single-numa-node, etc.).
Details on the various scopes and policies available today can be found below.
Note: To align CPU resources with other requested resources in a Pod Spec, the CPU Manager should be enabled and proper CPU Manager policy should be configured on a Node. See control CPU Management Policies.
Note: To align memory (and hugepages) resources with other requested resources in a Pod Spec, the Memory Manager should be enabled and proper Memory Manager policy should be configured on a Node. Examine Memory Manager documentation.
Topology Manager Scopes
The Topology Manager can deal with the alignment of resources in a couple of distinct scopes:
container (default)
pod
Either option can be selected at a time of the kubelet startup, with --topology-manager-scope flag.
container scope
The container scope is used by default.
Within this scope, the Topology Manager performs a number of sequential resource alignments, i.e., for each container (in a pod) a separate alignment is computed. In other words, there is no notion of grouping the containers to a specific set of NUMA nodes, for this particular scope. In effect, the Topology Manager performs an arbitrary alignment of individual containers to NUMA nodes.
The notion of grouping the containers was endorsed and implemented on purpose in the following scope, for example the pod scope.
pod scope
To select the pod scope, start the kubelet with the command line option --topology-manager-scope=pod.
This scope allows for grouping all containers in a pod to a common set of NUMA nodes. That is, the Topology Manager treats a pod as a whole and attempts to allocate the entire pod (all containers) to either a single NUMA node or a common set of NUMA nodes. The following examples illustrate the alignments produced by the Topology Manager on different occasions:
all containers can be and are allocated to a single NUMA node;
all containers can be and are allocated to a shared set of NUMA nodes.
The total amount of particular resource demanded for the entire pod is calculated according to effective requests/limits formula, and thus, this total value is equal to the maximum of:
the sum of all app container requests,
the maximum of init container requests,
for a resource.
Using the pod scope in tandem with single-numa-node Topology Manager policy is specifically valuable for workloads that are latency sensitive or for high-throughput applications that perform IPC. By combining both options, you are able to place all containers in a pod onto a single NUMA node; hence, the inter-NUMA communication overhead can be eliminated for that pod.
In the case of single-numa-node policy, a pod is accepted only if a suitable set of NUMA nodes is present among possible allocations. Reconsider the example above:
a set containing only a single NUMA node - it leads to pod being admitted,
whereas a set containing more NUMA nodes - it results in pod rejection (because instead of one NUMA node, two or more NUMA nodes are required to satisfy the allocation).
To recap, Topology Manager first computes a set of NUMA nodes and then tests it against Topology Manager policy, which either leads to the rejection or admission of the pod.
Topology Manager Policies
Topology Manager supports four allocation policies. You can set a policy via a Kubelet flag, --topology-manager-policy.
There are four supported policies:
none (default)
best-effort
restricted
single-numa-node
Note: If Topology Manager is configured with the pod scope, the container, which is considered by the policy, is reflecting requirements of the entire pod, and thus each container from the pod will result with the same topology alignment decision.
none policy
This is the default policy and does not perform any topology alignment.
best-effort policy
For each container in a Pod, the kubelet, with best-effort topology
management policy, calls each Hint Provider to discover their resource availability.
Using this information, the Topology Manager stores the
preferred NUMA Node affinity for that container. If the affinity is not preferred,
Topology Manager will store this and admit the pod to the node anyway.
The Hint Providers can then use this information when making the
resource allocation decision.
restricted policy
For each container in a Pod, the kubelet, with restricted topology
management policy, calls each Hint Provider to discover their resource availability.
Using this information, the Topology Manager stores the
preferred NUMA Node affinity for that container. If the affinity is not preferred,
Topology Manager will reject this pod from the node. This will result in a pod in a Terminated state with a pod admission failure.
Once the pod is in a Terminated state, the Kubernetes scheduler will not attempt to reschedule the pod. It is recommended to use a ReplicaSet or Deployment to trigger a redeploy of the pod.
An external control loop could be also implemented to trigger a redeployment of pods that have the Topology Affinity error.
If the pod is admitted, the Hint Providers can then use this information when making the
resource allocation decision.
single-numa-node policy
For each container in a Pod, the kubelet, with single-numa-node topology
management policy, calls each Hint Provider to discover their resource availability.
Using this information, the Topology Manager determines if a single NUMA Node affinity is possible.
If it is, Topology Manager will store this and the Hint Providers can then use this information when making the
resource allocation decision.
If, however, this is not possible then the Topology Manager will reject the pod from the node. This will result in a pod in a Terminated state with a pod admission failure.
Once the pod is in a Terminated state, the Kubernetes scheduler will not attempt to reschedule the pod. It is recommended to use a Deployment with replicas to trigger a redeploy of the Pod.
An external control loop could be also implemented to trigger a redeployment of pods that have the Topology Affinity error.
Pod Interactions with Topology Manager Policies
Consider the containers in the following pod specs:
spec:containers:- name:nginximage:nginx
This pod runs in the BestEffort QoS class because no resource requests or
limits are specified.
This pod runs in the Burstable QoS class because requests are less than limits.
If the selected policy is anything other than none, Topology Manager would consider these Pod specifications. The Topology Manager would consult the Hint Providers to get topology hints. In the case of the static, the CPU Manager policy would return default topology hint, because these Pods do not have explicitly request CPU resources.
This pod runs in the BestEffort QoS class because there are no CPU and memory requests.
The Topology Manager would consider the above pods. The Topology Manager would consult the Hint Providers, which are CPU and Device Manager to get topology hints for the pods.
In the case of the Guaranteed pod with integer CPU request, the static CPU Manager policy would return topology hints relating to the exclusive CPU and the Device Manager would send back hints for the requested device.
In the case of the Guaranteed pod with sharing CPU request, the static CPU Manager policy would return default topology hint as there is no exclusive CPU request and the Device Manager would send back hints for the requested device.
In the above two cases of the Guaranteed pod, the none CPU Manager policy would return default topology hint.
In the case of the BestEffort pod, the static CPU Manager policy would send back the default topology hint as there is no CPU request and the Device Manager would send back the hints for each of the requested devices.
Using this information the Topology Manager calculates the optimal hint for the pod and stores this information, which will be used by the Hint Providers when they are making their resource assignments.
Known Limitations
The maximum number of NUMA nodes that Topology Manager allows is 8. With more than 8 NUMA nodes there will be a state explosion when trying to enumerate the possible NUMA affinities and generating their hints.
The scheduler is not topology-aware, so it is possible to be scheduled on a node and then fail on the node due to the Topology Manager.
4.2.15 - Customizing DNS Service
This page explains how to configure your DNS
Pod(s) and customize the
DNS resolution process in your cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your cluster must be running the CoreDNS add-on.
Migrating to CoreDNS
explains how to use kubeadm to migrate from kube-dns.
Your Kubernetes server must be at or later than version v1.12.
To check the version, enter kubectl version.
Introduction
DNS is a built-in Kubernetes service launched automatically
using the addon managercluster add-on.
As of Kubernetes v1.12, CoreDNS is the recommended DNS Server, replacing kube-dns. If your cluster
originally used kube-dns, you may still have kube-dns deployed rather than CoreDNS.
Note: The CoreDNS Service is named kube-dns in the metadata.name field.
This is so that there is greater interoperability with workloads that relied on the legacy kube-dns Service name to resolve addresses internal to the cluster. Using a Service named kube-dns abstracts away the implementation detail of which DNS provider is running behind that common name.
If you are running CoreDNS as a Deployment, it will typically be exposed as a Kubernetes Service with a static IP address.
The kubelet passes DNS resolver information to each container with the --cluster-dns=<dns-service-ip> flag.
DNS names also need domains. You configure the local domain in the kubelet
with the flag --cluster-domain=<default-local-domain>.
The DNS server supports forward lookups (A and AAAA records), port lookups (SRV records), reverse IP address lookups (PTR records),
and more. For more information, see DNS for Services and Pods.
If a Pod's dnsPolicy is set to default, it inherits the name resolution
configuration from the node that the Pod runs on. The Pod's DNS resolution
should behave the same as the node.
But see Known issues.
If you don't want this, or if you want a different DNS config for pods, you can
use the kubelet's --resolv-conf flag. Set this flag to "" to prevent Pods from
inheriting DNS. Set it to a valid file path to specify a file other than
/etc/resolv.conf for DNS inheritance.
CoreDNS
CoreDNS is a general-purpose authoritative DNS server that can serve as cluster DNS, complying with the dns specifications.
CoreDNS ConfigMap options
CoreDNS is a DNS server that is modular and pluggable, and each plugin adds new functionality to CoreDNS.
This can be configured by maintaining a Corefile, which is the CoreDNS
configuration file. As a cluster administrator, you can modify the
ConfigMap for the CoreDNS Corefile to change how DNS service discovery
behaves for that cluster.
In Kubernetes, CoreDNS is installed with the following default Corefile configuration:
health: Health of CoreDNS is reported to http://localhost:8080/health. In this extended syntax lameduck will make the process unhealthy then wait for 5 seconds before the process is shut down.
ready: An HTTP endpoint on port 8181 will return 200 OK, when all plugins that are able to signal readiness have done so.
kubernetes: CoreDNS will reply to DNS queries based on IP of the services and pods of Kubernetes. You can find more details about that plugin on the CoreDNS website. ttl allows you to set a custom TTL for responses. The default is 5 seconds. The minimum TTL allowed is 0 seconds, and the maximum is capped at 3600 seconds. Setting TTL to 0 will prevent records from being cached.
The pods insecure option is provided for backward compatibility with kube-dns. You can use the pods verified option, which returns an A record only if there exists a pod in same namespace with matching IP. The pods disabled option can be used if you don't use pod records.
prometheus: Metrics of CoreDNS are available at http://localhost:9153/metrics in Prometheus format (also known as OpenMetrics).
forward: Any queries that are not within the cluster domain of Kubernetes will be forwarded to predefined resolvers (/etc/resolv.conf).
loop: Detects simple forwarding loops and halts the CoreDNS process if a loop is found.
reload: Allows automatic reload of a changed Corefile. After you edit the ConfigMap configuration, allow two minutes for your changes to take effect.
loadbalance: This is a round-robin DNS loadbalancer that randomizes the order of A, AAAA, and MX records in the answer.
You can modify the default CoreDNS behavior by modifying the ConfigMap.
Configuration of Stub-domain and upstream nameserver using CoreDNS
CoreDNS has the ability to configure stubdomains and upstream nameservers using the forward plugin.
Example
If a cluster operator has a Consul domain server located at 10.150.0.1, and all Consul names have the suffix .consul.local. To configure it in CoreDNS, the cluster administrator creates the following stanza in the CoreDNS ConfigMap.
To explicitly force all non-cluster DNS lookups to go through a specific nameserver at 172.16.0.1, point the forward to the nameserver instead of /etc/resolv.conf
forward . 172.16.0.1
The final ConfigMap along with the default Corefile configuration looks like:
The kubeadm tool supports automatic translation from the kube-dns ConfigMap
to the equivalent CoreDNS ConfigMap.
Note: While kube-dns accepts an FQDN for stubdomain and nameserver (eg: ns.foo.com), CoreDNS does not support this feature.
During translation, all FQDN nameservers will be omitted from the CoreDNS config.
CoreDNS configuration equivalent to kube-dns
CoreDNS supports the features of kube-dns and more.
A ConfigMap created for kube-dns to support StubDomainsand upstreamNameservers translates to the forward plugin in CoreDNS.
Example
This example ConfigMap for kube-dns specifies stubdomains and upstreamnameservers:
This page provides hints on diagnosing DNS problems.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Note: This example creates a pod in the default namespace. DNS name resolution for
services depends on the namespace of the pod. For more information, review
DNS for Services and Pods.
Use the kubectl get pods command to verify that the DNS pod is running.
kubectl get pods --namespace=kube-system -l k8s-app=kube-dns
NAME READY STATUS RESTARTS AGE
...
coredns-7b96bf9f76-5hsxb 1/1 Running 0 1h
coredns-7b96bf9f76-mvmmt 1/1 Running 0 1h
...
Note: The value for label k8s-app is kube-dns for both CoreDNS and kube-dns deployments.
If you see that no CoreDNS Pod is running or that the Pod has failed/completed,
the DNS add-on may not be deployed by default in your current environment and you
will have to deploy it manually.
Check for errors in the DNS pod
Use the kubectl logs command to see logs for the DNS containers.
See if there are any suspicious or unexpected messages in the logs.
Is DNS service up?
Verify that the DNS service is up by using the kubectl get service command.
kubectl get svc --namespace=kube-system
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
...
kube-dns ClusterIP 10.0.0.10 <none> 53/UDP,53/TCP 1h
...
Note: The service name is kube-dns for both CoreDNS and kube-dns deployments.
If you have created the Service or in the case it should be created by default
but it does not appear, see
debugging Services for
more information.
Are DNS endpoints exposed?
You can verify that DNS endpoints are exposed by using the kubectl get endpoints
command.
kubectl get endpoints kube-dns --namespace=kube-system
NAME ENDPOINTS AGE
kube-dns 10.180.3.17:53,10.180.3.17:53 1h
If you do not see the endpoints, see the endpoints section in the
debugging Services documentation.
For additional Kubernetes DNS examples, see the
cluster-dns examples
in the Kubernetes GitHub repository.
Are DNS queries being received/processed?
You can verify if queries are being received by CoreDNS by adding the log plugin to the CoreDNS configuration (aka Corefile).
The CoreDNS Corefile is held in a ConfigMap named coredns. To edit it, use the command:
kubectl -n kube-system edit configmap coredns
Then add log in the Corefile section per the example below:
After saving the changes, it may take up to minute or two for Kubernetes to propagate these changes to the CoreDNS pods.
Next, make some queries and view the logs per the sections above in this document. If CoreDNS pods are receiving the queries, you should see them in the logs.
Some Linux distributions (e.g. Ubuntu) use a local DNS resolver by default (systemd-resolved).
Systemd-resolved moves and replaces /etc/resolv.conf with a stub file that can cause a fatal forwarding
loop when resolving names in upstream servers. This can be fixed manually by using kubelet's --resolv-conf flag
to point to the correct resolv.conf (With systemd-resolved, this is /run/systemd/resolve/resolv.conf).
kubeadm automatically detects systemd-resolved, and adjusts the kubelet flags accordingly.
Kubernetes installs do not configure the nodes' resolv.conf files to use the
cluster DNS by default, because that process is inherently distribution-specific.
This should probably be implemented eventually.
Linux's libc (a.k.a. glibc) has a limit for the DNS nameserver records to 3 by default. What's more, for the glibc versions which are older than glibc-2.17-222 (the new versions update see this issue), the allowed number of DNS search records has been limited to 6 (see this bug from 2005). Kubernetes needs to consume 1 nameserver record and 3 search records. This means that if a local installation already uses 3 nameservers or uses more than 3 searches while your glibc version is in the affected list, some of those settings will be lost. To work around the DNS nameserver records limit, the node can run dnsmasq, which will provide more nameserver entries. You can also use kubelet's --resolv-conf flag. To fix the DNS search records limit, consider upgrading your linux distribution or upgrading to an unaffected version of glibc.
If you are using Alpine version 3.3 or earlier as your base image, DNS may not
work properly due to a known issue with Alpine.
Kubernetes issue 30215
details more information on this.
This document helps you get started using the Kubernetes NetworkPolicy API to declare network policies that govern how pods communicate with each other.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.8.
To check the version, enter kubectl version.
Make sure you've configured a network provider with network policy support. There are a number of network providers that support NetworkPolicy, including:
Create an nginx deployment and expose it via a service
To see how Kubernetes network policy works, start off by creating an nginx Deployment.
kubectl create deployment nginx --image=nginx
deployment.apps/nginx created
Expose the Deployment through a Service called nginx.
kubectl expose deployment nginx --port=80
service/nginx exposed
The above commands create a Deployment with an nginx Pod and expose the Deployment through a Service named nginx. The nginx Pod and Deployment are found in the default namespace.
kubectl get svc,pod
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/kubernetes 10.100.0.1 <none> 443/TCP 46m
service/nginx 10.100.0.16 <none> 80/TCP 33s
NAME READY STATUS RESTARTS AGE
pod/nginx-701339712-e0qfq 1/1 Running 0 35s
Test the service by accessing it from another Pod
You should be able to access the new nginx service from other Pods. To access the nginx Service from another Pod in the default namespace, start a busybox container:
kubectl run busybox --rm -ti --image=busybox:1.28 -- /bin/sh
In your shell, run the following command:
wget --spider --timeout=1 nginx
Connecting to nginx (10.100.0.16:80)
remote file exists
Limit access to the nginx service
To limit the access to the nginx service so that only Pods with the label access: true can query it, create a NetworkPolicy object as follows:
Note: NetworkPolicy includes a podSelector which selects the grouping of Pods to which the policy applies. You can see this policy selects Pods with the label app=nginx. The label was automatically added to the Pod in the nginx Deployment. An empty podSelector selects all pods in the namespace.
Assign the policy to the service
Use kubectl to create a NetworkPolicy from the above nginx-policy.yaml file:
networkpolicy.networking.k8s.io/access-nginx created
Test access to the service when access label is not defined
When you attempt to access the nginx Service from a Pod without the correct labels, the request times out:
kubectl run busybox --rm -ti --image=busybox:1.28 -- /bin/sh
In your shell, run the command:
wget --spider --timeout=1 nginx
Connecting to nginx (10.100.0.16:80)
wget: download timed out
Define access label and test again
You can create a Pod with the correct labels to see that the request is allowed:
kubectl run busybox --rm -ti --labels="access=true" --image=busybox:1.28 -- /bin/sh
In your shell, run the command:
wget --spider --timeout=1 nginx
Connecting to nginx (10.100.0.16:80)
remote file exists
4.2.18 - Developing Cloud Controller Manager
FEATURE STATE:Kubernetes v1.11 [beta]
The cloud-controller-manager is a Kubernetes control plane component
that embeds cloud-specific control logic. The cloud controller manager lets you link your
cluster into your cloud provider's API, and separates out the components that interact
with that cloud platform from components that only interact with your cluster.
By decoupling the interoperability logic between Kubernetes and the underlying cloud
infrastructure, the cloud-controller-manager component enables cloud providers to release
features at a different pace compared to the main Kubernetes project.
Background
Since cloud providers develop and release at a different pace compared to the Kubernetes project, abstracting the provider-specific code to the cloud-controller-manager binary allows cloud vendors to evolve independently from the core Kubernetes code.
The Kubernetes project provides skeleton cloud-controller-manager code with Go interfaces to allow you (or your cloud provider) to plug in your own implementations. This means that a cloud provider can implement a cloud-controller-manager by importing packages from Kubernetes core; each cloudprovider will register their own code by calling cloudprovider.RegisterCloudProvider to update a global variable of available cloud providers.
Developing
Out of tree
To build an out-of-tree cloud-controller-manager for your cloud:
Use main.go in cloud-controller-manager from Kubernetes core as a template for your main.go. As mentioned above, the only difference should be the cloud package that will be imported.
Many cloud providers publish their controller manager code as open source. If you are creating
a new cloud-controller-manager from scratch, you could take an existing out-of-tree cloud
controller manager as your starting point.
This page shows how to enable or disable an API version from your cluster's
control plane.
Specific API versions can be turned on or off by passing --runtime-config=api/<version> as a
command line argument to the API server. The values for this argument are a comma-separated
list of API versions. Later values override earlier values.
The runtime-config command line argument also supports 2 special keys:
api/all, representing all known APIs
api/legacy, representing only legacy APIs. Legacy APIs are any APIs that have been
explicitly deprecated.
For example, to turning off all API versions except v1, pass --runtime-config=api/all=false,api/v1=true
to the kube-apiserver.
This feature, specifically the alpha topologyKeys field, is deprecated since
Kubernetes v1.21.
Topology Aware Hints,
introduced in Kubernetes v1.21, provide similar functionality.
Service Topology enables a Service to route traffic based upon the Node
topology of the cluster. For example, a service can specify that traffic be
preferentially routed to endpoints that are on the same Node as the client, or
in the same availability zone.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
This page shows how to enable and configure encryption of secret data at rest.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version 1.13.
To check the version, enter kubectl version.
etcd v3.0 or later is required
Configuration and determining whether encryption at rest is already enabled
The kube-apiserver process accepts an argument --encryption-provider-config
that controls how API data is encrypted in etcd.
The configuration is provided as an API named
EncryptionConfiguration.
An example configuration is provided below.
Caution:IMPORTANT: For high-availability configurations (with two or more control plane nodes), the
encryption configuration file must be the same! Otherwise, the kube-apiserver component cannot
decrypt data stored in the etcd.
Understanding the encryption at rest configuration.
Each resources array item is a separate config and contains a complete configuration. The
resources.resources field is an array of Kubernetes resource names (resource or resource.group)
that should be encrypted. The providers array is an ordered list of the possible encryption
providers.
Only one provider type may be specified per entry (identity or aescbc may be provided,
but not both in the same item).
The first provider in the list is used to encrypt resources written into the storage. When reading
resources from storage, each provider that matches the stored data attempts in order to decrypt the
data. If no provider can read the stored data due to a mismatch in format or secret key, an error
is returned which prevents clients from accessing that resource.
For more detailed information about the EncryptionConfiguration struct, please refer to the
encryption configuration API.
Caution: If any resource is not readable via the encryption config (because keys were changed),
the only recourse is to delete that key from the underlying etcd directly. Calls that attempt to
read that resource will fail until it is deleted or a valid decryption key is provided.
Providers:
Providers for Kubernetes encryption at rest
Name
Encryption
Strength
Speed
Key Length
Other Considerations
identity
None
N/A
N/A
N/A
Resources written as-is without encryption. When set as the first provider, the resource will be decrypted as new values are written.
secretbox
XSalsa20 and Poly1305
Strong
Faster
32-byte
A newer standard and may not be considered acceptable in environments that require high levels of review.
aesgcm
AES-GCM with random nonce
Must be rotated every 200k writes
Fastest
16, 24, or 32-byte
Is not recommended for use except when an automated key rotation scheme is implemented.
Not recommended due to CBC's vulnerability to padding oracle attacks.
kms
Uses envelope encryption scheme: Data is encrypted by data encryption keys (DEKs) using AES-CBC with PKCS#7 padding (prior to v1.25), using AES-GCM starting from v1.25, DEKs are encrypted by key encryption keys (KEKs) according to configuration in Key Management Service (KMS)
Strongest
Fast
32-bytes
The recommended choice for using a third party tool for key management. Simplifies key rotation, with a new DEK generated for each encryption, and KEK rotation controlled by the user. Configure the KMS provider
Each provider supports multiple keys - the keys are tried in order for decryption, and if the provider
is the first provider, the first key is used for encryption.
Caution: Storing the raw encryption key in the EncryptionConfig only moderately improves your security
posture, compared to no encryption. Please use kms provider for additional security.
By default, the identity provider is used to protect Secrets in etcd, which provides no
encryption. EncryptionConfiguration was introduced to encrypt Secrets locally, with a locally
managed key.
Encrypting Secrets with a locally managed key protects against an etcd compromise, but it fails to
protect against a host compromise. Since the encryption keys are stored on the host in the
EncryptionConfiguration YAML file, a skilled attacker can access that file and extract the encryption
keys.
Envelope encryption creates dependence on a separate key, not stored in Kubernetes. In this case,
an attacker would need to compromise etcd, the kubeapi-server, and the third-party KMS provider to
retrieve the plaintext values, providing a higher level of security than locally stored encryption keys.
To create a new Secret, perform the following steps:
Generate a 32-byte random key and base64 encode it. If you're on Linux or macOS, run the following command:
head -c 32 /dev/urandom | base64
Place that value in the secret field of the EncryptionConfiguration struct.
Set the --encryption-provider-config flag on the kube-apiserver to point to
the location of the config file.
Restart your API server.
Caution: Your config file contains keys that can decrypt the contents in etcd, so you must properly restrict
permissions on your control-plane nodes so only the user who runs the kube-apiserver can read it.
Verifying that data is encrypted
Data is encrypted when written to etcd. After restarting your kube-apiserver, any newly created or
updated Secret should be encrypted when stored. To check this, you can use the etcdctl command line
program to retrieve the contents of your Secret.
Create a new Secret called secret1 in the default namespace:
The command above reads all Secrets and then updates them to apply server side encryption.
Note: If an error occurs due to a conflicting write, retry the command.
For larger clusters, you may wish to subdivide the secrets by namespace or script an update.
Rotating a decryption key
Changing a Secret without incurring downtime requires a multi-step operation, especially in
the presence of a highly-available deployment where multiple kube-apiserver processes are running.
Generate a new key and add it as the second key entry for the current provider on all servers
Restart all kube-apiserver processes to ensure each server can decrypt using the new key
Make the new key the first entry in the keys array so that it is used for encryption in the config
Restart all kube-apiserver processes to ensure each server now encrypts using the new key
Run kubectl get secrets --all-namespaces -o json | kubectl replace -f - to encrypt all
existing Secrets with the new key
Remove the old decryption key from the config after you have backed up etcd with the new key in use
and updated all Secrets
When running a single kube-apiserver instance, step 2 may be skipped.
Decrypting all data
To disable encryption at rest, place the identity provider as the first entry in the config
and restart all kube-apiserver processes.
4.2.22 - Guaranteed Scheduling For Critical Add-On Pods
Kubernetes core components such as the API server, scheduler, and controller-manager run on a control plane node. However, add-ons must run on a regular cluster node.
Some of these add-ons are critical to a fully functional cluster, such as metrics-server, DNS, and UI.
A cluster may stop working properly if a critical add-on is evicted (either manually or as a side effect of another operation like upgrade)
and becomes pending (for example when the cluster is highly utilized and either there are other pending pods that schedule into the space
vacated by the evicted critical add-on pod or the amount of resources available on the node changed for some other reason).
Note that marking a pod as critical is not meant to prevent evictions entirely; it only prevents the pod from becoming permanently unavailable.
A static pod marked as critical, can't be evicted. However, a non-static pods marked as critical are always rescheduled.
Marking pod as critical
To mark a Pod as critical, set priorityClassName for that Pod to system-cluster-critical or system-node-critical. system-node-critical is the highest available priority, even higher than system-cluster-critical.
4.2.23 - IP Masquerade Agent User Guide
This page shows how to configure and enable the ip-masq-agent.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The ip-masq-agent configures iptables rules to hide a pod's IP address behind the cluster node's IP address. This is typically done when sending traffic to destinations outside the cluster's pod CIDR range.
Key Terms
NAT (Network Address Translation)
Is a method of remapping one IP address to another by modifying either the source and/or destination address information in the IP header. Typically performed by a device doing IP routing.
Masquerading
A form of NAT that is typically used to perform a many to one address translation, where multiple source IP addresses are masked behind a single address, which is typically the device doing the IP routing. In Kubernetes this is the Node's IP address.
CIDR (Classless Inter-Domain Routing)
Based on the variable-length subnet masking, allows specifying arbitrary-length prefixes. CIDR introduced a new method of representation for IP addresses, now commonly known as CIDR notation, in which an address or routing prefix is written with a suffix indicating the number of bits of the prefix, such as 192.168.2.0/24.
Link Local
A link-local address is a network address that is valid only for communications within the network segment or the broadcast domain that the host is connected to. Link-local addresses for IPv4 are defined in the address block 169.254.0.0/16 in CIDR notation.
The ip-masq-agent configures iptables rules to handle masquerading node/pod IP addresses when sending traffic to destinations outside the cluster node's IP and the Cluster IP range. This essentially hides pod IP addresses behind the cluster node's IP address. In some environments, traffic to "external" addresses must come from a known machine address. For example, in Google Cloud, any traffic to the internet must come from a VM's IP. When containers are used, as in Google Kubernetes Engine, the Pod IP will be rejected for egress. To avoid this, we must hide the Pod IP behind the VM's own IP address - generally known as "masquerade". By default, the agent is configured to treat the three private IP ranges specified by RFC 1918 as non-masquerade CIDR. These ranges are 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16. The agent will also treat link-local (169.254.0.0/16) as a non-masquerade CIDR by default. The agent is configured to reload its configuration from the location /etc/config/ip-masq-agent every 60 seconds, which is also configurable.
The agent configuration file must be written in YAML or JSON syntax, and may contain three optional keys:
nonMasqueradeCIDRs: A list of strings in
CIDR notation that specify the non-masquerade ranges.
masqLinkLocal: A Boolean (true/false) which indicates whether to masquerade traffic to the
link local prefix 169.254.0.0/16. False by default.
resyncInterval: A time interval at which the agent attempts to reload config from disk.
For example: '30s', where 's' means seconds, 'ms' means milliseconds.
Traffic to 10.0.0.0/8, 172.16.0.0/12 and 192.168.0.0/16) ranges will NOT be masqueraded. Any other traffic (assumed to be internet) will be masqueraded. An example of a local destination from a pod could be its Node's IP address as well as another node's address or one of the IP addresses in Cluster's IP range. Any other traffic will be masqueraded by default. The below entries show the default set of rules that are applied by the ip-masq-agent:
iptables -t nat -L IP-MASQ-AGENT
RETURN all -- anywhere 169.254.0.0/16 /* ip-masq-agent: cluster-local traffic should not be subject to MASQUERADE */ ADDRTYPE match dst-type !LOCAL
RETURN all -- anywhere 10.0.0.0/8 /* ip-masq-agent: cluster-local traffic should not be subject to MASQUERADE */ ADDRTYPE match dst-type !LOCAL
RETURN all -- anywhere 172.16.0.0/12 /* ip-masq-agent: cluster-local traffic should not be subject to MASQUERADE */ ADDRTYPE match dst-type !LOCAL
RETURN all -- anywhere 192.168.0.0/16 /* ip-masq-agent: cluster-local traffic should not be subject to MASQUERADE */ ADDRTYPE match dst-type !LOCAL
MASQUERADE all -- anywhere anywhere /* ip-masq-agent: outbound traffic should be subject to MASQUERADE (this match must come after cluster-local CIDR matches) */ ADDRTYPE match dst-type !LOCAL
By default, in GCE/Google Kubernetes Engine, if network policy is enabled or
you are using a cluster CIDR not in the 10.0.0.0/8 range, the ip-masq-agent
will run in your cluster. If you are running in another environment,
you can add the ip-masq-agentDaemonSet
to your cluster.
Create an ip-masq-agent
To create an ip-masq-agent, run the following kubectl command:
More information can be found in the ip-masq-agent documentation here
In most cases, the default set of rules should be sufficient; however, if this is not the case for your cluster, you can create and apply a ConfigMap to customize the IP ranges that are affected. For example, to allow only 10.0.0.0/8 to be considered by the ip-masq-agent, you can create the following ConfigMap in a file called "config".
Note:
It is important that the file is called config since, by default, that will be used as the key for lookup by the ip-masq-agent:
nonMasqueradeCIDRs:- 10.0.0.0/8resyncInterval:60s
Run the following command to add the config map to your cluster:
This will update a file located at /etc/config/ip-masq-agent which is periodically checked every resyncInterval and applied to the cluster node.
After the resync interval has expired, you should see the iptables rules reflect your changes:
iptables -t nat -L IP-MASQ-AGENT
Chain IP-MASQ-AGENT (1 references)
target prot opt source destination
RETURN all -- anywhere 169.254.0.0/16 /* ip-masq-agent: cluster-local traffic should not be subject to MASQUERADE */ ADDRTYPE match dst-type !LOCAL
RETURN all -- anywhere 10.0.0.0/8 /* ip-masq-agent: cluster-local
MASQUERADE all -- anywhere anywhere /* ip-masq-agent: outbound traffic should be subject to MASQUERADE (this match must come after cluster-local CIDR matches) */ ADDRTYPE match dst-type !LOCAL
By default, the link local range (169.254.0.0/16) is also handled by the ip-masq agent, which sets up the appropriate iptables rules. To have the ip-masq-agent ignore link local, you can set masqLinkLocal to true in the ConfigMap.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The cluster-admin is operating a cluster on behalf of a user population and the admin wants to control
how much storage a single namespace can consume in order to control cost.
The admin would like to limit:
The number of persistent volume claims in a namespace
The amount of storage each claim can request
The amount of cumulative storage the namespace can have
LimitRange to limit requests for storage
Adding a LimitRange to a namespace enforces storage request sizes to a minimum and maximum. Storage is requested
via PersistentVolumeClaim. The admission controller that enforces limit ranges will reject any PVC that is above or below
the values set by the admin.
In this example, a PVC requesting 10Gi of storage would be rejected because it exceeds the 2Gi max.
Minimum storage requests are used when the underlying storage provider requires certain minimums. For example,
AWS EBS volumes have a 1Gi minimum requirement.
StorageQuota to limit PVC count and cumulative storage capacity
Admins can limit the number of PVCs in a namespace as well as the cumulative capacity of those PVCs. New PVCs that exceed
either maximum value will be rejected.
In this example, a 6th PVC in the namespace would be rejected because it exceeds the maximum count of 5. Alternatively,
a 5Gi maximum quota when combined with the 2Gi max limit above, cannot have 3 PVCs where each has 2Gi. That would be 6Gi requested
for a namespace capped at 5Gi.
A limit range can put a ceiling on how much storage is requested while a resource quota can effectively cap the storage
consumed by a namespace through claim counts and cumulative storage capacity. The allows a cluster-admin to plan their
cluster's storage budget without risk of any one project going over their allotment.
4.2.25 - Migrate Replicated Control Plane To Use Cloud Controller Manager
FEATURE STATE:Kubernetes v1.24 [stable]
The cloud-controller-manager is a Kubernetes control plane component
that embeds cloud-specific control logic. The cloud controller manager lets you link your
cluster into your cloud provider's API, and separates out the components that interact
with that cloud platform from components that only interact with your cluster.
By decoupling the interoperability logic between Kubernetes and the underlying cloud
infrastructure, the cloud-controller-manager component enables cloud providers to release
features at a different pace compared to the main Kubernetes project.
Background
As part of the cloud provider extraction effort, all cloud specific controllers must be moved out of the kube-controller-manager. All existing clusters that run cloud controllers in the kube-controller-manager must migrate to instead run the controllers in a cloud provider specific cloud-controller-manager.
Leader Migration provides a mechanism in which HA clusters can safely migrate "cloud specific" controllers between the kube-controller-manager and the cloud-controller-manager via a shared resource lock between the two components while upgrading the replicated control plane. For a single-node control plane, or if unavailability of controller managers can be tolerated during the upgrade, Leader Migration is not needed and this guide can be ignored.
Leader Migration can be enabled by setting --enable-leader-migration on kube-controller-manager or cloud-controller-manager. Leader Migration only applies during the upgrade and can be safely disabled or left enabled after the upgrade is complete.
This guide walks you through the manual process of upgrading the control plane from kube-controller-manager with built-in cloud provider to running both kube-controller-manager and cloud-controller-manager. If you use a tool to deploy and manage the cluster, please refer to the documentation of the tool and the cloud provider for specific instructions of the migration.
Before you begin
It is assumed that the control plane is running Kubernetes version N and to be upgraded to version N + 1. Although it is possible to migrate within the same version, ideally the migration should be performed as part of an upgrade so that changes of configuration can be aligned to each release. The exact versions of N and N + 1 depend on each cloud provider. For example, if a cloud provider builds a cloud-controller-manager to work with Kubernetes 1.24, then N can be 1.23 and N + 1 can be 1.24.
The control plane nodes should run kube-controller-manager with Leader Election enabled, which is the default. As of version N, an in-tree cloud provider must be set with --cloud-provider flag and cloud-controller-manager should not yet be deployed.
The out-of-tree cloud provider must have built a cloud-controller-manager with Leader Migration implementation. If the cloud provider imports k8s.io/cloud-provider and k8s.io/controller-manager of version v0.21.0 or later, Leader Migration will be available. However, for version before v0.22.0, Leader Migration is alpha and requires feature gate ControllerManagerLeaderMigration to be enabled in cloud-controller-manager.
This guide assumes that kubelet of each control plane node starts kube-controller-manager and cloud-controller-manager as static pods defined by their manifests. If the components run in a different setting, please adjust the steps accordingly.
For authorization, this guide assumes that the cluster uses RBAC. If another authorization mode grants permissions to kube-controller-manager and cloud-controller-manager components, please grant the needed access in a way that matches the mode.
Grant access to Migration Lease
The default permissions of the controller manager allow only accesses to their main Lease. In order for the migration to work, accesses to another Lease are required.
You can grant kube-controller-manager full access to the leases API by modifying the system::leader-locking-kube-controller-manager role. This task guide assumes that the name of the migration lease is cloud-provider-extraction-migration.
Leader Migration optionally takes a configuration file representing the state of controller-to-manager assignment. At this moment, with in-tree cloud provider, kube-controller-manager runs route, service, and cloud-node-lifecycle. The following example configuration shows the assignment.
Leader Migration can be enabled without a configuration. Please see Default Configuration for details.
Alternatively, because the controllers can run under either controller managers, setting component to *
for both sides makes the configuration file consistent between both parties of the migration.
On each control plane node, save the content to /etc/leadermigration.conf, and update the manifest of kube-controller-manager so that the file is mounted inside the container at the same location. Also, update the same manifest to add the following arguments:
--enable-leader-migration to enable Leader Migration on the controller manager
--leader-migration-config=/etc/leadermigration.conf to set configuration file
Restart kube-controller-manager on each node. At this moment, kube-controller-manager has leader migration enabled and is ready for the migration.
Deploy Cloud Controller Manager
In version N + 1, the desired state of controller-to-manager assignment can be represented by a new configuration file, shown as follows. Please note component field of each controllerLeaders changing from kube-controller-manager to cloud-controller-manager. Alternatively, use the wildcard version mentioned above, which has the same effect.
When creating control plane nodes of version N + 1, the content should be deployed to /etc/leadermigration.conf. The manifest of cloud-controller-manager should be updated to mount the configuration file in the same manner as kube-controller-manager of version N. Similarly, add --enable-leader-migration and --leader-migration-config=/etc/leadermigration.conf to the arguments of cloud-controller-manager.
Create a new control plane node of version N + 1 with the updated cloud-controller-manager manifest, and with the --cloud-provider flag set to external for kube-controller-manager. kube-controller-manager of version N + 1 MUST NOT have Leader Migration enabled because, with an external cloud provider, it does not run the migrated controllers anymore, and thus it is not involved in the migration.
The control plane now contains nodes of both version N and N + 1. The nodes of version N run kube-controller-manager only, and these of version N + 1 run both kube-controller-manager and cloud-controller-manager. The migrated controllers, as specified in the configuration, are running under either kube-controller-manager of version N or cloud-controller-manager of version N + 1 depending on which controller manager holds the migration lease. No controller will ever be running under both controller managers at any time.
In a rolling manner, create a new control plane node of version N + 1 and bring down one of version N + 1 until the control plane contains only nodes of version N + 1.
If a rollback from version N + 1 to N is required, add nodes of version N with Leader Migration enabled for kube-controller-manager back to the control plane, replacing one of version N + 1 each time until there are only nodes of version N.
(Optional) Disable Leader Migration
Now that the control plane has been upgraded to run both kube-controller-manager and cloud-controller-manager of version N + 1, Leader Migration has finished its job and can be safely disabled to save one Lease resource. It is safe to re-enable Leader Migration for the rollback in the future.
In a rolling manager, update manifest of cloud-controller-manager to unset both --enable-leader-migration and --leader-migration-config= flag, also remove the mount of /etc/leadermigration.conf, and finally remove /etc/leadermigration.conf. To re-enable Leader Migration, recreate the configuration file and add its mount and the flags that enable Leader Migration back to cloud-controller-manager.
Default Configuration
Starting Kubernetes 1.22, Leader Migration provides a default configuration suitable for the default controller-to-manager assignment.
The default configuration can be enabled by setting --enable-leader-migration but without --leader-migration-config=.
For kube-controller-manager and cloud-controller-manager, if there are no flags that enable any in-tree cloud provider or change ownership of controllers, the default configuration can be used to avoid manual creation of the configuration file.
Special case: migrating the Node IPAM controller
If your cloud provider provides an implementation of Node IPAM controller, you should switch to the implementation in cloud-controller-manager. Disable Node IPAM controller in kube-controller-manager of version N + 1 by adding --controllers=*,-nodeipam to its flags. Then add nodeipam to the list of migrated controllers.
# wildcard version, with nodeipamkind:LeaderMigrationConfigurationapiVersion:controllermanager.config.k8s.io/v1leaderName:cloud-provider-extraction-migrationcontrollerLeaders:- name:routecomponent:*- name:servicecomponent:*- name:cloud-node-lifecyclecomponent:*- name:nodeipam- component:*
A mechanism to attach authorization and policy to a subsection of the cluster.
Use of multiple namespaces is optional.
This example demonstrates how to use Kubernetes namespaces to subdivide your cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
By default, a Kubernetes cluster will instantiate a default namespace when provisioning the cluster to hold the default set of Pods,
Services, and Deployments used by the cluster.
Assuming you have a fresh cluster, you can inspect the available namespaces by doing the following:
kubectl get namespaces
NAME STATUS AGE
default Active 13m
Create new namespaces
For this exercise, we will create two additional Kubernetes namespaces to hold our content.
Let's imagine a scenario where an organization is using a shared Kubernetes cluster for development and production use cases.
The development team would like to maintain a space in the cluster where they can get a view on the list of Pods, Services, and Deployments
they use to build and run their application. In this space, Kubernetes resources come and go, and the restrictions on who can or cannot modify resources
are relaxed to enable agile development.
The operations team would like to maintain a space in the cluster where they can enforce strict procedures on who can or cannot manipulate the set of
Pods, Services, and Deployments that run the production site.
One pattern this organization could follow is to partition the Kubernetes cluster into two namespaces: development and production.
Let's create two new namespaces to hold our work.
Use the file namespace-dev.json which describes a development namespace:
The next step is to define a context for the kubectl client to work in each namespace. The value of "cluster" and "user" fields are copied from the current context.
By default, the above commands adds two contexts that are saved into file
.kube/config. You can now view the contexts and alternate against the two
new request contexts depending on which namespace you wish to work against.
We have created a deployment whose replica size is 2 that is running the pod called snowflake with a basic container that serves the hostname.
kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
snowflake 2/2 2 2 2m
kubectl get pods -l app=snowflake
NAME READY STATUS RESTARTS AGE
snowflake-3968820950-9dgr8 1/1 Running 0 2m
snowflake-3968820950-vgc4n 1/1 Running 0 2m
And this is great, developers are able to do what they want, and they do not have to worry about affecting content in the production namespace.
Let's switch to the production namespace and show how resources in one namespace are hidden from the other.
kubectl config use-context prod
The production namespace should be empty, and the following commands should return nothing.
kubectl get deployment
kubectl get pods
Production likes to run cattle, so let's create some cattle pods.
kubectl create deployment cattle --image=registry.k8s.io/serve_hostname --replicas=5kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
cattle 5/5 5 5 10s
kubectl get pods -l app=cattle
NAME READY STATUS RESTARTS AGE
cattle-2263376956-41xy6 1/1 Running 0 34s
cattle-2263376956-kw466 1/1 Running 0 34s
cattle-2263376956-n4v97 1/1 Running 0 34s
cattle-2263376956-p5p3i 1/1 Running 0 34s
cattle-2263376956-sxpth 1/1 Running 0 34s
At this point, it should be clear that the resources users create in one namespace are hidden from the other namespace.
As the policy support in Kubernetes evolves, we will extend this scenario to show how you can provide different
authorization rules for each namespace.
4.2.27 - Operating etcd clusters for Kubernetes
etcd is a consistent and highly-available key value store used as Kubernetes' backing store for all cluster data.
If your Kubernetes cluster uses etcd as its backing store, make sure you have a
back up plan
for those data.
You can find in-depth information about etcd in the official documentation.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
etcd is a leader-based distributed system. Ensure that the leader
periodically send heartbeats on time to all followers to keep the cluster
stable.
Ensure that no resource starvation occurs.
Performance and stability of the cluster is sensitive to network and disk
I/O. Any resource starvation can lead to heartbeat timeout, causing instability
of the cluster. An unstable etcd indicates that no leader is elected. Under
such circumstances, a cluster cannot make any changes to its current state,
which implies no new pods can be scheduled.
Keeping etcd clusters stable is critical to the stability of Kubernetes
clusters. Therefore, run etcd clusters on dedicated machines or isolated
environments for guaranteed resource requirements.
The minimum recommended version of etcd to run in production is 3.2.10+.
Resource requirements
Operating etcd with limited resources is suitable only for testing purposes.
For deploying in production, advanced hardware configuration is required.
Before deploying etcd in production, see
resource requirement reference.
Starting etcd clusters
This section covers starting a single-node and multi-node etcd cluster.
Single-node etcd cluster
Use a single-node etcd cluster only for testing purpose.
Start the Kubernetes API server with the flag
--etcd-servers=$PRIVATE_IP:2379.
Make sure PRIVATE_IP is set to your etcd client IP.
Multi-node etcd cluster
For durability and high availability, run etcd as a multi-node cluster in
production and back it up periodically. A five-member cluster is recommended
in production. For more information, see
FAQ documentation.
Configure an etcd cluster either by static member information or by dynamic
discovery. For more information on clustering, see
etcd clustering documentation.
For an example, consider a five-member etcd cluster running with the following
client URLs: http://$IP1:2379, http://$IP2:2379, http://$IP3:2379,
http://$IP4:2379, and http://$IP5:2379. To start a Kubernetes API server:
Start the Kubernetes API servers with the flag
--etcd-servers=$IP1:2379,$IP2:2379,$IP3:2379,$IP4:2379,$IP5:2379.
Make sure the IP<n> variables are set to your client IP addresses.
Multi-node etcd cluster with load balancer
To run a load balancing etcd cluster:
Set up an etcd cluster.
Configure a load balancer in front of the etcd cluster.
For example, let the address of the load balancer be $LB.
Start Kubernetes API Servers with the flag --etcd-servers=$LB:2379.
Securing etcd clusters
Access to etcd is equivalent to root permission in the cluster so ideally only
the API server should have access to it. Considering the sensitivity of the
data, it is recommended to grant permission to only those nodes that require
access to etcd clusters.
To secure etcd, either set up firewall rules or use the security features
provided by etcd. etcd security features depend on x509 Public Key
Infrastructure (PKI). To begin, establish secure communication channels by
generating a key and certificate pair. For example, use key pairs peer.key
and peer.cert for securing communication between etcd members, and
client.key and client.cert for securing communication between etcd and its
clients. See the example scripts
provided by the etcd project to generate key pairs and CA files for client
authentication.
Securing communication
To configure etcd with secure peer communication, specify flags
--peer-key-file=peer.key and --peer-cert-file=peer.cert, and use HTTPS as
the URL schema.
Similarly, to configure etcd with secure client communication, specify flags
--key-file=k8sclient.key and --cert-file=k8sclient.cert, and use HTTPS as
the URL schema. Here is an example on a client command that uses secure
communication:
ETCDCTL_API=3 etcdctl --endpoints 10.2.0.9:2379 \
--cert=/etc/kubernetes/pki/etcd/server.crt \
--key=/etc/kubernetes/pki/etcd/server.key \
--cacert=/etc/kubernetes/pki/etcd/ca.crt \
member list
Limiting access of etcd clusters
After configuring secure communication, restrict the access of etcd cluster to
only the Kubernetes API servers. Use TLS authentication to do so.
For example, consider key pairs k8sclient.key and k8sclient.cert that are
trusted by the CA etcd.ca. When etcd is configured with --client-cert-auth
along with TLS, it verifies the certificates from clients by using system CAs
or the CA passed in by --trusted-ca-file flag. Specifying flags
--client-cert-auth=true and --trusted-ca-file=etcd.ca will restrict the
access to clients with the certificate k8sclient.cert.
Once etcd is configured correctly, only clients with valid certificates can
access it. To give Kubernetes API servers the access, configure them with the
flags --etcd-certfile=k8sclient.cert, --etcd-keyfile=k8sclient.key and
--etcd-cafile=ca.cert.
Note: etcd authentication is not currently supported by Kubernetes. For more
information, see the related issue
Support Basic Auth for Etcd v2.
Replacing a failed etcd member
etcd cluster achieves high availability by tolerating minor member failures.
However, to improve the overall health of the cluster, replace failed members
immediately. When multiple members fail, replace them one by one. Replacing a
failed member involves two steps: removing the failed member and adding a new
member.
Though etcd keeps unique member IDs internally, it is recommended to use a
unique name for each member to avoid human errors. For example, consider a
three-member etcd cluster. Let the URLs be, member1=http://10.0.0.1,
member2=http://10.0.0.2, and member3=http://10.0.0.3. When member1 fails,
replace it with member4=http://10.0.0.4.
Get the member ID of the failed member1:
etcdctl --endpoints=http://10.0.0.2,http://10.0.0.3 member list
If each Kubernetes API server is configured to communicate with all etcd
members, remove the failed member from the --etcd-servers flag, then
restart each Kubernetes API server.
If each Kubernetes API server communicates with a single etcd member,
then stop the Kubernetes API server that communicates with the failed
etcd.
Stop the etcd server on the broken node. It is possible that other
clients besides the Kubernetes API server is causing traffic to etcd
and it is desirable to stop all traffic to prevent writes to the data
dir.
Remove the failed member:
etcdctl member remove 8211f1d0f64f3269
The following message is displayed:
Removed member 8211f1d0f64f3269 from cluster
Add the new member:
etcdctl member add member4 --peer-urls=http://10.0.0.4:2380
The following message is displayed:
Member 2be1eb8f84b7f63e added to cluster ef37ad9dc622a7c4
Start the newly added member on a machine with the IP 10.0.0.4:
If each Kubernetes API server is configured to communicate with all etcd
members, add the newly added member to the --etcd-servers flag, then
restart each Kubernetes API server.
If each Kubernetes API server communicates with a single etcd member,
start the Kubernetes API server that was stopped in step 2. Then
configure Kubernetes API server clients to again route requests to the
Kubernetes API server that was stopped. This can often be done by
configuring a load balancer.
All Kubernetes objects are stored on etcd. Periodically backing up the etcd
cluster data is important to recover Kubernetes clusters under disaster
scenarios, such as losing all control plane nodes. The snapshot file contains
all the Kubernetes states and critical information. In order to keep the
sensitive Kubernetes data safe, encrypt the snapshot files.
Backing up an etcd cluster can be accomplished in two ways: etcd built-in
snapshot and volume snapshot.
Built-in snapshot
etcd supports built-in snapshot. A snapshot may either be taken from a live
member with the etcdctl snapshot save command or by copying the
member/snap/db file from an etcd
data directory
that is not currently used by an etcd process. Taking the snapshot will
not affect the performance of the member.
Below is an example for taking a snapshot of the keyspace served by
$ENDPOINT to the file snapshotdb:
ETCDCTL_API=3 etcdctl --endpoints $ENDPOINT snapshot save snapshotdb
Verify the snapshot:
ETCDCTL_API=3 etcdctl --write-out=table snapshot status snapshotdb
If etcd is running on a storage volume that supports backup, such as Amazon
Elastic Block Store, back up etcd data by taking a snapshot of the storage
volume.
Snapshot using etcdctl options
We can also take the snapshot using various options given by etcdctl. For example
ETCDCTL_API=3 etcdctl -h
will list various options available from etcdctl. For example, you can take a snapshot by specifying
the endpoint, certificates etc as shown below:
where trusted-ca-file, cert-file and key-file can be obtained from the description of the etcd Pod.
Scaling up etcd clusters
Scaling up etcd clusters increases availability by trading off performance.
Scaling does not increase cluster performance nor capability. A general rule
is not to scale up or down etcd clusters. Do not configure any auto scaling
groups for etcd clusters. It is highly recommended to always run a static
five-member etcd cluster for production Kubernetes clusters at any officially
supported scale.
A reasonable scaling is to upgrade a three-member cluster to a five-member
one, when more reliability is desired. See
etcd reconfiguration documentation
for information on how to add members into an existing cluster.
Restoring an etcd cluster
etcd supports restoring from snapshots that are taken from an etcd process of
the major.minor version. Restoring a version from a
different patch version of etcd also is supported. A restore operation is
employed to recover the data of a failed cluster.
Before starting the restore operation, a snapshot file must be present. It can
either be a snapshot file from a previous backup operation, or from a remaining
data directory.
Here is an example:
If the access URLs of the restored cluster is changed from the previous
cluster, the Kubernetes API server must be reconfigured accordingly. In this
case, restart Kubernetes API servers with the flag
--etcd-servers=$NEW_ETCD_CLUSTER instead of the flag
--etcd-servers=$OLD_ETCD_CLUSTER. Replace $NEW_ETCD_CLUSTER and
$OLD_ETCD_CLUSTER with the respective IP addresses. If a load balancer is
used in front of an etcd cluster, you might need to update the load balancer
instead.
If the majority of etcd members have permanently failed, the etcd cluster is
considered failed. In this scenario, Kubernetes cannot make any changes to its
current state. Although the scheduled pods might continue to run, no new pods
can be scheduled. In such cases, recover the etcd cluster and potentially
reconfigure Kubernetes API servers to fix the issue.
Note:
If any API servers are running in your cluster, you should not attempt to
restore instances of etcd. Instead, follow these steps to restore etcd:
stop all API server instances
restore state in all etcd instances
restart all API server instances
We also recommend restarting any components (e.g. kube-scheduler,
kube-controller-manager, kubelet) to ensure that they don't rely on some
stale data. Note that in practice, the restore takes a bit of time. During the
restoration, critical components will lose leader lock and restart themselves.
Upgrading etcd clusters
For more details on etcd upgrade, please refer to the etcd upgrades documentation.
Note: Before you start an upgrade, please back up your etcd cluster first.
4.2.28 - Reconfigure a Node's Kubelet in a Live Cluster
FEATURE STATE:Kubernetes v1.22 [deprecated]
Caution: The Dynamic Kubelet Configuration
feature is deprecated in 1.22 and removed in 1.24.
Please switch to alternative means distributing configuration to the Nodes of your cluster.
Migrating from using Dynamic Kubelet Configuration
There is no recommended replacement for this feature that works generically
across various Kubernetes distributions. If you are using managed Kubernetes
version, please consult with the vendor hosting Kubernetes for the best
practices for customizing your Kubernetes. If you are using kubeadm, refer to
Configuring each kubelet in your cluster using kubeadm.
In order to migrate off the Dynamic Kubelet Configuration feature, the
alternative mechanism should be used to distribute kubelet configuration files.
In order to apply configuration, config file must be updated and kubelet restarted.
See the Set Kubelet parameters via a config file
for information.
Please note, the DynamicKubeletConfig feature gate cannot be set on a kubelet
starting v1.24 as it has no effect. However, the feature gate is not removed
from the API server or the controller manager before v1.26. This is designed for
the control plane to support nodes with older versions of kubelets and for
satisfying the Kubernetes version skew policy.
4.2.29 - Reserve Compute Resources for System Daemons
Kubernetes nodes can be scheduled to Capacity. Pods can consume all the
available capacity on a node by default. This is an issue because nodes
typically run quite a few system daemons that power the OS and Kubernetes
itself. Unless resources are set aside for these system daemons, pods and system
daemons compete for resources and lead to resource starvation issues on the
node.
The kubelet exposes a feature named 'Node Allocatable' that helps to reserve
compute resources for system daemons. Kubernetes recommends cluster
administrators to configure 'Node Allocatable' based on their workload density
on each node.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version 1.8.
To check the version, enter kubectl version.
Your Kubernetes server must be at or later than version 1.17 to use
the kubelet command line option --reserved-cpus to set an
explicitly reserved CPU list.
Node Allocatable
'Allocatable' on a Kubernetes node is defined as the amount of compute resources
that are available for pods. The scheduler does not over-subscribe
'Allocatable'. 'CPU', 'memory' and 'ephemeral-storage' are supported as of now.
Node Allocatable is exposed as part of v1.Node object in the API and as part
of kubectl describe node in the CLI.
Resources can be reserved for two categories of system daemons in the kubelet.
Enabling QoS and Pod level cgroups
To properly enforce node allocatable constraints on the node, you must
enable the new cgroup hierarchy via the --cgroups-per-qos flag. This flag is
enabled by default. When enabled, the kubelet will parent all end-user pods
under a cgroup hierarchy managed by the kubelet.
Configuring a cgroup driver
The kubelet supports manipulation of the cgroup hierarchy on
the host using a cgroup driver. The driver is configured via the
--cgroup-driver flag.
The supported values are the following:
cgroupfs is the default driver that performs direct manipulation of the
cgroup filesystem on the host in order to manage cgroup sandboxes.
systemd is an alternative driver that manages cgroup sandboxes using
transient slices for resources that are supported by that init system.
Depending on the configuration of the associated container runtime,
operators may have to choose a particular cgroup driver to ensure
proper system behavior. For example, if operators use the systemd
cgroup driver provided by the containerd runtime, the kubelet must
be configured to use the systemd cgroup driver.
kube-reserved is meant to capture resource reservation for kubernetes system
daemons like the kubelet, container runtime, node problem detector, etc.
It is not meant to reserve resources for system daemons that are run as pods.
kube-reserved is typically a function of pod density on the nodes.
In addition to cpu, memory, and ephemeral-storage, pid may be
specified to reserve the specified number of process IDs for
kubernetes system daemons.
To optionally enforce kube-reserved on kubernetes system daemons, specify the parent
control group for kube daemons as the value for --kube-reserved-cgroup kubelet
flag.
It is recommended that the kubernetes system daemons are placed under a top
level control group (runtime.slice on systemd machines for example). Each
system daemon should ideally run within its own child control group. Refer to
the design proposal
for more details on recommended control group hierarchy.
Note that Kubelet does not create --kube-reserved-cgroup if it doesn't
exist. Kubelet will fail if an invalid cgroup is specified. With systemd
cgroup driver, you should follow a specific pattern for the name of the cgroup you
define: the name should be the value you set for --kube-reserved-cgroup,
with .slice appended.
system-reserved is meant to capture resource reservation for OS system daemons
like sshd, udev, etc. system-reserved should reserve memory for the
kernel too since kernel memory is not accounted to pods in Kubernetes at this time.
Reserving resources for user login sessions is also recommended (user.slice in
systemd world).
In addition to cpu, memory, and ephemeral-storage, pid may be
specified to reserve the specified number of process IDs for OS system
daemons.
To optionally enforce system-reserved on system daemons, specify the parent
control group for OS system daemons as the value for --system-reserved-cgroup
kubelet flag.
It is recommended that the OS system daemons are placed under a top level
control group (system.slice on systemd machines for example).
Note that kubeletdoes not create --system-reserved-cgroup if it doesn't
exist. kubelet will fail if an invalid cgroup is specified. With systemd
cgroup driver, you should follow a specific pattern for the name of the cgroup you
define: the name should be the value you set for --system-reserved-cgroup,
with .slice appended.
Explicitly Reserved CPU List
FEATURE STATE:Kubernetes v1.17 [stable]
Kubelet Flag: --reserved-cpus=0-3
reserved-cpus is meant to define an explicit CPU set for OS system daemons and
kubernetes system daemons. reserved-cpus is for systems that do not intend to
define separate top level cgroups for OS system daemons and kubernetes system daemons
with regard to cpuset resource.
If the Kubelet does not have --system-reserved-cgroup and --kube-reserved-cgroup,
the explicit cpuset provided by reserved-cpus will take precedence over the CPUs
defined by --kube-reserved and --system-reserved options.
This option is specifically designed for Telco/NFV use cases where uncontrolled
interrupts/timers may impact the workload performance. you can use this option
to define the explicit cpuset for the system/kubernetes daemons as well as the
interrupts/timers, so the rest CPUs on the system can be used exclusively for
workloads, with less impact from uncontrolled interrupts/timers. To move the
system daemon, kubernetes daemons and interrupts/timers to the explicit cpuset
defined by this option, other mechanism outside Kubernetes should be used.
For example: in Centos, you can do this using the tuned toolset.
Memory pressure at the node level leads to System OOMs which affects the entire
node and all pods running on it. Nodes can go offline temporarily until memory
has been reclaimed. To avoid (or reduce the probability of) system OOMs kubelet
provides out of resource
management. Evictions are
supported for memory and ephemeral-storage only. By reserving some memory via
--eviction-hard flag, the kubelet attempts to evict pods whenever memory
availability on the node drops below the reserved value. Hypothetically, if
system daemons did not exist on a node, pods cannot use more than capacity - eviction-hard. For this reason, resources reserved for evictions are not
available for pods.
The scheduler treats 'Allocatable' as the available capacity for pods.
kubelet enforce 'Allocatable' across pods by default. Enforcement is performed
by evicting pods whenever the overall usage across all pods exceeds
'Allocatable'. More details on eviction policy can be found
on the node pressure eviction
page. This enforcement is controlled by
specifying pods value to the kubelet flag --enforce-node-allocatable.
Optionally, kubelet can be made to enforce kube-reserved and
system-reserved by specifying kube-reserved & system-reserved values in
the same flag. Note that to enforce kube-reserved or system-reserved,
--kube-reserved-cgroup or --system-reserved-cgroup needs to be specified
respectively.
General Guidelines
System daemons are expected to be treated similar to
Guaranteed pods.
System daemons can burst within their bounding control groups and this behavior needs
to be managed as part of kubernetes deployments. For example, kubelet should
have its own control group and share kube-reserved resources with the
container runtime. However, Kubelet cannot burst and use up all available Node
resources if kube-reserved is enforced.
Be extra careful while enforcing system-reserved reservation since it can lead
to critical system services being CPU starved, OOM killed, or unable
to fork on the node. The
recommendation is to enforce system-reserved only if a user has profiled their
nodes exhaustively to come up with precise estimates and is confident in their
ability to recover if any process in that group is oom-killed.
To begin with enforce 'Allocatable' on pods.
Once adequate monitoring and alerting is in place to track kube system
daemons, attempt to enforce kube-reserved based on usage heuristics.
If absolutely necessary, enforce system-reserved over time.
The resource requirements of kube system daemons may grow over time as more and
more features are added. Over time, kubernetes project will attempt to bring
down utilization of node system daemons, but that is not a priority as of now.
So expect a drop in Allocatable capacity in future releases.
Example Scenario
Here is an example to illustrate Node Allocatable computation:
Node has 32Gi of memory, 16 CPUs and 100Gi of Storage
--kube-reserved is set to cpu=1,memory=2Gi,ephemeral-storage=1Gi
--system-reserved is set to cpu=500m,memory=1Gi,ephemeral-storage=1Gi
--eviction-hard is set to memory.available<500Mi,nodefs.available<10%
Under this scenario, 'Allocatable' will be 14.5 CPUs, 28.5Gi of memory and
88Gi of local storage.
Scheduler ensures that the total memory requests across all pods on this node does
not exceed 28.5Gi and storage doesn't exceed 88Gi.
Kubelet evicts pods whenever the overall memory usage across pods exceeds 28.5Gi,
or if overall disk usage exceeds 88Gi If all processes on the node consume as
much CPU as they can, pods together cannot consume more than 14.5 CPUs.
If kube-reserved and/or system-reserved is not enforced and system daemons
exceed their reservation, kubelet evicts pods whenever the overall node memory
usage is higher than 31.5Gi or storage is greater than 90Gi.
4.2.30 - Running Kubernetes Node Components as a Non-root User
FEATURE STATE:Kubernetes v1.22 [alpha]
This document describes how to run Kubernetes Node components such as kubelet, CRI, OCI, and CNI
without root privileges, by using a user namespace.
This technique is also known as rootless mode.
Note:
This document describes how to run Kubernetes Node components (and hence pods) as a non-root user.
If you are just looking for how to run a pod as a non-root user, see SecurityContext.
Before you begin
Your Kubernetes server must be at or later than version 1.22.
To check the version, enter kubectl version.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
sysbox
Sysbox is an open-source container runtime
(similar to "runc") that supports running system-level workloads such as Docker
and Kubernetes inside unprivileged containers isolated with the Linux user
namespace.
Sysbox supports running Kubernetes inside unprivileged containers without
requiring Cgroup v2 and without the KubeletInUserNamespace feature gate. It
does this by exposing specially crafted /proc and /sys filesystems inside
the container plus several other advanced OS virtualization techniques.
Running Rootless Kubernetes directly on a host
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
If you are trying to run Kubernetes in a user-namespaced container such as
Rootless Docker/Podman or LXC/LXD, you are all set, and you can go to the next subsection.
Otherwise you have to create a user namespace by yourself, by calling unshare(2) with CLONE_NEWUSER.
A user namespace can be also unshared by using command line tools such as:
After unsharing the user namespace, you will also have to unshare other namespaces such as mount namespace.
You do not need to call chroot() nor pivot_root() after unsharing the mount namespace,
however, you have to mount writable filesystems on several directories in the namespace.
At least, the following directories need to be writable in the namespace (not outside the namespace):
/etc
/run
/var/logs
/var/lib/kubelet
/var/lib/cni
/var/lib/containerd (for containerd)
/var/lib/containers (for CRI-O)
Creating a delegated cgroup tree
In addition to the user namespace, you also need to have a writable cgroup tree with cgroup v2.
Note: Kubernetes support for running Node components in user namespaces requires cgroup v2.
Cgroup v1 is not supported.
If you are trying to run Kubernetes in Rootless Docker/Podman or LXC/LXD on a systemd-based host, you are all set.
Otherwise you have to create a systemd unit with Delegate=yes property to delegate a cgroup tree with writable permission.
On your node, systemd must already be configured to allow delegation; for more details, see
cgroup v2 in the Rootless
Containers documentation.
Configuring network
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
The network namespace of the Node components has to have a non-loopback interface, which can be for example configured with
slirp4netns,
VPNKit, or
lxc-user-nic(1).
The network namespaces of the Pods can be configured with regular CNI plugins.
For multi-node networking, Flannel (VXLAN, 8472/UDP) is known to work.
Ports such as the kubelet port (10250/TCP) and NodePort service ports have to be exposed from the Node network namespace to
the host with an external port forwarder, such as RootlessKit, slirp4netns, or
socat(1).
The kubelet relies on a container runtime. You should deploy a container runtime such as
containerd or CRI-O and ensure that it is running within the user namespace before the kubelet starts.
Running CRI plugin of containerd in a user namespace is supported since containerd 1.4.
Running containerd within a user namespace requires the following configurations.
version = 2[plugins."io.containerd.grpc.v1.cri"]
# Disable AppArmor disable_apparmor = true# Ignore an error during setting oom_score_adj restrict_oom_score_adj = true# Disable hugetlb cgroup v2 controller (because systemd does not support delegating hugetlb controller) disable_hugetlb_controller = true[plugins."io.containerd.grpc.v1.cri".containerd]
# Using non-fuse overlayfs is also possible for kernel >= 5.11, but requires SELinux to be disabled snapshotter = "fuse-overlayfs"[plugins."io.containerd.grpc.v1.cri".containerd.runtimes.runc.options]
# We use cgroupfs that is delegated by systemd, so we do not use SystemdCgroup driver# (unless you run another systemd in the namespace) SystemdCgroup = false
The default path of the configuration file is /etc/containerd/config.toml.
The path can be specified with containerd -c /path/to/containerd/config.toml.
Running CRI-O in a user namespace is supported since CRI-O 1.22.
CRI-O requires an environment variable _CRIO_ROOTLESS=1 to be set.
The following configurations are also recommended:
[crio]
storage_driver = "overlay"# Using non-fuse overlayfs is also possible for kernel >= 5.11, but requires SELinux to be disabled storage_option = ["overlay.mount_program=/usr/local/bin/fuse-overlayfs"]
[crio.runtime]
# We use cgroupfs that is delegated by systemd, so we do not use "systemd" driver# (unless you run another systemd in the namespace) cgroup_manager = "cgroupfs"
The default path of the configuration file is /etc/crio/crio.conf.
The path can be specified with crio --config /path/to/crio/crio.conf.
Configuring kubelet
Running kubelet in a user namespace requires the following configuration:
apiVersion:kubelet.config.k8s.io/v1beta1kind:KubeletConfigurationfeatureGates:KubeletInUserNamespace:true# We use cgroupfs that is delegated by systemd, so we do not use "systemd" driver# (unless you run another systemd in the namespace)cgroupDriver:"cgroupfs"
When the KubeletInUserNamespace feature gate is enabled, the kubelet ignores errors
that may happen during setting the following sysctl values on the node.
vm.overcommit_memory
vm.panic_on_oom
kernel.panic
kernel.panic_on_oops
kernel.keys.root_maxkeys
kernel.keys.root_maxbytes.
Within a user namespace, the kubelet also ignores any error raised from trying to open /dev/kmsg.
This feature gate also allows kube-proxy to ignore an error during setting RLIMIT_NOFILE.
The KubeletInUserNamespace feature gate was introduced in Kubernetes v1.22 with "alpha" status.
Running kubelet in a user namespace without using this feature gate is also possible
by mounting a specially crafted proc filesystem (as done by Sysbox), but not officially supported.
Configuring kube-proxy
Running kube-proxy in a user namespace requires the following configuration:
apiVersion:kubeproxy.config.k8s.io/v1alpha1kind:KubeProxyConfigurationmode:"iptables"# or "userspace"conntrack:# Skip setting sysctl value "net.netfilter.nf_conntrack_max"maxPerCore:0# Skip setting "net.netfilter.nf_conntrack_tcp_timeout_established"tcpEstablishedTimeout:0s# Skip setting "net.netfilter.nf_conntrack_tcp_timeout_close"tcpCloseWaitTimeout:0s
Caveats
Most of "non-local" volume drivers such as nfs and iscsi do not work.
Local volumes like local, hostPath, emptyDir, configMap, secret, and downwardAPI are known to work.
Some CNI plugins may not work. Flannel (VXLAN) is known to work.
To ensure that your workloads remain available during maintenance, you can
configure a PodDisruptionBudget.
If availability is important for any applications that run or could run on the node(s)
that you are draining, configure a PodDisruptionBudgets
first and then continue following this guide.
Use kubectl drain to remove a node from service
You can use kubectl drain to safely evict all of your pods from a
node before you perform maintenance on the node (e.g. kernel upgrade,
hardware maintenance, etc.). Safe evictions allow the pod's containers
to gracefully terminate
and will respect the PodDisruptionBudgets you have specified.
Note: By default kubectl drain ignores certain system pods on the node
that cannot be killed; see
the kubectl drain
documentation for more details.
When kubectl drain returns successfully, that indicates that all of
the pods (except the ones excluded as described in the previous paragraph)
have been safely evicted (respecting the desired graceful termination period,
and respecting the PodDisruptionBudget you have defined). It is then safe to
bring down the node by powering down its physical machine or, if running on a
cloud platform, deleting its virtual machine.
First, identify the name of the node you wish to drain. You can list all of the nodes in your cluster with
kubectl get nodes
Next, tell Kubernetes to drain the node:
kubectl drain <node name>
Once it returns (without giving an error), you can power down the node
(or equivalently, if on a cloud platform, delete the virtual machine backing the node).
If you leave the node in the cluster during the maintenance operation, you need to run
kubectl uncordon <node name>
afterwards to tell Kubernetes that it can resume scheduling new pods onto the node.
Draining multiple nodes in parallel
The kubectl drain command should only be issued to a single node at a
time. However, you can run multiple kubectl drain commands for
different nodes in parallel, in different terminals or in the
background. Multiple drain commands running concurrently will still
respect the PodDisruptionBudget you specify.
For example, if you have a StatefulSet with three replicas and have
set a PodDisruptionBudget for that set specifying minAvailable: 2,
kubectl drain only evicts a pod from the StatefulSet if all three
replicas pods are ready; if then you issue multiple drain commands in
parallel, Kubernetes respects the PodDisruptionBudget and ensure
that only 1 (calculated as replicas - minAvailable) Pod is unavailable
at any given time. Any drains that would cause the number of ready
replicas to fall below the specified budget are blocked.
The Eviction API
If you prefer not to use kubectl drain (such as
to avoid calling to an external command, or to get finer control over the pod
eviction process), you can also programmatically cause evictions using the
eviction API.
This document covers topics related to protecting a cluster from accidental or malicious access
and provides recommendations on overall security.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
As Kubernetes is entirely API-driven, controlling and limiting who can access the cluster and what actions
they are allowed to perform is the first line of defense.
Use Transport Layer Security (TLS) for all API traffic
Kubernetes expects that all API communication in the cluster is encrypted by default with TLS, and the
majority of installation methods will allow the necessary certificates to be created and distributed to
the cluster components. Note that some components and installation methods may enable local ports over
HTTP and administrators should familiarize themselves with the settings of each component to identify
potentially unsecured traffic.
API Authentication
Choose an authentication mechanism for the API servers to use that matches the common access patterns
when you install a cluster. For instance, small, single-user clusters may wish to use a simple certificate
or static Bearer token approach. Larger clusters may wish to integrate an existing OIDC or LDAP server that
allow users to be subdivided into groups.
All API clients must be authenticated, even those that are part of the infrastructure like nodes,
proxies, the scheduler, and volume plugins. These clients are typically service accounts or use x509 client certificates, and they are created automatically at cluster startup or are setup as part of the cluster installation.
Once authenticated, every API call is also expected to pass an authorization check. Kubernetes ships
an integrated Role-Based Access Control (RBAC) component that matches an incoming user or group to a
set of permissions bundled into roles. These permissions combine verbs (get, create, delete) with
resources (pods, services, nodes) and can be namespace-scoped or cluster-scoped. A set of out-of-the-box
roles are provided that offer reasonable default separation of responsibility depending on what
actions a client might want to perform. It is recommended that you use the
Node and
RBAC authorizers together, in combination with the
NodeRestriction admission plugin.
As with authentication, simple and broad roles may be appropriate for smaller clusters, but as
more users interact with the cluster, it may become necessary to separate teams into separate
namespaces with more limited roles.
With authorization, it is important to understand how updates on one object may cause actions in
other places. For instance, a user may not be able to create pods directly, but allowing them to
create a deployment, which creates pods on their behalf, will let them create those pods
indirectly. Likewise, deleting a node from the API will result in the pods scheduled to that node
being terminated and recreated on other nodes. The out-of-the box roles represent a balance
between flexibility and common use cases, but more limited roles should be carefully reviewed
to prevent accidental escalation. You can make roles specific to your use case if the out-of-box ones don't meet your needs.
Kubelets expose HTTPS endpoints which grant powerful control over the node and containers.
By default Kubelets allow unauthenticated access to this API.
Production clusters should enable Kubelet authentication and authorization.
Controlling the capabilities of a workload or user at runtime
Authorization in Kubernetes is intentionally high level, focused on coarse actions on resources.
More powerful controls exist as policies to limit by use case how those objects act on the
cluster, themselves, and other resources.
Limiting resource usage on a cluster
Resource quota limits the number or capacity of
resources granted to a namespace. This is most often used to limit the amount of CPU, memory,
or persistent disk a namespace can allocate, but can also control how many pods, services, or
volumes exist in each namespace.
Limit ranges restrict the maximum or minimum size of some of the
resources above, to prevent users from requesting unreasonably high or low values for commonly
reserved resources like memory, or to provide default limits when none are specified.
Controlling what privileges containers run with
A pod definition contains a security context
that allows it to request access to run as a specific Linux user on a node (like root),
access to run privileged or access the host network, and other controls that would otherwise
allow it to run unfettered on a hosting node.
Generally, most application workloads need limited access to host resources so they can
successfully run as a root process (uid 0) without access to host information. However,
considering the privileges associated with the root user, you should write application
containers to run as a non-root user. Similarly, administrators who wish to prevent
client applications from escaping their containers should apply the Baseline
or Restricted Pod Security Standard.
Preventing containers from loading unwanted kernel modules
The Linux kernel automatically loads kernel modules from disk if needed in certain
circumstances, such as when a piece of hardware is attached or a filesystem is mounted. Of
particular relevance to Kubernetes, even unprivileged processes can cause certain
network-protocol-related kernel modules to be loaded, just by creating a socket of the
appropriate type. This may allow an attacker to exploit a security hole in a kernel module
that the administrator assumed was not in use.
To prevent specific modules from being automatically loaded, you can uninstall them from
the node, or add rules to block them. On most Linux distributions, you can do that by
creating a file such as /etc/modprobe.d/kubernetes-blacklist.conf with contents like:
# DCCP is unlikely to be needed, has had multiple serious
# vulnerabilities, and is not well-maintained.
blacklist dccp
# SCTP is not used in most Kubernetes clusters, and has also had
# vulnerabilities in the past.
blacklist sctp
To block module loading more generically, you can use a Linux Security Module (such as
SELinux) to completely deny the module_request permission to containers, preventing the
kernel from loading modules for containers under any circumstances. (Pods would still be
able to use modules that had been loaded manually, or modules that were loaded by the
kernel on behalf of some more-privileged process.)
Restricting network access
The network policies for a namespace
allows application authors to restrict which pods in other namespaces may access pods and ports
within their namespaces. Many of the supported Kubernetes networking providers
now respect network policy.
Quota and limit ranges can also be used to control whether users may request node ports or
load-balanced services, which on many clusters can control whether those users applications
are visible outside of the cluster.
Additional protections may be available that control network rules on a per-plugin or per-
environment basis, such as per-node firewalls, physically separating cluster nodes to
prevent cross talk, or advanced networking policy.
Restricting cloud metadata API access
Cloud platforms (AWS, Azure, GCE, etc.) often expose metadata services locally to instances.
By default these APIs are accessible by pods running on an instance and can contain cloud
credentials for that node, or provisioning data such as kubelet credentials. These credentials
can be used to escalate within the cluster or to other cloud services under the same account.
When running Kubernetes on a cloud platform, limit permissions given to instance credentials, use
network policies to restrict pod access
to the metadata API, and avoid using provisioning data to deliver secrets.
As an administrator, a beta admission plugin PodNodeSelector can be used to force pods
within a namespace to default or require a specific node selector, and if end users cannot
alter namespaces, this can strongly limit the placement of all of the pods in a specific workload.
Protecting cluster components from compromise
This section describes some common patterns for protecting clusters from compromise.
Restrict access to etcd
Write access to the etcd backend for the API is equivalent to gaining root on the entire cluster,
and read access can be used to escalate fairly quickly. Administrators should always use strong
credentials from the API servers to their etcd server, such as mutual auth via TLS client certificates,
and it is often recommended to isolate the etcd servers behind a firewall that only the API servers
may access.
Caution: Allowing other components within the cluster to access the master etcd instance with
read or write access to the full keyspace is equivalent to granting cluster-admin access. Using
separate etcd instances for non-master components or using etcd ACLs to restrict read and write
access to a subset of the keyspace is strongly recommended.
Enable audit logging
The audit logger is a beta feature that records actions taken by the
API for later analysis in the event of a compromise. It is recommended to enable audit logging
and archive the audit file on a secure server.
Restrict access to alpha or beta features
Alpha and beta Kubernetes features are in active development and may have limitations or bugs
that result in security vulnerabilities. Always assess the value an alpha or beta feature may
provide against the possible risk to your security posture. When in doubt, disable features you
do not use.
Rotate infrastructure credentials frequently
The shorter the lifetime of a secret or credential the harder it is for an attacker to make
use of that credential. Set short lifetimes on certificates and automate their rotation. Use
an authentication provider that can control how long issued tokens are available and use short
lifetimes where possible. If you use service-account tokens in external integrations, plan to
rotate those tokens frequently. For example, once the bootstrap phase is complete, a bootstrap
token used for setting up nodes should be revoked or its authorization removed.
Review third party integrations before enabling them
Many third party integrations to Kubernetes may alter the security profile of your cluster. When
enabling an integration, always review the permissions that an extension requests before granting
it access. For example, many security integrations may request access to view all secrets on
your cluster which is effectively making that component a cluster admin. When in doubt,
restrict the integration to functioning in a single namespace if possible.
Components that create pods may also be unexpectedly powerful if they can do so inside namespaces
like the kube-system namespace, because those pods can gain access to service account secrets
or run with elevated permissions if those service accounts are granted access to permissive
PodSecurityPolicies.
If you use Pod Security admission and allow
any component to create Pods within a namespace that permits privileged Pods, those Pods may
be able to escape their containers and use this widened access to elevate their privileges.
You should not allow untrusted components to create Pods in any system namespace (those with
names that start with kube-) nor in any namespace where that access grant allows the possibility
of privilege escalation.
Encrypt secrets at rest
In general, the etcd database will contain any information accessible via the Kubernetes API
and may grant an attacker significant visibility into the state of your cluster. Always encrypt
your backups using a well reviewed backup and encryption solution, and consider using full disk
encryption where possible.
Kubernetes supports encryption at rest, a feature
introduced in 1.7, v1 beta since 1.13, and v2 alpha since 1.25. This will encrypt resources like Secret and ConfigMap in etcd, preventing
parties that gain access to your etcd backups from viewing the content of those secrets. While
this feature is currently beta, it offers an additional level of defense when backups
are not encrypted or an attacker gains read access to etcd.
Receiving alerts for security updates and reporting vulnerabilities
Join the kubernetes-announce
group for emails about security announcements. See the
security reporting
page for more on how to report vulnerabilities.
4.2.33 - Set Kubelet parameters via a config file
A subset of the Kubelet's configuration parameters may be
set via an on-disk config file, as a substitute for command-line flags.
Providing parameters via a config file is the recommended approach because
it simplifies node deployment and configuration management.
Create the config file
The subset of the Kubelet's configuration that can be configured via a file
is defined by the
KubeletConfiguration
struct.
The configuration file must be a JSON or YAML representation of the parameters
in this struct. Make sure the Kubelet has read permissions on the file.
Here is an example of what this file might look like:
In the example, the Kubelet is configured to serve on IP address 192.168.0.8 and port 20250, pull images in parallel,
and evict Pods when available memory drops below 200Mi.
All other Kubelet configuration values are left at their built-in defaults, unless overridden
by flags. Command line flags which target the same value as a config file will override that value.
Start a Kubelet process configured via the config file
Note: If you use kubeadm to initialize your cluster, use the kubelet-config while creating your cluster with kubeadmin init.
See configuring kubelet using kubeadm for details.
Start the Kubelet with the --config flag set to the path of the Kubelet's config file.
The Kubelet will then load its config from this file.
Note that command line flags which target the same value as a config file will override that value.
This helps ensure backwards compatibility with the command-line API.
Note that relative file paths in the Kubelet config file are resolved relative to the
location of the Kubelet config file, whereas relative paths in command line flags are resolved
relative to the Kubelet's current working directory.
Note that some default values differ between command-line flags and the Kubelet config file.
If --config is provided and the values are not specified via the command line, the
defaults for the KubeletConfiguration version apply.
In the above example, this version is kubelet.config.k8s.io/v1beta1.
What's next
Learn more about kubelet configuration by checking the
KubeletConfiguration
reference.
4.2.34 - Share a Cluster with Namespaces
This page shows how to view, work in, and delete namespaces. The page also shows how to use Kubernetes namespaces to subdivide your cluster.
NAME STATUS AGE
default Active 11d
kube-system Active 11d
kube-public Active 11d
Kubernetes starts with three initial namespaces:
default The default namespace for objects with no other namespace
kube-system The namespace for objects created by the Kubernetes system
kube-public This namespace is created automatically and is readable by all users (including those not authenticated). This namespace is mostly reserved for cluster usage, in case that some resources should be visible and readable publicly throughout the whole cluster. The public aspect of this namespace is only a convention, not a requirement.
You can also get the summary of a specific namespace using:
kubectl get namespaces <name>
Or you can get detailed information with:
kubectl describe namespaces <name>
Name: default
Labels: <none>
Annotations: <none>
Status: Active
No resource quota.
Resource Limits
Type Resource Min Max Default
---- -------- --- --- ---
Container cpu - - 100m
Note that these details show both resource quota (if present) as well as resource limit ranges.
Resource quota tracks aggregate usage of resources in the Namespace and allows cluster operators
to define Hard resource usage limits that a Namespace may consume.
A limit range defines min/max constraints on the amount of resources a single entity can consume in
a Namespace.
The name of your namespace must be a valid
DNS label.
There's an optional field finalizers, which allows observables to purge resources whenever the namespace is deleted. Keep in mind that if you specify a nonexistent finalizer, the namespace will be created but will get stuck in the Terminating state if the user tries to delete it.
More information on finalizers can be found in the namespace design doc.
Warning: This deletes everything under the namespace!
This delete is asynchronous, so for a time you will see the namespace in the Terminating state.
Subdividing your cluster using Kubernetes namespaces
Understand the default namespace
By default, a Kubernetes cluster will instantiate a default namespace when provisioning the cluster to hold the default set of Pods,
Services, and Deployments used by the cluster.
Assuming you have a fresh cluster, you can introspect the available namespaces by doing the following:
kubectl get namespaces
NAME STATUS AGE
default Active 13m
Create new namespaces
For this exercise, we will create two additional Kubernetes namespaces to hold our content.
In a scenario where an organization is using a shared Kubernetes cluster for development and production use cases:
The development team would like to maintain a space in the cluster where they can get a view on the list of Pods, Services, and Deployments
they use to build and run their application. In this space, Kubernetes resources come and go, and the restrictions on who can or cannot modify resources
are relaxed to enable agile development.
The operations team would like to maintain a space in the cluster where they can enforce strict procedures on who can or cannot manipulate the set of
Pods, Services, and Deployments that run the production site.
One pattern this organization could follow is to partition the Kubernetes cluster into two namespaces: development and production.
NAME READY UP-TO-DATE AVAILABLE AGE
cattle 5/5 5 5 10s
kubectl get pods -l app=cattle -n=production
NAME READY STATUS RESTARTS AGE
cattle-2263376956-41xy6 1/1 Running 0 34s
cattle-2263376956-kw466 1/1 Running 0 34s
cattle-2263376956-n4v97 1/1 Running 0 34s
cattle-2263376956-p5p3i 1/1 Running 0 34s
cattle-2263376956-sxpth 1/1 Running 0 34s
At this point, it should be clear that the resources users create in one namespace are hidden from the other namespace.
As the policy support in Kubernetes evolves, we will extend this scenario to show how you can provide different
authorization rules for each namespace.
Understanding the motivation for using namespaces
A single cluster should be able to satisfy the needs of multiple users or groups of users (henceforth a 'user community').
Kubernetes namespaces help different projects, teams, or customers to share a Kubernetes cluster.
policies (who can or cannot perform actions in their community)
constraints (this community is allowed this much quota, etc.)
A cluster operator may create a Namespace for each unique user community.
The Namespace provides a unique scope for:
named resources (to avoid basic naming collisions)
delegated management authority to trusted users
ability to limit community resource consumption
Use cases include:
As a cluster operator, I want to support multiple user communities on a single cluster.
As a cluster operator, I want to delegate authority to partitions of the cluster to trusted users
in those communities.
As a cluster operator, I want to limit the amount of resources each community can consume in order
to limit the impact to other communities using the cluster.
As a cluster user, I want to interact with resources that are pertinent to my user community in
isolation of what other user communities are doing on the cluster.
Understanding namespaces and DNS
When you create a Service, it creates a corresponding DNS entry.
This entry is of the form <service-name>.<namespace-name>.svc.cluster.local, which means
that if a container uses <service-name> it will resolve to the service which
is local to a namespace. This is useful for using the same configuration across
multiple namespaces such as Development, Staging and Production. If you want to reach
across namespaces, you need to use the fully qualified domain name (FQDN).
Adjust manifests and other resources based on the API changes that accompany the
new Kubernetes version
Before you begin
You must have an existing cluster. This page is about upgrading from Kubernetes
1.24 to Kubernetes 1.25. If your cluster
is not currently running Kubernetes 1.24 then please check
the documentation for the version of Kubernetes that you plan to upgrade to.
Upgrade approaches
kubeadm
If your cluster was deployed using the kubeadm tool, refer to
Upgrading kubeadm clusters
for detailed information on how to upgrade the cluster.
For each node in your cluster, drain
that node and then either replace it with a new node that uses the 1.25
kubelet, or upgrade the kubelet on that node and bring the node back into service.
Other deployments
Refer to the documentation for your cluster deployment tool to learn the recommended set
up steps for maintenance.
Post-upgrade tasks
Switch your cluster's storage API version
The objects that are serialized into etcd for a cluster's internal
representation of the Kubernetes resources active in the cluster are
written using a particular version of the API.
When the supported API changes, these objects may need to be rewritten
in the newer API. Failure to do this will eventually result in resources
that are no longer decodable or usable by the Kubernetes API server.
For each affected object, fetch it using the latest supported API and then
write it back also using the latest supported API.
Update manifests
Upgrading to a new Kubernetes version can provide new APIs.
You can use kubectl convert command to convert manifests between different API versions.
For example:
kubectl convert -f pod.yaml --output-version v1
The kubectl tool replaces the contents of pod.yaml with a manifest that sets kind to
Pod (unchanged), but with a revised apiVersion.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You also need to create a sample Deployment
to experiment with the different types of cascading deletion. You will need to
recreate the Deployment for each type.
Check owner references on your pods
Check that the ownerReferences field is present on your pods:
kubectl get pods -l app=nginx --output=yaml
The output has an ownerReferences field similar to this:
By default, Kubernetes uses background cascading deletion
to delete dependents of an object. You can switch to foreground cascading deletion
using either kubectl or the Kubernetes API, depending on the Kubernetes
version your cluster runs.
To check the version, enter kubectl version.
Use either kubectl or the Kubernetes API to delete the Deployment,
depending on the Kubernetes version your cluster runs.
To check the version, enter kubectl version.
You can delete objects using background cascading deletion using kubectl
or the Kubernetes API.
Kubernetes uses background cascading deletion by default, and does so
even if you run the following commands without the --cascade flag or the
propagationPolicy argument.
Kubernetes uses background cascading deletion by default, and does so
even if you run the following commands without the --cascade flag or the
propagationPolicy: Background argument.
By default, when you tell Kubernetes to delete an object, the
controller also deletes
dependent objects. You can make Kubernetes orphan these dependents using
kubectl or the Kubernetes API, depending on the Kubernetes version your
cluster runs.
To check the version, enter kubectl version.
This page shows how to configure a Key Management Service (KMS) provider and plugin to enable secret data encryption. Currently there are two KMS API versions. KMS v1 will continue to work while v2 develops in maturity. If you are not sure which KMS API version to pick, choose v1.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The version of Kubernetes that you need depends on which KMS API version
you have selected.
If you selected KMS API v1, any supported Kubernetes version will work fine.
If you selected KMS API v2, you should use Kubernetes v1.25
(if you are running a different version of Kubernetes that also supports the v2 KMS
API, switch to the documentation for that version of Kubernetes).
To check the version, enter kubectl version.
KMS v1
Kubernetes version 1.10.0 or later is required
Your cluster must use etcd v3 or later
FEATURE STATE:Kubernetes v1.12 [beta]
KMS v2
Kubernetes version 1.25.0 or later is required
Set kube-apiserver feature gate: --feature-gates=KMSv2=true to configure a KMS v2 provider
Your cluster must use etcd v3 or later
FEATURE STATE:Kubernetes v1.25 [alpha]
The KMS encryption provider uses an envelope encryption scheme to encrypt data in etcd.
The data is encrypted using a data encryption key (DEK); a new DEK is generated for each encryption.
The DEKs are encrypted with a key encryption key (KEK) that is stored and managed in a remote KMS.
The KMS provider uses gRPC to communicate with a specific KMS plugin.
The KMS plugin, which is implemented as a gRPC server and deployed on the same host(s)
as the Kubernetes control plane, is responsible for all communication with the remote KMS.
Configuring the KMS provider
To configure a KMS provider on the API server, include a provider of type kms in the
providers array in the encryption configuration file and set the following properties:
KMS v1
name: Display name of the KMS plugin. Cannot be changed once set.
endpoint: Listen address of the gRPC server (KMS plugin). The endpoint is a UNIX domain socket.
cachesize: Number of data encryption keys (DEKs) to be cached in the clear.
When cached, DEKs can be used without another call to the KMS;
whereas DEKs that are not cached require a call to the KMS to unwrap.
timeout: How long should kube-apiserver wait for kms-plugin to respond before
returning an error (default is 3 seconds).
KMS v2
apiVersion: API Version for KMS provider (Allowed values: v2, v1 or empty. Any other value will result in an error.) Must be set to v2 to use the KMS v2 APIs.
name: Display name of the KMS plugin. Cannot be changed once set.
endpoint: Listen address of the gRPC server (KMS plugin). The endpoint is a UNIX domain socket.
cachesize: Number of data encryption keys (DEKs) to be cached in the clear.
When cached, DEKs can be used without another call to the KMS;
whereas DEKs that are not cached require a call to the KMS to unwrap.
timeout: How long should kube-apiserver wait for kms-plugin to respond before
returning an error (default is 3 seconds).
To implement a KMS plugin, you can develop a new plugin gRPC server or enable a KMS plugin
already provided by your cloud provider.
You then integrate the plugin with the remote KMS and deploy it on the Kubernetes master.
Enabling the KMS supported by your cloud provider
Refer to your cloud provider for instructions on enabling the cloud provider-specific KMS plugin.
Developing a KMS plugin gRPC server
You can develop a KMS plugin gRPC server using a stub file available for Go. For other languages,
you use a proto file to create a stub file that you can use to develop the gRPC server code.
KMS v1
Using Go: Use the functions and data structures in the stub file:
api.pb.go
to develop the gRPC server code
Using languages other than Go: Use the protoc compiler with the proto file:
api.proto
to generate a stub file for the specific language
KMS v2
Using Go: Use the functions and data structures in the stub file:
api.pb.go
to develop the gRPC server code
Using languages other than Go: Use the protoc compiler with the proto file:
api.proto
to generate a stub file for the specific language
Then use the functions and data structures in the stub file to develop the server code.
Notes
KMS v1
kms plugin version: v1beta1
In response to procedure call Version, a compatible KMS plugin should return v1beta1 as VersionResponse.version.
message version: v1beta1
All messages from KMS provider have the version field set to current version v1beta1.
protocol: UNIX domain socket (unix)
The plugin is implemented as a gRPC server that listens at UNIX domain socket. The plugin deployment should create a file on the file system to run the gRPC unix domain socket connection. The API server (gRPC client) is configured with the KMS provider (gRPC server) unix domain socket endpoint in order to communicate with it. An abstract Linux socket may be used by starting the endpoint with /@, i.e. unix:///@foo. Care must be taken when using this type of socket as they do not have concept of ACL (unlike traditional file based sockets). However, they are subject to Linux networking namespace, so will only be accessible to containers within the same pod unless host networking is used.
KMS v2
kms plugin version: v2alpha1
In response to procedure call Status, a compatible KMS plugin should return v2alpha1 as StatusResponse.Version, "ok" as StatusResponse.Healthz and a keyID (KMS KEK ID) as StatusResponse.KeyID
protocol: UNIX domain socket (unix)
The plugin is implemented as a gRPC server that listens at UNIX domain socket. The plugin deployment should create a file on the file system to run the gRPC unix domain socket connection. The API server (gRPC client) is configured with the KMS provider (gRPC server) unix domain socket endpoint in order to communicate with it. An abstract Linux socket may be used by starting the endpoint with /@, i.e. unix:///@foo. Care must be taken when using this type of socket as they do not have concept of ACL (unlike traditional file based sockets). However, they are subject to Linux networking namespace, so will only be accessible to containers within the same pod unless host networking is used.
Integrating a KMS plugin with the remote KMS
The KMS plugin can communicate with the remote KMS using any protocol supported by the KMS.
All configuration data, including authentication credentials the KMS plugin uses to communicate with the remote KMS,
are stored and managed by the KMS plugin independently.
The KMS plugin can encode the ciphertext with additional metadata that may be required before sending it to the KMS for decryption.
Deploying the KMS plugin
Ensure that the KMS plugin runs on the same host(s) as the Kubernetes master(s).
Encrypting your data with the KMS provider
To encrypt the data:
Create a new EncryptionConfiguration file using the appropriate properties for the kms provider to encrypt resources like Secrets and ConfigMaps.
Set the --encryption-provider-config flag on the kube-apiserver to point to the location of the configuration file.
Until the steps defined in Ensuring all secrets are encrypted are performed, the providers list should end with the identity: {} provider to allow unencrypted data to be read. Once all resources are encrypted, the identity provider should be removed to prevent the API server from honoring unencrypted data.
Data is encrypted when written to etcd. After restarting your kube-apiserver,
any newly created or updated secret should be encrypted when stored. To verify,
you can use the etcdctl command line program to retrieve the contents of your secret.
Create a new secret called secret1 in the default namespace:
Using the etcdctl command line, read that secret out of etcd:
ETCDCTL_API=3 etcdctl get /kubernetes.io/secrets/default/secret1 [...] | hexdump -C
where [...] contains the additional arguments for connecting to the etcd server.
Verify the stored secret is prefixed with k8s:enc:kms:v1: for KMS v1 or prefixed with k8s:enc:kms:v2: for KMS v2, which indicates that the kms provider has encrypted the resulting data.
Verify that the secret is correctly decrypted when retrieved via the API:
kubectl describe secret secret1 -n default
The Secret should contain mykey: mydata
Ensuring all secrets are encrypted
Because secrets are encrypted on write, performing an update on a secret encrypts that content.
The following command reads all secrets and then updates them to apply server side encryption.
If an error occurs due to a conflicting write, retry the command.
For larger clusters, you may wish to subdivide the secrets by namespace or script an update.
This page describes the CoreDNS upgrade process and how to install CoreDNS instead of kube-dns.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.9.
To check the version, enter kubectl version.
About CoreDNS
CoreDNS is a flexible, extensible DNS server
that can serve as the Kubernetes cluster DNS.
Like Kubernetes, the CoreDNS project is hosted by the
CNCF.
You can use CoreDNS instead of kube-dns in your cluster by replacing
kube-dns in an existing deployment, or by using tools like kubeadm
that will deploy and upgrade the cluster for you.
Installing CoreDNS
For manual deployment or replacement of kube-dns, see the documentation at the
CoreDNS GitHub project.
Migrating to CoreDNS
Upgrading an existing cluster with kubeadm
In Kubernetes version 1.21, kubeadm removed its support for kube-dns as a DNS application.
For kubeadm v1.25, the only supported cluster DNS application
is CoreDNS.
You can move to CoreDNS when you use kubeadm to upgrade a cluster that is
using kube-dns. In this case, kubeadm generates the CoreDNS configuration
("Corefile") based upon the kube-dns ConfigMap, preserving configurations for
stub domains, and upstream name server.
Upgrading CoreDNS
You can check the version of CoreDNS that kubeadm installs for each version of
Kubernetes in the page
CoreDNS version in Kubernetes.
CoreDNS can be upgraded manually in case you want to only upgrade CoreDNS
or use your own custom image.
There is a helpful guideline and walkthrough
available to ensure a smooth upgrade.
Make sure the existing CoreDNS configuration ("Corefile") is retained when
upgrading your cluster.
If you are upgrading your cluster using the kubeadm tool, kubeadm
can take care of retaining the existing CoreDNS configuration automatically.
Tuning CoreDNS
When resource utilisation is a concern, it may be useful to tune the
configuration of CoreDNS. For more details, check out the
documentation on scaling CoreDNS.
What's next
You can configure CoreDNS to support many more use cases than
kube-dns does by modifying the CoreDNS configuration ("Corefile").
For more information, see the documentation
for the kubernetes CoreDNS plugin, or read the
Custom DNS Entries for Kubernetes.
in the CoreDNS blog.
4.2.39 - Using NodeLocal DNSCache in Kubernetes Clusters
FEATURE STATE:Kubernetes v1.18 [stable]
This page provides an overview of NodeLocal DNSCache feature in Kubernetes.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
NodeLocal DNSCache improves Cluster DNS performance by running a DNS caching agent
on cluster nodes as a DaemonSet. In today's architecture, Pods in 'ClusterFirst' DNS mode
reach out to a kube-dns serviceIP for DNS queries. This is translated to a
kube-dns/CoreDNS endpoint via iptables rules added by kube-proxy.
With this new architecture, Pods will reach out to the DNS caching agent
running on the same node, thereby avoiding iptables DNAT rules and connection tracking.
The local caching agent will query kube-dns service for cache misses of cluster
hostnames ("cluster.local" suffix by default).
Motivation
With the current DNS architecture, it is possible that Pods with the highest DNS QPS
have to reach out to a different node, if there is no local kube-dns/CoreDNS instance.
Having a local cache will help improve the latency in such scenarios.
Skipping iptables DNAT and connection tracking will help reduce
conntrack races
and avoid UDP DNS entries filling up conntrack table.
Connections from the local caching agent to kube-dns service can be upgraded to TCP.
TCP conntrack entries will be removed on connection close in contrast with
UDP entries that have to timeout
(defaultnf_conntrack_udp_timeout is 30 seconds)
Upgrading DNS queries from UDP to TCP would reduce tail latency attributed to
dropped UDP packets and DNS timeouts usually up to 30s (3 retries + 10s timeout).
Since the nodelocal cache listens for UDP DNS queries, applications don't need to be changed.
Metrics & visibility into DNS requests at a node level.
Negative caching can be re-enabled, thereby reducing the number of queries for the kube-dns service.
Architecture Diagram
This is the path followed by DNS Queries after NodeLocal DNSCache is enabled:
Nodelocal DNSCache flow
This image shows how NodeLocal DNSCache handles DNS queries.
Configuration
Note: The local listen IP address for NodeLocal DNSCache can be any address that
can be guaranteed to not collide with any existing IP in your cluster.
It's recommended to use an address with a local scope, for example,
from the 'link-local' range '169.254.0.0/16' for IPv4 or from the
'Unique Local Address' range in IPv6 'fd00::/8'.
This feature can be enabled using the following steps:
Prepare a manifest similar to the sample
nodelocaldns.yaml
and save it as nodelocaldns.yaml.
If using IPv6, the CoreDNS configuration file needs to enclose all the IPv6 addresses
into square brackets if used in 'IP:Port' format.
If you are using the sample manifest from the previous point, this will require you to modify
the configuration line L70
like this: "health [__PILLAR__LOCAL__DNS__]:8080"
Substitute the variables in the manifest with the right values:
kubedns=`kubectl get svc kube-dns -n kube-system -o jsonpath={.spec.clusterIP}`domain=<cluster-domain>
localdns=<node-local-address>
<cluster-domain> is "cluster.local" by default. <node-local-address> is the
local listen IP address chosen for NodeLocal DNSCache.
If kube-proxy is running in IPTABLES mode:
sed -i "s/__PILLAR__LOCAL__DNS__/$localdns/g; s/__PILLAR__DNS__DOMAIN__/$domain/g; s/__PILLAR__DNS__SERVER__/$kubedns/g" nodelocaldns.yaml
__PILLAR__CLUSTER__DNS__ and __PILLAR__UPSTREAM__SERVERS__ will be populated by
the node-local-dns pods.
In this mode, the node-local-dns pods listen on both the kube-dns service IP
as well as <node-local-address>, so pods can look up DNS records using either IP address.
If kube-proxy is running in IPVS mode:
sed -i "s/__PILLAR__LOCAL__DNS__/$localdns/g; s/__PILLAR__DNS__DOMAIN__/$domain/g; s/,__PILLAR__DNS__SERVER__//g; s/__PILLAR__CLUSTER__DNS__/$kubedns/g" nodelocaldns.yaml
In this mode, the node-local-dns pods listen only on <node-local-address>.
The node-local-dns interface cannot bind the kube-dns cluster IP since the
interface used for IPVS loadbalancing already uses this address.
__PILLAR__UPSTREAM__SERVERS__ will be populated by the node-local-dns pods.
Run kubectl create -f nodelocaldns.yaml
If using kube-proxy in IPVS mode, --cluster-dns flag to kubelet needs to be modified
to use <node-local-address> that NodeLocal DNSCache is listening on.
Otherwise, there is no need to modify the value of the --cluster-dns flag,
since NodeLocal DNSCache listens on both the kube-dns service IP as well as
<node-local-address>.
Once enabled, the node-local-dns Pods will run in the kube-system namespace
on each of the cluster nodes. This Pod runs CoreDNS
in cache mode, so all CoreDNS metrics exposed by the different plugins will
be available on a per-node basis.
You can disable this feature by removing the DaemonSet, using kubectl delete -f <manifest>.
You should also revert any changes you made to the kubelet configuration.
StubDomains and Upstream server Configuration
StubDomains and upstream servers specified in the kube-dns ConfigMap in the kube-system namespace
are automatically picked up by node-local-dns pods. The ConfigMap contents need to follow the format
shown in the example.
The node-local-dns ConfigMap can also be modified directly with the stubDomain configuration
in the Corefile format. Some cloud providers might not allow modifying node-local-dns ConfigMap directly.
In those cases, the kube-dns ConfigMap can be updated.
Setting memory limits
The node-local-dns Pods use memory for storing cache entries and processing queries.
Since they do not watch Kubernetes objects, the cluster size or the number of Services/Endpoints
do not directly affect memory usage. Memory usage is influenced by the DNS query pattern.
From CoreDNS docs,
The default cache size is 10000 entries, which uses about 30 MB when completely filled.
This would be the memory usage for each server block (if the cache gets completely filled).
Memory usage can be reduced by specifying smaller cache sizes.
The number of concurrent queries is linked to the memory demand, because each extra
goroutine used for handling a query requires an amount of memory. You can set an upper limit
using the max_concurrent option in the forward plugin.
If a node-local-dns Pod attempts to use more memory than is available (because of total system
resources, or because of a configured
resource limit), the operating system
may shut down that pod's container.
If this happens, the container that is terminated (“OOMKilled”) does not clean up the custom
packet filtering rules that it previously added during startup.
The node-local-dns container should get restarted (since managed as part of a DaemonSet), but this
will lead to a brief DNS downtime each time that the container fails: the packet filtering rules direct
DNS queries to a local Pod that is unhealthy.
You can determine a suitable memory limit by running node-local-dns pods without a limit and
measuring the peak usage. You can also set up and use a
VerticalPodAutoscaler
in recommender mode, and then check its recommendations.
4.2.40 - Using sysctls in a Kubernetes Cluster
FEATURE STATE:Kubernetes v1.21 [stable]
This document describes how to configure and use kernel parameters within a
Kubernetes cluster using the sysctl
interface.
Note: Starting from Kubernetes version 1.23, the kubelet supports the use of either / or .
as separators for sysctl names.
Starting from Kubernetes version 1.25, setting Sysctls for a Pod supports setting sysctls with slashes.
For example, you can represent the same sysctl name as kernel.shm_rmid_forced using a
period as the separator, or as kernel/shm_rmid_forced using a slash as a separator.
For more sysctl parameter conversion method details, please refer to
the page sysctl.d(5) from
the Linux man-pages project.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
For some steps, you also need to be able to reconfigure the command line
options for the kubelets running on your cluster.
Listing all Sysctl Parameters
In Linux, the sysctl interface allows an administrator to modify kernel
parameters at runtime. Parameters are available via the /proc/sys/ virtual
process file system. The parameters cover various subsystems such as:
Sysctls are grouped into safe and unsafe sysctls. In addition to proper
namespacing, a safe sysctl must be properly isolated between pods on the
same node. This means that setting a safe sysctl for one pod
must not have any influence on any other pod on the node
must not allow to harm the node's health
must not allow to gain CPU or memory resources outside of the resource limits
of a pod.
By far, most of the namespaced sysctls are not necessarily considered safe.
The following sysctls are supported in the safe set:
Note: The example net.ipv4.tcp_syncookies is not namespaced on Linux kernel version 4.4 or lower.
This list will be extended in future Kubernetes versions when the kubelet
supports better isolation mechanisms.
All safe sysctls are enabled by default.
All unsafe sysctls are disabled by default and must be allowed manually by the
cluster admin on a per-node basis. Pods with disabled unsafe sysctls will be
scheduled, but will fail to launch.
With the warning above in mind, the cluster admin can allow certain unsafe
sysctls for very special situations such as high-performance or real-time
application tuning. Unsafe sysctls are enabled on a node-by-node basis with a
flag of the kubelet; for example:
A number of sysctls are namespaced in today's Linux kernels. This means that
they can be set independently for each pod on a node. Only namespaced sysctls
are configurable via the pod securityContext within Kubernetes.
The following sysctls are known to be namespaced. This list could change
in future versions of the Linux kernel.
kernel.shm*,
kernel.msg*,
kernel.sem,
fs.mqueue.*,
The parameters under net.* that can be set in container networking
namespace. However, there are exceptions (e.g.,
net.netfilter.nf_conntrack_max and net.netfilter.nf_conntrack_expect_max
can be set in container networking namespace but they are unnamespaced).
Sysctls with no namespace are called node-level sysctls. If you need to set
them, you must manually configure them on each node's operating system, or by
using a DaemonSet with privileged containers.
Use the pod securityContext to configure namespaced sysctls. The securityContext
applies to all containers in the same pod.
This example uses the pod securityContext to set a safe sysctl
kernel.shm_rmid_forced and two unsafe sysctls net.core.somaxconn and
kernel.msgmax. There is no distinction between safe and unsafe sysctls in
the specification.
Warning: Only modify sysctl parameters after you understand their effects, to avoid
destabilizing your operating system.
Warning: Due to their nature of being unsafe, the use of unsafe sysctls
is at-your-own-risk and can lead to severe problems like wrong behavior of
containers, resource shortage or complete breakage of a node.
It is good practice to consider nodes with special sysctl settings as
tainted within a cluster, and only schedule pods onto them which need those
sysctl settings. It is suggested to use the Kubernetes taints and toleration
feature to implement this.
A pod with the unsafe sysctls will fail to launch on any node which has not
enabled those two unsafe sysctls explicitly. As with node-level sysctls it
is recommended to use
taints and toleration feature or
taints on nodes
to schedule those pods onto the right nodes.
4.2.41 - Utilizing the NUMA-aware Memory Manager
FEATURE STATE:Kubernetes v1.22 [beta]
The Kubernetes Memory Manager enables the feature of guaranteed memory (and hugepages)
allocation for pods in the GuaranteedQoS class.
The Memory Manager employs hint generation protocol to yield the most suitable NUMA affinity for a pod.
The Memory Manager feeds the central manager (Topology Manager) with these affinity hints.
Based on both the hints and Topology Manager policy, the pod is rejected or admitted to the node.
Moreover, the Memory Manager ensures that the memory which a pod requests
is allocated from a minimum number of NUMA nodes.
The Memory Manager is only pertinent to Linux based hosts.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.21.
To check the version, enter kubectl version.
To align memory resources with other requested resources in a Pod spec:
the CPU Manager should be enabled and proper CPU Manager policy should be configured on a Node.
See control CPU Management Policies;
the Topology Manager should be enabled and proper Topology Manager policy should be configured on a Node.
See control Topology Management Policies.
Starting from v1.22, the Memory Manager is enabled by default through MemoryManagerfeature gate.
Preceding v1.22, the kubelet must be started with the following flag:
--feature-gates=MemoryManager=true
in order to enable the Memory Manager feature.
How Memory Manager Operates?
The Memory Manager currently offers the guaranteed memory (and hugepages) allocation
for Pods in Guaranteed QoS class.
To immediately put the Memory Manager into operation follow the guidelines in the section
Memory Manager configuration, and subsequently,
prepare and deploy a Guaranteed pod as illustrated in the section
Placing a Pod in the Guaranteed QoS class.
The Memory Manager is a Hint Provider, and it provides topology hints for
the Topology Manager which then aligns the requested resources according to these topology hints.
It also enforces cgroups (i.e. cpuset.mems) for pods.
The complete flow diagram concerning pod admission and deployment process is illustrated in
Memory Manager KEP: Design Overview and below:
During this process, the Memory Manager updates its internal counters stored in
Node Map and Memory Maps to manage guaranteed memory allocation.
The Memory Manager updates the Node Map during the startup and runtime as follows.
The administrator must provide --reserved-memory flag when Static policy is configured.
Runtime
Reference Memory Manager KEP: Memory Maps at runtime (with examples) illustrates
how a successful pod deployment affects the Node Map, and it also relates to
how potential Out-of-Memory (OOM) situations are handled further by Kubernetes or operating system.
Other Managers should be first pre-configured. Next, the Memory Manger feature should be enabled
and be run with Static policy (section Static policy).
Optionally, some amount of memory can be reserved for system or kubelet processes to increase
node stability (section Reserved memory flag).
Policies
Memory Manager supports two policies. You can select a policy via a kubelet flag --memory-manager-policy:
None (default)
Static
None policy
This is the default policy and does not affect the memory allocation in any way.
It acts the same as if the Memory Manager is not present at all.
The None policy returns default topology hint. This special hint denotes that Hint Provider
(Memory Manger in this case) has no preference for NUMA affinity with any resource.
Static policy
In the case of the Guaranteed pod, the Static Memory Manger policy returns topology hints
relating to the set of NUMA nodes where the memory can be guaranteed,
and reserves the memory through updating the internal NodeMap object.
In the case of the BestEffort or Burstable pod, the Static Memory Manager policy sends back
the default topology hint as there is no request for the guaranteed memory,
and does not reserve the memory in the internal NodeMap object.
Reserved memory flag
The Node Allocatable mechanism
is commonly used by node administrators to reserve K8S node system resources for the kubelet
or operating system processes in order to enhance the node stability.
A dedicated set of flags can be used for this purpose to set the total amount of reserved memory
for a node. This pre-configured value is subsequently utilized to calculate
the real amount of node's "allocatable" memory available to pods.
The Kubernetes scheduler incorporates "allocatable" to optimise pod scheduling process.
The foregoing flags include --kube-reserved, --system-reserved and --eviction-threshold.
The sum of their values will account for the total amount of reserved memory.
A new --reserved-memory flag was added to Memory Manager to allow for this total reserved memory
to be split (by a node administrator) and accordingly reserved across many NUMA nodes.
The flag specifies a comma-separated list of memory reservations of different memory types per NUMA node.
Memory reservations across multiple NUMA nodes can be specified using semicolon as separator.
This parameter is only useful in the context of the Memory Manager feature.
The Memory Manager will not use this reserved memory for the allocation of container workloads.
For example, if you have a NUMA node "NUMA0" with 10Gi of memory available, and
the --reserved-memory was specified to reserve 1Gi of memory at "NUMA0",
the Memory Manager assumes that only 9Gi is available for containers.
You can omit this parameter, however, you should be aware that the quantity of reserved memory
from all NUMA nodes should be equal to the quantity of memory specified by the
Node Allocatable feature.
If at least one node allocatable parameter is non-zero, you will need to specify
--reserved-memory for at least one NUMA node.
In fact, eviction-hard threshold value is equal to 100Mi by default, so
if Static policy is used, --reserved-memory is obligatory.
Also, avoid the following configurations:
duplicates, i.e. the same NUMA node or memory type, but with a different value;
setting zero limit for any of memory types;
NUMA node IDs that do not exist in the machine hardware;
memory type names different than memory or hugepages-<size>
(hugepages of particular <size> should also exist).
When you specify values for --reserved-memory flag, you must comply with the setting that
you prior provided via Node Allocatable Feature flags.
That is, the following rule must be obeyed for each memory type:
An example of kubelet command-line arguments relevant to the node Allocatable configuration:
--kube-reserved=cpu=500m,memory=50Mi
--system-reserved=cpu=123m,memory=333Mi
--eviction-hard=memory.available<500Mi
Note: The default hard eviction threshold is 100MiB, and not zero.
Remember to increase the quantity of memory that you reserve by setting --reserved-memory
by that hard eviction threshold. Otherwise, the kubelet will not start Memory Manager and
display an error.
If the selected policy is anything other than None, the Memory Manager identifies pods
that are in the Guaranteed QoS class.
The Memory Manager provides specific topology hints to the Topology Manager for each Guaranteed pod.
For pods in a QoS class other than Guaranteed, the Memory Manager provides default topology hints
to the Topology Manager.
The following excerpts from pod manifests assign a pod to the Guaranteed QoS class.
Pod with integer CPU(s) runs in the Guaranteed QoS class, when requests are equal to limits:
Notice that both CPU and memory requests must be specified for a Pod to lend it to Guaranteed QoS class.
Troubleshooting
The following means can be used to troubleshoot the reason why a pod could not be deployed or
became rejected at a node:
pod status - indicates topology affinity errors
system logs - include valuable information for debugging, e.g., about generated hints
state file - the dump of internal state of the Memory Manager
(includes Node Map and Memory Maps)
starting from v1.22, the device plugin resource API can be used
to retrieve information about the memory reserved for containers
Pod status (TopologyAffinityError)
This error typically occurs in the following situations:
a node has not enough resources available to satisfy the pod's request
the pod's request is rejected due to particular Topology Manager policy constraints
The error appears in the status of a pod:
kubectl get pods
NAME READY STATUS RESTARTS AGE
guaranteed 0/1 TopologyAffinityError 0 113s
Use kubectl describe pod <id> or kubectl get events to obtain detailed error message:
Warning TopologyAffinityError 10m kubelet, dell8 Resources cannot be allocated with Topology locality
System logs
Search system logs with respect to a particular pod.
The set of hints that Memory Manager generated for the pod can be found in the logs.
Also, the set of hints generated by CPU Manager should be present in the logs.
Topology Manager merges these hints to calculate a single best hint.
The best hint should be also present in the logs.
The best hint indicates where to allocate all the resources.
Topology Manager tests this hint against its current policy, and based on the verdict,
it either admits the pod to the node or rejects it.
Also, search the logs for occurrences associated with the Memory Manager,
e.g. to find out information about cgroups and cpuset.mems updates.
Examine the memory manager state on a node
Let us first deploy a sample Guaranteed pod whose specification is as follows:
It can be deduced from the state file that the pod was pinned to both NUMA nodes, i.e.:
"numaAffinity":[
0,
1],
Pinned term means that pod's memory consumption is constrained (through cgroups configuration)
to these NUMA nodes.
This automatically implies that Memory Manager instantiated a new group that
comprises these two NUMA nodes, i.e. 0 and 1 indexed NUMA nodes.
Notice that the management of groups is handled in a relatively complex manner, and
further elaboration is provided in Memory Manager KEP in this and this sections.
In order to analyse memory resources available in a group,the corresponding entries from
NUMA nodes belonging to the group must be added up.
For example, the total amount of free "conventional" memory in the group can be computed
by adding up the free memory available at every NUMA node in the group,
i.e., in the "memory" section of NUMA node 0 ("free":0) and NUMA node 1 ("free":103739236352).
So, the total amount of free "conventional" memory in this group is equal to 0 + 103739236352 bytes.
The line "systemReserved":3221225472 indicates that the administrator of this node reserved
3221225472 bytes (i.e. 3Gi) to serve kubelet and system processes at NUMA node 0,
by using --reserved-memory flag.
Device plugin resource API
By employing the API,
the information about reserved memory for each container can be retrieved, which is contained
in protobuf ContainerMemory message.
This information can be retrieved solely for pods in Guaranteed QoS class.
These instructions are for Kubernetes 1.25. If you want
to check the integrity of components for a different version of Kubernetes,
check the documentation for that Kubernetes release.
You will need to have the following tools installed:
Note:COSIGN_EXPERIMENTAL=1 is used to allow verification of images signed
in KEYLESS mode. To learn more about keyless signing, please refer to
Keyless Signatures
.
Verifying images for all control plane components
To verify all signed control plane images, please run this command:
Once you have verified an image, specify that image by its digest in your Pod
manifests as per this
example: registry-url/image-name@sha256:45b23dee08af5e43a7fea6c4cf9c25ccf269ee113168c19722f87876677c5cb2
.
Verifying Image Signatures with Admission Controller
For non-control plane images (
e.g. conformance image)
, signatures can also be verified at deploy time using
sigstore policy-controller
admission controller. To get started with policy-controller here are a few helpful
resources:
Perform common configuration tasks for Pods and containers.
4.3.1 - Assign Memory Resources to Containers and Pods
This page shows how to assign a memory request and a memory limit to a
Container. A Container is guaranteed to have as much memory as it requests,
but is not allowed to use more memory than its limit.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Each node in your cluster must have at least 300 MiB of memory.
A few of the steps on this page require you to run the
metrics-server
service in your cluster. If you have the metrics-server
running, you can skip those steps.
If you are running Minikube, run the following command to enable the
metrics-server:
minikube addons enable metrics-server
To see whether the metrics-server is running, or another provider of the resource metrics
API (metrics.k8s.io), run the following command:
kubectl get apiservices
If the resource metrics API is available, the output includes a
reference to metrics.k8s.io.
NAME
v1beta1.metrics.k8s.io
Create a namespace
Create a namespace so that the resources you create in this exercise are
isolated from the rest of your cluster.
kubectl create namespace mem-example
Specify a memory request and a memory limit
To specify a memory request for a Container, include the resources:requests field
in the Container's resource manifest. To specify a memory limit, include resources:limits.
In this exercise, you create a Pod that has one Container. The Container has a memory
request of 100 MiB and a memory limit of 200 MiB. Here's the configuration file
for the Pod:
The args section in the configuration file provides arguments for the Container when it starts.
The "--vm-bytes", "150M" arguments tell the Container to attempt to allocate 150 MiB of memory.
kubectl top pod memory-demo --namespace=mem-example
The output shows that the Pod is using about 162,900,000 bytes of memory, which
is about 150 MiB. This is greater than the Pod's 100 MiB request, but within the
Pod's 200 MiB limit.
NAME CPU(cores) MEMORY(bytes)
memory-demo <something> 162856960
Delete your Pod:
kubectl delete pod memory-demo --namespace=mem-example
Exceed a Container's memory limit
A Container can exceed its memory request if the Node has memory available. But a Container
is not allowed to use more than its memory limit. If a Container allocates more memory than
its limit, the Container becomes a candidate for termination. If the Container continues to
consume memory beyond its limit, the Container is terminated. If a terminated Container can be
restarted, the kubelet restarts it, as with any other type of runtime failure.
In this exercise, you create a Pod that attempts to allocate more memory than its limit.
Here is the configuration file for a Pod that has one Container with a
memory request of 50 MiB and a memory limit of 100 MiB:
In the args section of the configuration file, you can see that the Container
will attempt to allocate 250 MiB of memory, which is well above the 100 MiB limit.
The Container in this exercise can be restarted, so the kubelet restarts it. Repeat
this command several times to see that the Container is repeatedly killed and restarted:
kubectl get pod memory-demo-2 --namespace=mem-example
The output shows that the Container is killed, restarted, killed again, restarted again, and so on:
kubectl get pod memory-demo-2 --namespace=mem-example
NAME READY STATUS RESTARTS AGE
memory-demo-2 0/1 OOMKilled 1 37s
kubectl get pod memory-demo-2 --namespace=mem-example
NAME READY STATUS RESTARTS AGE
memory-demo-2 1/1 Running 2 40s
View detailed information about the Pod history:
kubectl describe pod memory-demo-2 --namespace=mem-example
The output shows that the Container starts and fails repeatedly:
... Normal Created Created container with id 66a3a20aa7980e61be4922780bf9d24d1a1d8b7395c09861225b0eba1b1f8511
... Warning BackOff Back-off restarting failed container
View detailed information about your cluster's Nodes:
kubectl describe nodes
The output includes a record of the Container being killed because of an out-of-memory condition:
Warning OOMKilling Memory cgroup out of memory: Kill process 4481 (stress) score 1994 or sacrifice child
Delete your Pod:
kubectl delete pod memory-demo-2 --namespace=mem-example
Specify a memory request that is too big for your Nodes
Memory requests and limits are associated with Containers, but it is useful to think
of a Pod as having a memory request and limit. The memory request for the Pod is the
sum of the memory requests for all the Containers in the Pod. Likewise, the memory
limit for the Pod is the sum of the limits of all the Containers in the Pod.
Pod scheduling is based on requests. A Pod is scheduled to run on a Node only if the Node
has enough available memory to satisfy the Pod's memory request.
In this exercise, you create a Pod that has a memory request so big that it exceeds the
capacity of any Node in your cluster. Here is the configuration file for a Pod that has one
Container with a request for 1000 GiB of memory, which likely exceeds the capacity
of any Node in your cluster.
kubectl get pod memory-demo-3 --namespace=mem-example
The output shows that the Pod status is PENDING. That is, the Pod is not scheduled to run on any Node, and it will remain in the PENDING state indefinitely:
kubectl get pod memory-demo-3 --namespace=mem-example
NAME READY STATUS RESTARTS AGE
memory-demo-3 0/1 Pending 0 25s
View detailed information about the Pod, including events:
kubectl describe pod memory-demo-3 --namespace=mem-example
The output shows that the Container cannot be scheduled because of insufficient memory on the Nodes:
Events:
... Reason Message
------ -------
... FailedScheduling No nodes are available that match all of the following predicates:: Insufficient memory (3).
Memory units
The memory resource is measured in bytes. You can express memory as a plain integer or a
fixed-point integer with one of these suffixes: E, P, T, G, M, K, Ei, Pi, Ti, Gi, Mi, Ki.
For example, the following represent approximately the same value:
128974848, 129e6, 129M, 123Mi
Delete your Pod:
kubectl delete pod memory-demo-3 --namespace=mem-example
If you do not specify a memory limit
If you do not specify a memory limit for a Container, one of the following situations applies:
The Container has no upper bound on the amount of memory it uses. The Container
could use all of the memory available on the Node where it is running which in turn could invoke the OOM Killer. Further, in case of an OOM Kill, a container with no resource limits will have a greater chance of being killed.
The Container is running in a namespace that has a default memory limit, and the
Container is automatically assigned the default limit. Cluster administrators can use a
LimitRange
to specify a default value for the memory limit.
Motivation for memory requests and limits
By configuring memory requests and limits for the Containers that run in your
cluster, you can make efficient use of the memory resources available on your cluster's
Nodes. By keeping a Pod's memory request low, you give the Pod a good chance of being
scheduled. By having a memory limit that is greater than the memory request, you accomplish two things:
The Pod can have bursts of activity where it makes use of memory that happens to be available.
The amount of memory a Pod can use during a burst is limited to some reasonable amount.
Clean up
Delete your namespace. This deletes all the Pods that you created for this task:
4.3.2 - Assign CPU Resources to Containers and Pods
This page shows how to assign a CPU request and a CPU limit to
a container. Containers cannot use more CPU than the configured limit.
Provided the system has CPU time free, a container is guaranteed to be
allocated as much CPU as it requests.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your cluster must have at least 1 CPU available for use to run the task examples.
A few of the steps on this page require you to run the
metrics-server
service in your cluster. If you have the metrics-server
running, you can skip those steps.
If you are running Minikube, run the
following command to enable metrics-server:
minikube addons enable metrics-server
To see whether metrics-server (or another provider of the resource metrics
API, metrics.k8s.io) is running, type the following command:
kubectl get apiservices
If the resource metrics API is available, the output will include a
reference to metrics.k8s.io.
NAME
v1beta1.metrics.k8s.io
Create a namespace
Create a Namespace so that the resources you
create in this exercise are isolated from the rest of your cluster.
kubectl create namespace cpu-example
Specify a CPU request and a CPU limit
To specify a CPU request for a container, include the resources:requests field
in the Container resource manifest. To specify a CPU limit, include resources:limits.
In this exercise, you create a Pod that has one container. The container has a request
of 0.5 CPU and a limit of 1 CPU. Here is the configuration file for the Pod:
The args section of the configuration file provides arguments for the container when it starts.
The -cpus "2" argument tells the Container to attempt to use 2 CPUs.
kubectl get pod cpu-demo --output=yaml --namespace=cpu-example
The output shows that the one container in the Pod has a CPU request of 500 milliCPU
and a CPU limit of 1 CPU.
resources:limits:cpu:"1"requests:cpu:500m
Use kubectl top to fetch the metrics for the pod:
kubectl top pod cpu-demo --namespace=cpu-example
This example output shows that the Pod is using 974 milliCPU, which is
slightly less than the limit of 1 CPU specified in the Pod configuration.
NAME CPU(cores) MEMORY(bytes)
cpu-demo 974m <something>
Recall that by setting -cpu "2", you configured the Container to attempt to use 2 CPUs, but the Container is only being allowed to use about 1 CPU. The container's CPU use is being throttled, because the container is attempting to use more CPU resources than its limit.
Note: Another possible explanation for the CPU use being below 1.0 is that the Node might not have
enough CPU resources available. Recall that the prerequisites for this exercise require your cluster to have at least 1 CPU available for use. If your Container runs on a Node that has only 1 CPU, the Container cannot use more than 1 CPU regardless of the CPU limit specified for the Container.
CPU units
The CPU resource is measured in CPU units. One CPU, in Kubernetes, is equivalent to:
1 AWS vCPU
1 GCP Core
1 Azure vCore
1 Hyperthread on a bare-metal Intel processor with Hyperthreading
Fractional values are allowed. A Container that requests 0.5 CPU is guaranteed half as much
CPU as a Container that requests 1 CPU. You can use the suffix m to mean milli. For example
100m CPU, 100 milliCPU, and 0.1 CPU are all the same. Precision finer than 1m is not allowed.
CPU is always requested as an absolute quantity, never as a relative quantity; 0.1 is the same
amount of CPU on a single-core, dual-core, or 48-core machine.
Delete your Pod:
kubectl delete pod cpu-demo --namespace=cpu-example
Specify a CPU request that is too big for your Nodes
CPU requests and limits are associated with Containers, but it is useful to think
of a Pod as having a CPU request and limit. The CPU request for a Pod is the sum
of the CPU requests for all the Containers in the Pod. Likewise, the CPU limit for
a Pod is the sum of the CPU limits for all the Containers in the Pod.
Pod scheduling is based on requests. A Pod is scheduled to run on a Node only if
the Node has enough CPU resources available to satisfy the Pod CPU request.
In this exercise, you create a Pod that has a CPU request so big that it exceeds
the capacity of any Node in your cluster. Here is the configuration file for a Pod
that has one Container. The Container requests 100 CPU, which is likely to exceed the
capacity of any Node in your cluster.
kubectl get pod cpu-demo-2 --namespace=cpu-example
The output shows that the Pod status is Pending. That is, the Pod has not been
scheduled to run on any Node, and it will remain in the Pending state indefinitely:
NAME READY STATUS RESTARTS AGE
cpu-demo-2 0/1 Pending 0 7m
View detailed information about the Pod, including events:
kubectl describe pod cpu-demo-2 --namespace=cpu-example
The output shows that the Container cannot be scheduled because of insufficient
CPU resources on the Nodes:
Events:
Reason Message
------ -------
FailedScheduling No nodes are available that match all of the following predicates:: Insufficient cpu (3).
Delete your Pod:
kubectl delete pod cpu-demo-2 --namespace=cpu-example
If you do not specify a CPU limit
If you do not specify a CPU limit for a Container, then one of these situations applies:
The Container has no upper bound on the CPU resources it can use. The Container
could use all of the CPU resources available on the Node where it is running.
The Container is running in a namespace that has a default CPU limit, and the
Container is automatically assigned the default limit. Cluster administrators can use a
LimitRange
to specify a default value for the CPU limit.
If you specify a CPU limit but do not specify a CPU request
If you specify a CPU limit for a Container but do not specify a CPU request, Kubernetes automatically
assigns a CPU request that matches the limit. Similarly, if a Container specifies its own memory limit,
but does not specify a memory request, Kubernetes automatically assigns a memory request that matches
the limit.
Motivation for CPU requests and limits
By configuring the CPU requests and limits of the Containers that run in your
cluster, you can make efficient use of the CPU resources available on your cluster
Nodes. By keeping a Pod CPU request low, you give the Pod a good chance of being
scheduled. By having a CPU limit that is greater than the CPU request, you accomplish two things:
The Pod can have bursts of activity where it makes use of CPU resources that happen to be available.
The amount of CPU resources a Pod can use during a burst is limited to some reasonable amount.
4.3.3 - Configure GMSA for Windows Pods and containers
FEATURE STATE:Kubernetes v1.18 [stable]
This page shows how to configure Group Managed Service Accounts (GMSA) for Pods and containers that will run on Windows nodes. Group Managed Service Accounts are a specific type of Active Directory account that provides automatic password management, simplified service principal name (SPN) management, and the ability to delegate the management to other administrators across multiple servers.
In Kubernetes, GMSA credential specs are configured at a Kubernetes cluster-wide scope as Custom Resources. Windows Pods, as well as individual containers within a Pod, can be configured to use a GMSA for domain based functions (e.g. Kerberos authentication) when interacting with other Windows services.
Before you begin
You need to have a Kubernetes cluster and the kubectl command-line tool must be configured to communicate with your cluster. The cluster is expected to have Windows worker nodes. This section covers a set of initial steps required once for each cluster:
Install the GMSACredentialSpec CRD
A CustomResourceDefinition(CRD) for GMSA credential spec resources needs to be configured on the cluster to define the custom resource type GMSACredentialSpec. Download the GMSA CRD YAML and save it as gmsa-crd.yaml.
Next, install the CRD with kubectl apply -f gmsa-crd.yaml
Install webhooks to validate GMSA users
Two webhooks need to be configured on the Kubernetes cluster to populate and validate GMSA credential spec references at the Pod or container level:
A mutating webhook that expands references to GMSAs (by name from a Pod specification) into the full credential spec in JSON form within the Pod spec.
A validating webhook ensures all references to GMSAs are authorized to be used by the Pod service account.
Installing the above webhooks and associated objects require the steps below:
Create a certificate key pair (that will be used to allow the webhook container to communicate to the cluster)
Install a secret with the certificate from above.
Create a deployment for the core webhook logic.
Create the validating and mutating webhook configurations referring to the deployment.
A script can be used to deploy and configure the GMSA webhooks and associated objects mentioned above. The script can be run with a --dry-run=server option to allow you to review the changes that would be made to your cluster.
The YAML template used by the script may also be used to deploy the webhooks and associated objects manually (with appropriate substitutions for the parameters)
Configure GMSAs and Windows nodes in Active Directory
Before Pods in Kubernetes can be configured to use GMSAs, the desired GMSAs need to be provisioned in Active Directory as described in the Windows GMSA documentation. Windows worker nodes (that are part of the Kubernetes cluster) need to be configured in Active Directory to access the secret credentials associated with the desired GMSA as described in the Windows GMSA documentation
Create GMSA credential spec resources
With the GMSACredentialSpec CRD installed (as described earlier), custom resources containing GMSA credential specs can be configured. The GMSA credential spec does not contain secret or sensitive data. It is information that a container runtime can use to describe the desired GMSA of a container to Windows. GMSA credential specs can be generated in YAML format with a utility PowerShell script.
Following are the steps for generating a GMSA credential spec YAML manually in JSON format and then converting it:
Import the CredentialSpec module: ipmo CredentialSpec.psm1
Create a credential spec in JSON format using New-CredentialSpec. To create a GMSA credential spec named WebApp1, invoke New-CredentialSpec -Name WebApp1 -AccountName WebApp1 -Domain $(Get-ADDomain -Current LocalComputer)
Use Get-CredentialSpec to show the path of the JSON file.
Convert the credspec file from JSON to YAML format and apply the necessary header fields apiVersion, kind, metadata and credspec to make it a GMSACredentialSpec custom resource that can be configured in Kubernetes.
The following YAML configuration describes a GMSA credential spec named gmsa-WebApp1:
apiVersion:windows.k8s.io/v1kind:GMSACredentialSpecmetadata:name:gmsa-WebApp1 #This is an arbitrary name but it will be used as a referencecredspec:ActiveDirectoryConfig:GroupManagedServiceAccounts:- Name:WebApp1 #Username of the GMSA accountScope:CONTOSO #NETBIOS Domain Name- Name:WebApp1 #Username of the GMSA accountScope:contoso.com#DNS Domain NameCmsPlugins:- ActiveDirectoryDomainJoinConfig:DnsName:contoso.com #DNS Domain NameDnsTreeName:contoso.com#DNS Domain Name RootGuid:244818ae-87ac-4fcd-92ec-e79e5252348a #GUIDMachineAccountName:WebApp1#Username of the GMSA accountNetBiosName:CONTOSO #NETBIOS Domain NameSid:S-1-5-21-2126449477-2524075714-3094792973#SID of GMSA
The above credential spec resource may be saved as gmsa-Webapp1-credspec.yaml and applied to the cluster using: kubectl apply -f gmsa-Webapp1-credspec.yml
Configure cluster role to enable RBAC on specific GMSA credential specs
A cluster role needs to be defined for each GMSA credential spec resource. This authorizes the use verb on a specific GMSA resource by a subject which is typically a service account. The following example shows a cluster role that authorizes usage of the gmsa-WebApp1 credential spec from above. Save the file as gmsa-webapp1-role.yaml and apply using kubectl apply -f gmsa-webapp1-role.yaml
#Create the Role to read the credspecapiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:webapp1-rolerules:- apiGroups:["windows.k8s.io"]resources:["gmsacredentialspecs"]verbs:["use"]resourceNames:["gmsa-WebApp1"]
Assign role to service accounts to use specific GMSA credspecs
A service account (that Pods will be configured with) needs to be bound to the cluster role create above. This authorizes the service account to use the desired GMSA credential spec resource. The following shows the default service account being bound to a cluster role webapp1-role to use gmsa-WebApp1 credential spec resource created above.
Configure GMSA credential spec reference in Pod spec
The Pod spec field securityContext.windowsOptions.gmsaCredentialSpecName is used to specify references to desired GMSA credential spec custom resources in Pod specs. This configures all containers in the Pod spec to use the specified GMSA. A sample Pod spec with the annotation populated to refer to gmsa-WebApp1:
Individual containers in a Pod spec can also specify the desired GMSA credspec using a per-container securityContext.windowsOptions.gmsaCredentialSpecName field. For example:
As Pod specs with GMSA fields populated (as described above) are applied in a cluster, the following sequence of events take place:
The mutating webhook resolves and expands all references to GMSA credential spec resources to the contents of the GMSA credential spec.
The validating webhook ensures the service account associated with the Pod is authorized for the use verb on the specified GMSA credential spec.
The container runtime configures each Windows container with the specified GMSA credential spec so that the container can assume the identity of the GMSA in Active Directory and access services in the domain using that identity.
Authenticating to network shares using hostname or FQDN
If you are experiencing issues connecting to SMB shares from Pods using hostname or FQDN, but are able to access the shares via their IPv4 address then make sure the following registry key is set on the Windows nodes.
Running Pods will then need to be recreated to pick up the behavior changes.
More information on how this registry key is used can be found here
Troubleshooting
If you are having difficulties getting GMSA to work in your environment, there are a few troubleshooting steps you can take.
First, make sure the credspec has been passed to the Pod. To do this you will need to exec into one of your Pods and check the output of the nltest.exe /parentdomain command.
In the example below the Pod did not get the credspec correctly:
nltest.exe /parentdomain results in the following error:
Getting parent domain failed: Status = 1722 0x6ba RPC_S_SERVER_UNAVAILABLE
If your Pod did get the credspec correctly, then next check communication with the domain. First, from inside of your Pod, quickly do an nslookup to find the root of your domain.
This will tell us 3 things:
The Pod can reach the DC
The DC can reach the Pod
DNS is working correctly.
If the DNS and communication test passes, next you will need to check if the Pod has established secure channel communication with the domain. To do this, again, exec into your Pod and run the nltest.exe /query command.
nltest.exe /query
Results in the following output:
I_NetLogonControl failed: Status = 1722 0x6ba RPC_S_SERVER_UNAVAILABLE
This tells us that for some reason, the Pod was unable to logon to the domain using the account specified in the credspec. You can try to repair the secure channel by running the following:
nltest /sc_reset:domain.example
If the command is successful you will see and output similar to this:
Flags: 30 HAS_IP HAS_TIMESERV
Trusted DC Name \\dc10.domain.example
Trusted DC Connection Status Status = 0 0x0 NERR_Success
The command completed successfully
If the above corrects the error, you can automate the step by adding the following lifecycle hook to your Pod spec. If it did not correct the error, you will need to examine your credspec again and confirm that it is correct and complete.
If you add the lifecycle section show above to your Pod spec, the Pod will execute the commands listed to restart the netlogon service until the nltest.exe /query command exits without error.
4.3.4 - Configure RunAsUserName for Windows pods and containers
FEATURE STATE:Kubernetes v1.18 [stable]
This page shows how to use the runAsUserName setting for Pods and containers that will run on Windows nodes. This is roughly equivalent of the Linux-specific runAsUser setting, allowing you to run applications in a container as a different username than the default.
Before you begin
You need to have a Kubernetes cluster and the kubectl command-line tool must be configured to communicate with your cluster. The cluster is expected to have Windows worker nodes where pods with containers running Windows workloads will get scheduled.
Set the Username for a Pod
To specify the username with which to execute the Pod's container processes, include the securityContext field (PodSecurityContext) in the Pod specification, and within it, the windowsOptions (WindowsSecurityContextOptions) field containing the runAsUserName field.
The Windows security context options that you specify for a Pod apply to all Containers and init Containers in the Pod.
Here is a configuration file for a Windows Pod that has the runAsUserName field set:
Check that the shell is running user the correct username:
echo $env:USERNAME
The output should be:
ContainerUser
Set the Username for a Container
To specify the username with which to execute a Container's processes, include the securityContext field (SecurityContext) in the Container manifest, and within it, the windowsOptions (WindowsSecurityContextOptions) field containing the runAsUserName field.
The Windows security context options that you specify for a Container apply only to that individual Container, and they override the settings made at the Pod level.
Here is the configuration file for a Pod that has one Container, and the runAsUserName field is set at the Pod level and the Container level:
Check that the shell is running user the correct username (the one set at the Container level):
echo $env:USERNAME
The output should be:
ContainerAdministrator
Windows Username limitations
In order to use this feature, the value set in the runAsUserName field must be a valid username. It must have the following format: DOMAIN\USER, where DOMAIN\ is optional. Windows user names are case insensitive. Additionally, there are some restrictions regarding the DOMAIN and USER:
The runAsUserName field cannot be empty, and it cannot contain control characters (ASCII values: 0x00-0x1F, 0x7F)
The DOMAIN must be either a NetBios name, or a DNS name, each with their own restrictions:
NetBios names: maximum 15 characters, cannot start with . (dot), and cannot contain the following characters: \ / : * ? " < > |
DNS names: maximum 255 characters, contains only alphanumeric characters, dots, and dashes, and it cannot start or end with a . (dot) or - (dash).
The USER must have at most 20 characters, it cannot contain only dots or spaces, and it cannot contain the following characters: " / \ [ ] : ; | = , + * ? < > @.
Examples of acceptable values for the runAsUserName field: ContainerAdministrator, ContainerUser, NT AUTHORITY\NETWORK SERVICE, NT AUTHORITY\LOCAL SERVICE.
For more information about these limtations, check here and here.
Windows HostProcess containers enable you to run containerized
workloads on a Windows host. These containers operate as
normal processes but have access to the host network namespace,
storage, and devices when given the appropriate user privileges.
HostProcess containers can be used to deploy network plugins,
storage configurations, device plugins, kube-proxy, and other
components to Windows nodes without the need for dedicated proxies or
the direct installation of host services.
Administrative tasks such as installation of security patches, event
log collection, and more can be performed without requiring cluster operators to
log onto each Windows node. HostProcess containers can run as any user that is
available on the host or is in the domain of the host machine, allowing administrators
to restrict resource access through user permissions. While neither filesystem or process
isolation are supported, a new volume is created on the host upon starting the container
to give it a clean and consolidated workspace. HostProcess containers can also be built on
top of existing Windows base images and do not inherit the same
compatibility requirements
as Windows server containers, meaning that the version of the base images does not need
to match that of the host. It is, however, recommended that you use the same base image
version as your Windows Server container workloads to ensure you do not have any unused
images taking up space on the node. HostProcess containers also support
volume mounts within the container volume.
When should I use a Windows HostProcess container?
When you need to perform tasks which require the networking namespace of the host.
HostProcess containers have access to the host's network interfaces and IP addresses.
You need access to resources on the host such as the filesystem, event logs, etc.
Installation of specific device drivers or Windows services.
Consolidation of administrative tasks and security policies. This reduces the degree of
privileges needed by Windows nodes.
Before you begin
This task guide is specific to Kubernetes v1.25.
If you are not running Kubernetes v1.25, check the documentation for
that version of Kubernetes.
In Kubernetes 1.25, the HostProcess container feature is enabled by default. The kubelet will
communicate with containerd directly by passing the hostprocess flag via CRI. You can use the
latest version of containerd (v1.6+) to run HostProcess containers.
How to install containerd.
To disable HostProcess containers you need to pass the following feature gate flag to the
kubelet and kube-apiserver:
These limitations are relevant for Kubernetes v1.25:
HostProcess containers require containerd 1.6 or higher
container runtime.
HostProcess pods can only contain HostProcess containers. This is a current limitation
of the Windows OS; non-privileged Windows containers cannot share a vNIC with the host IP namespace.
HostProcess containers run as a process on the host and do not have any degree of
isolation other than resource constraints imposed on the HostProcess user account. Neither
filesystem or Hyper-V isolation are supported for HostProcess containers.
Volume mounts are supported and are mounted under the container volume. See
Volume Mounts
A limited set of host user accounts are available for HostProcess containers by default.
See Choosing a User Account.
Resource limits (disk, memory, cpu count) are supported in the same fashion as processes
on the host.
Both Named pipe mounts and Unix domain sockets are not supported and should instead
be accessed via their path on the host (e.g. \\.\pipe\*)
HostProcess Pod configuration requirements
Enabling a Windows HostProcess pod requires setting the right configurations in the pod security
configuration. Of the policies defined in the Pod Security Standards
HostProcess pods are disallowed by the baseline and restricted policies. It is therefore recommended
that HostProcess pods run in alignment with the privileged profile.
When running under the privileged policy, here are
the configurations which need to be set to enable the creation of a HostProcess pod:
Because HostProcess containers have privileged access to the host, the runAsNonRoot field cannot be set to true.
Allowed Values
Undefined/Nil
false
Example manifest (excerpt)
spec:securityContext:windowsOptions:hostProcess:truerunAsUserName:"NT AUTHORITY\\Local service"hostNetwork:truecontainers:- name:testimage:image1:latestcommand:- ping- -t- 127.0.0.1nodeSelector:"kubernetes.io/os": windows
Volume mounts
HostProcess containers support the ability to mount volumes within the container volume space.
Applications running inside the container can access volume mounts directly via relative or
absolute paths. An environment variable $CONTAINER_SANDBOX_MOUNT_POINT is set upon container
creation and provides the absolute host path to the container volume. Relative paths are based
upon the .spec.containers.volumeMounts.mountPath configuration.
Example
To access service account tokens the following path structures are supported within the container:
Resource limits (disk, memory, cpu count) are applied to the job and are job wide.
For example, with a limit of 10MB set, the memory allocated for any HostProcess job object
will be capped at 10MB. This is the same behavior as other Windows container types.
These limits would be specified the same way they are currently for whatever orchestrator
or runtime is being used. The only difference is in the disk resource usage calculation
used for resource tracking due to the difference in how HostProcess containers are bootstrapped.
Choosing a user account
HostProcess containers support the ability to run as one of three supported Windows service accounts:
You should select an appropriate Windows service account for each HostProcess
container, aiming to limit the degree of privileges so as to avoid accidental (or even
malicious) damage to the host. The LocalSystem service account has the highest level
of privilege of the three and should be used only if absolutely necessary. Where possible,
use the LocalService service account as it is the least privileged of the three options.
Troubleshooting HostProcess containers
HostProcess containers fail to start with failed to create user process token: failed to logon user: Access is denied.: unknown
Ensure containerd is running as LocalSystem or LocalService service accounts. User accounts (even Administrator accounts) do not have permissions to create logon tokens for any of the supported user accounts.
4.3.6 - Configure Quality of Service for Pods
This page shows how to configure Pods so that they will be assigned particular
Quality of Service (QoS) classes. Kubernetes uses QoS classes to make decisions about
scheduling and evicting Pods.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
When Kubernetes creates a Pod it assigns one of these QoS classes to the Pod:
Guaranteed
Burstable
BestEffort
Create a namespace
Create a namespace so that the resources you create in this exercise are
isolated from the rest of your cluster.
kubectl create namespace qos-example
Create a Pod that gets assigned a QoS class of Guaranteed
For a Pod to be given a QoS class of Guaranteed:
Every Container in the Pod must have a memory limit and a memory request.
For every Container in the Pod, the memory limit must equal the memory request.
Every Container in the Pod must have a CPU limit and a CPU request.
For every Container in the Pod, the CPU limit must equal the CPU request.
These restrictions apply to init containers and app containers equally.
Here is the configuration file for a Pod that has one Container. The Container has a memory limit and a
memory request, both equal to 200 MiB. The Container has a CPU limit and a CPU request, both equal to 700 milliCPU:
kubectl get pod qos-demo --namespace=qos-example --output=yaml
The output shows that Kubernetes gave the Pod a QoS class of Guaranteed. The output also
verifies that the Pod Container has a memory request that matches its memory limit, and it has
a CPU request that matches its CPU limit.
Note: If a Container specifies its own memory limit, but does not specify a memory request, Kubernetes
automatically assigns a memory request that matches the limit. Similarly, if a Container specifies its own
CPU limit, but does not specify a CPU request, Kubernetes automatically assigns a CPU request that matches
the limit.
Delete your Pod:
kubectl delete pod qos-demo --namespace=qos-example
Create a Pod that gets assigned a QoS class of Burstable
A Pod is given a QoS class of Burstable if:
The Pod does not meet the criteria for QoS class Guaranteed.
At least one Container in the Pod has a memory or CPU request or limit.
Here is the configuration file for a Pod that has one Container. The Container has a memory limit of 200 MiB
and a memory request of 100 MiB.
kubectl delete pod qos-demo-3 --namespace=qos-example
Create a Pod that has two Containers
Here is the configuration file for a Pod that has two Containers. One container specifies a memory
request of 200 MiB. The other Container does not specify any requests or limits.
Notice that this Pod meets the criteria for QoS class Burstable. That is, it does not meet the
criteria for QoS class Guaranteed, and one of its Containers has a memory request.
This page shows how to assign extended resources to a Container.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Before you do this exercise, do the exercise in
Advertise Extended Resources for a Node.
That will configure one of your Nodes to advertise a dongle resource.
Assign an extended resource to a Pod
To request an extended resource, include the resources:requests field in your
Container manifest. Extended resources are fully qualified with any domain outside of
*.kubernetes.io/. Valid extended resource names have the form example.com/foo where
example.com is replaced with your organization's domain and foo is a
descriptive resource name.
Here is the configuration file for a Pod that has one Container:
The output shows that the Pod cannot be scheduled, because there is no Node that has
2 dongles available:
Conditions:
Type Status
PodScheduled False
...
Events:
...
... Warning FailedScheduling pod (extended-resource-demo-2) failed to fit in any node
fit failure summary on nodes : Insufficient example.com/dongle (1)
View the Pod status:
kubectl get pod extended-resource-demo-2
The output shows that the Pod was created, but not scheduled to run on a Node.
It has a status of Pending:
NAME READY STATUS RESTARTS AGEextended-resource-demo-2 0/1 Pending 0 6m
Clean up
Delete the Pods that you created for this exercise:
kubectl delete pod extended-resource-demo
kubectl delete pod extended-resource-demo-2
4.3.8 - Configure a Pod to Use a Volume for Storage
This page shows how to configure a Pod to use a Volume for storage.
A Container's file system lives only as long as the Container does. So when a
Container terminates and restarts, filesystem changes are lost. For more
consistent storage that is independent of the Container, you can use a
Volume. This is especially important for stateful
applications, such as key-value stores (such as Redis) and databases.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this exercise, you create a Pod that runs one Container. This Pod has a
Volume of type
emptyDir
that lasts for the life of the Pod, even if the Container terminates and
restarts. Here is the configuration file for the Pod:
In addition to the local disk storage provided by emptyDir, Kubernetes
supports many different network-attached storage solutions, including PD on
GCE and EBS on EC2, which are preferred for critical data and will handle
details such as mounting and unmounting the devices on the nodes. See
Volumes for more details.
4.3.9 - Configure a Pod to Use a PersistentVolume for Storage
This page shows you how to configure a Pod to use a
PersistentVolumeClaim
for storage.
Here is a summary of the process:
You, as cluster administrator, create a PersistentVolume backed by physical
storage. You do not associate the volume with any Pod.
You, now taking the role of a developer / cluster user, create a
PersistentVolumeClaim that is automatically bound to a suitable
PersistentVolume.
You create a Pod that uses the above PersistentVolumeClaim for storage.
Before you begin
You need to have a Kubernetes cluster that has only one Node, and the
kubectl
command-line tool must be configured to communicate with your cluster. If you
do not already have a single-node cluster, you can create one by using
Minikube.
Open a shell to the single Node in your cluster. How you open a shell depends
on how you set up your cluster. For example, if you are using Minikube, you
can open a shell to your Node by entering minikube ssh.
In your shell on that Node, create a /mnt/data directory:
# This assumes that your Node uses "sudo" to run commands# as the superusersudo mkdir /mnt/data
In the /mnt/data directory, create an index.html file:
# This again assumes that your Node uses "sudo" to run commands# as the superusersudo sh -c "echo 'Hello from Kubernetes storage' > /mnt/data/index.html"
Note: If your Node uses a tool for superuser access other than sudo, you can
usually make this work if you replace sudo with the name of the other tool.
Test that the index.html file exists:
cat /mnt/data/index.html
The output should be:
Hello from Kubernetes storage
You can now close the shell to your Node.
Create a PersistentVolume
In this exercise, you create a hostPath PersistentVolume. Kubernetes supports
hostPath for development and testing on a single-node cluster. A hostPath
PersistentVolume uses a file or directory on the Node to emulate network-attached storage.
In a production cluster, you would not use hostPath. Instead a cluster administrator
would provision a network resource like a Google Compute Engine persistent disk,
an NFS share, or an Amazon Elastic Block Store volume. Cluster administrators can also
use StorageClasses
to set up
dynamic provisioning.
Here is the configuration file for the hostPath PersistentVolume:
The configuration file specifies that the volume is at /mnt/data on the
cluster's Node. The configuration also specifies a size of 10 gibibytes and
an access mode of ReadWriteOnce, which means the volume can be mounted as
read-write by a single Node. It defines the StorageClass namemanual for the PersistentVolume, which will be used to bind
PersistentVolumeClaim requests to this PersistentVolume.
The output shows that the PersistentVolume has a STATUS of Available. This
means it has not yet been bound to a PersistentVolumeClaim.
NAME CAPACITY ACCESSMODES RECLAIMPOLICY STATUS CLAIM STORAGECLASS REASON AGE
task-pv-volume 10Gi RWO Retain Available manual 4s
Create a PersistentVolumeClaim
The next step is to create a PersistentVolumeClaim. Pods use PersistentVolumeClaims
to request physical storage. In this exercise, you create a PersistentVolumeClaim
that requests a volume of at least three gibibytes that can provide read-write
access for at least one Node.
Here is the configuration file for the PersistentVolumeClaim:
After you create the PersistentVolumeClaim, the Kubernetes control plane looks
for a PersistentVolume that satisfies the claim's requirements. If the control
plane finds a suitable PersistentVolume with the same StorageClass, it binds the
claim to the volume.
Look again at the PersistentVolume:
kubectl get pv task-pv-volume
Now the output shows a STATUS of Bound.
NAME CAPACITY ACCESSMODES RECLAIMPOLICY STATUS CLAIM STORAGECLASS REASON AGE
task-pv-volume 10Gi RWO Retain Bound default/task-pv-claim manual 2m
Look at the PersistentVolumeClaim:
kubectl get pvc task-pv-claim
The output shows that the PersistentVolumeClaim is bound to your PersistentVolume,
task-pv-volume.
NAME STATUS VOLUME CAPACITY ACCESSMODES STORAGECLASS AGE
task-pv-claim Bound task-pv-volume 10Gi RWO manual 30s
Create a Pod
The next step is to create a Pod that uses your PersistentVolumeClaim as a volume.
Notice that the Pod's configuration file specifies a PersistentVolumeClaim, but
it does not specify a PersistentVolume. From the Pod's point of view, the claim
is a volume.
In your shell, verify that nginx is serving the index.html file from the
hostPath volume:
# Be sure to run these 3 commands inside the root shell that comes from# running "kubectl exec" in the previous stepapt update
apt install curl
curl http://localhost/
The output shows the text that you wrote to the index.html file on the
hostPath volume:
Hello from Kubernetes storage
If you see that message, you have successfully configured a Pod to
use storage from a PersistentVolumeClaim.
Clean up
Delete the Pod, the PersistentVolumeClaim and the PersistentVolume:
apiVersion:v1kind:Podmetadata:name:testspec:containers:- name:testimage:nginxvolumeMounts:# a mount for site-data- name:configmountPath:/usr/share/nginx/htmlsubPath:html# another mount for nginx config- name:configmountPath:/etc/nginx/nginx.confsubPath:nginx.confvolumes:- name:configpersistentVolumeClaim:claimName:test-nfs-claim
You can perform 2 volume mounts on your nginx container:
/usr/share/nginx/html for the static website
/etc/nginx/nginx.conf for the default config
Access control
Storage configured with a group ID (GID) allows writing only by Pods using the same
GID. Mismatched or missing GIDs cause permission denied errors. To reduce the
need for coordination with users, an administrator can annotate a PersistentVolume
with a GID. Then the GID is automatically added to any Pod that uses the
PersistentVolume.
Use the pv.beta.kubernetes.io/gid annotation as follows:
When a Pod consumes a PersistentVolume that has a GID annotation, the annotated GID
is applied to all containers in the Pod in the same way that GIDs specified in the
Pod's security context are. Every GID, whether it originates from a PersistentVolume
annotation or the Pod's specification, is applied to the first process run in
each container.
Note: When a Pod consumes a PersistentVolume, the GIDs associated with the
PersistentVolume are not present on the Pod resource itself.
4.3.10 - Configure a Pod to Use a Projected Volume for Storage
This page shows how to use a projected Volume to mount
several existing volume sources into the same directory. Currently, secret, configMap, downwardAPI,
and serviceAccountToken volumes can be projected.
Note:serviceAccountToken is not a volume type.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this exercise, you create username and password Secrets from local files. You then create a Pod that runs one container, using a projected Volume to mount the Secrets into the same shared directory.
allowPrivilegeEscalation: Controls whether a process can gain more privileges than
its parent process. This bool directly controls whether the
no_new_privs
flag gets set on the container process.
allowPrivilegeEscalation is always true when the container:
is run as privileged, or
has CAP_SYS_ADMIN
readOnlyRootFilesystem: Mounts the container's root filesystem as read-only.
The above bullets are not a complete set of security context settings -- please see
SecurityContext
for a comprehensive list.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
To specify security settings for a Pod, include the securityContext field
in the Pod specification. The securityContext field is a
PodSecurityContext object.
The security settings that you specify for a Pod apply to all Containers in the Pod.
Here is a configuration file for a Pod that has a securityContext and an emptyDir volume:
In the configuration file, the runAsUser field specifies that for any Containers in
the Pod, all processes run with user ID 1000. The runAsGroup field specifies the primary group ID of 3000 for
all processes within any containers of the Pod. If this field is omitted, the primary group ID of the containers
will be root(0). Any files created will also be owned by user 1000 and group 3000 when runAsGroup is specified.
Since fsGroup field is specified, all processes of the container are also part of the supplementary group ID 2000.
The owner for volume /data/demo and any files created in that volume will be Group ID 2000.
The output shows that the processes are running as user 1000, which is the value of runAsUser:
PID USER TIME COMMAND
1 1000 0:00 sleep 1h
6 1000 0:00 sh
...
In your shell, navigate to /data, and list the one directory:
cd /data
ls -l
The output shows that the /data/demo directory has group ID 2000, which is
the value of fsGroup.
drwxrwsrwx 2 root 2000 4096 Jun 6 20:08 demo
In your shell, navigate to /data/demo, and create a file:
cd demo
echo hello > testfile
List the file in the /data/demo directory:
ls -l
The output shows that testfile has group ID 2000, which is the value of fsGroup.
-rw-r--r-- 1 1000 2000 6 Jun 6 20:08 testfile
Run the following command:
id
The output is similar to this:
uid=1000 gid=3000 groups=2000
From the output, you can see that gid is 3000 which is same as the runAsGroup field.
If the runAsGroup was omitted, the gid would remain as 0 (root) and the process will
be able to interact with files that are owned by the root(0) group and groups that have
the required group permissions for the root (0) group.
Exit your shell:
exit
Configure volume permission and ownership change policy for Pods
FEATURE STATE:Kubernetes v1.23 [stable]
By default, Kubernetes recursively changes ownership and permissions for the contents of each
volume to match the fsGroup specified in a Pod's securityContext when that volume is
mounted.
For large volumes, checking and changing ownership and permissions can take a lot of time,
slowing Pod startup. You can use the fsGroupChangePolicy field inside a securityContext
to control the way that Kubernetes checks and manages ownership and permissions
for a volume.
fsGroupChangePolicy - fsGroupChangePolicy defines behavior for changing ownership
and permission of the volume before being exposed inside a Pod.
This field only applies to volume types that support fsGroup controlled ownership and permissions.
This field has two possible values:
OnRootMismatch: Only change permissions and ownership if permission and ownership of
root directory does not match with expected permissions of the volume.
This could help shorten the time it takes to change ownership and permission of a volume.
Always: Always change permission and ownership of the volume when volume is mounted.
Note: This field has no effect on ephemeral volume types such as
secret,
configMap,
and emptydir.
Delegating volume permission and ownership change to CSI driver
FEATURE STATE:Kubernetes v1.23 [beta]
If you deploy a Container Storage Interface (CSI)
driver which supports the VOLUME_MOUNT_GROUPNodeServiceCapability, the
process of setting file ownership and permissions based on the
fsGroup specified in the securityContext will be performed by the CSI driver
instead of Kubernetes, provided that the DelegateFSGroupToCSIDriver Kubernetes
feature gate is enabled. In this case, since Kubernetes doesn't perform any
ownership and permission change, fsGroupChangePolicy does not take effect, and
as specified by CSI, the driver is expected to mount the volume with the
provided fsGroup, resulting in a volume that is readable/writable by the
fsGroup.
Please refer to the KEP
and the description of the VolumeCapability.MountVolume.volume_mount_group
field in the CSI spec
for more information.
Set the security context for a Container
To specify security settings for a Container, include the securityContext field
in the Container manifest. The securityContext field is a
SecurityContext object.
Security settings that you specify for a Container apply only to
the individual Container, and they override settings made at the Pod level when
there is overlap. Container settings do not affect the Pod's Volumes.
Here is the configuration file for a Pod that has one Container. Both the Pod
and the Container have a securityContext field:
The output shows that the processes are running as user 2000. This is the value
of runAsUser specified for the Container. It overrides the value 1000 that is
specified for the Pod.
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
2000 1 0.0 0.0 4336 764 ? Ss 20:36 0:00 /bin/sh -c node server.js
2000 8 0.1 0.5 772124 22604 ? Sl 20:36 0:00 node server.js
...
Exit your shell:
exit
Set capabilities for a Container
With Linux capabilities,
you can grant certain privileges to a process without granting all the privileges
of the root user. To add or remove Linux capabilities for a Container, include the
capabilities field in the securityContext section of the Container manifest.
First, see what happens when you don't include a capabilities field.
Here is configuration file that does not add or remove any Container capabilities:
In the capability bitmap of the first container, bits 12 and 25 are clear. In the second container,
bits 12 and 25 are set. Bit 12 is CAP_NET_ADMIN, and bit 25 is CAP_SYS_TIME.
See capability.h
for definitions of the capability constants.
Note: Linux capability constants have the form CAP_XXX.
But when you list capabilities in your container manifest, you must
omit the CAP_ portion of the constant.
For example, to add CAP_SYS_TIME, include SYS_TIME in your list of capabilities.
Set the Seccomp Profile for a Container
To set the Seccomp profile for a Container, include the seccompProfile field
in the securityContext section of your Pod or Container manifest. The
seccompProfile field is a
SeccompProfile object consisting of type and localhostProfile.
Valid options for type include RuntimeDefault, Unconfined, and
Localhost. localhostProfile must only be set if type: Localhost. It
indicates the path of the pre-configured profile on the node, relative to the
kubelet's configured Seccomp profile location (configured with the --root-dir
flag).
Here is an example that sets the Seccomp profile to the node's container runtime
default profile:
To assign SELinux labels to a Container, include the seLinuxOptions field in
the securityContext section of your Pod or Container manifest. The
seLinuxOptions field is an
SELinuxOptions
object. Here's an example that applies an SELinux level:
Note: To assign SELinux labels, the SELinux security module must be loaded on the host operating system.
Efficient SELinux volume relabeling
FEATURE STATE:Kubernetes v1.25 [alpha]
By default, the contrainer runtime recursively assigns SELinux label to all
files on all Pod volumes. To speed up this process, Kubernetes can change the
SELinux label of a volume instantly by using a mount option
-o context=<label>.
To benefit from this speedup, all these conditions must be met:
Alpha feature gates ReadWriteOncePod and SELinuxMountReadWriteOncePod must
be enabled.
Pod must use PersistentVolumeClaim with accessModes: ["ReadWriteOncePod"].
Pod (or all its Containers that use the PersistentVolumeClaim) must
have seLinuxOptions set.
The corresponding PersistentVolume must be either a volume that uses a
CSI driver, or a volume that uses the
legacy iscsi volume type.
If you use a volume backed by a CSI driver, that CSI driver must announce that it
supports mounting with -o context by setting spec.seLinuxMount: true in
its CSIDriver instance.
For any other volume types, SELinux relabelling happens another way: the container
runtime recursively changes the SELinux label for all inodes (files and directories)
in the volume.
The more files and directories in the volume, the longer that relabelling takes.
Note: In Kubernetes 1.25, the kubelet loses track of volume labels after restart. In
other words, then kubelet may refuse to start Pods with errors similar to "conflicting
SELinux labels of volume", while there are no conflicting labels in Pods. Make sure
nodes are
fully drained
before restarting kubelet.
Discussion
The security context for a Pod applies to the Pod's Containers and also to
the Pod's Volumes when applicable. Specifically fsGroup and seLinuxOptions are
applied to Volumes as follows:
fsGroup: Volumes that support ownership management are modified to be owned
and writable by the GID specified in fsGroup. See the
Ownership Management design document
for more details.
seLinuxOptions: Volumes that support SELinux labeling are relabeled to be accessible
by the label specified under seLinuxOptions. Usually you only
need to set the level section. This sets the
Multi-Category Security (MCS)
label given to all Containers in the Pod as well as the Volumes.
Warning: After you specify an MCS label for a Pod, all Pods with the same label can access the Volume.
If you need inter-Pod protection, you must assign a unique MCS label to each Pod.
Clean up
Delete the Pod:
kubectl delete pod security-context-demo
kubectl delete pod security-context-demo-2
kubectl delete pod security-context-demo-3
kubectl delete pod security-context-demo-4
A service account provides an identity for processes that run in a Pod.
Note: This document is a user introduction to Service Accounts and describes how service accounts behave in a cluster set up
as recommended by the Kubernetes project. Your cluster administrator may have
customized the behavior in your cluster, in which case this documentation may
not apply.
When you (a human) access the cluster (for example, using kubectl), you are
authenticated by the apiserver as a particular User Account (currently this is
usually admin, unless your cluster administrator has customized your cluster). Processes in containers inside pods can also contact the apiserver.
When they do, they are authenticated as a particular Service Account (for example, default).
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Use the Default Service Account to access the API server
When you create a pod, if you do not specify a service account, it is
automatically assigned the default service account in the same namespace.
If you get the raw json or yaml for a pod you have created (for example, kubectl get pods/<podname> -o yaml),
you can see the spec.serviceAccountName field has been
automatically set.
You can access the API from inside a pod using automatically mounted service account credentials, as described in
Accessing the Cluster.
The API permissions of the service account depend on the
authorization plugin and policy in use.
You can opt out of automounting API credentials on /var/run/secrets/kubernetes.io/serviceaccount/token for a service account by setting automountServiceAccountToken: false on the ServiceAccount:
The pod spec takes precedence over the service account if both specify a automountServiceAccountToken value.
Use Multiple Service Accounts
Every namespace has a default service account resource called default.
You can list this and any other serviceAccount resources in the namespace with this command:
kubectl get serviceaccounts
The output is similar to this:
NAME SECRETS AGE
default 1 1d
You can create additional ServiceAccount objects like this:
Now, when a new Pod is created in the current namespace and using the default ServiceAccount, the new Pod has its spec.imagePullSecrets field set automatically:
kubectl run nginx --image=nginx --restart=Never
kubectl get pod nginx -o=jsonpath='{.spec.imagePullSecrets[0].name}{"\n"}'
The output is:
myregistrykey
Service Account Token Volume Projection
FEATURE STATE:Kubernetes v1.20 [stable]
Note:
To enable and use token request projection, you must specify each of the following
command line arguments to kube-apiserver:
--service-account-issuer
It can be used as the Identifier of the service account token issuer. You can specify the --service-account-issuer argument multiple times, this can be useful to enable a non-disruptive change of the issuer. When this flag is specified multiple times, the first is used to generate tokens and all are used to determine which issuers are accepted. You must be running Kubernetes v1.22 or later to be able to specify --service-account-issuer multiple times.
--service-account-key-file
File containing PEM-encoded x509 RSA or ECDSA private or public keys, used to verify ServiceAccount tokens. The specified file can contain multiple keys, and the flag can be specified multiple times with different files. If specified multiple times, tokens signed by any of the specified keys are considered valid by the Kubernetes API server.
--service-account-signing-key-file
Path to the file that contains the current private key of the service account token issuer. The issuer signs issued ID tokens with this private key.
--api-audiences (can be omitted)
The service account token authenticator validates that tokens used against the API are bound to at least one of these audiences. If api-audiences is specified multiple times, tokens for any of the specified audiences are considered valid by the Kubernetes API server. If the --service-account-issuer flag is configured and this flag is not, this field defaults to a single element list containing the issuer URL.
The kubelet can also project a service account token into a Pod. You can
specify desired properties of the token, such as the audience and the validity
duration. These properties are not configurable on the default service account
token. The service account token will also become invalid against the API when
the Pod or the ServiceAccount is deleted.
This behavior is configured on a PodSpec using a ProjectedVolume type called
ServiceAccountToken. To provide a
pod with a token with an audience of "vault" and a validity duration of two
hours, you would configure the following in your PodSpec:
The kubelet will request and store the token on behalf of the pod, make the
token available to the pod at a configurable file path, and refresh the token as it approaches expiration.
The kubelet proactively rotates the token if it is older than 80% of its total TTL, or if the token is older than 24 hours.
The application is responsible for reloading the token when it rotates. Periodic reloading (e.g. once every 5 minutes) is sufficient for most use cases.
Service Account Issuer Discovery
FEATURE STATE:Kubernetes v1.21 [stable]
The Service Account Issuer Discovery feature is enabled when the Service Account
Token Projection feature is enabled, as described
above.
Note:
The issuer URL must comply with the
OIDC Discovery Spec. In
practice, this means it must use the https scheme, and should serve an OpenID
provider configuration at {service-account-issuer}/.well-known/openid-configuration.
If the URL does not comply, the ServiceAccountIssuerDiscovery endpoints will
not be registered, even if the feature is enabled.
The Service Account Issuer Discovery feature enables federation of Kubernetes
service account tokens issued by a cluster (the identity provider) with
external systems (relying parties).
When enabled, the Kubernetes API server provides an OpenID Provider
Configuration document at /.well-known/openid-configuration and the associated
JSON Web Key Set (JWKS) at /openid/v1/jwks. The OpenID Provider Configuration
is sometimes referred to as the discovery document.
Clusters include a default RBAC ClusterRole called
system:service-account-issuer-discovery. A default RBAC ClusterRoleBinding
assigns this role to the system:serviceaccounts group, which all service
accounts implicitly belong to. This allows pods running on the cluster to access
the service account discovery document via their mounted service account token.
Administrators may, additionally, choose to bind the role to
system:authenticated or system:unauthenticated depending on their security
requirements and which external systems they intend to federate with.
Note: The responses served at /.well-known/openid-configuration and
/openid/v1/jwks are designed to be OIDC compatible, but not strictly OIDC
compliant. Those documents contain only the parameters necessary to perform
validation of Kubernetes service account tokens.
The JWKS response contains public keys that a relying party can use to validate
the Kubernetes service account tokens. Relying parties first query for the
OpenID Provider Configuration, and use the jwks_uri field in the response to
find the JWKS.
In many cases, Kubernetes API servers are not available on the public internet,
but public endpoints that serve cached responses from the API server can be made
available by users or service providers. In these cases, it is possible to
override the jwks_uri in the OpenID Provider Configuration so that it points
to the public endpoint, rather than the API server's address, by passing the
--service-account-jwks-uri flag to the API server. Like the issuer URL, the
JWKS URI is required to use the https scheme.
This page shows how to create a Pod that uses a
Secret to pull an image
from a private container image registry or repository. There are many private
registries in use. This task uses Docker Hub
as an example registry.
🛇 This item links to a third party project or product that is not part of Kubernetes itself. More information
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
If you need more control (for example, to set a namespace or a label on the new
secret) then you can customise the Secret before storing it.
Be sure to:
set the name of the data item to .dockerconfigjson
base64 encode the Docker configuration file and then paste that string, unbroken
as the value for field data[".dockerconfigjson"]
If you get the error message error: no objects passed to create, it may mean the base64 encoded string is invalid.
If you get an error message like Secret "myregistrykey" is invalid: data[.dockerconfigjson]: invalid value ..., it means
the base64 encoded string in the data was successfully decoded, but could not be parsed as a .docker/config.json file.
Create a Secret by providing credentials on the command line
<your-registry-server> is your Private Docker Registry FQDN.
Use https://index.docker.io/v1/ for DockerHub.
<your-name> is your Docker username.
<your-pword> is your Docker password.
<your-email> is your Docker email.
You have successfully set your Docker credentials in the cluster as a Secret called regcred.
Note: Typing secrets on the command line may store them in your shell history unprotected, and
those secrets might also be visible to other users on your PC during the time that
kubectl is running.
Inspecting the Secret regcred
To understand the contents of the regcred Secret you created, start by viewing the Secret in YAML format:
To pull the image from the private registry, Kubernetes needs credentials.
The imagePullSecrets field in the configuration file specifies that
Kubernetes should get the credentials from a Secret named regcred.
Create a Pod that uses your Secret, and verify that the Pod is running:
kubectl apply -f my-private-reg-pod.yaml
kubectl get pod private-reg
4.3.14 - Configure Liveness, Readiness and Startup Probes
This page shows how to configure liveness, readiness and startup probes for containers.
The kubelet uses liveness probes to know when to
restart a container. For example, liveness probes could catch a deadlock,
where an application is running, but unable to make progress. Restarting a
container in such a state can help to make the application more available
despite bugs.
The kubelet uses readiness probes to know when a container is ready to start
accepting traffic. A Pod is considered ready when all of its containers are ready.
One use of this signal is to control which Pods are used as backends for Services.
When a Pod is not ready, it is removed from Service load balancers.
The kubelet uses startup probes to know when a container application has started.
If such a probe is configured, it disables liveness and readiness checks until
it succeeds, making sure those probes don't interfere with the application startup.
This can be used to adopt liveness checks on slow starting containers, avoiding them
getting killed by the kubelet before they are up and running.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Many applications running for long periods of time eventually transition to
broken states, and cannot recover except by being restarted. Kubernetes provides
liveness probes to detect and remedy such situations.
In this exercise, you create a Pod that runs a container based on the
registry.k8s.io/busybox image. Here is the configuration file for the Pod:
In the configuration file, you can see that the Pod has a single Container.
The periodSeconds field specifies that the kubelet should perform a liveness
probe every 5 seconds. The initialDelaySeconds field tells the kubelet that it
should wait 5 seconds before performing the first probe. To perform a probe, the
kubelet executes the command cat /tmp/healthy in the target container. If the
command succeeds, it returns 0, and the kubelet considers the container to be alive and
healthy. If the command returns a non-zero value, the kubelet kills the container
and restarts it.
When the container starts, it executes this command:
For the first 30 seconds of the container's life, there is a /tmp/healthy file.
So during the first 30 seconds, the command cat /tmp/healthy returns a success
code. After 30 seconds, cat /tmp/healthy returns a failure code.
The output indicates that no liveness probes have failed yet:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 11s default-scheduler Successfully assigned default/liveness-exec to node01
Normal Pulling 9s kubelet, node01 Pulling image "registry.k8s.io/busybox"
Normal Pulled 7s kubelet, node01 Successfully pulled image "registry.k8s.io/busybox"
Normal Created 7s kubelet, node01 Created container liveness
Normal Started 7s kubelet, node01 Started container liveness
After 35 seconds, view the Pod events again:
kubectl describe pod liveness-exec
At the bottom of the output, there are messages indicating that the liveness
probes have failed, and the failed containers have been killed and recreated.
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 57s default-scheduler Successfully assigned default/liveness-exec to node01
Normal Pulling 55s kubelet, node01 Pulling image "registry.k8s.io/busybox"
Normal Pulled 53s kubelet, node01 Successfully pulled image "registry.k8s.io/busybox"
Normal Created 53s kubelet, node01 Created container liveness
Normal Started 53s kubelet, node01 Started container liveness
Warning Unhealthy 10s (x3 over 20s) kubelet, node01 Liveness probe failed: cat: can't open '/tmp/healthy': No such file or directory
Normal Killing 10s kubelet, node01 Container liveness failed liveness probe, will be restarted
Wait another 30 seconds, and verify that the container has been restarted:
kubectl get pod liveness-exec
The output shows that RESTARTS has been incremented. Note that the RESTARTS counter increments as soon as a failed container comes back to the running state:
NAME READY STATUS RESTARTS AGE
liveness-exec 1/1 Running 1 1m
Define a liveness HTTP request
Another kind of liveness probe uses an HTTP GET request. Here is the configuration
file for a Pod that runs a container based on the registry.k8s.io/liveness
image.
In the configuration file, you can see that the Pod has a single container.
The periodSeconds field specifies that the kubelet should perform a liveness
probe every 3 seconds. The initialDelaySeconds field tells the kubelet that it
should wait 3 seconds before performing the first probe. To perform a probe, the
kubelet sends an HTTP GET request to the server that is running in the container
and listening on port 8080. If the handler for the server's /healthz path
returns a success code, the kubelet considers the container to be alive and
healthy. If the handler returns a failure code, the kubelet kills the container
and restarts it.
Any code greater than or equal to 200 and less than 400 indicates success. Any
other code indicates failure.
You can see the source code for the server in
server.go.
For the first 10 seconds that the container is alive, the /healthz handler
returns a status of 200. After that, the handler returns a status of 500.
The kubelet starts performing health checks 3 seconds after the container starts.
So the first couple of health checks will succeed. But after 10 seconds, the health
checks will fail, and the kubelet will kill and restart the container.
After 10 seconds, view Pod events to verify that liveness probes have failed and
the container has been restarted:
kubectl describe pod liveness-http
In releases prior to v1.13 (including v1.13), if the environment variable
http_proxy (or HTTP_PROXY) is set on the node where a Pod is running,
the HTTP liveness probe uses that proxy.
In releases after v1.13, local HTTP proxy environment variable settings do not
affect the HTTP liveness probe.
Define a TCP liveness probe
A third type of liveness probe uses a TCP socket. With this configuration, the
kubelet will attempt to open a socket to your container on the specified port.
If it can establish a connection, the container is considered healthy, if it
can't it is considered a failure.
As you can see, configuration for a TCP check is quite similar to an HTTP check.
This example uses both readiness and liveness probes. The kubelet will send the
first readiness probe 5 seconds after the container starts. This will attempt to
connect to the goproxy container on port 8080. If the probe succeeds, the Pod
will be marked as ready. The kubelet will continue to run this check every 10
seconds.
In addition to the readiness probe, this configuration includes a liveness probe.
The kubelet will run the first liveness probe 15 seconds after the container
starts. Similar to the readiness probe, this will attempt to connect to the
goproxy container on port 8080. If the liveness probe fails, the container
will be restarted.
After 15 seconds, view Pod events to verify that liveness probes:
kubectl describe pod goproxy
Define a gRPC liveness probe
FEATURE STATE:Kubernetes v1.24 [beta]
If your application implements gRPC Health Checking Protocol,
kubelet can be configured to use it for application liveness checks.
You must enable the GRPCContainerProbefeature gate
in order to configure checks that rely on gRPC.
To use a gRPC probe, port must be configured. If the health endpoint is configured
on a non-default service, you must also specify the service.
Note: Unlike HTTP and TCP probes, named ports cannot be used and custom host cannot be configured.
Configuration problems (for example: incorrect port and service, unimplemented health checking protocol)
are considered a probe failure, similar to HTTP and TCP probes.
To try the gRPC liveness check, create a Pod using the command below.
In the example below, the etcd pod is configured to use gRPC liveness probe.
After 15 seconds, view Pod events to verify that the liveness check has not failed:
kubectl describe pod etcd-with-grpc
Before Kubernetes 1.23, gRPC health probes were often implemented using grpc-health-probe,
as described in the blog post Health checking gRPC servers on Kubernetes.
The built-in gRPC probes behavior is similar to one implemented by grpc-health-probe.
When migrating from grpc-health-probe to built-in probes, remember the following differences:
Built-in probes run against the pod IP address, unlike grpc-health-probe that often runs against 127.0.0.1.
Be sure to configure your gRPC endpoint to listen on the Pod's IP address.
Built-in probes do not support any authentication parameters (like -tls).
There are no error codes for built-in probes. All errors are considered as probe failures.
If ExecProbeTimeout feature gate is set to false, grpc-health-probe does not respect the timeoutSeconds setting (which defaults to 1s),
while built-in probe would fail on timeout.
Use a named port
You can use a named
port
for HTTP and TCP probes. (gRPC probes do not support named ports).
Protect slow starting containers with startup probes
Sometimes, you have to deal with legacy applications that might require
an additional startup time on their first initialization.
In such cases, it can be tricky to set up liveness probe parameters without
compromising the fast response to deadlocks that motivated such a probe.
The trick is to set up a startup probe with the same command, HTTP or TCP
check, with a failureThreshold * periodSeconds long enough to cover the
worse case startup time.
Thanks to the startup probe, the application will have a maximum of 5 minutes
(30 * 10 = 300s) to finish its startup.
Once the startup probe has succeeded once, the liveness probe takes over to
provide a fast response to container deadlocks.
If the startup probe never succeeds, the container is killed after 300s and
subject to the pod's restartPolicy.
Define readiness probes
Sometimes, applications are temporarily unable to serve traffic.
For example, an application might need to load large data or configuration
files during startup, or depend on external services after startup.
In such cases, you don't want to kill the application,
but you don't want to send it requests either. Kubernetes provides
readiness probes to detect and mitigate these situations. A pod with containers
reporting that they are not ready does not receive traffic through Kubernetes
Services.
Note: Readiness probes runs on the container during its whole lifecycle.
Caution: Liveness probes do not wait for readiness probes to succeed. If you want to wait before executing a liveness probe you should use initialDelaySeconds or a startupProbe.
Readiness probes are configured similarly to liveness probes. The only difference
is that you use the readinessProbe field instead of the livenessProbe field.
Configuration for HTTP and TCP readiness probes also remains identical to
liveness probes.
Readiness and liveness probes can be used in parallel for the same container.
Using both can ensure that traffic does not reach a container that is not ready
for it, and that containers are restarted when they fail.
Configure Probes
Probes have a number of fields that
you can use to more precisely control the behavior of liveness and readiness
checks:
initialDelaySeconds: Number of seconds after the container has started
before liveness or readiness probes are initiated. Defaults to 0 seconds. Minimum value is 0.
periodSeconds: How often (in seconds) to perform the probe. Default to 10
seconds. Minimum value is 1.
timeoutSeconds: Number of seconds after which the probe times out. Defaults
to 1 second. Minimum value is 1.
successThreshold: Minimum consecutive successes for the probe to be
considered successful after having failed. Defaults to 1. Must be 1 for liveness
and startup Probes. Minimum value is 1.
failureThreshold: When a probe fails, Kubernetes will
try failureThreshold times before giving up. Giving up in case of liveness probe means restarting the container. In case of readiness probe the Pod will be marked Unready.
Defaults to 3. Minimum value is 1.
Note:
Before Kubernetes 1.20, the field timeoutSeconds was not respected for exec probes:
probes continued running indefinitely, even past their configured deadline,
until a result was returned.
This defect was corrected in Kubernetes v1.20. You may have been relying on the previous behavior,
even without realizing it, as the default timeout is 1 second.
As a cluster administrator, you can disable the feature gateExecProbeTimeout (set it to false)
on each kubelet to restore the behavior from older versions, then remove that override
once all the exec probes in the cluster have a timeoutSeconds value set.
If you have pods that are impacted from the default 1 second timeout,
you should update their probe timeout so that you're ready for the
eventual removal of that feature gate.
With the fix of the defect, for exec probes, on Kubernetes 1.20+ with the dockershim container runtime,
the process inside the container may keep running even after probe returned failure because of the timeout.
Caution: Incorrect implementation of readiness probes may result in an ever growing number
of processes in the container, and resource starvation if this is left unchecked.
HTTP probes
HTTP probes
have additional fields that can be set on httpGet:
host: Host name to connect to, defaults to the pod IP. You probably want to
set "Host" in httpHeaders instead.
scheme: Scheme to use for connecting to the host (HTTP or HTTPS). Defaults to HTTP.
path: Path to access on the HTTP server. Defaults to /.
httpHeaders: Custom headers to set in the request. HTTP allows repeated headers.
port: Name or number of the port to access on the container. Number must be
in the range 1 to 65535.
For an HTTP probe, the kubelet sends an HTTP request to the specified path and
port to perform the check. The kubelet sends the probe to the pod's IP address,
unless the address is overridden by the optional host field in httpGet. If
scheme field is set to HTTPS, the kubelet sends an HTTPS request skipping the
certificate verification. In most scenarios, you do not want to set the host field.
Here's one scenario where you would set it. Suppose the container listens on 127.0.0.1
and the Pod's hostNetwork field is true. Then host, under httpGet, should be set
to 127.0.0.1. If your pod relies on virtual hosts, which is probably the more common
case, you should not use host, but rather set the Host header in httpHeaders.
For an HTTP probe, the kubelet sends two request headers in addition to the mandatory Host header:
User-Agent, and Accept. The default values for these headers are kube-probe/1.25
(where 1.25 is the version of the kubelet ), and */* respectively.
You can override the default headers by defining .httpHeaders for the probe; for example
For a TCP probe, the kubelet makes the probe connection at the node, not in the pod, which
means that you can not use a service name in the host parameter since the kubelet is unable
to resolve it.
Probe-level terminationGracePeriodSeconds
FEATURE STATE:Kubernetes v1.25 [beta]
Prior to release 1.21, the pod-level terminationGracePeriodSeconds was used
for terminating a container that failed its liveness or startup probe. This
coupling was unintended and may have resulted in failed containers taking an
unusually long time to restart when a pod-level terminationGracePeriodSeconds
was set.
In 1.25 and beyond, users can specify a probe-level terminationGracePeriodSeconds
as part of the probe specification. When both a pod- and probe-level
terminationGracePeriodSeconds are set, the kubelet will use the probe-level value.
Note:
Beginning in Kubernetes 1.25, the ProbeTerminationGracePeriod feature is enabled
by default. For users choosing to disable this feature, please note the following:
The ProbeTerminationGracePeriod feature gate is only available on the API Server.
The kubelet always honors the probe-level terminationGracePeriodSeconds field if
it is present on a Pod.
If you have existing Pods where the terminationGracePeriodSeconds field is set and
you no longer wish to use per-probe termination grace periods, you must delete
those existing Pods.
When you (or the control plane, or some other component) create replacement
Pods, and the feature gate ProbeTerminationGracePeriod is disabled, then the
API server ignores the Probe-level terminationGracePeriodSeconds field, even if
a Pod or pod template specifies it.
This page shows how to assign a Kubernetes Pod to a particular node in a
Kubernetes cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
List the nodes in your cluster, along with their labels:
kubectl get nodes --show-labels
The output is similar to this:
NAME STATUS ROLES AGE VERSION LABELS
worker0 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker0
worker1 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker1
worker2 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker2
Choose one of your nodes, and add a label to it:
kubectl label nodes <your-node-name> disktype=ssd
where <your-node-name> is the name of your chosen node.
Verify that your chosen node has a disktype=ssd label:
kubectl get nodes --show-labels
The output is similar to this:
NAME STATUS ROLES AGE VERSION LABELS
worker0 Ready <none> 1d v1.13.0 ...,disktype=ssd,kubernetes.io/hostname=worker0
worker1 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker1
worker2 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker2
In the preceding output, you can see that the worker0 node has a
disktype=ssd label.
Create a pod that gets scheduled to your chosen node
This pod configuration file describes a pod that has a node selector,
disktype: ssd. This means that the pod will get scheduled on a node that has
a disktype=ssd label.
apiVersion:v1kind:Podmetadata:name:nginxspec:nodeName:foo-node# schedule pod to specific nodecontainers:- name:nginximage:nginximagePullPolicy:IfNotPresent
Use the configuration file to create a pod that will get scheduled on foo-node only.
This page shows how to assign a Kubernetes Pod to a particular node using Node Affinity in a
Kubernetes cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.10.
To check the version, enter kubectl version.
Add a label to a node
List the nodes in your cluster, along with their labels:
kubectl get nodes --show-labels
The output is similar to this:
NAME STATUS ROLES AGE VERSION LABELS
worker0 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker0
worker1 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker1
worker2 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker2
Chose one of your nodes, and add a label to it:
kubectl label nodes <your-node-name> disktype=ssd
where <your-node-name> is the name of your chosen node.
Verify that your chosen node has a disktype=ssd label:
kubectl get nodes --show-labels
The output is similar to this:
NAME STATUS ROLES AGE VERSION LABELS
worker0 Ready <none> 1d v1.13.0 ...,disktype=ssd,kubernetes.io/hostname=worker0
worker1 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker1
worker2 Ready <none> 1d v1.13.0 ...,kubernetes.io/hostname=worker2
In the preceding output, you can see that the worker0 node has a
disktype=ssd label.
Schedule a Pod using required node affinity
This manifest describes a Pod that has a requiredDuringSchedulingIgnoredDuringExecution node affinity,disktype: ssd.
This means that the pod will get scheduled only on a node that has a disktype=ssd label.
Verify that the pod is running on your chosen node:
kubectl get pods --output=wide
The output is similar to this:
NAME READY STATUS RESTARTS AGE IP NODE
nginx 1/1 Running 0 13s 10.200.0.4 worker0
Schedule a Pod using preferred node affinity
This manifest describes a Pod that has a preferredDuringSchedulingIgnoredDuringExecution node affinity,disktype: ssd.
This means that the pod will prefer a node that has a disktype=ssd label.
This page shows how to use an Init Container to initialize a Pod before an
application Container runs.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this exercise you create a Pod that has one application Container and one
Init Container. The init container runs to completion before the application
container starts.
apiVersion:v1kind:Podmetadata:name:init-demospec:containers:- name:nginximage:nginxports:- containerPort:80volumeMounts:- name:workdirmountPath:/usr/share/nginx/html# These containers are run during pod initializationinitContainers:- name:installimage:busybox:1.28command:- wget- "-O"- "/work-dir/index.html"- http://info.cern.chvolumeMounts:- name:workdirmountPath:"/work-dir"dnsPolicy:Defaultvolumes:- name:workdiremptyDir:{}
In the configuration file, you can see that the Pod has a Volume that the init
container and the application container share.
The init container mounts the
shared Volume at /work-dir, and the application container mounts the shared
Volume at /usr/share/nginx/html. The init container runs the following command
and then terminates:
wget -O /work-dir/index.html http://info.cern.ch
Notice that the init container writes the index.html file in the root directory
of the nginx server.
The output shows that nginx is serving the web page that was written by the init container:
<html><head></head><body><header>
<title>http://info.cern.ch</title>
</header>
<h1>http://info.cern.ch - home of the first website</h1>
...
<li><a href="http://info.cern.ch/hypertext/WWW/TheProject.html">Browse the first website</a></li>
...
4.3.18 - Attach Handlers to Container Lifecycle Events
This page shows how to attach handlers to Container lifecycle events. Kubernetes supports
the postStart and preStop events. Kubernetes sends the postStart event immediately
after a Container is started, and it sends the preStop event immediately before the
Container is terminated. A Container may specify one handler per event.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
apiVersion:v1kind:Podmetadata:name:lifecycle-demospec:containers:- name:lifecycle-demo-containerimage:nginxlifecycle:postStart:exec:command:["/bin/sh","-c","echo Hello from the postStart handler > /usr/share/message"]preStop:exec:command:["/bin/sh","-c","nginx -s quit; while killall -0 nginx; do sleep 1; done"]
In the configuration file, you can see that the postStart command writes a message
file to the Container's /usr/share directory. The preStop command shuts down
nginx gracefully. This is helpful if the Container is being terminated because of a failure.
Get a shell into the Container running in your Pod:
kubectl exec -it lifecycle-demo -- /bin/bash
In your shell, verify that the postStart handler created the message file:
root@lifecycle-demo:/# cat /usr/share/message
The output shows the text written by the postStart handler:
Hello from the postStart handler
Discussion
Kubernetes sends the postStart event immediately after the Container is created.
There is no guarantee, however, that the postStart handler is called before
the Container's entrypoint is called. The postStart handler runs asynchronously
relative to the Container's code, but Kubernetes' management of the container
blocks until the postStart handler completes. The Container's status is not
set to RUNNING until the postStart handler completes.
Kubernetes sends the preStop event immediately before the Container is terminated.
Kubernetes' management of the Container blocks until the preStop handler completes,
unless the Pod's grace period expires. For more details, see
Pod Lifecycle.
Note: Kubernetes only sends the preStop event when a Pod is terminated.
This means that the preStop hook is not invoked when the Pod is completed.
This limitation is tracked in issue #55087.
Many applications rely on configuration which is used during either application initialization or runtime.
Most of the times there is a requirement to adjust values assigned to configuration parameters.
ConfigMaps are the Kubernetes way to inject application pods with configuration data.
ConfigMaps allow you to decouple configuration artifacts from image content to keep containerized applications portable. This page provides a series of usage examples demonstrating how to create ConfigMaps and configure Pods using data stored in ConfigMaps.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You can use either kubectl create configmap or a ConfigMap generator in kustomization.yaml to create a ConfigMap. Note that kubectl starts to support kustomization.yaml since 1.14.
where <map-name> is the name you want to assign to the ConfigMap and <data-source> is the directory, file, or literal value to draw the data from.
The name of a ConfigMap object must be a valid
DNS subdomain name.
When you are creating a ConfigMap based on a file, the key in the <data-source> defaults to the basename of the file, and the value defaults to the file content.
You can use kubectl create configmap to create a ConfigMap from multiple files in the same directory. When you are creating a ConfigMap based on a directory, kubectl identifies files whose basename is a valid key in the directory and packages each of those files into the new ConfigMap. Any directory entries except regular files are ignored (e.g. subdirectories, symlinks, devices, pipes, etc).
For example:
# Create the local directorymkdir -p configure-pod-container/configmap/
# Download the sample files into `configure-pod-container/configmap/` directorywget https://kubernetes.io/examples/configmap/game.properties -O configure-pod-container/configmap/game.properties
wget https://kubernetes.io/examples/configmap/ui.properties -O configure-pod-container/configmap/ui.properties
# Create the configmapkubectl create configmap game-config --from-file=configure-pod-container/configmap/
The above command packages each file, in this case, game.properties and ui.properties in the configure-pod-container/configmap/ directory into the game-config ConfigMap. You can display details of the ConfigMap using the following command:
When kubectl creates a ConfigMap from inputs that are not ASCII or UTF-8, the tool puts these into the binaryData field of the ConfigMap, and not in data. Both text and binary data sources can be combined in one ConfigMap.
If you want to view the binaryData keys (and their values) in a ConfigMap, you can run kubectl get configmap -o jsonpath='{.binaryData}' <name>.
Use the option --from-env-file to create a ConfigMap from an env-file, for example:
# Env-files contain a list of environment variables.# These syntax rules apply:# Each line in an env file has to be in VAR=VAL format.# Lines beginning with # (i.e. comments) are ignored.# Blank lines are ignored.# There is no special handling of quotation marks (i.e. they will be part of the ConfigMap value)).# Download the sample files into `configure-pod-container/configmap/` directorywget https://kubernetes.io/examples/configmap/game-env-file.properties -O configure-pod-container/configmap/game-env-file.properties
wget https://kubernetes.io/examples/configmap/ui-env-file.properties -O configure-pod-container/configmap/ui-env-file.properties
# The env-file `game-env-file.properties` looks like belowcat configure-pod-container/configmap/game-env-file.properties
enemies=aliens
lives=3allowed="true"# This comment and the empty line above it are ignored
Starting with Kubernetes v1.23, kubectl supports the --from-env-file argument to be
specified multiple times to create a ConfigMap from multiple data sources.
where <my-key-name> is the key you want to use in the ConfigMap and <path-to-file> is the location of the data source file you want the key to represent.
You can pass in multiple key-value pairs. Each pair provided on the command line is represented as a separate entry in the data section of the ConfigMap.
kubectl supports kustomization.yaml since 1.14.
You can also create a ConfigMap from generators and then apply it to create the object on
the Apiserver. The generators
should be specified in a kustomization.yaml inside a directory.
Generate ConfigMaps from files
For example, to generate a ConfigMap from files configure-pod-container/configmap/game.properties
# Create a kustomization.yaml file with ConfigMapGeneratorcat <<EOF >./kustomization.yaml
configMapGenerator:
- name: game-config-4
files:
- configure-pod-container/configmap/game.properties
EOF
Apply the kustomization directory to create the ConfigMap object.
kubectl apply -k .
configmap/game-config-4-m9dm2f92bt created
You can check that the ConfigMap was created like this:
kubectl get configmap
NAME DATA AGE
game-config-4-m9dm2f92bt 1 37s
kubectl describe configmaps/game-config-4-m9dm2f92bt
Name: game-config-4-m9dm2f92bt
Namespace: default
Labels: <none>
Annotations: kubectl.kubernetes.io/last-applied-configuration:
{"apiVersion":"v1","data":{"game.properties":"enemies=aliens\nlives=3\nenemies.cheat=true\nenemies.cheat.level=noGoodRotten\nsecret.code.p...
Data
====
game.properties:
----
enemies=aliens
lives=3
enemies.cheat=true
enemies.cheat.level=noGoodRotten
secret.code.passphrase=UUDDLRLRBABAS
secret.code.allowed=true
secret.code.lives=30
Events: <none>
Note that the generated ConfigMap name has a suffix appended by hashing the contents. This ensures that a
new ConfigMap is generated each time the content is modified.
Define the key to use when generating a ConfigMap from a file
You can define a key other than the file name to use in the ConfigMap generator.
For example, to generate a ConfigMap from files configure-pod-container/configmap/game.properties
with the key game-special-key
# Create a kustomization.yaml file with ConfigMapGeneratorcat <<EOF >./kustomization.yaml
configMapGenerator:
- name: game-config-5
files:
- game-special-key=configure-pod-container/configmap/game.properties
EOF
Apply the kustomization directory to create the ConfigMap object.
kubectl apply -k .
configmap/game-config-5-m67dt67794 created
Generate ConfigMaps from Literals
To generate a ConfigMap from literals special.type=charm and special.how=very,
you can specify the ConfigMap generator in kustomization.yaml as
# Create a kustomization.yaml file with ConfigMapGeneratorcat <<EOF >./kustomization.yaml
configMapGenerator:
- name: special-config-2
literals:
- special.how=very
- special.type=charm
EOF
Apply the kustomization directory to create the ConfigMap object.
kubectl apply -k .
configmap/special-config-2-c92b5mmcf2 created
Define container environment variables using ConfigMap data
Define a container environment variable with data from a single ConfigMap
Define an environment variable as a key-value pair in a ConfigMap:
apiVersion:v1kind:Podmetadata:name:dapi-test-podspec:containers:- name:test-containerimage:registry.k8s.io/busyboxcommand:["/bin/sh","-c","env"]env:# Define the environment variable- name:SPECIAL_LEVEL_KEYvalueFrom:configMapKeyRef:# The ConfigMap containing the value you want to assign to SPECIAL_LEVEL_KEYname:special-config# Specify the key associated with the valuekey:special.howrestartPolicy:Never
Use envFrom to define all of the ConfigMap's data as container environment variables. The key from the ConfigMap becomes the environment variable name in the Pod.
produces the following output in the test-container container:
very charm
Add ConfigMap data to a Volume
As explained in Create ConfigMaps from files, when you create a ConfigMap using --from-file, the filename becomes a key stored in the data section of the ConfigMap. The file contents become the key's value.
The examples in this section refer to a ConfigMap named special-config, shown below.
Add the ConfigMap name under the volumes section of the Pod specification.
This adds the ConfigMap data to the directory specified as volumeMounts.mountPath (in this case, /etc/config).
The command section lists directory files with names that match the keys in ConfigMap.
apiVersion:v1kind:Podmetadata:name:dapi-test-podspec:containers:- name:test-containerimage:registry.k8s.io/busyboxcommand:["/bin/sh","-c","ls /etc/config/"]volumeMounts:- name:config-volumemountPath:/etc/configvolumes:- name:config-volumeconfigMap:# Provide the name of the ConfigMap containing the files you want# to add to the containername:special-configrestartPolicy:Never
When the pod runs, the command ls /etc/config/ produces the output below:
SPECIAL_LEVEL
SPECIAL_TYPE
Caution: If there are some files in the /etc/config/ directory, they will be deleted.
Note: Text data is exposed as files using the UTF-8 character encoding. To use some other character encoding, use binaryData.
Add ConfigMap data to a specific path in the Volume
Use the path field to specify the desired file path for specific ConfigMap items.
In this case, the SPECIAL_LEVEL item will be mounted in the config-volume volume at /etc/config/keys.
When the pod runs, the command cat /etc/config/keys produces the output below:
very
Caution: Like before, all previous files in the /etc/config/ directory will be deleted.
Project keys to specific paths and file permissions
You can project keys to specific paths and specific permissions on a per-file
basis. The Secrets user guide explains the syntax.
Understanding ConfigMaps and Pods
The ConfigMap API resource stores configuration data as key-value pairs. The data can be consumed in pods or provide the configurations for system components such as controllers. ConfigMap is similar to Secrets, but provides a means of working with strings that don't contain sensitive information. Users and system components alike can store configuration data in ConfigMap.
Note: ConfigMaps should reference properties files, not replace them. Think of the ConfigMap as representing something similar to the Linux /etc directory and its contents. For example, if you create a Kubernetes Volume from a ConfigMap, each data item in the ConfigMap is represented by an individual file in the volume.
The ConfigMap's data field contains the configuration data. As shown in the example below, this can be simple -- like individual properties defined using --from-literal -- or complex -- like configuration files or JSON blobs defined using --from-file.
apiVersion:v1kind:ConfigMapmetadata:creationTimestamp:2016-02-18T19:14:38Zname:example-confignamespace:defaultdata:# example of a simple property defined using --from-literalexample.property.1:helloexample.property.2:world# example of a complex property defined using --from-fileexample.property.file:|- property.1=value-1
property.2=value-2
property.3=value-3
Restrictions
You must create the ConfigMap object before you reference it in a Pod specification. Alternatively, mark the ConfigMap reference as optional in the Pod spec (see Optional ConfigMaps). If you reference a ConfigMap that doesn't exist and you don't mark the reference as optional, the Pod won't start. Similarly, references to keys that don't exist in the ConfigMap will also prevent the Pod from starting, unless you mark the key references as optional.
If you use envFrom to define environment variables from ConfigMaps, keys that are considered invalid will be skipped. The pod will be allowed to start, but the invalid names will be recorded in the event log (InvalidVariableNames). The log message lists each skipped key. For example:
kubectl get events
The output is similar to this:
LASTSEEN FIRSTSEEN COUNT NAME KIND SUBOBJECT TYPE REASON SOURCE MESSAGE
0s 0s 1 dapi-test-pod Pod Warning InvalidEnvironmentVariableNames {kubelet, 127.0.0.1} Keys [1badkey, 2alsobad] from the EnvFrom configMap default/myconfig were skipped since they are considered invalid environment variable names.
ConfigMaps reside in a specific Namespace. A ConfigMap can only be referenced by pods residing in the same namespace.
You can't use ConfigMaps for static pods, because the Kubelet does not support this.
Optional ConfigMaps
You can mark a reference to a ConfigMap as optional in a Pod specification.
If the ConfigMap doesn't exist, the configuration for which it provides data in the Pod (e.g. environment variable, mounted volume) will be empty.
If the ConfigMap exists, but the referenced key is non-existent the data is also empty.
For example, the following Pod specification marks an environment variable from a ConfigMap as optional:
apiVersion:v1kind:Podmetadata:name:dapi-test-podspec:containers:- name:test-containerimage:gcr.io/google_containers/busyboxcommand:["/bin/sh","-c","env"]env:- name:SPECIAL_LEVEL_KEYvalueFrom:configMapKeyRef:name:a-configkey:akeyoptional:true# mark the variable as optionalrestartPolicy:Never
If you run this pod, and there is no ConfigMap named a-config, the output is empty.
If you run this pod, and there is a ConfigMap named a-config but that ConfigMap doesn't have
a key named akey, the output is also empty. If you do set a value for akey in the a-config
ConfigMap, this pod prints that value and then terminates.
You can also mark the volumes and files provided by a ConfigMap as optional. Kubernetes always creates the mount paths for the volume, even if the referenced ConfigMap or key doesn't exist. For example, the following
Pod specification marks a volume that references a ConfigMap as optional:
apiVersion:v1kind:Podmetadata:name:dapi-test-podspec:containers:- name:test-containerimage:gcr.io/google_containers/busyboxcommand:["/bin/sh","-c","ls /etc/config"]volumeMounts:- name:config-volumemountPath:/etc/configvolumes:- name:config-volumeconfigMap:name:no-configoptional:true# mark the source ConfigMap as optionalrestartPolicy:Never
Mounted ConfigMaps are updated automatically
When a mounted ConfigMap is updated, the projected content is eventually updated too. This applies in the case where an optionally referenced ConfigMap comes into
existence after a pod has started.
The kubelet checks whether the mounted ConfigMap is fresh on every periodic sync. However, it uses its local TTL-based cache for getting the current value of the
ConfigMap. As a result, the total delay from the moment when the ConfigMap is updated to the moment when new keys are projected to the pod can be as long as
kubelet sync period (1 minute by default) + TTL of ConfigMaps cache (1 minute by default) in kubelet.
Note: A container using a ConfigMap as a subPath volume will not receive ConfigMap updates.
4.3.20 - Share Process Namespace between Containers in a Pod
This page shows how to configure process namespace sharing for a pod. When
process namespace sharing is enabled, processes in a container are visible
to all other containers in the same pod.
You can use this feature to configure cooperating containers, such as a log
handler sidecar container, or to troubleshoot container images that don't
include debugging utilities like a shell.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
If you don't see a command prompt, try pressing enter. In the container shell:
# run this inside the "shell" containerps ax
The output is similar to this:
PID USER TIME COMMAND
1 root 0:00 /pause
8 root 0:00 nginx: master process nginx -g daemon off;
14 101 0:00 nginx: worker process
15 root 0:00 sh
21 root 0:00 ps ax
You can signal processes in other containers. For example, send SIGHUP to
nginx to restart the worker process. This requires the SYS_PTRACE capability.
# run this inside the "shell" containerkill -HUP 8# change "8" to match the PID of the nginx leader process, if necessaryps ax
The output is similar to this:
PID USER TIME COMMAND
1 root 0:00 /pause
8 root 0:00 nginx: master process nginx -g daemon off;
15 root 0:00 sh
22 101 0:00 nginx: worker process
23 root 0:00 ps ax
It's even possible to access the file system of another container using the
/proc/$pid/root link.
# run this inside the "shell" container# change "8" to the PID of the Nginx process, if necessaryhead /proc/8/root/etc/nginx/nginx.conf
Pods share many resources so it makes sense they would also share a process
namespace. Some containers may expect to be isolated from others, though,
so it's important to understand the differences:
The container process no longer has PID 1. Some containers refuse
to start without PID 1 (for example, containers using systemd) or run
commands like kill -HUP 1 to signal the container process. In pods with a
shared process namespace, kill -HUP 1 will signal the pod sandbox
(/pause in the above example).
Processes are visible to other containers in the pod. This includes all
information visible in /proc, such as passwords that were passed as arguments
or environment variables. These are protected only by regular Unix permissions.
Container filesystems are visible to other containers in the pod through the
/proc/$pid/root link. This makes debugging easier, but it also means
that filesystem secrets are protected only by filesystem permissions.
4.3.21 - Use a User Namespace With a Pod
FEATURE STATE:Kubernetes v1.25 [alpha]
This page shows how to configure a user namespace for stateless pods. This
allows to isolate the user running inside the container from the one in the
host.
A process running as root in a container can run as a different (non-root) user
in the host; in other words, the process has full privileges for operations
inside the user namespace, but is unprivileged for operations outside the
namespace.
You can use this feature to reduce the damage a compromised container can do to
the host or other pods in the same node. There are several security
vulnerabilities rated either HIGH or CRITICAL that were not
exploitable when user namespaces is active. It is expected user namespace will
mitigate some future vulnerabilities too.
Without using a user namespace a container running as root, in the case of a
container breakout, has root privileges on the node. And if some capability were
granted to the container, the capabilities are valid on the host too. None of
this is true when user namespaces are used.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be version v1.25.
To check the version, enter kubectl version.
🛇 This item links to a third party project or product that is not part of Kubernetes itself. More information
The node OS needs to be Linux
You need to exec commands in the host
You need to be able to exec into pods
Feature gate UserNamespacesStatelessPodsSupport need to be enabled.
In addition, support is needed in the
container runtime
to use this feature with Kubernetes stateless pods:
CRI-O: v1.25 has support for user namespaces.
Please note that if your container runtime doesn't support user namespaces, the
new pod.spec field will be silently ignored and the pod will be created without
user namespaces.
Run a Pod that uses a user namespace
A user namespace for a stateless pod is enabled setting the hostUsers field of
.spec to false. For example:
Then, open a shell in the host and run the same command.
The output must be different. This means the host and the pod are using a
different user namespace. When user namespaces are not enabled, the host and the
pod use the same user namespace.
If you are running the kubelet inside a user namespace, you need to compare the
output from running the command in the pod to the output of running in the host:
readlink /proc/$pid/ns/user
user:[4026534732]
replacing `$pid` with the kubelet PID.
4.3.22 - Create static Pods
Static Pods are managed directly by the kubelet daemon on a specific node,
without the API server
observing them.
Unlike Pods that are managed by the control plane (for example, a
Deployment);
instead, the kubelet watches each static Pod (and restarts it if it fails).
Static Pods are always bound to one Kubelet on a specific node.
The kubelet automatically tries to create a mirror Pod
on the Kubernetes API server for each static Pod.
This means that the Pods running on a node are visible on the API server,
but cannot be controlled from there.
The Pod names will be suffixed with the node hostname with a leading hyphen.
Note: If you are running clustered Kubernetes and are using static
Pods to run a Pod on every node, you should probably be using a
DaemonSet
instead.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
This page assumes you're using CRI-O to run Pods,
and that your nodes are running the Fedora operating system.
Instructions for other distributions or Kubernetes installations may vary.
Manifests are standard Pod definitions in JSON or YAML format in a specific directory. Use the staticPodPath: <the directory> field in the
kubelet configuration file,
which periodically scans the directory and creates/deletes static Pods as YAML/JSON files appear/disappear there.
Note that the kubelet will ignore files starting with dots when scanning the specified directory.
For example, this is how to start a simple web server as a static Pod:
Choose a node where you want to run the static Pod. In this example, it's my-node1.
ssh my-node1
Choose a directory, say /etc/kubernetes/manifests and place a web server Pod definition there, for example /etc/kubernetes/manifests/static-web.yaml:
# Run this command on the node where kubelet is runningmkdir -p /etc/kubernetes/manifests/
cat <<EOF >/etc/kubernetes/manifests/static-web.yaml
apiVersion: v1
kind: Pod
metadata:
name: static-web
labels:
role: myrole
spec:
containers:
- name: web
image: nginx
ports:
- name: web
containerPort: 80
protocol: TCP
EOF
Configure your kubelet on the node to use this directory by running it with --pod-manifest-path=/etc/kubernetes/manifests/ argument. On Fedora edit /etc/kubernetes/kubelet to include this line:
# Run this command on the node where the kubelet is runningsystemctl restart kubelet
Web-hosted static pod manifest
Kubelet periodically downloads a file specified by --manifest-url=<URL> argument
and interprets it as a JSON/YAML file that contains Pod definitions.
Similar to how filesystem-hosted manifests work, the kubelet
refetches the manifest on a schedule. If there are changes to the list of static
Pods, the kubelet applies them.
To use this approach:
Create a YAML file and store it on a web server so that you can pass the URL of that file to the kubelet.
Configure the kubelet on your selected node to use this web manifest by running it with --manifest-url=<manifest-url>. On Fedora, edit /etc/kubernetes/kubelet to include this line:
# Run this command on the node where the kubelet is runningsystemctl restart kubelet
Observe static pod behavior
When the kubelet starts, it automatically starts all defined static Pods. As you have
defined a static Pod and restarted the kubelet, the new static Pod should
already be running.
You can view running containers (including static Pods) by running (on the node):
# Run this command on the node where the kubelet is runningcrictl ps
The output might be something like:
CONTAINER IMAGE CREATED STATE NAME ATTEMPT POD ID
129fd7d382018 docker.io/library/nginx@sha256:... 11 minutes ago Running web 0 34533c6729106
Note:crictl outputs the image URI and SHA-256 checksum. NAME will look more like:
docker.io/library/nginx@sha256:0d17b565c37bcbd895e9d92315a05c1c3c9a29f762b011a10c54a66cd53c9b31.
You can see the mirror Pod on the API server:
kubectl get pods
NAME READY STATUS RESTARTS AGE
static-web 1/1 Running 0 2m
Note: Make sure the kubelet has permission to create the mirror Pod in the API server. If not, the creation request is rejected by the API server.
Labels from the static Pod are
propagated into the mirror Pod. You can use those labels as normal via
selectors, etc.
If you try to use kubectl to delete the mirror Pod from the API server,
the kubelet doesn't remove the static Pod:
kubectl delete pod static-web
pod "static-web" deleted
You can see that the Pod is still running:
kubectl get pods
NAME READY STATUS RESTARTS AGE
static-web 1/1 Running 0 4s
Back on your node where the kubelet is running, you can try to stop the container manually.
You'll see that, after a time, the kubelet will notice and will restart the Pod
automatically:
# Run these commands on the node where the kubelet is runningcrictl stop 129fd7d382018 # replace with the ID of your containersleep 20crictl ps
CONTAINER IMAGE CREATED STATE NAME ATTEMPT POD ID
89db4553e1eeb docker.io/library/nginx@sha256:... 19 seconds ago Running web 1 34533c6729106
Dynamic addition and removal of static pods
The running kubelet periodically scans the configured directory (/etc/kubernetes/manifests in our example) for changes and adds/removes Pods as files appear/disappear in this directory.
# This assumes you are using filesystem-hosted static Pod configuration# Run these commands on the node where the kubelet is running#mv /etc/kubernetes/manifests/static-web.yaml /tmp
sleep 20crictl ps
# You see that no nginx container is runningmv /tmp/static-web.yaml /etc/kubernetes/manifests/
sleep 20crictl ps
CONTAINER IMAGE CREATED STATE NAME ATTEMPT POD ID
f427638871c35 docker.io/library/nginx@sha256:... 19 seconds ago Running web 1 34533c6729106
4.3.23 - Translate a Docker Compose File to Kubernetes Resources
What's Kompose? It's a conversion tool for all things compose (namely Docker Compose) to container orchestrators (Kubernetes or OpenShift).
More information can be found on the Kompose website at http://kompose.io.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Installing using go get pulls from the master branch with the latest development changes.
go get -u github.com/kubernetes/kompose
Kompose is in EPEL CentOS repository.
If you don't have EPEL repository already installed and enabled you can do it by running sudo yum install epel-release
If you have EPEL enabled in your system, you can install Kompose like any other package.
sudo yum -y install kompose
Kompose is in Fedora 24, 25 and 26 repositories. You can install it like any other package.
sudo dnf -y install kompose
On macOS you can install latest release via Homebrew:
brew install kompose
Use Kompose
In a few steps, we'll take you from Docker Compose to Kubernetes. All
you need is an existing docker-compose.yml file.
Go to the directory containing your docker-compose.yml file. If you don't have one, test using this one.
To convert the docker-compose.yml file to files that you can use with
kubectl, run kompose convert and then kubectl apply -f <output file>.
kompose convert
The output is similar to:
INFO Kubernetes file "frontend-service.yaml" created
INFO Kubernetes file "frontend-service.yaml" created
INFO Kubernetes file "frontend-service.yaml" created
INFO Kubernetes file "redis-master-service.yaml" created
INFO Kubernetes file "redis-master-service.yaml" created
INFO Kubernetes file "redis-master-service.yaml" created
INFO Kubernetes file "redis-slave-service.yaml" created
INFO Kubernetes file "redis-slave-service.yaml" created
INFO Kubernetes file "redis-slave-service.yaml" created
INFO Kubernetes file "frontend-deployment.yaml" created
INFO Kubernetes file "frontend-deployment.yaml" created
INFO Kubernetes file "frontend-deployment.yaml" created
INFO Kubernetes file "redis-master-deployment.yaml" created
INFO Kubernetes file "redis-master-deployment.yaml" created
INFO Kubernetes file "redis-master-deployment.yaml" created
INFO Kubernetes file "redis-slave-deployment.yaml" created
INFO Kubernetes file "redis-slave-deployment.yaml" created
INFO Kubernetes file "redis-slave-deployment.yaml" created
service/frontend created
service/redis-master created
service/redis-slave created
deployment.apps/frontend created
deployment.apps/redis-master created
deployment.apps/redis-slave created
Your deployments are running in Kubernetes.
Access your application.
If you're already using minikube for your development process:
minikube service frontend
Otherwise, let's look up what IP your service is using!
Kompose has support for two providers: OpenShift and Kubernetes.
You can choose a targeted provider using global option --provider. If no provider is specified, Kubernetes is set by default.
kompose convert
Kompose supports conversion of V1, V2, and V3 Docker Compose files into Kubernetes and OpenShift objects.
Kubernetes kompose convert example
kompose --file docker-voting.yml convert
WARN Unsupported key networks - ignoring
WARN Unsupported key build - ignoring
INFO Kubernetes file "worker-svc.yaml" created
INFO Kubernetes file "db-svc.yaml" created
INFO Kubernetes file "redis-svc.yaml" created
INFO Kubernetes file "result-svc.yaml" created
INFO Kubernetes file "vote-svc.yaml" created
INFO Kubernetes file "redis-deployment.yaml" created
INFO Kubernetes file "result-deployment.yaml" created
INFO Kubernetes file "vote-deployment.yaml" created
INFO Kubernetes file "worker-deployment.yaml" created
INFO Kubernetes file "db-deployment.yaml" created
INFO Kubernetes file "frontend-service.yaml" created
INFO Kubernetes file "mlbparks-service.yaml" created
INFO Kubernetes file "mongodb-service.yaml" created
INFO Kubernetes file "redis-master-service.yaml" created
INFO Kubernetes file "redis-slave-service.yaml" created
INFO Kubernetes file "frontend-deployment.yaml" created
INFO Kubernetes file "mlbparks-deployment.yaml" created
INFO Kubernetes file "mongodb-deployment.yaml" created
INFO Kubernetes file "mongodb-claim0-persistentvolumeclaim.yaml" created
INFO Kubernetes file "redis-master-deployment.yaml" created
INFO Kubernetes file "redis-slave-deployment.yaml" created
WARN [worker] Service cannot be created because of missing port.
INFO OpenShift file "vote-service.yaml" created
INFO OpenShift file "db-service.yaml" created
INFO OpenShift file "redis-service.yaml" created
INFO OpenShift file "result-service.yaml" created
INFO OpenShift file "vote-deploymentconfig.yaml" created
INFO OpenShift file "vote-imagestream.yaml" created
INFO OpenShift file "worker-deploymentconfig.yaml" created
INFO OpenShift file "worker-imagestream.yaml" created
INFO OpenShift file "db-deploymentconfig.yaml" created
INFO OpenShift file "db-imagestream.yaml" created
INFO OpenShift file "redis-deploymentconfig.yaml" created
INFO OpenShift file "redis-imagestream.yaml" created
INFO OpenShift file "result-deploymentconfig.yaml" created
INFO OpenShift file "result-imagestream.yaml" created
It also supports creating buildconfig for build directive in a service. By default, it uses the remote repo for the current git branch as the source repo, and the current branch as the source branch for the build. You can specify a different source repo and branch using --build-repo and --build-branch options respectively.
WARN [foo] Service cannot be created because of missing port.
INFO OpenShift Buildconfig using git@github.com:rtnpro/kompose.git::master as source.
INFO OpenShift file "foo-deploymentconfig.yaml" created
INFO OpenShift file "foo-imagestream.yaml" created
INFO OpenShift file "foo-buildconfig.yaml" created
Note: If you are manually pushing the OpenShift artifacts using oc create -f, you need to ensure that you push the imagestream artifact before the buildconfig artifact, to workaround this OpenShift issue: https://github.com/openshift/origin/issues/4518 .
Alternative Conversions
The default kompose transformation will generate Kubernetes Deployments and Services, in yaml format. You have alternative option to generate json with -j. Also, you can alternatively generate Replication Controllers objects, Daemon Sets, or Helm charts.
kompose convert -j
INFO Kubernetes file "redis-svc.json" created
INFO Kubernetes file "web-svc.json" created
INFO Kubernetes file "redis-deployment.json" created
INFO Kubernetes file "web-deployment.json" created
The *-deployment.json files contain the Deployment objects.
kompose convert --replication-controller
INFO Kubernetes file "redis-svc.yaml" created
INFO Kubernetes file "web-svc.yaml" created
INFO Kubernetes file "redis-replicationcontroller.yaml" created
INFO Kubernetes file "web-replicationcontroller.yaml" created
The *-replicationcontroller.yaml files contain the Replication Controller objects. If you want to specify replicas (default is 1), use --replicas flag: kompose convert --replication-controller --replicas 3
kompose convert --daemon-set
INFO Kubernetes file "redis-svc.yaml" created
INFO Kubernetes file "web-svc.yaml" created
INFO Kubernetes file "redis-daemonset.yaml" created
INFO Kubernetes file "web-daemonset.yaml" created
The *-daemonset.yaml files contain the DaemonSet objects
If you want to generate a Chart to be used with Helm run:
kompose convert -c
INFO Kubernetes file "web-svc.yaml" created
INFO Kubernetes file "redis-svc.yaml" created
INFO Kubernetes file "web-deployment.yaml" created
INFO Kubernetes file "redis-deployment.yaml" created
chart created in "./docker-compose/"
kompose.service.expose defines if the service needs to be made accessible from outside the cluster or not. If the value is set to "true", the provider sets the endpoint automatically, and for any other value, the value is set as the hostname. If multiple ports are defined in a service, the first one is chosen to be the exposed.
For the Kubernetes provider, an ingress resource is created and it is assumed that an ingress controller has already been configured.
Note: The kompose.service.type label should be defined with ports only, otherwise kompose will fail.
Restart
If you want to create normal pods without controllers you can use restart construct of docker-compose to define that. Follow table below to see what happens on the restart value.
docker-composerestart
object created
Pod restartPolicy
""
controller object
Always
always
controller object
Always
on-failure
Pod
OnFailure
no
Pod
Never
Note: The controller object could be deployment or replicationcontroller.
For example, the pival service will become pod down here. This container calculated value of pi.
If the Docker Compose file has a volume specified for a service, the Deployment (Kubernetes) or DeploymentConfig (OpenShift) strategy is changed to "Recreate" instead of "RollingUpdate" (default). This is done to avoid multiple instances of a service from accessing a volume at the same time.
If the Docker Compose file has service name with _ in it (eg.web_service), then it will be replaced by - and the service name will be renamed accordingly (eg.web-service). Kompose does this because "Kubernetes" doesn't allow _ in object name.
Please note that changing service name might break some docker-compose files.
Docker Compose Versions
Kompose supports Docker Compose versions: 1, 2 and 3. We have limited support on versions 2.1 and 3.2 due to their experimental nature.
A full list on compatibility between all three versions is listed in our conversion document including a list of all incompatible Docker Compose keys.
4.3.24 - Enforce Pod Security Standards by Configuring the Built-in Admission Controller
apiVersion:apiserver.config.k8s.io/v1kind:AdmissionConfigurationplugins:- name:PodSecurityconfiguration:apiVersion:pod-security.admission.config.k8s.io/v1beta1kind:PodSecurityConfiguration# Defaults applied when a mode label is not set.## Level label values must be one of:# - "privileged" (default)# - "baseline"# - "restricted"## Version label values must be one of:# - "latest" (default) # - specific version like "v1.25"defaults:enforce:"privileged"enforce-version:"latest"audit:"privileged"audit-version:"latest"warn:"privileged"warn-version:"latest"exemptions:# Array of authenticated usernames to exempt.usernames:[]# Array of runtime class names to exempt.runtimeClasses:[]# Array of namespaces to exempt.namespaces:[]
Note: v1beta1 configuration requires v1.23+. For v1.22, use v1alpha1.
apiVersion:apiserver.config.k8s.io/v1kind:AdmissionConfigurationplugins:- name:PodSecurityconfiguration:apiVersion:pod-security.admission.config.k8s.io/v1alpha1kind:PodSecurityConfiguration# Defaults applied when a mode label is not set.## Level label values must be one of:# - "privileged" (default)# - "baseline"# - "restricted"## Version label values must be one of:# - "latest" (default) # - specific version like "v1.25"defaults:enforce:"privileged"enforce-version:"latest"audit:"privileged"audit-version:"latest"warn:"privileged"warn-version:"latest"exemptions:# Array of authenticated usernames to exempt.usernames:[]# Array of runtime class names to exempt.runtimeClasses:[]# Array of namespaces to exempt.namespaces:[]
4.3.25 - Enforce Pod Security Standards with Namespace Labels
Requiring the baseline Pod Security Standard with namespace labels
This manifest defines a Namespace my-baseline-namespace that:
Blocks any pods that don't satisfy the baseline policy requirements.
Generates a user-facing warning and adds an audit annotation to any created pod that does not
meet the restricted policy requirements.
Pins the versions of the baseline and restricted policies to v1.25.
apiVersion:v1kind:Namespacemetadata:name:my-baseline-namespacelabels:pod-security.kubernetes.io/enforce:baselinepod-security.kubernetes.io/enforce-version:v1.25# We are setting these to our _desired_ `enforce` level.pod-security.kubernetes.io/audit:restrictedpod-security.kubernetes.io/audit-version:v1.25pod-security.kubernetes.io/warn:restrictedpod-security.kubernetes.io/warn-version:v1.25
Add labels to existing namespaces with kubectl label
Note: When an enforce policy (or version) label is added or changed, the admission plugin will test
each pod in the namespace against the new policy. Violations are returned to the user as warnings.
It is helpful to apply the --dry-run flag when initially evaluating security profile changes for
namespaces. The Pod Security Standard checks will still be run in dry run mode, giving you
information about how the new policy would treat existing pods, without actually updating a policy.
If you're just getting started with the Pod Security Standards, a suitable first step would be to
configure all namespaces with audit annotations for a stricter level such as baseline:
Note that this is not setting an enforce level, so that namespaces that haven't been explicitly
evaluated can be distinguished. You can list namespaces without an explicitly set enforce level
using this command:
kubectl get namespaces --selector='!pod-security.kubernetes.io/enforce'
Applying to a single namespace
You can update a specific namespace as well. This command adds the enforce=restricted
policy to my-existing-namespace, pinning the restricted policy version to v1.25.
4.3.26 - Migrate from PodSecurityPolicy to the Built-In PodSecurity Admission Controller
This page describes the process of migrating from PodSecurityPolicies to the built-in PodSecurity
admission controller. This can be done effectively using a combination of dry-run and audit and
warn modes, although this becomes harder if mutating PSPs are used.
Before you begin
Your Kubernetes server must be at or later than version v1.22.
To check the version, enter kubectl version.
This page assumes you are already familiar with the basic Pod Security Admission
concepts.
Overall approach
There are multiple strategies you can take for migrating from PodSecurityPolicy to Pod Security
Admission. The following steps are one possible migration path, with a goal of minimizing both the
risks of a production outage and of a security gap.
Decide whether Pod Security Admission is the right fit for your use case.
Review namespace permissions
Simplify & standardize PodSecurityPolicies
Update namespaces
Identify an appropriate Pod Security level
Verify the Pod Security level
Enforce the Pod Security level
Bypass PodSecurityPolicy
Review namespace creation processes
Disable PodSecurityPolicy
0. Decide whether Pod Security Admission is right for you
Pod Security Admission was designed to meet the most common security needs out of the box, and to
provide a standard set of security levels across clusters. However, it is less flexible than
PodSecurityPolicy. Notably, the following features are supported by PodSecurityPolicy but not Pod
Security Admission:
Setting default security constraints - Pod Security Admission is a non-mutating admission
controller, meaning it won't modify pods before validating them. If you were relying on this
aspect of PSP, you will need to either modify your workloads to meet the Pod Security constraints,
or use a Mutating Admission Webhook
to make those changes. See Simplify & Standardize PodSecurityPolicies below for more detail.
Fine-grained control over policy definition - Pod Security Admission only supports
3 standard levels.
If you require more control over specific constraints, then you will need to use a
Validating Admission Webhook
to enforce those policies.
Sub-namespace policy granularity - PodSecurityPolicy lets you bind different policies to
different Service Accounts or users, even within a single namespace. This approach has many
pitfalls and is not recommended, but if you require this feature anyway you will
need to use a 3rd party webhook instead. The exception to this is if you only need to completely exempt
specific users or RuntimeClasses, in which case Pod
Security Admission does expose some
static configuration for exemptions.
Even if Pod Security Admission does not meet all of your needs it was designed to be complementary
to other policy enforcement mechanisms, and can provide a useful fallback running alongside other
admission webhooks.
1. Review namespace permissions
Pod Security Admission is controlled by labels on
namespaces.
This means that anyone who can update (or patch or create) a namespace can also modify the Pod
Security level for that namespace, which could be used to bypass a more restrictive policy. Before
proceeding, ensure that only trusted, privileged users have these namespace permissions. It is not
recommended to grant these powerful permissions to users that shouldn't have elevated permissions,
but if you must you will need to use an
admission webhook
to place additional restrictions on setting Pod Security labels on Namespace objects.
2. Simplify & standardize PodSecurityPolicies
In this section, you will reduce mutating PodSecurityPolicies and remove options that are outside
the scope of the Pod Security Standards. You should make the changes recommended here to an offline
copy of the original PodSecurityPolicy being modified. The cloned PSP should have a different
name that is alphabetically before the original (for example, prepend a 0 to it). Do not create the
new policies in Kubernetes yet - that will be covered in the Rollout the updated
policies section below.
2.a. Eliminate purely mutating fields
If a PodSecurityPolicy is mutating pods, then you could end up with pods that don't meet the Pod
Security level requirements when you finally turn PodSecurityPolicy off. In order to avoid this, you
should eliminate all PSP mutation prior to switching over. Unfortunately PSP does not cleanly
separate mutating & validating fields, so this is not a straightforward migration.
You can start by eliminating the fields that are purely mutating, and don't have any bearing on the
validating policy. These fields (also listed in the
Mapping PodSecurityPolicies to Pod Security Standards
reference) are:
.spec.defaultAddCapabilities - Although technically a mutating & validating field, these should
be merged into .spec.allowedCapabilities which performs the same validation without mutation.
Caution: Removing these could result in workloads missing required configuration, and cause problems. See
Rollout the updated policies below for advice on how to roll these changes
out safely.
2.b. Eliminate options not covered by the Pod Security Standards
There are several fields in PodSecurityPolicy that are not covered by the Pod Security Standards. If
you must enforce these options, you will need to supplement Pod Security Admission with an
admission webhook,
which is outside the scope of this guide.
First, you can remove the purely validating fields that the Pod Security Standards do not cover.
These fields (also listed in the
Mapping PodSecurityPolicies to Pod Security Standards
reference with "no opinion") are:
.spec.allowedHostPaths
.spec.allowedFlexVolumes
.spec.allowedCSIDrivers
.spec.forbiddenSysctls
.spec.runtimeClass
You can also remove the following fields, that are related to POSIX / UNIX group controls.
Caution: If any of these use the MustRunAs strategy they may be mutating! Removing these could result in
workloads not setting the required groups, and cause problems. See
Rollout the updated policies below for advice on how to roll these changes
out safely.
.spec.runAsGroup
.spec.supplementalGroups
.spec.fsGroup
The remaining mutating fields are required to properly support the Pod Security Standards, and will
need to be handled on a case-by-case basis later:
.spec.requiredDropCapabilities - Required to drop ALL for the Restricted profile.
.spec.seLinux - (Only mutating with the MustRunAs rule) required to enforce the SELinux
requirements of the Baseline & Restricted profiles.
.spec.runAsUser - (Non-mutating with the RunAsAny rule) required to enforce RunAsNonRoot for
the Restricted profile.
.spec.allowPrivilegeEscalation - (Only mutating if set to false) required for the Restricted
profile.
2.c. Rollout the updated PSPs
Next, you can rollout the updated policies to your cluster. You should proceed with caution, as
removing the mutating options may result in workloads missing required configuration.
For each updated PodSecurityPolicy:
Identify pods running under the original PSP. This can be done using the kubernetes.io/psp
annotation. For example, using kubectl:
PSP_NAME="original"# Set the name of the PSP you're checking forkubectl get pods --all-namespaces -o jsonpath="{range .items[?(@.metadata.annotations.kubernetes\.io\/psp=='$PSP_NAME')]}{.metadata.namespace} {.metadata.name}{'\n'}{end}"
Compare these running pods against the original pod spec to determine whether PodSecurityPolicy
has modified the pod. For pods created by a workload resource
you can compare the pod with the PodTemplate in the controller resource. If any changes are
identified, the original Pod or PodTemplate should be updated with the desired configuration.
The fields to review are:
.metadata.annotations['container.apparmor.security.beta.kubernetes.io/*'] (replace * with each container name)
.spec.runtimeClassName
.spec.securityContext.fsGroup
.spec.securityContext.seccompProfile
.spec.securityContext.seLinuxOptions
.spec.securityContext.supplementalGroups
On containers, under .spec.containers[*] and .spec.initContainers[*]:
.securityContext.allowPrivilegeEscalation
.securityContext.capabilities.add
.securityContext.capabilities.drop
.securityContext.readOnlyRootFilesystem
.securityContext.runAsGroup
.securityContext.runAsNonRoot
.securityContext.runAsUser
.securityContext.seccompProfile
.securityContext.seLinuxOptions
Create the new PodSecurityPolicies. If any Roles or ClusterRoles are granting use on all PSPs
this could cause the new PSPs to be used instead of their mutating counter-parts.
Update your authorization to grant access to the new PSPs. In RBAC this means updating any Roles
or ClusterRoles that grant the use permision on the original PSP to also grant it to the
updated PSP.
Verify: after some soak time, rerun the command from step 1 to see if any pods are still using
the original PSPs. Note that pods need to be recreated after the new policies have been rolled
out before they can be fully verified.
(optional) Once you have verified that the original PSPs are no longer in use, you can delete
them.
3. Update Namespaces
The following steps will need to be performed on every namespace in the cluster. Commands referenced
in these steps use the $NAMESPACE variable to refer to the namespace being updated.
3.a. Identify an appropriate Pod Security level
Start reviewing the Pod Security Standards and
familiarizing yourself with the 3 different levels.
There are several ways to choose a Pod Security level for your namespace:
By security requirements for the namespace - If you are familiar with the expected access
level for the namespace, you can choose an appropriate level based on those requirements, similar
to how one might approach this on a new cluster.
By existing PodSecurityPolicies - Using the
Mapping PodSecurityPolicies to Pod Security Standards
reference you can map each
PSP to a Pod Security Standard level. If your PSPs aren't based on the Pod Security Standards, you
may need to decide between choosing a level that is at least as permissive as the PSP, and a
level that is at least as restrictive. You can see which PSPs are in use for pods in a given
namespace with this command:
By existing pods - Using the strategies under Verify the Pod Security level,
you can test out both the Baseline and Restricted levels to see
whether they are sufficiently permissive for existing workloads, and chose the least-privileged
valid level.
Caution: Options 2 & 3 above are based on existing pods, and may miss workloads that aren't currently
running, such as CronJobs, scale-to-zero workloads, or other workloads that haven't rolled out.
3.b. Verify the Pod Security level
Once you have selected a Pod Security level for the namespace (or if you're trying several), it's a
good idea to test it out first (you can skip this step if using the Privileged level). Pod Security
includes several tools to help test and safely roll out profiles.
First, you can dry-run the policy, which will evaluate pods currently running in the namespace
against the applied policy, without making the new policy take effect:
# $LEVEL is the level to dry-run, either "baseline" or "restricted".kubectl label --dry-run=server --overwrite ns $NAMESPACE pod-security.kubernetes.io/enforce=$LEVEL
This command will return a warning for any existing pods that are not valid under the proposed
level.
The second option is better for catching workloads that are not currently running: audit mode. When
running under audit-mode (as opposed to enforcing), pods that violate the policy level are recorded
in the audit logs, which can be reviewed later after some soak time, but are not forbidden. Warning
mode works similarly, but returns the warning to the user immediately. You can set the audit level
on a namespace with this command:
If either of these approaches yield unexpected violations, you will need to either update the
violating workloads to meet the policy requirements, or relax the namespace Pod Security level.
3.c. Enforce the Pod Security level
When you are satisfied that the chosen level can safely be enforced on the namespace, you can update
the namespace to enforce the desired level:
Finally, you can effectively bypass PodSecurityPolicy at the namespace level by binding the fully
privileged PSP
to all service
accounts in the namespace.
# The following cluster-scoped commands are only needed once.kubectl apply -f privileged-psp.yaml
kubectl create clusterrole privileged-psp --verb use --resource podsecuritypolicies.policy --resource-name privileged
# Per-namespace disablekubectl create -n $NAMESPACE rolebinding disable-psp --clusterrole privileged-psp --group system:serviceaccounts:$NAMESPACE
Since the privileged PSP is non-mutating, and the PSP admission controller always
prefers non-mutating PSPs, this will ensure that pods in this namespace are no longer being modified
or restricted by PodSecurityPolicy.
The advantage to disabling PodSecurityPolicy on a per-namespace basis like this is if a problem
arises you can easily roll the change back by deleting the RoleBinding. Just make sure the
pre-existing PodSecurityPolicies are still in place!
Now that existing namespaces have been updated to enforce Pod Security Admission, you should ensure
that your processes and/or policies for creating new namespaces are updated to ensure that an
appropriate Pod Security profile is applied to new namespaces.
You can also statically configure the Pod Security admission controller to set a default enforce,
audit, and/or warn level for unlabeled namespaces. See
Configure the Admission Controller
for more information.
5. Disable PodSecurityPolicy
Finally, you're ready to disable PodSecurityPolicy. To do so, you will need to modify the admission
configuration of the API server:
How do I turn off an admission controller?.
To verify that the PodSecurityPolicy admission controller is no longer enabled, you can manually run
a test by impersonating a user without access to any PodSecurityPolicies (see the
PodSecurityPolicy example), or by verifying in
the API server logs. At startup, the API server outputs log lines listing the loaded admission
controller plugins:
I0218 00:59:44.903329 13 plugins.go:158] Loaded 16 mutating admission controller(s) successfully in the following order: NamespaceLifecycle,LimitRanger,ServiceAccount,NodeRestriction,TaintNodesByCondition,Priority,DefaultTolerationSeconds,ExtendedResourceToleration,PersistentVolumeLabel,DefaultStorageClass,StorageObjectInUseProtection,RuntimeClass,DefaultIngressClass,MutatingAdmissionWebhook.
I0218 00:59:44.903350 13 plugins.go:161] Loaded 14 validating admission controller(s) successfully in the following order: LimitRanger,ServiceAccount,PodSecurity,Priority,PersistentVolumeClaimResize,RuntimeClass,CertificateApproval,CertificateSigning,CertificateSubjectRestriction,DenyServiceExternalIPs,ValidatingAdmissionWebhook,ResourceQuota.
You should see PodSecurity (in the validating admission controllers), and neither list should
contain PodSecurityPolicy.
Once you are certain the PSP admission controller is disabled (and after sufficient soak time to be
confident you won't need to roll back), you are free to delete your PodSecurityPolicies and any
associated Roles, ClusterRoles, RoleBindings and ClusterRoleBindings (just make sure they don't
grant any other unrelated permissions).
4.4 - Monitoring, Logging, and Debugging
Set up monitoring and logging to troubleshoot a cluster, or debug a containerized application.
Sometimes things go wrong. This guide is aimed at making them right. It has
two sections:
Debugging your application - Useful
for users who are deploying code into Kubernetes and wondering why it is not working.
Debugging your cluster - Useful
for cluster administrators and people whose Kubernetes cluster is unhappy.
You should also check the known issues for the release
you're using.
Getting help
If your problem isn't answered by any of the guides above, there are variety of
ways for you to get help from the Kubernetes community.
Questions
The documentation on this site has been structured to provide answers to a wide
range of questions. Concepts explain the Kubernetes
architecture and how each component works, while Setup provides
practical instructions for getting started. Tasks show how to
accomplish commonly used tasks, and Tutorials are more
comprehensive walkthroughs of real-world, industry-specific, or end-to-end
development scenarios. The Reference section provides
detailed documentation on the Kubernetes API
and command-line interfaces (CLIs), such as kubectl.
Many people from the Kubernetes community hang out on Kubernetes Slack in the #kubernetes-users channel.
Slack requires registration; you can request an invitation,
and registration is open to everyone). Feel free to come and ask any and all questions.
Once registered, access the Kubernetes organisation in Slack
via your web browser or via Slack's own dedicated app.
Once you are registered, browse the growing list of channels for various subjects of
interest. For example, people new to Kubernetes may also want to join the
#kubernetes-novice channel. As another example, developers should join the
#kubernetes-dev channel.
There are also many country specific / local language channels. Feel free to join
these channels for localized support and info:
If you have what looks like a bug, or you would like to make a feature request,
please use the GitHub issue tracking system.
Before you file an issue, please search existing issues to see if your issue is
already covered.
If filing a bug, please include detailed information about how to reproduce the
problem, such as:
Kubernetes version: kubectl version
Cloud provider, OS distro, network configuration, and container runtime version
Steps to reproduce the problem
4.4.1 - Troubleshooting Applications
Debugging common containerized application issues.
This doc contains a set of resources for fixing issues with containerized applications. It covers things like common issues with Kubernetes resources (like Pods, Services, or StatefulSets), advice on making sense of container termination messages, and ways to debug running containers.
4.4.1.1 - Debug Pods
This guide is to help users debug applications that are deployed into Kubernetes and not behaving correctly.
This is not a guide for people who want to debug their cluster. For that you should check out
this guide.
Diagnosing the problem
The first step in troubleshooting is triage. What is the problem? Is it your Pods, your Replication Controller or
your Service?
The first step in debugging a Pod is taking a look at it. Check the current state of the Pod and recent events with the following command:
kubectl describe pods ${POD_NAME}
Look at the state of the containers in the pod. Are they all Running? Have there been recent restarts?
Continue debugging depending on the state of the pods.
My pod stays pending
If a Pod is stuck in Pending it means that it can not be scheduled onto a node. Generally this is because
there are insufficient resources of one type or another that prevent scheduling. Look at the output of the
kubectl describe ... command above. There should be messages from the scheduler about why it can not schedule
your pod. Reasons include:
You don't have enough resources: You may have exhausted the supply of CPU or Memory in your cluster, in this case
you need to delete Pods, adjust resource requests, or add new nodes to your cluster. See
Compute Resources document for more information.
You are using hostPort: When you bind a Pod to a hostPort there are a limited number of places that pod can be
scheduled. In most cases, hostPort is unnecessary, try using a Service object to expose your Pod. If you do require
hostPort then you can only schedule as many Pods as there are nodes in your Kubernetes cluster.
My pod stays waiting
If a Pod is stuck in the Waiting state, then it has been scheduled to a worker node, but it can't run on that machine.
Again, the information from kubectl describe ... should be informative. The most common cause of Waiting pods is a failure to pull the image. There are three things to check:
Make sure that you have the name of the image correct.
Have you pushed the image to the registry?
Try to manually pull the image to see if the image can be pulled. For example,
if you use Docker on your PC, run docker pull <image>.
My pod is crashing or otherwise unhealthy
Once your pod has been scheduled, the methods described in Debug Running Pods are available for debugging.
My pod is running but not doing what I told it to do
If your pod is not behaving as you expected, it may be that there was an error in your
pod description (e.g. mypod.yaml file on your local machine), and that the error
was silently ignored when you created the pod. Often a section of the pod description
is nested incorrectly, or a key name is typed incorrectly, and so the key is ignored.
For example, if you misspelled command as commnd then the pod will be created but
will not use the command line you intended it to use.
The first thing to do is to delete your pod and try creating it again with the --validate option.
For example, run kubectl apply --validate -f mypod.yaml.
If you misspelled command as commnd then will give an error like this:
I0805 10:43:25.129850 46757 schema.go:126] unknown field: commnd
I0805 10:43:25.129973 46757 schema.go:129] this may be a false alarm, see https://github.com/kubernetes/kubernetes/issues/6842
pods/mypod
The next thing to check is whether the pod on the apiserver
matches the pod you meant to create (e.g. in a yaml file on your local machine).
For example, run kubectl get pods/mypod -o yaml > mypod-on-apiserver.yaml and then
manually compare the original pod description, mypod.yaml with the one you got
back from apiserver, mypod-on-apiserver.yaml. There will typically be some
lines on the "apiserver" version that are not on the original version. This is
expected. However, if there are lines on the original that are not on the apiserver
version, then this may indicate a problem with your pod spec.
Debugging Replication Controllers
Replication controllers are fairly straightforward. They can either create Pods or they can't. If they can't
create pods, then please refer to the instructions above to debug your pods.
You can also use kubectl describe rc ${CONTROLLER_NAME} to introspect events related to the replication
controller.
Debugging Services
Services provide load balancing across a set of pods. There are several common problems that can make Services
not work properly. The following instructions should help debug Service problems.
First, verify that there are endpoints for the service. For every Service object, the apiserver makes an endpoints resource available.
You can view this resource with:
kubectl get endpoints ${SERVICE_NAME}
Make sure that the endpoints match up with the number of pods that you expect to be members of your service.
For example, if your Service is for an nginx container with 3 replicas, you would expect to see three different
IP addresses in the Service's endpoints.
My service is missing endpoints
If you are missing endpoints, try listing pods using the labels that Service uses. Imagine that you have
a Service where the labels are:
...spec:- selector:name:nginxtype:frontend
You can use:
kubectl get pods --selector=name=nginx,type=frontend
to list pods that match this selector. Verify that the list matches the Pods that you expect to provide your Service.
Verify that the pod's containerPort matches up with the Service's targetPort
If none of the above solves your problem, follow the instructions in
Debugging Service document
to make sure that your Service is running, has Endpoints, and your Pods are
actually serving; you have DNS working, iptables rules installed, and kube-proxy
does not seem to be misbehaving.
An issue that comes up rather frequently for new installations of Kubernetes is
that a Service is not working properly. You've run your Pods through a
Deployment (or other workload controller) and created a Service, but you
get no response when you try to access it. This document will hopefully help
you to figure out what's going wrong.
Running commands in a Pod
For many steps here you will want to see what a Pod running in the cluster
sees. The simplest way to do this is to run an interactive busybox Pod:
kubectl run -it --rm --restart=Never busybox --image=gcr.io/google-containers/busybox sh
Note: If you don't see a command prompt, try pressing enter.
If you already have a running Pod that you prefer to use, you can run a
command in it using:
For the purposes of this walk-through, let's run some Pods. Since you're
probably debugging your own Service you can substitute your own details, or you
can follow along and get a second data point.
The label "app" is automatically set by kubectl create deployment to the name of the
Deployment.
You can confirm your Pods are running:
kubectl get pods -l app=hostnames
NAME READY STATUS RESTARTS AGE
hostnames-632524106-bbpiw 1/1 Running 0 2m
hostnames-632524106-ly40y 1/1 Running 0 2m
hostnames-632524106-tlaok 1/1 Running 0 2m
You can also confirm that your Pods are serving. You can get the list of
Pod IP addresses and test them directly.
kubectl get pods -l app=hostnames \
-o go-template='{{range .items}}{{.status.podIP}}{{"\n"}}{{end}}'
10.244.0.5
10.244.0.6
10.244.0.7
The example container used for this walk-through serves its own hostname
via HTTP on port 9376, but if you are debugging your own app, you'll want to
use whatever port number your Pods are listening on.
From within a pod:
for ep in 10.244.0.5:9376 10.244.0.6:9376 10.244.0.7:9376; do wget -qO- $epdone
If you are not getting the responses you expect at this point, your Pods
might not be healthy or might not be listening on the port you think they are.
You might find kubectl logs to be useful for seeing what is happening, or
perhaps you need to kubectl exec directly into your Pods and debug from
there.
Assuming everything has gone to plan so far, you can start to investigate why
your Service doesn't work.
Does the Service exist?
The astute reader will have noticed that you did not actually create a Service
yet - that is intentional. This is a step that sometimes gets forgotten, and
is the first thing to check.
What would happen if you tried to access a non-existent Service? If
you have another Pod that consumes this Service by name you would get
something like:
wget -O- hostnames
Resolving hostnames (hostnames)... failed: Name or service not known.
wget: unable to resolve host address 'hostnames'
The first thing to check is whether that Service actually exists:
kubectl get svc hostnames
No resources found.
Error from server (NotFound): services "hostnames" not found
Let's create the Service. As before, this is for the walk-through - you can
use your own Service's details here.
In order to highlight the full range of configuration, the Service you created
here uses a different port number than the Pods. For many real-world
Services, these values might be the same.
Any Network Policy Ingress rules affecting the target Pods?
If you have deployed any Network Policy Ingress rules which may affect incoming
traffic to hostnames-* Pods, these need to be reviewed.
If this works, you'll need to adjust your app to use a cross-namespace name, or
run your app and Service in the same Namespace. If this still fails, try a
fully-qualified name:
Note the suffix here: "default.svc.cluster.local". The "default" is the
Namespace you're operating in. The "svc" denotes that this is a Service.
The "cluster.local" is your cluster domain, which COULD be different in your
own cluster.
You can also try this from a Node in the cluster:
Note: 10.0.0.10 is the cluster's DNS Service IP, yours might be different.
If you are able to do a fully-qualified name lookup but not a relative one, you
need to check that your /etc/resolv.conf file in your Pod is correct. From
within a Pod:
The nameserver line must indicate your cluster's DNS Service. This is
passed into kubelet with the --cluster-dns flag.
The search line must include an appropriate suffix for you to find the
Service name. In this case it is looking for Services in the local
Namespace ("default.svc.cluster.local"), Services in all Namespaces
("svc.cluster.local"), and lastly for names in the cluster ("cluster.local").
Depending on your own install you might have additional records after that (up
to 6 total). The cluster suffix is passed into kubelet with the
--cluster-domain flag. Throughout this document, the cluster suffix is
assumed to be "cluster.local". Your own clusters might be configured
differently, in which case you should change that in all of the previous
commands.
The options line must set ndots high enough that your DNS client library
considers search paths at all. Kubernetes sets this to 5 by default, which is
high enough to cover all of the DNS names it generates.
Does any Service work by DNS name?
If the above still fails, DNS lookups are not working for your Service. You
can take a step back and see what else is not working. The Kubernetes master
Service should always work. From within a Pod:
If this fails, please see the kube-proxy section
of this document, or even go back to the top of this document and start over,
but instead of debugging your own Service, debug the DNS Service.
Does the Service work by IP?
Assuming you have confirmed that DNS works, the next thing to test is whether your
Service works by its IP address. From a Pod in your cluster, access the
Service's IP (from kubectl get above).
for i in $(seq 1 3); do wget -qO- 10.0.1.175:80
done
If your Service is working, you should get correct responses. If not, there
are a number of things that could be going wrong. Read on.
Is the Service defined correctly?
It might sound silly, but you should really double and triple check that your
Service is correct and matches your Pod's port. Read back your Service
and verify it:
Is the Service port you are trying to access listed in spec.ports[]?
Is the targetPort correct for your Pods (some Pods use a different port than the Service)?
If you meant to use a numeric port, is it a number (9376) or a string "9376"?
If you meant to use a named port, do your Pods expose a port with the same name?
Is the port's protocol correct for your Pods?
Does the Service have any Endpoints?
If you got this far, you have confirmed that your Service is correctly
defined and is resolved by DNS. Now let's check that the Pods you ran are
actually being selected by the Service.
Earlier you saw that the Pods were running. You can re-check that:
kubectl get pods -l app=hostnames
NAME READY STATUS RESTARTS AGE
hostnames-632524106-bbpiw 1/1 Running 0 1h
hostnames-632524106-ly40y 1/1 Running 0 1h
hostnames-632524106-tlaok 1/1 Running 0 1h
The -l app=hostnames argument is a label selector configured on the Service.
The "AGE" column says that these Pods are about an hour old, which implies that
they are running fine and not crashing.
The "RESTARTS" column says that these pods are not crashing frequently or being
restarted. Frequent restarts could lead to intermittent connectivity issues.
If the restart count is high, read more about how to debug pods.
Inside the Kubernetes system is a control loop which evaluates the selector of
every Service and saves the results into a corresponding Endpoints object.
kubectl get endpoints hostnames
NAME ENDPOINTS
hostnames 10.244.0.5:9376,10.244.0.6:9376,10.244.0.7:9376
This confirms that the endpoints controller has found the correct Pods for
your Service. If the ENDPOINTS column is <none>, you should check that
the spec.selector field of your Service actually selects for
metadata.labels values on your Pods. A common mistake is to have a typo or
other error, such as the Service selecting for app=hostnames, but the
Deployment specifying run=hostnames, as in versions previous to 1.18, where
the kubectl run command could have been also used to create a Deployment.
Are the Pods working?
At this point, you know that your Service exists and has selected your Pods.
At the beginning of this walk-through, you verified the Pods themselves.
Let's check again that the Pods are actually working - you can bypass the
Service mechanism and go straight to the Pods, as listed by the Endpoints
above.
Note: These commands use the Pod port (9376), rather than the Service port (80).
From within a Pod:
for ep in 10.244.0.5:9376 10.244.0.6:9376 10.244.0.7:9376; do wget -qO- $epdone
You expect each Pod in the Endpoints list to return its own hostname. If
this is not what happens (or whatever the correct behavior is for your own
Pods), you should investigate what's happening there.
Is the kube-proxy working?
If you get here, your Service is running, has Endpoints, and your Pods
are actually serving. At this point, the whole Service proxy mechanism is
suspect. Let's confirm it, piece by piece.
The default implementation of Services, and the one used on most clusters, is
kube-proxy. This is a program that runs on every node and configures one of a
small set of mechanisms for providing the Service abstraction. If your
cluster does not use kube-proxy, the following sections will not apply, and you
will have to investigate whatever implementation of Services you are using.
Is kube-proxy running?
Confirm that kube-proxy is running on your Nodes. Running directly on a
Node, you should get something like the below:
Next, confirm that it is not failing something obvious, like contacting the
master. To do this, you'll have to look at the logs. Accessing the logs
depends on your Node OS. On some OSes it is a file, such as
/var/log/kube-proxy.log, while other OSes use journalctl to access logs. You
should see something like:
I1027 22:14:53.995134 5063 server.go:200] Running in resource-only container "/kube-proxy"
I1027 22:14:53.998163 5063 server.go:247] Using iptables Proxier.
I1027 22:14:53.999055 5063 server.go:255] Tearing down userspace rules. Errors here are acceptable.
I1027 22:14:54.038140 5063 proxier.go:352] Setting endpoints for "kube-system/kube-dns:dns-tcp" to [10.244.1.3:53]
I1027 22:14:54.038164 5063 proxier.go:352] Setting endpoints for "kube-system/kube-dns:dns" to [10.244.1.3:53]
I1027 22:14:54.038209 5063 proxier.go:352] Setting endpoints for "default/kubernetes:https" to [10.240.0.2:443]
I1027 22:14:54.038238 5063 proxier.go:429] Not syncing iptables until Services and Endpoints have been received from master
I1027 22:14:54.040048 5063 proxier.go:294] Adding new service "default/kubernetes:https" at 10.0.0.1:443/TCP
I1027 22:14:54.040154 5063 proxier.go:294] Adding new service "kube-system/kube-dns:dns" at 10.0.0.10:53/UDP
I1027 22:14:54.040223 5063 proxier.go:294] Adding new service "kube-system/kube-dns:dns-tcp" at 10.0.0.10:53/TCP
If you see error messages about not being able to contact the master, you
should double-check your Node configuration and installation steps.
One of the possible reasons that kube-proxy cannot run correctly is that the
required conntrack binary cannot be found. This may happen on some Linux
systems, depending on how you are installing the cluster, for example, you are
installing Kubernetes from scratch. If this is the case, you need to manually
install the conntrack package (e.g. sudo apt install conntrack on Ubuntu)
and then retry.
Kube-proxy can run in one of a few modes. In the log listed above, the
line Using iptables Proxier indicates that kube-proxy is running in
"iptables" mode. The most common other mode is "ipvs". The older "userspace"
mode has largely been replaced by these.
Iptables mode
In "iptables" mode, you should see something like the following on a Node:
iptables-save | grep hostnames
-A KUBE-SEP-57KPRZ3JQVENLNBR -s 10.244.3.6/32 -m comment --comment "default/hostnames:" -j MARK --set-xmark 0x00004000/0x00004000
-A KUBE-SEP-57KPRZ3JQVENLNBR -p tcp -m comment --comment "default/hostnames:" -m tcp -j DNAT --to-destination 10.244.3.6:9376
-A KUBE-SEP-WNBA2IHDGP2BOBGZ -s 10.244.1.7/32 -m comment --comment "default/hostnames:" -j MARK --set-xmark 0x00004000/0x00004000
-A KUBE-SEP-WNBA2IHDGP2BOBGZ -p tcp -m comment --comment "default/hostnames:" -m tcp -j DNAT --to-destination 10.244.1.7:9376
-A KUBE-SEP-X3P2623AGDH6CDF3 -s 10.244.2.3/32 -m comment --comment "default/hostnames:" -j MARK --set-xmark 0x00004000/0x00004000
-A KUBE-SEP-X3P2623AGDH6CDF3 -p tcp -m comment --comment "default/hostnames:" -m tcp -j DNAT --to-destination 10.244.2.3:9376
-A KUBE-SERVICES -d 10.0.1.175/32 -p tcp -m comment --comment "default/hostnames: cluster IP" -m tcp --dport 80 -j KUBE-SVC-NWV5X2332I4OT4T3
-A KUBE-SVC-NWV5X2332I4OT4T3 -m comment --comment "default/hostnames:" -m statistic --mode random --probability 0.33332999982 -j KUBE-SEP-WNBA2IHDGP2BOBGZ
-A KUBE-SVC-NWV5X2332I4OT4T3 -m comment --comment "default/hostnames:" -m statistic --mode random --probability 0.50000000000 -j KUBE-SEP-X3P2623AGDH6CDF3
-A KUBE-SVC-NWV5X2332I4OT4T3 -m comment --comment "default/hostnames:" -j KUBE-SEP-57KPRZ3JQVENLNBR
For each port of each Service, there should be 1 rule in KUBE-SERVICES and
one KUBE-SVC-<hash> chain. For each Pod endpoint, there should be a small
number of rules in that KUBE-SVC-<hash> and one KUBE-SEP-<hash> chain with
a small number of rules in it. The exact rules will vary based on your exact
config (including node-ports and load-balancers).
IPVS mode
In "ipvs" mode, you should see something like the following on a Node:
For each port of each Service, plus any NodePorts, external IPs, and
load-balancer IPs, kube-proxy will create a virtual server. For each Pod
endpoint, it will create corresponding real servers. In this example, service
hostnames(10.0.1.175:80) has 3 endpoints(10.244.0.5:9376,
10.244.0.6:9376, 10.244.0.7:9376).
Userspace mode
In rare cases, you may be using "userspace" mode. From your Node:
There should be 2 rules for each port of your Service (only one in this
example) - a "KUBE-PORTALS-CONTAINER" and a "KUBE-PORTALS-HOST".
Almost nobody should be using the "userspace" mode any more, so you won't spend
more time on it here.
Is kube-proxy proxying?
Assuming you do see one the above cases, try again to access your Service by
IP from one of your Nodes:
curl 10.0.1.175:80
hostnames-632524106-bbpiw
If this fails and you are using the userspace proxy, you can try accessing the
proxy directly. If you are using the iptables proxy, skip this section.
Look back at the iptables-save output above, and extract the
port number that kube-proxy is using for your Service. In the above
examples it is "48577". Now connect to that:
curl localhost:48577
hostnames-632524106-tlaok
If this still fails, look at the kube-proxy logs for specific lines like:
Setting endpoints for default/hostnames:default to [10.244.0.5:9376 10.244.0.6:9376 10.244.0.7:9376]
If you don't see those, try restarting kube-proxy with the -v flag set to 4, and
then look at the logs again.
Edge case: A Pod fails to reach itself via the Service IP
This might sound unlikely, but it does happen and it is supposed to work.
This can happen when the network is not properly configured for "hairpin"
traffic, usually when kube-proxy is running in iptables mode and Pods
are connected with bridge network. The Kubelet exposes a hairpin-modeflag that allows endpoints of a Service to loadbalance
back to themselves if they try to access their own Service VIP. The
hairpin-mode flag must either be set to hairpin-veth or
promiscuous-bridge.
The common steps to trouble shoot this are as follows:
Confirm hairpin-mode is set to hairpin-veth or promiscuous-bridge.
You should see something like the below. hairpin-mode is set to
promiscuous-bridge in the following example.
Confirm the effective hairpin-mode. To do this, you'll have to look at
kubelet log. Accessing the logs depends on your Node OS. On some OSes it
is a file, such as /var/log/kubelet.log, while other OSes use journalctl
to access logs. Please be noted that the effective hairpin mode may not
match --hairpin-mode flag due to compatibility. Check if there is any log
lines with key word hairpin in kubelet.log. There should be log lines
indicating the effective hairpin mode, like something below.
I0629 00:51:43.648698 3252 kubelet.go:380] Hairpin mode set to "promiscuous-bridge"
If the effective hairpin mode is hairpin-veth, ensure the Kubelet has
the permission to operate in /sys on node. If everything works properly,
you should see something like:
for intf in /sys/devices/virtual/net/cbr0/brif/*; do cat $intf/hairpin_mode; done
1
1
1
1
If the effective hairpin mode is promiscuous-bridge, ensure Kubelet
has the permission to manipulate linux bridge on node. If cbr0 bridge is
used and configured properly, you should see:
ifconfig cbr0 |grep PROMISC
UP BROADCAST RUNNING PROMISC MULTICAST MTU:1460 Metric:1
Seek help if none of above works out.
Seek help
If you get this far, something very strange is happening. Your Service is
running, has Endpoints, and your Pods are actually serving. You have DNS
working, and kube-proxy does not seem to be misbehaving. And yet your
Service is not working. Please let us know what is going on, so we can help
investigate!
You need to have a Kubernetes cluster, and the kubectl command-line tool must be configured to communicate with your cluster.
You should have a StatefulSet running that you want to investigate.
Debugging a StatefulSet
In order to list all the pods which belong to a StatefulSet, which have a label app.kubernetes.io/name=MyApp set on them,
you can use the following:
kubectl get pods -l app.kubernetes.io/name=MyApp
If you find that any Pods listed are in Unknown or Terminating state for an extended period of time,
refer to the Deleting StatefulSet Pods task for
instructions on how to deal with them.
You can debug individual Pods in a StatefulSet using the
Debugging Pods guide.
This page shows how to investigate problems related to the execution of
Init Containers. The example command lines below refer to the Pod as
<pod-name> and the Init Containers as <init-container-1> and
<init-container-2>.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You can also access the Init Container statuses programmatically by reading the
status.initContainerStatuses field on the Pod Spec:
kubectl get pod nginx --template '{{.status.initContainerStatuses}}'
This command will return the same information as above in raw JSON.
Accessing logs from Init Containers
Pass the Init Container name along with the Pod name
to access its logs.
kubectl logs <pod-name> -c <init-container-2>
Init Containers that run a shell script print
commands as they're executed. For example, you can do this in Bash by running
set -x at the beginning of the script.
Understanding Pod status
A Pod status beginning with Init: summarizes the status of Init Container
execution. The table below describes some example status values that you might
see while debugging Init Containers.
Status
Meaning
Init:N/M
The Pod has M Init Containers, and N have completed so far.
Init:Error
An Init Container has failed to execute.
Init:CrashLoopBackOff
An Init Container has failed repeatedly.
Pending
The Pod has not yet begun executing Init Containers.
PodInitializing or Running
The Pod has already finished executing Init Containers.
4.4.1.5 - Debug Running Pods
This page explains how to debug Pods running (or crashing) on a Node.
Before you begin
Your Pod should already be
scheduled and running. If your Pod is not yet running, start with Debugging
Pods.
For some of the advanced debugging steps you need to know on which Node the
Pod is running and have shell access to run commands on that Node. You don't
need that access to run the standard debug steps that use kubectl.
Using kubectl describe pod to fetch details about pods
For this example we'll use a Deployment to create two pods, similar to the earlier example.
NAME READY STATUS RESTARTS AGE
nginx-deployment-67d4bdd6f5-cx2nz 1/1 Running 0 13s
nginx-deployment-67d4bdd6f5-w6kd7 1/1 Running 0 13s
We can retrieve a lot more information about each of these pods using kubectl describe pod. For example:
kubectl describe pod nginx-deployment-67d4bdd6f5-w6kd7
Name: nginx-deployment-67d4bdd6f5-w6kd7
Namespace: default
Priority: 0
Node: kube-worker-1/192.168.0.113
Start Time: Thu, 17 Feb 2022 16:51:01 -0500
Labels: app=nginx
pod-template-hash=67d4bdd6f5
Annotations: <none>
Status: Running
IP: 10.88.0.3
IPs:
IP: 10.88.0.3
IP: 2001:db8::1
Controlled By: ReplicaSet/nginx-deployment-67d4bdd6f5
Containers:
nginx:
Container ID: containerd://5403af59a2b46ee5a23fb0ae4b1e077f7ca5c5fb7af16e1ab21c00e0e616462a
Image: nginx
Image ID: docker.io/library/nginx@sha256:2834dc507516af02784808c5f48b7cbe38b8ed5d0f4837f16e78d00deb7e7767
Port: 80/TCP
Host Port: 0/TCP
State: Running
Started: Thu, 17 Feb 2022 16:51:05 -0500
Ready: True
Restart Count: 0
Limits:
cpu: 500m
memory: 128Mi
Requests:
cpu: 500m
memory: 128Mi
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-bgsgp (ro)
Conditions:
Type Status
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes:
kube-api-access-bgsgp:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional: <nil>
DownwardAPI: true
QoS Class: Guaranteed
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 34s default-scheduler Successfully assigned default/nginx-deployment-67d4bdd6f5-w6kd7 to kube-worker-1
Normal Pulling 31s kubelet Pulling image "nginx"
Normal Pulled 30s kubelet Successfully pulled image "nginx" in 1.146417389s
Normal Created 30s kubelet Created container nginx
Normal Started 30s kubelet Started container nginx
Here you can see configuration information about the container(s) and Pod (labels, resource requirements, etc.), as well as status information about the container(s) and Pod (state, readiness, restart count, events, etc.).
The container state is one of Waiting, Running, or Terminated. Depending on the state, additional information will be provided -- here you can see that for a container in Running state, the system tells you when the container started.
Ready tells you whether the container passed its last readiness probe. (In this case, the container does not have a readiness probe configured; the container is assumed to be ready if no readiness probe is configured.)
Restart Count tells you how many times the container has been restarted; this information can be useful for detecting crash loops in containers that are configured with a restart policy of 'always.'
Currently the only Condition associated with a Pod is the binary Ready condition, which indicates that the pod is able to service requests and should be added to the load balancing pools of all matching services.
Lastly, you see a log of recent events related to your Pod. The system compresses multiple identical events by indicating the first and last time it was seen and the number of times it was seen. "From" indicates the component that is logging the event, "SubobjectPath" tells you which object (e.g. container within the pod) is being referred to, and "Reason" and "Message" tell you what happened.
Example: debugging Pending Pods
A common scenario that you can detect using events is when you've created a Pod that won't fit on any node. For example, the Pod might request more resources than are free on any node, or it might specify a label selector that doesn't match any nodes. Let's say we created the previous Deployment with 5 replicas (instead of 2) and requesting 600 millicores instead of 500, on a four-node cluster where each (virtual) machine has 1 CPU. In that case one of the Pods will not be able to schedule. (Note that because of the cluster addon pods such as fluentd, skydns, etc., that run on each node, if we requested 1000 millicores then none of the Pods would be able to schedule.)
kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-deployment-1006230814-6winp 1/1 Running 0 7m
nginx-deployment-1006230814-fmgu3 1/1 Running 0 7m
nginx-deployment-1370807587-6ekbw 1/1 Running 0 1m
nginx-deployment-1370807587-fg172 0/1 Pending 0 1m
nginx-deployment-1370807587-fz9sd 0/1 Pending 0 1m
To find out why the nginx-deployment-1370807587-fz9sd pod is not running, we can use kubectl describe pod on the pending Pod and look at its events:
kubectl describe pod nginx-deployment-1370807587-fz9sd
Name: nginx-deployment-1370807587-fz9sd
Namespace: default
Node: /
Labels: app=nginx,pod-template-hash=1370807587
Status: Pending
IP:
Controllers: ReplicaSet/nginx-deployment-1370807587
Containers:
nginx:
Image: nginx
Port: 80/TCP
QoS Tier:
memory: Guaranteed
cpu: Guaranteed
Limits:
cpu: 1
memory: 128Mi
Requests:
cpu: 1
memory: 128Mi
Environment Variables:
Volumes:
default-token-4bcbi:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-4bcbi
Events:
FirstSeen LastSeen Count From SubobjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
1m 48s 7 {default-scheduler } Warning FailedScheduling pod (nginx-deployment-1370807587-fz9sd) failed to fit in any node
fit failure on node (kubernetes-node-6ta5): Node didn't have enough resource: CPU, requested: 1000, used: 1420, capacity: 2000
fit failure on node (kubernetes-node-wul5): Node didn't have enough resource: CPU, requested: 1000, used: 1100, capacity: 2000
Here you can see the event generated by the scheduler saying that the Pod failed to schedule for reason FailedScheduling (and possibly others). The message tells us that there were not enough resources for the Pod on any of the nodes.
To correct this situation, you can use kubectl scale to update your Deployment to specify four or fewer replicas. (Or you could leave the one Pod pending, which is harmless.)
Events such as the ones you saw at the end of kubectl describe pod are persisted in etcd and provide high-level information on what is happening in the cluster. To list all events you can use
kubectl get events
but you have to remember that events are namespaced. This means that if you're interested in events for some namespaced object (e.g. what happened with Pods in namespace my-namespace) you need to explicitly provide a namespace to the command:
kubectl get events --namespace=my-namespace
To see events from all namespaces, you can use the --all-namespaces argument.
In addition to kubectl describe pod, another way to get extra information about a pod (beyond what is provided by kubectl get pod) is to pass the -o yaml output format flag to kubectl get pod. This will give you, in YAML format, even more information than kubectl describe pod--essentially all of the information the system has about the Pod. Here you will see things like annotations (which are key-value metadata without the label restrictions, that is used internally by Kubernetes system components), restart policy, ports, and volumes.
kubectl get pod nginx-deployment-1006230814-6winp -o yaml
If the container image includes
debugging utilities, as is the case with images built from Linux and Windows OS
base images, you can run commands inside a specific container with
kubectl exec:
Ephemeral containers
are useful for interactive troubleshooting when kubectl exec is insufficient
because a container has crashed or a container image doesn't include debugging
utilities, such as with distroless images.
Example debugging using ephemeral containers
You can use the kubectl debug command to add ephemeral containers to a
running Pod. First, create a pod for the example:
kubectl run ephemeral-demo --image=registry.k8s.io/pause:3.1 --restart=Never
The examples in this section use the pause container image because it does not
contain debugging utilities, but this method works with all container
images.
If you attempt to use kubectl exec to create a shell you will see an error
because there is no shell in this container image.
kubectl exec -it ephemeral-demo -- sh
OCI runtime exec failed: exec failed: container_linux.go:346: starting container process caused "exec: \"sh\": executable file not found in $PATH": unknown
You can instead add a debugging container using kubectl debug. If you
specify the -i/--interactive argument, kubectl will automatically attach
to the console of the Ephemeral Container.
Defaulting debug container name to debugger-8xzrl.
If you don't see a command prompt, try pressing enter.
/ #
This command adds a new busybox container and attaches to it. The --target
parameter targets the process namespace of another container. It's necessary
here because kubectl run does not enable process namespace sharing in the pod it
creates.
Note: The --target parameter must be supported by the Container Runtime. When not supported,
the Ephemeral Container may not be started, or it may be started with an
isolated process namespace so that ps does not reveal processes in other
containers.
You can view the state of the newly created ephemeral container using kubectl describe:
Use kubectl delete to remove the Pod when you're finished:
kubectl delete pod ephemeral-demo
Debugging using a copy of the Pod
Sometimes Pod configuration options make it difficult to troubleshoot in certain
situations. For example, you can't run kubectl exec to troubleshoot your
container if your container image does not include a shell or if your application
crashes on startup. In these situations you can use kubectl debug to create a
copy of the Pod with configuration values changed to aid debugging.
Copying a Pod while adding a new container
Adding a new container can be useful when your application is running but not
behaving as you expect and you'd like to add additional troubleshooting
utilities to the Pod.
For example, maybe your application's container images are built on busybox
but you need debugging utilities not included in busybox. You can simulate
this scenario using kubectl run:
kubectl run myapp --image=busybox:1.28 --restart=Never -- sleep 1d
Run this command to create a copy of myapp named myapp-debug that adds a
new Ubuntu container for debugging:
Defaulting debug container name to debugger-w7xmf.
If you don't see a command prompt, try pressing enter.
root@myapp-debug:/#
Note:
kubectl debug automatically generates a container name if you don't choose
one using the --container flag.
The -i flag causes kubectl debug to attach to the new container by
default. You can prevent this by specifying --attach=false. If your session
becomes disconnected you can reattach using kubectl attach.
The --share-processes allows the containers in this Pod to see processes
from the other containers in the Pod. For more information about how this
works, see Share Process Namespace between Containers in a Pod.
Don't forget to clean up the debugging Pod when you're finished with it:
kubectl delete pod myapp myapp-debug
Copying a Pod while changing its command
Sometimes it's useful to change the command for a container, for example to
add a debugging flag or because the application is crashing.
To simulate a crashing application, use kubectl run to create a container
that immediately exits:
kubectl run --image=busybox:1.28 myapp -- false
You can see using kubectl describe pod myapp that this container is crashing:
You can use kubectl debug to create a copy of this Pod with the command
changed to an interactive shell:
kubectl debug myapp -it --copy-to=myapp-debug --container=myapp -- sh
If you don't see a command prompt, try pressing enter.
/ #
Now you have an interactive shell that you can use to perform tasks like
checking filesystem paths or running the container command manually.
Note:
To change the command of a specific container you must
specify its name using --container or kubectl debug will instead
create a new container to run the command you specified.
The -i flag causes kubectl debug to attach to the container by default.
You can prevent this by specifying --attach=false. If your session becomes
disconnected you can reattach using kubectl attach.
Don't forget to clean up the debugging Pod when you're finished with it:
kubectl delete pod myapp myapp-debug
Copying a Pod while changing container images
In some situations you may want to change a misbehaving Pod from its normal
production container images to an image containing a debugging build or
additional utilities.
As an example, create a Pod using kubectl run:
kubectl run myapp --image=busybox:1.28 --restart=Never -- sleep 1d
Now use kubectl debug to make a copy and change its container image
to ubuntu:
The syntax of --set-image uses the same container_name=image syntax as
kubectl set image. *=ubuntu means change the image of all containers
to ubuntu.
Don't forget to clean up the debugging Pod when you're finished with it:
kubectl delete pod myapp myapp-debug
Debugging via a shell on the node
If none of these approaches work, you can find the Node on which the Pod is
running and create a privileged Pod running in the host namespaces. To create
an interactive shell on a node using kubectl debug, run:
kubectl debug node/mynode -it --image=ubuntu
Creating debugging pod node-debugger-mynode-pdx84 with container debugger on node mynode.
If you don't see a command prompt, try pressing enter.
root@ek8s:/#
When creating a debugging session on a node, keep in mind that:
kubectl debug automatically generates the name of the new Pod based on
the name of the Node.
The container runs in the host IPC, Network, and PID namespaces.
The root filesystem of the Node will be mounted at /host.
Don't forget to clean up the debugging Pod when you're finished with it:
kubectl delete pod node-debugger-mynode-pdx84
4.4.1.6 - Determine the Reason for Pod Failure
This page shows how to write and read a Container
termination message.
Termination messages provide a way for containers to write
information about fatal events to a location where it can
be easily retrieved and surfaced by tools like dashboards
and monitoring software. In most cases, information that you
put in a termination message should also be written to
the general
Kubernetes logs.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In the YAML file, in the command and args fields, you can see that the
container sleeps for 10 seconds and then writes "Sleep expired" to
the /dev/termination-log file. After the container writes
the "Sleep expired" message, it terminates.
Display information about the Pod:
kubectl get pod termination-demo
Repeat the preceding command until the Pod is no longer running.
Use a Go template to filter the output so that it includes
only the termination message:
kubectl get pod termination-demo -o go-template="{{range .status.containerStatuses}}{{.lastState.terminated.message}}{{end}}"
If you are running a multi-container pod, you can use a Go template to include the container's name. By doing so, you can discover which of the containers is failing:
kubectl get pod multi-container-pod -o go-template='{{range .status.containerStatuses}}{{printf "%s:\n%s\n\n" .name .lastState.terminated.message}}{{end}}'
Customizing the termination message
Kubernetes retrieves termination messages from the termination message file
specified in the terminationMessagePath field of a Container, which has a default
value of /dev/termination-log. By customizing this field, you can tell Kubernetes
to use a different file. Kubernetes use the contents from the specified file to
populate the Container's status message on both success and failure.
The termination message is intended to be brief final status, such as an assertion failure message.
The kubelet truncates messages that are longer than 4096 bytes.
The total message length across all containers is limited to 12KiB, divided equally among each container.
For example, if there are 12 containers (initContainers or containers), each has 1024 bytes of available termination message space.
The default termination message path is /dev/termination-log.
You cannot set the termination message path after a Pod is launched
In the following example, the container writes termination messages to
/tmp/my-log for Kubernetes to retrieve:
Moreover, users can set the terminationMessagePolicy field of a Container for
further customization. This field defaults to "File" which means the termination
messages are retrieved only from the termination message file. By setting the
terminationMessagePolicy to "FallbackToLogsOnError", you can tell Kubernetes
to use the last chunk of container log output if the termination message file
is empty and the container exited with an error. The log output is limited to
2048 bytes or 80 lines, whichever is smaller.
What's next
See the terminationMessagePath field in
Container.
This page shows how to use kubectl exec to get a shell to a
running container.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Note: The double dash (--) separates the arguments you want to pass to the command from the kubectl arguments.
In your shell, list the root directory:
# Run this inside the containerls /
In your shell, experiment with other commands. Here are
some examples:
# You can run these example commands inside the containerls /
cat /proc/mounts
cat /proc/1/maps
apt-get update
apt-get install -y tcpdump
tcpdump
apt-get install -y lsof
lsof
apt-get install -y procps
ps aux
ps aux | grep nginx
Writing the root page for nginx
Look again at the configuration file for your Pod. The Pod
has an emptyDir volume, and the container mounts the volume
at /usr/share/nginx/html.
In your shell, create an index.html file in the /usr/share/nginx/html
directory:
# Run this inside the containerecho'Hello shell demo' > /usr/share/nginx/html/index.html
In your shell, send a GET request to the nginx server:
# Run this in the shell inside your containerapt-get update
apt-get install curl
curl http://localhost/
The output shows the text that you wrote to the index.html file:
Hello shell demo
When you are finished with your shell, enter exit.
exit# To quit the shell in the container
Running individual commands in a container
In an ordinary command window, not your shell, list the environment
variables in the running container:
kubectl exec shell-demo env
Experiment with running other commands. Here are some examples:
kubectl exec shell-demo -- ps aux
kubectl exec shell-demo -- ls /
kubectl exec shell-demo -- cat /proc/1/mounts
Opening a shell when a Pod has more than one container
If a Pod has more than one container, use --container or -c to
specify a container in the kubectl exec command. For example,
suppose you have a Pod named my-pod, and the Pod has two containers
named main-app and helper-app. The following command would open a
shell to the main-app container.
This doc is about cluster troubleshooting; we assume you have already ruled out your application as the root cause of the
problem you are experiencing. See
the application troubleshooting guide for tips on application debugging.
You may also visit the troubleshooting overview document for more information.
Listing your cluster
The first thing to debug in your cluster is if your nodes are all registered correctly.
Run the following command:
kubectl get nodes
And verify that all of the nodes you expect to see are present and that they are all in the Ready state.
To get detailed information about the overall health of your cluster, you can run:
kubectl cluster-info dump
Example: debugging a down/unreachable node
Sometimes when debugging it can be useful to look at the status of a node -- for example, because
you've noticed strange behavior of a Pod that's running on the node, or to find out why a Pod
won't schedule onto the node. As with Pods, you can use kubectl describe node and kubectl get node -o yaml to retrieve detailed information about nodes. For example, here's what you'll see if
a node is down (disconnected from the network, or kubelet dies and won't restart, etc.). Notice
the events that show the node is NotReady, and also notice that the pods are no longer running
(they are evicted after five minutes of NotReady status).
kubectl get nodes
NAME STATUS ROLES AGE VERSION
kube-worker-1 NotReady <none> 1h v1.23.3
kubernetes-node-bols Ready <none> 1h v1.23.3
kubernetes-node-st6x Ready <none> 1h v1.23.3
kubernetes-node-unaj Ready <none> 1h v1.23.3
kubectl describe node kube-worker-1
Name: kube-worker-1
Roles: <none>
Labels: beta.kubernetes.io/arch=amd64
beta.kubernetes.io/os=linux
kubernetes.io/arch=amd64
kubernetes.io/hostname=kube-worker-1
kubernetes.io/os=linux
Annotations: kubeadm.alpha.kubernetes.io/cri-socket: /run/containerd/containerd.sock
node.alpha.kubernetes.io/ttl: 0
volumes.kubernetes.io/controller-managed-attach-detach: true
CreationTimestamp: Thu, 17 Feb 2022 16:46:30 -0500
Taints: node.kubernetes.io/unreachable:NoExecute
node.kubernetes.io/unreachable:NoSchedule
Unschedulable: false
Lease:
HolderIdentity: kube-worker-1
AcquireTime: <unset>
RenewTime: Thu, 17 Feb 2022 17:13:09 -0500
Conditions:
Type Status LastHeartbeatTime LastTransitionTime Reason Message
---- ------ ----------------- ------------------ ------ -------
NetworkUnavailable False Thu, 17 Feb 2022 17:09:13 -0500 Thu, 17 Feb 2022 17:09:13 -0500 WeaveIsUp Weave pod has set this
MemoryPressure Unknown Thu, 17 Feb 2022 17:12:40 -0500 Thu, 17 Feb 2022 17:13:52 -0500 NodeStatusUnknown Kubelet stopped posting node status.
DiskPressure Unknown Thu, 17 Feb 2022 17:12:40 -0500 Thu, 17 Feb 2022 17:13:52 -0500 NodeStatusUnknown Kubelet stopped posting node status.
PIDPressure Unknown Thu, 17 Feb 2022 17:12:40 -0500 Thu, 17 Feb 2022 17:13:52 -0500 NodeStatusUnknown Kubelet stopped posting node status.
Ready Unknown Thu, 17 Feb 2022 17:12:40 -0500 Thu, 17 Feb 2022 17:13:52 -0500 NodeStatusUnknown Kubelet stopped posting node status.
Addresses:
InternalIP: 192.168.0.113
Hostname: kube-worker-1
Capacity:
cpu: 2
ephemeral-storage: 15372232Ki
hugepages-2Mi: 0
memory: 2025188Ki
pods: 110
Allocatable:
cpu: 2
ephemeral-storage: 14167048988
hugepages-2Mi: 0
memory: 1922788Ki
pods: 110
System Info:
Machine ID: 9384e2927f544209b5d7b67474bbf92b
System UUID: aa829ca9-73d7-064d-9019-df07404ad448
Boot ID: 5a295a03-aaca-4340-af20-1327fa5dab5c
Kernel Version: 5.13.0-28-generic
OS Image: Ubuntu 21.10
Operating System: linux
Architecture: amd64
Container Runtime Version: containerd://1.5.9
Kubelet Version: v1.23.3
Kube-Proxy Version: v1.23.3
Non-terminated Pods: (4 in total)
Namespace Name CPU Requests CPU Limits Memory Requests Memory Limits Age
--------- ---- ------------ ---------- --------------- ------------- ---
default nginx-deployment-67d4bdd6f5-cx2nz 500m (25%) 500m (25%) 128Mi (6%) 128Mi (6%) 23m
default nginx-deployment-67d4bdd6f5-w6kd7 500m (25%) 500m (25%) 128Mi (6%) 128Mi (6%) 23m
kube-system kube-proxy-dnxbz 0 (0%) 0 (0%) 0 (0%) 0 (0%) 28m
kube-system weave-net-gjxxp 100m (5%) 0 (0%) 200Mi (10%) 0 (0%) 28m
Allocated resources:
(Total limits may be over 100 percent, i.e., overcommitted.)
Resource Requests Limits
-------- -------- ------
cpu 1100m (55%) 1 (50%)
memory 456Mi (24%) 256Mi (13%)
ephemeral-storage 0 (0%) 0 (0%)
hugepages-2Mi 0 (0%) 0 (0%)
Events:
...
kubectl get node kube-worker-1 -o yaml
apiVersion:v1kind:Nodemetadata:annotations:kubeadm.alpha.kubernetes.io/cri-socket:/run/containerd/containerd.socknode.alpha.kubernetes.io/ttl:"0"volumes.kubernetes.io/controller-managed-attach-detach:"true"creationTimestamp:"2022-02-17T21:46:30Z"labels:beta.kubernetes.io/arch:amd64beta.kubernetes.io/os:linuxkubernetes.io/arch:amd64kubernetes.io/hostname:kube-worker-1kubernetes.io/os:linuxname:kube-worker-1resourceVersion:"4026"uid:98efe7cb-2978-4a0b-842a-1a7bf12c05f8spec:{}status:addresses:- address:192.168.0.113type:InternalIP- address:kube-worker-1type:Hostnameallocatable:cpu:"2"ephemeral-storage:"14167048988"hugepages-2Mi:"0"memory:1922788Kipods:"110"capacity:cpu:"2"ephemeral-storage:15372232Kihugepages-2Mi:"0"memory:2025188Kipods:"110"conditions:- lastHeartbeatTime:"2022-02-17T22:20:32Z"lastTransitionTime:"2022-02-17T22:20:32Z"message:Weave pod has set thisreason:WeaveIsUpstatus:"False"type:NetworkUnavailable- lastHeartbeatTime:"2022-02-17T22:20:15Z"lastTransitionTime:"2022-02-17T22:13:25Z"message:kubelet has sufficient memory availablereason:KubeletHasSufficientMemorystatus:"False"type:MemoryPressure- lastHeartbeatTime:"2022-02-17T22:20:15Z"lastTransitionTime:"2022-02-17T22:13:25Z"message:kubelet has no disk pressurereason:KubeletHasNoDiskPressurestatus:"False"type:DiskPressure- lastHeartbeatTime:"2022-02-17T22:20:15Z"lastTransitionTime:"2022-02-17T22:13:25Z"message:kubelet has sufficient PID availablereason:KubeletHasSufficientPIDstatus:"False"type:PIDPressure- lastHeartbeatTime:"2022-02-17T22:20:15Z"lastTransitionTime:"2022-02-17T22:15:15Z"message:kubelet is posting ready status. AppArmor enabledreason:KubeletReadystatus:"True"type:ReadydaemonEndpoints:kubeletEndpoint:Port:10250nodeInfo:architecture:amd64bootID:22333234-7a6b-44d4-9ce1-67e31dc7e369containerRuntimeVersion:containerd://1.5.9kernelVersion:5.13.0-28-generickubeProxyVersion:v1.23.3kubeletVersion:v1.23.3machineID:9384e2927f544209b5d7b67474bbf92boperatingSystem:linuxosImage:Ubuntu 21.10systemUUID:aa829ca9-73d7-064d-9019-df07404ad448
Looking at logs
For now, digging deeper into the cluster requires logging into the relevant machines. Here are the locations
of the relevant log files. On systemd-based systems, you may need to use journalctl instead of examining log files.
Control Plane nodes
/var/log/kube-apiserver.log - API Server, responsible for serving the API
/var/log/kube-scheduler.log - Scheduler, responsible for making scheduling decisions
/var/log/kube-controller-manager.log - a component that runs most Kubernetes built-in
controllers, with the notable exception of scheduling
(the kube-scheduler handles scheduling).
Worker Nodes
/var/log/kubelet.log - logs from the kubelet, responsible for running containers on the node
/var/log/kube-proxy.log - logs from kube-proxy, which is responsible for directing traffic to Service endpoints
Cluster failure modes
This is an incomplete list of things that could go wrong, and how to adjust your cluster setup to mitigate the problems.
Contributing causes
VM(s) shutdown
Network partition within cluster, or between cluster and users
Crashes in Kubernetes software
Data loss or unavailability of persistent storage (e.g. GCE PD or AWS EBS volume)
Operator error, for example misconfigured Kubernetes software or application software
Specific scenarios
API server VM shutdown or apiserver crashing
Results
unable to stop, update, or start new pods, services, replication controller
existing pods and services should continue to work normally, unless they depend on the Kubernetes API
API server backing storage lost
Results
the kube-apiserver component fails to start successfully and become healthy
kubelets will not be able to reach it but will continue to run the same pods and provide the same service proxying
manual recovery or recreation of apiserver state necessary before apiserver is restarted
Supporting services (node controller, replication controller manager, scheduler, etc) VM shutdown or crashes
currently those are colocated with the apiserver, and their unavailability has similar consequences as apiserver
in future, these will be replicated as well and may not be co-located
they do not have their own persistent state
Individual node (VM or physical machine) shuts down
Results
pods on that Node stop running
Network partition
Results
partition A thinks the nodes in partition B are down; partition B thinks the apiserver is down.
(Assuming the master VM ends up in partition A.)
Kubelet software fault
Results
crashing kubelet cannot start new pods on the node
kubelet might delete the pods or not
node marked unhealthy
replication controllers start new pods elsewhere
Cluster operator error
Results
loss of pods, services, etc
lost of apiserver backing store
users unable to read API
etc.
Mitigations
Action: Use IaaS provider's automatic VM restarting feature for IaaS VMs
Mitigates: Apiserver VM shutdown or apiserver crashing
Mitigates: Supporting services VM shutdown or crashes
Action: Use IaaS providers reliable storage (e.g. GCE PD or AWS EBS volume) for VMs with apiserver+etcd
For Kubernetes, the Metrics API offers a basic set of metrics to support automatic scaling and
similar use cases. This API makes information available about resource usage for node and pod,
including metrics for CPU and memory. If you deploy the Metrics API into your cluster, clients of
the Kubernetes API can then query for this information, and you can use Kubernetes' access control
mechanisms to manage permissions to do so.
You can also view the resource metrics using the
kubectl top
command.
Note: The Metrics API, and the metrics pipeline that it enables, only offers the minimum
CPU and memory metrics to enable automatic scaling using HPA and / or VPA.
If you would like to provide a more complete set of metrics, you can complement
the simpler Metrics API by deploying a second
metrics pipeline
that uses the Custom Metrics API.
Figure 1 illustrates the architecture of the resource metrics pipeline.
flowchart RL
subgraph cluster[Cluster]
direction RL
S[
]
A[Metrics- Server]
subgraph B[Nodes]
direction TB
D[cAdvisor] --> C[kubelet]
E[Container runtime] --> D
E1[Container runtime] --> D
P[pod data] -.- C
end
L[API server]
W[HPA]
C ---->|Summary API| A -->|metrics API| L --> W
end
L ---> K[kubectl top]
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class W,B,P,K,cluster,D,E,E1 box
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class S spacewhite
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:#fff;
class A,L,C k8s
Figure 1. Resource Metrics Pipeline
The architecture components, from right to left in the figure, consist of the following:
cAdvisor: Daemon for collecting, aggregating and exposing
container metrics included in Kubelet.
kubelet: Node agent for managing container
resources. Resource metrics are accessible using the /metrics/resource and /stats kubelet
API endpoints.
Summary API: API provided by the kubelet for discovering and retrieving
per-node summarized stats available through the /stats endpoint.
metrics-server: Cluster addon component that collects and aggregates resource
metrics pulled from each kubelet. The API server serves Metrics API for use by HPA, VPA, and by
the kubectl top command. Metrics Server is a reference implementation of the Metrics API.
Metrics API: Kubernetes API supporting access to CPU and memory used for
workload autoscaling. To make this work in your cluster, you need an API extension server that
provides the Metrics API.
Note: cAdvisor supports reading metrics from cgroups, which works with typical container runtimes on Linux.
If you use a container runtime that uses another resource isolation mechanism, for example
virtualization, then that container runtime must support
CRI Container Metrics
in order for metrics to be available to the kubelet.
Metrics API
FEATURE STATE:Kubernetes 1.8 [beta]
The metrics-server implements the Metrics API. This API allows you to access CPU and memory usage
for the nodes and pods in your cluster. Its primary role is to feed resource usage metrics to K8s
autoscaler components.
Here is an example of the Metrics API request for a minikube node piped through jq for easier
reading:
kubectl get --raw "/apis/metrics.k8s.io/v1beta1/nodes/minikube" | jq '.'
Here is an example of the Metrics API request for a kube-scheduler-minikube pod contained in the
kube-system namespace and piped through jq for easier reading:
kubectl get --raw "/apis/metrics.k8s.io/v1beta1/namespaces/kube-system/pods/kube-scheduler-minikube" | jq '.'
Note: You must deploy the metrics-server or alternative adapter that serves the Metrics API to be able
to access it.
Measuring resource usage
CPU
CPU is reported as the average core usage measured in cpu units. One cpu, in Kubernetes, is
equivalent to 1 vCPU/Core for cloud providers, and 1 hyper-thread on bare-metal Intel processors.
This value is derived by taking a rate over a cumulative CPU counter provided by the kernel (in
both Linux and Windows kernels). The time window used to calculate CPU is shown under window field
in Metrics API.
To learn more about how Kubernetes allocates and measures CPU resources, see
meaning of CPU.
Memory
Memory is reported as the working set, measured in bytes, at the instant the metric was collected.
In an ideal world, the "working set" is the amount of memory in-use that cannot be freed under
memory pressure. However, calculation of the working set varies by host OS, and generally makes
heavy use of heuristics to produce an estimate.
The Kubernetes model for a container's working set expects that the container runtime counts
anonymous memory associated with the container in question. The working set metric typically also
includes some cached (file-backed) memory, because the host OS cannot always reclaim pages.
To learn more about how Kubernetes allocates and measures memory resources, see
meaning of memory.
Metrics Server
The metrics-server fetches resource metrics from the kubelets and exposes them in the Kubernetes
API server through the Metrics API for use by the HPA and VPA. You can also view these metrics
using the kubectl top command.
The metrics-server uses the Kubernetes API to track nodes and pods in your cluster. The
metrics-server queries each node over HTTP to fetch metrics. The metrics-server also builds an
internal view of pod metadata, and keeps a cache of pod health. That cached pod health information
is available via the extension API that the metrics-server makes available.
For example with an HPA query, the metrics-server needs to identify which pods fulfill the label
selectors in the deployment.
The metrics-server calls the kubelet API
to collect metrics from each node. Depending on the metrics-server version it uses:
Metrics resource endpoint /metrics/resource in version v0.6.0+ or
Summary API endpoint /stats/summary in older versions
Note: The summary API /stats/summary endpoint will be replaced by the /metrics/resource endpoint
beginning with metrics-server 0.6.x.
4.4.2.2 - Tools for Monitoring Resources
To scale an application and provide a reliable service, you need to
understand how the application behaves when it is deployed. You can examine
application performance in a Kubernetes cluster by examining the containers,
pods,
services, and
the characteristics of the overall cluster. Kubernetes provides detailed
information about an application's resource usage at each of these levels.
This information allows you to evaluate your application's performance and
where bottlenecks can be removed to improve overall performance.
In Kubernetes, application monitoring does not depend on a single monitoring solution.
On new clusters, you can use resource metrics or
full metrics pipelines to collect monitoring statistics.
Resource metrics pipeline
The resource metrics pipeline provides a limited set of metrics related to
cluster components such as the
Horizontal Pod Autoscaler
controller, as well as the kubectl top utility.
These metrics are collected by the lightweight, short-term, in-memory
metrics-server and
are exposed via the metrics.k8s.io API.
metrics-server discovers all nodes on the cluster and
queries each node's
kubelet for CPU and
memory usage. The kubelet acts as a bridge between the Kubernetes master and
the nodes, managing the pods and containers running on a machine. The kubelet
translates each pod into its constituent containers and fetches individual
container usage statistics from the container runtime through the container
runtime interface. If you use a container runtime that uses Linux cgroups and
namespaces to implement containers, and the container runtime does not publish
usage statistics, then the kubelet can look up those statistics directly
(using code from cAdvisor).
No matter how those statistics arrive, the kubelet then exposes the aggregated pod
resource usage statistics through the metrics-server Resource Metrics API.
This API is served at /metrics/resource/v1beta1 on the kubelet's authenticated and
read-only ports.
Full metrics pipeline
A full metrics pipeline gives you access to richer metrics. Kubernetes can
respond to these metrics by automatically scaling or adapting the cluster
based on its current state, using mechanisms such as the Horizontal Pod
Autoscaler. The monitoring pipeline fetches metrics from the kubelet and
then exposes them to Kubernetes via an adapter by implementing either the
custom.metrics.k8s.io or external.metrics.k8s.io API.
Prometheus, a CNCF project, can natively monitor Kubernetes, nodes, and Prometheus itself.
Full metrics pipeline projects that are not part of the CNCF are outside the scope of Kubernetes documentation.
What's next
Learn about additional debugging tools, including:
Node Problem Detector is a daemon for monitoring and reporting about a node's health.
You can run Node Problem Detector as a DaemonSet or as a standalone daemon.
Node Problem Detector collects information about node problems from various daemons
and reports these conditions to the API server as NodeCondition
and Event.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Node Problem Detector only supports file based kernel log.
Log tools such as journald are not supported.
Node Problem Detector uses the kernel log format for reporting kernel issues.
To learn how to extend the kernel log format, see Add support for another log format.
Enabling Node Problem Detector
Some cloud providers enable Node Problem Detector as an Addon.
You can also enable Node Problem Detector with kubectl or by creating an Addon pod.
Using kubectl to enable Node Problem Detector
kubectl provides the most flexible management of Node Problem Detector.
You can overwrite the default configuration to fit it into your environment or
to detect customized node problems. For example:
Create a Node Problem Detector configuration similar to node-problem-detector.yaml:
Using an Addon pod to enable Node Problem Detector
If you are using a custom cluster bootstrap solution and don't need
to overwrite the default configuration, you can leverage the Addon pod to
further automate the deployment.
Create node-problem-detector.yaml, and save the configuration in the Addon pod's
directory /etc/kubernetes/addons/node-problem-detector on a control plane node.
Overwrite the configuration
The default configuration
is embedded when building the Docker image of Node Problem Detector.
However, you can use a ConfigMap
to overwrite the configuration:
apiVersion:apps/v1kind:DaemonSetmetadata:name:node-problem-detector-v0.1namespace:kube-systemlabels:k8s-app:node-problem-detectorversion:v0.1kubernetes.io/cluster-service:"true"spec:selector:matchLabels:k8s-app:node-problem-detector version:v0.1kubernetes.io/cluster-service:"true"template:metadata:labels:k8s-app:node-problem-detectorversion:v0.1kubernetes.io/cluster-service:"true"spec:hostNetwork:truecontainers:- name:node-problem-detectorimage:registry.k8s.io/node-problem-detector:v0.1securityContext:privileged:trueresources:limits:cpu:"200m"memory:"100Mi"requests:cpu:"20m"memory:"20Mi"volumeMounts:- name:logmountPath:/logreadOnly:true- name:config# Overwrite the config/ directory with ConfigMap volumemountPath:/configreadOnly:truevolumes:- name:loghostPath:path:/var/log/- name:config# Define ConfigMap volumeconfigMap:name:node-problem-detector-config
Recreate the Node Problem Detector with the new configuration file:
# If you have a node-problem-detector running, delete before recreatingkubectl delete -f https://k8s.io/examples/debug/node-problem-detector.yaml
kubectl apply -f https://k8s.io/examples/debug/node-problem-detector-configmap.yaml
Note: This approach only applies to a Node Problem Detector started with kubectl.
Overwriting a configuration is not supported if a Node Problem Detector runs as a cluster Addon.
The Addon manager does not support ConfigMap.
Kernel Monitor
Kernel Monitor is a system log monitor daemon supported in the Node Problem Detector.
Kernel monitor watches the kernel log and detects known kernel issues following predefined rules.
The Kernel Monitor matches kernel issues according to a set of predefined rule list in
config/kernel-monitor.json. The rule list is extensible. You can expand the rule list by overwriting the
configuration.
Add new NodeConditions
To support a new NodeCondition, create a condition definition within the conditions field in
config/kernel-monitor.json, for example:
To detect new problems, you can extend the rules field in config/kernel-monitor.json
with a new rule definition:
{
"type": "temporary/permanent",
"condition": "NodeConditionOfPermanentIssue",
"reason": "CamelCaseShortReason",
"message": "regexp matching the issue in the kernel log"}
Configure path for the kernel log device
Check your kernel log path location in your operating system (OS) distribution.
The Linux kernel log device is usually presented as /dev/kmsg. However, the log path location varies by OS distribution.
The log field in config/kernel-monitor.json represents the log path inside the container.
You can configure the log field to match the device path as seen by the Node Problem Detector.
Add support for another log format
Kernel monitor uses the
Translator plugin to translate the internal data structure of the kernel log.
You can implement a new translator for a new log format.
Recommendations and restrictions
It is recommended to run the Node Problem Detector in your cluster to monitor node health.
When running the Node Problem Detector, you can expect extra resource overhead on each node.
Usually this is fine, because:
The kernel log grows relatively slowly.
A resource limit is set for the Node Problem Detector.
Even under high load, the resource usage is acceptable. For more information, see the Node Problem Detector
benchmark result.
4.4.2.4 - Debugging Kubernetes nodes with crictl
FEATURE STATE:Kubernetes v1.11 [stable]
crictl is a command-line interface for CRI-compatible container runtimes.
You can use it to inspect and debug container runtimes and applications on a
Kubernetes node. crictl and its source are hosted in the
cri-tools repository.
Before you begin
crictl requires a Linux operating system with a CRI runtime.
Installing crictl
You can download a compressed archive crictl from the cri-tools
release page, for several
different architectures. Download the version that corresponds to your version
of Kubernetes. Extract it and move it to a location on your system path, such as
/usr/local/bin/.
General usage
The crictl command has several subcommands and runtime flags. Use
crictl help or crictl <subcommand> help for more details.
You can set the endpoint for crictl by doing one of the following:
Set the --runtime-endpoint and --image-endpoint flags.
Set the CONTAINER_RUNTIME_ENDPOINT and IMAGE_SERVICE_ENDPOINT environment
variables.
Set the endpoint in the configuration file /etc/crictl.yaml. To specify a
different file, use the --config=PATH_TO_FILE flag when you run crictl.
Note: If you don't set an endpoint, crictl attempts to connect to a list of known
endpoints, which might result in an impact to performance.
You can also specify timeout values when connecting to the server and enable or
disable debugging, by specifying timeout or debug values in the configuration
file or using the --timeout and --debug command-line flags.
To view or edit the current configuration, view or edit the contents of
/etc/crictl.yaml. For example, the configuration when using the containerd
container runtime would be similar to this:
The following examples show some crictl commands and example output.
Warning: If you use crictl to create pod sandboxes or containers on a running
Kubernetes cluster, the Kubelet will eventually delete them. crictl is not a
general purpose workflow tool, but a tool that is useful for debugging.
List pods
List all pods:
crictl pods
The output is similar to this:
POD ID CREATED STATE NAME NAMESPACE ATTEMPT
926f1b5a1d33a About a minute ago Ready sh-84d7dcf559-4r2gq default 0
4dccb216c4adb About a minute ago Ready nginx-65899c769f-wv2gp default 0
a86316e96fa89 17 hours ago Ready kube-proxy-gblk4 kube-system 0
919630b8f81f1 17 hours ago Ready nvidia-device-plugin-zgbbv kube-system 0
List pods by name:
crictl pods --name nginx-65899c769f-wv2gp
The output is similar to this:
POD ID CREATED STATE NAME NAMESPACE ATTEMPT
4dccb216c4adb 2 minutes ago Ready nginx-65899c769f-wv2gp default 0
List pods by label:
crictl pods --label run=nginx
The output is similar to this:
POD ID CREATED STATE NAME NAMESPACE ATTEMPT
4dccb216c4adb 2 minutes ago Ready nginx-65899c769f-wv2gp default 0
CONTAINER ID IMAGE CREATED STATE NAME ATTEMPT
1f73f2d81bf98 busybox@sha256:141c253bc4c3fd0a201d32dc1f493bcf3fff003b6df416dea4f41046e0f37d47 7 minutes ago Running sh 1
9c5951df22c78 busybox@sha256:141c253bc4c3fd0a201d32dc1f493bcf3fff003b6df416dea4f41046e0f37d47 8 minutes ago Exited sh 0
87d3992f84f74 nginx@sha256:d0a8828cccb73397acb0073bf34f4d7d8aa315263f1e7806bf8c55d8ac139d5f 8 minutes ago Running nginx 0
1941fb4da154f k8s-gcrio.azureedge.net/hyperkube-amd64@sha256:00d814b1f7763f4ab5be80c58e98140dfc69df107f253d7fdd714b30a714260a 18 hours ago Running kube-proxy 0
List running containers:
crictl ps
The output is similar to this:
CONTAINER ID IMAGE CREATED STATE NAME ATTEMPT
1f73f2d81bf98 busybox@sha256:141c253bc4c3fd0a201d32dc1f493bcf3fff003b6df416dea4f41046e0f37d47 6 minutes ago Running sh 1
87d3992f84f74 nginx@sha256:d0a8828cccb73397acb0073bf34f4d7d8aa315263f1e7806bf8c55d8ac139d5f 7 minutes ago Running nginx 0
1941fb4da154f k8s-gcrio.azureedge.net/hyperkube-amd64@sha256:00d814b1f7763f4ab5be80c58e98140dfc69df107f253d7fdd714b30a714260a 17 hours ago Running kube-proxy 0
Using crictl to run a pod sandbox is useful for debugging container runtimes.
On a running Kubernetes cluster, the sandbox will eventually be stopped and
deleted by the Kubelet.
Use the crictl runp command to apply the JSON and run the sandbox.
crictl runp pod-config.json
The ID of the sandbox is returned.
Create a container
Using crictl to create a container is useful for debugging container runtimes.
On a running Kubernetes cluster, the sandbox will eventually be stopped and
deleted by the Kubelet.
Pull a busybox image
crictl pull busybox
Image is up to date for busybox@sha256:141c253bc4c3fd0a201d32dc1f493bcf3fff003b6df416dea4f41046e0f37d47
Create the container, passing the ID of the previously-created pod, the
container config file, and the pod config file. The ID of the container is
returned.
Kubernetes auditing provides a security-relevant, chronological set of records documenting
the sequence of actions in a cluster. The cluster audits the activities generated by users,
by applications that use the Kubernetes API, and by the control plane itself.
Auditing allows cluster administrators to answer the following questions:
what happened?
when did it happen?
who initiated it?
on what did it happen?
where was it observed?
from where was it initiated?
to where was it going?
Audit records begin their lifecycle inside the
kube-apiserver
component. Each request on each stage
of its execution generates an audit event, which is then pre-processed according to
a certain policy and written to a backend. The policy determines what's recorded
and the backends persist the records. The current backend implementations
include logs files and webhooks.
Each request can be recorded with an associated stage. The defined stages are:
RequestReceived - The stage for events generated as soon as the audit
handler receives the request, and before it is delegated down the handler
chain.
ResponseStarted - Once the response headers are sent, but before the
response body is sent. This stage is only generated for long-running requests
(e.g. watch).
ResponseComplete - The response body has been completed and no more bytes
will be sent.
The audit logging feature increases the memory consumption of the API server
because some context required for auditing is stored for each request.
Memory consumption depends on the audit logging configuration.
Audit policy
Audit policy defines rules about what events should be recorded and what data
they should include. The audit policy object structure is defined in the
audit.k8s.io API group.
When an event is processed, it's
compared against the list of rules in order. The first matching rule sets the
audit level of the event. The defined audit levels are:
None - don't log events that match this rule.
Metadata - log request metadata (requesting user, timestamp, resource,
verb, etc.) but not request or response body.
Request - log event metadata and request body but not response body.
This does not apply for non-resource requests.
RequestResponse - log event metadata, request and response bodies.
This does not apply for non-resource requests.
You can pass a file with the policy to kube-apiserver
using the --audit-policy-file flag. If the flag is omitted, no events are logged.
Note that the rules field must be provided in the audit policy file.
A policy with no (0) rules is treated as illegal.
apiVersion:audit.k8s.io/v1# This is required.kind:Policy# Don't generate audit events for all requests in RequestReceived stage.omitStages:- "RequestReceived"rules:# Log pod changes at RequestResponse level- level:RequestResponseresources:- group:""# Resource "pods" doesn't match requests to any subresource of pods,# which is consistent with the RBAC policy.resources:["pods"]# Log "pods/log", "pods/status" at Metadata level- level:Metadataresources:- group:""resources:["pods/log","pods/status"]# Don't log requests to a configmap called "controller-leader"- level:Noneresources:- group:""resources:["configmaps"]resourceNames:["controller-leader"]# Don't log watch requests by the "system:kube-proxy" on endpoints or services- level:Noneusers:["system:kube-proxy"]verbs:["watch"]resources:- group:""# core API groupresources:["endpoints","services"]# Don't log authenticated requests to certain non-resource URL paths.- level:NoneuserGroups:["system:authenticated"]nonResourceURLs:- "/api*"# Wildcard matching.- "/version"# Log the request body of configmap changes in kube-system.- level:Requestresources:- group:""# core API groupresources:["configmaps"]# This rule only applies to resources in the "kube-system" namespace.# The empty string "" can be used to select non-namespaced resources.namespaces:["kube-system"]# Log configmap and secret changes in all other namespaces at the Metadata level.- level:Metadataresources:- group:""# core API groupresources:["secrets","configmaps"]# Log all other resources in core and extensions at the Request level.- level:Requestresources:- group:""# core API group- group:"extensions"# Version of group should NOT be included.# A catch-all rule to log all other requests at the Metadata level.- level:Metadata# Long-running requests like watches that fall under this rule will not# generate an audit event in RequestReceived.omitStages:- "RequestReceived"
You can use a minimal audit policy file to log all requests at the Metadata level:
# Log all requests at the Metadata level.apiVersion:audit.k8s.io/v1kind:Policyrules:- level:Metadata
If you're crafting your own audit profile, you can use the audit profile for Google Container-Optimized OS as a starting point. You can check the
configure-helper.sh
script, which generates an audit policy file. You can see most of the audit policy file by looking directly at the script.
Audit backends persist audit events to an external storage.
Out of the box, the kube-apiserver provides two backends:
Log backend, which writes events into the filesystem
Webhook backend, which sends events to an external HTTP API
In all cases, audit events follow a structure defined by the Kubernetes API in the
audit.k8s.io API group.
Note:
In case of patches, request body is a JSON array with patch operations, not a JSON object
with an appropriate Kubernetes API object. For example, the following request body is a valid patch
request to /apis/batch/v1/namespaces/some-namespace/jobs/some-job-name:
The log backend writes audit events to a file in JSONlines format.
You can configure the log audit backend using the following kube-apiserver flags:
--audit-log-path specifies the log file path that log backend uses to write
audit events. Not specifying this flag disables log backend. - means standard out
--audit-log-maxage defined the maximum number of days to retain old audit log files
--audit-log-maxbackup defines the maximum number of audit log files to retain
--audit-log-maxsize defines the maximum size in megabytes of the audit log file before it gets rotated
If your cluster's control plane runs the kube-apiserver as a Pod, remember to mount the hostPath
to the location of the policy file and log file, so that audit records are persisted. For example:
The webhook audit backend sends audit events to a remote web API, which is assumed to
be a form of the Kubernetes API, including means of authentication. You can configure
a webhook audit backend using the following kube-apiserver flags:
--audit-webhook-config-file specifies the path to a file with a webhook
configuration. The webhook configuration is effectively a specialized
kubeconfig.
--audit-webhook-initial-backoff specifies the amount of time to wait after the first failed
request before retrying. Subsequent requests are retried with exponential backoff.
The webhook config file uses the kubeconfig format to specify the remote address of
the service and credentials used to connect to it.
Event batching
Both log and webhook backends support batching. Using webhook as an example, here's the list of
available flags. To get the same flag for log backend, replace webhook with log in the flag
name. By default, batching is enabled in webhook and disabled in log. Similarly, by default
throttling is enabled in webhook and disabled in log.
--audit-webhook-mode defines the buffering strategy. One of the following:
batch - buffer events and asynchronously process them in batches. This is the default.
blocking - block API server responses on processing each individual event.
blocking-strict - Same as blocking, but when there is a failure during audit logging at the
RequestReceived stage, the whole request to the kube-apiserver fails.
The following flags are used only in the batch mode:
--audit-webhook-batch-buffer-size defines the number of events to buffer before batching.
If the rate of incoming events overflows the buffer, events are dropped.
--audit-webhook-batch-max-size defines the maximum number of events in one batch.
--audit-webhook-batch-max-wait defines the maximum amount of time to wait before unconditionally
batching events in the queue.
--audit-webhook-batch-throttle-qps defines the maximum average number of batches generated
per second.
--audit-webhook-batch-throttle-burst defines the maximum number of batches generated at the same
moment if the allowed QPS was underutilized previously.
Parameter tuning
Parameters should be set to accommodate the load on the API server.
For example, if kube-apiserver receives 100 requests each second, and each request is audited only
on ResponseStarted and ResponseComplete stages, you should account for ≅200 audit
events being generated each second. Assuming that there are up to 100 events in a batch,
you should set throttling level at least 2 queries per second. Assuming that the backend can take up to
5 seconds to write events, you should set the buffer size to hold up to 5 seconds of events;
that is: 10 batches, or 1000 events.
In most cases however, the default parameters should be sufficient and you don't have to worry about
setting them manually. You can look at the following Prometheus metrics exposed by kube-apiserver
and in the logs to monitor the state of the auditing subsystem.
apiserver_audit_event_total metric contains the total number of audit events exported.
apiserver_audit_error_total metric contains the total number of events dropped due to an error
during exporting.
Log entry truncation
Both log and webhook backends support limiting the size of events that are logged.
As an example, the following is the list of flags available for the log backend:
audit-log-truncate-enabled whether event and batch truncating is enabled.
audit-log-truncate-max-batch-size maximum size in bytes of the batch sent to the underlying backend.
audit-log-truncate-max-event-size maximum size in bytes of the audit event sent to the underlying backend.
By default truncate is disabled in both webhook and log, a cluster administrator should set
audit-log-truncate-enabled or audit-webhook-truncate-enabled to enable the feature.
Learn more about Event
and the Policy
resource types by reading the Audit configuration reference.
4.4.2.6 - Developing and debugging services locally using telepresence
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Kubernetes applications usually consist of multiple, separate services, each running in its own container. Developing and debugging these services on a remote Kubernetes cluster can be cumbersome, requiring you to get a shell on a running container in order to run debugging tools.
telepresence is a tool to ease the process of developing and debugging services locally while proxying the service to a remote Kubernetes cluster. Using telepresence allows you to use custom tools, such as a debugger and IDE, for a local service and provides the service full access to ConfigMap, secrets, and the services running on the remote cluster.
This document describes using telepresence to develop and debug services running on a remote cluster locally.
Before you begin
Kubernetes cluster is installed
kubectl is configured to communicate with the cluster
Connecting your local machine to a remote Kubernetes cluster
After installing telepresence, run telepresence connect to launch its Daemon and connect your local workstation to the cluster.
$ telepresence connect
Launching Telepresence Daemon
...
Connected to context default (https://<cluster public IP>)
You can curl services using the Kubernetes syntax e.g. curl -ik https://kubernetes.default
Developing or debugging an existing service
When developing an application on Kubernetes, you typically program or debug a single service. The service might require access to other services for testing and debugging. One option is to use the continuous deployment pipeline, but even the fastest deployment pipeline introduces a delay in the program or debug cycle.
Use the telepresence intercept $SERVICE_NAME --port $LOCAL_PORT:REMOTE_PORT command to create an "intercept" for rerouting remote service traffic.
Where:
$SERVICE_NAME is the name of your local service
$LOCAL_PORT is the port that your service is running on your local workstation
And $REMOTE_PORT is the port your service listens to in the cluster
Running this command tells Telepresence to send remote traffic to your local service instead of the service in the remote Kubernetes cluster. Make edits to your service source code locally, save, and see the corresponding changes when accessing your remote application take effect immediately. You can also run your local service using a debugger or any other local development tool.
How does Telepresence work?
Telepresence installs a traffic-agent sidecar next to your existing application's container running in the remote cluster. It then captures all traffic requests going into the Pod, and instead of forwarding this to the application in the remote cluster, it routes all traffic (when you create a global intercept) or a subset of the traffic (when you create a personal intercept) to your local development environment.
What's next
If you're interested in a hands-on tutorial, check out this tutorial that walks through locally developing the Guestbook application on Google Kubernetes Engine.
My Pods are stuck at "Container Creating" or restarting over and over
Ensure that your pause image is compatible with your Windows OS version.
See Pause container
to see the latest / recommended pause image and/or get more information.
Note: If using containerd as your container runtime the pause image is specified in the
plugins.plugins.cri.sandbox_image field of the of config.toml configration file.
My pods show status as ErrImgPull or ImagePullBackOff
Ensure that your Pod is getting scheduled to a
compatible
Windows Node.
More information on how to specify a compatible node for your Pod can be found in
this guide.
Network troubleshooting
My Windows Pods do not have network connectivity
If you are using virtual machines, ensure that MAC spoofing is enabled on all
the VM network adapter(s).
My Windows Pods cannot ping external resources
Windows Pods do not have outbound rules programmed for the ICMP protocol. However,
TCP/UDP is supported. When trying to demonstrate connectivity to resources
outside of the cluster, substitute ping <IP> with corresponding
curl <IP> commands.
If you are still facing problems, most likely your network configuration in
cni.conf
deserves some extra attention. You can always edit this static file. The
configuration update will apply to any new Kubernetes resources.
One of the Kubernetes networking requirements
(see Kubernetes model) is
for cluster communication to occur without
NAT internally. To honor this requirement, there is an
ExceptionList
for all the communication where you do not want outbound NAT to occur. However,
this also means that you need to exclude the external IP you are trying to query
from the ExceptionList. Only then will the traffic originating from your Windows
pods be SNAT'ed correctly to receive a response from the outside world. In this
regard, your ExceptionList in cni.conf should look as follows:
My Windows node cannot access NodePort type Services
Local NodePort access from the node itself fails. This is a known
limitation. NodePort access works from other nodes or external clients.
vNICs and HNS endpoints of containers are being deleted
This issue can be caused when the hostname-override parameter is not passed to
kube-proxy. To resolve
it, users need to pass the hostname to kube-proxy as follows:
My Windows node cannot access my services using the service IP
This is a known limitation of the networking stack on Windows. However, Windows Pods can access the Service IP.
No network adapter is found when starting the kubelet
The Windows networking stack needs a virtual adapter for Kubernetes networking to work.
If the following commands return no results (in an admin shell),
virtual network creation — a necessary prerequisite for the kubelet to work — has failed:
Get-HnsNetwork | ? Name -ieq"cbr0"Get-NetAdapter | ? Name -Like"vEthernet (Ethernet*"
Often it is worthwhile to modify the InterfaceName
parameter of the start.ps1 script, in cases where the host's network adapter isn't "Ethernet".
Otherwise, consult the output of the start-kubelet.ps1 script to see if there are errors during virtual network creation.
DNS resolution is not properly working
Check the DNS limitations for Windows in this section.
kubectl port-forward fails with "unable to do port forwarding: wincat not found"
This was implemented in Kubernetes 1.15 by including wincat.exe in the pause infrastructure container
mcr.microsoft.com/oss/kubernetes/pause:3.6.
Be sure to use a supported version of Kubernetes.
If you would like to build your own pause infrastructure container be sure to include
wincat.
My Kubernetes installation is failing because my Windows Server node is behind a proxy
If you are behind a proxy, the following PowerShell environment variables must be defined:
With Flannel, my nodes are having issues after rejoining a cluster
Whenever a previously deleted node is being re-joined to the cluster, flannelD
tries to assign a new pod subnet to the node. Users should remove the old pod
subnet configuration files in the following paths:
Flanneld is stuck in "Waiting for the Network to be created"
There are numerous reports of this issue;
most likely it is a timing issue for when the management IP of the flannel network is set.
A workaround is to relaunch start.ps1 or relaunch it manually as follows:
My Windows Pods cannot launch because of missing /run/flannel/subnet.env
This indicates that Flannel didn't launch correctly. You can either try
to restart flanneld.exe or you can copy the files over manually from
/run/flannel/subnet.env on the Kubernetes master to C:\run\flannel\subnet.env
on the Windows worker node and modify the FLANNEL_SUBNET row to a different
number. For example, if node subnet 10.244.4.1/24 is desired:
Declarative and imperative paradigms for interacting with the Kubernetes API.
4.5.1 - Declarative Management of Kubernetes Objects Using Configuration Files
Kubernetes objects can be created, updated, and deleted by storing multiple
object configuration files in a directory and using kubectl apply to
recursively create and update those objects as needed. This method
retains writes made to live objects without merging the changes
back into the object configuration files. kubectl diff also gives you a
preview of what changes apply will make.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The kubectl tool supports three kinds of object management:
Imperative commands
Imperative object configuration
Declarative object configuration
See Kubernetes Object Management
for a discussion of the advantages and disadvantage of each kind of object management.
Overview
Declarative object configuration requires a firm understanding of
the Kubernetes object definitions and configuration. Read and complete
the following documents if you have not already:
Following are definitions for terms used in this document:
object configuration file / configuration file: A file that defines the
configuration for a Kubernetes object. This topic shows how to pass configuration
files to kubectl apply. Configuration files are typically stored in source control, such as Git.
live object configuration / live configuration: The live configuration
values of an object, as observed by the Kubernetes cluster. These are kept in the Kubernetes
cluster storage, typically etcd.
declarative configuration writer / declarative writer: A person or software component
that makes updates to a live object. The live writers referred to in this topic make changes
to object configuration files and run kubectl apply to write the changes.
How to create objects
Use kubectl apply to create all objects, except those that already exist,
defined by configuration files in a specified directory:
kubectl apply -f <directory>/
This sets the kubectl.kubernetes.io/last-applied-configuration: '{...}'
annotation on each object. The annotation contains the contents of the object
configuration file that was used to create the object.
Note: Add the -R flag to recursively process directories.
Here's an example of an object configuration file:
Since diff performs a server-side apply request in dry-run mode,
it requires granting PATCH, CREATE, and UPDATE permissions.
See Dry-Run Authorization
for details.
kubectl get -f https://k8s.io/examples/application/simple_deployment.yaml -o yaml
The output shows that the kubectl.kubernetes.io/last-applied-configuration annotation
was written to the live configuration, and it matches the configuration file:
kind:Deploymentmetadata:annotations:# ...# This is the json representation of simple_deployment.yaml# It was written by kubectl apply when the object was createdkubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"minReadySeconds":5,"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.14.2","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:# ...minReadySeconds:5selector:matchLabels:# ...app:nginxtemplate:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.14.2# ...name:nginxports:- containerPort:80# ...# ...# ...# ...
How to update objects
You can also use kubectl apply to update all objects defined in a directory, even
if those objects already exist. This approach accomplishes the following:
Sets fields that appear in the configuration file in the live configuration.
Clears fields removed from the configuration file in the live configuration.
Note: For purposes of illustration, the preceding command refers to a single
configuration file instead of a directory.
Print the live configuration using kubectl get:
kubectl get -f https://k8s.io/examples/application/simple_deployment.yaml -o yaml
The output shows that the kubectl.kubernetes.io/last-applied-configuration annotation
was written to the live configuration, and it matches the configuration file:
kind:Deploymentmetadata:annotations:# ...# This is the json representation of simple_deployment.yaml# It was written by kubectl apply when the object was createdkubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"minReadySeconds":5,"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.14.2","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:# ...minReadySeconds:5selector:matchLabels:# ...app:nginxtemplate:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.14.2# ...name:nginxports:- containerPort:80# ...# ...# ...# ...
Directly update the replicas field in the live configuration by using kubectl scale.
This does not use kubectl apply:
The output shows that the replicas field has been set to 2, and the last-applied-configuration
annotation does not contain a replicas field:
apiVersion:apps/v1kind:Deploymentmetadata:annotations:# ...# note that the annotation does not contain replicas# because it was not updated through applykubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"minReadySeconds":5,"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.14.2","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:replicas:2# written by scale# ...minReadySeconds:5selector:matchLabels:# ...app:nginxtemplate:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.14.2# ...name:nginxports:- containerPort:80# ...
Update the simple_deployment.yaml configuration file to change the image from
nginx:1.14.2 to nginx:1.16.1, and delete the minReadySeconds field:
kubectl get -f https://k8s.io/examples/application/update_deployment.yaml -o yaml
The output shows the following changes to the live configuration:
The replicas field retains the value of 2 set by kubectl scale.
This is possible because it is omitted from the configuration file.
The image field has been updated to nginx:1.16.1 from nginx:1.14.2.
The last-applied-configuration annotation has been updated with the new image.
The minReadySeconds field has been cleared.
The last-applied-configuration annotation no longer contains the minReadySeconds field.
apiVersion:apps/v1kind:Deploymentmetadata:annotations:# ...# The annotation contains the updated image to nginx 1.11.9,# but does not contain the updated replicas to 2kubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.16.1","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:replicas:2# Set by `kubectl scale`. Ignored by `kubectl apply`.# minReadySeconds cleared by `kubectl apply`# ...selector:matchLabels:# ...app:nginxtemplate:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.16.1# Set by `kubectl apply`# ...name:nginxports:- containerPort:80# ...# ...# ...# ...
Warning: Mixing kubectl apply with the imperative object configuration commands
create and replace is not supported. This is because create
and replace do not retain the kubectl.kubernetes.io/last-applied-configuration
that kubectl apply uses to compute updates.
How to delete objects
There are two approaches to delete objects managed by kubectl apply.
Recommended: kubectl delete -f <filename>
Manually deleting objects using the imperative command is the recommended
approach, as it is more explicit about what is being deleted, and less likely
to result in the user deleting something unintentionally:
Warning:kubectl apply --prune is in alpha, and backwards incompatible
changes might be introduced in subsequent releases.
Warning: You must be careful when using this command, so that you
do not delete objects unintentionally.
As an alternative to kubectl delete, you can use kubectl apply to identify objects to be deleted after their
configuration files have been removed from the directory. Apply with --prune
queries the API server for all objects matching a set of labels, and attempts
to match the returned live object configurations against the object
configuration files. If an object matches the query, and it does not have a
configuration file in the directory, and it has a last-applied-configuration annotation,
it is deleted.
kubectl apply -f <directory/> --prune -l <labels>
Warning: Apply with prune should only be run against the root directory
containing the object configuration files. Running against sub-directories
can cause objects to be unintentionally deleted if they are returned
by the label selector query specified with -l <labels> and
do not appear in the subdirectory.
How to view an object
You can use kubectl get with -o yaml to view the configuration of a live object:
kubectl get -f <filename|url> -o yaml
How apply calculates differences and merges changes
Caution: A patch is an update operation that is scoped to specific fields of an object
instead of the entire object. This enables updating only a specific set of fields
on an object without reading the object first.
When kubectl apply updates the live configuration for an object,
it does so by sending a patch request to the API server. The
patch defines updates scoped to specific fields of the live object
configuration. The kubectl apply command calculates this patch request
using the configuration file, the live configuration, and the
last-applied-configuration annotation stored in the live configuration.
Merge patch calculation
The kubectl apply command writes the contents of the configuration file to the
kubectl.kubernetes.io/last-applied-configuration annotation. This
is used to identify fields that have been removed from the configuration
file and need to be cleared from the live configuration. Here are the steps used
to calculate which fields should be deleted or set:
Calculate the fields to delete. These are the fields present in last-applied-configuration and missing from the configuration file.
Calculate the fields to add or set. These are the fields present in the configuration file whose values don't match the live configuration.
Here's an example. Suppose this is the configuration file for a Deployment object:
apiVersion:apps/v1kind:Deploymentmetadata:name:nginx-deploymentspec:selector:matchLabels:app:nginxtemplate:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.16.1# update the imageports:- containerPort:80
Also, suppose this is the live configuration for the same Deployment object:
apiVersion:apps/v1kind:Deploymentmetadata:annotations:# ...# note that the annotation does not contain replicas# because it was not updated through applykubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"minReadySeconds":5,"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.14.2","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:replicas:2# written by scale# ...minReadySeconds:5selector:matchLabels:# ...app:nginxtemplate:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.14.2# ...name:nginxports:- containerPort:80# ...
Here are the merge calculations that would be performed by kubectl apply:
Calculate the fields to delete by reading values from
last-applied-configuration and comparing them to values in the
configuration file.
Clear fields explicitly set to null in the local object configuration file
regardless of whether they appear in the last-applied-configuration.
In this example, minReadySeconds appears in the
last-applied-configuration annotation, but does not appear in the configuration file.
Action: Clear minReadySeconds from the live configuration.
Calculate the fields to set by reading values from the configuration
file and comparing them to values in the live configuration. In this example,
the value of image in the configuration file does not match
the value in the live configuration. Action: Set the value of image in the live configuration.
Set the last-applied-configuration annotation to match the value
of the configuration file.
Merge the results from 1, 2, 3 into a single patch request to the API server.
Here is the live configuration that is the result of the merge:
apiVersion:apps/v1kind:Deploymentmetadata:annotations:# ...# The annotation contains the updated image to nginx 1.11.9,# but does not contain the updated replicas to 2kubectl.kubernetes.io/last-applied-configuration:| {"apiVersion":"apps/v1","kind":"Deployment",
"metadata":{"annotations":{},"name":"nginx-deployment","namespace":"default"},
"spec":{"selector":{"matchLabels":{"app":nginx}},"template":{"metadata":{"labels":{"app":"nginx"}},
"spec":{"containers":[{"image":"nginx:1.16.1","name":"nginx",
"ports":[{"containerPort":80}]}]}}}}# ...spec:selector:matchLabels:# ...app:nginxreplicas:2# Set by `kubectl scale`. Ignored by `kubectl apply`.# minReadySeconds cleared by `kubectl apply`# ...template:metadata:# ...labels:app:nginxspec:containers:- image:nginx:1.16.1# Set by `kubectl apply`# ...name:nginxports:- containerPort:80# ...# ...# ...# ...
How different types of fields are merged
How a particular field in a configuration file is merged with
the live configuration depends on the
type of the field. There are several types of fields:
primitive: A field of type string, integer, or boolean.
For example, image and replicas are primitive fields. Action: Replace.
map, also called object: A field of type map or a complex type that contains subfields. For example, labels,
annotations,spec and metadata are all maps. Action: Merge elements or subfields.
list: A field containing a list of items that can be either primitive types or maps.
For example, containers, ports, and args are lists. Action: Varies.
When kubectl apply updates a map or list field, it typically does
not replace the entire field, but instead updates the individual subelements.
For instance, when merging the spec on a Deployment, the entire spec is
not replaced. Instead the subfields of spec, such as replicas, are compared
and merged.
Merging changes to primitive fields
Primitive fields are replaced or cleared.
Note:- is used for "not applicable" because the value is not used.
Field in object configuration file
Field in live object configuration
Field in last-applied-configuration
Action
Yes
Yes
-
Set live to configuration file value.
Yes
No
-
Set live to local configuration.
No
-
Yes
Clear from live configuration.
No
-
No
Do nothing. Keep live value.
Merging changes to map fields
Fields that represent maps are merged by comparing each of the subfields or elements of the map:
Note:- is used for "not applicable" because the value is not used.
Key in object configuration file
Key in live object configuration
Field in last-applied-configuration
Action
Yes
Yes
-
Compare sub fields values.
Yes
No
-
Set live to local configuration.
No
-
Yes
Delete from live configuration.
No
-
No
Do nothing. Keep live value.
Merging changes for fields of type list
Merging changes to a list uses one of three strategies:
Replace the list if all its elements are primitives.
Merge individual elements in a list of complex elements.
Merge a list of primitive elements.
The choice of strategy is made on a per-field basis.
Replace the list if all its elements are primitives
Treat the list the same as a primitive field. Replace or delete the
entire list. This preserves ordering.
Example: Use kubectl apply to update the args field of a Container in a Pod. This sets
the value of args in the live configuration to the value in the configuration file.
Any args elements that had previously been added to the live configuration are lost.
The order of the args elements defined in the configuration file is
retained in the live configuration.
# last-applied-configuration valueargs:["a","b"]# configuration file valueargs:["a","c"]# live configurationargs:["a","b","d"]# result after mergeargs:["a","c"]
Explanation: The merge used the configuration file value as the new list value.
Merge individual elements of a list of complex elements:
Treat the list as a map, and treat a specific field of each element as a key.
Add, delete, or update individual elements. This does not preserve ordering.
This merge strategy uses a special tag on each field called a patchMergeKey. The
patchMergeKey is defined for each field in the Kubernetes source code:
types.go
When merging a list of maps, the field specified as the patchMergeKey for a given element
is used like a map key for that element.
Example: Use kubectl apply to update the containers field of a PodSpec.
This merges the list as though it was a map where each element is keyed
by name.
# last-applied-configuration valuecontainers:- name:nginximage:nginx:1.16- name: nginx-helper-a # key:nginx-helper-a; will be deleted in resultimage:helper:1.3- name: nginx-helper-b # key:nginx-helper-b; will be retainedimage:helper:1.3# configuration file valuecontainers:- name:nginximage:nginx:1.16- name:nginx-helper-bimage:helper:1.3- name: nginx-helper-c # key:nginx-helper-c; will be added in resultimage:helper:1.3# live configurationcontainers:- name:nginximage:nginx:1.16- name:nginx-helper-aimage:helper:1.3- name:nginx-helper-bimage:helper:1.3args:["run"]# Field will be retained- name: nginx-helper-d # key:nginx-helper-d; will be retainedimage:helper:1.3# result after mergecontainers:- name:nginximage:nginx:1.16# Element nginx-helper-a was deleted- name:nginx-helper-bimage:helper:1.3args:["run"]# Field was retained- name:nginx-helper-c# Element was addedimage:helper:1.3- name:nginx-helper-d# Element was ignoredimage:helper:1.3
Explanation:
The container named "nginx-helper-a" was deleted because no container
named "nginx-helper-a" appeared in the configuration file.
The container named "nginx-helper-b" retained the changes to args
in the live configuration. kubectl apply was able to identify
that "nginx-helper-b" in the live configuration was the same
"nginx-helper-b" as in the configuration file, even though their fields
had different values (no args in the configuration file). This is
because the patchMergeKey field value (name) was identical in both.
The container named "nginx-helper-c" was added because no container
with that name appeared in the live configuration, but one with
that name appeared in the configuration file.
The container named "nginx-helper-d" was retained because
no element with that name appeared in the last-applied-configuration.
Merge a list of primitive elements
As of Kubernetes 1.5, merging lists of primitive elements is not supported.
Note: Which of the above strategies is chosen for a given field is controlled by
the patchStrategy tag in types.go
If no patchStrategy is specified for a field of type list, then
the list is replaced.
Default field values
The API server sets certain fields to default values in the live configuration if they are
not specified when the object is created.
Here's a configuration file for a Deployment. The file does not specify strategy:
kubectl get -f https://k8s.io/examples/application/simple_deployment.yaml -o yaml
The output shows that the API server set several fields to default values in the live
configuration. These fields were not specified in the configuration file.
apiVersion:apps/v1kind:Deployment# ...spec:selector:matchLabels:app:nginxminReadySeconds:5replicas:1# defaulted by apiserverstrategy:rollingUpdate:# defaulted by apiserver - derived from strategy.typemaxSurge:1maxUnavailable:1type:RollingUpdate# defaulted by apiservertemplate:metadata:creationTimestamp:nulllabels:app:nginxspec:containers:- image:nginx:1.14.2imagePullPolicy:IfNotPresent# defaulted by apiservername:nginxports:- containerPort:80protocol:TCP# defaulted by apiserverresources:{}# defaulted by apiserverterminationMessagePath:/dev/termination-log# defaulted by apiserverdnsPolicy:ClusterFirst# defaulted by apiserverrestartPolicy:Always# defaulted by apiserversecurityContext:{}# defaulted by apiserverterminationGracePeriodSeconds:30# defaulted by apiserver# ...
In a patch request, defaulted fields are not re-defaulted unless they are explicitly cleared
as part of a patch request. This can cause unexpected behavior for
fields that are defaulted based
on the values of other fields. When the other fields are later changed,
the values defaulted from them will not be updated unless they are
explicitly cleared.
For this reason, it is recommended that certain fields defaulted
by the server are explicitly defined in the configuration file, even
if the desired values match the server defaults. This makes it
easier to recognize conflicting values that will not be re-defaulted
by the server.
Example:
# last-applied-configurationspec:template:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80# configuration filespec:strategy:type:Recreate# updated valuetemplate:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80# live configurationspec:strategy:type:RollingUpdate# defaulted valuerollingUpdate:# defaulted value derived from typemaxSurge :1maxUnavailable:1template:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80# result after merge - ERROR!spec:strategy:type: Recreate # updated value:incompatible with rollingUpdaterollingUpdate: # defaulted value:incompatible with "type: Recreate"maxSurge :1maxUnavailable:1template:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80
Explanation:
The user creates a Deployment without defining strategy.type.
The server defaults strategy.type to RollingUpdate and defaults the
strategy.rollingUpdate values.
The user changes strategy.type to Recreate. The strategy.rollingUpdate
values remain at their defaulted values, though the server expects them to be cleared.
If the strategy.rollingUpdate values had been defined initially in the configuration file,
it would have been more clear that they needed to be deleted.
Apply fails because strategy.rollingUpdate is not cleared. The strategy.rollingupdate
field cannot be defined with a strategy.type of Recreate.
Recommendation: These fields should be explicitly defined in the object configuration file:
Selectors and PodTemplate labels on workloads, such as Deployment, StatefulSet, Job, DaemonSet,
ReplicaSet, and ReplicationController
Deployment rollout strategy
How to clear server-defaulted fields or fields set by other writers
Fields that do not appear in the configuration file can be cleared by
setting their values to null and then applying the configuration file.
For fields defaulted by the server, this triggers re-defaulting
the values.
How to change ownership of a field between the configuration file and direct imperative writers
These are the only methods you should use to change an individual object field:
Use kubectl apply.
Write directly to the live configuration without modifying the configuration file:
for example, use kubectl scale.
Changing the owner from a direct imperative writer to a configuration file
Add the field to the configuration file. For the field, discontinue direct updates to
the live configuration that do not go through kubectl apply.
Changing the owner from a configuration file to a direct imperative writer
As of Kubernetes 1.5, changing ownership of a field from a configuration file to
an imperative writer requires manual steps:
Remove the field from the configuration file.
Remove the field from the kubectl.kubernetes.io/last-applied-configuration annotation on the live object.
Changing management methods
Kubernetes objects should be managed using only one method at a time.
Switching from one method to another is possible, but is a manual process.
Note: It is OK to use imperative deletion with declarative management.
Migrating from imperative command management to declarative object configuration
Migrating from imperative command management to declarative object
configuration involves several manual steps:
Export the live object to a local configuration file:
kubectl get <kind>/<name> -o yaml > <kind>_<name>.yaml
Manually remove the status field from the configuration file.
Note: This step is optional, as kubectl apply does not update the status field
even if it is present in the configuration file.
Set the kubectl.kubernetes.io/last-applied-configuration annotation on the object:
Since 1.14, Kubectl also
supports the management of Kubernetes objects using a kustomization file.
To view Resources found in a directory containing a kustomization file, run the following command:
kubectl kustomize <kustomization_directory>
To apply those Resources, run kubectl apply with --kustomize or -k flag:
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Kustomize is a tool for customizing Kubernetes configurations. It has the following features to manage application configuration files:
generating resources from other sources
setting cross-cutting fields for resources
composing and customizing collections of resources
Generating Resources
ConfigMaps and Secrets hold configuration or sensitive data that are used by other Kubernetes objects, such as Pods. The source of truth of ConfigMaps or Secrets are usually external to a cluster, such as a .properties file or an SSH keyfile.
Kustomize has secretGenerator and configMapGenerator, which generate Secret and ConfigMap from files or literals.
configMapGenerator
To generate a ConfigMap from a file, add an entry to the files list in configMapGenerator. Here is an example of generating a ConfigMap with a data item from a .properties file:
To generate a ConfigMap from an env file, add an entry to the envs list in configMapGenerator. Here is an example of generating a ConfigMap with a data item from a .env file:
Note: Each variable in the .env file becomes a separate key in the ConfigMap that you generate. This is different from the previous example which embeds a file named .properties (and all its entries) as the value for a single key.
ConfigMaps can also be generated from literal key-value pairs. To generate a ConfigMap from a literal key-value pair, add an entry to the literals list in configMapGenerator. Here is an example of generating a ConfigMap with a data item from a key-value pair:
To use a generated ConfigMap in a Deployment, reference it by the name of the configMapGenerator. Kustomize will automatically replace this name with the generated name.
This is an example deployment that uses a generated ConfigMap:
You can generate Secrets from files or literal key-value pairs. To generate a Secret from a file, add an entry to the files list in secretGenerator. Here is an example of generating a Secret with a data item from a file:
To generate a Secret from a literal key-value pair, add an entry to literals list in secretGenerator. Here is an example of generating a Secret with a data item from a key-value pair:
The generated ConfigMaps and Secrets have a content hash suffix appended. This ensures that a new ConfigMap or Secret is generated when the contents are changed. To disable the behavior of appending a suffix, one can use generatorOptions. Besides that, it is also possible to specify cross-cutting options for generated ConfigMaps and Secrets.
It is common to compose a set of Resources in a project and manage them inside
the same file or directory.
Kustomize offers composing Resources from different files and applying patches or other customization to them.
Composing
Kustomize supports composition of different resources. The resources field, in the kustomization.yaml file, defines the list of resources to include in a configuration. Set the path to a resource's configuration file in the resources list.
Here is an example of an NGINX application comprised of a Deployment and a Service:
The Resources from kubectl kustomize ./ contain both the Deployment and the Service objects.
Customizing
Patches can be used to apply different customizations to Resources. Kustomize supports different patching
mechanisms through patchesStrategicMerge and patchesJson6902. patchesStrategicMerge is a list of file paths. Each file should be resolved to a strategic merge patch. The names inside the patches must match Resource names that are already loaded. Small patches that do one thing are recommended. For example, create one patch for increasing the deployment replica number and another patch for setting the memory limit.
Not all Resources or fields support strategic merge patches. To support modifying arbitrary fields in arbitrary Resources,
Kustomize offers applying JSON patch through patchesJson6902.
To find the correct Resource for a Json patch, the group, version, kind and name of that Resource need to be
specified in kustomization.yaml. For example, increasing the replica number of a Deployment object can also be done
through patchesJson6902.
In addition to patches, Kustomize also offers customizing container images or injecting field values from other objects into containers
without creating patches. For example, you can change the image used inside containers by specifying the new image in images field in kustomization.yaml.
Sometimes, the application running in a Pod may need to use configuration values from other objects. For example,
a Pod from a Deployment object need to read the corresponding Service name from Env or as a command argument.
Since the Service name may change as namePrefix or nameSuffix is added in the kustomization.yaml file. It is
not recommended to hard code the Service name in the command argument. For this usage, Kustomize can inject the Service name into containers through vars.
Kustomize has the concepts of bases and overlays. A base is a directory with a kustomization.yaml, which contains a
set of resources and associated customization. A base could be either a local directory or a directory from a remote repo,
as long as a kustomization.yaml is present inside. An overlay is a directory with a kustomization.yaml that refers to other
kustomization directories as its bases. A base has no knowledge of an overlay and can be used in multiple overlays.
An overlay may have multiple bases and it composes all resources
from bases and may also have customization on top of them.
Here is an example of a base:
# Create a directory to hold the basemkdir base
# Create a base/deployment.yamlcat <<EOF > base/deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: my-nginx
spec:
selector:
matchLabels:
run: my-nginx
replicas: 2
template:
metadata:
labels:
run: my-nginx
spec:
containers:
- name: my-nginx
image: nginx
EOF# Create a base/service.yaml filecat <<EOF > base/service.yaml
apiVersion: v1
kind: Service
metadata:
name: my-nginx
labels:
run: my-nginx
spec:
ports:
- port: 80
protocol: TCP
selector:
run: my-nginx
EOF# Create a base/kustomization.yamlcat <<EOF > base/kustomization.yaml
resources:
- deployment.yaml
- service.yaml
EOF
This base can be used in multiple overlays. You can add different namePrefix or other cross-cutting fields
in different overlays. Here are two overlays using the same base.
Use --kustomize or -k in kubectl commands to recognize Resources managed by kustomization.yaml.
Note that -k should point to a kustomization directory, such as
4.5.3 - Managing Kubernetes Objects Using Imperative Commands
Kubernetes objects can quickly be created, updated, and deleted directly using
imperative commands built into the kubectl command-line tool. This document
explains how those commands are organized and how to use them to manage live objects.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The kubectl tool supports three kinds of object management:
Imperative commands
Imperative object configuration
Declarative object configuration
See Kubernetes Object Management
for a discussion of the advantages and disadvantage of each kind of object management.
How to create objects
The kubectl tool supports verb-driven commands for creating some of the most common
object types. The commands are named to be recognizable to users unfamiliar with
the Kubernetes object types.
run: Create a new Pod to run a Container.
expose: Create a new Service object to load balance traffic across Pods.
autoscale: Create a new Autoscaler object to automatically horizontally scale a controller, such as a Deployment.
The kubectl tool also supports creation commands driven by object type.
These commands support more object types and are more explicit about
their intent, but require users to know the type of objects they intend
to create.
create <objecttype> [<subtype>] <instancename>
Some objects types have subtypes that you can specify in the create command.
For example, the Service object has several subtypes including ClusterIP,
LoadBalancer, and NodePort. Here's an example that creates a Service with
subtype NodePort:
kubectl create service nodeport <myservicename>
In the preceding example, the create service nodeport command is called
a subcommand of the create service command.
You can use the -h flag to find the arguments and flags supported by
a subcommand:
kubectl create service nodeport -h
How to update objects
The kubectl command supports verb-driven commands for some common update operations.
These commands are named to enable users unfamiliar with Kubernetes
objects to perform updates without knowing the specific fields
that must be set:
scale: Horizontally scale a controller to add or remove Pods by updating the replica count of the controller.
annotate: Add or remove an annotation from an object.
label: Add or remove a label from an object.
The kubectl command also supports update commands driven by an aspect of the object.
Setting this aspect may set different fields for different object types:
set<field>: Set an aspect of an object.
Note: In Kubernetes version 1.5, not every verb-driven command has an associated aspect-driven command.
The kubectl tool supports these additional ways to update a live object directly,
however they require a better understanding of the Kubernetes object schema.
edit: Directly edit the raw configuration of a live object by opening its configuration in an editor.
patch: Directly modify specific fields of a live object by using a patch string.
For more details on patch strings, see the patch section in
API Conventions.
How to delete objects
You can use the delete command to delete an object from a cluster:
delete <type>/<name>
Note: You can use kubectl delete for both imperative commands and imperative object
configuration. The difference is in the arguments passed to the command. To use
kubectl delete as an imperative command, pass the object to be deleted as
an argument. Here's an example that passes a Deployment object named nginx:
kubectl delete deployment/nginx
How to view an object
There are several commands for printing information about an object:
get: Prints basic information about matching objects. Use get -h to see a list of options.
describe: Prints aggregated detailed information about matching objects.
logs: Prints the stdout and stderr for a container running in a Pod.
Using set commands to modify objects before creation
There are some object fields that don't have a flag you can use
in a create command. In some of those cases, you can use a combination of
set and create to specify a value for the field before object
creation. This is done by piping the output of the create command to the
set command, and then back to the create command. Here's an example:
The kubectl create service -o yaml --dry-run=client command creates the configuration for the Service, but prints it to stdout as YAML instead of sending it to the Kubernetes API server.
The kubectl set selector --local -f - -o yaml command reads the configuration from stdin, and writes the updated configuration to stdout as YAML.
The kubectl create -f - command creates the object using the configuration provided via stdin.
Using --edit to modify objects before creation
You can use kubectl create --edit to make arbitrary changes to an object
before it is created. Here's an example:
4.5.4 - Imperative Management of Kubernetes Objects Using Configuration Files
Kubernetes objects can be created, updated, and deleted by using the kubectl
command-line tool along with an object configuration file written in YAML or JSON.
This document explains how to define and manage objects using configuration files.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The kubectl tool supports three kinds of object management:
Imperative commands
Imperative object configuration
Declarative object configuration
See Kubernetes Object Management
for a discussion of the advantages and disadvantage of each kind of object management.
How to create objects
You can use kubectl create -f to create an object from a configuration file.
Refer to the kubernetes API reference
for details.
kubectl create -f <filename|url>
How to update objects
Warning: Updating objects with the replace command drops all
parts of the spec not specified in the configuration file. This
should not be used with objects whose specs are partially managed
by the cluster, such as Services of type LoadBalancer, where
the externalIPs field is managed independently from the configuration
file. Independently managed fields must be copied to the configuration
file to prevent replace from dropping them.
You can use kubectl replace -f to update a live object according to a
configuration file.
kubectl replace -f <filename|url>
How to delete objects
You can use kubectl delete -f to delete an object that is described in a
configuration file.
kubectl delete -f <filename|url>
Note:
If configuration file has specified the generateName field in the metadata
section instead of the name field, you cannot delete the object using
kubectl delete -f <filename|url>.
You will have to use other flags for deleting the object. For example:
You can use kubectl get -f to view information about an object that is
described in a configuration file.
kubectl get -f <filename|url> -o yaml
The -o yaml flag specifies that the full object configuration is printed.
Use kubectl get -h to see a list of options.
Limitations
The create, replace, and delete commands work well when each object's
configuration is fully defined and recorded in its configuration
file. However when a live object is updated, and the updates are not merged
into its configuration file, the updates will be lost the next time a replace
is executed. This can happen if a controller, such as
a HorizontalPodAutoscaler, makes updates directly to a live object. Here's
an example:
You create an object from a configuration file.
Another source updates the object by changing some field.
You replace the object from the configuration file. Changes made by
the other source in step 2 are lost.
If you need to support multiple writers to the same object, you can use
kubectl apply to manage the object.
Creating and editing an object from a URL without saving the configuration
Suppose you have the URL of an object configuration file. You can use
kubectl create --edit to make changes to the configuration before the
object is created. This is particularly useful for tutorials and tasks
that point to a configuration file that could be modified by the reader.
kubectl create -f <url> --edit
Migrating from imperative commands to imperative object configuration
Migrating from imperative commands to imperative object configuration involves
several manual steps.
Export the live object to a local object configuration file:
kubectl get <kind>/<name> -o yaml > <kind>_<name>.yaml
Manually remove the status field from the object configuration file.
For subsequent object management, use replace exclusively.
kubectl replace -f <kind>_<name>.yaml
Defining controller selectors and PodTemplate labels
Warning: Updating selectors on controllers is strongly discouraged.
The recommended approach is to define a single, immutable PodTemplate label
used only by the controller selector with no other semantic meaning.
4.5.5 - Update API Objects in Place Using kubectl patch
Use kubectl patch to update Kubernetes API objects in place. Do a strategic merge patch or a JSON merge patch.
This task shows how to use kubectl patch to update an API object in place. The exercises
in this task demonstrate a strategic merge patch and a JSON merge patch.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The output shows that the Deployment has two Pods. The 1/1 indicates that
each Pod has one container:
NAME READY STATUS RESTARTS AGE
patch-demo-28633765-670qr 1/1 Running 0 23s
patch-demo-28633765-j5qs3 1/1 Running 0 23s
Make a note of the names of the running Pods. Later, you will see that these Pods
get terminated and replaced by new ones.
At this point, each Pod has one Container that runs the nginx image. Now suppose
you want each Pod to have two containers: one that runs nginx and one that runs redis.
Create a file named patch-file.yaml that has this content:
View the Pods associated with your patched Deployment:
kubectl get pods
The output shows that the running Pods have different names from the Pods that
were running previously. The Deployment terminated the old Pods and created two
new Pods that comply with the updated Deployment spec. The 2/2 indicates that
each Pod has two Containers:
NAME READY STATUS RESTARTS AGE
patch-demo-1081991389-2wrn5 2/2 Running 0 1m
patch-demo-1081991389-jmg7b 2/2 Running 0 1m
Take a closer look at one of the patch-demo Pods:
kubectl get pod <your-pod-name> --output yaml
The output shows that the Pod has two Containers: one running nginx and one running redis:
containers:
- image: redis
...
- image: nginx
...
Notes on the strategic merge patch
The patch you did in the preceding exercise is called a strategic merge patch.
Notice that the patch did not replace the containers list. Instead it added a new
Container to the list. In other words, the list in the patch was merged with the
existing list. This is not always what happens when you use a strategic merge patch on a list.
In some cases, the list is replaced, not merged.
With a strategic merge patch, a list is either replaced or merged depending on its
patch strategy. The patch strategy is specified by the value of the patchStrategy key
in a field tag in the Kubernetes source code. For example, the Containers field of PodSpec
struct has a patchStrategy of merge:
You can also see the patch strategy in the
OpenApi spec:
"io.k8s.api.core.v1.PodSpec": {...,"containers": {"description": "List of containers belonging to the pod. ...."},"x-kubernetes-patch-merge-key": "name","x-kubernetes-patch-strategy": "merge"}
Notice that the tolerations list in the PodSpec was replaced, not merged. This is because
the Tolerations field of PodSpec does not have a patchStrategy key in its field tag. So the
strategic merge patch uses the default patch strategy, which is replace.
type PodSpec struct {
... Tolerations []Toleration `json:"tolerations,omitempty" protobuf:"bytes,22,opt,name=tolerations"`...}
Use a JSON merge patch to update a Deployment
A strategic merge patch is different from a
JSON merge patch.
With a JSON merge patch, if you
want to update a list, you have to specify the entire new list. And the new list completely
replaces the existing list.
The kubectl patch command has a type parameter that you can set to one of these values:
The containers list that you specified in the patch has only one Container.
The output shows that your list of one Container replaced the existing containers list.
In the output, you can see that the existing Pods were terminated, and new Pods
were created. The 1/1 indicates that each new Pod is running only one Container.
NAME READY STATUS RESTARTS AGE
patch-demo-1307768864-69308 1/1 Running 0 1m
patch-demo-1307768864-c86dc 1/1 Running 0 1m
Use strategic merge patch to update a Deployment using the retainKeys strategy
Here's the configuration file for a Deployment that uses the RollingUpdate strategy:
In the output, you can see that it is not possible to set type as Recreate when a value is defined for spec.strategy.rollingUpdate:
The Deployment "retainkeys-demo" is invalid: spec.strategy.rollingUpdate: Forbidden: may not be specified when strategy `type` is 'Recreate'
The way to remove the value for spec.strategy.rollingUpdate when updating the value for type is to use the retainKeys strategy for the strategic merge.
Create another file named patch-file-retainkeys.yaml that has this content:
spec:strategy:$retainKeys:- typetype:Recreate
With this patch, we indicate that we want to retain only the type key of the strategy object. Thus, the rollingUpdate will be removed during the patch operation.
kubectl get deployment retainkeys-demo --output yaml
The output shows that the strategy object in the Deployment does not contain the rollingUpdate key anymore:
spec:strategy:type:Recreatetemplate:
Notes on the strategic merge patch using the retainKeys strategy
The patch you did in the preceding exercise is called a strategic merge patch with retainKeys strategy. This method introduces a new directive $retainKeys that has the following strategies:
It contains a list of strings.
All fields needing to be preserved must be present in the $retainKeys list.
The fields that are present will be merged with live object.
All of the missing fields will be cleared when patching.
All fields in the $retainKeys list must be a superset or the same as the fields present in the patch.
The retainKeys strategy does not work for all objects. It only works when the value of the patchStrategy key in a field tag in the Kubernetes source code contains retainKeys. For example, the Strategy field of the DeploymentSpec struct has a patchStrategy of retainKeys:
You can also see the retainKeys strategy in the OpenApi spec:
"io.k8s.api.apps.v1.DeploymentSpec": {...,"strategy": {"$ref": "#/definitions/io.k8s.api.apps.v1.DeploymentStrategy","description": "The deployment strategy to use to replace existing pods with new ones.","x-kubernetes-patch-strategy": "retainKeys"},....}
In this exercise, you used kubectl patch to change the live configuration
of a Deployment object. You did not change the configuration file that you originally used to
create the Deployment object. Other commands for updating API objects include
kubectl annotate,
kubectl edit,
kubectl replace,
kubectl scale,
and
kubectl apply.
Note: Strategic merge patch is not supported for custom resources.
Managing confidential settings data using Secrets.
4.6.1 - Managing Secrets using kubectl
Creating Secret objects using kubectl command line.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
A Secret can contain user credentials required by pods to access a database.
For example, a database connection string consists of a username and password.
You can store the username in a file ./username.txt and the password in a
file ./password.txt on your local machine.
In these commands, the -n flag ensures that the generated files do not have
an extra newline character at the end of the text. This is important because
when kubectl reads a file and encodes the content into a base64 string, the
extra newline character gets encoded too.
The kubectl create secret command packages these files into a Secret and creates
the object on the API server.
You do not need to escape special characters in password strings that you
include in a file.
You can also provide Secret data using the --from-literal=<key>=<value> tag.
This tag can be specified more than once to provide multiple key-value pairs.
Note that special characters such as $, \, *, =, and ! will be
interpreted by your shell
and require escaping.
In most shells, the easiest way to escape the password is to surround it with
single quotes ('). For example, if your password is S!B\*d$zDsb=,
run the following command:
The commands kubectl get and kubectl describe avoid showing the contents
of a Secret by default. This is to protect the Secret from being exposed
accidentally, or from being stored in a terminal log.
To check the actual content of the encoded data, refer to Decoding the Secret.
Decoding the Secret
To view the contents of the Secret you created, run the following command:
kubectl get secret db-user-pass -o jsonpath='{.data}'
# This is an example for documentation purposes.# If you did things this way, the data 'MWYyZDFlMmU2N2Rm' could be stored in# your shell history.# Someone with access to you computer could find that remembered command# and base-64 decode the secret, perhaps without your knowledge.# It's usually better to combine the steps, as shown later in the page.echo'MWYyZDFlMmU2N2Rm' | base64 --decode
The output is similar to:
1f2d1e2e67df
In order to avoid storing a secret encoded value in your shell history, you can
run the following command:
kubectl get secret db-user-pass -o jsonpath='{.data.password}' | base64 --decode
Creating Secret objects using resource configuration file.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You can define the Secret object in a manifest first, in JSON or YAML format,
and then create that object. The
Secret
resource contains two maps: data and stringData.
The data field is used to store arbitrary data, encoded using base64. The
stringData field is provided for convenience, and it allows you to provide
the same data as unencoded strings.
The keys of data and stringData must consist of alphanumeric characters,
-, _ or ..
The following example stores two strings in a Secret using the data field.
Note: The serialized JSON and YAML values of Secret data are encoded as base64 strings. Newlines are not valid within these strings and must be omitted. When using the base64 utility on Darwin/macOS, users should avoid using the -b option to split long lines. Conversely, Linux users should add the option -w 0 to base64 commands or the pipeline base64 | tr -d '\n' if the -w option is not available.
For certain scenarios, you may wish to use the stringData field instead. This
field allows you to put a non-base64 encoded string directly into the Secret,
and the string will be encoded for you when the Secret is created or updated.
A practical example of this might be where you are deploying an application
that uses a Secret to store a configuration file, and you want to populate
parts of that configuration file during your deployment process.
For example, if your application uses the following configuration file:
Creating Secret objects using kustomization.yaml file.
Since Kubernetes v1.14, kubectl supports
managing objects using Kustomize.
Kustomize provides resource Generators to create Secrets and ConfigMaps. The
Kustomize generators should be specified in a kustomization.yaml file inside
a directory. After generating the Secret, you can create the Secret on the API
server with kubectl apply.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You can generate a Secret by defining a secretGenerator in a
kustomization.yaml file that references other existing files.
For example, the following kustomization file references the
./username.txt and the ./password.txt files:
You can also define the secretGenerator in the kustomization.yaml
file by providing some literals.
For example, the following kustomization.yaml file contains two literals
for username and password respectively:
You can also define the secretGenerator in the kustomization.yaml
file by providing .env files.
For example, the following kustomization.yaml file pulls in data from
.env.secret file:
Note that in all cases, you don't need to base64 encode the values.
Create the Secret
Apply the directory containing the kustomization.yaml to create the Secret.
kubectl apply -k .
The output is similar to:
secret/db-user-pass-96mffmfh4k created
Note that when a Secret is generated, the Secret name is created by hashing
the Secret data and appending the hash value to the name. This ensures that
a new Secret is generated each time the data is modified.
Check the Secret created
You can check that the secret was created:
kubectl get secrets
The output is similar to:
NAME TYPE DATA AGE
db-user-pass-96mffmfh4k Opaque 2 51s
The commands kubectl get and kubectl describe avoid showing the contents of a Secret by
default. This is to protect the Secret from being exposed accidentally to an onlooker,
or from being stored in a terminal log.
To check the actual content of the encoded data, please refer to
decoding secret.
Specify configuration and other data for the Pods that run your workload.
4.7.1 - Define a Command and Arguments for a Container
This page shows how to define commands and arguments when you run a container
in a Pod.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Define a command and arguments when you create a Pod
When you create a Pod, you can define a command and arguments for the
containers that run in the Pod. To define a command, include the command
field in the configuration file. To define arguments for the command, include
the args field in the configuration file. The command and arguments that
you define cannot be changed after the Pod is created.
The command and arguments that you define in the configuration file
override the default command and arguments provided by the container image.
If you define args, but do not define a command, the default command is used
with your new arguments.
Note: The command field corresponds to entrypoint in some container runtimes.
In this exercise, you create a Pod that runs one container. The configuration
file for the Pod defines a command and two arguments:
The output shows that the container that ran in the command-demo Pod has
completed.
To see the output of the command that ran in the container, view the logs
from the Pod:
kubectl logs command-demo
The output shows the values of the HOSTNAME and KUBERNETES_PORT environment
variables:
command-demo
tcp://10.3.240.1:443
Use environment variables to define arguments
In the preceding example, you defined the arguments directly by
providing strings. As an alternative to providing strings directly,
you can define arguments by using environment variables:
This means you can define an argument for a Pod using any of
the techniques available for defining environment variables, including
ConfigMaps
and
Secrets.
Note: The environment variable appears in parentheses, "$(VAR)". This is
required for the variable to be expanded in the command or args field.
Run a command in a shell
In some cases, you need your command to run in a shell. For example, your
command might consist of several commands piped together, or it might be a shell
script. To run your command in a shell, wrap it like this:
command: ["/bin/sh"]args: ["-c", "while true; do echo hello; sleep 10;done"]
This page shows how to define dependent environment variables for a container
in a Kubernetes Pod.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Define an environment dependent variable for a container
When you create a Pod, you can set dependent environment variables for the containers that run in the Pod. To set dependent environment variables, you can use $(VAR_NAME) in the value of env in the configuration file.
In this exercise, you create a Pod that runs one container. The configuration
file for the Pod defines a dependent environment variable with common usage defined. Here is the configuration manifest for the
Pod:
As shown above, you have defined the correct dependency reference of SERVICE_ADDRESS, bad dependency reference of UNCHANGED_REFERENCE and skip dependent references of ESCAPED_REFERENCE.
When an environment variable is already defined when being referenced,
the reference can be correctly resolved, such as in the SERVICE_ADDRESS case.
When the environment variable is undefined or only includes some variables, the undefined environment variable is treated as a normal string, such as UNCHANGED_REFERENCE. Note that incorrectly parsed environment variables, in general, will not block the container from starting.
The $(VAR_NAME) syntax can be escaped with a double $, ie: $$(VAR_NAME).
Escaped references are never expanded, regardless of whether the referenced variable
is defined or not. This can be seen from the ESCAPED_REFERENCE case above.
4.7.3 - Define Environment Variables for a Container
This page shows how to define environment variables for a container
in a Kubernetes Pod.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
When you create a Pod, you can set environment variables for the containers
that run in the Pod. To set environment variables, include the env or
envFrom field in the configuration file.
In this exercise, you create a Pod that runs one container. The configuration
file for the Pod defines an environment variable with name DEMO_GREETING and
value "Hello from the environment". Here is the configuration manifest for the
Pod:
apiVersion:v1kind:Podmetadata:name:envar-demolabels:purpose:demonstrate-envarsspec:containers:- name:envar-demo-containerimage:gcr.io/google-samples/node-hello:1.0env:- name:DEMO_GREETINGvalue:"Hello from the environment"- name:DEMO_FAREWELLvalue:"Such a sweet sorrow"
NAME READY STATUS RESTARTS AGE
envar-demo 1/1 Running 0 9s
List the Pod's container environment variables:
kubectl exec envar-demo -- printenv
The output is similar to this:
NODE_VERSION=4.4.2
EXAMPLE_SERVICE_PORT_8080_TCP_ADDR=10.3.245.237
HOSTNAME=envar-demo
...
DEMO_GREETING=Hello from the environment
DEMO_FAREWELL=Such a sweet sorrow
Note: The environment variables set using the env or envFrom field
override any environment variables specified in the container image.
Note: Environment variables may reference each other, however ordering is important.
Variables making use of others defined in the same context must come later in
the list. Similarly, avoid circular references.
Using environment variables inside of your config
Environment variables that you define in a Pod's configuration can be used
elsewhere in the configuration, for example in commands and arguments that
you set for the Pod's containers.
In the example configuration below, the GREETING, HONORIFIC, and
NAME environment variables are set to Warm greetings to, The Most Honorable, and Kubernetes, respectively. Those environment variables
are then used in the CLI arguments passed to the env-print-demo
container.
4.7.4 - Expose Pod Information to Containers Through Environment Variables
This page shows how a Pod can use environment variables to expose information
about itself to containers running in the Pod, using the downward API.
You can use environment variables to expose Pod fields, container fields, or both.
In Kubernetes, there are two ways to expose Pod and container fields to a running container:
Together, these two ways of exposing Pod and container fields are called the
downward API.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Use Pod fields as values for environment variables
In this part of exercise, you create a Pod that has one container, and you
project Pod-level fields into the running container as environment variables.
In that manifest, you can see five environment variables. The env
field is an array of
environment variable definitions.
The first element in the array specifies that the MY_NODE_NAME environment
variable gets its value from the Pod's spec.nodeName field. Similarly, the
other environment variables get their names from Pod fields.
Note: The fields in this example are Pod fields. They are not fields of the
container in the Pod.
To see why these values are in the log, look at the command and args fields
in the configuration file. When the container starts, it writes the values of
five environment variables to stdout. It repeats this every ten seconds.
Next, get a shell into the container that is running in your Pod:
kubectl exec -it dapi-envars-fieldref -- sh
In your shell, view the environment variables:
# Run this in a shell inside the containerprintenv
The output shows that certain environment variables have been assigned the
values of Pod fields:
Use container fields as values for environment variables
In the preceding exercise, you used information from Pod-level fields as the values
for environment variables.
In this next exercise, you are going to pass fields that are part of the Pod
definition, but taken from the specific
container
rather than from the Pod overall.
Here is a manifest for another Pod that again has just one container:
In this manifest, you can see four environment variables. The env
field is an array of
environment variable definitions.
The first element in the array specifies that the MY_CPU_REQUEST environment
variable gets its value from the requests.cpu field of a container named
test-container. Similarly, the other environment variables get their values
from fields that are specific to this container.
4.7.5 - Expose Pod Information to Containers Through Files
This page shows how a Pod can use a
downwardAPI volume,
to expose information about itself to containers running in the Pod.
A downwardAPI volume can expose Pod fields and container fields.
In Kubernetes, there are two ways to expose Pod and container fields to a running container:
Together, these two ways of exposing Pod and container fields are called the
downward API.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this part of exercise, you create a Pod that has one container, and you
project Pod-level fields into the running container as files.
Here is the manifest for the Pod:
In the manifest, you can see that the Pod has a downwardAPI Volume,
and the container mounts the volume at /etc/podinfo.
Look at the items array under downwardAPI. Each element of the array
defines a downwardAPI volume.
The first element specifies that the value of the Pod's
metadata.labels field should be stored in a file named labels.
The second element specifies that the value of the Pod's annotations
field should be stored in a file named annotations.
Note: The fields in this example are Pod fields. They are not
fields of the container in the Pod.
In the output, you can see that the labels and annotations files
are in a temporary subdirectory: in this example,
..2982_06_02_21_47_53.299460680. In the /etc/podinfo directory, ..data is
a symbolic link to the temporary subdirectory. Also in the /etc/podinfo directory,
labels and annotations are symbolic links.
drwxr-xr-x ... Feb 6 21:47 ..2982_06_02_21_47_53.299460680
lrwxrwxrwx ... Feb 6 21:47 ..data -> ..2982_06_02_21_47_53.299460680
lrwxrwxrwx ... Feb 6 21:47 annotations -> ..data/annotations
lrwxrwxrwx ... Feb 6 21:47 labels -> ..data/labels
/etc/..2982_06_02_21_47_53.299460680:
total 8
-rw-r--r-- ... Feb 6 21:47 annotations
-rw-r--r-- ... Feb 6 21:47 labels
Using symbolic links enables dynamic atomic refresh of the metadata; updates are
written to a new temporary directory, and the ..data symlink is updated
atomically using rename(2).
Note: A container using Downward API as a
subPath volume mount will not
receive Downward API updates.
Exit the shell:
/# exit
Store container fields
The preceding exercise, you made Pod-level fields accessible using the
downward API.
In this next exercise, you are going to pass fields that are part of the Pod
definition, but taken from the specific
container
rather than from the Pod overall. Here is a manifest for a Pod that again has
just one container:
In the manifest, you can see that the Pod has a
downwardAPI volume,
and that the single container in that Pod mounts the volume at /etc/podinfo.
Look at the items array under downwardAPI. Each element of the array
defines a file in the downward API volume.
The first element specifies that in the container named client-container,
the value of the limits.cpu field in the format specified by 1m should be
published as a file named cpu_limit. The divisor field is optional and has the
default value of 1. A divisor of 1 means cores for cpu resources, or
bytes for memory resources.
Get a shell into the container that is running in your Pod:
kubectl exec -it kubernetes-downwardapi-volume-example-2 -- sh
In your shell, view the cpu_limit file:
# Run this in a shell inside the containercat /etc/podinfo/cpu_limit
You can use similar commands to view the cpu_request, mem_limit and
mem_request files.
Project keys to specific paths and file permissions
You can project keys to specific paths and specific permissions on a per-file
basis. For more information, see
Secrets.
What's next
Read the spec
API definition for Pod. This includes the definition of Container (part of Pod).
Read the list of available fields that you
can expose using the downward API.
Read about volumes in the legacy API reference:
Check the Volume
API definition which defines a generic volume in a Pod for containers to access.
Check the DownwardAPIVolumeSource
API definition which defines a volume that contains Downward API information.
Check the DownwardAPIVolumeFile
API definition which contains references to object or resource fields for
populating a file in the Downward API volume.
Check the ResourceFieldSelector
API definition which specifies the container resources and their output format.
4.7.6 - Distribute Credentials Securely Using Secrets
This page shows how to securely inject sensitive data, such as passwords and
encryption keys, into Pods.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Convert your secret data to a base-64 representation
Suppose you want to have two pieces of secret data: a username my-app and a password
39528$vdg7Jb. First, use a base64 encoding tool to convert your username and password to a base64 representation. Here's an example using the commonly available base64 program:
apiVersion:v1kind:Podmetadata:name:secret-test-podspec:containers:- name:test-containerimage:nginxvolumeMounts:# name must match the volume name below- name:secret-volumemountPath:/etc/secret-volume# The secret data is exposed to Containers in the Pod through a Volume.volumes:- name:secret-volumesecret:secretName:test-secret
Use envFrom to define all of the Secret's data as container environment variables. The key from the Secret becomes the environment variable name in the Pod.
Run and manage both stateless and stateful applications.
4.8.1 - Run a Stateless Application Using a Deployment
This page shows how to run an application using a Kubernetes Deployment object.
Objectives
Create an nginx deployment.
Use kubectl to list information about the deployment.
Update the deployment.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.9.
To check the version, enter kubectl version.
Creating and exploring an nginx deployment
You can run an application by creating a Kubernetes Deployment object, and you
can describe a Deployment in a YAML file. For example, this YAML file describes
a Deployment that runs the nginx:1.14.2 Docker image:
apiVersion:apps/v1kind:Deploymentmetadata:name:nginx-deploymentspec:selector:matchLabels:app:nginxreplicas:2# tells deployment to run 2 pods matching the templatetemplate:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80
apiVersion:apps/v1kind:Deploymentmetadata:name:nginx-deploymentspec:selector:matchLabels:app:nginxreplicas:2template:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.16.1# Update the version of nginx from 1.14.2 to 1.16.1ports:- containerPort:80
Watch the deployment create pods with new names and delete the old pods:
kubectl get pods -l app=nginx
Scaling the application by increasing the replica count
You can increase the number of Pods in your Deployment by applying a new YAML
file. This YAML file sets replicas to 4, which specifies that the Deployment
should have four Pods:
apiVersion:apps/v1kind:Deploymentmetadata:name:nginx-deploymentspec:selector:matchLabels:app:nginxreplicas:4# Update the replicas from 2 to 4template:metadata:labels:app:nginxspec:containers:- name:nginximage:nginx:1.14.2ports:- containerPort:80
NAME READY STATUS RESTARTS AGE
nginx-deployment-148880595-4zdqq 1/1 Running 0 25s
nginx-deployment-148880595-6zgi1 1/1 Running 0 25s
nginx-deployment-148880595-fxcez 1/1 Running 0 2m
nginx-deployment-148880595-rwovn 1/1 Running 0 2m
Deleting a deployment
Delete the deployment by name:
kubectl delete deployment nginx-deployment
ReplicationControllers -- the Old Way
The preferred way to create a replicated application is to use a Deployment,
which in turn uses a ReplicaSet. Before the Deployment and ReplicaSet were
added to Kubernetes, replicated applications were configured using a
ReplicationController.
4.8.2 - Run a Single-Instance Stateful Application
This page shows you how to run a single-instance stateful application
in Kubernetes using a PersistentVolume and a Deployment. The
application is MySQL.
Objectives
Create a PersistentVolume referencing a disk in your environment.
Create a MySQL Deployment.
Expose MySQL to other pods in the cluster at a known DNS name.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
You can run a stateful application by creating a Kubernetes Deployment
and connecting it to an existing PersistentVolume using a
PersistentVolumeClaim. For example, this YAML file describes a
Deployment that runs MySQL and references the PersistentVolumeClaim. The file
defines a volume mount for /var/lib/mysql, and then creates a
PersistentVolumeClaim that looks for a 20G volume. This claim is
satisfied by any existing volume that meets the requirements,
or by a dynamic provisioner.
Note: The password is defined in the config yaml, and this is insecure. See
Kubernetes Secrets
for a secure solution.
apiVersion:v1kind:Servicemetadata:name:mysqlspec:ports:- port:3306selector:app:mysqlclusterIP:None---apiVersion:apps/v1kind:Deploymentmetadata:name:mysqlspec:selector:matchLabels:app:mysqlstrategy:type:Recreatetemplate:metadata:labels:app:mysqlspec:containers:- image:mysql:5.6name:mysqlenv:# Use secret in real usage- name:MYSQL_ROOT_PASSWORDvalue:passwordports:- containerPort:3306name:mysqlvolumeMounts:- name:mysql-persistent-storagemountPath:/var/lib/mysqlvolumes:- name:mysql-persistent-storagepersistentVolumeClaim:claimName:mysql-pv-claim
The preceding YAML file creates a service that
allows other Pods in the cluster to access the database. The Service option
clusterIP: None lets the Service DNS name resolve directly to the
Pod's IP address. This is optimal when you have only one Pod
behind a Service and you don't intend to increase the number of Pods.
Run a MySQL client to connect to the server:
kubectl run -it --rm --image=mysql:5.6 --restart=Never mysql-client -- mysql -h mysql -ppassword
This command creates a new Pod in the cluster running a MySQL client
and connects it to the server through the Service. If it connects, you
know your stateful MySQL database is up and running.
Waiting for pod default/mysql-client-274442439-zyp6i to be running, status is Pending, pod ready: false
If you don't see a command prompt, try pressing enter.
mysql>
Updating
The image or any other part of the Deployment can be updated as usual
with the kubectl apply command. Here are some precautions that are
specific to stateful apps:
Don't scale the app. This setup is for single-instance apps
only. The underlying PersistentVolume can only be mounted to one
Pod. For clustered stateful apps, see the
StatefulSet documentation.
Use strategy:type: Recreate in the Deployment configuration
YAML file. This instructs Kubernetes to not use rolling
updates. Rolling updates will not work, as you cannot have more than
one Pod running at a time. The Recreate strategy will stop the
first pod before creating a new one with the updated configuration.
If you manually provisioned a PersistentVolume, you also need to manually
delete it, as well as release the underlying resource.
If you used a dynamic provisioner, it automatically deletes the
PersistentVolume when it sees that you deleted the PersistentVolumeClaim.
Some dynamic provisioners (such as those for EBS and PD) also release the
underlying resource upon deleting the PersistentVolume.
This page shows how to run a replicated stateful application using a
StatefulSet.
This application is a replicated MySQL database. The example topology has a
single primary server and multiple replicas, using asynchronous row-based
replication.
Note:This is not a production configuration. MySQL settings remain on insecure defaults to keep the focus
on general patterns for running stateful applications in Kubernetes.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
apiVersion:v1kind:ConfigMapmetadata:name:mysqllabels:app:mysqlapp.kubernetes.io/name:mysqldata:primary.cnf:| # Apply this config only on the primary.
[mysqld]
log-binreplica.cnf:| # Apply this config only on replicas.
[mysqld]
super-read-only
This ConfigMap provides my.cnf overrides that let you independently control
configuration on the primary MySQL server and its replicas.
In this case, you want the primary server to be able to serve replication logs to replicas
and you want replicas to reject any writes that don't come via replication.
There's nothing special about the ConfigMap itself that causes different
portions to apply to different Pods.
Each Pod decides which portion to look at as it's initializing,
based on information provided by the StatefulSet controller.
Create Services
Create the Services from the following YAML configuration file:
# Headless service for stable DNS entries of StatefulSet members.apiVersion:v1kind:Servicemetadata:name:mysqllabels:app:mysqlapp.kubernetes.io/name:mysqlspec:ports:- name:mysqlport:3306clusterIP:Noneselector:app:mysql---# Client service for connecting to any MySQL instance for reads.# For writes, you must instead connect to the primary: mysql-0.mysql.apiVersion:v1kind:Servicemetadata:name:mysql-readlabels:app:mysqlapp.kubernetes.io/name:mysqlreadonly:"true"spec:ports:- name:mysqlport:3306selector:app:mysql
The headless Service provides a home for the DNS entries that the StatefulSet
controllers creates for each
Pod that's part of the set.
Because the headless Service is named mysql, the Pods are accessible by
resolving <pod-name>.mysql from within any other Pod in the same Kubernetes
cluster and namespace.
The client Service, called mysql-read, is a normal Service with its own
cluster IP that distributes connections across all MySQL Pods that report
being Ready. The set of potential endpoints includes the primary MySQL server and all
replicas.
Note that only read queries can use the load-balanced client Service.
Because there is only one primary MySQL server, clients should connect directly to the
primary MySQL Pod (through its DNS entry within the headless Service) to execute
writes.
Create the StatefulSet
Finally, create the StatefulSet from the following YAML configuration file:
After a while, you should see all 3 Pods become Running:
NAME READY STATUS RESTARTS AGE
mysql-0 2/2 Running 0 2m
mysql-1 2/2 Running 0 1m
mysql-2 2/2 Running 0 1m
Press Ctrl+C to cancel the watch.
Note: If you don't see any progress, make sure you have a dynamic PersistentVolume
provisioner enabled, as mentioned in the prerequisites.
This manifest uses a variety of techniques for managing stateful Pods as part of
a StatefulSet. The next section highlights some of these techniques to explain
what happens as the StatefulSet creates Pods.
Understanding stateful Pod initialization
The StatefulSet controller starts Pods one at a time, in order by their
ordinal index.
It waits until each Pod reports being Ready before starting the next one.
In addition, the controller assigns each Pod a unique, stable name of the form
<statefulset-name>-<ordinal-index>, which results in Pods named mysql-0,
mysql-1, and mysql-2.
The Pod template in the above StatefulSet manifest takes advantage of these
properties to perform orderly startup of MySQL replication.
Generating configuration
Before starting any of the containers in the Pod spec, the Pod first runs any
init containers
in the order defined.
The first init container, named init-mysql, generates special MySQL config
files based on the ordinal index.
The script determines its own ordinal index by extracting it from the end of
the Pod name, which is returned by the hostname command.
Then it saves the ordinal (with a numeric offset to avoid reserved values)
into a file called server-id.cnf in the MySQL conf.d directory.
This translates the unique, stable identity provided by the StatefulSet
into the domain of MySQL server IDs, which require the same properties.
The script in the init-mysql container also applies either primary.cnf or
replica.cnf from the ConfigMap by copying the contents into conf.d.
Because the example topology consists of a single primary MySQL server and any number of
replicas, the script assigns ordinal 0 to be the primary server, and everyone
else to be replicas.
Combined with the StatefulSet controller's
deployment order guarantee,
this ensures the primary MySQL server is Ready before creating replicas, so they can begin
replicating.
Cloning existing data
In general, when a new Pod joins the set as a replica, it must assume the primary MySQL
server might already have data on it. It also must assume that the replication
logs might not go all the way back to the beginning of time.
These conservative assumptions are the key to allow a running StatefulSet
to scale up and down over time, rather than being fixed at its initial size.
The second init container, named clone-mysql, performs a clone operation on
a replica Pod the first time it starts up on an empty PersistentVolume.
That means it copies all existing data from another running Pod,
so its local state is consistent enough to begin replicating from the primary server.
MySQL itself does not provide a mechanism to do this, so the example uses a
popular open-source tool called Percona XtraBackup.
During the clone, the source MySQL server might suffer reduced performance.
To minimize impact on the primary MySQL server, the script instructs each Pod to clone
from the Pod whose ordinal index is one lower.
This works because the StatefulSet controller always ensures Pod N is
Ready before starting Pod N+1.
Starting replication
After the init containers complete successfully, the regular containers run.
The MySQL Pods consist of a mysql container that runs the actual mysqld
server, and an xtrabackup container that acts as a
sidecar.
The xtrabackup sidecar looks at the cloned data files and determines if
it's necessary to initialize MySQL replication on the replica.
If so, it waits for mysqld to be ready and then executes the
CHANGE MASTER TO and START SLAVE commands with replication parameters
extracted from the XtraBackup clone files.
Once a replica begins replication, it remembers its primary MySQL server and
reconnects automatically if the server restarts or the connection dies.
Also, because replicas look for the primary server at its stable DNS name
(mysql-0.mysql), they automatically find the primary server even if it gets a new
Pod IP due to being rescheduled.
Lastly, after starting replication, the xtrabackup container listens for
connections from other Pods requesting a data clone.
This server remains up indefinitely in case the StatefulSet scales up, or in
case the next Pod loses its PersistentVolumeClaim and needs to redo the clone.
Sending client traffic
You can send test queries to the primary MySQL server (hostname mysql-0.mysql)
by running a temporary container with the mysql:5.7 image and running the
mysql client binary.
kubectl run mysql-client --image=mysql:5.7 -i --rm --restart=Never --\
mysql -h mysql-0.mysql <<EOF
CREATE DATABASE test;
CREATE TABLE test.messages (message VARCHAR(250));
INSERT INTO test.messages VALUES ('hello');
EOF
Use the hostname mysql-read to send test queries to any server that reports
being Ready:
kubectl run mysql-client --image=mysql:5.7 -i -t --rm --restart=Never --\
mysql -h mysql-read -e "SELECT * FROM test.messages"
You should get output like this:
Waiting for pod default/mysql-client to be running, status is Pending, pod ready: false
+---------+
| message |
+---------+
| hello |
+---------+
pod "mysql-client" deleted
To demonstrate that the mysql-read Service distributes connections across
servers, you can run SELECT @@server_id in a loop:
kubectl run mysql-client-loop --image=mysql:5.7 -i -t --rm --restart=Never --\
bash -ic "while sleep 1; do mysql -h mysql-read -e 'SELECT @@server_id,NOW()'; done"
You should see the reported @@server_id change randomly, because a different
endpoint might be selected upon each connection attempt:
You can press Ctrl+C when you want to stop the loop, but it's useful to keep
it running in another window so you can see the effects of the following steps.
Simulate Pod and Node failure
To demonstrate the increased availability of reading from the pool of replicas
instead of a single server, keep the SELECT @@server_id loop from above
running while you force a Pod out of the Ready state.
Break the Readiness probe
The readiness probe
for the mysql container runs the command mysql -h 127.0.0.1 -e 'SELECT 1'
to make sure the server is up and able to execute queries.
One way to force this readiness probe to fail is to break that command:
kubectl exec mysql-2 -c mysql -- mv /usr/bin/mysql /usr/bin/mysql.off
This reaches into the actual container's filesystem for Pod mysql-2 and
renames the mysql command so the readiness probe can't find it.
After a few seconds, the Pod should report one of its containers as not Ready,
which you can check by running:
kubectl get pod mysql-2
Look for 1/2 in the READY column:
NAME READY STATUS RESTARTS AGE
mysql-2 1/2 Running 0 3m
At this point, you should see your SELECT @@server_id loop continue to run,
although it never reports 102 anymore.
Recall that the init-mysql script defined server-id as 100 + $ordinal,
so server ID 102 corresponds to Pod mysql-2.
Now repair the Pod and it should reappear in the loop output
after a few seconds:
kubectl exec mysql-2 -c mysql -- mv /usr/bin/mysql.off /usr/bin/mysql
Delete Pods
The StatefulSet also recreates Pods if they're deleted, similar to what a
ReplicaSet does for stateless Pods.
kubectl delete pod mysql-2
The StatefulSet controller notices that no mysql-2 Pod exists anymore,
and creates a new one with the same name and linked to the same
PersistentVolumeClaim.
You should see server ID 102 disappear from the loop output for a while
and then return on its own.
Drain a Node
If your Kubernetes cluster has multiple Nodes, you can simulate Node downtime
(such as when Nodes are upgraded) by issuing a
drain.
First determine which Node one of the MySQL Pods is on:
kubectl get pod mysql-2 -o wide
The Node name should show up in the last column:
NAME READY STATUS RESTARTS AGE IP NODE
mysql-2 2/2 Running 0 15m 10.244.5.27 kubernetes-node-9l2t
Then, drain the Node by running the following command, which cordons it so
no new Pods may schedule there, and then evicts any existing Pods.
Replace <node-name> with the name of the Node you found in the last step.
Caution: Draining a Node can impact other workloads and applications
running on the same node. Only perform the following step in a test
cluster.
# See above advice about impact on other workloadskubectl drain <node-name> --force --delete-emptydir-data --ignore-daemonsets
Now you can watch as the Pod reschedules on a different Node:
And again, you should see server ID 102 disappear from the
SELECT @@server_id loop output for a while and then return.
Now uncordon the Node to return it to a normal state:
kubectl uncordon <node-name>
Scaling the number of replicas
When you use MySQL replication, you can scale your read query capacity by
adding replicas.
For a StatefulSet, you can achieve this with a single command:
kubectl scale statefulset mysql --replicas=5
Watch the new Pods come up by running:
kubectl get pods -l app=mysql --watch
Once they're up, you should see server IDs 103 and 104 start appearing in
the SELECT @@server_id loop output.
You can also verify that these new servers have the data you added before they
existed:
kubectl run mysql-client --image=mysql:5.7 -i -t --rm --restart=Never --\
mysql -h mysql-3.mysql -e "SELECT * FROM test.messages"
Waiting for pod default/mysql-client to be running, status is Pending, pod ready: false
+---------+
| message |
+---------+
| hello |
+---------+
pod "mysql-client" deleted
Scaling back down is also seamless:
kubectl scale statefulset mysql --replicas=3
Note:
Although scaling up creates new PersistentVolumeClaims
automatically, scaling down does not automatically delete these PVCs.
This gives you the choice to keep those initialized PVCs around to make
scaling back up quicker, or to extract data before deleting them.
You can see this by running:
kubectl get pvc -l app=mysql
Which shows that all 5 PVCs still exist, despite having scaled the
StatefulSet down to 3:
Cancel the SELECT @@server_id loop by pressing Ctrl+C in its terminal,
or running the following from another terminal:
kubectl delete pod mysql-client-loop --now
Delete the StatefulSet. This also begins terminating the Pods.
kubectl delete statefulset mysql
Verify that the Pods disappear.
They might take some time to finish terminating.
kubectl get pods -l app=mysql
You'll know the Pods have terminated when the above returns:
No resources found.
Delete the ConfigMap, Services, and PersistentVolumeClaims.
kubectl delete configmap,service,pvc -l app=mysql
If you manually provisioned PersistentVolumes, you also need to manually
delete them, as well as release the underlying resources.
If you used a dynamic provisioner, it automatically deletes the
PersistentVolumes when it sees that you deleted the PersistentVolumeClaims.
Some dynamic provisioners (such as those for EBS and PD) also release the
underlying resources upon deleting the PersistentVolumes.
This task shows how to scale a StatefulSet. Scaling a StatefulSet refers to increasing or decreasing the number of replicas.
Before you begin
StatefulSets are only available in Kubernetes version 1.5 or later.
To check your version of Kubernetes, run kubectl version.
Not all stateful applications scale nicely. If you are unsure about whether to scale your StatefulSets, see StatefulSet concepts or StatefulSet tutorial for further information.
You should perform scaling only when you are confident that your stateful application
cluster is completely healthy.
Scaling StatefulSets
Use kubectl to scale StatefulSets
First, find the StatefulSet you want to scale.
kubectl get statefulsets <stateful-set-name>
Change the number of replicas of your StatefulSet:
You cannot scale down a StatefulSet when any of the stateful Pods it manages is unhealthy. Scaling down only takes place
after those stateful Pods become running and ready.
If spec.replicas > 1, Kubernetes cannot determine the reason for an unhealthy Pod. It might be the result of a permanent fault or of a transient fault. A transient fault can be caused by a restart required by upgrading or maintenance.
If the Pod is unhealthy due to a permanent fault, scaling
without correcting the fault may lead to a state where the StatefulSet membership
drops below a certain minimum number of replicas that are needed to function
correctly. This may cause your StatefulSet to become unavailable.
If the Pod is unhealthy due to a transient fault and the Pod might become available again,
the transient error may interfere with your scale-up or scale-down operation. Some distributed
databases have issues when nodes join and leave at the same time. It is better
to reason about scaling operations at the application level in these cases, and
perform scaling only when you are sure that your stateful application cluster is
completely healthy.
This task assumes you have an application running on your cluster represented by a StatefulSet.
Deleting a StatefulSet
You can delete a StatefulSet in the same way you delete other resources in Kubernetes: use the kubectl delete command, and specify the StatefulSet either by file or by name.
kubectl delete -f <file.yaml>
kubectl delete statefulsets <statefulset-name>
You may need to delete the associated headless service separately after the StatefulSet itself is deleted.
kubectl delete service <service-name>
When deleting a StatefulSet through kubectl, the StatefulSet scales down to 0. All Pods that are part of this workload are also deleted. If you want to delete only the StatefulSet and not the Pods, use --cascade=orphan.
For example:
kubectl delete -f <file.yaml> --cascade=orphan
By passing --cascade=orphan to kubectl delete, the Pods managed by the StatefulSet are left behind even after the StatefulSet object itself is deleted. If the pods have a label app.kubernetes.io/name=MyApp, you can then delete them as follows:
Deleting the Pods in a StatefulSet will not delete the associated volumes. This is to ensure that you have the chance to copy data off the volume before deleting it. Deleting the PVC after the pods have terminated might trigger deletion of the backing Persistent Volumes depending on the storage class and reclaim policy. You should never assume ability to access a volume after claim deletion.
Note: Use caution when deleting a PVC, as it may lead to data loss.
Complete deletion of a StatefulSet
To delete everything in a StatefulSet, including the associated pods, you can run a series of commands similar to the following:
In the example above, the Pods have the label app.kubernetes.io/name=MyApp; substitute your own label as appropriate.
Force deletion of StatefulSet pods
If you find that some pods in your StatefulSet are stuck in the 'Terminating' or 'Unknown' states for an extended period of time, you may need to manually intervene to forcefully delete the pods from the apiserver. This is a potentially dangerous task. Refer to Force Delete StatefulSet Pods for details.
This page shows how to delete Pods which are part of a stateful set, and explains the considerations to keep in mind when doing so.
Before you begin
This is a fairly advanced task and has the potential to violate some of the properties inherent to StatefulSet.
Before proceeding, make yourself familiar with the considerations enumerated below.
StatefulSet considerations
In normal operation of a StatefulSet, there is never a need to force delete a StatefulSet Pod. The StatefulSet controller is responsible for creating, scaling and deleting members of the StatefulSet. It tries to ensure that the specified number of Pods from ordinal 0 through N-1 are alive and ready. StatefulSet ensures that, at any time, there is at most one Pod with a given identity running in a cluster. This is referred to as at most one semantics provided by a StatefulSet.
Manual force deletion should be undertaken with caution, as it has the potential to violate the at most one semantics inherent to StatefulSet. StatefulSets may be used to run distributed and clustered applications which have a need for a stable network identity and stable storage. These applications often have configuration which relies on an ensemble of a fixed number of members with fixed identities. Having multiple members with the same identity can be disastrous and may lead to data loss (e.g. split brain scenario in quorum-based systems).
Delete Pods
You can perform a graceful pod deletion with the following command:
kubectl delete pods <pod>
For the above to lead to graceful termination, the Pod must not specify a
pod.Spec.TerminationGracePeriodSeconds of 0. The practice of setting a
pod.Spec.TerminationGracePeriodSeconds of 0 seconds is unsafe and strongly discouraged
for StatefulSet Pods. Graceful deletion is safe and will ensure that the Pod
shuts down gracefully
before the kubelet deletes the name from the apiserver.
A Pod is not deleted automatically when a node is unreachable.
The Pods running on an unreachable Node enter the 'Terminating' or 'Unknown' state after a
timeout.
Pods may also enter these states when the user attempts graceful deletion of a Pod
on an unreachable Node.
The only ways in which a Pod in such a state can be removed from the apiserver are as follows:
The Node object is deleted (either by you, or by the Node Controller).
The kubelet on the unresponsive Node starts responding, kills the Pod and removes the entry from the apiserver.
Force deletion of the Pod by the user.
The recommended best practice is to use the first or second approach. If a Node is confirmed to be dead (e.g. permanently disconnected from the network, powered down, etc), then delete the Node object. If the Node is suffering from a network partition, then try to resolve this or wait for it to resolve. When the partition heals, the kubelet will complete the deletion of the Pod and free up its name in the apiserver.
Normally, the system completes the deletion once the Pod is no longer running on a Node, or the Node is deleted by an administrator. You may override this by force deleting the Pod.
Force Deletion
Force deletions do not wait for confirmation from the kubelet that the Pod has been terminated. Irrespective of whether a force deletion is successful in killing a Pod, it will immediately free up the name from the apiserver. This would let the StatefulSet controller create a replacement Pod with that same identity; this can lead to the duplication of a still-running Pod, and if said Pod can still communicate with the other members of the StatefulSet, will violate the at most one semantics that StatefulSet is designed to guarantee.
When you force delete a StatefulSet pod, you are asserting that the Pod in question will never again make contact with other Pods in the StatefulSet and its name can be safely freed up for a replacement to be created.
If you want to delete a Pod forcibly using kubectl version >= 1.5, do the following:
In Kubernetes, a HorizontalPodAutoscaler automatically updates a workload resource (such as
a Deployment or
StatefulSet), with the
aim of automatically scaling the workload to match demand.
Horizontal scaling means that the response to increased load is to deploy more
Pods.
This is different from vertical scaling, which for Kubernetes would mean
assigning more resources (for example: memory or CPU) to the Pods that are already
running for the workload.
If the load decreases, and the number of Pods is above the configured minimum,
the HorizontalPodAutoscaler instructs the workload resource (the Deployment, StatefulSet,
or other similar resource) to scale back down.
Horizontal pod autoscaling does not apply to objects that can't be scaled (for example:
a DaemonSet.)
The HorizontalPodAutoscaler is implemented as a Kubernetes API resource and a
controller.
The resource determines the behavior of the controller.
The horizontal pod autoscaling controller, running within the Kubernetes
control plane, periodically adjusts the
desired scale of its target (for example, a Deployment) to match observed metrics such as average
CPU utilization, average memory utilization, or any other custom metric you specify.
HorizontalPodAutoscaler controls the scale of a Deployment and its ReplicaSet
Kubernetes implements horizontal pod autoscaling as a control loop that runs intermittently
(it is not a continuous process). The interval is set by the
--horizontal-pod-autoscaler-sync-period parameter to the
kube-controller-manager
(and the default interval is 15 seconds).
Once during each period, the controller manager queries the resource utilization against the
metrics specified in each HorizontalPodAutoscaler definition. The controller manager
finds the target resource defined by the scaleTargetRef,
then selects the pods based on the target resource's .spec.selector labels, and obtains the metrics from either the resource metrics API (for per-pod resource metrics),
or the custom metrics API (for all other metrics).
For per-pod resource metrics (like CPU), the controller fetches the metrics
from the resource metrics API for each Pod targeted by the HorizontalPodAutoscaler.
Then, if a target utilization value is set, the controller calculates the utilization
value as a percentage of the equivalent
resource request
on the containers in each Pod. If a target raw value is set, the raw metric values are used directly.
The controller then takes the mean of the utilization or the raw value (depending on the type
of target specified) across all targeted Pods, and produces a ratio used to scale
the number of desired replicas.
Please note that if some of the Pod's containers do not have the relevant resource request set,
CPU utilization for the Pod will not be defined and the autoscaler will
not take any action for that metric. See the algorithm details section below
for more information about how the autoscaling algorithm works.
For per-pod custom metrics, the controller functions similarly to per-pod resource metrics,
except that it works with raw values, not utilization values.
For object metrics and external metrics, a single metric is fetched, which describes
the object in question. This metric is compared to the target
value, to produce a ratio as above. In the autoscaling/v2 API
version, this value can optionally be divided by the number of Pods before the
comparison is made.
The common use for HorizontalPodAutoscaler is to configure it to fetch metrics from
aggregated APIs
(metrics.k8s.io, custom.metrics.k8s.io, or external.metrics.k8s.io). The metrics.k8s.io API is
usually provided by an add-on named Metrics Server, which needs to be launched separately.
For more information about resource metrics, see
Metrics Server.
Support for metrics APIs explains the stability guarantees and support status for these
different APIs.
The HorizontalPodAutoscaler controller accesses corresponding workload resources that support scaling (such as Deployments
and StatefulSet). These resources each have a subresource named scale, an interface that allows you to dynamically set the
number of replicas and examine each of their current states.
For general information about subresources in the Kubernetes API, see
Kubernetes API Concepts.
Algorithm details
From the most basic perspective, the HorizontalPodAutoscaler controller
operates on the ratio between desired metric value and current metric
value:
For example, if the current metric value is 200m, and the desired value
is 100m, the number of replicas will be doubled, since 200.0 / 100.0 == 2.0 If the current value is instead 50m, you'll halve the number of
replicas, since 50.0 / 100.0 == 0.5. The control plane skips any scaling
action if the ratio is sufficiently close to 1.0 (within a globally-configurable
tolerance, 0.1 by default).
When a targetAverageValue or targetAverageUtilization is specified,
the currentMetricValue is computed by taking the average of the given
metric across all Pods in the HorizontalPodAutoscaler's scale target.
Before checking the tolerance and deciding on the final values, the control
plane also considers whether any metrics are missing, and how many Pods
are Ready.
All Pods with a deletion timestamp set (objects with a deletion timestamp are
in the process of being shut down / removed) are ignored, and all failed Pods
are discarded.
If a particular Pod is missing metrics, it is set aside for later; Pods
with missing metrics will be used to adjust the final scaling amount.
When scaling on CPU, if any pod has yet to become ready (it's still
initializing, or possibly is unhealthy) or the most recent metric point for
the pod was before it became ready, that pod is set aside as well.
Due to technical constraints, the HorizontalPodAutoscaler controller
cannot exactly determine the first time a pod becomes ready when
determining whether to set aside certain CPU metrics. Instead, it
considers a Pod "not yet ready" if it's unready and transitioned to
unready within a short, configurable window of time since it started.
This value is configured with the --horizontal-pod-autoscaler-initial-readiness-delay flag, and its default is 30
seconds. Once a pod has become ready, it considers any transition to
ready to be the first if it occurred within a longer, configurable time
since it started. This value is configured with the --horizontal-pod-autoscaler-cpu-initialization-period flag, and its
default is 5 minutes.
The currentMetricValue / desiredMetricValue base scale ratio is then
calculated using the remaining pods not set aside or discarded from above.
If there were any missing metrics, the control plane recomputes the average more
conservatively, assuming those pods were consuming 100% of the desired
value in case of a scale down, and 0% in case of a scale up. This dampens
the magnitude of any potential scale.
Furthermore, if any not-yet-ready pods were present, and the workload would have
scaled up without factoring in missing metrics or not-yet-ready pods,
the controller conservatively assumes that the not-yet-ready pods are consuming 0%
of the desired metric, further dampening the magnitude of a scale up.
After factoring in the not-yet-ready pods and missing metrics, the
controller recalculates the usage ratio. If the new ratio reverses the scale
direction, or is within the tolerance, the controller doesn't take any scaling
action. In other cases, the new ratio is used to decide any change to the
number of Pods.
Note that the original value for the average utilization is reported
back via the HorizontalPodAutoscaler status, without factoring in the
not-yet-ready pods or missing metrics, even when the new usage ratio is
used.
If multiple metrics are specified in a HorizontalPodAutoscaler, this
calculation is done for each metric, and then the largest of the desired
replica counts is chosen. If any of these metrics cannot be converted
into a desired replica count (e.g. due to an error fetching the metrics
from the metrics APIs) and a scale down is suggested by the metrics which
can be fetched, scaling is skipped. This means that the HPA is still capable
of scaling up if one or more metrics give a desiredReplicas greater than
the current value.
Finally, right before HPA scales the target, the scale recommendation is recorded. The
controller considers all recommendations within a configurable window choosing the
highest recommendation from within that window. This value can be configured using the --horizontal-pod-autoscaler-downscale-stabilization flag, which defaults to 5 minutes.
This means that scaledowns will occur gradually, smoothing out the impact of rapidly
fluctuating metric values.
API Object
The Horizontal Pod Autoscaler is an API resource in the Kubernetes
autoscaling API group. The current stable version can be found in
the autoscaling/v2 API version which includes support for scaling on
memory and custom metrics. The new fields introduced in
autoscaling/v2 are preserved as annotations when working with
autoscaling/v1.
When you create a HorizontalPodAutoscaler API object, make sure the name specified is a valid
DNS subdomain name.
More details about the API object can be found at
HorizontalPodAutoscaler Object.
Stability of workload scale
When managing the scale of a group of replicas using the HorizontalPodAutoscaler,
it is possible that the number of replicas keeps fluctuating frequently due to the
dynamic nature of the metrics evaluated. This is sometimes referred to as thrashing,
or flapping. It's similar to the concept of hysteresis in cybernetics.
Autoscaling during rolling update
Kubernetes lets you perform a rolling update on a Deployment. In that
case, the Deployment manages the underlying ReplicaSets for you.
When you configure autoscaling for a Deployment, you bind a
HorizontalPodAutoscaler to a single Deployment. The HorizontalPodAutoscaler
manages the replicas field of the Deployment. The deployment controller is responsible
for setting the replicas of the underlying ReplicaSets so that they add up to a suitable
number during the rollout and also afterwards.
If you perform a rolling update of a StatefulSet that has an autoscaled number of
replicas, the StatefulSet directly manages its set of Pods (there is no intermediate resource
similar to ReplicaSet).
Support for resource metrics
Any HPA target can be scaled based on the resource usage of the pods in the scaling target.
When defining the pod specification the resource requests like cpu and memory should
be specified. This is used to determine the resource utilization and used by the HPA controller
to scale the target up or down. To use resource utilization based scaling specify a metric source
like this:
With this metric the HPA controller will keep the average utilization of the pods in the scaling
target at 60%. Utilization is the ratio between the current usage of resource to the requested
resources of the pod. See Algorithm for more details about how the utilization
is calculated and averaged.
Note: Since the resource usages of all the containers are summed up the total pod utilization may not
accurately represent the individual container resource usage. This could lead to situations where
a single container might be running with high usage and the HPA will not scale out because the overall
pod usage is still within acceptable limits.
Container resource metrics
FEATURE STATE:Kubernetes v1.20 [alpha]
The HorizontalPodAutoscaler API also supports a container metric source where the HPA can track the
resource usage of individual containers across a set of Pods, in order to scale the target resource.
This lets you configure scaling thresholds for the containers that matter most in a particular Pod.
For example, if you have a web application and a logging sidecar, you can scale based on the resource
use of the web application, ignoring the sidecar container and its resource use.
If you revise the target resource to have a new Pod specification with a different set of containers,
you should revise the HPA spec if that newly added container should also be used for
scaling. If the specified container in the metric source is not present or only present in a subset
of the pods then those pods are ignored and the recommendation is recalculated. See Algorithm
for more details about the calculation. To use container resources for autoscaling define a metric
source as follows:
In the above example the HPA controller scales the target such that the average utilization of the cpu
in the application container of all the pods is 60%.
Note:
If you change the name of a container that a HorizontalPodAutoscaler is tracking, you can
make that change in a specific order to ensure scaling remains available and effective
whilst the change is being applied. Before you update the resource that defines the container
(such as a Deployment), you should update the associated HPA to track both the new and
old container names. This way, the HPA is able to calculate a scaling recommendation
throughout the update process.
Once you have rolled out the container name change to the workload resource, tidy up by removing
the old container name from the HPA specification.
Scaling on custom metrics
FEATURE STATE:Kubernetes v1.23 [stable]
(the autoscaling/v2beta2 API version previously provided this ability as a beta feature)
Provided that you use the autoscaling/v2 API version, you can configure a HorizontalPodAutoscaler
to scale based on a custom metric (that is not built in to Kubernetes or any Kubernetes component).
The HorizontalPodAutoscaler controller then queries for these custom metrics from the Kubernetes
API.
(the autoscaling/v2beta2 API version previously provided this ability as a beta feature)
Provided that you use the autoscaling/v2 API version, you can specify multiple metrics for a
HorizontalPodAutoscaler to scale on. Then, the HorizontalPodAutoscaler controller evaluates each metric,
and proposes a new scale based on that metric. The HorizontalPodAutoscaler takes the maximum scale
recommended for each metric and sets the workload to that size (provided that this isn't larger than the
overall maximum that you configured).
Support for metrics APIs
By default, the HorizontalPodAutoscaler controller retrieves metrics from a series of APIs. In order for it to access these
APIs, cluster administrators must ensure that:
For resource metrics, this is the metrics.k8s.io API, generally provided by metrics-server.
It can be launched as a cluster add-on.
For custom metrics, this is the custom.metrics.k8s.io API. It's provided by "adapter" API servers provided by metrics solution vendors.
Check with your metrics pipeline to see if there is a Kubernetes metrics adapter available.
For external metrics, this is the external.metrics.k8s.io API. It may be provided by the custom metrics adapters provided above.
(the autoscaling/v2beta2 API version previously provided this ability as a beta feature)
If you use the v2 HorizontalPodAutoscaler API, you can use the behavior field
(see the API reference)
to configure separate scale-up and scale-down behaviors.
You specify these behaviours by setting scaleUp and / or scaleDown
under the behavior field.
You can specify a stabilization window that prevents flapping
the replica count for a scaling target. Scaling policies also let you controls the
rate of change of replicas while scaling.
Scaling policies
One or more scaling policies can be specified in the behavior section of the spec.
When multiple policies are specified the policy which allows the highest amount of
change is the policy which is selected by default. The following example shows this behavior
while scaling down:
periodSeconds indicates the length of time in the past for which the policy must hold true.
The first policy (Pods) allows at most 4 replicas to be scaled down in one minute. The second policy
(Percent) allows at most 10% of the current replicas to be scaled down in one minute.
Since by default the policy which allows the highest amount of change is selected, the second policy will
only be used when the number of pod replicas is more than 40. With 40 or less replicas, the first policy will be applied.
For instance if there are 80 replicas and the target has to be scaled down to 10 replicas
then during the first step 8 replicas will be reduced. In the next iteration when the number
of replicas is 72, 10% of the pods is 7.2 but the number is rounded up to 8. On each loop of
the autoscaler controller the number of pods to be change is re-calculated based on the number
of current replicas. When the number of replicas falls below 40 the first policy (Pods) is applied
and 4 replicas will be reduced at a time.
The policy selection can be changed by specifying the selectPolicy field for a scaling
direction. By setting the value to Min which would select the policy which allows the
smallest change in the replica count. Setting the value to Disabled completely disables
scaling in that direction.
Stabilization window
The stabilization window is used to restrict the flapping of
replica count when the metrics used for scaling keep fluctuating. The autoscaling algorithm
uses this window to infer a previous desired state and avoid unwanted changes to workload
scale.
For example, in the following example snippet, a stabilization window is specified for scaleDown.
behavior:scaleDown:stabilizationWindowSeconds:300
When the metrics indicate that the target should be scaled down the algorithm looks
into previously computed desired states, and uses the highest value from the specified
interval. In the above example, all desired states from the past 5 minutes will be considered.
This approximates a rolling maximum, and avoids having the scaling algorithm frequently
remove Pods only to trigger recreating an equivalent Pod just moments later.
Default Behavior
To use the custom scaling not all fields have to be specified. Only values which need to be
customized can be specified. These custom values are merged with default values. The default values
match the existing behavior in the HPA algorithm.
For scaling down the stabilization window is 300 seconds (or the value of the
--horizontal-pod-autoscaler-downscale-stabilization flag if provided). There is only a single policy
for scaling down which allows a 100% of the currently running replicas to be removed which
means the scaling target can be scaled down to the minimum allowed replicas.
For scaling up there is no stabilization window. When the metrics indicate that the target should be
scaled up the target is scaled up immediately. There are 2 policies where 4 pods or a 100% of the currently
running replicas will be added every 15 seconds till the HPA reaches its steady state.
Example: change downscale stabilization window
To provide a custom downscale stabilization window of 1 minute, the following
behavior would be added to the HPA:
behavior:scaleDown:stabilizationWindowSeconds:60
Example: limit scale down rate
To limit the rate at which pods are removed by the HPA to 10% per minute, the
following behavior would be added to the HPA:
To ensure that no more than 5 Pods are removed per minute, you can add a second scale-down
policy with a fixed size of 5, and set selectPolicy to minimum. Setting selectPolicy to Min means
that the autoscaler chooses the policy that affects the smallest number of Pods:
The selectPolicy value of Disabled turns off scaling the given direction.
So to prevent downscaling the following policy would be used:
behavior:scaleDown:selectPolicy:Disabled
Support for HorizontalPodAutoscaler in kubectl
HorizontalPodAutoscaler, like every API resource, is supported in a standard way by kubectl.
You can create a new autoscaler using kubectl create command.
You can list autoscalers by kubectl get hpa or get detailed description by kubectl describe hpa.
Finally, you can delete an autoscaler using kubectl delete hpa.
In addition, there is a special kubectl autoscale command for creating a HorizontalPodAutoscaler object.
For instance, executing kubectl autoscale rs foo --min=2 --max=5 --cpu-percent=80
will create an autoscaler for ReplicaSet foo, with target CPU utilization set to 80%
and the number of replicas between 2 and 5.
Implicit maintenance-mode deactivation
You can implicitly deactivate the HPA for a target without the
need to change the HPA configuration itself. If the target's desired replica count
is set to 0, and the HPA's minimum replica count is greater than 0, the HPA
stops adjusting the target (and sets the ScalingActive Condition on itself
to false) until you reactivate it by manually adjusting the target's desired
replica count or HPA's minimum replica count.
Migrating Deployments and StatefulSets to horizontal autoscaling
When an HPA is enabled, it is recommended that the value of spec.replicas of
the Deployment and / or StatefulSet be removed from their
manifest(s). If this isn't done, any time
a change to that object is applied, for example via kubectl apply -f deployment.yaml, this will instruct Kubernetes to scale the current number of Pods
to the value of the spec.replicas key. This may not be
desired and could be troublesome when an HPA is active.
Keep in mind that the removal of spec.replicas may incur a one-time
degradation of Pod counts as the default value of this key is 1 (reference
Deployment Replicas).
Upon the update, all Pods except 1 will begin their termination procedures. Any
deployment application afterwards will behave as normal and respect a rolling
update configuration as desired. You can avoid this degradation by choosing one of the following two
methods based on how you are modifying your deployments:
In the editor, remove spec.replicas. When you save and exit the editor, kubectl
applies the update. No changes to Pod counts happen at this step.
You can now remove spec.replicas from the manifest. If you use source code management,
also commit your changes or take whatever other steps for revising the source code
are appropriate for how you track updates.
From here on out you can run kubectl apply -f deployment.yaml
A HorizontalPodAutoscaler
(HPA for short)
automatically updates a workload resource (such as
a Deployment or
StatefulSet), with the
aim of automatically scaling the workload to match demand.
Horizontal scaling means that the response to increased load is to deploy more
Pods.
This is different from vertical scaling, which for Kubernetes would mean
assigning more resources (for example: memory or CPU) to the Pods that are already
running for the workload.
If the load decreases, and the number of Pods is above the configured minimum,
the HorizontalPodAutoscaler instructs the workload resource (the Deployment, StatefulSet,
or other similar resource) to scale back down.
This document walks you through an example of enabling HorizontalPodAutoscaler to
automatically manage scale for an example web app. This example workload is Apache
httpd running some PHP code.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version 1.23.
To check the version, enter kubectl version.
If you're running an older
release of Kubernetes, refer to the version of the documentation for that release (see
available documentation versions).
To follow this walkthrough, you also need to use a cluster that has a
Metrics Server deployed and configured.
The Kubernetes Metrics Server collects resource metrics from
the kubelets in your cluster, and exposes those metrics
through the Kubernetes API,
using an APIService to add
new kinds of resource that represent metric readings.
To demonstrate a HorizontalPodAutoscaler, you will first start a Deployment that runs a container using the
hpa-example image, and expose it as a Service
using the following manifest:
deployment.apps/php-apache created
service/php-apache created
Create the HorizontalPodAutoscaler
Now that the server is running, create the autoscaler using kubectl. There is
kubectl autoscale subcommand,
part of kubectl, that helps you do this.
You will shortly run a command that creates a HorizontalPodAutoscaler that maintains
between 1 and 10 replicas of the Pods controlled by the php-apache Deployment that
you created in the first step of these instructions.
Roughly speaking, the HPA controller will increase and decrease
the number of replicas (by updating the Deployment) to maintain an average CPU utilization across all Pods of 50%.
The Deployment then updates the ReplicaSet - this is part of how all Deployments work in Kubernetes -
and then the ReplicaSet either adds or removes Pods based on the change to its .spec.
Since each pod requests 200 milli-cores by kubectl run, this means an average CPU usage of 100 milli-cores.
See Algorithm details for more details
on the algorithm.
You can check the current status of the newly-made HorizontalPodAutoscaler, by running:
# You can use "hpa" or "horizontalpodautoscaler"; either name works OK.kubectl get hpa
The output is similar to:
NAME REFERENCE TARGET MINPODS MAXPODS REPLICAS AGE
php-apache Deployment/php-apache/scale 0% / 50% 1 10 1 18s
(if you see other HorizontalPodAutoscalers with different names, that means they already existed,
and isn't usually a problem).
Please note that the current CPU consumption is 0% as there are no clients sending requests to the server
(the TARGET column shows the average across all the Pods controlled by the corresponding deployment).
Increase the load
Next, see how the autoscaler reacts to increased load.
To do this, you'll start a different Pod to act as a client. The container within the client Pod
runs in an infinite loop, sending queries to the php-apache service.
# Run this in a separate terminal# so that the load generation continues and you can carry on with the rest of the stepskubectl run -i --tty load-generator --rm --image=busybox:1.28 --restart=Never -- /bin/sh -c "while sleep 0.01; do wget -q -O- http://php-apache; done"
Now run:
# type Ctrl+C to end the watch when you're readykubectl get hpa php-apache --watch
Within a minute or so, you should see the higher CPU load; for example:
NAME REFERENCE TARGET MINPODS MAXPODS REPLICAS AGE
php-apache Deployment/php-apache/scale 305% / 50% 1 10 1 3m
and then, more replicas. For example:
NAME REFERENCE TARGET MINPODS MAXPODS REPLICAS AGE
php-apache Deployment/php-apache/scale 305% / 50% 1 10 7 3m
Here, CPU consumption has increased to 305% of the request.
As a result, the Deployment was resized to 7 replicas:
kubectl get deployment php-apache
You should see the replica count matching the figure from the HorizontalPodAutoscaler
NAME READY UP-TO-DATE AVAILABLE AGE
php-apache 7/7 7 7 19m
Note: It may take a few minutes to stabilize the number of replicas. Since the amount
of load is not controlled in any way it may happen that the final number of replicas
will differ from this example.
Stop generating load
To finish the example, stop sending the load.
In the terminal where you created the Pod that runs a busybox image, terminate
the load generation by typing <Ctrl> + C.
Then verify the result state (after a minute or so):
# type Ctrl+C to end the watch when you're readykubectl get hpa php-apache --watch
The output is similar to:
NAME REFERENCE TARGET MINPODS MAXPODS REPLICAS AGE
php-apache Deployment/php-apache/scale 0% / 50% 1 10 1 11m
and the Deployment also shows that it has scaled down:
kubectl get deployment php-apache
NAME READY UP-TO-DATE AVAILABLE AGE
php-apache 1/1 1 1 27m
Once CPU utilization dropped to 0, the HPA automatically scaled the number of replicas back down to 1.
Autoscaling the replicas may take a few minutes.
Autoscaling on multiple metrics and custom metrics
You can introduce additional metrics to use when autoscaling the php-apache Deployment
by making use of the autoscaling/v2 API version.
First, get the YAML of your HorizontalPodAutoscaler in the autoscaling/v2 form:
kubectl get hpa php-apache -o yaml > /tmp/hpa-v2.yaml
Open the /tmp/hpa-v2.yaml file in an editor, and you should see YAML which looks like this:
Notice that the targetCPUUtilizationPercentage field has been replaced with an array called metrics.
The CPU utilization metric is a resource metric, since it is represented as a percentage of a resource
specified on pod containers. Notice that you can specify other resource metrics besides CPU. By default,
the only other supported resource metric is memory. These resources do not change names from cluster
to cluster, and should always be available, as long as the metrics.k8s.io API is available.
You can also specify resource metrics in terms of direct values, instead of as percentages of the
requested value, by using a target.type of AverageValue instead of Utilization, and
setting the corresponding target.averageValue field instead of the target.averageUtilization.
There are two other types of metrics, both of which are considered custom metrics: pod metrics and
object metrics. These metrics may have names which are cluster specific, and require a more
advanced cluster monitoring setup.
The first of these alternative metric types is pod metrics. These metrics describe Pods, and
are averaged together across Pods and compared with a target value to determine the replica count.
They work much like resource metrics, except that they only support a target type of AverageValue.
Pod metrics are specified using a metric block like this:
The second alternative metric type is object metrics. These metrics describe a different
object in the same namespace, instead of describing Pods. The metrics are not necessarily
fetched from the object; they only describe it. Object metrics support target types of
both Value and AverageValue. With Value, the target is compared directly to the returned
metric from the API. With AverageValue, the value returned from the custom metrics API is divided
by the number of Pods before being compared to the target. The following example is the YAML
representation of the requests-per-second metric.
If you provide multiple such metric blocks, the HorizontalPodAutoscaler will consider each metric in turn.
The HorizontalPodAutoscaler will calculate proposed replica counts for each metric, and then choose the
one with the highest replica count.
For example, if you had your monitoring system collecting metrics about network traffic,
you could update the definition above using kubectl edit to look like this:
Then, your HorizontalPodAutoscaler would attempt to ensure that each pod was consuming roughly
50% of its requested CPU, serving 1000 packets per second, and that all pods behind the main-route
Ingress were serving a total of 10000 requests per second.
Autoscaling on more specific metrics
Many metrics pipelines allow you to describe metrics either by name or by a set of additional
descriptors called labels. For all non-resource metric types (pod, object, and external,
described below), you can specify an additional label selector which is passed to your metric
pipeline. For instance, if you collect a metric http_requests with the verb
label, you can specify the following metric block to scale only on GET requests:
This selector uses the same syntax as the full Kubernetes label selectors. The monitoring pipeline
determines how to collapse multiple series into a single value, if the name and selector
match multiple series. The selector is additive, and cannot select metrics
that describe objects that are not the target object (the target pods in the case of the Pods
type, and the described object in the case of the Object type).
Autoscaling on metrics not related to Kubernetes objects
Applications running on Kubernetes may need to autoscale based on metrics that don't have an obvious
relationship to any object in the Kubernetes cluster, such as metrics describing a hosted service with
no direct correlation to Kubernetes namespaces. In Kubernetes 1.10 and later, you can address this use case
with external metrics.
Using external metrics requires knowledge of your monitoring system; the setup is
similar to that required when using custom metrics. External metrics allow you to autoscale your cluster
based on any metric available in your monitoring system. Provide a metric block with a
name and selector, as above, and use the External metric type instead of Object.
If multiple time series are matched by the metricSelector,
the sum of their values is used by the HorizontalPodAutoscaler.
External metrics support both the Value and AverageValue target types, which function exactly the same
as when you use the Object type.
For example if your application processes tasks from a hosted queue service, you could add the following
section to your HorizontalPodAutoscaler manifest to specify that you need one worker per 30 outstanding tasks.
When possible, it's preferable to use the custom metric target types instead of external metrics, since it's
easier for cluster administrators to secure the custom metrics API. The external metrics API potentially allows
access to any metric, so cluster administrators should take care when exposing it.
Appendix: Horizontal Pod Autoscaler Status Conditions
When using the autoscaling/v2 form of the HorizontalPodAutoscaler, you will be able to see
status conditions set by Kubernetes on the HorizontalPodAutoscaler. These status conditions indicate
whether or not the HorizontalPodAutoscaler is able to scale, and whether or not it is currently restricted
in any way.
The conditions appear in the status.conditions field. To see the conditions affecting a HorizontalPodAutoscaler,
we can use kubectl describe hpa:
kubectl describe hpa cm-test
Name: cm-test
Namespace: prom
Labels: <none>
Annotations: <none>
CreationTimestamp: Fri, 16 Jun 2017 18:09:22 +0000
Reference: ReplicationController/cm-test
Metrics: ( current / target )
"http_requests" on pods: 66m / 500m
Min replicas: 1
Max replicas: 4
ReplicationController pods: 1 current / 1 desired
Conditions:
Type Status Reason Message
---- ------ ------ -------
AbleToScale True ReadyForNewScale the last scale time was sufficiently old as to warrant a new scale
ScalingActive True ValidMetricFound the HPA was able to successfully calculate a replica count from pods metric http_requests
ScalingLimited False DesiredWithinRange the desired replica count is within the acceptable range
Events:
For this HorizontalPodAutoscaler, you can see several conditions in a healthy state. The first,
AbleToScale, indicates whether or not the HPA is able to fetch and update scales, as well as
whether or not any backoff-related conditions would prevent scaling. The second, ScalingActive,
indicates whether or not the HPA is enabled (i.e. the replica count of the target is not zero) and
is able to calculate desired scales. When it is False, it generally indicates problems with
fetching metrics. Finally, the last condition, ScalingLimited, indicates that the desired scale
was capped by the maximum or minimum of the HorizontalPodAutoscaler. This is an indication that
you may wish to raise or lower the minimum or maximum replica count constraints on your
HorizontalPodAutoscaler.
Quantities
All metrics in the HorizontalPodAutoscaler and metrics APIs are specified using
a special whole-number notation known in Kubernetes as a
quantity. For example,
the quantity 10500m would be written as 10.5 in decimal notation. The metrics APIs
will return whole numbers without a suffix when possible, and will generally return
quantities in milli-units otherwise. This means you might see your metric value fluctuate
between 1 and 1500m, or 1 and 1.5 when written in decimal notation.
Other possible scenarios
Creating the autoscaler declaratively
Instead of using kubectl autoscale command to create a HorizontalPodAutoscaler imperatively we
can use the following manifest to create it declaratively:
horizontalpodautoscaler.autoscaling/php-apache created
4.8.9 - Specifying a Disruption Budget for your Application
FEATURE STATE:Kubernetes v1.21 [stable]
This page shows how to limit the number of concurrent disruptions
that your application experiences, allowing for higher availability
while permitting the cluster administrator to manage the clusters
nodes.
Before you begin
Your Kubernetes server must be at or later than version v1.21.
To check the version, enter kubectl version.
You are the owner of an application running on a Kubernetes cluster that requires
high availability.
You should confirm with your cluster owner or service provider that they respect
Pod Disruption Budgets.
Protecting an Application with a PodDisruptionBudget
Identify what application you want to protect with a PodDisruptionBudget (PDB).
Think about how your application reacts to disruptions.
Create a PDB definition as a YAML file.
Create the PDB object from the YAML file.
Identify an Application to Protect
The most common use case when you want to protect an application
specified by one of the built-in Kubernetes controllers:
Deployment
ReplicationController
ReplicaSet
StatefulSet
In this case, make a note of the controller's .spec.selector; the same
selector goes into the PDBs .spec.selector.
From version 1.15 PDBs support custom controllers where the scale subresource is enabled.
You can also use PDBs with pods which are not controlled by one of the above
controllers, or arbitrary groups of pods, but there are some restrictions,
described in Arbitrary Controllers and Selectors.
Think about how your application reacts to disruptions
Decide how many instances can be down at the same time for a short period
due to a voluntary disruption.
Stateless frontends:
Concern: don't reduce serving capacity by more than 10%.
Solution: use PDB with minAvailable 90% for example.
Single-instance Stateful Application:
Concern: do not terminate this application without talking to me.
Possible Solution 1: Do not use a PDB and tolerate occasional downtime.
Possible Solution 2: Set PDB with maxUnavailable=0. Have an understanding
(outside of Kubernetes) that the cluster operator needs to consult you before
termination. When the cluster operator contacts you, prepare for downtime,
and then delete the PDB to indicate readiness for disruption. Recreate afterwards.
Multiple-instance Stateful application such as Consul, ZooKeeper, or etcd:
Concern: Do not reduce number of instances below quorum, otherwise writes fail.
Possible Solution 1: set maxUnavailable to 1 (works with varying scale of application).
Possible Solution 2: set minAvailable to quorum-size (e.g. 3 when scale is 5). (Allows more disruptions at once).
Restartable Batch Job:
Concern: Job needs to complete in case of voluntary disruption.
Possible solution: Do not create a PDB. The Job controller will create a replacement pod.
Rounding logic when specifying percentages
Values for minAvailable or maxUnavailable can be expressed as integers or as a percentage.
When you specify an integer, it represents a number of Pods. For instance, if you set minAvailable to 10, then 10
Pods must always be available, even during a disruption.
When you specify a percentage by setting the value to a string representation of a percentage (eg. "50%"), it represents a percentage of
total Pods. For instance, if you set maxUnavailable to "50%", then only 50% of the Pods can be unavailable during a
disruption.
When you specify the value as a percentage, it may not map to an exact number of Pods. For example, if you have 7 Pods and
you set minAvailable to "50%", it's not immediately obvious whether that means 3 Pods or 4 Pods must be available.
Kubernetes rounds up to the nearest integer, so in this case, 4 Pods must be available. You can examine the
code
that controls this behavior.
Specifying a PodDisruptionBudget
A PodDisruptionBudget has three fields:
A label selector .spec.selector to specify the set of
pods to which it applies. This field is required.
.spec.minAvailable which is a description of the number of pods from that
set that must still be available after the eviction, even in the absence
of the evicted pod. minAvailable can be either an absolute number or a percentage.
.spec.maxUnavailable (available in Kubernetes 1.7 and higher) which is a description
of the number of pods from that set that can be unavailable after the eviction.
It can be either an absolute number or a percentage.
Note: The behavior for an empty selector differs between the policy/v1beta1 and policy/v1 APIs for
PodDisruptionBudgets. For policy/v1beta1 an empty selector matches zero pods, while
for policy/v1 an empty selector matches every pod in the namespace.
You can specify only one of maxUnavailable and minAvailable in a single PodDisruptionBudget.
maxUnavailable can only be used to control the eviction of pods
that have an associated controller managing them. In the examples below, "desired replicas"
is the scale of the controller managing the pods being selected by the
PodDisruptionBudget.
Example 1: With a minAvailable of 5, evictions are allowed as long as they leave behind
5 or more healthy pods among those selected by the PodDisruptionBudget's selector.
Example 2: With a minAvailable of 30%, evictions are allowed as long as at least 30%
of the number of desired replicas are healthy.
Example 3: With a maxUnavailable of 5, evictions are allowed as long as there are at most 5
unhealthy replicas among the total number of desired replicas.
Example 4: With a maxUnavailable of 30%, evictions are allowed as long as no more than 30%
of the desired replicas are unhealthy.
In typical usage, a single budget would be used for a collection of pods managed by
a controller—for example, the pods in a single ReplicaSet or StatefulSet.
Note: A disruption budget does not truly guarantee that the specified
number/percentage of pods will always be up. For example, a node that hosts a
pod from the collection may fail when the collection is at the minimum size
specified in the budget, thus bringing the number of available pods from the
collection below the specified size. The budget can only protect against
voluntary evictions, not all causes of unavailability.
If you set maxUnavailable to 0% or 0, or you set minAvailable to 100% or the number of replicas,
you are requiring zero voluntary evictions. When you set zero voluntary evictions for a workload
object such as ReplicaSet, then you cannot successfully drain a Node running one of those Pods.
If you try to drain a Node where an unevictable Pod is running, the drain never completes. This is permitted as per the
semantics of PodDisruptionBudget.
You can find examples of pod disruption budgets defined below. They match pods with the label
app: zookeeper.
For example, if the above zk-pdb object selects the pods of a StatefulSet of size 3, both
specifications have the exact same meaning. The use of maxUnavailable is recommended as it
automatically responds to changes in the number of replicas of the corresponding controller.
Create the PDB object
You can create or update the PDB object using kubectl.
kubectl apply -f mypdb.yaml
Check the status of the PDB
Use kubectl to check that your PDB is created.
Assuming you don't actually have pods matching app: zookeeper in your namespace,
then you'll see something like this:
kubectl get poddisruptionbudgets
NAME MIN AVAILABLE MAX UNAVAILABLE ALLOWED DISRUPTIONS AGE
zk-pdb 2 N/A 0 7s
If there are matching pods (say, 3), then you would see something like this:
kubectl get poddisruptionbudgets
NAME MIN AVAILABLE MAX UNAVAILABLE ALLOWED DISRUPTIONS AGE
zk-pdb 2 N/A 1 7s
The non-zero value for ALLOWED DISRUPTIONS means that the disruption controller has seen the pods,
counted the matching pods, and updated the status of the PDB.
You can get more information about the status of a PDB with this command:
You can skip this section if you only use PDBs with the built-in
application controllers (Deployment, ReplicationController, ReplicaSet, and StatefulSet),
with the PDB selector matching the controller's selector.
You can use a PDB with pods controlled by another type of controller, by an
"operator", or bare pods, but with these restrictions:
only .spec.minAvailable can be used, not .spec.maxUnavailable.
only an integer value can be used with .spec.minAvailable, not a percentage.
You can use a selector which selects a subset or superset of the pods belonging to a built-in
controller. The eviction API will disallow eviction of any pod covered by multiple PDBs,
so most users will want to avoid overlapping selectors. One reasonable use of overlapping
PDBs is when pods are being transitioned from one PDB to another.
4.8.10 - Accessing the Kubernetes API from a Pod
This guide demonstrates how to access the Kubernetes API from within a pod.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
When accessing the API from within a Pod, locating and authenticating
to the API server are slightly different to the external client case.
The easiest way to use the Kubernetes API from a Pod is to use
one of the official client libraries. These
libraries can automatically discover the API server and authenticate.
Using Official Client Libraries
From within a Pod, the recommended ways to connect to the Kubernetes API are:
For a Go client, use the official Go client library.
The rest.InClusterConfig() function handles API host discovery and authentication automatically.
See an example here.
For a Python client, use the official Python client library.
The config.load_incluster_config() function handles API host discovery and authentication automatically.
See an example here.
There are a number of other libraries available, please refer to the Client Libraries page.
In each case, the service account credentials of the Pod are used to communicate
securely with the API server.
Directly accessing the REST API
While running in a Pod, the Kubernetes apiserver is accessible via a Service named
kubernetes in the default namespace. Therefore, Pods can use the
kubernetes.default.svc hostname to query the API server. Official client libraries
do this automatically.
The recommended way to authenticate to the API server is with a
service account
credential. By default, a Pod
is associated with a service account, and a credential (token) for that
service account is placed into the filesystem tree of each container in that Pod,
at /var/run/secrets/kubernetes.io/serviceaccount/token.
If available, a certificate bundle is placed into the filesystem tree of each
container at /var/run/secrets/kubernetes.io/serviceaccount/ca.crt, and should be
used to verify the serving certificate of the API server.
Finally, the default namespace to be used for namespaced API operations is placed in a file
at /var/run/secrets/kubernetes.io/serviceaccount/namespace in each container.
Using kubectl proxy
If you would like to query the API without an official client library, you can run kubectl proxy
as the command
of a new sidecar container in the Pod. This way, kubectl proxy will authenticate
to the API and expose it on the localhost interface of the Pod, so that other containers
in the Pod can use it directly.
Without using a proxy
It is possible to avoid using the kubectl proxy by passing the authentication token
directly to the API server. The internal certificate secures the connection.
# Point to the internal API server hostnameAPISERVER=https://kubernetes.default.svc
# Path to ServiceAccount tokenSERVICEACCOUNT=/var/run/secrets/kubernetes.io/serviceaccount
# Read this Pod's namespaceNAMESPACE=$(cat ${SERVICEACCOUNT}/namespace)# Read the ServiceAccount bearer tokenTOKEN=$(cat ${SERVICEACCOUNT}/token)# Reference the internal certificate authority (CA)CACERT=${SERVICEACCOUNT}/ca.crt
# Explore the API with TOKENcurl --cacert ${CACERT} --header "Authorization: Bearer ${TOKEN}" -X GET ${APISERVER}/api
CronJobs was promoted to general availability in Kubernetes v1.21. If you are using an older version of
Kubernetes, please refer to the documentation for the version of Kubernetes that you are using,
so that you see accurate information. Older Kubernetes versions do not support the batch/v1 CronJob API.
You can use a CronJob to run Jobs
on a time-based schedule.
These automated jobs run like Cron tasks on a Linux or UNIX system.
Cron jobs are useful for creating periodic and recurring tasks, like running backups or sending emails.
Cron jobs can also schedule individual tasks for a specific time, such as if you want to schedule a job for a low activity period.
Cron jobs have limitations and idiosyncrasies.
For example, in certain circumstances, a single cron job can create multiple jobs.
Therefore, jobs should be idempotent.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
After creating the cron job, get its status using this command:
kubectl get cronjob hello
The output is similar to this:
NAME SCHEDULE SUSPEND ACTIVE LAST SCHEDULE AGE
hello */1 * * * * False 0 <none> 10s
As you can see from the results of the command, the cron job has not scheduled or run any jobs yet.
Watch for the job to be created in around one minute:
kubectl get jobs --watch
The output is similar to this:
NAME COMPLETIONS DURATION AGE
hello-4111706356 0/1 0s
hello-4111706356 0/1 0s 0s
hello-4111706356 1/1 5s 5s
Now you've seen one running job scheduled by the "hello" cron job.
You can stop watching the job and view the cron job again to see that it scheduled the job:
kubectl get cronjob hello
The output is similar to this:
NAME SCHEDULE SUSPEND ACTIVE LAST SCHEDULE AGE
hello */1 * * * * False 0 50s 75s
You should see that the cron job hello successfully scheduled a job at the time specified in
LAST SCHEDULE. There are currently 0 active jobs, meaning that the job has completed or failed.
Now, find the pods that the last scheduled job created and view the standard output of one of the pods.
Note: The job name is different from the pod name.
# Replace "hello-4111706356" with the job name in your systempods=$(kubectl get pods --selector=job-name=hello-4111706356 --output=jsonpath={.items[*].metadata.name})
Show the pod log:
kubectl logs $pods
The output is similar to this:
Fri Feb 22 11:02:09 UTC 2019
Hello from the Kubernetes cluster
Deleting a CronJob
When you don't need a cron job any more, delete it with kubectl delete cronjob <cronjob name>:
kubectl delete cronjob hello
Deleting the cron job removes all the jobs and pods it created and stops it from creating additional jobs.
You can read more about removing jobs in garbage collection.
Writing a CronJob Spec
As with all other Kubernetes objects, a CronJob must have apiVersion, kind, and metadata fields.
For more information about working with Kubernetes objects and their
manifests, see the
managing resources,
and using kubectl to manage resources documents.
Each manifest for a CronJob also needs a .spec section.
Note: If you modify a CronJob, the changes you make will apply to new jobs that start to run after your modification
is complete. Jobs (and their Pods) that have already started continue to run without changes.
That is, the CronJob does not update existing jobs, even if those remain running.
Schedule
The .spec.schedule is a required field of the .spec.
It takes a Cron format string, such as 0 * * * * or @hourly,
as schedule time of its jobs to be created and executed.
The format also includes extended "Vixie cron" step values. As explained in the
FreeBSD manual:
Step values can be used in conjunction with ranges. Following a range
with /<number> specifies skips of the number's value through the
range. For example, 0-23/2 can be used in the hours field to specify
command execution every other hour (the alternative in the V7 standard is
0,2,4,6,8,10,12,14,16,18,20,22). Steps are also permitted after an
asterisk, so if you want to say "every two hours", just use */2.
Note: A question mark (?) in the schedule has the same meaning as an asterisk *, that is,
it stands for any of available value for a given field.
Job Template
The .spec.jobTemplate is the template for the job, and it is required.
It has exactly the same schema as a Job, except that
it is nested and does not have an apiVersion or kind.
For information about writing a job .spec, see Writing a Job Spec.
Starting Deadline
The .spec.startingDeadlineSeconds field is optional.
It stands for the deadline in seconds for starting the job if it misses its scheduled time for any reason.
After the deadline, the cron job does not start the job.
Jobs that do not meet their deadline in this way count as failed jobs.
If this field is not specified, the jobs have no deadline.
If the .spec.startingDeadlineSeconds field is set (not null), the CronJob
controller measures the time between when a job is expected to be created and
now. If the difference is higher than that limit, it will skip this execution.
For example, if it is set to 200, it allows a job to be created for up to 200
seconds after the actual schedule.
Concurrency Policy
The .spec.concurrencyPolicy field is also optional.
It specifies how to treat concurrent executions of a job that is created by this cron job.
The spec may specify only one of the following concurrency policies:
Allow (default): The cron job allows concurrently running jobs
Forbid: The cron job does not allow concurrent runs; if it is time for a new job run and the
previous job run hasn't finished yet, the cron job skips the new job run
Replace: If it is time for a new job run and the previous job run hasn't finished yet, the
cron job replaces the currently running job run with a new job run
Note that concurrency policy only applies to the jobs created by the same cron job.
If there are multiple cron jobs, their respective jobs are always allowed to run concurrently.
Suspend
The .spec.suspend field is also optional.
If it is set to true, all subsequent executions are suspended.
This setting does not apply to already started executions.
Defaults to false.
Caution: Executions that are suspended during their scheduled time count as missed jobs.
When .spec.suspend changes from true to false on an existing cron job without a
starting deadline, the missed jobs are scheduled immediately.
Jobs History Limits
The .spec.successfulJobsHistoryLimit and .spec.failedJobsHistoryLimit fields are optional.
These fields specify how many completed and failed jobs should be kept.
By default, they are set to 3 and 1 respectively. Setting a limit to 0 corresponds to keeping
none of the corresponding kind of jobs after they finish.
4.9.2 - Coarse Parallel Processing Using a Work Queue
In this example, we will run a Kubernetes Job with multiple parallel
worker processes.
In this example, as each pod is created, it picks up one unit of work
from a task queue, completes it, deletes it from the queue, and exits.
Here is an overview of the steps in this example:
Start a message queue service. In this example, we use RabbitMQ, but you could use another
one. In practice you would set up a message queue service once and reuse it for many jobs.
Create a queue, and fill it with messages. Each message represents one task to be done. In
this example, a message is an integer that we will do a lengthy computation on.
Start a Job that works on tasks from the queue. The Job starts several pods. Each pod takes
one task from the message queue, processes it, and repeats until the end of the queue is reached.
Before you begin
Be familiar with the basic,
non-parallel, use of Job.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Now, we can experiment with accessing the message queue. We will
create a temporary interactive pod, install some tools on it,
and experiment with queues.
First create a temporary interactive Pod.
# Create a temporary interactive containerkubectl run -i --tty temp --image ubuntu:18.04
Waiting for pod default/temp-loe07 to be running, status is Pending, pod ready: false
... [ previous line repeats several times .. hit return when it stops ] ...
Note that your pod name and command prompt will be different.
Next install the amqp-tools so we can work with message queues.
# Install some toolsroot@temp-loe07:/# apt-get update
.... [ lots of output ] ....
root@temp-loe07:/# apt-get install -y curl ca-certificates amqp-tools python dnsutils
.... [ lots of output ] ....
Later, we will make a docker image that includes these packages.
Next, we will check that we can discover the rabbitmq service:
# Note the rabbitmq-service has a DNS name, provided by Kubernetes:
root@temp-loe07:/# nslookup rabbitmq-service
Server: 10.0.0.10
Address: 10.0.0.10#53
Name: rabbitmq-service.default.svc.cluster.local
Address: 10.0.147.152
# Your address will vary.
If Kube-DNS is not set up correctly, the previous step may not work for you.
You can also find the service IP in an env var:
# env | grep RABBIT | grep HOST
RABBITMQ_SERVICE_SERVICE_HOST=10.0.147.152
# Your address will vary.
Next we will verify we can create a queue, and publish and consume messages.
# In the next line, rabbitmq-service is the hostname where the rabbitmq-service# can be reached. 5672 is the standard port for rabbitmq.root@temp-loe07:/# exportBROKER_URL=amqp://guest:guest@rabbitmq-service:5672
# If you could not resolve "rabbitmq-service" in the previous step,# then use this command instead:# root@temp-loe07:/# BROKER_URL=amqp://guest:guest@$RABBITMQ_SERVICE_SERVICE_HOST:5672# Now create a queue:root@temp-loe07:/# /usr/bin/amqp-declare-queue --url=$BROKER_URL -q foo -d
foo
# Publish one message to it:root@temp-loe07:/# /usr/bin/amqp-publish --url=$BROKER_URL -r foo -p -b Hello
# And get it back.root@temp-loe07:/# /usr/bin/amqp-consume --url=$BROKER_URL -q foo -c 1 cat &&echoHello
root@temp-loe07:/#
In the last command, the amqp-consume tool takes one message (-c 1)
from the queue, and passes that message to the standard input of an arbitrary command. In this case, the program cat prints out the characters read from standard input, and the echo adds a carriage
return so the example is readable.
Filling the Queue with tasks
Now let's fill the queue with some "tasks". In our example, our tasks are strings to be
printed.
In a practice, the content of the messages might be:
names of files to that need to be processed
extra flags to the program
ranges of keys in a database table
configuration parameters to a simulation
frame numbers of a scene to be rendered
In practice, if there is large data that is needed in a read-only mode by all pods
of the Job, you will typically put that in a shared file system like NFS and mount
that readonly on all the pods, or the program in the pod will natively read data from
a cluster file system like HDFS.
For our example, we will create the queue and fill it using the amqp command line tools.
In practice, you might write a program to fill the queue using an amqp client library.
#!/usr/bin/env python# Just prints standard out and sleeps for 10 seconds.importsysimporttimeprint("Processing "+ sys.stdin.readlines()[0])
time.sleep(10)
Give the script execution permission:
chmod +x worker.py
Now, build an image. If you are working in the source
tree, then change directory to examples/job/work-queue-1.
Otherwise, make a temporary directory, change to it,
download the Dockerfile,
and worker.py. In either case,
build the image with this command:
docker build -t job-wq-1 .
For the Docker Hub, tag your app image with
your username and push to the Hub with the below commands. Replace
<username> with your Hub username.
docker tag job-wq-1 <username>/job-wq-1
docker push <username>/job-wq-1
If you are using Google Container
Registry, tag
your app image with your project ID, and push to GCR. Replace
<project> with your project ID.
docker tag job-wq-1 gcr.io/<project>/job-wq-1
gcloud docker -- push gcr.io/<project>/job-wq-1
Defining a Job
Here is a job definition. You'll need to make a copy of the Job and edit the
image to match the name you used, and call it ./job.yaml.
In this example, each pod works on one item from the queue and then exits.
So, the completion count of the Job corresponds to the number of work items
done. So we set, .spec.completions: 8 for the example, since we put 8 items in the queue.
Running the Job
So, now run the Job:
kubectl apply -f ./job.yaml
Now wait a bit, then check on the job.
kubectl describe jobs/job-wq-1
Name: job-wq-1
Namespace: default
Selector: controller-uid=41d75705-92df-11e7-b85e-fa163ee3c11f
Labels: controller-uid=41d75705-92df-11e7-b85e-fa163ee3c11f
job-name=job-wq-1
Annotations: <none>
Parallelism: 2
Completions: 8
Start Time: Wed, 06 Sep 2017 16:42:02 +0800
Pods Statuses: 0 Running / 8 Succeeded / 0 Failed
Pod Template:
Labels: controller-uid=41d75705-92df-11e7-b85e-fa163ee3c11f
job-name=job-wq-1
Containers:
c:
Image: gcr.io/causal-jigsaw-637/job-wq-1
Port:
Environment:
BROKER_URL: amqp://guest:guest@rabbitmq-service:5672
QUEUE: job1
Mounts: <none>
Volumes: <none>
Events:
FirstSeen LastSeen Count From SubobjectPath Type Reason Message
───────── ──────── ───── ──── ───────────── ────── ────── ───────
27s 27s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-hcobb
27s 27s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-weytj
27s 27s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-qaam5
27s 27s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-b67sr
26s 26s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-xe5hj
15s 15s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-w2zqe
14s 14s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-d6ppa
14s 14s 1 {job } Normal SuccessfulCreate Created pod: job-wq-1-p17e0
All our pods succeeded. Yay.
Alternatives
This approach has the advantage that you
do not need to modify your "worker" program to be aware that there is a work queue.
It does require that you run a message queue service.
If running a queue service is inconvenient, you may
want to consider one of the other job patterns.
This approach creates a pod for every work item. If your work items only take a few seconds,
though, creating a Pod for every work item may add a lot of overhead. Consider another
example, that executes multiple work items per Pod.
In this example, we use the amqp-consume utility to read the message
from the queue and run our actual program. This has the advantage that you
do not need to modify your program to be aware of the queue.
A different example, shows how to
communicate with the work queue using a client library.
Caveats
If the number of completions is set to less than the number of items in the queue, then
not all items will be processed.
If the number of completions is set to more than the number of items in the queue,
then the Job will not appear to be completed, even though all items in the queue
have been processed. It will start additional pods which will block waiting
for a message.
There is an unlikely race with this pattern. If the container is killed in between the time
that the message is acknowledged by the amqp-consume command and the time that the container
exits with success, or if the node crashes before the kubelet is able to post the success of the pod
back to the api-server, then the Job will not appear to be complete, even though all items
in the queue have been processed.
4.9.3 - Fine Parallel Processing Using a Work Queue
In this example, we will run a Kubernetes Job with multiple parallel
worker processes in a given pod.
In this example, as each pod is created, it picks up one unit of work
from a task queue, processes it, and repeats until the end of the queue is reached.
Here is an overview of the steps in this example:
Start a storage service to hold the work queue. In this example, we use Redis to store
our work items. In the previous example, we used RabbitMQ. In this example, we use Redis and
a custom work-queue client library because AMQP does not provide a good way for clients to
detect when a finite-length work queue is empty. In practice you would set up a store such
as Redis once and reuse it for the work queues of many jobs, and other things.
Create a queue, and fill it with messages. Each message represents one task to be done. In
this example, a message is an integer that we will do a lengthy computation on.
Start a Job that works on tasks from the queue. The Job starts several pods. Each pod takes
one task from the message queue, processes it, and repeats until the end of the queue is reached.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Be familiar with the basic,
non-parallel, use of Job.
Starting Redis
For this example, for simplicity, we will start a single instance of Redis.
See the Redis Example for an example
of deploying Redis scalably and redundantly.
You could also download the following files directly:
Now let's fill the queue with some "tasks". In our example, our tasks are strings to be
printed.
Start a temporary interactive pod for running the Redis CLI.
kubectl run -i --tty temp --image redis --command "/bin/sh"Waiting for pod default/redis2-c7h78 to be running, status is Pending, pod ready: falseHit enter forcommand prompt
Now hit enter, start the redis CLI, and create a list with some work items in it.
#!/usr/bin/env pythonimporttimeimportrediswqhost="redis"# Uncomment next two lines if you do not have Kube-DNS working.# import os# host = os.getenv("REDIS_SERVICE_HOST")q = rediswq.RedisWQ(name="job2", host=host)
print("Worker with sessionID: "+ q.sessionID())
print("Initial queue state: empty="+str(q.empty()))
whilenot q.empty():
item = q.lease(lease_secs=10, block=True, timeout=2)
if item isnotNone:
itemstr = item.decode("utf-8")
print("Working on "+ itemstr)
time.sleep(10) # Put your actual work here instead of sleep. q.complete(item)
else:
print("Waiting for work")
print("Queue empty, exiting")
Be sure to edit the job template to
change gcr.io/myproject to your own path.
In this example, each pod works on several items from the queue and then exits when there are no more items.
Since the workers themselves detect when the workqueue is empty, and the Job controller does not
know about the workqueue, it relies on the workers to signal when they are done working.
The workers signal that the queue is empty by exiting with success. So, as soon as any worker
exits with success, the controller knows the work is done, and the Pods will exit soon.
So, we set the completion count of the Job to 1. The job controller will wait for the other pods to complete
too.
Running the Job
So, now run the Job:
kubectl apply -f ./job.yaml
Now wait a bit, then check on the job.
kubectl describe jobs/job-wq-2
Name: job-wq-2
Namespace: default
Selector: controller-uid=b1c7e4e3-92e1-11e7-b85e-fa163ee3c11f
Labels: controller-uid=b1c7e4e3-92e1-11e7-b85e-fa163ee3c11f
job-name=job-wq-2
Annotations: <none>
Parallelism: 2Completions: <unset>
Start Time: Mon, 11 Jan 2016 17:07:59 -0800
Pods Statuses: 1 Running / 0 Succeeded / 0 Failed
Pod Template:
Labels: controller-uid=b1c7e4e3-92e1-11e7-b85e-fa163ee3c11f
job-name=job-wq-2
Containers:
c:
Image: gcr.io/exampleproject/job-wq-2
Port:
Environment: <none>
Mounts: <none>
Volumes: <none>
Events:
FirstSeen LastSeen Count From SubobjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
33s 33s 1{job-controller } Normal SuccessfulCreate Created pod: job-wq-2-lglf8
kubectl logs pods/job-wq-2-7r7b2
Worker with sessionID: bbd72d0a-9e5c-4dd6-abf6-416cc267991f
Initial queue state: empty=False
Working on banana
Working on date
Working on lemon
As you can see, one of our pods worked on several work units.
Alternatives
If running a queue service or modifying your containers to use a work queue is inconvenient, you may
want to consider one of the other
job patterns.
If you have a continuous stream of background processing work to run, then
consider running your background workers with a ReplicaSet instead,
and consider running a background processing library such as
https://github.com/resque/resque.
4.9.4 - Indexed Job for Parallel Processing with Static Work Assignment
FEATURE STATE:Kubernetes v1.24 [stable]
In this example, you will run a Kubernetes Job that uses multiple parallel
worker processes.
Each worker is a different container running in its own Pod. The Pods have an
index number that the control plane sets automatically, which allows each Pod
to identify which part of the overall task to work on.
The pod index is available in the annotationbatch.kubernetes.io/job-completion-index as a string representing its
decimal value. In order for the containerized task process to obtain this index,
you can publish the value of the annotation using the downward API
mechanism.
For convenience, the control plane automatically sets the downward API to
expose the index in the JOB_COMPLETION_INDEX environment variable.
Here is an overview of the steps in this example:
Define a Job manifest using indexed completion.
The downward API allows you to pass the pod index annotation as an
environment variable or file to the container.
Start an Indexed Job based on that manifest.
Before you begin
You should already be familiar with the basic,
non-parallel, use of Job.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.21.
To check the version, enter kubectl version.
Choose an approach
To access the work item from the worker program, you have a few options:
Read the JOB_COMPLETION_INDEX environment variable. The Job
controller
automatically links this variable to the annotation containing the completion
index.
Read a file that contains the completion index.
Assuming that you can't modify the program, you can wrap it with a script
that reads the index using any of the methods above and converts it into
something that the program can use as input.
For this example, imagine that you chose option 3 and you want to run the
rev utility. This
program accepts a file as an argument and prints its content reversed.
rev data.txt
You'll use the rev tool from the
busybox container image.
As this is only an example, each Pod only does a tiny piece of work (reversing a short
string). In a real workload you might, for example, create a Job that represents
the
task of producing 60 seconds of video based on scene data.
Each work item in the video rendering Job would be to render a particular
frame of that video clip. Indexed completion would mean that each Pod in
the Job knows which frame to render and publish, by counting frames from
the start of the clip.
Define an Indexed Job
Here is a sample Job manifest that uses Indexed completion mode:
# This uses the first approach (relying on $JOB_COMPLETION_INDEX)kubectl apply -f https://kubernetes.io/examples/application/job/indexed-job.yaml
When you create this Job, the control plane creates a series of Pods, one for each index you specified. The value of .spec.parallelism determines how many can run at once whereas .spec.completions determines how many Pods the Job creates in total.
Because .spec.parallelism is less than .spec.completions, the control plane waits for some of the first Pods to complete before starting more of them.
Once you have created the Job, wait a moment then check on progress:
kubectl describe jobs/indexed-job
The output is similar to:
Name: indexed-job
Namespace: default
Selector: controller-uid=bf865e04-0b67-483b-9a90-74cfc4c3e756
Labels: controller-uid=bf865e04-0b67-483b-9a90-74cfc4c3e756
job-name=indexed-job
Annotations: <none>
Parallelism: 3
Completions: 5
Start Time: Thu, 11 Mar 2021 15:47:34 +0000
Pods Statuses: 2 Running / 3 Succeeded / 0 Failed
Completed Indexes: 0-2
Pod Template:
Labels: controller-uid=bf865e04-0b67-483b-9a90-74cfc4c3e756
job-name=indexed-job
Init Containers:
input:
Image: docker.io/library/bash
Port: <none>
Host Port: <none>
Command:
bash
-c
items=(foo bar baz qux xyz)
echo ${items[$JOB_COMPLETION_INDEX]} > /input/data.txt
Environment: <none>
Mounts:
/input from input (rw)
Containers:
worker:
Image: docker.io/library/busybox
Port: <none>
Host Port: <none>
Command:
rev
/input/data.txt
Environment: <none>
Mounts:
/input from input (rw)
Volumes:
input:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium:
SizeLimit: <unset>
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 4s job-controller Created pod: indexed-job-njkjj
Normal SuccessfulCreate 4s job-controller Created pod: indexed-job-9kd4h
Normal SuccessfulCreate 4s job-controller Created pod: indexed-job-qjwsz
Normal SuccessfulCreate 1s job-controller Created pod: indexed-job-fdhq5
Normal SuccessfulCreate 1s job-controller Created pod: indexed-job-ncslj
In this example, you run the Job with custom values for each index. You can
inspect the output of one of the pods:
kubectl logs indexed-job-fdhq5 # Change this to match the name of a Pod from that Job
The output is similar to:
xuq
4.9.5 - Parallel Processing using Expansions
This task demonstrates running multiple Jobs
based on a common template. You can use this approach to process batches of work in
parallel.
For this example there are only three items: apple, banana, and cherry.
The sample Jobs process each item by printing a string then pausing.
You should be familiar with the basic,
non-parallel, use of Job.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
# Use curl to download job-tmpl.yamlcurl -L -s -O https://k8s.io/examples/application/job/job-tmpl.yaml
The file you downloaded is not yet a valid Kubernetes
manifest.
Instead that template is a YAML representation of a Job object with some placeholders
that need to be filled in before it can be used. The $ITEM syntax is not meaningful to Kubernetes.
Create manifests from the template
The following shell snippet uses sed to replace the string $ITEM with the loop
variable, writing into a temporary directory named jobs. Run this now:
# Expand the template into multiple files, one for each item to be processed.mkdir ./jobs
for i in apple banana cherry
do cat job-tmpl.yaml | sed "s/\$ITEM/$i/" > ./jobs/job-$i.yaml
done
Check if it worked:
ls jobs/
The output is similar to this:
job-apple.yaml
job-banana.yaml
job-cherry.yaml
You could use any type of template language (for example: Jinja2; ERB), or
write a program to generate the Job manifests.
Create Jobs from the manifests
Next, create all the Jobs with one kubectl command:
kubectl create -f ./jobs
The output is similar to this:
job.batch/process-item-apple created
job.batch/process-item-banana created
job.batch/process-item-cherry created
Now, check on the jobs:
kubectl get jobs -l jobgroup=jobexample
The output is similar to this:
NAME COMPLETIONS DURATION AGE
process-item-apple 1/1 14s 22s
process-item-banana 1/1 12s 21s
process-item-cherry 1/1 12s 20s
Using the -l option to kubectl selects only the Jobs that are part
of this group of jobs (there might be other unrelated jobs in the system).
You can check on the Pods as well using the same
label selector:
kubectl get pods -l jobgroup=jobexample
The output is similar to:
NAME READY STATUS RESTARTS AGE
process-item-apple-kixwv 0/1 Completed 0 4m
process-item-banana-wrsf7 0/1 Completed 0 4m
process-item-cherry-dnfu9 0/1 Completed 0 4m
We can use this single command to check on the output of all jobs at once:
kubectl logs -f -l jobgroup=jobexample
The output should be:
Processing item apple
Processing item banana
Processing item cherry
Clean up
# Remove the Jobs you created# Your cluster automatically cleans up their Podskubectl delete job -l jobgroup=jobexample
Use advanced template parameters
In the first example, each instance of the template had one
parameter, and that parameter was also used in the Job's name. However,
names are restricted
to contain only certain characters.
This slightly more complex example uses the
Jinja template language to generate manifests
and then objects from those manifests, with a multiple parameters for each Job.
For this part of the task, you are going to use a one-line Python script to
convert the template to a set of manifests.
First, copy and paste the following template of a Job object, into a file called job.yaml.jinja2:
{% set params = [{ "name": "apple", "url": "http://dbpedia.org/resource/Apple", },
{ "name": "banana", "url": "http://dbpedia.org/resource/Banana", },
{ "name": "cherry", "url": "http://dbpedia.org/resource/Cherry" }]
%}
{% for p in params %}
{% set name = p["name"] %}
{% set url = p["url"] %}
---
apiVersion: batch/v1
kind: Job
metadata:
name: jobexample-{{ name }}
labels:
jobgroup: jobexample
spec:
template:
metadata:
name: jobexample
labels:
jobgroup: jobexample
spec:
containers:
- name: c
image: busybox:1.28
command: ["sh", "-c", "echo Processing URL {{ url }} && sleep 5"]
restartPolicy: Never
{% endfor %}
The above template defines two parameters for each Job object using a list of
python dicts (lines 1-4). A for loop emits one Job manifest for each
set of parameters (remaining lines).
This example relies on a feature of YAML. One YAML file can contain multiple
documents (Kubernetes manifests, in this case), separated by --- on a line
by itself.
You can pipe the output directly to kubectl to create the Jobs.
Next, use this one-line Python program to expand the template:
# Remove the Jobs you created# Your cluster automatically cleans up their Podskubectl delete job -l jobgroup=jobexample
Using Jobs in real workloads
In a real use case, each Job performs some substantial computation, such as rendering a frame
of a movie, or processing a range of rows in a database. If you were rendering a movie
you would set $ITEM to the frame number. If you were processing rows from a database
table, you would set $ITEM to represent the range of database rows to process.
In the task, you ran a command to collect the output from Pods by fetching
their logs. In a real use case, each Pod for a Job writes its output to
durable storage before completing. You can use a PersistentVolume for each Job,
or an external storage service. For example, if you are rendering frames for a movie,
use HTTP to PUT the rendered frame data to a URL, using a different URL for each
frame.
Labels on Jobs and Pods
After you create a Job, Kubernetes automatically adds additional
labels that
distinguish one Job's pods from another Job's pods.
In this example, each Job and its Pod template have a label:
jobgroup=jobexample.
Kubernetes itself pays no attention to labels named jobgroup. Setting a label
for all the Jobs you create from a template makes it convenient to operate on all
those Jobs at once.
In the first example you used a template to
create several Jobs. The template ensures that each Pod also gets the same label, so
you can check on all Pods for these templated Jobs with a single command.
Note: The label key jobgroup is not special or reserved.
You can pick your own labelling scheme.
There are recommended labels
that you can use if you wish.
Alternatives
If you plan to create a large number of Job objects, you may find that:
Even using labels, managing so many Jobs is cumbersome.
If you create many Jobs in a batch, you might place high load
on the Kubernetes control plane. Alternatively, the Kubernetes API
server could rate limit you, temporarily rejecting your requests with a 429 status.
You are limited by a resource quota
on Jobs: the API server permanently rejects some of your requests
when you create a great deal of work in one batch.
There are other job patterns
that you can use to process large amounts of work without creating very many Job
objects.
You could also consider writing your own controller
to manage Job objects automatically.
4.9.6 - Handling retriable and non-retriable pod failures with Pod failure policy
FEATURE STATE:Kubernetes v1.25 [alpha]
This document shows you how to use the
Pod failure policy,
in combination with the default
Pod backoff failure policy,
to improve the control over the handling of container- or Pod-level failure
within a Job.
The definition of Pod failure policy may help you to:
better utilize the computational resources by avoiding unnecessary Pod retries.
You should already be familiar with the basic use of Job.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be version v1.25.
To check the version, enter kubectl version.
Note: As the features are in Alpha, prepare the Kubernetes cluster with the two
feature gates
enabled: JobPodFailurePolicy and PodDisruptionsCondition.
Using Pod failure policy to avoid unnecessary Pod retries
With the following example, you can learn how to use Pod failure policy to
avoid unnecessary Pod restarts when a Pod failure indicates a non-retriable
software bug.
After around 30s the entire Job should be terminated. Inspect the status of the Job by running:
kubectl get jobs -l job-name=job-pod-failure-policy-failjob -o yaml
In the Job status, see a job Failed condition with the field reason
equal PodFailurePolicy. Additionally, the message field contains a
more detailed information about the Job termination, such as:
Container main for pod default/job-pod-failure-policy-failjob-8ckj8 failed with exit code 42 matching FailJob rule at index 0.
For comparison, if the Pod failure policy was disabled it would take 6 retries
of the Pod, taking at least 2 minutes.
Using Pod failure policy to ignore Pod disruptions
With the following example, you can learn how to use Pod failure policy to
ignore Pod disruptions from incrementing the Pod retry counter towards the
.spec.backoffLimit limit.
Caution: Timing is important for this example, so you may want to read the steps before
execution. In order to trigger a Pod disruption it is important to drain the
node while the Pod is running on it (within 90s since the Pod is scheduled).
Inspect the .status.failed to check the counter for the Job is not incremented:
kubectl get jobs -l job-name=job-pod-failure-policy-ignore -o yaml
Uncordon the node:
kubectl uncordon nodes/$nodeName
The Job resumes and succeeds.
For comparison, if the Pod failure policy was disabled the Pod disruption would
result in terminating the entire Job (as the .spec.backoffLimit is set to 0).
Cleaning up
Delete the Job you created:
kubectl delete jobs/job-pod-failure-policy-ignore
The cluster automatically cleans up the Pods.
Alternatives
You could rely solely on the
Pod backoff failure policy,
by specifying the Job's .spec.backoffLimit field. However, in many situations
it is problematic to find a balance between setting the a low value for .spec.backoffLimit
to avoid unnecessary Pod retries, yet high enough to make sure the Job would
not be terminated by Pod disruptions.
4.10 - Access Applications in a Cluster
Configure load balancing, port forwarding, or setup firewall or DNS configurations to access applications in a cluster.
4.10.1 - Deploy and Access the Kubernetes Dashboard
Deploy the web UI (Kubernetes Dashboard) and access it.
Dashboard is a web-based Kubernetes user interface.
You can use Dashboard to deploy containerized applications to a Kubernetes cluster,
troubleshoot your containerized application, and manage the cluster resources.
You can use Dashboard to get an overview of applications running on your cluster,
as well as for creating or modifying individual Kubernetes resources
(such as Deployments, Jobs, DaemonSets, etc).
For example, you can scale a Deployment, initiate a rolling update, restart a pod
or deploy new applications using a deploy wizard.
Dashboard also provides information on the state of Kubernetes resources in your cluster and on any errors that may have occurred.
Deploying the Dashboard UI
The Dashboard UI is not deployed by default. To deploy it, run the following command:
To protect your cluster data, Dashboard deploys with a minimal RBAC configuration by default.
Currently, Dashboard only supports logging in with a Bearer Token.
To create a token for this demo, you can follow our guide on
creating a sample user.
Warning: The sample user created in the tutorial will have administrative privileges and is for educational purposes only.
Command line proxy
You can enable access to the Dashboard using the kubectl command-line tool,
by running the following command:
The UI can only be accessed from the machine where the command is executed. See kubectl proxy --help for more options.
Note: The kubeconfig authentication method does not support external identity providers
or X.509 certificate-based authentication.
Welcome view
When you access Dashboard on an empty cluster, you'll see the welcome page.
This page contains a link to this document as well as a button to deploy your first application.
In addition, you can view which system applications are running by default in the kube-systemnamespace of your cluster, for example the Dashboard itself.
Deploying containerized applications
Dashboard lets you create and deploy a containerized application as a Deployment and optional Service with a simple wizard.
You can either manually specify application details, or upload a YAML or JSON manifest file containing application configuration.
Click the CREATE button in the upper right corner of any page to begin.
Specifying application details
The deploy wizard expects that you provide the following information:
App name (mandatory): Name for your application.
A label with the name will be
added to the Deployment and Service, if any, that will be deployed.
The application name must be unique within the selected Kubernetes namespace.
It must start with a lowercase character, and end with a lowercase character or a number,
and contain only lowercase letters, numbers and dashes (-). It is limited to 24 characters.
Leading and trailing spaces are ignored.
Container image (mandatory):
The URL of a public Docker container image on any registry,
or a private image (commonly hosted on the Google Container Registry or Docker Hub).
The container image specification must end with a colon.
Number of pods (mandatory): The target number of Pods you want your application to be deployed in.
The value must be a positive integer.
A Deployment will be created to
maintain the desired number of Pods across your cluster.
Service (optional): For some parts of your application (e.g. frontends) you may want to expose a
Service onto an external,
maybe public IP address outside of your cluster (external Service).
Note: For external Services, you may need to open up one or more ports to do so.
Other Services that are only visible from inside the cluster are called internal Services.
Irrespective of the Service type, if you choose to create a Service and your container listens
on a port (incoming), you need to specify two ports.
The Service will be created mapping the port (incoming) to the target port seen by the container.
This Service will route to your deployed Pods. Supported protocols are TCP and UDP.
The internal DNS name for this Service will be the value you specified as application name above.
If needed, you can expand the Advanced options section where you can specify more settings:
Description: The text you enter here will be added as an
annotation
to the Deployment and displayed in the application's details.
Labels: Default labels to be used
for your application are application name and version.
You can specify additional labels to be applied to the Deployment, Service (if any), and Pods,
such as release, environment, tier, partition, and release track.
Namespace: Kubernetes supports multiple virtual clusters backed by the same physical cluster.
These virtual clusters are called namespaces.
They let you partition resources into logically named groups.
Dashboard offers all available namespaces in a dropdown list, and allows you to create a new namespace.
The namespace name may contain a maximum of 63 alphanumeric characters and dashes (-) but can not contain capital letters.
Namespace names should not consist of only numbers.
If the name is set as a number, such as 10, the pod will be put in the default namespace.
In case the creation of the namespace is successful, it is selected by default.
If the creation fails, the first namespace is selected.
Image Pull Secret:
In case the specified Docker container image is private, it may require
pull secret credentials.
Dashboard offers all available secrets in a dropdown list, and allows you to create a new secret.
The secret name must follow the DNS domain name syntax, for example new.image-pull.secret.
The content of a secret must be base64-encoded and specified in a
.dockercfg file.
The secret name may consist of a maximum of 253 characters.
In case the creation of the image pull secret is successful, it is selected by default. If the creation fails, no secret is applied.
CPU requirement (cores) and Memory requirement (MiB):
You can specify the minimum resource limits
for the container. By default, Pods run with unbounded CPU and memory limits.
Run command and Run command arguments:
By default, your containers run the specified Docker image's default
entrypoint command.
You can use the command options and arguments to override the default.
Run as privileged: This setting determines whether processes in
privileged containers
are equivalent to processes running as root on the host.
Privileged containers can make use of capabilities like manipulating the network stack and accessing devices.
Environment variables: Kubernetes exposes Services through
environment variables.
You can compose environment variable or pass arguments to your commands using the values of environment variables.
They can be used in applications to find a Service.
Values can reference other variables using the $(VAR_NAME) syntax.
Uploading a YAML or JSON file
Kubernetes supports declarative configuration.
In this style, all configuration is stored in manifests (YAML or JSON configuration files).
The manifests use Kubernetes API resource schemas.
As an alternative to specifying application details in the deploy wizard,
you can define your application in one or more manifests, and upload the files using Dashboard.
Using Dashboard
Following sections describe views of the Kubernetes Dashboard UI; what they provide and how can they be used.
Navigation
When there are Kubernetes objects defined in the cluster, Dashboard shows them in the initial view.
By default only objects from the default namespace are shown and
this can be changed using the namespace selector located in the navigation menu.
Dashboard shows most Kubernetes object kinds and groups them in a few menu categories.
Admin overview
For cluster and namespace administrators, Dashboard lists Nodes, Namespaces and PersistentVolumes and has detail views for them.
Node list view contains CPU and memory usage metrics aggregated across all Nodes.
The details view shows the metrics for a Node, its specification, status,
allocated resources, events and pods running on the node.
Workloads
Shows all applications running in the selected namespace.
The view lists applications by workload kind (for example: Deployments, ReplicaSets, StatefulSets).
Each workload kind can be viewed separately.
The lists summarize actionable information about the workloads,
such as the number of ready pods for a ReplicaSet or current memory usage for a Pod.
Detail views for workloads show status and specification information and
surface relationships between objects.
For example, Pods that ReplicaSet is controlling or new ReplicaSets and HorizontalPodAutoscalers for Deployments.
Services
Shows Kubernetes resources that allow for exposing services to external world and
discovering them within a cluster.
For that reason, Service and Ingress views show Pods targeted by them,
internal endpoints for cluster connections and external endpoints for external users.
Storage
Storage view shows PersistentVolumeClaim resources which are used by applications for storing data.
ConfigMaps and Secrets
Shows all Kubernetes resources that are used for live configuration of applications running in clusters.
The view allows for editing and managing config objects and displays secrets hidden by default.
Logs viewer
Pod lists and detail pages link to a logs viewer that is built into Dashboard.
The viewer allows for drilling down logs from containers belonging to a single Pod.
This topic discusses multiple ways to interact with clusters.
Accessing for the first time with kubectl
When accessing the Kubernetes API for the first time, we suggest using the
Kubernetes CLI, kubectl.
To access a cluster, you need to know the location of the cluster and have credentials
to access it. Typically, this is automatically set-up when you work through
a Getting started guide,
or someone else set up the cluster and provided you with credentials and a location.
Check the location and credentials that kubectl knows about with this command:
kubectl config view
Many of the examples provide an introduction to using
kubectl, and complete documentation is found in the
kubectl reference.
Directly accessing the REST API
Kubectl handles locating and authenticating to the apiserver.
If you want to directly access the REST API with an http client like
curl or wget, or a browser, there are several ways to locate and authenticate:
Run kubectl in proxy mode.
Recommended approach.
Uses stored apiserver location.
Verifies identity of apiserver using self-signed cert. No MITM possible.
Authenticates to apiserver.
In future, may do intelligent client-side load-balancing and failover.
Provide the location and credentials directly to the http client.
Alternate approach.
Works with some types of client code that are confused by using a proxy.
Need to import a root cert into your browser to protect against MITM.
Using kubectl proxy
The following command runs kubectl in a mode where it acts as a reverse proxy. It handles
locating the apiserver and authenticating.
Run it like this:
The above examples use the --insecure flag. This leaves it subject to MITM
attacks. When kubectl accesses the cluster it uses a stored root certificate
and client certificates to access the server. (These are installed in the
~/.kube directory). Since cluster certificates are typically self-signed, it
may take special configuration to get your http client to use root
certificate.
On some clusters, the apiserver does not require authentication; it may serve
on localhost, or be protected by a firewall. There is not a standard
for this. Controlling Access to the API
describes how a cluster admin can configure this.
Programmatic access to the API
Kubernetes officially supports Go and Python
client libraries.
Go client
To get the library, run the following command: go get k8s.io/client-go@kubernetes-<kubernetes-version-number>, see INSTALL.md for detailed installation instructions. See https://github.com/kubernetes/client-go to see which versions are supported.
Write an application atop of the client-go clients. Note that client-go defines its own API objects, so if needed, please import API definitions from client-go rather than from the main repository, e.g., import "k8s.io/client-go/kubernetes" is correct.
The Go client can use the same kubeconfig file
as the kubectl CLI does to locate and authenticate to the apiserver. See this example.
If the application is deployed as a Pod in the cluster, please refer to the next section.
The previous section describes how to connect to the Kubernetes API server.
For information about connecting to other services running on a Kubernetes cluster, see
Access Cluster Services.
Requesting redirects
The redirect capabilities have been deprecated and removed. Please use a proxy (see below) instead.
So Many Proxies
There are several different proxies you may encounter when using Kubernetes:
existence and implementation varies from cluster to cluster (e.g. nginx)
sits between all clients and one or more apiservers
acts as load balancer if there are several apiservers.
Cloud Load Balancers on external services:
are provided by some cloud providers (e.g. AWS ELB, Google Cloud Load Balancer)
are created automatically when the Kubernetes service has type LoadBalancer
use UDP/TCP only
implementation varies by cloud provider.
Kubernetes users will typically not need to worry about anything other than the first two types. The cluster admin
will typically ensure that the latter types are set up correctly.
4.10.3 - Configure Access to Multiple Clusters
This page shows how to configure access to multiple clusters by using
configuration files. After your clusters, users, and contexts are defined in
one or more configuration files, you can quickly switch between clusters by using the
kubectl config use-context command.
Note: A file that is used to configure access to a cluster is sometimes called
a kubeconfig file. This is a generic way of referring to configuration files.
It does not mean that there is a file named kubeconfig.
Warning: Only use kubeconfig files from trusted sources. Using a specially-crafted kubeconfig
file could result in malicious code execution or file exposure.
If you must use an untrusted kubeconfig file, inspect it carefully first, much as you would a shell script.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
To check that kubectl is installed,
run kubectl version --client. The kubectl version should be
within one minor version of your
cluster's API server.
Define clusters, users, and contexts
Suppose you have two clusters, one for development work and one for scratch work.
In the development cluster, your frontend developers work in a namespace called frontend,
and your storage developers work in a namespace called storage. In your scratch cluster,
developers work in the default namespace, or they create auxiliary namespaces as they
see fit. Access to the development cluster requires authentication by certificate. Access
to the scratch cluster requires authentication by username and password.
Create a directory named config-exercise. In your
config-exercise directory, create a file named config-demo with this content:
A configuration file describes clusters, users, and contexts. Your config-demo file
has the framework to describe two clusters, two users, and three contexts.
Go to your config-exercise directory. Enter these commands to add cluster details to
your configuration file:
The fake-ca-file, fake-cert-file and fake-key-file above are the placeholders
for the pathnames of the certificate files. You need to change these to the actual pathnames
of certificate files in your environment.
Sometimes you may want to use Base64-encoded data embedded here instead of separate
certificate files; in that case you need to add the suffix -data to the keys, for example,
certificate-authority-data, client-certificate-data, client-key-data.
Each context is a triple (cluster, user, namespace). For example, the
dev-frontend context says, "Use the credentials of the developer
user to access the frontend namespace of the development cluster".
Now whenever you enter a kubectl command, the action will apply to the cluster,
and namespace listed in the dev-frontend context. And the command will use
the credentials of the user listed in the dev-frontend context.
To see only the configuration information associated with
the current context, use the --minify flag.
Now any kubectl command you give will apply to the default namespace of
the scratch cluster. And the command will use the credentials of the user
listed in the exp-scratch context.
View configuration associated with the new current context, exp-scratch.
The preceding configuration file defines a new context named dev-ramp-up.
Set the KUBECONFIG environment variable
See whether you have an environment variable named KUBECONFIG. If so, save the
current value of your KUBECONFIG environment variable, so you can restore it later.
For example:
Linux
exportKUBECONFIG_SAVED="$KUBECONFIG"
Windows PowerShell
$Env:KUBECONFIG_SAVED=$ENV:KUBECONFIG
The KUBECONFIG environment variable is a list of paths to configuration files. The list is
colon-delimited for Linux and Mac, and semicolon-delimited for Windows. If you have
a KUBECONFIG environment variable, familiarize yourself with the configuration files
in the list.
Temporarily append two paths to your KUBECONFIG environment variable. For example:
In your config-exercise directory, enter this command:
kubectl config view
The output shows merged information from all the files listed in your KUBECONFIG
environment variable. In particular, notice that the merged information has the
dev-ramp-up context from the config-demo-2 file and the three contexts from
the config-demo file:
If you already have a cluster, and you can use kubectl to interact with
the cluster, then you probably have a file named config in the $HOME/.kube
directory.
Go to $HOME/.kube, and see what files are there. Typically, there is a file named
config. There might also be other configuration files in this directory. Briefly
familiarize yourself with the contents of these files.
Append $HOME/.kube/config to your KUBECONFIG environment variable
If you have a $HOME/.kube/config file, and it's not already listed in your
KUBECONFIG environment variable, append it to your KUBECONFIG environment variable now.
For example:
View configuration information merged from all the files that are now listed
in your KUBECONFIG environment variable. In your config-exercise directory, enter:
kubectl config view
Clean up
Return your KUBECONFIG environment variable to its original value. For example:
4.10.4 - Use Port Forwarding to Access Applications in a Cluster
This page shows how to use kubectl port-forward to connect to a MongoDB
server running in a Kubernetes cluster. This type of connection can be useful
for database debugging.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The output of a successful command verifies that the Service was created:
service/mongo created
Check the Service created:
kubectl get service mongo
The output displays the service created:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
mongo ClusterIP 10.96.41.183 <none> 27017/TCP 11s
Verify that the MongoDB server is running in the Pod, and listening on port 27017:
# Change mongo-75f59d57f4-4nd6q to the name of the Podkubectl get pod mongo-75f59d57f4-4nd6q --template='{{(index (index .spec.containers 0).ports 0).containerPort}}{{"\n"}}'
The output displays the port for MongoDB in that Pod:
27017
27017 is the TCP port allocated to MongoDB on the internet.
Forward a local port to a port on the Pod
kubectl port-forward allows using resource name, such as a pod name, to select a matching pod to port forward to.
# Change mongo-75f59d57f4-4nd6q to the name of the Podkubectl port-forward mongo-75f59d57f4-4nd6q 28015:27017
Any of the above commands works. The output is similar to this:
Forwarding from 127.0.0.1:28015 -> 27017
Forwarding from [::1]:28015 -> 27017
Note:kubectl port-forward does not return. To continue with the exercises, you will need to open another terminal.
Start the MongoDB command line interface:
mongosh --port 28015
At the MongoDB command line prompt, enter the ping command:
db.runCommand( { ping: 1 } )
A successful ping request returns:
{ ok: 1 }
Optionally let kubectl choose the local port
If you don't need a specific local port, you can let kubectl choose and allocate
the local port and thus relieve you from having to manage local port conflicts, with
the slightly simpler syntax:
kubectl port-forward deployment/mongo :27017
The kubectl tool finds a local port number that is not in use (avoiding low ports numbers,
because these might be used by other applications). The output is similar to:
Forwarding from 127.0.0.1:63753 -> 27017
Forwarding from [::1]:63753 -> 27017
Discussion
Connections made to local port 28015 are forwarded to port 27017 of the Pod that
is running the MongoDB server. With this connection in place, you can use your
local workstation to debug the database that is running in the Pod.
Note:kubectl port-forward is implemented for TCP ports only.
The support for UDP protocol is tracked in
issue 47862.
4.10.5 - Use a Service to Access an Application in a Cluster
This page shows how to create a Kubernetes Service object that external
clients can use to access an application running in a cluster. The Service
provides load balancing for an application that has two running instances.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Make a note of the NodePort value for the service. For example,
in the preceding output, the NodePort value is 31496.
List the pods that are running the Hello World application:
kubectl get pods --selector="run=load-balancer-example" --output=wide
The output is similar to this:
NAME READY STATUS ... IP NODE
hello-world-2895499144-bsbk5 1/1 Running ... 10.200.1.4 worker1
hello-world-2895499144-m1pwt 1/1 Running ... 10.200.2.5 worker2
Get the public IP address of one of your nodes that is running
a Hello World pod. How you get this address depends on how you set
up your cluster. For example, if you are using Minikube, you can
see the node address by running kubectl cluster-info. If you are
using Google Compute Engine instances, you can use the
gcloud compute instances list command to see the public addresses of your
nodes.
On your chosen node, create a firewall rule that allows TCP traffic
on your node port. For example, if your Service has a NodePort value of
31568, create a firewall rule that allows TCP traffic on port 31568. Different
cloud providers offer different ways of configuring firewall rules.
Use the node address and node port to access the Hello World application:
curl http://<public-node-ip>:<node-port>
where <public-node-ip> is the public IP address of your node,
and <node-port> is the NodePort value for your service. The
response to a successful request is a hello message:
Hello Kubernetes!
Using a service configuration file
As an alternative to using kubectl expose, you can use a
service configuration file
to create a Service.
Cleaning up
To delete the Service, enter this command:
kubectl delete services example-service
To delete the Deployment, the ReplicaSet, and the Pods that are running
the Hello World application, enter this command:
4.10.6 - Connect a Frontend to a Backend Using Services
This task shows how to create a frontend and a backend microservice. The backend
microservice is a hello greeter. The frontend exposes the backend using nginx and a
Kubernetes Service object.
Objectives
Create and run a sample hello backend microservice using a
Deployment object.
Use a Service object to send traffic to the backend microservice's multiple replicas.
Create and run a nginx frontend microservice, also using a Deployment object.
Configure the frontend microservice to send traffic to the backend microservice.
Use a Service object of type=LoadBalancer to expose the frontend microservice
outside the cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
This task uses
Services with external load balancers, which
require a supported environment. If your environment does not support this, you can use a Service of type
NodePort instead.
Creating the backend using a Deployment
The backend is a simple hello greeter microservice. Here is the configuration
file for the backend Deployment:
Name: backend
Namespace: default
CreationTimestamp: Mon, 24 Oct 2016 14:21:02 -0700
Labels: app=hello
tier=backend
track=stable
Annotations: deployment.kubernetes.io/revision=1
Selector: app=hello,tier=backend,track=stable
Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 1 max unavailable, 1 max surge
Pod Template:
Labels: app=hello
tier=backend
track=stable
Containers:
hello:
Image: "gcr.io/google-samples/hello-go-gke:1.0"
Port: 80/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets: <none>
NewReplicaSet: hello-3621623197 (3/3 replicas created)
Events:
...
Creating the hello Service object
The key to sending requests from a frontend to a backend is the backend
Service. A Service creates a persistent IP address and DNS name entry
so that the backend microservice can always be reached. A Service uses
selectors to find
the Pods that it routes traffic to.
At this point, you have a backend Deployment running three replicas of your hello
application, and you have a Service that can route traffic to them. However, this
service is neither available nor resolvable outside the cluster.
Creating the frontend
Now that you have your backend running, you can create a frontend that is accessible
outside the cluster, and connects to the backend by proxying requests to it.
The frontend sends requests to the backend worker Pods by using the DNS name
given to the backend Service. The DNS name is hello, which is the value
of the name field in the examples/service/access/backend-service.yaml
configuration file.
The Pods in the frontend Deployment run a nginx image that is configured
to proxy requests to the hello backend Service. Here is the nginx configuration file:
# The identifier Backend is internal to nginx, and used to name this specific upstream
upstream Backend {
# hello is the internal DNS name used by the backend Service inside Kubernetes
server hello;
}
server {
listen 80;
location / {
# The following statement will proxy traffic to the upstream named Backend
proxy_pass http://Backend;
}
}
Similar to the backend, the frontend has a Deployment and a Service. An important
difference to notice between the backend and frontend services, is that the
configuration for the frontend Service has type: LoadBalancer, which means that
the Service uses a load balancer provisioned by your cloud provider and will be
accessible from outside the cluster.
The output verifies that both resources were created:
deployment.apps/frontend created
service/frontend created
Note: The nginx configuration is baked into the
container image. A better way to do this would
be to use a
ConfigMap,
so that you can change the configuration more easily.
Interact with the frontend Service
Once you've created a Service of type LoadBalancer, you can use this
command to find the external IP:
kubectl get service frontend --watch
This displays the configuration for the frontend Service and watches for
changes. Initially, the external IP is listed as <pending>:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
frontend LoadBalancer 10.51.252.116 <pending> 80/TCP 10s
As soon as an external IP is provisioned, however, the configuration updates
to include the new IP under the EXTERNAL-IP heading:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
frontend LoadBalancer 10.51.252.116 XXX.XXX.XXX.XXX 80/TCP 1m
That IP can now be used to interact with the frontend service from outside the
cluster.
Send traffic through the frontend
The frontend and backend are now connected. You can hit the endpoint
by using the curl command on the external IP of your frontend Service.
curl http://${EXTERNAL_IP}# replace this with the EXTERNAL-IP you saw earlier
The output shows the message generated by the backend:
{"message":"Hello"}
Cleaning up
To delete the Services, enter this command:
kubectl delete services frontend backend
To delete the Deployments, the ReplicaSets and the Pods that are running the backend and frontend applications, enter this command:
This page shows how to create an external load balancer.
When creating a Service, you have
the option of automatically creating a cloud load balancer. This provides an
externally-accessible IP address that sends traffic to the correct port on your cluster
nodes,
provided your cluster runs in a supported environment and is configured with
the correct cloud load balancer provider package.
You can also use an Ingress in place of Service.
For more information, check the Ingress
documentation.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
The load balancer's IP address is listed next to LoadBalancer Ingress.
Note:
If you are running your service on Minikube, you can find the assigned IP address and port with:
minikube service example-service --url
Preserving the client source IP
By default, the source IP seen in the target container is not the original
source IP of the client. To enable preservation of the client IP, the following
fields can be configured in the .spec of the Service:
.spec.externalTrafficPolicy - denotes if this Service desires to route
external traffic to node-local or cluster-wide endpoints. There are two available
options: Cluster (default) and Local. Cluster obscures the client source
IP and may cause a second hop to another node, but should have good overall
load-spreading. Local preserves the client source IP and avoids a second hop
for LoadBalancer and NodePort type Services, but risks potentially imbalanced
traffic spreading.
.spec.healthCheckNodePort - specifies the health check node port
(numeric port number) for the service. If you don't specify
healthCheckNodePort, the service controller allocates a port from your
cluster's NodePort range.
You can configure that range by setting an API server command line option,
--service-node-port-range. The Service will use the user-specified
healthCheckNodePort value if you specify it, provided that the
Service type is set to LoadBalancer and externalTrafficPolicy is set
to Local.
Setting externalTrafficPolicy to Local in the Service manifest
activates this feature. For example:
Caveats and limitations when preserving source IPs
Load balancing services from some cloud providers do not let you configure different weights for each target.
With each target weighted equally in terms of sending traffic to Nodes, external
traffic is not equally load balanced across different Pods. The external load balancer
is unaware of the number of Pods on each node that are used as a target.
Where NumServicePods << _NumNodes or NumServicePods >> NumNodes, a fairly close-to-equal
distribution will be seen, even without weights.
Internal pod to pod traffic should behave similar to ClusterIP services, with equal probability across all pods.
Garbage collecting load balancers
FEATURE STATE:Kubernetes v1.17 [stable]
In usual case, the correlating load balancer resources in cloud provider should
be cleaned up soon after a LoadBalancer type Service is deleted. But it is known
that there are various corner cases where cloud resources are orphaned after the
associated Service is deleted. Finalizer Protection for Service LoadBalancers was
introduced to prevent this from happening. By using finalizers, a Service resource
will never be deleted until the correlating load balancer resources are also deleted.
Specifically, if a Service has type LoadBalancer, the service controller will attach
a finalizer named service.kubernetes.io/load-balancer-cleanup.
The finalizer will only be removed after the load balancer resource is cleaned up.
This prevents dangling load balancer resources even in corner cases such as the
service controller crashing.
External load balancer providers
It is important to note that the datapath for this functionality is provided by a load balancer external to the Kubernetes cluster.
When the Service type is set to LoadBalancer, Kubernetes provides functionality equivalent to type equals ClusterIP to pods
within the cluster and extends it by programming the (external to Kubernetes) load balancer with entries for the nodes
hosting the relevant Kubernetes pods. The Kubernetes control plane automates the creation of the external load balancer,
health checks (if needed), and packet filtering rules (if needed). Once the cloud provider allocates an IP address for the load
balancer, the control plane looks up that external IP address and populates it into the Service object.
4.10.8 - List All Container Images Running in a Cluster
This page shows how to use kubectl to list all of the Container images
for Pods running in a cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this exercise you will use kubectl to fetch all of the Pods
running in a cluster, and format the output to pull out the list
of Containers for each.
List all Container images in all namespaces
Fetch all Pods in all namespaces using kubectl get pods --all-namespaces
Format the output to include only the list of Container image names
using -o jsonpath={.items[*].spec.containers[*].image}. This will recursively parse out the
image field from the returned json.
See the jsonpath reference
for further information on how to use jsonpath.
Format the output using standard tools: tr, sort, uniq
The above command will recursively return all fields named image
for all items returned.
As an alternative, it is possible to use the absolute path to the image
field within the Pod. This ensures the correct field is retrieved
even when the field name is repeated,
e.g. many fields are called name within a given item:
kubectl get pods --all-namespaces -o jsonpath="{.items[*].spec.containers[*].image}"
The jsonpath is interpreted as follows:
.items[*]: for each returned value
.spec: get the spec
.containers[*]: for each container
.image: get the image
Note: When fetching a single Pod by name, for example kubectl get pod nginx,
the .items[*] portion of the path should be omitted because a single
Pod is returned instead of a list of items.
List Container images by Pod
The formatting can be controlled further by using the range operation to
iterate over elements individually.
4.10.9 - Set up Ingress on Minikube with the NGINX Ingress Controller
An Ingress is an API object that defines rules which allow external access
to services in a cluster. An Ingress controller fulfills the rules set in the Ingress.
This page shows you how to set up a simple Ingress which routes requests to Service web or web2 depending on the HTTP URI.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version 1.19.
To check the version, enter kubectl version.
If you are using an older Kubernetes version, switch to the documentation
for that version.
Create a Minikube cluster
Using Katacoda
Locally
If you already installed Minikube
locally, run minikube start to create a cluster.
Enable the Ingress controller
To enable the NGINX Ingress controller, run the following command:
minikube addons enable ingress
Verify that the NGINX Ingress controller is running
Make sure that you see a Pod with a name that starts with nginx-ingress-controller-.
Deploy a hello, world app
Create a Deployment using the following command:
kubectl create deployment web --image=gcr.io/google-samples/hello-app:1.0
The output should be:
deployment.apps/web created
Expose the Deployment:
kubectl expose deployment web --type=NodePort --port=8080
The output should be:
service/web exposed
Verify the Service is created and is available on a node port:
kubectl get service web
The output is similar to:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
web NodePort 10.104.133.249 <none> 8080:31637/TCP 12m
Visit the Service via NodePort:
minikube service web --url
The output is similar to:
http://172.17.0.15:31637
Note: Katacoda environment only: at the top of the terminal panel, click the plus sign, and then click Select port to view on Host 1. Enter the NodePort, in this case 31637, and then click Display Port.
You should see an IPv4 address in the ADDRESS column; for example:
NAME CLASS HOSTS ADDRESS PORTS AGE
example-ingress <none> hello-world.info 172.17.0.15 80 38s
Add the following line to the bottom of the /etc/hosts file on
your computer (you will need administrator access):
172.17.0.15 hello-world.info
Note: If you are running Minikube locally, use minikube ip to get the external IP. The IP address displayed within the ingress list will be the internal IP.
After you make this change, your web browser sends requests for
hello-world.info URLs to Minikube.
Verify that the Ingress controller is directing traffic:
4.10.10 - Communicate Between Containers in the Same Pod Using a Shared Volume
This page shows how to use a Volume to communicate between two Containers running
in the same Pod. See also how to allow processes to communicate by
sharing process namespace
between containers.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In this exercise, you create a Pod that runs two Containers. The two containers
share a Volume that they can use to communicate. Here is the configuration file
for the Pod:
apiVersion:v1kind:Podmetadata:name:two-containersspec:restartPolicy:Nevervolumes:- name:shared-dataemptyDir:{}containers:- name:nginx-containerimage:nginxvolumeMounts:- name:shared-datamountPath:/usr/share/nginx/html- name:debian-containerimage:debianvolumeMounts:- name:shared-datamountPath:/pod-datacommand:["/bin/sh"]args:["-c","echo Hello from the debian container > /pod-data/index.html"]
In the configuration file, you can see that the Pod has a Volume named
shared-data.
The first container listed in the configuration file runs an nginx server. The
mount path for the shared Volume is /usr/share/nginx/html.
The second container is based on the debian image, and has a mount path of
/pod-data. The second container runs the following command and then terminates.
echo Hello from the debian container > /pod-data/index.html
Notice that the second container writes the index.html file in the root
directory of the nginx server.
USER PID ... STAT START TIME COMMAND
root 1 ... Ss 21:12 0:00 nginx: master process nginx -g daemon off;
Recall that the debian Container created the index.html file in the nginx root
directory. Use curl to send a GET request to the nginx server:
root@two-containers:/# curl localhost
The output shows that nginx serves a web page written by the debian container:
Hello from the debian container
Discussion
The primary reason that Pods can have multiple containers is to support
helper applications that assist a primary application. Typical examples of
helper applications are data pullers, data pushers, and proxies.
Helper and primary applications often need to communicate with each other.
Typically this is done through a shared filesystem, as shown in this exercise,
or through the loopback network interface, localhost. An example of this pattern is a
web server along with a helper program that polls a Git repository for new updates.
The Volume in this exercise provides a way for Containers to communicate during
the life of the Pod. If the Pod is deleted and recreated, any data stored in
the shared Volume is lost.
Kubernetes offers a DNS cluster addon, which most of the supported environments enable by default. In Kubernetes version 1.11 and later, CoreDNS is recommended and is installed by default with kubeadm.
For more information on how to configure CoreDNS for a Kubernetes cluster, see the Customizing DNS Service. An example demonstrating how to use Kubernetes DNS with kube-dns, see the Kubernetes DNS sample plugin.
4.10.12 - Access Services Running on Clusters
This page shows how to connect to services running on the Kubernetes cluster.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
In Kubernetes, nodes,
pods and services all have
their own IPs. In many cases, the node IPs, pod IPs, and some service IPs on a cluster will not be
routable, so they will not be reachable from a machine outside the cluster,
such as your desktop machine.
Ways to connect
You have several options for connecting to nodes, pods and services from outside the cluster:
Access services through public IPs.
Use a service with type NodePort or LoadBalancer to make the service reachable outside
the cluster. See the services and
kubectl expose documentation.
Depending on your cluster environment, this may only expose the service to your corporate network,
or it may expose it to the internet. Think about whether the service being exposed is secure.
Does it do its own authentication?
Place pods behind services. To access one specific pod from a set of replicas, such as for debugging,
place a unique label on the pod and create a new service which selects this label.
In most cases, it should not be necessary for application developer to directly access
nodes via their nodeIPs.
Access services, nodes, or pods using the Proxy Verb.
Does apiserver authentication and authorization prior to accessing the remote service.
Use this if the services are not secure enough to expose to the internet, or to gain
access to ports on the node IP, or for debugging.
Proxies may cause problems for some web applications.
Run a pod, and then connect to a shell in it using kubectl exec.
Connect to other nodes, pods, and services from that shell.
Some clusters may allow you to ssh to a node in the cluster. From there you may be able to
access cluster services. This is a non-standard method, and will work on some clusters but
not others. Browsers and other tools may or may not be installed. Cluster DNS may not work.
Discovering builtin services
Typically, there are several services which are started on a cluster by kube-system. Get a list of these
with the kubectl cluster-info command:
kubectl cluster-info
The output is similar to this:
Kubernetes master is running at https://192.0.2.1
elasticsearch-logging is running at https://192.0.2.1/api/v1/namespaces/kube-system/services/elasticsearch-logging/proxy
kibana-logging is running at https://192.0.2.1/api/v1/namespaces/kube-system/services/kibana-logging/proxy
kube-dns is running at https://192.0.2.1/api/v1/namespaces/kube-system/services/kube-dns/proxy
grafana is running at https://192.0.2.1/api/v1/namespaces/kube-system/services/monitoring-grafana/proxy
heapster is running at https://192.0.2.1/api/v1/namespaces/kube-system/services/monitoring-heapster/proxy
This shows the proxy-verb URL for accessing each service.
For example, this cluster has cluster-level logging enabled (using Elasticsearch), which can be reached
at https://192.0.2.1/api/v1/namespaces/kube-system/services/elasticsearch-logging/proxy/ if suitable credentials are passed, or through a kubectl proxy at, for example:
http://localhost:8080/api/v1/namespaces/kube-system/services/elasticsearch-logging/proxy/.
As mentioned above, you use the kubectl cluster-info command to retrieve the service's proxy URL. To create proxy URLs that include service endpoints, suffixes, and parameters, you append to the service's proxy URL:
http://kubernetes_master_address/api/v1/namespaces/namespace_name/services/[https:]service_name[:port_name]/proxy
If you haven't specified a name for your port, you don't have to specify port_name in the URL. You can also use the port number in place of the port_name for both named and unnamed ports.
By default, the API server proxies to your service using HTTP. To use HTTPS, prefix the service name with https::
http://<kubernetes_master_address>/api/v1/namespaces/<namespace_name>/services/<service_name>/proxy
The supported formats for the <service_name> segment of the URL are:
<service_name> - proxies to the default or unnamed port using http
<service_name>:<port_name> - proxies to the specified port name or port number using http
https:<service_name>: - proxies to the default or unnamed port using https (note the trailing colon)
https:<service_name>:<port_name> - proxies to the specified port name or port number using https
Examples
To access the Elasticsearch service endpoint _search?q=user:kimchy, you would use:
Using web browsers to access services running on the cluster
You may be able to put an apiserver proxy URL into the address bar of a browser. However:
Web browsers cannot usually pass tokens, so you may need to use basic (password) auth. Apiserver can be configured to accept basic auth,
but your cluster may not be configured to accept basic auth.
Some web apps may not work, particularly those with client side javascript that construct URLs in a
way that is unaware of the proxy path prefix.
4.11 - Extend Kubernetes
Understand advanced ways to adapt your Kubernetes cluster to the needs of your work environment.
4.11.1 - Configure the Aggregation Layer
Configuring the aggregation layer allows the Kubernetes apiserver to be extended with additional APIs, which are not part of the core Kubernetes APIs.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Note: There are a few setup requirements for getting the aggregation layer working in your environment to support mutual TLS auth between the proxy and extension apiservers. Kubernetes and the kube-apiserver have multiple CAs, so make sure that the proxy is signed by the aggregation layer CA and not by something else, like the Kubernetes general CA.
Caution: Reusing the same CA for different client types can negatively impact the cluster's ability to function. For more information, see CA Reusage and Conflicts.
Authentication Flow
Unlike Custom Resource Definitions (CRDs), the Aggregation API involves another server - your Extension apiserver - in addition to the standard Kubernetes apiserver. The Kubernetes apiserver will need to communicate with your extension apiserver, and your extension apiserver will need to communicate with the Kubernetes apiserver. In order for this communication to be secured, the Kubernetes apiserver uses x509 certificates to authenticate itself to the extension apiserver.
This section describes how the authentication and authorization flows work, and how to configure them.
The high-level flow is as follows:
Kubernetes apiserver: authenticate the requesting user and authorize their rights to the requested API path.
Kubernetes apiserver: proxy the request to the extension apiserver
Extension apiserver: authenticate the request from the Kubernetes apiserver
Extension apiserver: authorize the request from the original user
Extension apiserver: execute
The rest of this section describes these steps in detail.
The flow can be seen in the following diagram.
.
The source for the above swimlanes can be found in the source of this document.
Kubernetes Apiserver Authentication and Authorization
A request to an API path that is served by an extension apiserver begins the same way as all API requests: communication to the Kubernetes apiserver. This path already has been registered with the Kubernetes apiserver by the extension apiserver.
The user communicates with the Kubernetes apiserver, requesting access to the path. The Kubernetes apiserver uses standard authentication and authorization configured with the Kubernetes apiserver to authenticate the user and authorize access to the specific path.
Everything to this point has been standard Kubernetes API requests, authentication and authorization.
The Kubernetes apiserver now is prepared to send the request to the extension apiserver.
Kubernetes Apiserver Proxies the Request
The Kubernetes apiserver now will send, or proxy, the request to the extension apiserver that registered to handle the request. In order to do so, it needs to know several things:
How should the Kubernetes apiserver authenticate to the extension apiserver, informing the extension apiserver that the request, which comes over the network, is coming from a valid Kubernetes apiserver?
How should the Kubernetes apiserver inform the extension apiserver of the username and group for which the original request was authenticated?
In order to provide for these two, you must configure the Kubernetes apiserver using several flags.
Kubernetes Apiserver Client Authentication
The Kubernetes apiserver connects to the extension apiserver over TLS, authenticating itself using a client certificate. You must provide the following to the Kubernetes apiserver upon startup, using the provided flags:
private key file via --proxy-client-key-file
signed client certificate file via --proxy-client-cert-file
certificate of the CA that signed the client certificate file via --requestheader-client-ca-file
valid Common Name values (CNs) in the signed client certificate via --requestheader-allowed-names
The Kubernetes apiserver will use the files indicated by --proxy-client-*-file to authenticate to the extension apiserver. In order for the request to be considered valid by a compliant extension apiserver, the following conditions must be met:
The connection must be made using a client certificate that is signed by the CA whose certificate is in --requestheader-client-ca-file.
The connection must be made using a client certificate whose CN is one of those listed in --requestheader-allowed-names.
Note: You can set this option to blank as --requestheader-allowed-names="". This will indicate to an extension apiserver that any CN is acceptable.
When started with these options, the Kubernetes apiserver will:
Use them to authenticate to the extension apiserver.
Create a configmap in the kube-system namespace called extension-apiserver-authentication, in which it will place the CA certificate and the allowed CNs. These in turn can be retrieved by extension apiservers to validate requests.
Note that the same client certificate is used by the Kubernetes apiserver to authenticate against all extension apiservers. It does not create a client certificate per extension apiserver, but rather a single one to authenticate as the Kubernetes apiserver. This same one is reused for all extension apiserver requests.
Original Request Username and Group
When the Kubernetes apiserver proxies the request to the extension apiserver, it informs the extension apiserver of the username and group with which the original request successfully authenticated. It provides these in http headers of its proxied request. You must inform the Kubernetes apiserver of the names of the headers to be used.
the header in which to store the username via --requestheader-username-headers
the header in which to store the group via --requestheader-group-headers
the prefix to append to all extra headers via --requestheader-extra-headers-prefix
These header names are also placed in the extension-apiserver-authentication configmap, so they can be retrieved and used by extension apiservers.
Extension Apiserver Authenticates the Request
The extension apiserver, upon receiving a proxied request from the Kubernetes apiserver, must validate that the request actually did come from a valid authenticating proxy, which role the Kubernetes apiserver is fulfilling. The extension apiserver validates it via:
Retrieve the following from the configmap in kube-system, as described above:
Client CA certificate
List of allowed names (CNs)
Header names for username, group and extra info
Check that the TLS connection was authenticated using a client certificate which:
Was signed by the CA whose certificate matches the retrieved CA certificate.
Has a CN in the list of allowed CNs, unless the list is blank, in which case all CNs are allowed.
Extract the username and group from the appropriate headers
If the above passes, then the request is a valid proxied request from a legitimate authenticating proxy, in this case the Kubernetes apiserver.
Note that it is the responsibility of the extension apiserver implementation to provide the above. Many do it by default, leveraging the k8s.io/apiserver/ package. Others may provide options to override it using command-line options.
In order to have permission to retrieve the configmap, an extension apiserver requires the appropriate role. There is a default role named extension-apiserver-authentication-reader in the kube-system namespace which can be assigned.
Extension Apiserver Authorizes the Request
The extension apiserver now can validate that the user/group retrieved from the headers are authorized to execute the given request. It does so by sending a standard SubjectAccessReview request to the Kubernetes apiserver.
In order for the extension apiserver to be authorized itself to submit the SubjectAccessReview request to the Kubernetes apiserver, it needs the correct permissions. Kubernetes includes a default ClusterRole named system:auth-delegator that has the appropriate permissions. It can be granted to the extension apiserver's service account.
Extension Apiserver Executes
If the SubjectAccessReview passes, the extension apiserver executes the request.
Enable Kubernetes Apiserver flags
Enable the aggregation layer via the following kube-apiserver flags. They may have already been taken care of by your provider.
--requestheader-client-ca-file=<path to aggregator CA cert>
--requestheader-allowed-names=front-proxy-client
--requestheader-extra-headers-prefix=X-Remote-Extra-
--requestheader-group-headers=X-Remote-Group
--requestheader-username-headers=X-Remote-User
--proxy-client-cert-file=<path to aggregator proxy cert>
--proxy-client-key-file=<path to aggregator proxy key>
CA Reusage and Conflicts
The Kubernetes apiserver has two client CA options:
--client-ca-file
--requestheader-client-ca-file
Each of these functions independently and can conflict with each other, if not used correctly.
--client-ca-file: When a request arrives to the Kubernetes apiserver, if this option is enabled, the Kubernetes apiserver checks the certificate of the request. If it is signed by one of the CA certificates in the file referenced by --client-ca-file, then the request is treated as a legitimate request, and the user is the value of the common name CN=, while the group is the organization O=. See the documentation on TLS authentication.
--requestheader-client-ca-file: When a request arrives to the Kubernetes apiserver, if this option is enabled, the Kubernetes apiserver checks the certificate of the request. If it is signed by one of the CA certificates in the file reference by --requestheader-client-ca-file, then the request is treated as a potentially legitimate request. The Kubernetes apiserver then checks if the common name CN= is one of the names in the list provided by --requestheader-allowed-names. If the name is allowed, the request is approved; if it is not, the request is not.
If both--client-ca-file and --requestheader-client-ca-file are provided, then the request first checks the --requestheader-client-ca-file CA and then the --client-ca-file. Normally, different CAs, either root CAs or intermediate CAs, are used for each of these options; regular client requests match against --client-ca-file, while aggregation requests match against --requestheader-client-ca-file. However, if both use the same CA, then client requests that normally would pass via --client-ca-file will fail, because the CA will match the CA in --requestheader-client-ca-file, but the common name CN= will not match one of the acceptable common names in --requestheader-allowed-names. This can cause your kubelets and other control plane components, as well as end-users, to be unable to authenticate to the Kubernetes apiserver.
For this reason, use different CA certs for the --client-ca-file option - to authorize control plane components and end-users - and the --requestheader-client-ca-file option - to authorize aggregation apiserver requests.
Warning: Do not reuse a CA that is used in a different context unless you understand the risks and the mechanisms to protect the CA's usage.
If you are not running kube-proxy on a host running the API server, then you must make sure that the system is enabled with the following kube-apiserver flag:
--enable-aggregator-routing=true
Register APIService objects
You can dynamically configure what client requests are proxied to extension
apiserver. The following is an example registration:
apiVersion:apiregistration.k8s.io/v1kind:APIServicemetadata:name:<name of the registration object>spec:group:<API group name this extension apiserver hosts>version:<API version this extension apiserver hosts>groupPriorityMinimum:<priority this APIService for this group, see API documentation>versionPriority:<prioritizes ordering of this version within a group, see API documentation>service:namespace:<namespace of the extension apiserver service>name:<name of the extension apiserver service>caBundle:<pem encoded ca cert that signs the server cert used by the webhook>
Once the Kubernetes apiserver has determined a request should be sent to an extension apiserver,
it needs to know how to contact it.
The service stanza is a reference to the service for an extension apiserver.
The service namespace and name are required. The port is optional and defaults to 443.
Here is an example of an extension apiserver that is configured to be called on port "1234",
and to verify the TLS connection against the ServerName
my-service-name.my-service-namespace.svc using a custom CA bundle.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version 1.16.
To check the version, enter kubectl version.
If you are using an older version of Kubernetes that is still supported, switch to
the documentation for that version to see advice that is relevant for your cluster.
Create a CustomResourceDefinition
When you create a new CustomResourceDefinition (CRD), the Kubernetes API Server
creates a new RESTful resource path for each version you specify. The custom
resource created from a CRD object can be either namespaced or cluster-scoped,
as specified in the CRD's spec.scope field. As with existing built-in
objects, deleting a namespace deletes all custom objects in that namespace.
CustomResourceDefinitions themselves are non-namespaced and are available to
all namespaces.
For example, if you save the following CustomResourceDefinition to resourcedefinition.yaml:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:# name must match the spec fields below, and be in the form: <plural>.<group>name:crontabs.stable.example.comspec:# group name to use for REST API: /apis/<group>/<version>group:stable.example.com# list of versions supported by this CustomResourceDefinitionversions:- name:v1# Each version can be enabled/disabled by Served flag.served:true# One and only one version must be marked as the storage version.storage:trueschema:openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringimage:type:stringreplicas:type:integer# either Namespaced or Clusterscope:Namespacednames:# plural name to be used in the URL: /apis/<group>/<version>/<plural>plural:crontabs# singular name to be used as an alias on the CLI and for displaysingular:crontab# kind is normally the CamelCased singular type. Your resource manifests use this.kind:CronTab# shortNames allow shorter string to match your resource on the CLIshortNames:- ct
and create it:
kubectl apply -f resourcedefinition.yaml
Then a new namespaced RESTful API endpoint is created at:
This endpoint URL can then be used to create and manage custom objects.
The kind of these objects will be CronTab from the spec of the
CustomResourceDefinition object you created above.
It might take a few seconds for the endpoint to be created.
You can watch the Established condition of your CustomResourceDefinition
to be true or watch the discovery information of the API server for your
resource to show up.
Create custom objects
After the CustomResourceDefinition object has been created, you can create
custom objects. Custom objects can contain custom fields. These fields can
contain arbitrary JSON.
In the following example, the cronSpec and image custom fields are set in a
custom object of kind CronTab. The kind CronTab comes from the spec of the
CustomResourceDefinition object you created above.
If you save the following YAML to my-crontab.yaml:
You can then manage your CronTab objects using kubectl. For example:
kubectl get crontab
Should print a list like this:
NAME AGE
my-new-cron-object 6s
Resource names are not case-sensitive when using kubectl, and you can use either
the singular or plural forms defined in the CRD, as well as any short names.
You can also view the raw YAML data:
kubectl get ct -o yaml
You should see that it contains the custom cronSpec and image fields
from the YAML you used to create it:
When you delete a CustomResourceDefinition, the server will uninstall the RESTful API endpoint
and delete all custom objects stored in it.
kubectl delete -f resourcedefinition.yaml
kubectl get crontabs
Error from server (NotFound): Unable to list {"stable.example.com" "v1" "crontabs"}: the server could not
find the requested resource (get crontabs.stable.example.com)
If you later recreate the same CustomResourceDefinition, it will start out empty.
Specifying a structural schema
CustomResources store structured data in custom fields (alongside the built-in
fields apiVersion, kind and metadata, which the API server validates
implicitly). With OpenAPI v3.0 validation a schema can be
specified, which is validated during creation and updates, compare below for
details and limits of such a schema.
With apiextensions.k8s.io/v1 the definition of a structural schema is
mandatory for CustomResourceDefinitions. In the beta version of
CustomResourceDefinition, the structural schema was optional.
specifies a non-empty type (via type in OpenAPI) for the root, for each specified field of an object node
(via properties or additionalProperties in OpenAPI) and for each item in an array node
(via items in OpenAPI), with the exception of:
a node with x-kubernetes-int-or-string: true
a node with x-kubernetes-preserve-unknown-fields: true
for each field in an object and each item in an array which is specified within any of allOf, anyOf,
oneOf or not, the schema also specifies the field/item outside of those logical junctors (compare example 1 and 2).
does not set description, type, default, additionalProperties, nullable within an allOf, anyOf,
oneOf or not, with the exception of the two pattern for x-kubernetes-int-or-string: true (see below).
if metadata is specified, then only restrictions on metadata.name and metadata.generateName are allowed.
Non-structural example 1:
allOf:- properties:foo:...
conflicts with rule 2. The following would be correct:
properties:foo:...allOf:- properties:foo:...
Non-structural example 2:
allOf:- items:properties:foo:...
conflicts with rule 2. The following would be correct:
properties:foo:pattern:"abc"metadata:type:objectproperties:name:type:stringpattern:"^a"finalizers:type:arrayitems:type:stringpattern:"my-finalizer"anyOf:- properties:bar:type:integerminimum:42required:["bar"]description:"foo bar object"
is not a structural schema because of the following violations:
the type at the root is missing (rule 1).
the type of foo is missing (rule 1).
bar inside of anyOf is not specified outside (rule 2).
bar's type is within anyOf (rule 3).
the description is set within anyOf (rule 3).
metadata.finalizers might not be restricted (rule 4).
In contrast, the following, corresponding schema is structural:
type:objectdescription:"foo bar object"properties:foo:type:stringpattern:"abc"bar:type:integermetadata:type:objectproperties:name:type:stringpattern:"^a"anyOf:- properties:bar:minimum:42required:["bar"]
Violations of the structural schema rules are reported in the NonStructural condition in the CustomResourceDefinition.
Field pruning
CustomResourceDefinitions store validated resource data in the cluster's persistence store, etcd.
As with native Kubernetes resources such as ConfigMap,
if you specify a field that the API server does not recognize, the unknown field is pruned (removed) before being persisted.
Note:
CRDs converted from apiextensions.k8s.io/v1beta1 to apiextensions.k8s.io/v1 might lack structural schemas,
and spec.preserveUnknownFields might be true.
For legacy CustomResourceDefinition objects created as
apiextensions.k8s.io/v1beta1 with spec.preserveUnknownFields set to
true, the following is also true:
Pruning is not enabled.
You can store arbitrary data.
For compatibility with apiextensions.k8s.io/v1, update your custom
resource definitions to:
Use a structural OpenAPI schema.
Set spec.preserveUnknownFields to false.
If you save the following YAML to my-crontab.yaml:
This example turned off client-side validation to demonstrate the API server's behavior, by adding the --validate=false command line option.
Because the OpenAPI validation schemas are also published
to clients, kubectl also checks for unknown fields and rejects those objects well before they would be sent to the API server.
Controlling pruning
By default, all unspecified fields for a custom resource, across all versions, are pruned. It is possible though to
opt-out of that for specifc sub-trees of fields by adding x-kubernetes-preserve-unknown-fields: true in the
structural OpenAPI v3 validation schema.
Also those nodes are partially excluded from rule 3 in the sense that the following two patterns are allowed
(exactly those, without variations in order to additional fields):
x-kubernetes-int-or-string:trueallOf:- anyOf:- type:integer- type:string- ...# zero or more...
With one of those specification, both an integer and a string validate.
In Validation Schema Publishing,
x-kubernetes-int-or-string: true is unfolded to one of the two patterns shown above.
RawExtension
RawExtensions (as in runtime.RawExtension defined in
k8s.io/apimachinery)
holds complete Kubernetes objects, i.e. with apiVersion and kind fields.
It is possible to specify those embedded objects (both completely without constraints or partially specified)
by setting x-kubernetes-embedded-resource: true. For example:
Here, the field foo holds a complete object, e.g.:
foo:apiVersion:v1kind:Podspec:...
Because x-kubernetes-preserve-unknown-fields: true is specified alongside, nothing is pruned.
The use of x-kubernetes-preserve-unknown-fields: true is optional though.
With x-kubernetes-embedded-resource: true, the apiVersion, kind and metadata are implicitly specified and validated.
Serving multiple versions of a CRD
See Custom resource definition versioning
for more information about serving multiple versions of your
CustomResourceDefinition and migrating your objects from one version to another.
Advanced topics
Finalizers
Finalizers allow controllers to implement asynchronous pre-delete hooks.
Custom objects support finalizers similar to built-in objects.
You can add a finalizer to a custom object like this:
Identifiers of custom finalizers consist of a domain name, a forward slash and the name of
the finalizer. Any controller can add a finalizer to any object's list of finalizers.
The first delete request on an object with finalizers sets a value for the
metadata.deletionTimestamp field but does not delete it. Once this value is set,
entries in the finalizers list can only be removed. While any finalizers remain it is also
impossible to force the deletion of an object.
When the metadata.deletionTimestamp field is set, controllers watching the object execute any
finalizers they handle and remove the finalizer from the list after they are done. It is the
responsibility of each controller to remove its finalizer from the list.
The value of metadata.deletionGracePeriodSeconds controls the interval between polling updates.
Once the list of finalizers is empty, meaning all finalizers have been executed, the resource is
deleted by Kubernetes.
The default field can be set when the Defaulting feature is enabled,
which is the case with apiextensions.k8s.io/v1 CustomResourceDefinitions.
Defaulting is in GA since 1.17 (beta since 1.16 with the CustomResourceDefaultingfeature gate
enabled, which is the case automatically for many clusters for beta features).
Refer to the structural schemas section for other
restrictions and CustomResourceDefinition features.
The schema is defined in the CustomResourceDefinition. In the following example, the
CustomResourceDefinition applies the following validations on the custom object:
spec.cronSpec must be a string and must be of the form described by the regular expression.
spec.replicas must be an integer and must have a minimum value of 1 and a maximum value of 10.
Save the CustomResourceDefinition to resourcedefinition.yaml:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:name:crontabs.stable.example.comspec:group:stable.example.comversions:- name:v1served:truestorage:trueschema:# openAPIV3Schema is the schema for validating custom objects.openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringpattern:'^(\d+|\*)(/\d+)?(\s+(\d+|\*)(/\d+)?){4}$'image:type:stringreplicas:type:integerminimum:1maximum:10scope:Namespacednames:plural:crontabssingular:crontabkind:CronTabshortNames:- ct
and create it:
kubectl apply -f resourcedefinition.yaml
A request to create a custom object of kind CronTab is rejected if there are invalid values in its fields.
In the following example, the custom object contains fields with invalid values:
spec.cronSpec does not match the regular expression.
spec.replicas is greater than 10.
If you save the following YAML to my-crontab.yaml:
The CronTab "my-new-cron-object" is invalid: []: Invalid value: map[string]interface {}{"apiVersion":"stable.example.com/v1", "kind":"CronTab", "metadata":map[string]interface {}{"name":"my-new-cron-object", "namespace":"default", "deletionTimestamp":interface {}(nil), "deletionGracePeriodSeconds":(*int64)(nil), "creationTimestamp":"2017-09-05T05:20:07Z", "uid":"e14d79e7-91f9-11e7-a598-f0761cb232d1", "clusterName":""}, "spec":map[string]interface {}{"cronSpec":"* * * *", "image":"my-awesome-cron-image", "replicas":15}}:
validation failure list:
spec.cronSpec in body should match '^(\d+|\*)(/\d+)?(\s+(\d+|\*)(/\d+)?){4}$'
spec.replicas in body should be less than or equal to 10
If the fields contain valid values, the object creation request is accepted.
kubectl apply -f my-crontab.yaml
crontab "my-new-cron-object" created
Validation rules
FEATURE STATE:Kubernetes v1.25 [beta]
Validation rules are in beta since 1.25 and the
CustomResourceValidationExpressionsfeature
gate is enabled by default to validate custom resource
based on validation rules. You can disable this feature by explicitly setting the
CustomResourceValidationExpressions feature gate to false,
for the [kube-apiserver](https://kubernetes.io/docs/reference/command-line-tools-reference/kube-apiserver/) component.
This feature is only available if the schema is a
structural schema.
Validation rules use the Common Expression Language (CEL)
to validate custom resource values. Validation rules are included in
CustomResourceDefinition schemas using the x-kubernetes-validations extension.
The Rule is scoped to the location of the x-kubernetes-validations extension in the schema.
And self variable in the CEL expression is bound to the scoped value.
All validation rules are scoped to the current object: no cross-object or stateful validation rules are supported.
For example:
...openAPIV3Schema:type:objectproperties:spec:type:objectx-kubernetes-validations:- rule:"self.minReplicas <= self.replicas"message:"replicas should be greater than or equal to minReplicas."- rule:"self.replicas <= self.maxReplicas"message:"replicas should be smaller than or equal to maxReplicas."properties:...minReplicas:type:integerreplicas:type:integermaxReplicas:type:integerrequired:- minReplicas- replicas- maxReplicas
will reject a request to create this custom resource:
The CronTab "my-new-cron-object" is invalid:
* spec: Invalid value: map[string]interface {}{"maxReplicas":10, "minReplicas":0, "replicas":20}: replicas should be smaller than or equal to maxReplicas.
x-kubernetes-validations could have multiple rules.
The rule under x-kubernetes-validations represents the expression which will be evaluated by CEL.
The message represents the message displayed when validation fails. If message is unset, the above response would be:
The CronTab "my-new-cron-object" is invalid:
* spec: Invalid value: map[string]interface {}{"maxReplicas":10, "minReplicas":0, "replicas":20}: failed rule: self.replicas <= self.maxReplicas
Validation rules are compiled when CRDs are created/updated.
The request of CRDs create/update will fail if compilation of validation rules fail.
Compilation process includes type checking as well.
The compilation failure:
no_matching_overload: this function has no overload for the types of the arguments.
e.g. Rule like self == true against a field of integer type will get error:
Invalid value: apiextensions.ValidationRule{Rule:"self == true", Message:""}: compilation failed: ERROR: \<input>:1:6: found no matching overload for '_==_' applied to '(int, bool)'
no_such_field: does not contain the desired field.
e.g. Rule like self.nonExistingField > 0 against a non-existing field will return the error:
If the Rule is scoped to the root of a resource, it may make field selection into any fields
declared in the OpenAPIv3 schema of the CRD as well as apiVersion, kind, metadata.name and
metadata.generateName. This includes selection of fields in both the spec and status in the
same expression:
If the Rule is scoped to an object with properties, the accessible properties of the object are field selectable
via self.field and field presence can be checked via has(self.field). Null valued fields are treated as
absent fields in CEL expressions.
If the Rule is scoped to an object with additionalProperties (i.e. a map) the value of the map
are accessible via self[mapKey], map containment can be checked via mapKey in self and all entries of the map
are accessible via CEL macros and functions such as self.all(...).
The apiVersion, kind, metadata.name and metadata.generateName are always accessible from the root of the
object and from any x-kubernetes-embedded-resource annotated objects. No other metadata properties are accessible.
Unknown data preserved in custom resources via x-kubernetes-preserve-unknown-fields is not accessible in CEL
expressions. This includes:
Unknown field values that are preserved by object schemas with x-kubernetes-preserve-unknown-fields.
Object properties where the property schema is of an "unknown type". An "unknown type" is recursively defined as:
A schema with no type and x-kubernetes-preserve-unknown-fields set to true
An array where the items schema is of an "unknown type"
An object where the additionalProperties schema is of an "unknown type"
Only property names of the form [a-zA-Z_.-/][a-zA-Z0-9_.-/]* are accessible.
Accessible property names are escaped according to the following rules when accessed in the expression:
Note: CEL RESERVED keyword needs to match the exact property name to be escaped (e.g. int in the word sprint would not be escaped).
Examples on escaping:
property name
rule with escaped property name
namespace
self.__namespace__ > 0
x-prop
self.x__dash__prop > 0
redact__d
self.redact__underscores__d > 0
string
self.startsWith('kube')
Equality on arrays with x-kubernetes-list-type of set or map ignores element order, i.e. [1, 2] == [2, 1].
Concatenation on arrays with x-kubernetes-list-type use the semantics of the list type:
set: X + Y performs a union where the array positions of all elements in X are preserved and
non-intersecting elements in Y are appended, retaining their partial order.
map: X + Y performs a merge where the array positions of all keys in X are preserved but the values
are overwritten by values in Y when the key sets of X and Y intersect. Elements in Y with
non-intersecting keys are appended, retaining their partial order.
Here is the declarations type mapping between OpenAPIv3 and CEL type:
OpenAPIv3 type
CEL type
'object' with Properties
object / "message type"
'object' with AdditionalProperties
map
'object' with x-kubernetes-embedded-type
object / "message type", 'apiVersion', 'kind', 'metadata.name' and 'metadata.generateName' are implicitly included in schema
'object' with x-kubernetes-preserve-unknown-fields
object / "message type", unknown fields are NOT accessible in CEL expression
x-kubernetes-int-or-string
dynamic object that is either an int or a string, type(value) can be used to check the type
'array
list
'array' with x-kubernetes-list-type=map
list with map based Equality & unique key guarantees
'array' with x-kubernetes-list-type=set
list with set based Equality & unique entry guarantees
A rule that contains an expression referencing the identifier oldSelf is implicitly considered a
transition rule. Transition rules allow schema authors to prevent certain transitions between two
otherwise valid states. For example:
Unlike other rules, transition rules apply only to operations meeting the following criteria:
The operation updates an existing object. Transition rules never apply to create operations.
Both an old and a new value exist. It remains possible to check if a value has been added or
removed by placing a transition rule on the parent node. Transition rules are never applied to
custom resource creation. When placed on an optional field, a transition rule will not apply to
update operations that set or unset the field.
The path to the schema node being validated by a transition rule must resolve to a node that is
comparable between the old object and the new object. For example, list items and their
descendants (spec.foo[10].bar) can't necessarily be correlated between an existing object and a
later update to the same object.
Errors will be generated on CRD writes if a schema node contains a transition rule that can never be
applied, e.g. "path: update rule rule cannot be set on schema because the schema or its parent
schema is not mergeable".
Transition rules are only allowed on correlatable portions of a schema.
A portion of the schema is correlatable if all array parent schemas are of type x-kubernetes-list-type=map;
any setor atomicarray parent schemas make it impossible to unambiguously correlate a self with oldSelf.
If previous value was X, new value can only be A or B, not Y or Z
oldSelf != 'X' || self in ['A', 'B']
Monotonic (non-decreasing) counters
self >= oldSelf
Resource use by validation functions
When you create or update a CustomResourceDefinition that uses validation rules,
the API server checks the likely impact of running those validation rules. If a rule is
estimated to be prohibitively expensive to execute, the API server rejects the create
or update operation, and returns an error message.
A similar system is used at runtime that observes the actions the interpreter takes. If the interpreter executes
too many instructions, execution of the rule will be halted, and an error will result.
Each CustomResourceDefinition is also allowed a certain amount of resources to finish executing all of
its validation rules. If the sum total of its rules are estimated at creation time to go over that limit,
then a validation error will also occur.
You are unlikely to encounter issues with the resource budget for validation if you only
specify rules that always take the same amount of time regardless of how large their input is.
For example, a rule that asserts that self.foo == 1 does not by itself have any
risk of rejection on validation resource budget groups.
But if foo is a string and you define a validation rule self.foo.contains("someString"), that rule takes
longer to execute depending on how long foo is.
Another example would be if foo were an array, and you specified a validation rule self.foo.all(x, x > 5).
The cost system always assumes the worst-case scenario if a limit on the length of foo is not
given, and this will happen for anything that can be iterated over (lists, maps, etc.).
Because of this, it is considered best practice to put a limit via maxItems, maxProperties, and
maxLength for anything that will be processed in a validation rule in order to prevent validation
errors during cost estimation. For example, given this schema with one rule:
then the API server rejects this rule on validation budget grounds with error:
spec.validation.openAPIV3Schema.properties[spec].properties[foo].x-kubernetes-validations[0].rule: Forbidden:
CEL rule exceeded budget by more than 100x (try simplifying the rule, or adding maxItems, maxProperties, and
maxLength where arrays, maps, and strings are used)
The rejection happens because self.all implies calling contains() on every string in foo,
which in turn will check the given string to see if it contains 'a string'. Without limits, this is a very
expensive rule.
If you do not specify any validation limit, the estimated cost of this rule will exceed the per-rule cost limit. But if you
add limits in the appropriate places, the rule will be allowed:
The cost estimation system takes into account how many times the rule will be executed in addition to the
estimated cost of the rule itself. For instance, the following rule will have the same estimated cost as the
previous example (despite the rule now being defined on the individual array items):
If a list inside of a list has a validation rule that uses self.all, that is significantly more expensive
than a non-nested list with the same rule. A rule that would have been allowed on a non-nested list might need
lower limits set on both nested lists in order to be allowed. For example, even without having limits set,
the following rule is allowed:
openAPIV3Schema:type:objectproperties:foo:type:arrayitems:type:integerx-kubernetes-validations:- rule:"self.all(x, x == 5)"
But the same rule on the following schema (with a nested array added) produces a validation error:
openAPIV3Schema:type:objectproperties:foo:type:arrayitems:type:arrayitems:type:integerx-kubernetes-validations:- rule:"self.all(x, x == 5)"
This is because each item of foo is itself an array, and each subarray in turn calls self.all. Avoid nested
lists and maps if possible where validation rules are used.
Defaulting
Note: To use defaulting, your CustomResourceDefinition must use API version apiextensions.k8s.io/v1.
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:name:crontabs.stable.example.comspec:group:stable.example.comversions:- name:v1served:truestorage:trueschema:# openAPIV3Schema is the schema for validating custom objects.openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringpattern:'^(\d+|\*)(/\d+)?(\s+(\d+|\*)(/\d+)?){4}$'default:"5 0 * * *"image:type:stringreplicas:type:integerminimum:1maximum:10default:1scope:Namespacednames:plural:crontabssingular:crontabkind:CronTabshortNames:- ct
With this both cronSpec and replicas are defaulted:
in the request to the API server using the request version defaults,
when reading from etcd using the storage version defaults,
after mutating admission plugins with non-empty patches using the admission webhook object version defaults.
Defaults applied when reading data from etcd are not automatically written back to etcd.
An update request via the API is required to persist those defaults back into etcd.
Default values must be pruned (with the exception of defaults for metadata fields) and must validate against a provided schema.
Default values for metadata fields of x-kubernetes-embedded-resources: true nodes (or parts of
a default value covering metadata) are not pruned during CustomResourceDefinition creation, but
through the pruning step during handling of requests.
Defaulting and Nullable
New in 1.20: null values for fields that either don't specify the nullable flag, or give it a
false value, will be pruned before defaulting happens. If a default is present, it will be
applied. When nullable is true, null values will be conserved and won't be defaulted.
creating an object with null values for foo and bar and baz
spec:foo:nullbar:nullbaz:null
leads to
spec:foo:"default"bar:null
with foo pruned and defaulted because the field is non-nullable, bar maintaining the null
value due to nullable: true, and baz pruned because the field is non-nullable and has no
default.
The kubectl command-line tool consumes the published schema to perform
client-side validation (kubectl create and kubectl apply), schema explanation (kubectl explain)
on custom resources. The published schema can be consumed for other purposes as well, like client generation or documentation.
The OpenAPI v3 validation schema is converted to OpenAPI v2 schema, and
show up in definitions and paths fields in the OpenAPI v2 spec.
The following modifications are applied during the conversion to keep backwards compatibility with
kubectl in previous 1.13 version. These modifications prevent kubectl from being over-strict and rejecting
valid OpenAPI schemas that it doesn't understand. The conversion won't modify the validation schema defined in CRD,
and therefore won't affect validation in the API server.
The following fields are removed as they aren't supported by OpenAPI v2
(in future versions OpenAPI v3 will be used without these restrictions)
The fields allOf, anyOf, oneOf and not are removed
If nullable: true is set, we drop type, nullable, items and properties because OpenAPI v2 is
not able to express nullable. To avoid kubectl to reject good objects, this is necessary.
Additional printer columns
The kubectl tool relies on server-side output formatting. Your cluster's API server decides which
columns are shown by the kubectl get command. You can customize these columns for a
CustomResourceDefinition. The following example adds the Spec, Replicas, and Age
columns.
Save the CustomResourceDefinition to resourcedefinition.yaml:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:name:crontabs.stable.example.comspec:group:stable.example.comscope:Namespacednames:plural:crontabssingular:crontabkind:CronTabshortNames:- ctversions:- name:v1served:truestorage:trueschema:openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringimage:type:stringreplicas:type:integeradditionalPrinterColumns:- name:Spectype:stringdescription:The cron spec defining the interval a CronJob is runjsonPath:.spec.cronSpec- name:Replicastype:integerdescription:The number of jobs launched by the CronJobjsonPath:.spec.replicas- name:Agetype:datejsonPath:.metadata.creationTimestamp
Create the CustomResourceDefinition:
kubectl apply -f resourcedefinition.yaml
Create an instance using the my-crontab.yaml from the previous section.
Invoke the server-side printing:
kubectl get crontab my-new-cron-object
Notice the NAME, SPEC, REPLICAS, and AGE columns in the output:
NAME SPEC REPLICAS AGE
my-new-cron-object * * * * * 1 7s
Note: The NAME column is implicit and does not need to be defined in the CustomResourceDefinition.
Priority
Each column includes a priority field. Currently, the priority
differentiates between columns shown in standard view or wide view (using the -o wide flag).
Columns with priority 0 are shown in standard view.
Columns with priority greater than 0 are shown only in wide view.
date – rendered differentially as time since this timestamp.
If the value inside a CustomResource does not match the type specified for the column,
the value is omitted. Use CustomResource validation to ensure that the value
types are correct.
Format
A column's format field can be any of the following:
int32
int64
float
double
byte
date
date-time
password
The column's format controls the style used when kubectl prints the value.
Subresources
Custom resources support /status and /scale subresources.
The status and scale subresources can be optionally enabled by
defining them in the CustomResourceDefinition.
Status subresource
When the status subresource is enabled, the /status subresource for the custom resource is exposed.
The status and the spec stanzas are represented by the .status and .spec JSONPaths respectively inside of a custom resource.
PUT requests to the /status subresource take a custom resource object and ignore changes to anything except the status stanza.
PUT requests to the /status subresource only validate the status stanza of the custom resource.
PUT/POST/PATCH requests to the custom resource ignore changes to the status stanza.
The .metadata.generation value is incremented for all changes, except for changes to .metadata or .status.
Only the following constructs are allowed at the root of the CRD OpenAPI validation schema:
description
example
exclusiveMaximum
exclusiveMinimum
externalDocs
format
items
maximum
maxItems
maxLength
minimum
minItems
minLength
multipleOf
pattern
properties
required
title
type
uniqueItems
Scale subresource
When the scale subresource is enabled, the /scale subresource for the custom resource is exposed.
The autoscaling/v1.Scale object is sent as the payload for /scale.
To enable the scale subresource, the following fields are defined in the CustomResourceDefinition.
specReplicasPath defines the JSONPath inside of a custom resource that corresponds to scale.spec.replicas.
It is a required value.
Only JSONPaths under .spec and with the dot notation are allowed.
If there is no value under the specReplicasPath in the custom resource,
the /scale subresource will return an error on GET.
statusReplicasPath defines the JSONPath inside of a custom resource that corresponds to scale.status.replicas.
It is a required value.
Only JSONPaths under .status and with the dot notation are allowed.
If there is no value under the statusReplicasPath in the custom resource,
the status replica value in the /scale subresource will default to 0.
labelSelectorPath defines the JSONPath inside of a custom resource that corresponds to Scale.Status.Selector.
It is an optional value.
It must be set to work with HPA.
Only JSONPaths under .status or .spec and with the dot notation are allowed.
If there is no value under the labelSelectorPath in the custom resource,
the status selector value in the /scale subresource will default to the empty string.
The field pointed by this JSON path must be a string field (not a complex selector struct)
which contains a serialized label selector in string form.
In the following example, both status and scale subresources are enabled.
Save the CustomResourceDefinition to resourcedefinition.yaml:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:name:crontabs.stable.example.comspec:group:stable.example.comversions:- name:v1served:truestorage:trueschema:openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringimage:type:stringreplicas:type:integerstatus:type:objectproperties:replicas:type:integerlabelSelector:type:string# subresources describes the subresources for custom resources.subresources:# status enables the status subresource.status:{}# scale enables the scale subresource.scale:# specReplicasPath defines the JSONPath inside of a custom resource that corresponds to Scale.Spec.Replicas.specReplicasPath:.spec.replicas# statusReplicasPath defines the JSONPath inside of a custom resource that corresponds to Scale.Status.Replicas.statusReplicasPath:.status.replicas# labelSelectorPath defines the JSONPath inside of a custom resource that corresponds to Scale.Status.Selector.labelSelectorPath:.status.labelSelectorscope:Namespacednames:plural:crontabssingular:crontabkind:CronTabshortNames:- ct
And create it:
kubectl apply -f resourcedefinition.yaml
After the CustomResourceDefinition object has been created, you can create custom objects.
If you save the following YAML to my-crontab.yaml:
A custom resource can be scaled using the kubectl scale command.
For example, the following command sets .spec.replicas of the
custom resource created above to 5:
You can use a PodDisruptionBudget to protect custom
resources that have the scale subresource enabled.
Categories
Categories is a list of grouped resources the custom resource belongs to (eg. all).
You can use kubectl get <category-name> to list the resources belonging to the category.
The following example adds all in the list of categories in the CustomResourceDefinition
and illustrates how to output the custom resource using kubectl get all.
Save the following CustomResourceDefinition to resourcedefinition.yaml:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:name:crontabs.stable.example.comspec:group:stable.example.comversions:- name:v1served:truestorage:trueschema:openAPIV3Schema:type:objectproperties:spec:type:objectproperties:cronSpec:type:stringimage:type:stringreplicas:type:integerscope:Namespacednames:plural:crontabssingular:crontabkind:CronTabshortNames:- ct# categories is a list of grouped resources the custom resource belongs to.categories:- all
and create it:
kubectl apply -f resourcedefinition.yaml
After the CustomResourceDefinition object has been created, you can create custom objects.
This page explains how to add versioning information to
CustomResourceDefinitions, to indicate the stability
level of your CustomResourceDefinitions or advance your API to a new version with conversion between API representations. It also describes how to upgrade an object from one version to another.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.16.
To check the version, enter kubectl version.
Overview
The CustomResourceDefinition API provides a workflow for introducing and upgrading
to new versions of a CustomResourceDefinition.
When a CustomResourceDefinition is created, the first version is set in the
CustomResourceDefinition spec.versions list to an appropriate stability level
and a version number. For example v1beta1 would indicate that the first
version is not yet stable. All custom resource objects will initially be stored
at this version.
Once the CustomResourceDefinition is created, clients may begin using the
v1beta1 API.
Later it might be necessary to add new version such as v1.
Adding a new version:
Pick a conversion strategy. Since custom resource objects need to be able to
be served at both versions, that means they will sometimes be served at a
different version than their storage version. In order for this to be
possible, the custom resource objects must sometimes be converted between the
version they are stored at and the version they are served at. If the
conversion involves schema changes and requires custom logic, a conversion
webhook should be used. If there are no schema changes, the default None
conversion strategy may be used and only the apiVersion field will be
modified when serving different versions.
If using conversion webhooks, create and deploy the conversion webhook. See
the Webhook conversion for more details.
Update the CustomResourceDefinition to include the new version in the
spec.versions list with served:true. Also, set spec.conversion field
to the selected conversion strategy. If using a conversion webhook, configure
spec.conversion.webhookClientConfig field to call the webhook.
Once the new version is added, clients may incrementally migrate to the new
version. It is perfectly safe for some clients to use the old version while
others use the new version.
It is safe for clients to use both the old and new version before, during and
after upgrading the objects to a new stored version.
Removing an old version:
Ensure all clients are fully migrated to the new version. The kube-apiserver
logs can be reviewed to help identify any clients that are still accessing via
the old version.
Set served to false for the old version in the spec.versions list. If
any clients are still unexpectedly using the old version they may begin reporting
errors attempting to access the custom resource objects at the old version.
If this occurs, switch back to using served:true on the old version, migrate the
remaining clients to the new version and repeat this step.
Verify that the storage is set to true for the new version in the spec.versions list in the CustomResourceDefinition.
Verify that the old version is no longer listed in the CustomResourceDefinition status.storedVersions.
Remove the old version from the CustomResourceDefinition spec.versions list.
Drop conversion support for the old version in conversion webhooks.
Specify multiple versions
The CustomResourceDefinition API versions field can be used to support multiple versions of custom resources that you
have developed. Versions can have different schemas, and conversion webhooks can convert custom resources between versions.
Webhook conversions should follow the Kubernetes API conventions wherever applicable.
Specifically, See the API change documentation for a set of useful gotchas and suggestions.
Note: In apiextensions.k8s.io/v1beta1, there was a version field instead of versions. The
version field is deprecated and optional, but if it is not empty, it must
match the first item in the versions field.
This example shows a CustomResourceDefinition with two versions. For the first
example, the assumption is all versions share the same schema with no conversion
between them. The comments in the YAML provide more context.
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:# name must match the spec fields below, and be in the form: <plural>.<group>name:crontabs.example.comspec:# group name to use for REST API: /apis/<group>/<version>group:example.com# list of versions supported by this CustomResourceDefinitionversions:- name:v1beta1# Each version can be enabled/disabled by Served flag.served:true# One and only one version must be marked as the storage version.storage:true# A schema is requiredschema:openAPIV3Schema:type:objectproperties:host:type:stringport:type:string- name:v1served:truestorage:falseschema:openAPIV3Schema:type:objectproperties:host:type:stringport:type:string# The conversion section is introduced in Kubernetes 1.13+ with a default value of# None conversion (strategy sub-field set to None).conversion:# None conversion assumes the same schema for all versions and only sets the apiVersion# field of custom resources to the proper valuestrategy:None# either Namespaced or Clusterscope:Namespacednames:# plural name to be used in the URL: /apis/<group>/<version>/<plural>plural:crontabs# singular name to be used as an alias on the CLI and for displaysingular:crontab# kind is normally the CamelCased singular type. Your resource manifests use this.kind:CronTab# shortNames allow shorter string to match your resource on the CLIshortNames:- ct
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinitionmetadata:# name must match the spec fields below, and be in the form: <plural>.<group>name:crontabs.example.comspec:# group name to use for REST API: /apis/<group>/<version>group:example.com# list of versions supported by this CustomResourceDefinitionversions:- name:v1beta1# Each version can be enabled/disabled by Served flag.served:true# One and only one version must be marked as the storage version.storage:true- name:v1served:truestorage:falsevalidation:openAPIV3Schema:type:objectproperties:host:type:stringport:type:string# The conversion section is introduced in Kubernetes 1.13+ with a default value of# None conversion (strategy sub-field set to None).conversion:# None conversion assumes the same schema for all versions and only sets the apiVersion# field of custom resources to the proper valuestrategy:None# either Namespaced or Clusterscope:Namespacednames:# plural name to be used in the URL: /apis/<group>/<version>/<plural>plural:crontabs# singular name to be used as an alias on the CLI and for displaysingular:crontab# kind is normally the PascalCased singular type. Your resource manifests use this.kind:CronTab# shortNames allow shorter string to match your resource on the CLIshortNames:- ct
You can save the CustomResourceDefinition in a YAML file, then use
kubectl apply to create it.
kubectl apply -f my-versioned-crontab.yaml
After creation, the API server starts to serve each enabled version at an HTTP
REST endpoint. In the above example, the API versions are available at
/apis/example.com/v1beta1 and /apis/example.com/v1.
Version priority
Regardless of the order in which versions are defined in a
CustomResourceDefinition, the version with the highest priority is used by
kubectl as the default version to access objects. The priority is determined
by parsing the name field to determine the version number, the stability
(GA, Beta, or Alpha), and the sequence within that stability level.
The algorithm used for sorting the versions is designed to sort versions in the
same way that the Kubernetes project sorts Kubernetes versions. Versions start with a
v followed by a number, an optional beta or alpha designation, and
optional additional numeric versioning information. Broadly, a version string might look
like v2 or v2beta1. Versions are sorted using the following algorithm:
Entries that follow Kubernetes version patterns are sorted before those that
do not.
For entries that follow Kubernetes version patterns, the numeric portions of
the version string is sorted largest to smallest.
If the strings beta or alpha follow the first numeric portion, they sorted
in that order, after the equivalent string without the beta or alpha
suffix (which is presumed to be the GA version).
If another number follows the beta, or alpha, those numbers are also
sorted from largest to smallest.
Strings that don't fit the above format are sorted alphabetically and the
numeric portions are not treated specially. Notice that in the example below,
foo1 is sorted above foo10. This is different from the sorting of the
numeric portion of entries that do follow the Kubernetes version patterns.
This might make sense if you look at the following sorted version list:
For the example in Specify multiple versions, the
version sort order is v1, followed by v1beta1. This causes the kubectl
command to use v1 as the default version unless the provided object specifies
the version.
Version deprecation
FEATURE STATE:Kubernetes v1.19 [stable]
Starting in v1.19, a CustomResourceDefinition can indicate a particular version of the resource it defines is deprecated.
When API requests to a deprecated version of that resource are made, a warning message is returned in the API response as a header.
The warning message for each deprecated version of the resource can be customized if desired.
A customized warning message should indicate the deprecated API group, version, and kind,
and should indicate what API group, version, and kind should be used instead, if applicable.
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionname:crontabs.example.comspec:group:example.comnames:plural:crontabssingular:crontabkind:CronTabscope:Namespacedversions:- name:v1alpha1served:true# This indicates the v1alpha1 version of the custom resource is deprecated.# API requests to this version receive a warning header in the server response.deprecated:true# This overrides the default warning returned to API clients making v1alpha1 API requests.deprecationWarning:"example.com/v1alpha1 CronTab is deprecated; see http://example.com/v1alpha1-v1 for instructions to migrate to example.com/v1 CronTab"schema:...- name:v1beta1served:true# This indicates the v1beta1 version of the custom resource is deprecated.# API requests to this version receive a warning header in the server response.# A default warning message is returned for this version.deprecated:trueschema:...- name:v1served:truestorage:trueschema:...
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinitionmetadata:name:crontabs.example.comspec:group:example.comnames:plural:crontabssingular:crontabkind:CronTabscope:Namespacedvalidation:...versions:- name:v1alpha1served:true# This indicates the v1alpha1 version of the custom resource is deprecated.# API requests to this version receive a warning header in the server response.deprecated:true# This overrides the default warning returned to API clients making v1alpha1 API requests.deprecationWarning:"example.com/v1alpha1 CronTab is deprecated; see http://example.com/v1alpha1-v1 for instructions to migrate to example.com/v1 CronTab"- name:v1beta1served:true# This indicates the v1beta1 version of the custom resource is deprecated.# API requests to this version receive a warning header in the server response.# A default warning message is returned for this version.deprecated:true- name:v1served:truestorage:true
Version removal
An older API version cannot be dropped from a CustomResourceDefinition manifest until existing persisted data has been migrated to the newer API version for all clusters that served the older version of the custom resource, and the old version is removed from the status.storedVersions of the CustomResourceDefinition.
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionname:crontabs.example.comspec:group:example.comnames:plural:crontabssingular:crontabkind:CronTabscope:Namespacedversions:- name:v1beta1# This indicates the v1beta1 version of the custom resource is no longer served.# API requests to this version receive a not found error in the server response.served:falseschema:...- name:v1served:true# The new served version should be set as the storage versionstorage:trueschema:...
Webhook conversion
FEATURE STATE:Kubernetes v1.16 [stable]
Note: Webhook conversion is available as beta since 1.15, and as alpha since Kubernetes 1.13. The
CustomResourceWebhookConversion feature must be enabled, which is the case automatically for many clusters for beta features. Please refer to the feature gate documentation for more information.
The above example has a None conversion between versions which only sets the apiVersion field
on conversion and does not change the rest of the object. The API server also supports webhook
conversions that call an external service in case a conversion is required. For example when:
custom resource is requested in a different version than stored version.
Watch is created in one version but the changed object is stored in another version.
custom resource PUT request is in a different version than storage version.
To cover all of these cases and to optimize conversion by the API server,
the conversion requests may contain multiple objects in order to minimize the external calls.
The webhook should perform these conversions independently.
Write a conversion webhook server
Please refer to the implementation of the custom resource conversion webhook
server
that is validated in a Kubernetes e2e test. The webhook handles the
ConversionReview requests sent by the API servers, and sends back conversion
results wrapped in ConversionResponse. Note that the request
contains a list of custom resources that need to be converted independently without
changing the order of objects.
The example server is organized in a way to be reused for other conversions.
Most of the common code are located in the
framework file
that leaves only
one function
to be implemented for different conversions.
Note: The example conversion webhook server leaves the ClientAuth field
empty,
which defaults to NoClientCert. This means that the webhook server does not
authenticate the identity of the clients, supposedly API servers. If you need
mutual TLS or other ways to authenticate the clients, see
how to authenticate API servers.
Permissible mutations
A conversion webhook must not mutate anything inside of metadata of the converted object
other than labels and annotations.
Attempted changes to name, UID and namespace are rejected and fail the request
which caused the conversion. All other changes are ignored.
Deploy the conversion webhook service
Documentation for deploying the conversion webhook is the same as for the
admission webhook example service.
The assumption for next sections is that the conversion webhook server is deployed to a service
named example-conversion-webhook-server in default namespace and serving traffic on path /crdconvert.
Note: When the webhook server is deployed into the Kubernetes cluster as a
service, it has to be exposed via a service on port 443 (The server
itself can have an arbitrary port but the service object should map it to port 443).
The communication between the API server and the webhook service may fail
if a different port is used for the service.
Configure CustomResourceDefinition to use conversion webhooks
The None conversion example can be extended to use the conversion webhook by modifying conversion
section of the spec:
apiVersion:apiextensions.k8s.io/v1kind:CustomResourceDefinitionmetadata:# name must match the spec fields below, and be in the form: <plural>.<group>name:crontabs.example.comspec:# group name to use for REST API: /apis/<group>/<version>group:example.com# list of versions supported by this CustomResourceDefinitionversions:- name:v1beta1# Each version can be enabled/disabled by Served flag.served:true# One and only one version must be marked as the storage version.storage:true# Each version can define its own schema when there is no top-level# schema is defined.schema:openAPIV3Schema:type:objectproperties:hostPort:type:string- name:v1served:truestorage:falseschema:openAPIV3Schema:type:objectproperties:host:type:stringport:type:stringconversion:# a Webhook strategy instruct API server to call an external webhook for any conversion between custom resources.strategy:Webhook# webhook is required when strategy is `Webhook` and it configures the webhook endpoint to be called by API server.webhook:# conversionReviewVersions indicates what ConversionReview versions are understood/preferred by the webhook.# The first version in the list understood by the API server is sent to the webhook.# The webhook must respond with a ConversionReview object in the same version it received.conversionReviewVersions:["v1","v1beta1"]clientConfig:service:namespace:defaultname:example-conversion-webhook-serverpath:/crdconvertcaBundle:"Ci0tLS0tQk...<base64-encoded PEM bundle>...tLS0K"# either Namespaced or Clusterscope:Namespacednames:# plural name to be used in the URL: /apis/<group>/<version>/<plural>plural:crontabs# singular name to be used as an alias on the CLI and for displaysingular:crontab# kind is normally the CamelCased singular type. Your resource manifests use this.kind:CronTab# shortNames allow shorter string to match your resource on the CLIshortNames:- ct
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinitionmetadata:# name must match the spec fields below, and be in the form: <plural>.<group>name:crontabs.example.comspec:# group name to use for REST API: /apis/<group>/<version>group:example.com# prunes object fields that are not specified in OpenAPI schemas below.preserveUnknownFields:false# list of versions supported by this CustomResourceDefinitionversions:- name:v1beta1# Each version can be enabled/disabled by Served flag.served:true# One and only one version must be marked as the storage version.storage:true# Each version can define its own schema when there is no top-level# schema is defined.schema:openAPIV3Schema:type:objectproperties:hostPort:type:string- name:v1served:truestorage:falseschema:openAPIV3Schema:type:objectproperties:host:type:stringport:type:stringconversion:# a Webhook strategy instruct API server to call an external webhook for any conversion between custom resources.strategy:Webhook# webhookClientConfig is required when strategy is `Webhook` and it configures the webhook endpoint to be called by API server.webhookClientConfig:service:namespace:defaultname:example-conversion-webhook-serverpath:/crdconvertcaBundle:"Ci0tLS0tQk...<base64-encoded PEM bundle>...tLS0K"# either Namespaced or Clusterscope:Namespacednames:# plural name to be used in the URL: /apis/<group>/<version>/<plural>plural:crontabs# singular name to be used as an alias on the CLI and for displaysingular:crontab# kind is normally the CamelCased singular type. Your resource manifests use this.kind:CronTab# shortNames allow shorter string to match your resource on the CLIshortNames:- ct
You can save the CustomResourceDefinition in a YAML file, then use
kubectl apply to apply it.
Make sure the conversion service is up and running before applying new changes.
Contacting the webhook
Once the API server has determined a request should be sent to a conversion webhook,
it needs to know how to contact the webhook. This is specified in the webhookClientConfig
stanza of the webhook configuration.
Conversion webhooks can either be called via a URL or a service reference,
and can optionally include a custom CA bundle to use to verify the TLS connection.
URL
url gives the location of the webhook, in standard URL form
(scheme://host:port/path).
The host should not refer to a service running in the cluster; use
a service reference by specifying the service field instead.
The host might be resolved via external DNS in some apiservers
(i.e., kube-apiserver cannot resolve in-cluster DNS as that would
be a layering violation). host may also be an IP address.
Please note that using localhost or 127.0.0.1 as a host is
risky unless you take great care to run this webhook on all hosts
which run an apiserver which might need to make calls to this
webhook. Such installations are likely to be non-portable or not readily run in a new cluster.
The scheme must be "https"; the URL must begin with "https://".
Attempting to use a user or basic auth (for example "user:password@") is not allowed.
Fragments ("#...") and query parameters ("?...") are also not allowed.
Here is an example of a conversion webhook configured to call a URL
(and expects the TLS certificate to be verified using system trust roots, so does not specify a caBundle):
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinition...spec:...conversion:strategy:WebhookwebhookClientConfig:url:"https://my-webhook.example.com:9443/my-webhook-path"...
Service Reference
The service stanza inside webhookClientConfig is a reference to the service for a conversion webhook.
If the webhook is running within the cluster, then you should use service instead of url.
The service namespace and name are required. The port is optional and defaults to 443.
The path is optional and defaults to "/".
Here is an example of a webhook that is configured to call a service on port "1234"
at the subpath "/my-path", and to verify the TLS connection against the ServerName
my-service-name.my-service-namespace.svc using a custom CA bundle.
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinition...spec:...conversion:strategy:WebhookwebhookClientConfig:service:namespace:my-service-namespacename:my-service-namepath:/my-pathport:1234caBundle:"Ci0tLS0tQk...<base64-encoded PEM bundle>...tLS0K"...
Webhook request and response
Request
Webhooks are sent a POST request, with Content-Type: application/json,
with a ConversionReview API object in the apiextensions.k8s.io API group
serialized to JSON as the body.
Webhooks can specify what versions of ConversionReview objects they accept
with the conversionReviewVersions field in their CustomResourceDefinition:
conversionReviewVersions is a required field when creating
apiextensions.k8s.io/v1 custom resource definitions.
Webhooks are required to support at least one ConversionReview
version understood by the current and previous API server.
# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1apiVersion:apiextensions.k8s.io/v1beta1kind:CustomResourceDefinition...spec:...conversion:strategy:WebhookconversionReviewVersions:["v1","v1beta1"]...
If no conversionReviewVersions are specified, the default when creating
apiextensions.k8s.io/v1beta1 custom resource definitions is v1beta1.
API servers send the first ConversionReview version in the conversionReviewVersions list they support.
If none of the versions in the list are supported by the API server, the custom resource definition will not be allowed to be created.
If an API server encounters a conversion webhook configuration that was previously created and does not support any of the ConversionReview
versions the API server knows how to send, attempts to call to the webhook will fail.
This example shows the data contained in an ConversionReview object
for a request to convert CronTab objects to example.com/v1:
{"apiVersion": "apiextensions.k8s.io/v1","kind": "ConversionReview","request": {# Random uid uniquely identifying this conversion call"uid": "705ab4f5-6393-11e8-b7cc-42010a800002",# The API group and version the objects should be converted to"desiredAPIVersion": "example.com/v1",# The list of objects to convert.# May contain one or more objects, in one or more versions."objects": [{"kind": "CronTab","apiVersion": "example.com/v1beta1","metadata": {"creationTimestamp": "2019-09-04T14:03:02Z","name": "local-crontab","namespace": "default","resourceVersion": "143","uid": "3415a7fc-162b-4300-b5da-fd6083580d66"},"hostPort": "localhost:1234"},{"kind": "CronTab","apiVersion": "example.com/v1beta1","metadata": {"creationTimestamp": "2019-09-03T13:02:01Z","name": "remote-crontab","resourceVersion": "12893","uid": "359a83ec-b575-460d-b553-d859cedde8a0"},"hostPort": "example.com:2345"}]}}
{# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1"apiVersion": "apiextensions.k8s.io/v1beta1","kind": "ConversionReview","request": {# Random uid uniquely identifying this conversion call"uid": "705ab4f5-6393-11e8-b7cc-42010a800002",# The API group and version the objects should be converted to"desiredAPIVersion": "example.com/v1",# The list of objects to convert.# May contain one or more objects, in one or more versions."objects": [{"kind": "CronTab","apiVersion": "example.com/v1beta1","metadata": {"creationTimestamp": "2019-09-04T14:03:02Z","name": "local-crontab","namespace": "default","resourceVersion": "143","uid": "3415a7fc-162b-4300-b5da-fd6083580d66"},"hostPort": "localhost:1234"},{"kind": "CronTab","apiVersion": "example.com/v1beta1","metadata": {"creationTimestamp": "2019-09-03T13:02:01Z","name": "remote-crontab","resourceVersion": "12893","uid": "359a83ec-b575-460d-b553-d859cedde8a0"},"hostPort": "example.com:2345"}]}}
Response
Webhooks respond with a 200 HTTP status code, Content-Type: application/json,
and a body containing a ConversionReview object (in the same version they were sent),
with the response stanza populated, serialized to JSON.
If conversion succeeds, a webhook should return a response stanza containing the following fields:
uid, copied from the request.uid sent to the webhook
result, set to {"status":"Success"}
convertedObjects, containing all of the objects from request.objects, converted to request.desiredVersion
Example of a minimal successful response from a webhook:
{"apiVersion": "apiextensions.k8s.io/v1","kind": "ConversionReview","response": {# must match <request.uid>"uid": "705ab4f5-6393-11e8-b7cc-42010a800002","result": {"status": "Success"},# Objects must match the order of request.objects, and have apiVersion set to <request.desiredAPIVersion>.# kind, metadata.uid, metadata.name, and metadata.namespace fields must not be changed by the webhook.# metadata.labels and metadata.annotations fields may be changed by the webhook.# All other changes to metadata fields by the webhook are ignored."convertedObjects": [{"kind": "CronTab","apiVersion": "example.com/v1","metadata": {"creationTimestamp": "2019-09-04T14:03:02Z","name": "local-crontab","namespace": "default","resourceVersion": "143","uid": "3415a7fc-162b-4300-b5da-fd6083580d66"},"host": "localhost","port": "1234"},{"kind": "CronTab","apiVersion": "example.com/v1","metadata": {"creationTimestamp": "2019-09-03T13:02:01Z","name": "remote-crontab","resourceVersion": "12893","uid": "359a83ec-b575-460d-b553-d859cedde8a0"},"host": "example.com","port": "2345"}]}}
{# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1"apiVersion": "apiextensions.k8s.io/v1beta1","kind": "ConversionReview","response": {# must match <request.uid>"uid": "705ab4f5-6393-11e8-b7cc-42010a800002","result": {"status": "Failed"},# Objects must match the order of request.objects, and have apiVersion set to <request.desiredAPIVersion>.# kind, metadata.uid, metadata.name, and metadata.namespace fields must not be changed by the webhook.# metadata.labels and metadata.annotations fields may be changed by the webhook.# All other changes to metadata fields by the webhook are ignored."convertedObjects": [{"kind": "CronTab","apiVersion": "example.com/v1","metadata": {"creationTimestamp": "2019-09-04T14:03:02Z","name": "local-crontab","namespace": "default","resourceVersion": "143","uid": "3415a7fc-162b-4300-b5da-fd6083580d66"},"host": "localhost","port": "1234"},{"kind": "CronTab","apiVersion": "example.com/v1","metadata": {"creationTimestamp": "2019-09-03T13:02:01Z","name": "remote-crontab","resourceVersion": "12893","uid": "359a83ec-b575-460d-b553-d859cedde8a0"},"host": "example.com","port": "2345"}]}}
If conversion fails, a webhook should return a response stanza containing the following fields:
uid, copied from the request.uid sent to the webhook
result, set to {"status":"Failed"}
Warning: Failing conversion can disrupt read and write access to the custom resources,
including the ability to update or delete the resources. Conversion failures
should be avoided whenever possible, and should not be used to enforce validation
constraints (use validation schemas or webhook admission instead).
Example of a response from a webhook indicating a conversion request failed, with an optional message:
{"apiVersion": "apiextensions.k8s.io/v1","kind": "ConversionReview","response": {"uid": "<value from request.uid>","result": {"status": "Failed","message": "hostPort could not be parsed into a separate host and port"}}}
{# Deprecated in v1.16 in favor of apiextensions.k8s.io/v1"apiVersion": "apiextensions.k8s.io/v1beta1","kind": "ConversionReview","response": {"uid": "<value from request.uid>","result": {"status": "Failed","message": "hostPort could not be parsed into a separate host and port"}}}
Writing, reading, and updating versioned CustomResourceDefinition objects
When an object is written, it is persisted at the version designated as the
storage version at the time of the write. If the storage version changes,
existing objects are never converted automatically. However, newly-created
or updated objects are written at the new storage version. It is possible for an
object to have been written at a version that is no longer served.
When you read an object, you specify the version as part of the path. If you
specify a version that is different from the object's persisted version,
Kubernetes returns the object to you at the version you requested, but the
persisted object is neither changed on disk, nor converted in any way
(other than changing the apiVersion string) while serving the request.
You can request an object at any version that is currently served.
If you update an existing object, it is rewritten at the version that is
currently the storage version. This is the only way that objects can change from
one version to another.
To illustrate this, consider the following hypothetical series of events:
The storage version is v1beta1. You create an object. It is persisted in
storage at version v1beta1
You add version v1 to your CustomResourceDefinition and designate it as
the storage version.
You read your object at version v1beta1, then you read the object again at
version v1. Both returned objects are identical except for the apiVersion
field.
You create a new object. It is persisted in storage at version v1. You now
have two objects, one of which is at v1beta1, and the other of which is at
v1.
You update the first object. It is now persisted at version v1 since that
is the current storage version.
Previous storage versions
The API server records each version which has ever been marked as the storage
version in the status field storedVersions. Objects may have been persisted
at any version that has ever been designated as a storage version. No objects
can exist in storage at a version that has never been a storage version.
Upgrade existing objects to a new stored version
When deprecating versions and dropping support, select a storage upgrade
procedure.
Remove the old version from the CustomResourceDefinition status.storedVersions field.
Option 2: Manually upgrade the existing objects to a new stored version
The following is an example procedure to upgrade from v1beta1 to v1.
Set v1 as the storage in the CustomResourceDefinition file and apply it
using kubectl. The storedVersions is now v1beta1, v1.
Write an upgrade procedure to list all existing objects and write them with
the same content. This forces the backend to write objects in the current
storage version, which is v1.
Remove v1beta1 from the CustomResourceDefinition status.storedVersions field.
Note:
The kubectl tool currently cannot be used to edit or patch the status subresource on a CRD: see the Kubectl Subresource Support KEP for more details.
The easier way to patch the status subresource from the CLI is directly interacting with the API server using the curl tool, in this example:
Setting up an extension API server to work with the aggregation layer allows the Kubernetes apiserver to be extended with additional APIs, which are not part of the core Kubernetes APIs.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Set up an extension api-server to work with the aggregation layer
The following steps describe how to set up an extension-apiserver at a high level. These steps apply regardless if you're using YAML configs or using APIs. An attempt is made to specifically identify any differences between the two. For a concrete example of how they can be implemented using YAML configs, you can look at the sample-apiserver in the Kubernetes repo.
Alternatively, you can use an existing 3rd party solution, such as apiserver-builder, which should generate a skeleton and automate all of the following steps for you.
Make sure the APIService API is enabled (check --runtime-config). It should be on by default, unless it's been deliberately turned off in your cluster.
You may need to make an RBAC rule allowing you to add APIService objects, or get your cluster administrator to make one. (Since API extensions affect the entire cluster, it is not recommended to do testing/development/debug of an API extension in a live cluster.)
Create the Kubernetes namespace you want to run your extension api-service in.
Create/get a CA cert to be used to sign the server cert the extension api-server uses for HTTPS.
Create a server cert/key for the api-server to use for HTTPS. This cert should be signed by the above CA. It should also have a CN of the Kube DNS name. This is derived from the Kubernetes service and be of the form <service name>.<service name namespace>.svc
Create a Kubernetes secret with the server cert/key in your namespace.
Create a Kubernetes deployment for the extension api-server and make sure you are loading the secret as a volume. It should contain a reference to a working image of your extension api-server. The deployment should also be in your namespace.
Make sure that your extension-apiserver loads those certs from that volume and that they are used in the HTTPS handshake.
Create a Kubernetes service account in your namespace.
Create a Kubernetes cluster role for the operations you want to allow on your resources.
Create a Kubernetes cluster role binding from the service account in your namespace to the cluster role you created.
Create a Kubernetes cluster role binding from the service account in your namespace to the system:auth-delegator cluster role to delegate auth decisions to the Kubernetes core API server.
Create a Kubernetes role binding from the service account in your namespace to the extension-apiserver-authentication-reader role. This allows your extension api-server to access the extension-apiserver-authentication configmap.
Create a Kubernetes apiservice. The CA cert above should be base64 encoded, stripped of new lines and used as the spec.caBundle in the apiservice. This should not be namespaced. If using the kube-aggregator API, only pass in the PEM encoded CA bundle because the base 64 encoding is done for you.
Use kubectl to get your resource. When run, kubectl should return "No resources found.". This message
indicates that everything worked but you currently have no objects of that resource type created.
Kubernetes ships with a default scheduler that is described
here.
If the default scheduler does not suit your needs you can implement your own scheduler.
Moreover, you can even run multiple schedulers simultaneously alongside the default
scheduler and instruct Kubernetes what scheduler to use for each of your pods. Let's
learn how to run multiple schedulers in Kubernetes with an example.
A detailed description of how to implement a scheduler is outside the scope of this
document. Please refer to the kube-scheduler implementation in
pkg/scheduler
in the Kubernetes source directory for a canonical example.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Package your scheduler binary into a container image. For the purposes of this example,
you can use the default scheduler (kube-scheduler) as your second scheduler.
Clone the Kubernetes source code from GitHub
and build the source.
git clone https://github.com/kubernetes/kubernetes.git
cd kubernetes
make
Create a container image containing the kube-scheduler binary. Here is the Dockerfile
to build the image:
FROM busyboxADD ./_output/local/bin/linux/amd64/kube-scheduler /usr/local/bin/kube-scheduler
Save the file as Dockerfile, build the image and push it to a registry. This example
pushes the image to
Google Container Registry (GCR).
For more details, please read the GCR
documentation.
Now that you have your scheduler in a container image, create a pod
configuration for it and run it in your Kubernetes cluster. But instead of creating a pod
directly in the cluster, you can use a Deployment
for this example. A Deployment manages a
Replica Set which in turn manages the pods,
thereby making the scheduler resilient to failures. Here is the deployment
config. Save it as my-scheduler.yaml:
In the above manifest, you use a KubeSchedulerConfiguration
to customize the behavior of your scheduler implementation. This configuration has been passed to
the kube-scheduler during initialization with the --config option. The my-scheduler-config ConfigMap stores the configuration file. The Pod of themy-scheduler Deployment mounts the my-scheduler-config ConfigMap as a volume.
In the aforementioned Scheduler Configuration, your scheduler implementation is represented via
a KubeSchedulerProfile.
Note: To determine if a scheduler is responsible for scheduling a specific Pod, the spec.schedulerName field in a
PodTemplate or Pod manifest must match the schedulerName field of the KubeSchedulerProfile.
All schedulers running in the cluster must have unique names.
Also, note that you create a dedicated service account my-scheduler and bind the ClusterRole
system:kube-scheduler to it so that it can acquire the same privileges as kube-scheduler.
In order to run your scheduler in a Kubernetes cluster, create the deployment
specified in the config above in a Kubernetes cluster:
kubectl create -f my-scheduler.yaml
Verify that the scheduler pod is running:
kubectl get pods --namespace=kube-system
NAME READY STATUS RESTARTS AGE
....
my-scheduler-lnf4s-4744f 1/1 Running 0 2m
...
You should see a "Running" my-scheduler pod, in addition to the default kube-scheduler
pod in this list.
Enable leader election
To run multiple-scheduler with leader election enabled, you must do the following:
Update the following fields for the KubeSchedulerConfiguration in the my-scheduler-config ConfigMap in your YAML file:
leaderElection.leaderElect to true
leaderElection.resourceNamespace to <lock-object-namespace>
leaderElection.resourceName to <lock-object-name>
Note: The control plane creates the lock objects for you, but the namespace must already exist.
You can use the kube-system namespace.
If RBAC is enabled on your cluster, you must update the system:kube-scheduler cluster role.
Add your scheduler name to the resourceNames of the rule applied for endpoints and leases resources, as in the following example:
Now that your second scheduler is running, create some pods, and direct them
to be scheduled by either the default scheduler or the one you deployed.
In order to schedule a given pod using a specific scheduler, specify the name of the
scheduler in that pod spec. Let's look at three examples.
A scheduler is specified by supplying the scheduler name as a value to spec.schedulerName. In this case, we supply the name of the
default scheduler which is default-scheduler.
Save this file as pod2.yaml and submit it to the Kubernetes cluster.
In this case, we specify that this pod should be scheduled using the scheduler that we
deployed - my-scheduler. Note that the value of spec.schedulerName should match the name supplied for the scheduler
in the schedulerName field of the mapping KubeSchedulerProfile.
Save this file as pod3.yaml and submit it to the Kubernetes cluster.
kubectl create -f pod3.yaml
Verify that all three pods are running.
kubectl get pods
Verifying that the pods were scheduled using the desired schedulers
In order to make it easier to work through these examples, we did not verify that the
pods were actually scheduled using the desired schedulers. We can verify that by
changing the order of pod and deployment config submissions above. If we submit all the
pod configs to a Kubernetes cluster before submitting the scheduler deployment config,
we see that the pod annotation-second-scheduler remains in "Pending" state forever
while the other two pods get scheduled. Once we submit the scheduler deployment config
and our new scheduler starts running, the annotation-second-scheduler pod gets
scheduled as well.
Alternatively, you can look at the "Scheduled" entries in the event logs to
verify that the pods were scheduled by the desired schedulers.
kubectl get events
You can also use a custom scheduler configuration
or a custom container image for the cluster's main scheduler by modifying its static pod manifest
on the relevant control plane nodes.
4.11.5 - Use an HTTP Proxy to Access the Kubernetes API
This page shows how to use an HTTP proxy to access the Kubernetes API.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
4.11.6 - Use a SOCKS5 Proxy to Access the Kubernetes API
FEATURE STATE:Kubernetes v1.24 [stable]
This page shows how to use a SOCKS5 proxy to access the API of a remote Kubernetes cluster.
This is useful when the cluster you want to access does not expose its API directly on the public internet.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Your Kubernetes server must be at or later than version v1.24.
To check the version, enter kubectl version.
You need SSH client software (the ssh tool), and an SSH service running on the remote server.
You must be able to log in to the SSH service on the remote server.
Task context
Note: This example tunnels traffic using SSH, with the SSH client and server acting as a SOCKS proxy.
You can instead use any other kind of SOCKS5 proxies.
Figure 1 represents what you're going to achieve in this task.
You have a client computer, referred to as local in the steps ahead, from where you're going to create requests to talk to the Kubernetes API.
The Kubernetes server/API is hosted on a remote server.
You will use SSH client and server software to create a secure SOCKS5 tunnel between the local and
the remote server. The HTTPS traffic between the client and the Kubernetes API will flow over the SOCKS5
tunnel, which is itself tunnelled over SSH.
graph LR;
subgraph local[Local client machine]
client([client])-- local traffic .-> local_ssh[Local SSH SOCKS5 proxy];
end
local_ssh[SSH SOCKS5 proxy]-- SSH Tunnel -->sshd
subgraph remote[Remote server]
sshd[SSH server]-- local traffic -->service1;
end
client([client])-. proxied HTTPs traffic going through the proxy .->service1[Kubernetes API];
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class ingress,service1,service2,pod1,pod2,pod3,pod4 k8s;
class client plain;
class cluster cluster;
Figure 1. SOCKS5 tutorial components
Using ssh to create a SOCKS5 proxy
This command starts a SOCKS5 proxy between your client machine and the remote server.
The SOCKS5 proxy lets you connect to your cluster's API server.
# The SSH tunnel continues running in the foreground after you run thisssh -D 1080 -q -N username@kubernetes-remote-server.example
-D 1080: opens a SOCKS proxy on local port :1080.
-q: quiet mode. Causes most warning and diagnostic messages to be suppressed.
-N: Do not execute a remote command. Useful for just forwarding ports.
username@kubernetes-remote-server.example: the remote SSH server where the Kubernetes cluster is running.
Client configuration
To explore the Kubernetes API you'll first need to instruct your clients to send their queries through
the SOCKS5 proxy we created earlier.
For command-line tools, set the https_proxy environment variable and pass it to commands that you run.
exporthttps_proxy=socks5h://localhost:1080
When you set the https_proxy variable, tools such as curl route HTTPS traffic through the proxy
you configured. For this to work, the tool must support SOCKS5 proxying.
Note: In the URL https://localhost:6443/api, localhost does not refer to your local client computer.
Instead, it refers to the endpoint on the remote server known as localhost.
The curl tool sends the hostname from the HTTPS URL over SOCKS, and the remote server
resolves that locally (to an address that belongs to its loopback interface).
curl -k -v https://localhost:6443/api
To use the official Kubernetes client kubectl with a proxy, set the proxy-url element
for the relevant cluster entry within your ~/.kube/config file. For example:
apiVersion:v1clusters:- cluster:certificate-authority-data:LRMEMMW2# shortened for readability server:https://<API_SERVER_IP_ADRESS>:6443 # the "Kubernetes API" server, in other words the IP address of kubernetes-remote-server.exampleproxy-url:socks5://localhost:1080 # the "SSH SOCKS5 proxy" in the diagram above (DNS resolution over socks is built-in)name:defaultcontexts:- context:cluster:defaultuser:defaultname:defaultcurrent-context:defaultkind:Configpreferences:{}users:- name:defaultuser:client-certificate-data:LS0tLS1CR==# shortened for readabilityclient-key-data:LS0tLS1CRUdJT= # shortened for readability
If the tunnel is operating and you use kubectl with a context that uses this cluster, you can interact with your cluster through that proxy. For example:
kubectl get pods
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system coredns-85cb69466-klwq8 1/1 Running 0 5m46s
Clean up
Stop the ssh port-forwarding process by pressing CTRL+C on the terminal where it is running.
Type unset https_proxy in a terminal to stop forwarding http traffic through the proxy.
The Konnectivity service provides a TCP level proxy for the control plane to cluster
communication.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this
tutorial on a cluster with at least two nodes that are not acting as control
plane hosts. If you do not already have a cluster, you can create one by using
minikube.
Configure the Konnectivity service
The following steps require an egress configuration, for example:
apiVersion:apiserver.k8s.io/v1beta1kind:EgressSelectorConfigurationegressSelections:# Since we want to control the egress traffic to the cluster, we use the# "cluster" as the name. Other supported values are "etcd", and "controlplane".- name:clusterconnection:# This controls the protocol between the API Server and the Konnectivity# server. Supported values are "GRPC" and "HTTPConnect". There is no# end user visible difference between the two modes. You need to set the# Konnectivity server to work in the same mode.proxyProtocol:GRPCtransport:# This controls what transport the API Server uses to communicate with the# Konnectivity server. UDS is recommended if the Konnectivity server# locates on the same machine as the API Server. You need to configure the# Konnectivity server to listen on the same UDS socket.# The other supported transport is "tcp". You will need to set up TLS # config to secure the TCP transport.uds:udsName:/etc/kubernetes/konnectivity-server/konnectivity-server.socket
You need to configure the API Server to use the Konnectivity service
and direct the network traffic to the cluster nodes:
Generate or obtain a certificate and kubeconfig for konnectivity-server.
For example, you can use the OpenSSL command line tool to issue a X.509 certificate,
using the cluster CA certificate /etc/kubernetes/pki/ca.crt from a control-plane host.
Deploy the Konnectivity server on your control plane node. The provided
konnectivity-server.yaml manifest assumes
that the Kubernetes components are deployed as a static Pod in your cluster. If not, you can deploy the Konnectivity
server as a DaemonSet.
apiVersion:v1kind:Podmetadata:name:konnectivity-servernamespace:kube-systemspec:priorityClassName:system-cluster-criticalhostNetwork:truecontainers:- name:konnectivity-server-containerimage:us.gcr.io/k8s-artifacts-prod/kas-network-proxy/proxy-server:v0.0.16command:["/proxy-server"]args:["--logtostderr=true",# This needs to be consistent with the value set in egressSelectorConfiguration."--uds-name=/etc/kubernetes/konnectivity-server/konnectivity-server.socket",# The following two lines assume the Konnectivity server is# deployed on the same machine as the apiserver, and the certs and# key of the API Server are at the specified location."--cluster-cert=/etc/kubernetes/pki/apiserver.crt","--cluster-key=/etc/kubernetes/pki/apiserver.key",# This needs to be consistent with the value set in egressSelectorConfiguration."--mode=grpc","--server-port=0","--agent-port=8132","--admin-port=8133","--health-port=8134","--agent-namespace=kube-system","--agent-service-account=konnectivity-agent","--kubeconfig=/etc/kubernetes/konnectivity-server.conf","--authentication-audience=system:konnectivity-server"]livenessProbe:httpGet:scheme:HTTPhost:127.0.0.1port:8134path:/healthzinitialDelaySeconds:30timeoutSeconds:60ports:- name:agentportcontainerPort:8132hostPort:8132- name:adminportcontainerPort:8133hostPort:8133- name:healthportcontainerPort:8134hostPort:8134volumeMounts:- name:k8s-certsmountPath:/etc/kubernetes/pkireadOnly:true- name:kubeconfigmountPath:/etc/kubernetes/konnectivity-server.confreadOnly:true- name:konnectivity-udsmountPath:/etc/kubernetes/konnectivity-serverreadOnly:falsevolumes:- name:k8s-certshostPath:path:/etc/kubernetes/pki- name:kubeconfighostPath:path:/etc/kubernetes/konnectivity-server.conftype:FileOrCreate- name:konnectivity-udshostPath:path:/etc/kubernetes/konnectivity-servertype:DirectoryOrCreate
Then deploy the Konnectivity agents in your cluster:
apiVersion:apps/v1# Alternatively, you can deploy the agents as Deployments. It is not necessary# to have an agent on each node.kind:DaemonSetmetadata:labels:addonmanager.kubernetes.io/mode:Reconcilek8s-app:konnectivity-agentnamespace:kube-systemname:konnectivity-agentspec:selector:matchLabels:k8s-app:konnectivity-agenttemplate:metadata:labels:k8s-app:konnectivity-agentspec:priorityClassName:system-cluster-criticaltolerations:- key:"CriticalAddonsOnly"operator:"Exists"containers:- image:us.gcr.io/k8s-artifacts-prod/kas-network-proxy/proxy-agent:v0.0.16name:konnectivity-agentcommand:["/proxy-agent"]args:["--logtostderr=true","--ca-cert=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt",# Since the konnectivity server runs with hostNetwork=true,# this is the IP address of the master machine."--proxy-server-host=35.225.206.7","--proxy-server-port=8132","--admin-server-port=8133","--health-server-port=8134","--service-account-token-path=/var/run/secrets/tokens/konnectivity-agent-token"]volumeMounts:- mountPath:/var/run/secrets/tokensname:konnectivity-agent-tokenlivenessProbe:httpGet:port:8134path:/healthzinitialDelaySeconds:15timeoutSeconds:15serviceAccountName:konnectivity-agentvolumes:- name:konnectivity-agent-tokenprojected:sources:- serviceAccountToken:path:konnectivity-agent-tokenaudience:system:konnectivity-server
Last, if RBAC is enabled in your cluster, create the relevant RBAC rules:
Understand how to protect traffic within your cluster using Transport Layer Security (TLS).
4.12.1 - Configure Certificate Rotation for the Kubelet
This page shows how to enable and configure certificate rotation for the kubelet.
FEATURE STATE:Kubernetes v1.19 [stable]
Before you begin
Kubernetes version 1.8.0 or later is required
Overview
The kubelet uses certificates for authenticating to the Kubernetes API. By
default, these certificates are issued with one year expiration so that they do
not need to be renewed too frequently.
Kubernetes contains kubelet certificate
rotation,
that will automatically generate a new key and request a new certificate from
the Kubernetes API as the current certificate approaches expiration. Once the
new certificate is available, it will be used for authenticating connections to
the Kubernetes API.
Enabling client certificate rotation
The kubelet process accepts an argument --rotate-certificates that controls
if the kubelet will automatically request a new certificate as the expiration of
the certificate currently in use approaches.
The kube-controller-manager process accepts an argument
--cluster-signing-duration (--experimental-cluster-signing-duration prior to 1.19)
that controls how long certificates will be issued for.
Understanding the certificate rotation configuration
When a kubelet starts up, if it is configured to bootstrap (using the
--bootstrap-kubeconfig flag), it will use its initial certificate to connect
to the Kubernetes API and issue a certificate signing request. You can view the
status of certificate signing requests using:
kubectl get csr
Initially a certificate signing request from the kubelet on a node will have a
status of Pending. If the certificate signing requests meets specific
criteria, it will be auto approved by the controller manager, then it will have
a status of Approved. Next, the controller manager will sign a certificate,
issued for the duration specified by the
--cluster-signing-duration parameter, and the signed certificate
will be attached to the certificate signing request.
The kubelet will retrieve the signed certificate from the Kubernetes API and
write that to disk, in the location specified by --cert-dir. Then the kubelet
will use the new certificate to connect to the Kubernetes API.
As the expiration of the signed certificate approaches, the kubelet will
automatically issue a new certificate signing request, using the Kubernetes API.
This can happen at any point between 30% and 10% of the time remaining on the
certificate. Again, the controller manager will automatically approve the certificate
request and attach a signed certificate to the certificate signing request. The
kubelet will retrieve the new signed certificate from the Kubernetes API and
write that to disk. Then it will update the connections it has to the
Kubernetes API to reconnect using the new certificate.
4.12.2 - Manage TLS Certificates in a Cluster
Kubernetes provides a certificates.k8s.io API, which lets you provision TLS
certificates signed by a Certificate Authority (CA) that you control. These CA
and certificates can be used by your workloads to establish trust.
certificates.k8s.io API uses a protocol that is similar to the ACME
draft.
Note: Certificates created using the certificates.k8s.io API are signed by a
dedicated CA. It is possible to configure your cluster to use the cluster root
CA for this purpose, but you should never rely on this. Do not assume that
these certificates will validate against the cluster root CA.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Some steps in this page use the jq tool. If you don't have jq, you can
install it via your operating system's software sources, or fetch it from
https://stedolan.github.io/jq/.
Trusting TLS in a cluster
Trusting the custom CA from an application running as a pod usually requires
some extra application configuration. You will need to add the CA certificate
bundle to the list of CA certificates that the TLS client or server trusts. For
example, you would do this with a golang TLS config by parsing the certificate
chain and adding the parsed certificates to the RootCAs field in the
tls.Config struct.
Note:
Even though the custom CA certificate may be included in the filesystem (in the
ConfigMap kube-root-ca.crt),
you should not use that certificate authority for any purpose other than to verify internal
Kubernetes endpoints. An example of an internal Kubernetes endpoint is the
Service named kubernetes in the default namespace.
If you want to use a custom certificate authority for your workloads, you should generate
that CA separately, and distribute its CA certificate using a
ConfigMap that your pods
have access to read.
Requesting a certificate
The following section demonstrates how to create a TLS certificate for a
Kubernetes service accessed through DNS.
Note: This tutorial uses CFSSL: Cloudflare's PKI and TLS toolkit click here to know more.
Create a certificate signing request
Generate a private key and certificate signing request (or CSR) by running
the following command:
Where 192.0.2.24 is the service's cluster IP,
my-svc.my-namespace.svc.cluster.local is the service's DNS name,
10.0.34.2 is the pod's IP and my-pod.my-namespace.pod.cluster.local
is the pod's DNS name. You should see the output similar to:
This command generates two files; it generates server.csr containing the PEM
encoded PKCS#10 certification request,
and server-key.pem containing the PEM encoded key to the certificate that
is still to be created.
Create a CertificateSigningRequest object to send to the Kubernetes API
Generate a CSR manifest (in YAML), and send it to the API server. You can do that by
running the following command:
Notice that the server.csr file created in step 1 is base64 encoded
and stashed in the .spec.request field. You are also requesting a
certificate with the "digital signature", "key encipherment", and "server
auth" key usages, signed by an example example.com/serving signer.
A specific signerName must be requested.
View documentation for supported signer names
for more information.
The CSR should now be visible from the API in a Pending state. You can see
it by running:
kubectl describe csr my-svc.my-namespace
Name: my-svc.my-namespace
Labels: <none>
Annotations: <none>
CreationTimestamp: Tue, 01 Feb 2022 11:49:15 -0500
Requesting User: yourname@example.com
Signer: example.com/serving
Status: Pending
Subject:
Common Name: my-pod.my-namespace.pod.cluster.local
Serial Number:
Subject Alternative Names:
DNS Names: my-pod.my-namespace.pod.cluster.local
my-svc.my-namespace.svc.cluster.local
IP Addresses: 192.0.2.24
10.0.34.2
Events: <none>
Get the CertificateSigningRequest approved
Approving the certificate signing request
is either done by an automated approval process or on a one off basis by a cluster
administrator. If you're authorized to approve a certificate request, you can do that
manually using kubectl; for example:
NAME AGE SIGNERNAME REQUESTOR REQUESTEDDURATION CONDITION
my-svc.my-namespace 10m example.com/serving yourname@example.com <none> Approved
This means the certificate request has been approved and is waiting for the
requested signer to sign it.
Sign the CertificateSigningRequest
Next, you'll play the part of a certificate signer, issue the certificate, and upload it to the API.
A signer would typically watch the CertificateSigningRequest API for objects with its signerName,
check that they have been approved, sign certificates for those requests,
and update the API object status with the issued certificate.
Create a Certificate Authority
You need an authority to provide the digital signature on the new certificate.
First, create a signing certificate by running the following:
cat <<EOF | cfssl gencert -initca - | cfssljson -bare ca
{
"CN": "My Example Signer",
"key": {
"algo": "rsa",
"size": 2048
}
}
EOF
You should see output similar to:
2022/02/01 11:50:39 [INFO] generating a new CA key and certificate from CSR
2022/02/01 11:50:39 [INFO] generate received request
2022/02/01 11:50:39 [INFO] received CSR
2022/02/01 11:50:39 [INFO] generating key: rsa-2048
2022/02/01 11:50:39 [INFO] encoded CSR
2022/02/01 11:50:39 [INFO] signed certificate with serial number 263983151013686720899716354349605500797834580472
This produces a certificate authority key file (ca-key.pem) and certificate (ca.pem).
Note: This uses the command line tool jq to populate the base64-encoded
content in the .status.certificate field.
If you do not have jq, you can also save the JSON output to a file, populate this field manually, and
upload the resulting file.
Once the CSR is approved and the signed certificate is uploaded, run:
kubectl get csr
The output is similar to:
NAME AGE SIGNERNAME REQUESTOR REQUESTEDDURATION CONDITION
my-svc.my-namespace 20m example.com/serving yourname@example.com <none> Approved,Issued
Download the certificate and use it
Now, as the requesting user, you can download the issued certificate
and save it to a server.crt file by running the following:
Now you can populate server.crt and server-key.pem in a
Secret
that you could later mount into a Pod (for example, to use with a webserver
that serves HTTPS).
kubectl create secret tls server --cert server.crt --key server-key.pem
secret/server created
Finally, you can populate ca.pem into a ConfigMap
and use it as the trust root to verify the serving certificate:
A Kubernetes administrator (with appropriate permissions) can manually approve
(or deny) CertificateSigningRequests by using the kubectl certificate approve and kubectl certificate deny commands. However if you intend
to make heavy usage of this API, you might consider writing an automated
certificates controller.
Caution:
The ability to approve CSRs decides who trusts whom within your environment. The
ability to approve CSRs should not be granted broadly or lightly.
You should make sure that you confidently understand both the verification requirements
that fall on the approver and the repercussions of issuing a specific certificate
before you grant the approve permission.
Whether a machine or a human using kubectl as above, the role of the approver is
to verify that the CSR satisfies two requirements:
The subject of the CSR controls the private key used to sign the CSR. This
addresses the threat of a third party masquerading as an authorized subject.
In the above example, this step would be to verify that the pod controls the
private key used to generate the CSR.
The subject of the CSR is authorized to act in the requested context. This
addresses the threat of an undesired subject joining the cluster. In the
above example, this step would be to verify that the pod is allowed to
participate in the requested service.
If and only if these two requirements are met, the approver should approve
the CSR and otherwise should deny the CSR.
For more information on certificate approval and access control, read
the Certificate Signing Requests
reference page.
Configuring your cluster to provide signing
This page assumes that a signer is set up to serve the certificates API. The
Kubernetes controller manager provides a default implementation of a signer. To
enable it, pass the --cluster-signing-cert-file and
--cluster-signing-key-file parameters to the controller manager with paths to
your Certificate Authority's keypair.
4.12.3 - Manual Rotation of CA Certificates
This page shows how to manually rotate the certificate authority (CA) certificates.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
For more information about authentication in Kubernetes, see
Authenticating.
For more information about best practices for CA certificates, see
Single root CA.
Rotate the CA certificates manually
Caution:
Make sure to back up your certificate directory along with configuration files and any other necessary files.
This approach assumes operation of the Kubernetes control plane in a HA configuration with multiple API servers.
Graceful termination of the API server is also assumed so clients can cleanly disconnect from one API server and
reconnect to another.
Configurations with a single API server will experience unavailability while the API server is being restarted.
Distribute the new CA certificates and private keys (for example: ca.crt, ca.key, front-proxy-ca.crt,
and front-proxy-ca.key) to all your control plane nodes in the Kubernetes certificates directory.
Update the --root-ca-file flag for the kube-controller-manager to include
both old and new CA, then restart the kube-controller-manager.
Any ServiceAccount created after this point will get
Secrets that include both old and new CAs.
Note:
The files specified by the kube-controller-manager flags --client-ca-file and --cluster-signing-cert-file
cannot be CA bundles. If these flags and --root-ca-file point to the same ca.crt file which is now a
bundle (includes both old and new CA) you will face an error. To workaround this problem you can copy the new CA
to a separate file and make the flags --client-ca-file and --cluster-signing-cert-file point to the copy.
Once ca.crt is no longer a bundle you can restore the problem flags to point to ca.crt and delete the copy.
Issue 1350 for kubeadm tracks an bug with the
kube-controller-manager being unable to accept a CA bundle.
Wait for the controller manager to update ca.crt in the service account Secrets to include both old and new CA certificates.
If any Pods are started before new CA is used by API servers, the new Pods get this update and will trust both
old and new CAs.
Restart all pods using in-cluster configurations (for example: kube-proxy, CoreDNS, etc) so they can use the
updated certificate authority data from Secrets that link to ServiceAccounts.
Make sure CoreDNS, kube-proxy and other Pods using in-cluster configurations are working as expected.
Append the both old and new CA to the file against --client-ca-file and --kubelet-certificate-authority
flag in the kube-apiserver configuration.
Append the both old and new CA to the file against --client-ca-file flag in the kube-scheduler configuration.
Update certificates for user accounts by replacing the content of client-certificate-data and client-key-data
respectively.
Additionally, update the certificate-authority-data section in the kubeconfig files,
respectively with Base64-encoded old and new certificate authority data
Update the --root-ca-file flag for the Cloud Controller Manager to include
both old and new CA, then restart the cloud-controller-manager.
Note: If your cluster does not have a cloud-controller-manager, you can skip this step.
Follow the steps below in a rolling fashion.
Restart any other
aggregated API servers or
webhook handlers to trust the new CA certificates.
Restart the kubelet by update the file against clientCAFile in kubelet configuration and
certificate-authority-data in kubelet.conf to use both the old and new CA on all nodes.
If your kubelet is not using client certificate rotation, update client-certificate-data and
client-key-data in kubelet.conf on all nodes along with the kubelet client certificate file
usually found in /var/lib/kubelet/pki.
Restart API servers with the certificates (apiserver.crt, apiserver-kubelet-client.crt and
front-proxy-client.crt) signed by new CA.
You can use the existing private keys or new private keys.
If you changed the private keys then update these in the Kubernetes certificates directory as well.
Since the Pods in your cluster trust both old and new CAs, there will be a momentarily disconnection
after which pods' Kubernetes clients reconnect to the new API server.
The new API server uses a certificate signed by the new CA.
Restart the kube-scheduler to use and
trust the new CAs.
Make sure control plane components logs no TLS errors.
Note: To generate certificates and private keys for your cluster using the openssl command line tool,
see Certificates (openssl).
You can also use cfssl.
Annotate any DaemonSets and Deployments to trigger pod replacement in a safer rolling fashion.
for namespace in $(kubectl get namespace -o jsonpath='{.items[*].metadata.name}'); dofor name in $(kubectl get deployments -n $namespace -o jsonpath='{.items[*].metadata.name}'); do kubectl patch deployment -n ${namespace}${name} -p '{"spec":{"template":{"metadata":{"annotations":{"ca-rotation": "1"}}}}}';
donefor name in $(kubectl get daemonset -n $namespace -o jsonpath='{.items[*].metadata.name}'); do kubectl patch daemonset -n ${namespace}${name} -p '{"spec":{"template":{"metadata":{"annotations":{"ca-rotation": "1"}}}}}';
donedone
Check the logs from control plane components, along with the kubelet and the kube-proxy.
Ensure those components are not reporting any TLS errors; see
looking at the logs for more details.
Validate logs from any aggregated api servers and pods using in-cluster config.
Once the cluster functionality is successfully verified:
Update all service account tokens to include new CA certificate only.
All pods using an in-cluster kubeconfig will eventually need to be restarted to pick up the new Secret,
so that no Pods are relying on the old cluster CA.
Restart the control plane components by removing the old CA from the kubeconfig files and the files against
--client-ca-file, --root-ca-file flags resp.
On each node, restart the kubelet by removing the old CA from file against the clientCAFile flag
and from the kubelet kubeconfig file. You should carry this out as a rolling update.
If your cluster lets you make this change, you can also roll it out by replacing nodes rather than
reconfiguring them.
4.13 - Manage Cluster Daemons
Perform common tasks for managing a DaemonSet, such as performing a rolling update.
4.13.1 - Perform a Rolling Update on a DaemonSet
This page shows how to perform a rolling update on a DaemonSet.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
OnDelete: With OnDelete update strategy, after you update a DaemonSet template, new
DaemonSet pods will only be created when you manually delete old DaemonSet
pods. This is the same behavior of DaemonSet in Kubernetes version 1.5 or
before.
RollingUpdate: This is the default update strategy.
With RollingUpdate update strategy, after you update a
DaemonSet template, old DaemonSet pods will be killed, and new DaemonSet pods
will be created automatically, in a controlled fashion. At most one pod of
the DaemonSet will be running on each node during the whole update process.
Performing a Rolling Update
To enable the rolling update feature of a DaemonSet, you must set its
.spec.updateStrategy.type to RollingUpdate.
apiVersion:apps/v1kind:DaemonSetmetadata:name:fluentd-elasticsearchnamespace:kube-systemlabels:k8s-app:fluentd-loggingspec:selector:matchLabels:name:fluentd-elasticsearchupdateStrategy:type:RollingUpdaterollingUpdate:maxUnavailable:1template:metadata:labels:name:fluentd-elasticsearchspec:tolerations:# these tolerations are to have the daemonset runnable on control plane nodes# remove them if your control plane nodes should not run pods- key:node-role.kubernetes.io/control-planeoperator:Existseffect:NoSchedule- key:node-role.kubernetes.io/masteroperator:Existseffect:NoSchedulecontainers:- name:fluentd-elasticsearchimage:quay.io/fluentd_elasticsearch/fluentd:v2.5.2volumeMounts:- name:varlogmountPath:/var/log- name:varlibdockercontainersmountPath:/var/lib/docker/containersreadOnly:trueterminationGracePeriodSeconds:30volumes:- name:varloghostPath:path:/var/log- name:varlibdockercontainershostPath:path:/var/lib/docker/containers
After verifying the update strategy of the DaemonSet manifest, create the DaemonSet:
If the output isn't RollingUpdate, go back and modify the DaemonSet object or
manifest accordingly.
Updating a DaemonSet template
Any updates to a RollingUpdate DaemonSet .spec.template will trigger a rolling
update. Let's update the DaemonSet by applying a new YAML file. This can be done with several different kubectl commands.
apiVersion:apps/v1kind:DaemonSetmetadata:name:fluentd-elasticsearchnamespace:kube-systemlabels:k8s-app:fluentd-loggingspec:selector:matchLabels:name:fluentd-elasticsearchupdateStrategy:type:RollingUpdaterollingUpdate:maxUnavailable:1template:metadata:labels:name:fluentd-elasticsearchspec:tolerations:# these tolerations are to have the daemonset runnable on control plane nodes# remove them if your control plane nodes should not run pods- key:node-role.kubernetes.io/control-planeoperator:Existseffect:NoSchedule- key:node-role.kubernetes.io/masteroperator:Existseffect:NoSchedulecontainers:- name:fluentd-elasticsearchimage:quay.io/fluentd_elasticsearch/fluentd:v2.5.2resources:limits:memory:200Mirequests:cpu:100mmemory:200MivolumeMounts:- name:varlogmountPath:/var/log- name:varlibdockercontainersmountPath:/var/lib/docker/containersreadOnly:trueterminationGracePeriodSeconds:30volumes:- name:varloghostPath:path:/var/log- name:varlibdockercontainershostPath:path:/var/lib/docker/containers
If you only need to update the container image in the DaemonSet template, i.e.
.spec.template.spec.containers[*].image, use kubectl set image:
kubectl set image ds/fluentd-elasticsearch fluentd-elasticsearch=quay.io/fluentd_elasticsearch/fluentd:v2.6.0 -n kube-system
Watching the rolling update status
Finally, watch the rollout status of the latest DaemonSet rolling update:
kubectl rollout status ds/fluentd-elasticsearch -n kube-system
When the rollout is complete, the output is similar to this:
daemonset "fluentd-elasticsearch" successfully rolled out
Troubleshooting
DaemonSet rolling update is stuck
Sometimes, a DaemonSet rolling update may be stuck. Here are some possible
causes:
Some nodes run out of resources
The rollout is stuck because new DaemonSet pods can't be scheduled on at least one
node. This is possible when the node is
running out of resources.
When this happens, find the nodes that don't have the DaemonSet pods scheduled on
by comparing the output of kubectl get nodes and the output of:
kubectl get pods -l name=fluentd-elasticsearch -o wide -n kube-system
Once you've found those nodes, delete some non-DaemonSet pods from the node to
make room for new DaemonSet pods.
Note: This will cause service disruption when deleted pods are not controlled by any controllers or pods are not
replicated. This does not respect PodDisruptionBudget
either.
Broken rollout
If the recent DaemonSet template update is broken, for example, the container is
crash looping, or the container image doesn't exist (often due to a typo),
DaemonSet rollout won't progress.
To fix this, update the DaemonSet template again. New rollout won't be
blocked by previous unhealthy rollouts.
Clock skew
If .spec.minReadySeconds is specified in the DaemonSet, clock skew between
master and nodes will make DaemonSet unable to detect the right rollout
progress.
This page shows how to perform a rollback on a DaemonSet.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Change cause is copied from DaemonSet annotation kubernetes.io/change-cause
to its revisions upon creation. You may specify --record=true in kubectl
to record the command executed in the change cause annotation.
To see the details of a specific revision:
kubectl rollout history daemonset <daemonset-name> --revision=1
# Specify the revision number you get from Step 1 in --to-revisionkubectl rollout undo daemonset <daemonset-name> --to-revision=<revision>
If it succeeds, the command returns:
daemonset "<daemonset-name>" rolled back
Note: If --to-revision flag is not specified, kubectl picks the most recent revision.
Step 3: Watch the progress of the DaemonSet rollback
kubectl rollout undo daemonset tells the server to start rolling back the
DaemonSet. The real rollback is done asynchronously inside the cluster
control plane.
To watch the progress of the rollback:
kubectl rollout status ds/<daemonset-name>
When the rollback is complete, the output is similar to:
daemonset "<daemonset-name>" successfully rolled out
Understanding DaemonSet revisions
In the previous kubectl rollout history step, you got a list of DaemonSet
revisions. Each revision is stored in a resource named ControllerRevision.
To see what is stored in each revision, find the DaemonSet revision raw
resources:
kubectl get controllerrevision -l <daemonset-selector-key>=<daemonset-selector-value>
This returns a list of ControllerRevisions:
NAME CONTROLLER REVISION AGE
<daemonset-name>-<revision-hash> DaemonSet/<daemonset-name> 1 1h
<daemonset-name>-<revision-hash> DaemonSet/<daemonset-name> 2 1h
Each ControllerRevision stores the annotations and template of a DaemonSet
revision.
kubectl rollout undo takes a specific ControllerRevision and replaces
DaemonSet template with the template stored in the ControllerRevision.
kubectl rollout undo is equivalent to updating DaemonSet template to a
previous revision through other commands, such as kubectl edit or kubectl apply.
Note: DaemonSet revisions only roll forward. That is to say, after a
rollback completes, the revision number (.revision field) of the
ControllerRevision being rolled back to will advance. For example, if you
have revision 1 and 2 in the system, and roll back from revision 2 to revision
1, the ControllerRevision with .revision: 1 will become .revision: 3.
Learn how to configure networking for your cluster.
4.14.1 - Adding entries to Pod /etc/hosts with HostAliases
Adding entries to a Pod's /etc/hosts file provides Pod-level override of hostname resolution when DNS and other options are not applicable. You can add these custom entries with the HostAliases field in PodSpec.
Modification not using HostAliases is not suggested because the file is managed by the kubelet and can be overwritten on during Pod creation/restart.
Default hosts file content
Start an Nginx Pod which is assigned a Pod IP:
kubectl run nginx --image nginx
pod/nginx created
Examine a Pod IP:
kubectl get pods --output=wide
NAME READY STATUS RESTARTS AGE IP NODE
nginx 1/1 Running 0 13s 10.200.0.4 worker0
By default, the hosts file only includes IPv4 and IPv6 boilerplates like
localhost and its own hostname.
Adding additional entries with hostAliases
In addition to the default boilerplate, you can add additional entries to the
hosts file.
For example: to resolve foo.local, bar.local to 127.0.0.1 and foo.remote,
bar.remote to 10.1.2.3, you can configure HostAliases for a Pod under
.spec.hostAliases:
with the additional entries specified at the bottom.
Why does the kubelet manage the hosts file?
The kubelet manages the
hosts file for each container of the Pod to prevent the container runtime from
modifying the file after the containers have already been started.
Historically, Kubernetes always used Docker Engine as its container runtime, and Docker Engine would
then modify the /etc/hosts file after each container had started.
Current Kubernetes can use a variety of container runtimes; even so, the kubelet manages the
hosts file within each container so that the outcome is as intended regardless of which
container runtime you use.
Caution:
Avoid making manual changes to the hosts file inside a container.
If you make manual changes to the hosts file,
those changes are lost when the container exits.
4.14.2 - Validate IPv4/IPv6 dual-stack
This document shares how to validate IPv4/IPv6 dual-stack enabled Kubernetes clusters.
Before you begin
Provider support for dual-stack networking (Cloud provider or otherwise must be able to provide Kubernetes nodes with routable IPv4/IPv6 network interfaces)
Your Kubernetes server must be at or later than version v1.23.
To check the version, enter kubectl version.
Note: While you can validate with an earlier version, the feature is only GA and officially supported since v1.23.
Validate addressing
Validate node addressing
Each dual-stack Node should have a single IPv4 block and a single IPv6 block allocated. Validate that IPv4/IPv6 Pod address ranges are configured by running the following command. Replace the sample node name with a valid dual-stack Node from your cluster. In this example, the Node's name is k8s-linuxpool1-34450317-0:
kubectl get nodes k8s-linuxpool1-34450317-0 -o go-template --template='{{range .spec.podCIDRs}}{{printf "%s\n" .}}{{end}}'
10.244.1.0/24
a00:100::/24
There should be one IPv4 block and one IPv6 block allocated.
Validate that the node has an IPv4 and IPv6 interface detected. Replace node name with a valid node from the cluster. In this example the node name is k8s-linuxpool1-34450317-0:
Validate that a Pod has an IPv4 and IPv6 address assigned. Replace the Pod name with a valid Pod in your cluster. In this example the Pod name is pod01:
kubectl get pods pod01 -o go-template --template='{{range .status.podIPs}}{{printf "%s\n" .ip}}{{end}}'
10.244.1.4
a00:100::4
You can also validate Pod IPs using the Downward API via the status.podIPs fieldPath. The following snippet demonstrates how you can expose the Pod IPs via an environment variable called MY_POD_IPS within a container.
The following command prints the value of the MY_POD_IPS environment variable from within a container. The value is a comma separated list that corresponds to the Pod's IPv4 and IPv6 addresses.
kubectl exec -it pod01 -- set | grep MY_POD_IPS
MY_POD_IPS=10.244.1.4,a00:100::4
The Pod's IP addresses will also be written to /etc/hosts within a container. The following command executes a cat on /etc/hosts on a dual stack Pod. From the output you can verify both the IPv4 and IPv6 IP address for the Pod.
Create the following Service that does not explicitly define .spec.ipFamilyPolicy. Kubernetes will assign a cluster IP for the Service from the first configured service-cluster-ip-range and set the .spec.ipFamilyPolicy to SingleStack.
The Service has .spec.ipFamilyPolicy set to SingleStack and .spec.clusterIP set to an IPv4 address from the first configured range set via --service-cluster-ip-range flag on kube-controller-manager.
Create the following Service that explicitly defines IPv6 as the first array element in .spec.ipFamilies. Kubernetes will assign a cluster IP for the Service from the IPv6 range configured service-cluster-ip-range and set the .spec.ipFamilyPolicy to SingleStack.
The Service has .spec.ipFamilyPolicy set to SingleStack and .spec.clusterIP set to an IPv6 address from the IPv6 range set via --service-cluster-ip-range flag on kube-controller-manager.
Create the following Service that explicitly defines PreferDualStack in .spec.ipFamilyPolicy. Kubernetes will assign both IPv4 and IPv6 addresses (as this cluster has dual-stack enabled) and select the .spec.ClusterIP from the list of .spec.ClusterIPs based on the address family of the first element in the .spec.ipFamilies array.
The kubectl get svc command will only show the primary IP in the CLUSTER-IP field.
kubectl get svc -l app.kubernetes.io/name=MyApp
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-service ClusterIP 10.0.216.242 <none> 80/TCP 5s
Validate that the Service gets cluster IPs from the IPv4 and IPv6 address blocks using kubectl describe. You may then validate access to the service via the IPs and ports.
If the cloud provider supports the provisioning of IPv6 enabled external load balancers, create the following Service with PreferDualStack in .spec.ipFamilyPolicy, IPv6 as the first element of the .spec.ipFamilies array and the type field set to LoadBalancer.
Validate that the Service receives a CLUSTER-IP address from the IPv6 address block along with an EXTERNAL-IP. You may then validate access to the service via the IP and port.
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-service LoadBalancer fd00::7ebc 2603:1030:805::5 80:30790/TCP 35s
4.15 - Configure a kubelet image credential provider
Configure the kubelet's image credential provider plugin
FEATURE STATE:Kubernetes v1.24 [beta]
Starting from Kubernetes v1.20, the kubelet can dynamically retrieve credentials for a container image registry
using exec plugins. The kubelet and the exec plugin communicate through stdio (stdin, stdout, and stderr) using
Kubernetes versioned APIs. These plugins allow the kubelet to request credentials for a container registry dynamically
as opposed to storing static credentials on disk. For example, the plugin may talk to a local metadata server to retrieve
short-lived credentials for an image that is being pulled by the kubelet.
You may be interested in using this capability if any of the below are true:
API calls to a cloud provider service are required to retrieve authentication information for a registry.
Credentials have short expiration times and requesting new credentials frequently is required.
Storing registry credentials on disk or in imagePullSecrets is not acceptable.
This guide demonstrates how to configure the kubelet's image credential provider plugin mechanism.
Before you begin
The kubelet image credential provider is introduced in v1.20 as an alpha feature. As with other alpha features,
a feature gate KubeletCredentialProviders must be enabled on only the kubelet for the feature to work.
A working implementation of a credential provider exec plugin. You can build your own plugin or use one provided by cloud providers.
Installing Plugins on Nodes
A credential provider plugin is an executable binary that will be run by the kubelet. Ensure that the plugin binary exists on
every node in your cluster and stored in a known directory. The directory will be required later when configuring kubelet flags.
Configuring the Kubelet
In order to use this feature, the kubelet expects two flags to be set:
--image-credential-provider-config - the path to the credential provider plugin config file.
--image-credential-provider-bin-dir - the path to the directory where credential provider plugin binaries are located.
Configure a kubelet credential provider
The configuration file passed into --image-credential-provider-config is read by the kubelet to determine which exec plugins
should be invoked for which container images. Here's an example configuration file you may end up using if you are using the
ECR-based plugin:
apiVersion:kubelet.config.k8s.io/v1alpha1kind:CredentialProviderConfig# providers is a list of credential provider plugins that will be enabled by the kubelet.# Multiple providers may match against a single image, in which case credentials# from all providers will be returned to the kubelet. If multiple providers are called# for a single image, the results are combined. If providers return overlapping# auth keys, the value from the provider earlier in this list is used.providers:# name is the required name of the credential provider. It must match the name of the# provider executable as seen by the kubelet. The executable must be in the kubelet's# bin directory (set by the --image-credential-provider-bin-dir flag).- name:ecr# matchImages is a required list of strings used to match against images in order to# determine if this provider should be invoked. If one of the strings matches the# requested image from the kubelet, the plugin will be invoked and given a chance# to provide credentials. Images are expected to contain the registry domain# and URL path.## Each entry in matchImages is a pattern which can optionally contain a port and a path.# Globs can be used in the domain, but not in the port or the path. Globs are supported# as subdomains like '*.k8s.io' or 'k8s.*.io', and top-level-domains such as 'k8s.*'.# Matching partial subdomains like 'app*.k8s.io' is also supported. Each glob can only match# a single subdomain segment, so *.io does not match *.k8s.io.## A match exists between an image and a matchImage when all of the below are true:# - Both contain the same number of domain parts and each part matches.# - The URL path of an imageMatch must be a prefix of the target image URL path.# - If the imageMatch contains a port, then the port must match in the image as well.## Example values of matchImages:# - 123456789.dkr.ecr.us-east-1.amazonaws.com# - *.azurecr.io# - gcr.io# - *.*.registry.io# - registry.io:8080/pathmatchImages:- "*.dkr.ecr.*.amazonaws.com"- "*.dkr.ecr.*.amazonaws.cn"- "*.dkr.ecr-fips.*.amazonaws.com"- "*.dkr.ecr.us-iso-east-1.c2s.ic.gov"- "*.dkr.ecr.us-isob-east-1.sc2s.sgov.gov"# defaultCacheDuration is the default duration the plugin will cache credentials in-memory# if a cache duration is not provided in the plugin response. This field is required.defaultCacheDuration:"12h"# Required input version of the exec CredentialProviderRequest. The returned CredentialProviderResponse# MUST use the same encoding version as the input. Current supported values are:# - credentialprovider.kubelet.k8s.io/v1alpha1apiVersion:credentialprovider.kubelet.k8s.io/v1alpha1# Arguments to pass to the command when executing it.# +optionalargs:- get-credentials# Env defines additional environment variables to expose to the process. These# are unioned with the host's environment, as well as variables client-go uses# to pass argument to the plugin.# +optionalenv:- name:AWS_PROFILEvalue:example_profile
The providers field is a list of enabled plugins used by the kubelet. Each entry has a few required fields:
name: the name of the plugin which MUST match the name of the executable binary that exists
in the directory passed into --image-credential-provider-bin-dir.
matchImages: a list of strings used to match against images in order to determine
if this provider should be invoked. More on this below.
defaultCacheDuration: the default duration the kubelet will cache credentials in-memory
if a cache duration was not specified by the plugin.
apiVersion: the API version that the kubelet and the exec plugin will use when communicating.
Each credential provider can also be given optional args and environment variables as well.
Consult the plugin implementors to determine what set of arguments and environment variables are required for a given plugin.
Configure image matching
The matchImages field for each credential provider is used by the kubelet to determine whether a plugin should be invoked
for a given image that a Pod is using. Each entry in matchImages is an image pattern which can optionally contain a port and a path.
Globs can be used in the domain, but not in the port or the path. Globs are supported as subdomains like *.k8s.io or k8s.*.io,
and top-level domains such as k8s.*. Matching partial subdomains like app*.k8s.io is also supported. Each glob can only match
a single subdomain segment, so *.io does NOT match *.k8s.io.
A match exists between an image name and a matchImage entry when all of the below are true:
Both contain the same number of domain parts and each part matches.
The URL path of match image must be a prefix of the target image URL path.
If the imageMatch contains a port, then the port must match in the image as well.
Extend kubectl by creating and installing kubectl plugins.
This guide demonstrates how to install and write extensions for kubectl. By thinking of core kubectl commands as essential building blocks for interacting with a Kubernetes cluster, a cluster administrator can think
of plugins as a means of utilizing these building blocks to create more complex behavior. Plugins extend kubectl with new sub-commands, allowing for new and custom features not included in the main distribution of kubectl.
Before you begin
You need to have a working kubectl binary installed.
Installing kubectl plugins
A plugin is a standalone executable file, whose name begins with kubectl-. To install a plugin, move its executable file to anywhere on your PATH.
You can also discover and install kubectl plugins available in the open source
using Krew. Krew is a plugin manager maintained by
the Kubernetes SIG CLI community.
Caution: Kubectl plugins available via the Krew plugin index
are not audited for security. You should install and run third-party plugins at your
own risk, since they are arbitrary programs running on your machine.
Discovering plugins
kubectl provides a command kubectl plugin list that searches your PATH for valid plugin executables.
Executing this command causes a traversal of all files in your PATH. Any files that are executable, and begin with kubectl- will show up in the order in which they are present in your PATH in this command's output.
A warning will be included for any files beginning with kubectl- that are not executable.
A warning will also be included for any valid plugin files that overlap each other's name.
You can use Krew to discover and install kubectl
plugins from a community-curated
plugin index.
Limitations
It is currently not possible to create plugins that overwrite existing kubectl commands. For example, creating a plugin kubectl-version will cause that plugin to never be executed, as the existing kubectl version command will always take precedence over it. Due to this limitation, it is also not possible to use plugins to add new subcommands to existing kubectl commands. For example, adding a subcommand kubectl create foo by naming your plugin kubectl-create-foo will cause that plugin to be ignored.
kubectl plugin list shows warnings for any valid plugins that attempt to do this.
Writing kubectl plugins
You can write a plugin in any programming language or script that allows you to write command-line commands.
There is no plugin installation or pre-loading required. Plugin executables receive
the inherited environment from the kubectl binary.
A plugin determines which command path it wishes to implement based on its name.
For example, a plugin named kubectl-foo provides a command kubectl foo. You must
install the plugin executable somewhere in your PATH.
Example plugin
#!/bin/bash
# optional argument handlingif[["$1"=="version"]]thenecho"1.0.0"exit0fi# optional argument handlingif[["$1"=="config"]]thenecho"$KUBECONFIG"exit0fiecho"I am a plugin named kubectl-foo"
Using a plugin
To use a plugin, make the plugin executable:
sudo chmod +x ./kubectl-foo
and place it anywhere in your PATH:
sudo mv ./kubectl-foo /usr/local/bin
You may now invoke your plugin as a kubectl command:
kubectl foo
I am a plugin named kubectl-foo
All args and flags are passed as-is to the executable:
kubectl foo version
1.0.0
All environment variables are also passed as-is to the executable:
Additionally, the first argument that is passed to a plugin will always be the full path to the location where it was invoked ($0 would equal /usr/local/bin/kubectl-foo in the example above).
Naming a plugin
As seen in the example above, a plugin determines the command path that it will implement based on its filename. Every sub-command in the command path that a plugin targets, is separated by a dash (-).
For example, a plugin that wishes to be invoked whenever the command kubectl foo bar baz is invoked by the user, would have the filename of kubectl-foo-bar-baz.
Flags and argument handling
Note:
The plugin mechanism does not create any custom, plugin-specific values or environment variables for a plugin process.
An older kubectl plugin mechanism provided environment variables such as KUBECTL_PLUGINS_CURRENT_NAMESPACE; that no longer happens.
kubectl plugins must parse and validate all of the arguments passed to them.
See using the command line runtime package for details of a Go library aimed at plugin authors.
Here are some additional cases where users invoke your plugin while providing additional flags and arguments. This builds upon the kubectl-foo-bar-baz plugin from the scenario above.
If you run kubectl foo bar baz arg1 --flag=value arg2, kubectl's plugin mechanism will first try to find the plugin with the longest possible name, which in this case
would be kubectl-foo-bar-baz-arg1. Upon not finding that plugin, kubectl then treats the last dash-separated value as an argument (arg1 in this case), and attempts to find the next longest possible name, kubectl-foo-bar-baz.
Upon having found a plugin with this name, kubectl then invokes that plugin, passing all args and flags after the plugin's name as arguments to the plugin process.
Example:
# create a pluginecho -e '#!/bin/bash\n\necho "My first command-line argument was $1"' > kubectl-foo-bar-baz
sudo chmod +x ./kubectl-foo-bar-baz
# "install" your plugin by moving it to a directory in your $PATHsudo mv ./kubectl-foo-bar-baz /usr/local/bin
# check that kubectl recognizes your pluginkubectl plugin list
The following kubectl-compatible plugins are available:
/usr/local/bin/kubectl-foo-bar-baz
# test that calling your plugin via a "kubectl" command works
# even when additional arguments and flags are passed to your
# plugin executable by the user.
kubectl foo bar baz arg1 --meaningless-flag=true
My first command-line argument was arg1
As you can see, your plugin was found based on the kubectl command specified by a user, and all extra arguments and flags were passed as-is to the plugin executable once it was found.
Names with dashes and underscores
Although the kubectl plugin mechanism uses the dash (-) in plugin filenames to separate the sequence of sub-commands processed by the plugin, it is still possible to create a plugin
command containing dashes in its commandline invocation by using underscores (_) in its filename.
Example:
# create a plugin containing an underscore in its filenameecho -e '#!/bin/bash\n\necho "I am a plugin with a dash in my name"' > ./kubectl-foo_bar
sudo chmod +x ./kubectl-foo_bar
# move the plugin into your $PATHsudo mv ./kubectl-foo_bar /usr/local/bin
# You can now invoke your plugin via kubectl:kubectl foo-bar
I am a plugin with a dash in my name
Note that the introduction of underscores to a plugin filename does not prevent you from having commands such as kubectl foo_bar.
The command from the above example, can be invoked using either a dash (-) or an underscore (_):
# You can invoke your custom command with a dashkubectl foo-bar
I am a plugin with a dash in my name
# You can also invoke your custom command with an underscorekubectl foo_bar
I am a plugin with a dash in my name
Name conflicts and overshadowing
It is possible to have multiple plugins with the same filename in different locations throughout your PATH.
For example, given a PATH with the following value: PATH=/usr/local/bin/plugins:/usr/local/bin/moreplugins, a copy of plugin kubectl-foo could exist in /usr/local/bin/plugins and /usr/local/bin/moreplugins,
such that the output of the kubectl plugin list command is:
PATH=/usr/local/bin/plugins:/usr/local/bin/moreplugins kubectl plugin list
The following kubectl-compatible plugins are available:
/usr/local/bin/plugins/kubectl-foo
/usr/local/bin/moreplugins/kubectl-foo
- warning: /usr/local/bin/moreplugins/kubectl-foo is overshadowed by a similarly named plugin: /usr/local/bin/plugins/kubectl-foo
error: one plugin warning was found
In the above scenario, the warning under /usr/local/bin/moreplugins/kubectl-foo tells you that this plugin will never be executed. Instead, the executable that appears first in your PATH, /usr/local/bin/plugins/kubectl-foo, will always be found and executed first by the kubectl plugin mechanism.
A way to resolve this issue is to ensure that the location of the plugin that you wish to use with kubectl always comes first in your PATH. For example, if you want to always use /usr/local/bin/moreplugins/kubectl-foo anytime that the kubectl command kubectl foo was invoked, change the value of your PATH to be /usr/local/bin/moreplugins:/usr/local/bin/plugins.
Invocation of the longest executable filename
There is another kind of overshadowing that can occur with plugin filenames. Given two plugins present in a user's PATH: kubectl-foo-bar and kubectl-foo-bar-baz, the kubectl plugin mechanism will always choose the longest possible plugin name for a given user command. Some examples below, clarify this further:
# for a given kubectl command, the plugin with the longest possible filename will always be preferredkubectl foo bar baz
Plugin kubectl-foo-bar-baz is executed
kubectl foo bar
Plugin kubectl-foo-bar is executed
kubectl foo bar baz buz
Plugin kubectl-foo-bar-baz is executed, with "buz" as its first argument
kubectl foo bar buz
Plugin kubectl-foo-bar is executed, with "buz" as its first argument
This design choice ensures that plugin sub-commands can be implemented across multiple files, if needed, and that these sub-commands can be nested under a "parent" plugin command:
You can use the aforementioned kubectl plugin list command to ensure that your plugin is visible by kubectl, and verify that there are no warnings preventing it from being called as a kubectl command.
kubectl plugin list
The following kubectl-compatible plugins are available:
test/fixtures/pkg/kubectl/plugins/kubectl-foo
/usr/local/bin/kubectl-foo
- warning: /usr/local/bin/kubectl-foo is overshadowed by a similarly named plugin: test/fixtures/pkg/kubectl/plugins/kubectl-foo
plugins/kubectl-invalid
- warning: plugins/kubectl-invalid identified as a kubectl plugin, but it is not executable
error: 2 plugin warnings were found
Using the command line runtime package
If you're writing a plugin for kubectl and you're using Go, you can make use
of the
cli-runtime utility libraries.
These libraries provide helpers for parsing or updating a user's
kubeconfig
file, for making REST-style requests to the API server, or to bind flags
associated with configuration and printing.
See the Sample CLI Plugin for
an example usage of the tools provided in the CLI Runtime repo.
Distributing kubectl plugins
If you have developed a plugin for others to use, you should consider how you
package it, distribute it and deliver updates to your users.
Krew
Krew offers a cross-platform way to package and
distribute your plugins. This way, you use a single packaging format for all
target platforms (Linux, Windows, macOS etc) and deliver updates to your users.
Krew also maintains a plugin
index so that other people can
discover your plugin and install it.
Native / platform specific package management
Alternatively, you can use traditional package managers such as, apt or yum
on Linux, Chocolatey on Windows, and Homebrew on macOS. Any package
manager will be suitable if it can place new executables placed somewhere
in the user's PATH.
As a plugin author, if you pick this option then you also have the burden
of updating your kubectl plugin's distribution package across multiple
platforms for each release.
Source code
You can publish the source code; for example, as a Git repository. If you
choose this option, someone who wants to use that plugin must fetch the code,
set up a build environment (if it needs compiling), and deploy the plugin.
If you also make compiled packages available, or use Krew, that will make
installs easier.
What's next
Check the Sample CLI Plugin repository for a
detailed example of a
plugin written in Go.
In case of any questions, feel free to reach out to the
SIG CLI team.
Read about Krew, a package manager for kubectl plugins.
4.17 - Manage HugePages
Configure and manage huge pages as a schedulable resource in a cluster.
FEATURE STATE:Kubernetes v1.25 [stable]
Kubernetes supports the allocation and consumption of pre-allocated huge pages
by applications in a Pod. This page describes how users can consume huge pages.
Before you begin
Kubernetes nodes must pre-allocate huge pages in order for the node to report
its huge page capacity. A node can pre-allocate huge pages for multiple
sizes.
The nodes will automatically discover and report all huge page resources as
schedulable resources.
API
Huge pages can be consumed via container level resource requirements using the
resource name hugepages-<size>, where <size> is the most compact binary
notation using integer values supported on a particular node. For example, if a
node supports 2048KiB and 1048576KiB page sizes, it will expose a schedulable
resources hugepages-2Mi and hugepages-1Gi. Unlike CPU or memory, huge pages
do not support overcommit. Note that when requesting hugepage resources, either
memory or CPU resources must be requested as well.
A pod may consume multiple huge page sizes in a single pod spec. In this case it
must use medium: HugePages-<hugepagesize> notation for all volume mounts.
Huge page requests must equal the limits. This is the default if limits are
specified, but requests are not.
Huge pages are isolated at a container scope, so each container has own
limit on their cgroup sandbox as requested in a container spec.
EmptyDir volumes backed by huge pages may not consume more huge page memory
than the pod request.
Applications that consume huge pages via shmget() with SHM_HUGETLB must
run with a supplemental group that matches proc/sys/vm/hugetlb_shm_group.
Huge page usage in a namespace is controllable via ResourceQuota similar
to other compute resources like cpu or memory using the hugepages-<size>
token.
4.18 - Schedule GPUs
Configure and schedule GPUs for use as a resource by nodes in a cluster.
FEATURE STATE:Kubernetes v1.10 [beta]
Kubernetes includes experimental support for managing AMD and NVIDIA GPUs
(graphical processing units) across several nodes.
This page describes how users can consume GPUs across different Kubernetes versions
and the current limitations.
Using device plugins
Kubernetes implements Device Plugins
to let Pods access specialized hardware features such as GPUs.
As an administrator, you have to install GPU drivers from the corresponding
hardware vendor on the nodes and run the corresponding device plugin from the
GPU vendor:
When the above conditions are true, Kubernetes will expose amd.com/gpu or
nvidia.com/gpu as a schedulable resource.
You can consume these GPUs from your containers by requesting
<vendor>.com/gpu the same way you request cpu or memory.
However, there are some limitations in how you specify the resource requirements
when using GPUs:
GPUs are only supposed to be specified in the limits section, which means:
You can specify GPU limits without specifying requests because
Kubernetes will use the limit as the request value by default.
You can specify GPU in both limits and requests but these two values
must be equal.
You cannot specify GPU requests without specifying limits.
Containers (and Pods) do not share GPUs. There's no overcommitting of GPUs.
Each container can request one or more GPUs. It is not possible to request a
fraction of a GPU.
You can report issues with this third-party device plugin by logging an issue in
NVIDIA/k8s-device-plugin.
NVIDIA GPU device plugin used by GCE
The NVIDIA GPU device plugin used by GCE
doesn't require using nvidia-docker and should work with any container runtime
that is compatible with the Kubernetes Container Runtime Interface (CRI). It's tested
on Container-Optimized OS
and has experimental code for Ubuntu from 1.9 onwards.
You can use the following commands to install the NVIDIA drivers and device plugin:
# Install NVIDIA drivers on Container-Optimized OS:kubectl create -f https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/stable/daemonset.yaml
# Install NVIDIA drivers on Ubuntu (experimental):kubectl create -f https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/stable/nvidia-driver-installer/ubuntu/daemonset.yaml
# Install the device plugin:kubectl create -f https://raw.githubusercontent.com/kubernetes/kubernetes/release-1.14/cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml
Google publishes its own instructions for using NVIDIA GPUs on GKE .
Clusters containing different types of GPUs
If different nodes in your cluster have different types of GPUs, then you
can use Node Labels and Node Selectors
to schedule pods to appropriate nodes.
For example:
# Label your nodes with the accelerator type they have.kubectl label nodes <node-with-k80> accelerator=nvidia-tesla-k80
kubectl label nodes <node-with-p100> accelerator=nvidia-tesla-p100
Automatic node labelling
If you're using AMD GPU devices, you can deploy
Node Labeller.
Node Labeller is a controller that automatically
labels your nodes with GPU device properties.
At the moment, that controller can add labels for:
With the Node Labeller in use, you can specify the GPU type in the Pod spec:
apiVersion:v1kind:Podmetadata:name:cuda-vector-addspec:restartPolicy:OnFailurecontainers:- name:cuda-vector-add# https://github.com/kubernetes/kubernetes/blob/v1.7.11/test/images/nvidia-cuda/Dockerfileimage:"registry.k8s.io/cuda-vector-add:v0.1"resources:limits:nvidia.com/gpu:1nodeSelector:accelerator:nvidia-tesla-p100# or nvidia-tesla-k80 etc.
This will ensure that the Pod will be scheduled to a node that has the GPU type
you specified.
5 - Tutorials
This section of the Kubernetes documentation contains tutorials.
A tutorial shows how to accomplish a goal that is larger than a single
task. Typically a tutorial has several sections,
each of which has a sequence of steps.
Before walking through each tutorial, you may want to bookmark the
Standardized Glossary page for later references.
Basics
Kubernetes Basics is an in-depth interactive tutorial that helps you understand the Kubernetes system and try out some basic Kubernetes features.
If you would like to write a tutorial, see
Content Page Types
for information about the tutorial page type.
5.1 - Hello Minikube
This tutorial shows you how to run a sample app
on Kubernetes using minikube and Katacoda.
Katacoda provides a free, in-browser Kubernetes environment.
Note: You can also follow this tutorial if you've installed minikube locally.
See minikube start for installation instructions.
Objectives
Deploy a sample application to minikube.
Run the app.
View application logs.
Before you begin
This tutorial provides a container image that uses NGINX to echo back all the requests.
Create a minikube cluster
Click Launch Terminal
Note: If you installed minikube locally, run minikube start. Before you run minikube dashboard, you should open a new terminal, start minikube dashboard there, and then switch back to the main terminal.
Open the Kubernetes dashboard in a browser:
minikube dashboard
Katacoda environment only: At the top of the terminal pane, click the plus sign, and then click Select port to view on Host 1.
Katacoda environment only: Type 30000, and then click Display Port.
Note:
The dashboard command enables the dashboard add-on and opens the proxy in the default web browser.
You can create Kubernetes resources on the dashboard such as Deployment and Service.
By default, the dashboard is only accessible from within the internal Kubernetes virtual network.
The dashboard command creates a temporary proxy to make the dashboard accessible from outside the Kubernetes virtual network.
To stop the proxy, run Ctrl+C to exit the process.
After the command exits, the dashboard remains running in the Kubernetes cluster.
You can run the dashboard command again to create another proxy to access the dashboard.
Open Dashboard with URL
If you don't want to open a web browser, run the dashboard command with the --url flag to emit a URL:
minikube dashboard --url
Create a Deployment
A Kubernetes Pod is a group of one or more Containers,
tied together for the purposes of administration and networking. The Pod in this
tutorial has only one Container. A Kubernetes
Deployment checks on the health of your
Pod and restarts the Pod's Container if it terminates. Deployments are the
recommended way to manage the creation and scaling of Pods.
Use the kubectl create command to create a Deployment that manages a Pod. The
Pod runs a Container based on the provided Docker image.
NAME READY UP-TO-DATE AVAILABLE AGE
hello-node 1/1 1 1 1m
View the Pod:
kubectl get pods
The output is similar to:
NAME READY STATUS RESTARTS AGE
hello-node-5f76cf6ccf-br9b5 1/1 Running 0 1m
View cluster events:
kubectl get events
View the kubectl configuration:
kubectl config view
Note: For more information about kubectl commands, see the kubectl overview.
Create a Service
By default, the Pod is only accessible by its internal IP address within the
Kubernetes cluster. To make the hello-node Container accessible from outside the
Kubernetes virtual network, you have to expose the Pod as a
Kubernetes Service.
Expose the Pod to the public internet using the kubectl expose command:
The --type=LoadBalancer flag indicates that you want to expose your Service
outside of the cluster.
The application code inside the image registry.k8s.io/echoserver only listens on TCP port 8080. If you used
kubectl expose to expose a different port, clients could not connect to that other port.
View the Service you created:
kubectl get services
The output is similar to:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-node LoadBalancer 10.108.144.78 <pending> 8080:30369/TCP 21s
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 23m
On cloud providers that support load balancers,
an external IP address would be provisioned to access the Service. On minikube,
the LoadBalancer type makes the Service accessible through the minikube service
command.
Run the following command:
minikube service hello-node
Katacoda environment only: Click the plus sign, and then click Select port to view on Host 1.
Katacoda environment only: Note the 5-digit port number displayed opposite to 8080 in services output. This port number is randomly generated and it can be different for you. Type your number in the port number text box, then click Display Port. Using the example from earlier, you would type 30369.
This opens up a browser window that serves your app and shows the app's response.
Enable addons
The minikube tool includes a set of built-in addons that can be enabled, disabled and opened in the local Kubernetes environment.
This tutorial provides a walkthrough of the basics of the Kubernetes cluster orchestration system. Each module contains some background information on major Kubernetes features and concepts, and includes an interactive online tutorial. These interactive tutorials let you manage a simple cluster and its containerized applications for yourself.
Using the interactive tutorials, you can learn to:
Deploy a containerized application on a cluster.
Scale the deployment.
Update the containerized application with a new software version.
Debug the containerized application.
The tutorials use Katacoda to run a virtual terminal in your web browser that runs Minikube, a small-scale local deployment of Kubernetes that can run anywhere. There's no need to install any software or configure anything; each interactive tutorial runs directly out of your web browser itself.
What can Kubernetes do for you?
With modern web services, users expect applications to be available 24/7, and developers expect to deploy new versions of those applications several times a day. Containerization helps package software to serve these goals, enabling applications to be released and updated without downtime. Kubernetes helps you make sure those containerized applications run where and when you want, and helps them find the resources and tools they need to work. Kubernetes is a production-ready, open source platform designed with Google's accumulated experience in container orchestration, combined with best-of-breed ideas from the community.
Learn about Kubernetes cluster and create a simple cluster using Minikube.
5.2.1.1 - Using Minikube to Create a Cluster
Objectives
Learn what a Kubernetes cluster is.
Learn what Minikube is.
Start a Kubernetes cluster using an online terminal.
Kubernetes Clusters
Kubernetes coordinates a highly available cluster of computers that are connected to work as a single unit. The abstractions in Kubernetes allow you to deploy containerized applications to a cluster without tying them specifically to individual machines. To make use of this new model of deployment, applications need to be packaged in a way that decouples them from individual hosts: they need to be containerized. Containerized applications are more flexible and available than in past deployment models, where applications were installed directly onto specific machines as packages deeply integrated into the host. Kubernetes automates the distribution and scheduling of application containers across a cluster in a more efficient way. Kubernetes is an open-source platform and is production-ready.
A Kubernetes cluster consists of two types of resources:
The Control Plane coordinates the cluster
Nodes are the workers that run applications
Summary:
Kubernetes cluster
Minikube
Kubernetes is a production-grade, open-source platform that orchestrates the placement (scheduling) and execution of application containers within and across computer clusters.
Cluster Diagram
The Control Plane is responsible for managing the cluster. The Control Plane coordinates all activities in your cluster, such as scheduling applications, maintaining applications' desired state, scaling applications, and rolling out new updates.
A node is a VM or a physical computer that serves as a worker machine in a Kubernetes cluster. Each node has a Kubelet, which is an agent for managing the node and communicating with the Kubernetes control plane. The node should also have tools for handling container operations, such as containerd or Docker. A Kubernetes cluster that handles production traffic should have a minimum of three nodes because if one node goes down, both an etcd member and a control plane instance are lost, and redundancy is compromised. You can mitigate this risk by adding more control plane nodes.
Control Planes manage the cluster and the nodes that are used to host the running applications.
When you deploy applications on Kubernetes, you tell the control plane to start the application containers. The control plane schedules the containers to run on the cluster's nodes. The nodes communicate with the control plane using the Kubernetes API, which the control plane exposes. End users can also use the Kubernetes API directly to interact with the cluster.
A Kubernetes cluster can be deployed on either physical or virtual machines. To get started with Kubernetes development, you can use Minikube. Minikube is a lightweight Kubernetes implementation that creates a VM on your local machine and deploys a simple cluster containing only one node. Minikube is available for Linux, macOS, and Windows systems. The Minikube CLI provides basic bootstrapping operations for working with your cluster, including start, stop, status, and delete. For this tutorial, however, you'll use a provided online terminal with Minikube pre-installed.
Now that you know what Kubernetes is, let's go to the online tutorial and start our first cluster!
Once you have a running Kubernetes cluster, you can deploy your containerized applications on top of it.
To do so, you create a Kubernetes Deployment configuration. The Deployment instructs Kubernetes
how to create and update instances of your application. Once you've created a Deployment, the Kubernetes
control plane schedules the application instances included in that Deployment to run on individual Nodes in the
cluster.
Once the application instances are created, a Kubernetes Deployment Controller continuously monitors those instances. If the Node hosting an instance goes down or is deleted, the Deployment controller replaces the instance with an instance on another Node in the cluster. This provides a self-healing mechanism to address machine failure or maintenance.
In a pre-orchestration world, installation scripts would often be used to start applications, but they did not allow recovery from machine failure. By both creating your application instances and keeping them running across Nodes, Kubernetes Deployments provide a fundamentally different approach to application management.
Summary:
Deployments
Kubectl
A Deployment is responsible for creating and updating instances of your application
Deploying your first app on Kubernetes
You can create and manage a Deployment by using the Kubernetes command line interface, Kubectl. Kubectl uses the Kubernetes API to interact with the cluster. In this module, you'll learn the most common Kubectl commands needed to create Deployments that run your applications on a Kubernetes cluster.
When you create a Deployment, you'll need to specify the container image for your application and the number of replicas that you want to run. You can change that information later by updating your Deployment; Modules 5 and 6 of the bootcamp discuss how you can scale and update your Deployments.
Applications need to be packaged into one of the supported container formats in order to be deployed on Kubernetes
For your first Deployment, you'll use a hello-node application packaged in a Docker container that uses NGINX to echo back all the requests. (If you didn't already try creating a hello-node application and deploying it using a container, you can do that first by following the instructions from the Hello Minikube tutorial).
Now that you know what Deployments are, let's go to the online tutorial and deploy our first app!
A Pod is the basic execution unit of a Kubernetes application. Each Pod represents a part of a workload that is running on your cluster. Learn more about Pods.
To interact with the Terminal, please use the desktop/tablet version
When you created a Deployment in Module 2, Kubernetes created a Pod to host your application instance. A Pod is a Kubernetes abstraction that represents a group of one or more application containers (such as Docker), and some shared resources for those containers. Those resources include:
Shared storage, as Volumes
Networking, as a unique cluster IP address
Information about how to run each container, such as the container image version or specific ports to use
A Pod models an application-specific "logical host" and can contain different application containers which are relatively tightly coupled. For example, a Pod might include both the container with your Node.js app as well as a different container that feeds the data to be published by the Node.js webserver. The containers in a Pod share an IP Address and port space, are always co-located and co-scheduled, and run in a shared context on the same Node.
Pods are the atomic unit on the Kubernetes platform. When we create a Deployment on Kubernetes, that Deployment creates Pods with containers inside them (as opposed to creating containers directly). Each Pod is tied to the Node where it is scheduled, and remains there until termination (according to restart policy) or deletion. In case of a Node failure, identical Pods are scheduled on other available Nodes in the cluster.
Summary:
Pods
Nodes
Kubectl main commands
A Pod is a group of one or more application containers (such as Docker) and includes shared storage (volumes), IP address and information about how to run them.
Pods overview
Nodes
A Pod always runs on a Node. A Node is a worker machine in Kubernetes and may be either a virtual or a physical machine, depending on the cluster. Each Node is managed by the control plane. A Node can have multiple pods, and the Kubernetes control plane automatically handles scheduling the pods across the Nodes in the cluster. The control plane's automatic scheduling takes into account the available resources on each Node.
Every Kubernetes Node runs at least:
Kubelet, a process responsible for communication between the Kubernetes control plane and the Node; it manages the Pods and the containers running on a machine.
A container runtime (like Docker) responsible for pulling the container image from a registry, unpacking the container, and running the application.
Containers should only be scheduled together in a single Pod if they are tightly coupled and need to share resources such as disk.
Node overview
Troubleshooting with kubectl
In Module 2, you used Kubectl command-line interface. You'll continue to use it in Module 3 to get information about deployed applications and their environments. The most common operations can be done with the following kubectl commands:
kubectl get - list resources
kubectl describe - show detailed information about a resource
kubectl logs - print the logs from a container in a pod
kubectl exec - execute a command on a container in a pod
You can use these commands to see when applications were deployed, what their current statuses are, where they are running and what their configurations are.
Now that we know more about our cluster components and the command line, let's explore our application.
A node is a worker machine in Kubernetes and may be a VM or physical machine, depending on the cluster. Multiple Pods can run on one Node.
Understand how labels and LabelSelector objects relate to a Service
Expose an application outside a Kubernetes cluster using a Service
Overview of Kubernetes Services
Kubernetes Pods are mortal. Pods in fact have a lifecycle. When a worker node dies, the Pods running on the Node are also lost. A ReplicaSet might then dynamically drive the cluster back to desired state via creation of new Pods to keep your application running. As another example, consider an image-processing backend with 3 replicas. Those replicas are exchangeable; the front-end system should not care about backend replicas or even if a Pod is lost and recreated. That said, each Pod in a Kubernetes cluster has a unique IP address, even Pods on the same Node, so there needs to be a way of automatically reconciling changes among Pods so that your applications continue to function.
A Service in Kubernetes is an abstraction which defines a logical set of Pods and a policy by which to access them. Services enable a loose coupling between dependent Pods. A Service is defined using YAML (preferred) or JSON, like all Kubernetes objects. The set of Pods targeted by a Service is usually determined by a LabelSelector (see below for why you might want a Service without including selector in the spec).
Although each Pod has a unique IP address, those IPs are not exposed outside the cluster without a Service. Services allow your applications to receive traffic. Services can be exposed in different ways by specifying a type in the ServiceSpec:
ClusterIP (default) - Exposes the Service on an internal IP in the cluster. This type makes the Service only reachable from within the cluster.
NodePort - Exposes the Service on the same port of each selected Node in the cluster using NAT. Makes a Service accessible from outside the cluster using <NodeIP>:<NodePort>. Superset of ClusterIP.
LoadBalancer - Creates an external load balancer in the current cloud (if supported) and assigns a fixed, external IP to the Service. Superset of NodePort.
ExternalName - Maps the Service to the contents of the externalName field (e.g. foo.bar.example.com), by returning a CNAME record with its value. No proxying of any kind is set up. This type requires v1.7 or higher of kube-dns, or CoreDNS version 0.0.8 or higher.
Additionally, note that there are some use cases with Services that involve not defining selector in the spec. A Service created without selector will also not create the corresponding Endpoints object. This allows users to manually map a Service to specific endpoints. Another possibility why there may be no selector is you are strictly using type: ExternalName.
Summary
Exposing Pods to external traffic
Load balancing traffic across multiple Pods
Using labels
A Kubernetes Service is an abstraction layer which defines a logical set of Pods and enables external traffic exposure, load balancing and service discovery for those Pods.
Services and Labels
A Service routes traffic across a set of Pods. Services are the abstraction that allows pods to die and replicate in Kubernetes without impacting your application. Discovery and routing among dependent Pods (such as the frontend and backend components in an application) are handled by Kubernetes Services.
Services match a set of Pods using labels and selectors, a grouping primitive that allows logical operation on objects in Kubernetes. Labels are key/value pairs attached to objects and can be used in any number of ways:
Designate objects for development, test, and production
Embed version tags
Classify an object using tags
Labels can be attached to objects at creation time or later on. They can be modified at any time. Let's expose our application now using a Service and apply some labels.
In the previous modules we created a Deployment, and then exposed it publicly via a Service. The Deployment created only one Pod for running our application. When traffic increases, we will need to scale the application to keep up with user demand.
Scaling is accomplished by changing the number of replicas in a Deployment
Summary:
Scaling a Deployment
You can create from the start a Deployment with multiple instances using the --replicas parameter for the kubectl create deployment command
Scaling out a Deployment will ensure new Pods are created and scheduled to Nodes with available resources. Scaling will increase the number of Pods to the new desired state. Kubernetes also supports autoscaling of Pods, but it is outside of the scope of this tutorial. Scaling to zero is also possible, and it will terminate all Pods of the specified Deployment.
Running multiple instances of an application will require a way to distribute the traffic to all of them. Services have an integrated load-balancer that will distribute network traffic to all Pods of an exposed Deployment. Services will monitor continuously the running Pods using endpoints, to ensure the traffic is sent only to available Pods.
Scaling is accomplished by changing the number of replicas in a Deployment.
Once you have multiple instances of an Application running, you would be able to do Rolling updates without downtime. We'll cover that in the next module. Now, let's go to the online terminal and scale our application.
Users expect applications to be available all the time and developers are expected to deploy new versions of them several times a day. In Kubernetes this is done with rolling updates. Rolling updates allow Deployments' update to take place with zero downtime by incrementally updating Pods instances with new ones. The new Pods will be scheduled on Nodes with available resources.
In the previous module we scaled our application to run multiple instances. This is a requirement for performing updates without affecting application availability. By default, the maximum number of Pods that can be unavailable during the update and the maximum number of new Pods that can be created, is one. Both options can be configured to either numbers or percentages (of Pods).
In Kubernetes, updates are versioned and any Deployment update can be reverted to a previous (stable) version.
Summary:
Updating an app
Rolling updates allow Deployments' update to take place with zero downtime by incrementally updating Pods instances with new ones.
Similar to application Scaling, if a Deployment is exposed publicly, the Service will load-balance the traffic only to available Pods during the update. An available Pod is an instance that is available to the users of the application.
Rolling updates allow the following actions:
Promote an application from one environment to another (via container image updates)
Rollback to previous versions
Continuous Integration and Continuous Delivery of applications with zero downtime
If a Deployment is exposed publicly, the Service will load-balance the traffic only to available Pods during the update.
In the following interactive tutorial, we'll update our application to a new version, and also perform a rollback.
5.3.1.1 - Externalizing config using MicroProfile, ConfigMaps and Secrets
In this tutorial you will learn how and why to externalize your microservice’s configuration.
Specifically, you will learn how to use Kubernetes ConfigMaps and Secrets to set environment
variables and then consume them using MicroProfile Config.
Before you begin
Creating Kubernetes ConfigMaps & Secrets
There are several ways to set environment variables for a Docker container in Kubernetes,
including: Dockerfile, kubernetes.yml, Kubernetes ConfigMaps, and Kubernetes Secrets. In the
tutorial, you will learn how to use the latter two for setting your environment variables whose
values will be injected into your microservices. One of the benefits for using ConfigMaps and
Secrets is that they can be re-used across multiple containers, including being assigned to
different environment variables for the different containers.
ConfigMaps are API Objects that store non-confidential key-value pairs. In the Interactive
Tutorial you will learn how to use a ConfigMap to store the application's name. For more
information regarding ConfigMaps, you can find the documentation
here.
Although Secrets are also used to store key-value pairs, they differ from ConfigMaps in that
they're intended for confidential/sensitive information and are stored using Base64 encoding.
This makes secrets the appropriate choice for storing such things as credentials, keys, and
tokens, the former of which you'll do in the Interactive Tutorial. For more information on
Secrets, you can find the documentation here.
Externalizing Config from Code
Externalized application configuration is useful because configuration usually changes depending
on your environment. In order to accomplish this, we'll use Java's Contexts and Dependency
Injection (CDI) and MicroProfile Config. MicroProfile Config is a feature of MicroProfile, a set
of open Java technologies for developing and deploying cloud-native microservices.
CDI provides a standard dependency injection capability enabling an application to be assembled
from collaborating, loosely-coupled beans. MicroProfile Config provides apps and microservices a
standard way to obtain config properties from various sources, including the application, runtime,
and environment. Based on the source's defined priority, the properties are automatically
combined into a single set of properties that the application can access via an API. Together,
CDI & MicroProfile will be used in the Interactive Tutorial to retrieve the externally provided
properties from the Kubernetes ConfigMaps and Secrets and get injected into your application code.
Many open source frameworks and runtimes implement and support MicroProfile Config. Throughout
the interactive tutorial, you'll be using Open Liberty, a flexible open-source Java runtime for
building and running cloud-native apps and microservices. However, any MicroProfile compatible
runtime could be used instead.
Objectives
Create a Kubernetes ConfigMap and Secret
Inject microservice configuration using MicroProfile Config
Example: Externalizing config using MicroProfile, ConfigMaps and Secrets
5.3.1.2 - Interactive Tutorial - Configuring a Java Microservice
To interact with the Terminal, please use the desktop/tablet version
5.3.2 - Configuring Redis using a ConfigMap
This page provides a real world example of how to configure Redis using a ConfigMap and builds upon the Configure a Pod to Use a ConfigMap task.
Objectives
Create a ConfigMap with Redis configuration values
Create a Redis Pod that mounts and uses the created ConfigMap
Verify that the configuration was correctly applied.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Examine the contents of the Redis pod manifest and note the following:
A volume named config is created by spec.volumes[1]
The key and path under spec.volumes[1].items[0] exposes the redis-config key from the
example-redis-config ConfigMap as a file named redis.conf on the config volume.
The config volume is then mounted at /redis-master by spec.containers[0].volumeMounts[1].
This has the net effect of exposing the data in data.redis-config from the example-redis-config
ConfigMap above as /redis-master/redis.conf inside the Pod.
Check the Redis Pod again using redis-cli via kubectl exec to see if the configuration was applied:
kubectl exec -it redis -- redis-cli
Check maxmemory:
127.0.0.1:6379> CONFIG GET maxmemory
It remains at the default value of 0:
1)"maxmemory"2)"0"
Similarly, maxmemory-policy remains at the noeviction default setting:
127.0.0.1:6379> CONFIG GET maxmemory-policy
Returns:
1)"maxmemory-policy"2)"noeviction"
The configuration values have not changed because the Pod needs to be restarted to grab updated
values from associated ConfigMaps. Let's delete and recreate the Pod:
kubectl delete pod redis
kubectl apply -f https://raw.githubusercontent.com/kubernetes/website/main/content/en/examples/pods/config/redis-pod.yaml
Now re-check the configuration values one last time:
kubectl exec -it redis -- redis-cli
Check maxmemory:
127.0.0.1:6379> CONFIG GET maxmemory
It should now return the updated value of 2097152:
1)"maxmemory"2)"2097152"
Similarly, maxmemory-policy has also been updated:
127.0.0.1:6379> CONFIG GET maxmemory-policy
It now reflects the desired value of allkeys-lru:
1)"maxmemory-policy"2)"allkeys-lru"
Clean up your work by deleting the created resources:
5.4.1 - Apply Pod Security Standards at the Cluster Level
Note
This tutorial applies only for new clusters.
Pod Security admission (PSA) is enabled by default in v1.23 and later, as it has
graduated to beta.
Pod Security
is an admission controller that carries out checks against the Kubernetes
Pod Security Standards when new pods are
created. This tutorial shows you how to enforce the baseline Pod Security
Standard at the cluster level which applies a standard configuration
to all namespaces in a cluster.
Kubernetes control plane is running at https://127.0.0.1:61350
CoreDNS is running at https://127.0.0.1:61350/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
Get a list of namespaces in the cluster:
kubectl get ns
The output is similar to this:
NAME STATUS AGE
default Active 9m30s
kube-node-lease Active 9m32s
kube-public Active 9m32s
kube-system Active 9m32s
local-path-storage Active 9m26s
Use --dry-run=server to understand what happens when different Pod Security Standards
are applied:
From the previous output, you'll notice that applying the privileged Pod Security Standard shows no warnings
for any namespaces. However, baseline and restricted standards both have
warnings, specifically in the kube-system namespace.
Set modes, versions and standards
In this section, you apply the following Pod Security Standards to the latest version:
baseline standard in enforce mode.
restricted standard in warn and audit mode.
The baseline Pod Security Standard provides a convenient
middle ground that allows keeping the exemption list short and prevents known
privilege escalations.
Additionally, to prevent pods from failing in kube-system, you'll exempt the namespace
from having Pod Security Standards applied.
When you implement Pod Security Admission in your own environment, consider the
following:
Based on the risk posture applied to a cluster, a stricter Pod Security
Standard like restricted might be a better choice.
Exempting the kube-system namespace allows pods to run as
privileged in this namespace. For real world use, the Kubernetes project
strongly recommends that you apply strict RBAC
policies that limit access to kube-system, following the principle of least
privilege.
To implement the preceding standards, do the following:
Create a configuration file that can be consumed by the Pod Security
Admission Controller to implement these Pod Security Standards:
Creating cluster "psa-with-cluster-pss" ...
✓ Ensuring node image (kindest/node:v1.24.0) 🖼
✓ Preparing nodes 📦
✓ Writing configuration 📜
✓ Starting control-plane 🕹️
✓ Installing CNI 🔌
✓ Installing StorageClass 💾
Set kubectl context to "kind-psa-with-cluster-pss"
You can now use your cluster with:
kubectl cluster-info --context kind-psa-with-cluster-pss
Have a question, bug, or feature request? Let us know! https://kind.sigs.k8s.io/#community 🙂
Kubernetes control plane is running at https://127.0.0.1:63855
CoreDNS is running at https://127.0.0.1:63855/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
Create the following Pod specification for a minimal configuration in the default namespace:
Warning: would violate PodSecurity "restricted:latest": allowPrivilegeEscalation != false (container "nginx" must set securityContext.allowPrivilegeEscalation=false), unrestricted capabilities (container "nginx" must set securityContext.capabilities.drop=["ALL"]), runAsNonRoot != true (pod or container "nginx" must set securityContext.runAsNonRoot=true), seccompProfile (pod or container "nginx" must set securityContext.seccompProfile.type to "RuntimeDefault" or "Localhost")
pod/nginx created
Clean up
Run kind delete cluster --name psa-with-cluster-pss and
kind delete cluster --name psa-wo-cluster-pss to delete the clusters you
created.
What's next
Run a
shell script
to perform all the preceding steps at once:
Create a Pod Security Standards based cluster level Configuration
Create a file to let API server consumes this configuration
Create a cluster that creates an API server with this configuration
Set kubectl context to this new cluster
Create a minimal pod yaml file
Apply this file to create a Pod in the new cluster
5.4.2 - Apply Pod Security Standards at the Namespace Level
Note
This tutorial applies only for new clusters.
Pod Security admission (PSA) is enabled by default in v1.23 and later, as it
graduated to beta. Pod Security Admission
is an admission controller that applies
Pod Security Standards
when pods are created. In this tutorial, you will enforce the baseline Pod Security Standard,
one namespace at a time.
Creating cluster "psa-ns-level" ...
✓ Ensuring node image (kindest/node:v1.23.0) 🖼
✓ Preparing nodes 📦
✓ Writing configuration 📜
✓ Starting control-plane 🕹️
✓ Installing CNI 🔌
✓ Installing StorageClass 💾
Set kubectl context to "kind-psa-ns-level"
You can now use your cluster with:
kubectl cluster-info --context kind-psa-ns-level
Not sure what to do next? 😅 Check out https://kind.sigs.k8s.io/docs/user/quick-start/
Set the kubectl context to the new cluster:
kubectl cluster-info --context kind-psa-ns-level
The output is similar to this:
Kubernetes control plane is running at https://127.0.0.1:50996
CoreDNS is running at https://127.0.0.1:50996/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
Create a namespace
Create a new namespace called example:
kubectl create ns example
The output is similar to this:
namespace/example created
Apply Pod Security Standards
Enable Pod Security Standards on this namespace using labels supported by
built-in Pod Security Admission. In this step we will warn on baseline pod
security standard as per the latest version (default value)
kubectl label --overwrite ns example \
pod-security.kubernetes.io/warn=baseline \
pod-security.kubernetes.io/warn-version=latest
Multiple pod security standards can be enabled on any namespace, using labels.
Following command will enforce the baseline Pod Security Standard, but
warn and audit for restricted Pod Security Standards as per the latest
version (default value)
Apply the pod spec to the cluster in example namespace:
kubectl apply -n example -f /tmp/pss/nginx-pod.yaml
The output is similar to this:
Warning: would violate PodSecurity "restricted:latest": allowPrivilegeEscalation != false (container "nginx" must set securityContext.allowPrivilegeEscalation=false), unrestricted capabilities (container "nginx" must set securityContext.capabilities.drop=["ALL"]), runAsNonRoot != true (pod or container "nginx" must set securityContext.runAsNonRoot=true), seccompProfile (pod or container "nginx" must set securityContext.seccompProfile.type to "RuntimeDefault" or "Localhost")
pod/nginx created
Apply the pod spec to the cluster in default namespace:
5.4.3 - Restrict a Container's Access to Resources with AppArmor
FEATURE STATE:Kubernetes v1.4 [beta]
AppArmor is a Linux kernel security module that supplements the standard Linux user and group based
permissions to confine programs to a limited set of resources. AppArmor can be configured for any
application to reduce its potential attack surface and provide greater in-depth defense. It is
configured through profiles tuned to allow the access needed by a specific program or container,
such as Linux capabilities, network access, file permissions, etc. Each profile can be run in either
enforcing mode, which blocks access to disallowed resources, or complain mode, which only reports
violations.
AppArmor can help you to run a more secure deployment by restricting what containers are allowed to
do, and/or provide better auditing through system logs. However, it is important to keep in mind
that AppArmor is not a silver bullet and can only do so much to protect against exploits in your
application code. It is important to provide good, restrictive profiles, and harden your
applications and cluster from other angles as well.
Objectives
See an example of how to load a profile on a node
Learn how to enforce the profile on a Pod
Learn how to check that the profile is loaded
See what happens when a profile is violated
See what happens when a profile cannot be loaded
Before you begin
Make sure:
Kubernetes version is at least v1.4 -- Kubernetes support for AppArmor was added in
v1.4. Kubernetes components older than v1.4 are not aware of the new AppArmor annotations, and
will silently ignore any AppArmor settings that are provided. To ensure that your Pods are
receiving the expected protections, it is important to verify the Kubelet version of your nodes:
kubectl get nodes -o=jsonpath=$'{range .items[*]}{@.metadata.name}: {@.status.nodeInfo.kubeletVersion}\n{end}'
AppArmor kernel module is enabled -- For the Linux kernel to enforce an AppArmor profile, the
AppArmor kernel module must be installed and enabled. Several distributions enable the module by
default, such as Ubuntu and SUSE, and many others provide optional support. To check whether the
module is enabled, check the /sys/module/apparmor/parameters/enabled file:
cat /sys/module/apparmor/parameters/enabled
Y
If the Kubelet contains AppArmor support (>= v1.4), it will refuse to run a Pod with AppArmor
options if the kernel module is not enabled.
Note: Ubuntu carries many AppArmor patches that have not been merged into the upstream Linux
kernel, including patches that add additional hooks and features. Kubernetes has only been
tested with the upstream version, and does not promise support for other features.
Container runtime supports AppArmor -- Currently all common Kubernetes-supported container
runtimes should support AppArmor, like Docker,
CRI-O or containerd.
Please refer to the corresponding runtime documentation and verify that the cluster fulfills
the requirements to use AppArmor.
Profile is loaded -- AppArmor is applied to a Pod by specifying an AppArmor profile that each
container should be run with. If any of the specified profiles is not already loaded in the
kernel, the Kubelet (>= v1.4) will reject the Pod. You can view which profiles are loaded on a
node by checking the /sys/kernel/security/apparmor/profiles file. For example:
As long as the Kubelet version includes AppArmor support (>= v1.4), the Kubelet will reject a Pod
with AppArmor options if any of the prerequisites are not met. You can also verify AppArmor support
on nodes by checking the node ready condition message (though this is likely to be removed in a
later release):
kubectl get nodes -o=jsonpath=$'{range .items[*]}{@.metadata.name}: {.status.conditions[?(@.reason=="KubeletReady")].message}\n{end}'
gke-test-default-pool-239f5d02-gyn2: kubelet is posting ready status. AppArmor enabled
gke-test-default-pool-239f5d02-x1kf: kubelet is posting ready status. AppArmor enabled
gke-test-default-pool-239f5d02-xwux: kubelet is posting ready status. AppArmor enabled
Securing a Pod
Note: AppArmor is currently in beta, so options are specified as annotations. Once support graduates to
general availability, the annotations will be replaced with first-class fields (more details in
Upgrade path to GA).
AppArmor profiles are specified per-container. To specify the AppArmor profile to run a Pod
container with, add an annotation to the Pod's metadata:
Where <container_name> is the name of the container to apply the profile to, and <profile_ref>
specifies the profile to apply. The profile_ref can be one of:
runtime/default to apply the runtime's default profile
localhost/<profile_name> to apply the profile loaded on the host with the name <profile_name>
unconfined to indicate that no profiles will be loaded
See the API Reference for the full details on the annotation and profile name formats.
Kubernetes AppArmor enforcement works by first checking that all the prerequisites have been
met, and then forwarding the profile selection to the container runtime for enforcement. If the
prerequisites have not been met, the Pod will be rejected, and will not run.
To verify that the profile was applied, you can look for the AppArmor security option listed in the container created event:
kubectl get events | grep Created
22s 22s 1 hello-apparmor Pod spec.containers{hello} Normal Created {kubelet e2e-test-stclair-node-pool-31nt} Created container with docker id 269a53b202d3; Security:[seccomp=unconfined apparmor=k8s-apparmor-example-deny-write]
You can also verify directly that the container's root process is running with the correct profile by checking its proc attr:
kubectl exec <pod_name> cat /proc/1/attr/current
k8s-apparmor-example-deny-write (enforce)
Example
This example assumes you have already set up a cluster with AppArmor support.
First, we need to load the profile we want to use onto our nodes. This profile denies all file writes:
Since we don't know where the Pod will be scheduled, we'll need to load the profile on all our
nodes. For this example we'll use SSH to install the profiles, but other approaches are
discussed in Setting up nodes with profiles.
NODES=(# The SSH-accessible domain names of your nodes gke-test-default-pool-239f5d02-gyn2.us-central1-a.my-k8s
gke-test-default-pool-239f5d02-x1kf.us-central1-a.my-k8s
gke-test-default-pool-239f5d02-xwux.us-central1-a.my-k8s)for NODE in ${NODES[*]}; do ssh $NODE'sudo apparmor_parser -q <<EOF
#include <tunables/global>
profile k8s-apparmor-example-deny-write flags=(attach_disconnected) {
#include <abstractions/base>
file,
# Deny all file writes.
deny /** w,
}
EOF'done
Next, we'll run a simple "Hello AppArmor" pod with the deny-write profile:
apiVersion:v1kind:Podmetadata:name:hello-apparmorannotations:# Tell Kubernetes to apply the AppArmor profile "k8s-apparmor-example-deny-write".# Note that this is ignored if the Kubernetes node is not running version 1.4 or greater.container.apparmor.security.beta.kubernetes.io/hello:localhost/k8s-apparmor-example-deny-writespec:containers:- name:helloimage:busybox:1.28command:["sh","-c","echo 'Hello AppArmor!' && sleep 1h"]
kubectl create -f ./hello-apparmor.yaml
If we look at the pod events, we can see that the Pod container was created with the AppArmor
profile "k8s-apparmor-example-deny-write":
kubectl get events | grep hello-apparmor
14s 14s 1 hello-apparmor Pod Normal Scheduled {default-scheduler } Successfully assigned hello-apparmor to gke-test-default-pool-239f5d02-gyn2
14s 14s 1 hello-apparmor Pod spec.containers{hello} Normal Pulling {kubelet gke-test-default-pool-239f5d02-gyn2} pulling image "busybox"
13s 13s 1 hello-apparmor Pod spec.containers{hello} Normal Pulled {kubelet gke-test-default-pool-239f5d02-gyn2} Successfully pulled image "busybox"
13s 13s 1 hello-apparmor Pod spec.containers{hello} Normal Created {kubelet gke-test-default-pool-239f5d02-gyn2} Created container with docker id 06b6cd1c0989; Security:[seccomp=unconfined apparmor=k8s-apparmor-example-deny-write]
13s 13s 1 hello-apparmor Pod spec.containers{hello} Normal Started {kubelet gke-test-default-pool-239f5d02-gyn2} Started container with docker id 06b6cd1c0989
We can verify that the container is actually running with that profile by checking its proc attr:
To wrap up, let's look at what happens if we try to specify a profile that hasn't been loaded:
kubectl create -f /dev/stdin <<EOF
apiVersion:v1kind:Podmetadata:name:hello-apparmor-2annotations:container.apparmor.security.beta.kubernetes.io/hello:localhost/k8s-apparmor-example-allow-writespec:containers:- name:helloimage:busybox:1.28command:["sh","-c","echo 'Hello AppArmor!' && sleep 1h"]EOFpod/hello-apparmor-2 created
kubectl describe pod hello-apparmor-2
Name: hello-apparmor-2
Namespace: default
Node: gke-test-default-pool-239f5d02-x1kf/
Start Time: Tue, 30 Aug 2016 17:58:56 -0700
Labels: <none>
Annotations: container.apparmor.security.beta.kubernetes.io/hello=localhost/k8s-apparmor-example-allow-write
Status: Pending
Reason: AppArmor
Message: Pod Cannot enforce AppArmor: profile "k8s-apparmor-example-allow-write" is not loaded
IP:
Controllers: <none>
Containers:
hello:
Container ID:
Image: busybox
Image ID:
Port:
Command:
sh
-c
echo 'Hello AppArmor!' && sleep 1h
State: Waiting
Reason: Blocked
Ready: False
Restart Count: 0
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from default-token-dnz7v (ro)
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
Volumes:
default-token-dnz7v:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-dnz7v
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: <none>
Events:
FirstSeen LastSeen Count From SubobjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
23s 23s 1 {default-scheduler } Normal Scheduled Successfully assigned hello-apparmor-2 to e2e-test-stclair-node-pool-t1f5
23s 23s 1 {kubelet e2e-test-stclair-node-pool-t1f5} Warning AppArmor Cannot enforce AppArmor: profile "k8s-apparmor-example-allow-write" is not loaded
Note the pod status is Pending, with a helpful error message: Pod Cannot enforce AppArmor: profile "k8s-apparmor-example-allow-write" is not loaded. An event was also recorded with the same message.
Administration
Setting up nodes with profiles
Kubernetes does not currently provide any native mechanisms for loading AppArmor profiles onto
nodes. There are lots of ways to set up the profiles though, such as:
Through a DaemonSet that runs a Pod on each node to
ensure the correct profiles are loaded. An example implementation can be found
here.
At node initialization time, using your node initialization scripts (e.g. Salt, Ansible, etc.) or
image.
By copying the profiles to each node and loading them through SSH, as demonstrated in the
Example.
The scheduler is not aware of which profiles are loaded onto which node, so the full set of profiles
must be loaded onto every node. An alternative approach is to add a node label for each profile (or
class of profiles) on the node, and use a
node selector to ensure the Pod is run on a
node with the required profile.
Disabling AppArmor
If you do not want AppArmor to be available on your cluster, it can be disabled by a command-line flag:
--feature-gates=AppArmor=false
When disabled, any Pod that includes an AppArmor profile will fail validation with a "Forbidden"
error.
Note: Even if the Kubernetes feature is disabled, runtimes may still enforce the default profile. The
option to disable the AppArmor feature will be removed when AppArmor graduates to general
availability (GA).
Authoring Profiles
Getting AppArmor profiles specified correctly can be a tricky business. Fortunately there are some
tools to help with that:
aa-genprof and aa-logprof generate profile rules by monitoring an application's activity and
logs, and admitting the actions it takes. Further instructions are provided by the
AppArmor documentation.
bane is an AppArmor profile generator for Docker that uses a
simplified profile language.
To debug problems with AppArmor, you can check the system logs to see what, specifically, was
denied. AppArmor logs verbose messages to dmesg, and errors can usually be found in the system
logs or through journalctl. More information is provided in
AppArmor failures.
API Reference
Pod Annotation
Specifying the profile a container will run with:
key: container.apparmor.security.beta.kubernetes.io/<container_name>
Where <container_name> matches the name of a container in the Pod.
A separate profile can be specified for each container in the Pod.
value: a profile reference, described below
Profile Reference
runtime/default: Refers to the default runtime profile.
Equivalent to not specifying a profile, except it still
requires AppArmor to be enabled.
5.4.4 - Restrict a Container's Syscalls with seccomp
FEATURE STATE:Kubernetes v1.19 [stable]
Seccomp stands for secure computing mode and has been a feature of the Linux
kernel since version 2.6.12. It can be used to sandbox the privileges of a
process, restricting the calls it is able to make from userspace into the
kernel. Kubernetes lets you automatically apply seccomp profiles loaded onto a
node to your Pods and containers.
Identifying the privileges required for your workloads can be difficult. In this
tutorial, you will go through how to load seccomp profiles into a local
Kubernetes cluster, how to apply them to a Pod, and how you can begin to craft
profiles that give only the necessary privileges to your container processes.
Objectives
Learn how to load seccomp profiles on a node
Learn how to apply a seccomp profile to a container
Observe auditing of syscalls made by a container process
Observe behavior when a missing profile is specified
Observe a violation of a seccomp profile
Learn how to create fine-grained seccomp profiles
Learn how to apply a container runtime default seccomp profile
Before you begin
In order to complete all steps in this tutorial, you must install
kind and kubectl.
This tutorial shows some examples that are still beta (since v1.25) and
others that use only generally available seccomp functionality. You should
make sure that your cluster is
configured correctly
for the version you are using.
The tutorial also uses the curl tool for downloading examples to your computer.
You can adapt the steps to use a different tool if you prefer.
Note: It is not possible to apply a seccomp profile to a container running with
privileged: true set in the container's securityContext. Privileged containers always
run as Unconfined.
Download example seccomp profiles
The contents of these profiles will be explored later on, but for now go ahead
and download them into a directory named profiles/ so that they can be loaded
into the cluster.
You should see three profiles listed at the end of the final step:
audit.json fine-grained.json violation.json
Create a local Kubernetes cluster with kind
For simplicity, kind can be used to create a single
node cluster with the seccomp profiles loaded. Kind runs Kubernetes in Docker,
so each node of the cluster is a container. This allows for files
to be mounted in the filesystem of each container similar to loading files
onto a node.
You can set a specific Kubernetes version by setting the node's container image.
See Nodes within the
kind documentation about configuration for more details on this.
This tutorial assumes you are using Kubernetes v1.25.
Once you have a kind configuration in place, create the kind cluster with
that configuration:
kind create cluster --config=kind.yaml
After the new Kubernetes cluster is ready, identify the Docker container running
as the single node cluster:
docker ps
You should see output indicating that a container is running with name
kind-control-plane. The output is similar to:
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
6a96207fed4b kindest/node:v1.18.2 "/usr/local/bin/entr…" 27 seconds ago Up 24 seconds 127.0.0.1:42223->6443/tcp kind-control-plane
If observing the filesystem of that container, you should see that the
profiles/ directory has been successfully loaded into the default seccomp path
of the kubelet. Use docker exec to run a command in the Pod:
# Change 6a96207fed4b to the container ID you saw from "docker ps"docker exec -it 6a96207fed4b ls /var/lib/kubelet/seccomp/profiles
audit.json fine-grained.json violation.json
You have verified that these seccomp profiles are available to the kubelet
running within kind.
Enable the use of RuntimeDefault as the default seccomp profile for all workloads
FEATURE STATE:Kubernetes v1.25 [beta]
To use seccomp profile defaulting, you must run the kubelet with the SeccompDefaultfeature gate enabled
(this is the default). You must also explicitly enable the defaulting behavior for each
node where you want to use this with the corresponding --seccomp-defaultcommand line flag.
Both have to be enabled simultaneously to use the feature.
If enabled, the kubelet will use the RuntimeDefault seccomp profile by default, which is
defined by the container runtime, instead of using the Unconfined (seccomp disabled) mode.
The default profiles aim to provide a strong set
of security defaults while preserving the functionality of the workload. It is
possible that the default profiles differ between container runtimes and their
release versions, for example when comparing those from CRI-O and containerd.
Note: Enabling the feature will neither change the Kubernetes
securityContext.seccompProfile API field nor add the deprecated annotations of
the workload. This provides users the possibility to rollback anytime without
actually changing the workload configuration. Tools like
crictl inspect can be used to
verify which seccomp profile is being used by a container.
Some workloads may require a lower amount of syscall restrictions than others.
This means that they can fail during runtime even with the RuntimeDefault
profile. To mitigate such a failure, you can:
Run the workload explicitly as Unconfined.
Disable the SeccompDefault feature for the nodes. Also making sure that
workloads get scheduled on nodes where the feature is disabled.
Create a custom seccomp profile for the workload.
If you were introducing this feature into production-like cluster, the Kubernetes project
recommends that you enable this feature gate on a subset of your nodes and then
test workload execution before rolling the change out cluster-wide.
You can find more detailed information about a possible upgrade and downgrade strategy
in the related Kubernetes Enhancement Proposal (KEP):
Enable seccomp by default.
Kubernetes 1.25 lets you configure the seccomp profile
that applies when the spec for a Pod doesn't define a specific seccomp profile.
This is a beta feature and the corresponding SeccompDefaultfeature
gate is enabled by
default. However, you still need to enable this defaulting for each node where
you would like to use it.
If you are running a Kubernetes 1.25 cluster and want to
enable the feature, either run the kubelet with the --seccomp-default command
line flag, or enable it through the kubelet configuration
file. To enable the
feature gate in kind, ensure that kind provides
the minimum required Kubernetes version and enables the SeccompDefault feature
in the kind configuration:
kubectl run --rm -it --restart=Never --image=alpine alpine -- sh
Should now have the default seccomp profile attached. This can be verified by
using docker exec to run crictl inspect for the container on the kind
worker:
apiVersion:v1kind:Podmetadata:name:audit-podlabels:app:audit-podspec:securityContext:seccompProfile:type:LocalhostlocalhostProfile:profiles/audit.jsoncontainers:- name:test-containerimage:hashicorp/http-echo:0.2.3args:- "-text=just made some syscalls!"securityContext:allowPrivilegeEscalation:false
Note:
The functional support for the already deprecated seccomp annotations
seccomp.security.alpha.kubernetes.io/pod (for the whole pod) and
container.seccomp.security.alpha.kubernetes.io/[name] (for a single container)
is going to be removed with a future release of Kubernetes. Please always use
the native API fields in favor of the annotations.
Since Kubernetes v1.25, kubelets no longer support the annotations, use of the
annotations in static pods is no longer supported, and the seccomp annotations
are no longer auto-populated when pods with seccomp fields are created.
Auto-population of the seccomp fields from the annotations is planned to be
removed in a future release.
This profile does not restrict any syscalls, so the Pod should start
successfully.
kubectl get pod/audit-pod
NAME READY STATUS RESTARTS AGE
audit-pod 1/1 Running 0 30s
In order to be able to interact with this endpoint exposed by this
container, create a NodePort Services
that allows access to the endpoint from inside the kind control plane container.
kubectl expose pod audit-pod --type NodePort --port 5678
Check what port the Service has been assigned on the node.
kubectl get service audit-pod
The output is similar to:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
audit-pod NodePort 10.111.36.142 <none> 5678:32373/TCP 72s
Now you can use curl to access that endpoint from inside the kind control plane container,
at the port exposed by this Service. Use docker exec to run the curl command within the
container belonging to that control plane container:
# Change 6a96207fed4b to the control plane container ID you saw from "docker ps"docker exec -it 6a96207fed4b curl localhost:32373
just made some syscalls!
You can see that the process is running, but what syscalls did it actually make?
Because this Pod is running in a local cluster, you should be able to see those
in /var/log/syslog. Open up a new terminal window and tail the output for
calls from http-echo:
tail -f /var/log/syslog | grep 'http-echo'
You should already see some logs of syscalls made by http-echo, and if you
curl the endpoint in the control plane container you will see more written.
You can begin to understand the syscalls required by the http-echo process by
looking at the syscall= entry on each line. While these are unlikely to
encompass all syscalls it uses, it can serve as a basis for a seccomp profile
for this container.
Clean up that Pod and Service before moving to the next section:
kubectl delete service audit-pod --wait
kubectl delete pod audit-pod --wait --now
Create Pod with a seccomp profile that causes violation
For demonstration, apply a profile to the Pod that does not allow for any
syscalls.
apiVersion:v1kind:Podmetadata:name:violation-podlabels:app:violation-podspec:securityContext:seccompProfile:type:LocalhostlocalhostProfile:profiles/violation.jsoncontainers:- name:test-containerimage:hashicorp/http-echo:0.2.3args:- "-text=just made some syscalls!"securityContext:allowPrivilegeEscalation:false
The Pod creates, but there is an issue.
If you check the status of the Pod, you should see that it failed to start.
kubectl get pod/violation-pod
NAME READY STATUS RESTARTS AGE
violation-pod 0/1 CrashLoopBackOff 1 6s
As seen in the previous example, the http-echo process requires quite a few
syscalls. Here seccomp has been instructed to error on any syscall by setting
"defaultAction": "SCMP_ACT_ERRNO". This is extremely secure, but removes the
ability to do anything meaningful. What you really want is to give workloads
only the privileges they need.
Clean up that Pod before moving to the next section:
kubectl delete pod violation-pod --wait --now
Create Pod with a seccomp profile that only allows necessary syscalls
If you take a look at the fine-grained.json profile, you will notice some of the syscalls
seen in syslog of the first example where the profile set "defaultAction": "SCMP_ACT_LOG". Now the profile is setting "defaultAction": "SCMP_ACT_ERRNO",
but explicitly allowing a set of syscalls in the "action": "SCMP_ACT_ALLOW"
block. Ideally, the container will run successfully and you will see no messages
sent to syslog.
apiVersion:v1kind:Podmetadata:name:fine-podlabels:app:fine-podspec:securityContext:seccompProfile:type:LocalhostlocalhostProfile:profiles/fine-grained.jsoncontainers:- name:test-containerimage:hashicorp/http-echo:0.2.3args:- "-text=just made some syscalls!"securityContext:allowPrivilegeEscalation:false
The Pod should be showing as having started successfully:
NAME READY STATUS RESTARTS AGE
fine-pod 1/1 Running 0 30s
Open up a new terminal window and use tail to monitor for log entries that
mention calls from http-echo:
# The log path on your computer might be different from "/var/log/syslog"tail -f /var/log/syslog | grep 'http-echo'
Next, expose the Pod with a NodePort Service:
kubectl expose pod fine-pod --type NodePort --port 5678
Check what port the Service has been assigned on the node:
kubectl get service fine-pod
The output is similar to:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
fine-pod NodePort 10.111.36.142 <none> 5678:32373/TCP 72s
Use curl to access that endpoint from inside the kind control plane container:
# Change 6a96207fed4b to the control plane container ID you saw from "docker ps"docker exec -it 6a96207fed4b curl localhost:32373
just made some syscalls!
You should see no output in the syslog. This is because the profile allowed all
necessary syscalls and specified that an error should occur if one outside of
the list is invoked. This is an ideal situation from a security perspective, but
required some effort in analyzing the program. It would be nice if there was a
simple way to get closer to this security without requiring as much effort.
Clean up that Pod and Service before moving to the next section:
kubectl delete service fine-pod --wait
kubectl delete pod fine-pod --wait --now
Create Pod that uses the container runtime default seccomp profile
Most container runtimes provide a sane set of default syscalls that are allowed
or not. You can adopt these defaults for your workload by setting the seccomp
type in the security context of a pod or container to RuntimeDefault.
Note: If you have the SeccompDefaultfeature gate enabled, then Pods use the RuntimeDefault seccomp profile whenever
no other seccomp profile is specified. Otherwise, the default is Unconfined.
Here's a manifest for a Pod that requests the RuntimeDefault seccomp profile
for all its containers:
apiVersion:v1kind:Podmetadata:name:default-podlabels:app:default-podspec:securityContext:seccompProfile:type:RuntimeDefaultcontainers:- name:test-containerimage:hashicorp/http-echo:0.2.3args:- "-text=just made some more syscalls!"securityContext:allowPrivilegeEscalation:false
Use a cloud provider like Google Kubernetes Engine or Amazon Web Services to
create a Kubernetes cluster. This tutorial creates an
external load balancer,
which requires a cloud provider.
Configure kubectl to communicate with your Kubernetes API server. For instructions, see the
documentation for your cloud provider.
Objectives
Run five instances of a Hello World application.
Create a Service object that exposes an external IP address.
Use the Service object to access the running application.
Creating a service for an application running in five pods
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my-service LoadBalancer 10.3.245.137 104.198.205.71 8080/TCP 54s
Note: The type=LoadBalancer service is backed by external cloud providers, which is not covered in this example, please refer to this page for the details.
Note: If the external IP address is shown as <pending>, wait for a minute and enter the same command again.
Make a note of the external IP address (LoadBalancer Ingress) exposed by
your service. In this example, the external IP address is 104.198.205.71.
Also note the value of Port and NodePort. In this example, the Port
is 8080 and the NodePort is 32377.
In the preceding output, you can see that the service has several endpoints:
10.0.0.6:8080,10.0.1.6:8080,10.0.1.7:8080 + 2 more. These are internal
addresses of the pods that are running the Hello World application. To
verify these are pod addresses, enter this command:
Use the external IP address (LoadBalancer Ingress) to access the Hello
World application:
curl http://<external-ip>:<port>
where <external-ip> is the external IP address (LoadBalancer Ingress)
of your Service, and <port> is the value of Port in your Service
description.
If you are using minikube, typing minikube service my-service will
automatically open the Hello World application in a browser.
The response to a successful request is a hello message:
Hello Kubernetes!
Cleaning up
To delete the Service, enter this command:
kubectl delete services my-service
To delete the Deployment, the ReplicaSet, and the Pods that are running
the Hello World application, enter this command:
5.5.2 - Example: Deploying PHP Guestbook application with Redis
This tutorial shows you how to build and deploy a simple (not production
ready), multi-tier web application using Kubernetes and
Docker. This example consists of the following
components:
A single-instance Redis to store guestbook entries
Multiple web frontend instances
Objectives
Start up a Redis leader.
Start up two Redis followers.
Start up the guestbook frontend.
Expose and view the Frontend Service.
Clean up.
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Query the list of Pods to verify that the Redis Pod is running:
kubectl get pods
The response should be similar to this:
NAME READY STATUS RESTARTS AGE
redis-leader-fb76b4755-xjr2n 1/1 Running 0 13s
Run the following command to view the logs from the Redis leader Pod:
kubectl logs -f deployment/redis-leader
Creating the Redis leader Service
The guestbook application needs to communicate to the Redis to write its data.
You need to apply a Service to
proxy the traffic to the Redis Pod. A Service defines a policy to access the
Pods.
Query the list of Services to verify that the Redis Service is running:
kubectl get service
The response should be similar to this:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 1m
redis-leader ClusterIP 10.103.78.24 <none> 6379/TCP 16s
Note: This manifest file creates a Service named redis-leader with a set of labels
that match the labels previously defined, so the Service routes network
traffic to the Redis Pod.
Set up Redis followers
Although the Redis leader is a single Pod, you can make it highly available
and meet traffic demands by adding a few Redis followers, or replicas.
Verify that the two Redis follower replicas are running by querying the list of Pods:
kubectl get pods
The response should be similar to this:
NAME READY STATUS RESTARTS AGE
redis-follower-dddfbdcc9-82sfr 1/1 Running 0 37s
redis-follower-dddfbdcc9-qrt5k 1/1 Running 0 38s
redis-leader-fb76b4755-xjr2n 1/1 Running 0 11m
Creating the Redis follower service
The guestbook application needs to communicate with the Redis followers to
read data. To make the Redis followers discoverable, you must set up another
Service.
# SOURCE: https://cloud.google.com/kubernetes-engine/docs/tutorials/guestbookapiVersion:v1kind:Servicemetadata:name:redis-followerlabels:app:redisrole:followertier:backendspec:ports:# the port that this service should serve on- port:6379selector:app:redisrole:followertier:backend
Apply the Redis Service from the following redis-follower-service.yaml file:
Query the list of Services to verify that the Redis Service is running:
kubectl get service
The response should be similar to this:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 3d19h
redis-follower ClusterIP 10.110.162.42 <none> 6379/TCP 9s
redis-leader ClusterIP 10.103.78.24 <none> 6379/TCP 6m10s
Note: This manifest file creates a Service named redis-follower with a set of
labels that match the labels previously defined, so the Service routes network
traffic to the Redis Pod.
Set up and Expose the Guestbook Frontend
Now that you have the Redis storage of your guestbook up and running, start
the guestbook web servers. Like the Redis followers, the frontend is deployed
using a Kubernetes Deployment.
The guestbook app uses a PHP frontend. It is configured to communicate with
either the Redis follower or leader Services, depending on whether the request
is a read or a write. The frontend exposes a JSON interface, and serves a
jQuery-Ajax-based UX.
Query the list of Pods to verify that the three frontend replicas are running:
kubectl get pods -l app=guestbook -l tier=frontend
The response should be similar to this:
NAME READY STATUS RESTARTS AGE
frontend-85595f5bf9-5tqhb 1/1 Running 0 47s
frontend-85595f5bf9-qbzwm 1/1 Running 0 47s
frontend-85595f5bf9-zchwc 1/1 Running 0 47s
Creating the Frontend Service
The Redis Services you applied is only accessible within the Kubernetes
cluster because the default type for a Service is
ClusterIP.
ClusterIP provides a single IP address for the set of Pods the Service is
pointing to. This IP address is accessible only within the cluster.
If you want guests to be able to access your guestbook, you must configure the
frontend Service to be externally visible, so a client can request the Service
from outside the Kubernetes cluster. However a Kubernetes user can use
kubectl port-forward to access the service even though it uses a
ClusterIP.
Note: Some cloud providers, like Google Compute Engine or Google Kubernetes Engine,
support external load balancers. If your cloud provider supports load
balancers and you want to use it, uncomment type: LoadBalancer.
# SOURCE: https://cloud.google.com/kubernetes-engine/docs/tutorials/guestbookapiVersion:v1kind:Servicemetadata:name:frontendlabels:app:guestbooktier:frontendspec:# if your cluster supports it, uncomment the following to automatically create# an external load-balanced IP for the frontend service.# type: LoadBalancer#type: LoadBalancerports:# the port that this service should serve on- port:80selector:app:guestbooktier:frontend
Apply the frontend Service from the frontend-service.yaml file:
If you deployed the frontend-service.yaml manifest with type: LoadBalancer
you need to find the IP address to view your Guestbook.
Run the following command to get the IP address for the frontend Service.
kubectl get service frontend
The response should be similar to this:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
frontend LoadBalancer 10.51.242.136 109.197.92.229 80:32372/TCP 1m
Copy the external IP address, and load the page in your browser to view your guestbook.
Note: Try adding some guestbook entries by typing in a message, and clicking Submit.
The message you typed appears in the frontend. This message indicates that
data is successfully added to Redis through the Services you created earlier.
Scale the Web Frontend
You can scale up or down as needed because your servers are defined as a
Service that uses a Deployment controller.
Run the following command to scale up the number of frontend Pods:
kubectl scale deployment frontend --replicas=5
Query the list of Pods to verify the number of frontend Pods running:
This tutorial provides an introduction to managing applications with
StatefulSets.
It demonstrates how to create, delete, scale, and update the Pods of StatefulSets.
Before you begin
Before you begin this tutorial, you should familiarize yourself with the
following Kubernetes concepts:
Note: This tutorial assumes that your cluster is configured to dynamically provision
PersistentVolumes. If your cluster is not configured to do so, you
will have to manually provision two 1 GiB volumes prior to starting this
tutorial.
Objectives
StatefulSets are intended to be used with stateful applications and distributed
systems. However, the administration of stateful applications and
distributed systems on Kubernetes is a broad, complex topic. In order to
demonstrate the basic features of a StatefulSet, and not to conflate the former
topic with the latter, you will deploy a simple web application using a StatefulSet.
After this tutorial, you will be familiar with the following.
How to create a StatefulSet
How a StatefulSet manages its Pods
How to delete a StatefulSet
How to scale a StatefulSet
How to update a StatefulSet's Pods
Creating a StatefulSet
Begin by creating a StatefulSet using the example below. It is similar to the
example presented in the
StatefulSets concept.
It creates a headless Service,
nginx, to publish the IP addresses of Pods in the StatefulSet, web.
Download the example above, and save it to a file named web.yaml
You will need to use two terminal windows. In the first terminal, use
kubectl get to watch the creation
of the StatefulSet's Pods.
kubectl get pods -w -l app=nginx
In the second terminal, use
kubectl apply to create the
headless Service and StatefulSet defined in web.yaml.
kubectl apply -f web.yaml
service/nginx created
statefulset.apps/web created
The command above creates two Pods, each running an
NGINX webserver. Get the nginx Service...
kubectl get service nginx
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
nginx ClusterIP None <none> 80/TCP 12s
...then get the web StatefulSet, to verify that both were created successfully:
kubectl get statefulset web
NAME DESIRED CURRENT AGE
web 2 1 20s
Ordered Pod Creation
For a StatefulSet with n replicas, when Pods are being deployed, they are
created sequentially, ordered from {0..n-1}. Examine the output of the
kubectl get command in the first terminal. Eventually, the output will
look like the example below.
Notice that the web-1 Pod is not launched until the web-0 Pod is
Running (see Pod Phase)
and Ready (see type in Pod Conditions).
Pods in a StatefulSet
Pods in a StatefulSet have a unique ordinal index and a stable network identity.
Examining the Pod's Ordinal Index
Get the StatefulSet's Pods:
kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 1m
web-1 1/1 Running 0 1m
As mentioned in the StatefulSets
concept, the Pods in a StatefulSet have a sticky, unique identity. This identity
is based on a unique ordinal index that is assigned to each Pod by the
StatefulSet controller.
The Pods' names take the form <statefulset name>-<ordinal index>.
Since the web StatefulSet has two replicas, it creates two Pods, web-0 and web-1.
Using Stable Network Identities
Each Pod has a stable hostname based on its ordinal index. Use
kubectl exec to execute the
hostname command in each Pod:
for i in 0 1; do kubectl exec"web-$i" -- sh -c 'hostname'; done
web-0
web-1
Use kubectl run to execute
a container that provides the nslookup command from the dnsutils package.
Using nslookup on the Pods' hostnames, you can examine their in-cluster DNS
addresses:
kubectl run -i --tty --image busybox:1.28 dns-test --restart=Never --rm
which starts a new shell. In that new shell, run:
# Run this in the dns-test container shellnslookup web-0.nginx
The CNAME of the headless service points to SRV records (one for each Pod that
is Running and Ready). The SRV records point to A record entries that
contain the Pods' IP addresses.
In one terminal, watch the StatefulSet's Pods:
kubectl get pod -w -l app=nginx
In a second terminal, use
kubectl delete to delete all
the Pods in the StatefulSet:
kubectl delete pod -l app=nginx
pod "web-0" deleted
pod "web-1" deleted
Wait for the StatefulSet to restart them, and for both Pods to transition to
Running and Ready:
kubectl get pod -w -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 0/1 ContainerCreating 0 0s
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 2s
web-1 0/1 Pending 0 0s
web-1 0/1 Pending 0 0s
web-1 0/1 ContainerCreating 0 0s
web-1 1/1 Running 0 34s
Use kubectl exec and kubectl run to view the Pods' hostnames and in-cluster
DNS entries. First, view the Pods' hostnames:
for i in 0 1; do kubectl exec web-$i -- sh -c 'hostname'; done
web-0
web-1
then, run:
kubectl run -i --tty --image busybox:1.28 dns-test --restart=Never --rm /bin/sh
which starts a new shell.
In that new shell, run:
# Run this in the dns-test container shellnslookup web-0.nginx
The Pods' ordinals, hostnames, SRV records, and A record names have not changed,
but the IP addresses associated with the Pods may have changed. In the cluster
used for this tutorial, they have. This is why it is important not to configure
other applications to connect to Pods in a StatefulSet by IP address.
If you need to find and connect to the active members of a StatefulSet, you
should query the CNAME of the headless Service
(nginx.default.svc.cluster.local). The SRV records associated with the
CNAME will contain only the Pods in the StatefulSet that are Running and
Ready.
If your application already implements connection logic that tests for
liveness and readiness, you can use the SRV records of the Pods (
web-0.nginx.default.svc.cluster.local,
web-1.nginx.default.svc.cluster.local), as they are stable, and your
application will be able to discover the Pods' addresses when they transition
to Running and Ready.
Writing to Stable Storage
Get the PersistentVolumeClaims for web-0 and web-1:
kubectl get pvc -l app=nginx
The output is similar to:
NAME STATUS VOLUME CAPACITY ACCESSMODES AGE
www-web-0 Bound pvc-15c268c7-b507-11e6-932f-42010a800002 1Gi RWO 48s
www-web-1 Bound pvc-15c79307-b507-11e6-932f-42010a800002 1Gi RWO 48s
As the cluster used in this tutorial is configured to dynamically provision PersistentVolumes,
the PersistentVolumes were created and bound automatically.
The NGINX webserver, by default, serves an index file from
/usr/share/nginx/html/index.html. The volumeMounts field in the
StatefulSet's spec ensures that the /usr/share/nginx/html directory is
backed by a PersistentVolume.
Write the Pods' hostnames to their index.html files and verify that the NGINX
webservers serve the hostnames:
for i in 0 1; do kubectl exec"web-$i" -- sh -c 'echo "$(hostname)" > /usr/share/nginx/html/index.html'; donefor i in 0 1; do kubectl exec -i -t "web-$i" -- curl http://localhost/; done
web-0
web-1
Note:
If you instead see 403 Forbidden responses for the above curl command,
you will need to fix the permissions of the directory mounted by the volumeMounts
(due to a bug when using hostPath volumes),
by running:
for i in 0 1; do kubectl exec web-$i -- chmod 755 /usr/share/nginx/html; done
before retrying the curl command above.
In one terminal, watch the StatefulSet's Pods:
kubectl get pod -w -l app=nginx
In a second terminal, delete all of the StatefulSet's Pods:
kubectl delete pod -l app=nginx
pod "web-0" deleted
pod "web-1" deleted
Examine the output of the kubectl get command in the first terminal, and wait
for all of the Pods to transition to Running and Ready.
kubectl get pod -w -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 0/1 ContainerCreating 0 0s
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 2s
web-1 0/1 Pending 0 0s
web-1 0/1 Pending 0 0s
web-1 0/1 ContainerCreating 0 0s
web-1 1/1 Running 0 34s
Verify the web servers continue to serve their hostnames:
for i in 0 1; do kubectl exec -i -t "web-$i" -- curl http://localhost/; done
web-0
web-1
Even though web-0 and web-1 were rescheduled, they continue to serve their
hostnames because the PersistentVolumes associated with their
PersistentVolumeClaims are remounted to their volumeMounts. No matter what
node web-0and web-1 are scheduled on, their PersistentVolumes will be
mounted to the appropriate mount points.
Scaling a StatefulSet
Scaling a StatefulSet refers to increasing or decreasing the number of replicas.
This is accomplished by updating the replicas field. You can use either
kubectl scale or
kubectl patch to scale a StatefulSet.
Scaling Up
In one terminal window, watch the Pods in the StatefulSet:
kubectl get pods -w -l app=nginx
In another terminal window, use kubectl scale to scale the number of replicas
to 5:
kubectl scale sts web --replicas=5
statefulset.apps/web scaled
Examine the output of the kubectl get command in the first terminal, and wait
for the three additional Pods to transition to Running and Ready.
The StatefulSet controller scaled the number of replicas. As with
StatefulSet creation, the StatefulSet controller
created each Pod sequentially with respect to its ordinal index, and it
waited for each Pod's predecessor to be Running and Ready before launching the
subsequent Pod.
Scaling Down
In one terminal, watch the StatefulSet's Pods:
kubectl get pods -w -l app=nginx
In another terminal, use kubectl patch to scale the StatefulSet back down to
three replicas:
kubectl patch sts web -p '{"spec":{"replicas":3}}'
statefulset.apps/web patched
Wait for web-4 and web-3 to transition to Terminating.
kubectl get pods -w -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 3h
web-1 1/1 Running 0 3h
web-2 1/1 Running 0 55s
web-3 1/1 Running 0 36s
web-4 0/1 ContainerCreating 0 18s
NAME READY STATUS RESTARTS AGE
web-4 1/1 Running 0 19s
web-4 1/1 Terminating 0 24s
web-4 1/1 Terminating 0 24s
web-3 1/1 Terminating 0 42s
web-3 1/1 Terminating 0 42s
Ordered Pod Termination
The controller deleted one Pod at a time, in reverse order with respect to its
ordinal index, and it waited for each to be completely shutdown before
deleting the next.
There are still five PersistentVolumeClaims and five PersistentVolumes.
When exploring a Pod's stable storage, we saw that the PersistentVolumes mounted to the Pods of a StatefulSet are not deleted when the StatefulSet's Pods are deleted. This is still true when Pod deletion is caused by scaling the StatefulSet down.
Updating StatefulSets
In Kubernetes 1.7 and later, the StatefulSet controller supports automated updates. The
strategy used is determined by the spec.updateStrategy field of the
StatefulSet API Object. This feature can be used to upgrade the container
images, resource requests and/or limits, labels, and annotations of the Pods in a
StatefulSet. There are two valid update strategies, RollingUpdate and
OnDelete.
RollingUpdate update strategy is the default for StatefulSets.
Rolling Update
The RollingUpdate update strategy will update all Pods in a StatefulSet, in
reverse ordinal order, while respecting the StatefulSet guarantees.
Patch the web StatefulSet to apply the RollingUpdate update strategy:
kubectl patch statefulset web -p '{"spec":{"updateStrategy":{"type":"RollingUpdate"}}}'
statefulset.apps/web patched
In one terminal window, patch the web StatefulSet to change the container
image again:
kubectl patch statefulset web --type='json' -p='[{"op": "replace", "path": "/spec/template/spec/containers/0/image", "value":"gcr.io/google_containers/nginx-slim:0.8"}]'
statefulset.apps/web patched
In another terminal, watch the Pods in the StatefulSet:
The Pods in the StatefulSet are updated in reverse ordinal order. The
StatefulSet controller terminates each Pod, and waits for it to transition to Running and
Ready prior to updating the next Pod. Note that, even though the StatefulSet
controller will not proceed to update the next Pod until its ordinal successor
is Running and Ready, it will restore any Pod that fails during the update to
its current version.
Pods that have already received the update will be restored to the updated version,
and Pods that have not yet received the update will be restored to the previous
version. In this way, the controller attempts to continue to keep the application
healthy and the update consistent in the presence of intermittent failures.
Get the Pods to view their container images:
for p in 01 2; do kubectl get pod "web-$p" --template '{{range $i, $c := .spec.containers}}{{$c.image}}{{end}}'; echo; done
All the Pods in the StatefulSet are now running the previous container image.
Note: You can also use kubectl rollout status sts/<name> to view
the status of a rolling update to a StatefulSet
Staging an Update
You can stage an update to a StatefulSet by using the partition parameter of
the RollingUpdate update strategy. A staged update will keep all of the Pods
in the StatefulSet at the current version while allowing mutations to the
StatefulSet's .spec.template.
Patch the web StatefulSet to add a partition to the updateStrategy field:
kubectl patch statefulset web -p '{"spec":{"updateStrategy":{"type":"RollingUpdate","rollingUpdate":{"partition":3}}}}'
statefulset.apps/web patched
Patch the StatefulSet again to change the container's image:
kubectl patch statefulset web --type='json' -p='[{"op": "replace", "path": "/spec/template/spec/containers/0/image", "value":"registry.k8s.io/nginx-slim:0.7"}]'
statefulset.apps/web patched
Delete a Pod in the StatefulSet:
kubectl delete pod web-2
pod "web-2" deleted
Wait for the Pod to be Running and Ready.
kubectl get pod -l app=nginx -w
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 4m
web-1 1/1 Running 0 4m
web-2 0/1 ContainerCreating 0 11s
web-2 1/1 Running 0 18s
Get the Pod's container image:
kubectl get pod web-2 --template '{{range $i, $c := .spec.containers}}{{$c.image}}{{end}}'
registry.k8s.io/nginx-slim:0.8
Notice that, even though the update strategy is RollingUpdate the StatefulSet
restored the Pod with its original container. This is because the
ordinal of the Pod is less than the partition specified by the
updateStrategy.
Rolling Out a Canary
You can roll out a canary to test a modification by decrementing the partition
you specified above.
Patch the StatefulSet to decrement the partition:
kubectl patch statefulset web -p '{"spec":{"updateStrategy":{"type":"RollingUpdate","rollingUpdate":{"partition":2}}}}'
statefulset.apps/web patched
Wait for web-2 to be Running and Ready.
kubectl get pod -l app=nginx -w
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 4m
web-1 1/1 Running 0 4m
web-2 0/1 ContainerCreating 0 11s
web-2 1/1 Running 0 18s
Get the Pod's container:
kubectl get pod web-2 --template '{{range $i, $c := .spec.containers}}{{$c.image}}{{end}}'
registry.k8s.io/nginx-slim:0.7
When you changed the partition, the StatefulSet controller automatically
updated the web-2 Pod because the Pod's ordinal was greater than or equal to
the partition.
kubectl get pod web-1 --template '{{range $i, $c := .spec.containers}}{{$c.image}}{{end}}'
registry.k8s.io/nginx-slim:0.8
web-1 was restored to its original configuration because the Pod's ordinal
was less than the partition. When a partition is specified, all Pods with an
ordinal that is greater than or equal to the partition will be updated when the
StatefulSet's .spec.template is updated. If a Pod that has an ordinal less
than the partition is deleted or otherwise terminated, it will be restored to
its original configuration.
Phased Roll Outs
You can perform a phased roll out (e.g. a linear, geometric, or exponential
roll out) using a partitioned rolling update in a similar manner to how you
rolled out a canary. To perform a phased roll out, set
the partition to the ordinal at which you want the controller to pause the
update.
The partition is currently set to 2. Set the partition to 0:
kubectl patch statefulset web -p '{"spec":{"updateStrategy":{"type":"RollingUpdate","rollingUpdate":{"partition":0}}}}'
statefulset.apps/web patched
Wait for all of the Pods in the StatefulSet to become Running and Ready.
By moving the partition to 0, you allowed the StatefulSet to
continue the update process.
On Delete
The OnDelete update strategy implements the legacy (1.6 and prior) behavior,
When you select this update strategy, the StatefulSet controller will not
automatically update Pods when a modification is made to the StatefulSet's
.spec.template field. This strategy can be selected by setting the
.spec.template.updateStrategy.type to OnDelete.
Deleting StatefulSets
StatefulSet supports both Non-Cascading and Cascading deletion. In a
Non-Cascading Delete, the StatefulSet's Pods are not deleted when the StatefulSet is deleted. In a Cascading Delete, both the StatefulSet and its Pods are
deleted.
Non-Cascading Delete
In one terminal window, watch the Pods in the StatefulSet.
kubectl get pods -w -l app=nginx
Use kubectl delete to delete the
StatefulSet. Make sure to supply the --cascade=orphan parameter to the
command. This parameter tells Kubernetes to only delete the StatefulSet, and to
not delete any of its Pods.
kubectl delete statefulset web --cascade=orphan
statefulset.apps "web" deleted
Get the Pods, to examine their status:
kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 6m
web-1 1/1 Running 0 7m
web-2 1/1 Running 0 5m
Even though web has been deleted, all of the Pods are still Running and Ready.
Delete web-0:
kubectl delete pod web-0
pod "web-0" deleted
Get the StatefulSet's Pods:
kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
web-1 1/1 Running 0 10m
web-2 1/1 Running 0 7m
As the web StatefulSet has been deleted, web-0 has not been relaunched.
In one terminal, watch the StatefulSet's Pods.
kubectl get pods -w -l app=nginx
In a second terminal, recreate the StatefulSet. Note that, unless
you deleted the nginx Service (which you should not have), you will see
an error indicating that the Service already exists.
kubectl apply -f web.yaml
statefulset.apps/web created
service/nginx unchanged
Ignore the error. It only indicates that an attempt was made to create the nginx
headless Service even though that Service already exists.
Examine the output of the kubectl get command running in the first terminal.
kubectl get pods -w -l app=nginx
NAME READY STATUS RESTARTS AGE
web-1 1/1 Running 0 16m
web-2 1/1 Running 0 2m
NAME READY STATUS RESTARTS AGE
web-0 0/1 Pending 0 0s
web-0 0/1 Pending 0 0s
web-0 0/1 ContainerCreating 0 0s
web-0 1/1 Running 0 18s
web-2 1/1 Terminating 0 3m
web-2 0/1 Terminating 0 3m
web-2 0/1 Terminating 0 3m
web-2 0/1 Terminating 0 3m
When the web StatefulSet was recreated, it first relaunched web-0.
Since web-1 was already Running and Ready, when web-0 transitioned to
Running and Ready, it adopted this Pod. Since you recreated the StatefulSet
with replicas equal to 2, once web-0 had been recreated, and once
web-1 had been determined to already be Running and Ready, web-2 was
terminated.
Let's take another look at the contents of the index.html file served by the
Pods' webservers:
for i in 0 1; do kubectl exec -i -t "web-$i" -- curl http://localhost/; done
web-0
web-1
Even though you deleted both the StatefulSet and the web-0 Pod, it still
serves the hostname originally entered into its index.html file. This is
because the StatefulSet never deletes the PersistentVolumes associated with a
Pod. When you recreated the StatefulSet and it relaunched web-0, its original
PersistentVolume was remounted.
Cascading Delete
In one terminal window, watch the Pods in the StatefulSet.
kubectl get pods -w -l app=nginx
In another terminal, delete the StatefulSet again. This time, omit the
--cascade=orphan parameter.
kubectl delete statefulset web
statefulset.apps "web" deleted
Examine the output of the kubectl get command running in the first terminal,
and wait for all of the Pods to transition to Terminating.
kubectl get pods -w -l app=nginx
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 11m
web-1 1/1 Running 0 27m
NAME READY STATUS RESTARTS AGE
web-0 1/1 Terminating 0 12m
web-1 1/1 Terminating 0 29m
web-0 0/1 Terminating 0 12m
web-0 0/1 Terminating 0 12m
web-0 0/1 Terminating 0 12m
web-1 0/1 Terminating 0 29m
web-1 0/1 Terminating 0 29m
web-1 0/1 Terminating 0 29m
As you saw in the Scaling Down section, the Pods
are terminated one at a time, with respect to the reverse order of their ordinal
indices. Before terminating a Pod, the StatefulSet controller waits for
the Pod's successor to be completely terminated.
Note: Although a cascading delete removes a StatefulSet together with its Pods,
the cascade does not delete the headless Service associated with the StatefulSet.
You must delete the nginx Service manually.
kubectl delete service nginx
service "nginx" deleted
Recreate the StatefulSet and headless Service one more time:
kubectl apply -f web.yaml
service/nginx created
statefulset.apps/web created
When all of the StatefulSet's Pods transition to Running and Ready, retrieve
the contents of their index.html files:
for i in 0 1; do kubectl exec -i -t "web-$i" -- curl http://localhost/; done
web-0
web-1
Even though you completely deleted the StatefulSet, and all of its Pods, the
Pods are recreated with their PersistentVolumes mounted, and web-0 and
web-1 continue to serve their hostnames.
Finally, delete the nginx Service...
kubectl delete service nginx
service "nginx" deleted
...and the web StatefulSet:
kubectl delete statefulset web
statefulset "web" deleted
Pod Management Policy
For some distributed systems, the StatefulSet ordering guarantees are
unnecessary and/or undesirable. These systems require only uniqueness and
identity. To address this, in Kubernetes 1.7, we introduced
.spec.podManagementPolicy to the StatefulSet API Object.
OrderedReady Pod Management
OrderedReady pod management is the default for StatefulSets. It tells the
StatefulSet controller to respect the ordering guarantees demonstrated
above.
Parallel Pod Management
Parallel pod management tells the StatefulSet controller to launch or
terminate all Pods in parallel, and not to wait for Pods to become Running
and Ready or completely terminated prior to launching or terminating another
Pod. This option only affects the behavior for scaling operations. Updates are not affected.
During deletion, a StatefulSet removes all Pods concurrently; it does not wait for
a Pod's ordinal successor to terminate prior to deleting that Pod.
Close the terminal where the kubectl get command is running and delete the nginx
Service:
kubectl delete svc nginx
Note:
You also need to delete the persistent storage media for the PersistentVolumes
used in this tutorial.
Follow the necessary steps, based on your environment, storage configuration,
and provisioning method, to ensure that all storage is reclaimed.
5.6.2 - Example: Deploying WordPress and MySQL with Persistent Volumes
This tutorial shows you how to deploy a WordPress site and a MySQL database using Minikube. Both applications use PersistentVolumes and PersistentVolumeClaims to store data.
A PersistentVolume (PV) is a piece of storage in the cluster that has been manually provisioned by an administrator, or dynamically provisioned by Kubernetes using a StorageClass. A PersistentVolumeClaim (PVC) is a request for storage by a user that can be fulfilled by a PV. PersistentVolumes and PersistentVolumeClaims are independent from Pod lifecycles and preserve data through restarting, rescheduling, and even deleting Pods.
Warning: This deployment is not suitable for production use cases, as it uses single instance WordPress and MySQL Pods. Consider using WordPress Helm Chart to deploy WordPress in production.
Note: The files provided in this tutorial are using GA Deployment APIs and are specific to kubernetes version 1.9 and later. If you wish to use this tutorial with an earlier version of Kubernetes, please update the API version appropriately, or reference earlier versions of this tutorial.
Objectives
Create PersistentVolumeClaims and PersistentVolumes
Create a kustomization.yaml with
a Secret generator
MySQL resource configs
WordPress resource configs
Apply the kustomization directory by kubectl apply -k ./
Clean up
Before you begin
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Create PersistentVolumeClaims and PersistentVolumes
MySQL and Wordpress each require a PersistentVolume to store data. Their PersistentVolumeClaims will be created at the deployment step.
Many cluster environments have a default StorageClass installed. When a StorageClass is not specified in the PersistentVolumeClaim, the cluster's default StorageClass is used instead.
When a PersistentVolumeClaim is created, a PersistentVolume is dynamically provisioned based on the StorageClass configuration.
Warning: In local clusters, the default StorageClass uses the hostPath provisioner. hostPath volumes are only suitable for development and testing. With hostPath volumes, your data lives in /tmp on the node the Pod is scheduled onto and does not move between nodes. If a Pod dies and gets scheduled to another node in the cluster, or the node is rebooted, the data is lost.
Note: If you are bringing up a cluster that needs to use the hostPath provisioner, the --enable-hostpath-provisioner flag must be set in the controller-manager component.
Note: If you have a Kubernetes cluster running on Google Kubernetes Engine, please follow this guide.
Create a kustomization.yaml
Add a Secret generator
A Secret is an object that stores a piece of sensitive data like a password or key. Since 1.14, kubectl supports the management of Kubernetes objects using a kustomization file. You can create a Secret by generators in kustomization.yaml.
Add a Secret generator in kustomization.yaml from the following command. You will need to replace YOUR_PASSWORD with the password you want to use.
The following manifest describes a single-instance MySQL Deployment. The MySQL container mounts the PersistentVolume at /var/lib/mysql. The MYSQL_ROOT_PASSWORD environment variable sets the database password from the Secret.
The following manifest describes a single-instance WordPress Deployment. The WordPress container mounts the
PersistentVolume at /var/www/html for website data files. The WORDPRESS_DB_HOST environment variable sets
the name of the MySQL Service defined above, and WordPress will access the database by Service. The
WORDPRESS_DB_PASSWORD environment variable sets the database password from the Secret kustomize generated.
The kustomization.yaml contains all the resources for deploying a WordPress site and a
MySQL database. You can apply the directory by
kubectl apply -k ./
Now you can verify that all objects exist.
Verify that the Secret exists by running the following command:
kubectl get secrets
The response should be like this:
NAME TYPE DATA AGE
mysql-pass-c57bb4t7mf Opaque 1 9s
Verify that a PersistentVolume got dynamically provisioned.
kubectl get pvc
Note: It can take up to a few minutes for the PVs to be provisioned and bound.
The response should be like this:
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
mysql-pv-claim Bound pvc-8cbd7b2e-4044-11e9-b2bb-42010a800002 20Gi RWO standard 77s
wp-pv-claim Bound pvc-8cd0df54-4044-11e9-b2bb-42010a800002 20Gi RWO standard 77s
Verify that the Pod is running by running the following command:
kubectl get pods
Note: It can take up to a few minutes for the Pod's Status to be RUNNING.
The response should be like this:
NAME READY STATUS RESTARTS AGE
wordpress-mysql-1894417608-x5dzt 1/1 Running 0 40s
Verify that the Service is running by running the following command:
kubectl get services wordpress
The response should be like this:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
wordpress LoadBalancer 10.0.0.89 <pending> 80:32406/TCP 4m
Note: Minikube can only expose Services through NodePort. The EXTERNAL-IP is always pending.
Run the following command to get the IP Address for the WordPress Service:
minikube service wordpress --url
The response should be like this:
http://1.2.3.4:32406
Copy the IP address, and load the page in your browser to view your site.
You should see the WordPress set up page similar to the following screenshot.
Warning: Do not leave your WordPress installation on this page. If another user finds it, they can set up a website on your instance and use it to serve malicious content.
Either install WordPress by creating a username and password or delete your instance.
Cleaning up
Run the following command to delete your Secret, Deployments, Services and PersistentVolumeClaims:
5.6.3 - Example: Deploying Cassandra with a StatefulSet
This tutorial shows you how to run Apache Cassandra on Kubernetes.
Cassandra, a database, needs persistent storage to provide data durability (application state).
In this example, a custom Cassandra seed provider lets the database discover new Cassandra instances as they join the Cassandra cluster.
StatefulSets make it easier to deploy stateful applications into your Kubernetes cluster.
For more information on the features used in this tutorial, see
StatefulSet.
Note:
Cassandra and Kubernetes both use the term node to mean a member of a cluster. In this
tutorial, the Pods that belong to the StatefulSet are Cassandra nodes and are members
of the Cassandra cluster (called a ring). When those Pods run in your Kubernetes cluster,
the Kubernetes control plane schedules those Pods onto Kubernetes
Nodes.
When a Cassandra node starts, it uses a seed list to bootstrap discovery of other
nodes in the ring.
This tutorial deploys a custom Cassandra seed provider that lets the database discover
new Cassandra Pods as they appear inside your Kubernetes cluster.
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
To complete this tutorial, you should already have a basic familiarity with
Pods,
Services, and
StatefulSets.
Additional Minikube setup instructions
Caution:
Minikube defaults to 2048MB of memory and 2 CPU.
Running Minikube with the default resource configuration results in insufficient resource
errors during this tutorial. To avoid these errors, start Minikube with the following settings:
minikube start --memory 5120 --cpus=4
Creating a headless Service for Cassandra
In Kubernetes, a Service describes a set of
Pods that perform the same task.
The following Service is used for DNS lookups between Cassandra Pods and clients within your cluster:
apiVersion:apps/v1kind:StatefulSetmetadata:name:cassandralabels:app:cassandraspec:serviceName:cassandrareplicas:3selector:matchLabels:app:cassandratemplate:metadata:labels:app:cassandraspec:terminationGracePeriodSeconds:1800containers:- name:cassandraimage:gcr.io/google-samples/cassandra:v13imagePullPolicy:Alwaysports:- containerPort:7000name:intra-node- containerPort:7001name:tls-intra-node- containerPort:7199name:jmx- containerPort:9042name:cqlresources:limits:cpu:"500m"memory:1Girequests:cpu:"500m"memory:1GisecurityContext:capabilities:add:- IPC_LOCKlifecycle:preStop:exec:command:- /bin/sh- -c- nodetool drainenv:- name:MAX_HEAP_SIZEvalue:512M- name:HEAP_NEWSIZEvalue:100M- name:CASSANDRA_SEEDSvalue:"cassandra-0.cassandra.default.svc.cluster.local"- name:CASSANDRA_CLUSTER_NAMEvalue:"K8Demo"- name:CASSANDRA_DCvalue:"DC1-K8Demo"- name:CASSANDRA_RACKvalue:"Rack1-K8Demo"- name:POD_IPvalueFrom:fieldRef:fieldPath:status.podIPreadinessProbe:exec:command:- /bin/bash- -c- /ready-probe.shinitialDelaySeconds:15timeoutSeconds:5# These volume mounts are persistent. They are like inline claims,# but not exactly because the names need to match exactly one of# the stateful pod volumes.volumeMounts:- name:cassandra-datamountPath:/cassandra_data# These are converted to volume claims by the controller# and mounted at the paths mentioned above.# do not use these in production until ssd GCEPersistentDisk or other ssd pdvolumeClaimTemplates:- metadata:name:cassandra-dataspec:accessModes:["ReadWriteOnce"]storageClassName:fastresources:requests:storage:1Gi---kind:StorageClassapiVersion:storage.k8s.io/v1metadata:name:fastprovisioner:k8s.io/minikube-hostpathparameters:type:pd-ssd
Create the Cassandra StatefulSet from the cassandra-statefulset.yaml file:
# Use this if you are able to apply cassandra-statefulset.yaml unmodifiedkubectl apply -f https://k8s.io/examples/application/cassandra/cassandra-statefulset.yaml
# Use this if you needed to modify cassandra-statefulset.yaml locallykubectl apply -f cassandra-statefulset.yaml
Validating the Cassandra StatefulSet
Get the Cassandra StatefulSet:
kubectl get statefulset cassandra
The response should be similar to:
NAME DESIRED CURRENT AGE
cassandra 3 0 13s
The StatefulSet resource deploys Pods sequentially.
Get the Pods to see the ordered creation status:
kubectl get pods -l="app=cassandra"
The response should be similar to:
NAME READY STATUS RESTARTS AGE
cassandra-0 1/1 Running 0 1m
cassandra-1 0/1 ContainerCreating 0 8s
It can take several minutes for all three Pods to deploy. Once they are deployed, the same command
returns output similar to:
NAME READY STATUS RESTARTS AGE
cassandra-0 1/1 Running 0 10m
cassandra-1 1/1 Running 0 9m
cassandra-2 1/1 Running 0 8m
Run the Cassandra nodetool inside the first Pod, to
display the status of the ring.
kubectl exec -it cassandra-0 -- nodetool status
The response should look something like:
Datacenter: DC1-K8Demo
======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
-- Address Load Tokens Owns (effective) Host ID Rack
UN 172.17.0.5 83.57 KiB 32 74.0% e2dd09e6-d9d3-477e-96c5-45094c08db0f Rack1-K8Demo
UN 172.17.0.4 101.04 KiB 32 58.8% f89d6835-3a42-4419-92b3-0e62cae1479c Rack1-K8Demo
UN 172.17.0.6 84.74 KiB 32 67.1% a6a1e8c2-3dc5-4417-b1a0-26507af2aaad Rack1-K8Demo
Modifying the Cassandra StatefulSet
Use kubectl edit to modify the size of a Cassandra StatefulSet.
Run the following command:
kubectl edit statefulset cassandra
This command opens an editor in your terminal. The line you need to change is the replicas field.
The following sample is an excerpt of the StatefulSet file:
# Please edit the object below. Lines beginning with a '#' will be ignored,# and an empty file will abort the edit. If an error occurs while saving this file will be# reopened with the relevant failures.#apiVersion:apps/v1kind:StatefulSetmetadata:creationTimestamp:2016-08-13T18:40:58Zgeneration:1labels:app:cassandraname:cassandranamespace:defaultresourceVersion:"323"uid:7a219483-6185-11e6-a910-42010a8a0fc0spec:replicas:3
Change the number of replicas to 4, and then save the manifest.
The StatefulSet now scales to run with 4 Pods.
Get the Cassandra StatefulSet to verify your change:
kubectl get statefulset cassandra
The response should be similar to:
NAME DESIRED CURRENT AGE
cassandra 4 4 36m
Cleaning up
Deleting or scaling a StatefulSet down does not delete the volumes associated with the StatefulSet.
This setting is for your safety because your data is more valuable than automatically purging all related StatefulSet resources.
Warning: Depending on the storage class and reclaim policy, deleting the PersistentVolumeClaims may cause the associated volumes
to also be deleted. Never assume you'll be able to access data if its volume claims are deleted.
Run the following commands (chained together into a single command) to delete everything in the Cassandra StatefulSet:
This image includes a standard Cassandra installation from the Apache Debian repo.
By using environment variables you can change values that are inserted into cassandra.yaml.
You must have a cluster with at least four nodes, and each node requires at least 2 CPUs and 4 GiB of memory. In this tutorial you will cordon and drain the cluster's nodes. This means that the cluster will terminate and evict all Pods on its nodes, and the nodes will temporarily become unschedulable. You should use a dedicated cluster for this tutorial, or you should ensure that the disruption you cause will not interfere with other tenants.
This tutorial assumes that you have configured your cluster to dynamically provision
PersistentVolumes. If your cluster is not configured to do so, you
will have to manually provision three 20 GiB volumes before starting this
tutorial.
Objectives
After this tutorial, you will know the following.
How to deploy a ZooKeeper ensemble using StatefulSet.
How to consistently configure the ensemble.
How to spread the deployment of ZooKeeper servers in the ensemble.
How to use PodDisruptionBudgets to ensure service availability during planned maintenance.
ZooKeeper
Apache ZooKeeper is a
distributed, open-source coordination service for distributed applications.
ZooKeeper allows you to read, write, and observe updates to data. Data are
organized in a file system like hierarchy and replicated to all ZooKeeper
servers in the ensemble (a set of ZooKeeper servers). All operations on data
are atomic and sequentially consistent. ZooKeeper ensures this by using the
Zab
consensus protocol to replicate a state machine across all servers in the ensemble.
The ensemble uses the Zab protocol to elect a leader, and the ensemble cannot write data until that election is complete. Once complete, the ensemble uses Zab to ensure that it replicates all writes to a quorum before it acknowledges and makes them visible to clients. Without respect to weighted quorums, a quorum is a majority component of the ensemble containing the current leader. For instance, if the ensemble has three servers, a component that contains the leader and one other server constitutes a quorum. If the ensemble can not achieve a quorum, the ensemble cannot write data.
ZooKeeper servers keep their entire state machine in memory, and write every mutation to a durable WAL (Write Ahead Log) on storage media. When a server crashes, it can recover its previous state by replaying the WAL. To prevent the WAL from growing without bound, ZooKeeper servers will periodically snapshot them in memory state to storage media. These snapshots can be loaded directly into memory, and all WAL entries that preceded the snapshot may be discarded.
The StatefulSet controller creates three Pods, and each Pod has a container with
a ZooKeeper server.
Facilitating leader election
Because there is no terminating algorithm for electing a leader in an anonymous network, Zab requires explicit membership configuration to perform leader election. Each server in the ensemble needs to have a unique identifier, all servers need to know the global set of identifiers, and each identifier needs to be associated with a network address.
Use kubectl exec to get the hostnames
of the Pods in the zk StatefulSet.
for i in 01 2; do kubectl exec zk-$i -- hostname; done
The StatefulSet controller provides each Pod with a unique hostname based on its ordinal index. The hostnames take the form of <statefulset name>-<ordinal index>. Because the replicas field of the zk StatefulSet is set to 3, the Set's controller creates three Pods with their hostnames set to zk-0, zk-1, and
zk-2.
zk-0
zk-1
zk-2
The servers in a ZooKeeper ensemble use natural numbers as unique identifiers, and store each server's identifier in a file called myid in the server's data directory.
To examine the contents of the myid file for each server use the following command.
for i in 01 2; doecho"myid zk-$i";kubectl exec zk-$i -- cat /var/lib/zookeeper/data/myid; done
Because the identifiers are natural numbers and the ordinal indices are non-negative integers, you can generate an identifier by adding 1 to the ordinal.
myid zk-0
1
myid zk-1
2
myid zk-2
3
To get the Fully Qualified Domain Name (FQDN) of each Pod in the zk StatefulSet use the following command.
for i in 01 2; do kubectl exec zk-$i -- hostname -f; done
The zk-hs Service creates a domain for all of the Pods,
zk-hs.default.svc.cluster.local.
The A records in Kubernetes DNS resolve the FQDNs to the Pods' IP addresses. If Kubernetes reschedules the Pods, it will update the A records with the Pods' new IP addresses, but the A records names will not change.
ZooKeeper stores its application configuration in a file named zoo.cfg. Use kubectl exec to view the contents of the zoo.cfg file in the zk-0 Pod.
In the server.1, server.2, and server.3 properties at the bottom of
the file, the 1, 2, and 3 correspond to the identifiers in the
ZooKeeper servers' myid files. They are set to the FQDNs for the Pods in
the zk StatefulSet.
Consensus protocols require that the identifiers of each participant be unique. No two participants in the Zab protocol should claim the same unique identifier. This is necessary to allow the processes in the system to agree on which processes have committed which data. If two Pods are launched with the same ordinal, two ZooKeeper servers would both identify themselves as the same server.
The A records for each Pod are entered when the Pod becomes Ready. Therefore,
the FQDNs of the ZooKeeper servers will resolve to a single endpoint, and that
endpoint will be the unique ZooKeeper server claiming the identity configured
in its myid file.
When the servers use the Zab protocol to attempt to commit a value, they will either achieve consensus and commit the value (if leader election has succeeded and at least two of the Pods are Running and Ready), or they will fail to do so (if either of the conditions are not met). No state will arise where one server acknowledges a write on behalf of another.
Sanity testing the ensemble
The most basic sanity test is to write data to one ZooKeeper server and
to read the data from another.
The command below executes the zkCli.sh script to write world to the path /hello on the zk-0 Pod in the ensemble.
kubectl exec zk-0 -- zkCli.sh create /hello world
WATCHER::
WatchedEvent state:SyncConnected type:None path:null
Created /hello
To get the data from the zk-1 Pod use the following command.
kubectl exec zk-1 -- zkCli.sh get /hello
The data that you created on zk-0 is available on all the servers in the
ensemble.
WATCHER::
WatchedEvent state:SyncConnected type:None path:null
world
cZxid = 0x100000002
ctime = Thu Dec 08 15:13:30 UTC 2016
mZxid = 0x100000002
mtime = Thu Dec 08 15:13:30 UTC 2016
pZxid = 0x100000002
cversion = 0
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 5
numChildren = 0
Providing durable storage
As mentioned in the ZooKeeper Basics section,
ZooKeeper commits all entries to a durable WAL, and periodically writes snapshots
in memory state, to storage media. Using WALs to provide durability is a common
technique for applications that use consensus protocols to achieve a replicated
state machine.
Use the kubectl delete command to delete the
zk StatefulSet.
kubectl delete statefulset zk
statefulset.apps "zk" deleted
Watch the termination of the Pods in the StatefulSet.
kubectl get pods -w -l app=zk
When zk-0 if fully terminated, use CTRL-C to terminate kubectl.
When a Pod in the zkStatefulSet is (re)scheduled, it will always have the
same PersistentVolume mounted to the ZooKeeper server's data directory.
Even when the Pods are rescheduled, all the writes made to the ZooKeeper
servers' WALs, and all their snapshots, remain durable.
Ensuring consistent configuration
As noted in the Facilitating Leader Election and
Achieving Consensus sections, the servers in a
ZooKeeper ensemble require consistent configuration to elect a leader
and form a quorum. They also require consistent configuration of the Zab protocol
in order for the protocol to work correctly over a network. In our example we
achieve consistent configuration by embedding the configuration directly into
the manifest.
The command used to start the ZooKeeper servers passed the configuration as command line parameter. You can also use environment variables to pass configuration to the ensemble.
Configuring logging
One of the files generated by the zkGenConfig.sh script controls ZooKeeper's logging.
ZooKeeper uses Log4j, and, by default,
it uses a time and size based rolling file appender for its logging configuration.
Use the command below to get the logging configuration from one of Pods in the zkStatefulSet.
This is the simplest possible way to safely log inside the container.
Because the applications write logs to standard out, Kubernetes will handle log rotation for you.
Kubernetes also implements a sane retention policy that ensures application logs written to
standard out and standard error do not exhaust local storage media.
Use kubectl logs to retrieve the last 20 log lines from one of the Pods.
kubectl logs zk-0 --tail 20
You can view application logs written to standard out or standard error using kubectl logs and from the Kubernetes Dashboard.
2016-12-06 19:34:16,236 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52740
2016-12-06 19:34:16,237 [myid:1] - INFO [Thread-1136:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52740 (no session established for client)
2016-12-06 19:34:26,155 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52749
2016-12-06 19:34:26,155 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52749
2016-12-06 19:34:26,156 [myid:1] - INFO [Thread-1137:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52749 (no session established for client)
2016-12-06 19:34:26,222 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52750
2016-12-06 19:34:26,222 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52750
2016-12-06 19:34:26,226 [myid:1] - INFO [Thread-1138:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52750 (no session established for client)
2016-12-06 19:34:36,151 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52760
2016-12-06 19:34:36,152 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52760
2016-12-06 19:34:36,152 [myid:1] - INFO [Thread-1139:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52760 (no session established for client)
2016-12-06 19:34:36,230 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52761
2016-12-06 19:34:36,231 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52761
2016-12-06 19:34:36,231 [myid:1] - INFO [Thread-1140:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52761 (no session established for client)
2016-12-06 19:34:46,149 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52767
2016-12-06 19:34:46,149 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52767
2016-12-06 19:34:46,149 [myid:1] - INFO [Thread-1141:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52767 (no session established for client)
2016-12-06 19:34:46,230 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:52768
2016-12-06 19:34:46,230 [myid:1] - INFO [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@827] - Processing ruok command from /127.0.0.1:52768
2016-12-06 19:34:46,230 [myid:1] - INFO [Thread-1142:NIOServerCnxn@1008] - Closed socket connection for client /127.0.0.1:52768 (no session established for client)
Kubernetes integrates with many logging solutions. You can choose a logging solution
that best fits your cluster and applications. For cluster-level logging and aggregation,
consider deploying a sidecar container to rotate and ship your logs.
Configuring a non-privileged user
The best practices to allow an application to run as a privileged
user inside of a container are a matter of debate. If your organization requires
that applications run as a non-privileged user you can use a
SecurityContext to control the user that
the entry point runs as.
The zkStatefulSet's Pod template contains a SecurityContext.
securityContext:runAsUser:1000fsGroup:1000
In the Pods' containers, UID 1000 corresponds to the zookeeper user and GID 1000
corresponds to the zookeeper group.
Get the ZooKeeper process information from the zk-0 Pod.
kubectl exec zk-0 -- ps -elf
As the runAsUser field of the securityContext object is set to 1000,
instead of running as root, the ZooKeeper process runs as the zookeeper user.
F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD
4 S zookeep+ 1 0 0 80 0 - 1127 - 20:46 ? 00:00:00 sh -c zkGenConfig.sh && zkServer.sh start-foreground
0 S zookeep+ 27 1 0 80 0 - 1155556 - 20:46 ? 00:00:19 /usr/lib/jvm/java-8-openjdk-amd64/bin/java -Dzookeeper.log.dir=/var/log/zookeeper -Dzookeeper.root.logger=INFO,CONSOLE -cp /usr/bin/../build/classes:/usr/bin/../build/lib/*.jar:/usr/bin/../share/zookeeper/zookeeper-3.4.9.jar:/usr/bin/../share/zookeeper/slf4j-log4j12-1.6.1.jar:/usr/bin/../share/zookeeper/slf4j-api-1.6.1.jar:/usr/bin/../share/zookeeper/netty-3.10.5.Final.jar:/usr/bin/../share/zookeeper/log4j-1.2.16.jar:/usr/bin/../share/zookeeper/jline-0.9.94.jar:/usr/bin/../src/java/lib/*.jar:/usr/bin/../etc/zookeeper: -Xmx2G -Xms2G -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.local.only=false org.apache.zookeeper.server.quorum.QuorumPeerMain /usr/bin/../etc/zookeeper/zoo.cfg
By default, when the Pod's PersistentVolumes is mounted to the ZooKeeper server's data directory, it is only accessible by the root user. This configuration prevents the ZooKeeper process from writing to its WAL and storing its snapshots.
Use the command below to get the file permissions of the ZooKeeper data directory on the zk-0 Pod.
kubectl exec -ti zk-0 -- ls -ld /var/lib/zookeeper/data
Because the fsGroup field of the securityContext object is set to 1000, the ownership of the Pods' PersistentVolumes is set to the zookeeper group, and the ZooKeeper process is able to read and write its data.
drwxr-sr-x 3 zookeeper zookeeper 4096 Dec 5 20:45 /var/lib/zookeeper/data
Managing the ZooKeeper process
The ZooKeeper documentation
mentions that "You will want to have a supervisory process that
manages each of your ZooKeeper server processes (JVM)." Utilizing a watchdog
(supervisory process) to restart failed processes in a distributed system is a
common pattern. When deploying an application in Kubernetes, rather than using
an external utility as a supervisory process, you should use Kubernetes as the
watchdog for your application.
Updating the ensemble
The zkStatefulSet is configured to use the RollingUpdate update strategy.
You can use kubectl patch to update the number of cpus allocated to the servers.
Use kubectl rollout status to watch the status of the update.
kubectl rollout status sts/zk
waiting for statefulset rolling update to complete 0 pods at revision zk-5db4499664...
Waiting for 1 pods to be ready...
Waiting for 1 pods to be ready...
waiting for statefulset rolling update to complete 1 pods at revision zk-5db4499664...
Waiting for 1 pods to be ready...
Waiting for 1 pods to be ready...
waiting for statefulset rolling update to complete 2 pods at revision zk-5db4499664...
Waiting for 1 pods to be ready...
Waiting for 1 pods to be ready...
statefulset rolling update complete 3 pods at revision zk-5db4499664...
This terminates the Pods, one at a time, in reverse ordinal order, and recreates them with the new configuration. This ensures that quorum is maintained during a rolling update.
Use the kubectl rollout history command to view a history or previous configurations.
kubectl rollout history sts/zk
The output is similar to this:
statefulsets "zk"
REVISION
1
2
Use the kubectl rollout undo command to roll back the modification.
kubectl rollout undo sts/zk
The output is similar to this:
statefulset.apps/zk rolled back
Handling process failure
Restart Policies control how
Kubernetes handles process failures for the entry point of the container in a Pod.
For Pods in a StatefulSet, the only appropriate RestartPolicy is Always, and this
is the default value. For stateful applications you should never override
the default policy.
Use the following command to examine the process tree for the ZooKeeper server running in the zk-0 Pod.
kubectl exec zk-0 -- ps -ef
The command used as the container's entry point has PID 1, and
the ZooKeeper process, a child of the entry point, has PID 27.
In another terminal watch the Pods in the zkStatefulSet with the following command.
kubectl get pod -w -l app=zk
In another terminal, terminate the ZooKeeper process in Pod zk-0 with the following command.
kubectl exec zk-0 -- pkill java
The termination of the ZooKeeper process caused its parent process to terminate. Because the RestartPolicy of the container is Always, it restarted the parent process.
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 0 21m
zk-1 1/1 Running 0 20m
zk-2 1/1 Running 0 19m
NAME READY STATUS RESTARTS AGE
zk-0 0/1 Error 0 29m
zk-0 0/1 Running 1 29m
zk-0 1/1 Running 1 29m
If your application uses a script (such as zkServer.sh) to launch the process
that implements the application's business logic, the script must terminate with the
child process. This ensures that Kubernetes will restart the application's
container when the process implementing the application's business logic fails.
Testing for liveness
Configuring your application to restart failed processes is not enough to
keep a distributed system healthy. There are scenarios where
a system's processes can be both alive and unresponsive, or otherwise
unhealthy. You should use liveness probes to notify Kubernetes
that your application's processes are unhealthy and it should restart them.
The Pod template for the zkStatefulSet specifies a liveness probe.
When the liveness probe for the ZooKeeper process fails, Kubernetes will
automatically restart the process for you, ensuring that unhealthy processes in
the ensemble are restarted.
kubectl get pod -w -l app=zk
NAME READY STATUS RESTARTS AGE
zk-0 1/1 Running 0 1h
zk-1 1/1 Running 0 1h
zk-2 1/1 Running 0 1h
NAME READY STATUS RESTARTS AGE
zk-0 0/1 Running 0 1h
zk-0 0/1 Running 1 1h
zk-0 1/1 Running 1 1h
Testing for readiness
Readiness is not the same as liveness. If a process is alive, it is scheduled
and healthy. If a process is ready, it is able to process input. Liveness is
a necessary, but not sufficient, condition for readiness. There are cases,
particularly during initialization and termination, when a process can be
alive but not ready.
If you specify a readiness probe, Kubernetes will ensure that your application's
processes will not receive network traffic until their readiness checks pass.
For a ZooKeeper server, liveness implies readiness. Therefore, the readiness
probe from the zookeeper.yaml manifest is identical to the liveness probe.
Even though the liveness and readiness probes are identical, it is important
to specify both. This ensures that only healthy servers in the ZooKeeper
ensemble receive network traffic.
Tolerating Node failure
ZooKeeper needs a quorum of servers to successfully commit mutations
to data. For a three server ensemble, two servers must be healthy for
writes to succeed. In quorum based systems, members are deployed across failure
domains to ensure availability. To avoid an outage, due to the loss of an
individual machine, best practices preclude co-locating multiple instances of the
application on the same machine.
By default, Kubernetes may co-locate Pods in a StatefulSet on the same node.
For the three server ensemble you created, if two servers are on the same node, and that node fails,
the clients of your ZooKeeper service will experience an outage until at least one of the Pods can be rescheduled.
You should always provision additional capacity to allow the processes of critical
systems to be rescheduled in the event of node failures. If you do so, then the
outage will only last until the Kubernetes scheduler reschedules one of the ZooKeeper
servers. However, if you want your service to tolerate node failures with no downtime,
you should set podAntiAffinity.
Use the command below to get the nodes for Pods in the zkStatefulSet.
for i in 01 2; do kubectl get pod zk-$i --template {{.spec.nodeName}}; echo""; done
All of the Pods in the zkStatefulSet are deployed on different nodes.
The requiredDuringSchedulingIgnoredDuringExecution field tells the
Kubernetes Scheduler that it should never co-locate two Pods which have app label
as zk in the domain defined by the topologyKey. The topologyKeykubernetes.io/hostname indicates that the domain is an individual node. Using
different rules, labels, and selectors, you can extend this technique to spread
your ensemble across physical, network, and power failure domains.
Surviving maintenance
In this section you will cordon and drain nodes. If you are using this tutorial
on a shared cluster, be sure that this will not adversely affect other tenants.
The previous section showed you how to spread your Pods across nodes to survive
unplanned node failures, but you also need to plan for temporary node failures
that occur due to planned maintenance.
Use this command to get the nodes in your cluster.
kubectl get nodes
This tutorial assumes a cluster with at least four nodes. If the cluster has more than four, use kubectl cordon to cordon all but four nodes. Constraining to four nodes will ensure Kubernetes encounters affinity and PodDisruptionBudget constraints when scheduling zookeeper Pods in the following maintenance simulation.
kubectl cordon <node-name>
Use this command to get the zk-pdbPodDisruptionBudget.
kubectl get pdb zk-pdb
The max-unavailable field indicates to Kubernetes that at most one Pod from
zkStatefulSet can be unavailable at any time.
NAME MIN-AVAILABLE MAX-UNAVAILABLE ALLOWED-DISRUPTIONS AGE
zk-pdb N/A 1 1
In one terminal, use this command to watch the Pods in the zkStatefulSet.
kubectl get pods -w -l app=zk
In another terminal, use this command to get the nodes that the Pods are currently scheduled on.
for i in 01 2; do kubectl get pod zk-$i --template {{.spec.nodeName}}; echo""; done
Keep watching the StatefulSet's Pods in the first terminal and drain the node on which
zk-1 is scheduled.
kubectl drain $(kubectl get pod zk-1 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-emptydir-data
The output is similar to this:
"kubernetes-node-ixsl" cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-node-ixsl, kube-proxy-kubernetes-node-ixsl; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-voc74
pod "zk-1" deleted
node "kubernetes-node-ixsl" drained
The zk-1 Pod cannot be scheduled because the zkStatefulSet contains a PodAntiAffinity rule preventing
co-location of the Pods, and as only two nodes are schedulable, the Pod will remain in a Pending state.
Continue to watch the Pods of the StatefulSet, and drain the node on which
zk-2 is scheduled.
kubectl drain $(kubectl get pod zk-2 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-emptydir-data
The output is similar to this:
node "kubernetes-node-i4c4" cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-node-i4c4, kube-proxy-kubernetes-node-i4c4; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog
WARNING: Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog; Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-node-i4c4, kube-proxy-kubernetes-node-i4c4
There are pending pods when an error occurred: Cannot evict pod as it would violate the pod's disruption budget.
pod/zk-2
Use CTRL-C to terminate to kubectl.
You cannot drain the third node because evicting zk-2 would violate zk-budget. However, the node will remain cordoned.
Use zkCli.sh to retrieve the value you entered during the sanity test from zk-0.
kubectl exec zk-0 zkCli.sh get /hello
The service is still available because its PodDisruptionBudget is respected.
WatchedEvent state:SyncConnected type:None path:null
world
cZxid = 0x200000002
ctime = Wed Dec 07 00:08:59 UTC 2016
mZxid = 0x200000002
mtime = Wed Dec 07 00:08:59 UTC 2016
pZxid = 0x200000002
cversion = 0
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 5
numChildren = 0
Attempt to drain the node on which zk-2 is scheduled.
kubectl drain $(kubectl get pod zk-2 --template {{.spec.nodeName}}) --ignore-daemonsets --force --delete-emptydir-data
The output is similar to this:
node "kubernetes-node-i4c4" already cordoned
WARNING: Deleting pods not managed by ReplicationController, ReplicaSet, Job, or DaemonSet: fluentd-cloud-logging-kubernetes-node-i4c4, kube-proxy-kubernetes-node-i4c4; Ignoring DaemonSet-managed pods: node-problem-detector-v0.1-dyrog
pod "heapster-v1.2.0-2604621511-wht1r" deleted
pod "zk-2" deleted
node "kubernetes-node-i4c4" drained
This time kubectl drain succeeds.
Uncordon the second node to allow zk-2 to be rescheduled.
kubectl uncordon kubernetes-node-ixsl
The output is similar to this:
node "kubernetes-node-ixsl" uncordoned
You can use kubectl drain in conjunction with PodDisruptionBudgets to ensure that your services remain available during maintenance.
If drain is used to cordon nodes and evict pods prior to taking the node offline for maintenance,
services that express a disruption budget will have that budget respected.
You should always allocate additional capacity for critical services so that their Pods can be immediately rescheduled.
Cleaning up
Use kubectl uncordon to uncordon all the nodes in your cluster.
You must delete the persistent storage media for the PersistentVolumes used in this tutorial.
Follow the necessary steps, based on your environment, storage configuration,
and provisioning method, to ensure that all storage is reclaimed.
5.7 - Services
5.7.1 - Using Source IP
Applications running in a Kubernetes cluster find and communicate with each
other, and the outside world, through the Service abstraction. This document
explains what happens to the source IP of packets sent to different types
of Services, and how you can toggle this behavior according to your needs.
a network daemon that orchestrates Service VIP management on every node
Prerequisites
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
Expose a simple application through various types of Services
Understand how each Service type handles source IP NAT
Understand the tradeoffs involved in preserving source IP
Source IP for Services with Type=ClusterIP
Packets sent to ClusterIP from within the cluster are never source NAT'd if
you're running kube-proxy in
iptables mode,
(the default). You can query the kube-proxy mode by fetching
http://localhost:10249/proxyMode on the node where kube-proxy is running.
kubectl get nodes
The output is similar to this:
NAME STATUS ROLES AGE VERSION
kubernetes-node-6jst Ready <none> 2h v1.13.0
kubernetes-node-cx31 Ready <none> 2h v1.13.0
kubernetes-node-jj1t Ready <none> 2h v1.13.0
Get the proxy mode on one of the nodes (kube-proxy listens on port 10249):
# Run this in a shell on the node you want to query.curl http://localhost:10249/proxyMode
The output is:
iptables
You can test source IP preservation by creating a Service over the source IP app:
NODEPORT=$(kubectl get -o jsonpath="{.spec.ports[0].nodePort}" services nodeport)NODES=$(kubectl get nodes -o jsonpath='{ $.items[*].status.addresses[?(@.type=="InternalIP")].address }')
If you're running on a cloud provider, you may need to open up a firewall-rule
for the nodes:nodeport reported above.
Now you can try reaching the Service from outside the cluster through the node
port allocated above.
for node in $NODES; do curl -s $node:$NODEPORT | grep -i client_address; done
Note that these are not the correct client IPs, they're cluster internal IPs. This is what happens:
Client sends packet to node2:nodePort
node2 replaces the source IP address (SNAT) in the packet with its own IP address
node2 replaces the destination IP on the packet with the pod IP
packet is routed to node 1, and then to the endpoint
the pod's reply is routed back to node2
the pod's reply is sent back to the client
Visually:
Figure. Source IP Type=NodePort using SNAT
To avoid this, Kubernetes has a feature to
preserve the client source IP.
If you set service.spec.externalTrafficPolicy to the value Local,
kube-proxy only proxies proxy requests to local endpoints, and does not
forward traffic to other nodes. This approach preserves the original
source IP address. If there are no local endpoints, packets sent to the
node are dropped, so you can rely on the correct source-ip in any packet
processing rules you might apply a packet that make it through to the
endpoint.
Set the service.spec.externalTrafficPolicy field as follows:
for node in $NODES; do curl --connect-timeout 1 -s $node:$NODEPORT | grep -i client_address; done
The output is similar to:
client_address=198.51.100.79
Note that you only got one reply, with the right client IP, from the one node on which the endpoint pod
is running.
This is what happens:
client sends packet to node2:nodePort, which doesn't have any endpoints
packet is dropped
client sends packet to node1:nodePort, which does have endpoints
node1 routes packet to endpoint with the correct source IP
Visually:
Figure. Source IP Type=NodePort preserves client source IP address
Source IP for Services with Type=LoadBalancer
Packets sent to Services with
Type=LoadBalancer
are source NAT'd by default, because all schedulable Kubernetes nodes in the
Ready state are eligible for load-balanced traffic. So if packets arrive
at a node without an endpoint, the system proxies it to a node with an
endpoint, replacing the source IP on the packet with the IP of the node (as
described in the previous section).
You can test this by exposing the source-ip-app through a load balancer:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
loadbalancer LoadBalancer 10.0.65.118 203.0.113.140 80/TCP 5m
Next, send a request to this Service's external-ip:
curl 203.0.113.140
The output is similar to this:
CLIENT VALUES:
client_address=10.240.0.5
...
However, if you're running on Google Kubernetes Engine/GCE, setting the same service.spec.externalTrafficPolicy
field to Local forces nodes without Service endpoints to remove
themselves from the list of nodes eligible for loadbalanced traffic by
deliberately failing health checks.
You should immediately see the service.spec.healthCheckNodePort field allocated
by Kubernetes:
kubectl get svc loadbalancer -o yaml | grep -i healthCheckNodePort
The output is similar to this:
healthCheckNodePort:32122
The service.spec.healthCheckNodePort field points to a port on every node
serving the health check at /healthz. You can test this:
kubectl get pod -o wide -l app=source-ip-app
The output is similar to this:
NAME READY STATUS RESTARTS AGE IP NODE
source-ip-app-826191075-qehz4 1/1 Running 0 20h 10.180.1.136 kubernetes-node-6jst
Use curl to fetch the /healthz endpoint on various nodes:
# Run this locally on a node you choosecurl localhost:32122/healthz
1 Service Endpoints found
On a different node you might get a different result:
# Run this locally on a node you choosecurl localhost:32122/healthz
No Service Endpoints Found
A controller running on the
control plane is
responsible for allocating the cloud load balancer. The same controller also
allocates HTTP health checks pointing to this port/path on each node. Wait
about 10 seconds for the 2 nodes without endpoints to fail health checks,
then use curl to query the IPv4 address of the load balancer:
curl 203.0.113.140
The output is similar to this:
CLIENT VALUES:
client_address=198.51.100.79
...
Cross-platform support
Only some cloud providers offer support for source IP preservation through
Services with Type=LoadBalancer.
The cloud provider you're running on might fulfill the request for a loadbalancer
in a few different ways:
With a proxy that terminates the client connection and opens a new connection
to your nodes/endpoints. In such cases the source IP will always be that of the
cloud LB, not that of the client.
With a packet forwarder, such that requests from the client sent to the
loadbalancer VIP end up at the node with the source IP of the client, not
an intermediate proxy.
Load balancers in the first category must use an agreed upon
protocol between the loadbalancer and backend to communicate the true client IP
such as the HTTP Forwarded
or X-FORWARDED-FOR
headers, or the
proxy protocol.
Load balancers in the second category can leverage the feature described above
by creating an HTTP health check pointing at the port stored in
the service.spec.healthCheckNodePort field on the Service.
kubeadm - CLI tool to easily provision a secure Kubernetes cluster.
Components
kubelet - The
primary agent that runs on each node. The kubelet takes a set of PodSpecs
and ensures that the described containers are running and healthy.
kube-apiserver -
REST API that validates and configures data for API objects such as pods,
services, replication controllers.
kube-controller-manager - Daemon that embeds the core control loops shipped with Kubernetes.
kube-proxy - Can
do simple TCP/UDP stream forwarding or round-robin TCP/UDP forwarding across
a set of back-ends.
kube-scheduler - Scheduler that manages availability, performance, and capacity.
List of ports and protocols that
should be open on control plane and worker nodes
Config APIs
This section hosts the documentation for "unpublished" APIs which are used to
configure kubernetes components or tools. Most of these APIs are not exposed
by the API server in a RESTful way though they are essential for a user or an
operator to use or manage a cluster.
This section provides reference information for the Kubernetes API.
The REST API is the fundamental fabric of Kubernetes. All operations and
communications between components, and external user commands are REST API
calls that the API Server handles. Consequently, everything in the Kubernetes
platform is treated as an API object and has a corresponding entry in the
API.
The JSON and Protobuf serialization schemas follow the same guidelines for
schema changes. The following descriptions cover both formats.
The API versioning and software versioning are indirectly related.
The API and release versioning proposal
describes the relationship between API versioning and software versioning.
Different API versions indicate different levels of stability and support. You
can find more information about the criteria for each level in the
API Changes documentation.
Here's a summary of each level:
Alpha:
The version names contain alpha (for example, v1alpha1).
The software may contain bugs. Enabling a feature may expose bugs. A
feature may be disabled by default.
The support for a feature may be dropped at any time without notice.
The API may change in incompatible ways in a later software release without notice.
The software is recommended for use only in short-lived testing clusters,
due to increased risk of bugs and lack of long-term support.
Beta:
The version names contain beta (for example, v2beta3).
The software is well tested. Enabling a feature is considered safe.
Features are enabled by default.
The support for a feature will not be dropped, though the details may change.
The schema and/or semantics of objects may change in incompatible ways in
a subsequent beta or stable release. When this happens, migration
instructions are provided. Schema changes may require deleting, editing, and
re-creating API objects. The editing process may not be straightforward.
The migration may require downtime for applications that rely on the feature.
The software is not recommended for production uses. Subsequent releases
may introduce incompatible changes. If you have multiple clusters which
can be upgraded independently, you may be able to relax this restriction.
Note: Please try beta features and provide feedback. After the features exit beta, it
may not be practical to make more changes.
Stable:
The version name is vX where X is an integer.
The stable versions of features appear in released software for many subsequent versions.
API groups
API groups
make it easier to extend the Kubernetes API.
The API group is specified in a REST path and in the apiVersion field of a
serialized object.
There are several API groups in Kubernetes:
The core (also called legacy) group is found at REST path /api/v1.
The core group is not specified as part of the apiVersion field, for
example, apiVersion: v1.
The named groups are at REST path /apis/$GROUP_NAME/$VERSION and use
apiVersion: $GROUP_NAME/$VERSION (for example, apiVersion: batch/v1).
You can find the full list of supported API groups in
Kubernetes API reference.
Enabling or disabling API groups
Certain resources and API groups are enabled by default. You can enable or
disable them by setting --runtime-config on the API server. The
--runtime-config flag accepts comma separated <key>[=<value>] pairs
describing the runtime configuration of the API server. If the =<value>
part is omitted, it is treated as if =true is specified. For example:
to disable batch/v1, set --runtime-config=batch/v1=false
to enable batch/v2alpha1, set --runtime-config=batch/v2alpha1
to enable a specific version of an API, such as storage.k8s.io/v1beta1/csistoragecapacities, set --runtime-config=storage.k8s.io/v1beta1/csistoragecapacities
Note: When you enable or disable groups or resources, you need to restart the API
server and controller manager to pick up the --runtime-config changes.
Persistence
Kubernetes stores its serialized state in terms of the API resources by writing them into
etcd.
The Kubernetes API is a resource-based (RESTful) programmatic interface
provided via HTTP. It supports retrieving, creating, updating, and deleting
primary resources via the standard HTTP verbs (POST, PUT, PATCH, DELETE,
GET).
For some resources, the API includes additional subresources that allow
fine grained authorization (such as separate views for Pod details and
log retrievals), and can accept and serve those resources in different
representations for convenience or efficiency.
Kubernetes supports efficient change notifications on resources via watches.
Kubernetes also provides consistent list operations so that API clients can
effectively cache, track, and synchronize the state of resources.
You can view the API reference online,
or read on to learn about the API in general.
Kubernetes API terminology
Kubernetes generally leverages common RESTful terminology to describe the
API concepts:
A resource type is the name used in the URL (pods, namespaces, services)
All resource types have a concrete representation (their object schema) which is called a kind
A list of instances of a resource is known as a collection
A single instance of a resource type is called a resource, and also usually represents an object
For some resource types, the API includes one or more sub-resources, which are represented as URI paths below the resource
Most Kubernetes API resource types are
objects:
they represent a concrete instance of a concept on the cluster, like a
pod or namespace. A smaller number of API resource types are virtual in
that they often represent operations on objects, rather than objects, such
as a permission check
(use a POST with a JSON-encoded body of SubjectAccessReview to the
subjectaccessreviews resource), or the eviction sub-resource of a Pod
(used to trigger
API-initiated eviction).
Object names
All objects you can create via the API have a unique object
name to allow idempotent creation and
retrieval, except that virtual resource types may not have unique names if they are
not retrievable, or do not rely on idempotency.
Within a namespace, only one object
of a given kind can have a given name at a time. However, if you delete the object,
you can make a new object with the same name. Some objects are not namespaced (for
example: Nodes), and so their names must be unique across the whole cluster.
API verbs
Almost all object resource types support the standard HTTP verbs - GET, POST, PUT, PATCH,
and DELETE. Kubernetes also uses its own verbs, which are often written lowercase to distinguish
them from HTTP verbs.
Kubernetes uses the term list to describe returning a collection of
resources to distinguish from retrieving a single resource which is usually called
a get. If you sent an HTTP GET request with the ?watch query parameter,
Kubernetes calls this a watch and not a get (see
Efficient detection of changes for more details).
For PUT requests, Kubernetes internally classifies these as either create or update
based on the state of the existing object. An update is different from a patch; the
HTTP verb for a patch is PATCH.
Resource URIs
All resource types are either scoped by the cluster (/apis/GROUP/VERSION/*) or to a
namespace (/apis/GROUP/VERSION/namespaces/NAMESPACE/*). A namespace-scoped resource
type will be deleted when its namespace is deleted and access to that resource type
is controlled by authorization checks on the namespace scope.
You can also access collections of resources (for example: listing all Nodes).
The following paths are used to retrieve collections and resources:
Cluster-scoped resources:
GET /apis/GROUP/VERSION/RESOURCETYPE - return the collection of resources of the resource type
GET /apis/GROUP/VERSION/RESOURCETYPE/NAME - return the resource with NAME under the resource type
Namespace-scoped resources:
GET /apis/GROUP/VERSION/RESOURCETYPE - return the collection of all instances of the resource type across all namespaces
GET /apis/GROUP/VERSION/namespaces/NAMESPACE/RESOURCETYPE - return collection of all instances of the resource type in NAMESPACE
GET /apis/GROUP/VERSION/namespaces/NAMESPACE/RESOURCETYPE/NAME - return the instance of the resource type with NAME in NAMESPACE
Since a namespace is a cluster-scoped resource type, you can retrieve the list
(“collection”) of all namespaces with GET /api/v1/namespaces and details about
a particular namespace with GET /api/v1/namespaces/NAME.
Cluster-scoped subresource: GET /apis/GROUP/VERSION/RESOURCETYPE/NAME/SUBRESOURCE
Namespace-scoped subresource: GET /apis/GROUP/VERSION/namespaces/NAMESPACE/RESOURCETYPE/NAME/SUBRESOURCE
The verbs supported for each subresource will differ depending on the object -
see the API reference for more information. It
is not possible to access sub-resources across multiple resources - generally a new
virtual resource type would be used if that becomes necessary.
Efficient detection of changes
The Kubernetes API allows clients to make an initial request for an object or a
collection, and then to track changes since that initial request: a watch. Clients
can send a list or a get and then make a follow-up watch request.
To make this change tracking possible, every Kubernetes object has a resourceVersion
field representing the version of that resource as stored in the underlying persistence
layer. When retrieving a collection of resources (either namespace or cluster scoped),
the response from the API server contains a resourceVersion value. The client can
use that resourceVersion to initiate a watch against the API server.
When you send a watch request, the API server responds with a stream of
changes. These changes itemize the outcome of operations (such as create, delete,
and update) that occurred after the resourceVersion you specified as a parameter
to the watch request. The overall watch mechanism allows a client to fetch
the current state and then subscribe to subsequent changes, without missing any events.
If a client watch is disconnected then that client can start a new watch from
the last returned resourceVersion; the client could also perform a fresh get /
list request and begin again. See Resource Version Semantics
for more detail.
For example:
List all of the pods in a given namespace.
GET /api/v1/namespaces/test/pods
---
200 OK
Content-Type: application/json
{
"kind": "PodList",
"apiVersion": "v1",
"metadata": {"resourceVersion":"10245"},
"items": [...]
}
Starting from resource version 10245, receive notifications of any API operations
(such as create, delete, apply or update) that affect Pods in the
test namespace. Each change notification is a JSON document. The HTTP response body
(served as application/json) consists a series of JSON documents.
A given Kubernetes server will only preserve a historical record of changes for a
limited time. Clusters using etcd 3 preserve changes in the last 5 minutes by default.
When the requested watch operations fail because the historical version of that
resource is not available, clients must handle the case by recognizing the status code
410 Gone, clearing their local cache, performing a new get or list operation,
and starting the watch from the resourceVersion that was returned.
For subscribing to collections, Kubernetes client libraries typically offer some form
of standard tool for this list-then-watch logic. (In the Go client library,
this is called a Reflector and is located in the k8s.io/client-go/tools/cache package.)
Watch bookmarks
To mitigate the impact of short history window, the Kubernetes API provides a watch
event named BOOKMARK. It is a special kind of event to mark that all changes up
to a given resourceVersion the client is requesting have already been sent. The
document representing the BOOKMARK event is of the type requested by the request,
but only includes a .metadata.resourceVersion field. For example:
As a client, you can request BOOKMARK events by setting the
allowWatchBookmarks=true query parameter to a watch request, but you shouldn't
assume bookmarks are returned at any specific interval, nor can clients assume that
the API server will send any BOOKMARK event even when requested.
Retrieving large results sets in chunks
FEATURE STATE:Kubernetes v1.9 [beta]
On large clusters, retrieving the collection of some resource types may result in
very large responses that can impact the server and client. For instance, a cluster
may have tens of thousands of Pods, each of which is equivalent to roughly 2 KiB of
encoded JSON. Retrieving all pods across all namespaces may result in a very large
response (10-20MB) and consume a large amount of server resources.
Provided that you don't explicitly disable the APIListChunkingfeature gate, the
Kubernetes API server supports the ability to break a single large collection request
into many smaller chunks while preserving the consistency of the total request. Each
chunk can be returned sequentially which reduces both the total size of the request and
allows user-oriented clients to display results incrementally to improve responsiveness.
You can request that the API server handles a list by serving single collection
using pages (which Kubernetes calls chunks). To retrieve a single collection in
chunks, two query parameters limit and continue are supported on requests against
collections, and a response field continue is returned from all list operations
in the collection's metadata field. A client should specify the maximum results they
wish to receive in each chunk with limit and the server will return up to limit
resources in the result and include a continue value if there are more resources
in the collection.
As an API client, you can then pass this continue value to the API server on the
next request, to instruct the server to return the next page (chunk) of results. By
continuing until the server returns an empty continue value, you can retrieve the
entire collection.
Like a watch operation, a continue token will expire after a short amount
of time (by default 5 minutes) and return a 410 Gone if more results cannot be
returned. In this case, the client will need to start from the beginning or omit the
limit parameter.
For example, if there are 1,253 pods on the cluster and you wants to receive chunks
of 500 pods at a time, request those chunks as follows:
List all of the pods on a cluster, retrieving up to 500 pods each time.
Continue the previous call, retrieving the last 253 pods.
GET /api/v1/pods?limit=500&continue=ENCODED_CONTINUE_TOKEN_2
---
200 OK
Content-Type: application/json
{
"kind": "PodList",
"apiVersion": "v1",
"metadata": {
"resourceVersion":"10245",
"continue": "", // continue token is empty because we have reached the end of the list
...
},
"items": [...] // returns pods 1001-1253
}
Notice that the resourceVersion of the collection remains constant across each request,
indicating the server is showing you a consistent snapshot of the pods. Pods that
are created, updated, or deleted after version 10245 would not be shown unless
you make a separate list request without the continue token. This allows you
to break large requests into smaller chunks and then perform a watch operation
on the full set without missing any updates.
remainingItemCount is the number of subsequent items in the collection that are not
included in this response. If the list request contained label or field
selectors then the number of
remaining items is unknown and the API server does not include a remainingItemCount
field in its response.
If the list is complete (either because it is not chunking, or because this is the
last chunk), then there are no more remaining items and the API server does not include a
remainingItemCount field in its response. The intended use of the remainingItemCount
is estimating the size of a collection.
Collections
In Kubernetes terminology, the response you get from a list is
a collection. However, Kubernetes defines concrete kinds for
collections of different types of resource. Collections have a kind
named for the resource kind, with List appended.
When you query the API for a particular type, all items returned by that query are
of that type.
For example, when you list Services, the collection response
has kind set to
ServiceList; each item in that collection represents a single Service. For example:
There are dozens of collection types (such as PodList, ServiceList,
and NodeList) defined in the Kubernetes API.
You can get more information about each collection type from the
Kubernetes API documentation.
Some tools, such as kubectl, represent the Kubernetes collection
mechanism slightly differently from the Kubernetes API itself.
Because the output of kubectl might include the response from
multiple list operations at the API level, kubectl represents
a list of items using kind: List. For example:
Keep in mind that the Kubernetes API does not have a kind named List.
kind: List is a client-side, internal implementation detail for processing
collections that might be of different kinds of object. Avoid depending on
kind: List in automation or other code.
Receiving resources as Tables
When you run kubectl get, the default output format is a simple tabular
representation of one or more instances of a particular resource type. In the past,
clients were required to reproduce the tabular and describe output implemented in
kubectl to perform simple lists of objects.
A few limitations of that approach include non-trivial logic when dealing with
certain objects. Additionally, types provided by API aggregation or third party
resources are not known at compile time. This means that generic implementations
had to be in place for types unrecognized by a client.
In order to avoid potential limitations as described above, clients may request
the Table representation of objects, delegating specific details of printing to the
server. The Kubernetes API implements standard HTTP content type negotiation: passing
an Accept header containing a value of application/json;as=Table;g=meta.k8s.io;v=v1
with a GET call will request that the server return objects in the Table content
type.
For example, list all of the pods on a cluster in the Table format.
GET /api/v1/pods
Accept: application/json;as=Table;g=meta.k8s.io;v=v1
---
200 OK
Content-Type: application/json
{
"kind": "Table",
"apiVersion": "meta.k8s.io/v1",
...
"columnDefinitions": [
...
]
}
For API resource types that do not have a custom Table definition known to the control
plane, the API server returns a default Table response that consists of the resource's
name and creationTimestamp fields.
Not all API resource types support a Table response; for example, a
CustomResourceDefinitions
might not define field-to-table mappings, and an APIService that
extends the core Kubernetes API
might not serve Table responses at all. If you are implementing a client that
uses the Table information and must work against all resource types, including
extensions, you should make requests that specify multiple content types in the
Accept header. For example:
By default, Kubernetes returns objects serialized to JSON with content type
application/json. This is the default serialization format for the API. However,
clients may request the more efficient
Protobuf representation of these objects for better performance at scale.
The Kubernetes API implements standard HTTP content type negotiation: passing an
Accept header with a GET call will request that the server tries to return
a response in your preferred media type, while sending an object in Protobuf to
the server for a PUT or POST call means that you must set the Content-Type
header appropriately.
The server will return a response with a Content-Type header if the requested
format is supported, or the 406 Not acceptable error if none of the media types you
requested are supported. All built-in resource types support the application/json
media type.
See the Kubernetes API reference for a list of
supported content types for each API.
For example:
List all of the pods on a cluster in Protobuf format.
GET /api/v1/pods
Accept: application/vnd.kubernetes.protobuf
---
200 OK
Content-Type: application/vnd.kubernetes.protobuf
... binary encoded PodList object
Create a pod by sending Protobuf encoded data to the server, but request a response
in JSON.
POST /api/v1/namespaces/test/pods
Content-Type: application/vnd.kubernetes.protobuf
Accept: application/json
... binary encoded Pod object
---
200 OK
Content-Type: application/json
{
"kind": "Pod",
"apiVersion": "v1",
...
}
Not all API resource types support Protobuf; specifically, Protobuf isn't available for
resources that are defined as
CustomResourceDefinitions
or are served via the
aggregation layer.
As a client, if you might need to work with extension types you should specify multiple
content types in the request Accept header to support fallback to JSON.
For example:
Kubernetes uses an envelope wrapper to encode Protobuf responses. That wrapper starts
with a 4 byte magic number to help identify content in disk or in etcd as Protobuf
(as opposed to JSON), and then is followed by a Protobuf encoded wrapper message, which
describes the encoding and type of the underlying object and then contains the object.
The wrapper format is:
A four byte magic number prefix:
Bytes 0-3: "k8s\x00" [0x6b, 0x38, 0x73, 0x00]
An encoded Protobuf message with the following IDL:
message Unknown {
// typeMeta should have the string values for "kind" and "apiVersion" as set on the JSON object
optional TypeMeta typeMeta = 1;
// raw will hold the complete serialized object in protobuf. See the protobuf definitions in the client libraries for a given kind.
optional bytes raw = 2;
// contentEncoding is encoding used for the raw data. Unspecified means no encoding.
optional string contentEncoding = 3;
// contentType is the serialization method used to serialize 'raw'. Unspecified means application/vnd.kubernetes.protobuf and is usually
// omitted.
optional string contentType = 4;
}
message TypeMeta {
// apiVersion is the group/version for this type
optional string apiVersion = 1;
// kind is the name of the object schema. A protobuf definition should exist for this object.
optional string kind = 2;
}
Note: Clients that receive a response in application/vnd.kubernetes.protobuf that does
not match the expected prefix should reject the response, as future versions may need
to alter the serialization format in an incompatible way and will do so by changing
the prefix.
Resource deletion
When you delete a resource this takes place in two phases.
When a client first sends a delete to request removal of a resource, the .metadata.deletionTimestamp is set to the current time.
Once the .metadata.deletionTimestamp is set, external controllers that act on finalizers
may start performing their cleanup work at any time, in any order.
Order is not enforced between finalizers because it would introduce significant
risk of stuck .metadata.finalizers.
The .metadata.finalizers field is shared: any actor with permission can reorder it.
If the finalizer list were processed in order, then this might lead to a situation
in which the component responsible for the first finalizer in the list is
waiting for some signal (field value, external system, or other) produced by a
component responsible for a finalizer later in the list, resulting in a deadlock.
Without enforced ordering, finalizers are free to order amongst themselves and are
not vulnerable to ordering changes in the list.
Once the last finalizer is removed, the resource is actually removed from etcd.
Single resource API
The Kubernetes API verbs get, create, apply, update, patch,
delete and proxy support single resources only.
These verbs with single resource support have no support for submitting multiple
resources together in an ordered or unordered list or transaction.
When clients (including kubectl) act on a set of resources, the client makes a series
of single-resource API requests, then aggregates the responses if needed.
By contrast, the Kubernetes API verbs list and watch allow getting multiple
resources, and deletecollection allows deleting multiple resources.
Field validation
Kubernetes always validates the type of fields. For example, if a field in the
API is defined as a number, you cannot set the field to a text value. If a field
is defined as an array of strings, you can only provide an array. Some fields
allow you to omit them, other fields are required. Omitting a required field
from an API request is an error.
If you make a request with an extra field, one that the cluster's control plane
does not recognize, then the behavior of the API server is more complicated.
By default, the API server drops fields that it does not recognize
from an input that it receives (for example, the JSON body of a PUT request).
There are two situations where the API server drops fields that you supplied in
an HTTP request.
These situations are:
The field is unrecognized because it is not in the resource's OpenAPI schema. (One
exception to this is for CRDs that explicitly choose not to prune unknown
fields via x-kubernetes-preserve-unknown-fields).
The field is duplicated in the object.
Setting the field validation level
FEATURE STATE:Kubernetes v1.25 [beta]
Provided that the ServerSideFieldValidationfeature gate is enabled (disabled
by default in 1.23 and 1.24, enabled by default starting in 1.25), you can take
advantage of server side field validation to catch these unrecognized fields.
When you use HTTP verbs that can submit data (POST, PUT, and PATCH), field
validation gives you the option to choose how you would like to be notified of
these fields that are being dropped by the API server. Possible levels of
validation are Ignore, Warn, and Strict.
Note:
If you submit a request that specifies an unrecognized field, and that is also invalid for
a different reason (for example, the request provides a string value where the API expects
an integer), then the API server responds with a 400 Bad Request error response.
You always receive an error response in this case, no matter what field validation level you requested.
Field validation is set by the fieldValidation query parameter. The three
values that you can provide for this parameter are:
Ignore
The API server succeeds in handling the request as it would without the erroneous fields
being set, dropping all unknown and duplicate fields and giving no indication it
has done so.
Warn
(Default) The API server succeeds in handling the request, and reports a
warning to the client. The warning is sent using the Warning: response header,
adding one warning item for each unknown or duplicate field. For more
information about warnings and the Kubernetes API, see the blog article
Warning: Helpful Warnings Ahead.
Strict
The API server rejects the request with a 400 Bad Request error when it
detects any unknown or duplicate fields. The response message from the API
server specifies all the unknown or duplicate fields that the API server has
detected.
Tools that submit requests to the server (such as kubectl), might set their own
defaults that are different from the Warn validation level that the API server uses
by default.
The kubectl tool uses the --validate flag to set the level of field validation.
Historically --validate was used to toggle client-side validation on or off as
a boolean flag. Since Kubernetes 1.25, kubectl uses
server-side field validation when sending requests to a serer with this feature
enabled. Validation will fall back to client-side only when it cannot connect
to an API server with field validation enabled.
It accepts the values ignore, warn,
and strict while also accepting the values true (equivalent to strict) and false
(equivalent to ignore). The default validation setting for kubectl is --validate=true,
which means strict server-side field validation.
Dry-run
FEATURE STATE:Kubernetes v1.18 [stable]
When you use HTTP verbs that can modify resources (POST, PUT, PATCH, and
DELETE), you can submit your request in a dry run mode. Dry run mode helps to
evaluate a request through the typical request stages (admission chain, validation,
merge conflicts) up until persisting objects to storage. The response body for the
request is as close as possible to a non-dry-run response. Kubernetes guarantees that
dry-run requests will not be persisted in storage or have any other side effects.
Make a dry-run request
Dry-run is triggered by setting the dryRun query parameter. This parameter is a
string, working as an enum, and the only accepted values are:
[no value set]
Allow side effects. You request this with a query string such as ?dryRun
or ?dryRun&pretty=true. The response is the final object that would have been
persisted, or an error if the request could not be fulfilled.
All
Every stage runs as normal, except for the final storage stage where side effects
are prevented.
When you set ?dryRun=All, any relevant
admission controllers
are run, validating admission controllers check the request post-mutation, merge is
performed on PATCH, fields are defaulted, and schema validation occurs. The changes
are not persisted to the underlying storage, but the final object which would have
been persisted is still returned to the user, along with the normal status code.
If the non-dry-run version of a request would trigger an admission controller that has
side effects, the request will be failed rather than risk an unwanted side effect. All
built in admission control plugins support dry-run. Additionally, admission webhooks can
declare in their
configuration object
that they do not have side effects, by setting their sideEffects field to None.
Note: If a webhook actually does have side effects, then the sideEffects field should be
set to "NoneOnDryRun". That change is appropriate provided that the webhook is also
be modified to understand the DryRun field in AdmissionReview, and to prevent side
effects on any request marked as dry runs.
Here is an example dry-run request that uses ?dryRun=All:
POST /api/v1/namespaces/test/pods?dryRun=All
Content-Type: application/json
Accept: application/json
The response would look the same as for non-dry-run request, but the values of some
generated fields may differ.
Generated values
Some values of an object are typically generated before the object is persisted. It
is important not to rely upon the values of these fields set by a dry-run request,
since these values will likely be different in dry-run mode from when the real
request is made. Some of these fields are:
name: if generateName is set, name will have a unique random name
creationTimestamp / deletionTimestamp: records the time of creation/deletion
UID: uniquely identifies the object and is randomly generated (non-deterministic)
resourceVersion: tracks the persisted version of the object
Any field set by a mutating admission controller
For the Service resource: Ports or IP addresses that the kube-apiserver assigns to Service objects
Dry-run authorization
Authorization for dry-run and non-dry-run requests is identical. Thus, to make
a dry-run request, you must be authorized to make the non-dry-run request.
For example, to run a dry-run patch for a Deployment, you must be authorized
to perform that patch. Here is an example of a rule for Kubernetes
RBAC that allows patching
Deployments:
Kubernetes' Server Side Apply
feature allows the control plane to track managed fields for newly created objects.
Server Side Apply provides a clear pattern for managing field conflicts,
offers server-side Apply and Update operations, and replaces the
client-side functionality of kubectl apply.
The API verb for Server-Side Apply is apply.
See Server Side Apply for more details.
Resource versions
Resource versions are strings that identify the server's internal version of an
object. Resource versions can be used by clients to determine when objects have
changed, or to express data consistency requirements when getting, listing and
watching resources. Resource versions must be treated as opaque by clients and passed
unmodified back to the server.
You must not assume resource versions are numeric or collatable. API clients may
only compare two resource versions for equality (this means that you must not compare
resource versions for greater-than or less-than relationships).
resourceVersion fields in metadata
Clients find resource versions in resources, including the resources from the response
stream for a watch, or when using list to enumerate resources.
v1.meta/ObjectMeta - The metadata.resourceVersion of a resource instance identifies the resource version the instance was last modified at.
v1.meta/ListMeta - The metadata.resourceVersion of a resource collection (the response to a list) identifies the resource version at which the collection was constructed.
resourceVersion parameters in query strings
The get, list, and watch operations support the resourceVersion parameter.
From version v1.19, Kubernetes API servers also support the resourceVersionMatch
parameter on list requests.
The API server interprets the resourceVersion parameter differently depending
on the operation you request, and on the value of resourceVersion. If you set
resourceVersionMatch then this also affects the way matching happens.
Semantics for get and list
For get and list, the semantics of resourceVersion are:
get:
resourceVersion unset
resourceVersion="0"
resourceVersion="{value other than 0}"
Most Recent
Any
Not older than
list:
From version v1.19, Kubernetes API servers support the resourceVersionMatch parameter
on list requests. If you set both resourceVersion and resourceVersionMatch, the
resourceVersionMatch parameter determines how the API server interprets
resourceVersion.
You should always set the resourceVersionMatch parameter when setting
resourceVersion on a list request. However, be prepared to handle the case
where the API server that responds is unaware of resourceVersionMatch
and ignores it.
Unless you have strong consistency requirements, using resourceVersionMatch=NotOlderThan and
a known resourceVersion is preferable since it can achieve better performance and scalability
of your cluster than leaving resourceVersion and resourceVersionMatch unset, which requires
quorum read to be served.
Setting the resourceVersionMatch parameter without setting resourceVersion is not valid.
This table explains the behavior of list requests with various combinations of
resourceVersion and resourceVersionMatch:
resourceVersionMatch and paging parameters for list
resourceVersionMatch param
paging params
resourceVersion not set
resourceVersion="0"
resourceVersion="{value other than 0}"
unset
limit unset
Most Recent
Any
Not older than
unset
limit=<n>, continue unset
Most Recent
Any
Exact
unset
limit=<n>, continue=<token>
Continue Token, Exact
Invalid, treated as Continue Token, Exact
Invalid, HTTP 400 Bad Request
resourceVersionMatch=Exact
limit unset
Invalid
Invalid
Exact
resourceVersionMatch=Exact
limit=<n>, continue unset
Invalid
Invalid
Exact
resourceVersionMatch=NotOlderThan
limit unset
Invalid
Any
Not older than
resourceVersionMatch=NotOlderThan
limit=<n>, continue unset
Invalid
Any
Not older than
Note: If your cluster's API server does not honor the resourceVersionMatch parameter,
the behavior is the same as if you did not set it.
The meaning of the get and list semantics are:
Any
Return data at any resource version. The newest available resource version is preferred,
but strong consistency is not required; data at any resource version may be served. It is possible
for the request to return data at a much older resource version that the client has previously
observed, particularly in high availability configurations, due to partitions or stale
caches. Clients that cannot tolerate this should not use this semantic.
Most recent
Return data at the most recent resource version. The returned data must be
consistent (in detail: served from etcd via a quorum read).
Not older than
Return data at least as new as the provided resourceVersion. The newest
available data is preferred, but any data not older than the provided resourceVersion may be
served. For list requests to servers that honor the resourceVersionMatch parameter, this
guarantees that the collection's .metadata.resourceVersion is not older than the requested
resourceVersion, but does not make any guarantee about the .metadata.resourceVersion of any
of the items in that collection.
Exact
Return data at the exact resource version provided. If the provided resourceVersion is
unavailable, the server responds with HTTP 410 "Gone". For list requests to servers that honor the
resourceVersionMatch parameter, this guarantees that the collection's .metadata.resourceVersion
is the same as the resourceVersion you requested in the query string. That guarantee does
not apply to the .metadata.resourceVersion of any items within that collection.
Continue Token, Exact
Return data at the resource version of the initial paginated list call. The returned continue
tokens are responsible for keeping track of the initially provided resource version for all paginated
list calls after the initial paginated list.
Note: When you list resources and receive a collection response, the response includes the
metadata of the collection as
well as object metadata
for each item in that collection. For individual objects found within a collection response,
.metadata.resourceVersion tracks when that object was last updated, and not how up-to-date
the object is when served.
When using resourceVersionMatch=NotOlderThan and limit is set, clients must
handle HTTP 410 "Gone" responses. For example, the client might retry with a
newer resourceVersion or fall back to resourceVersion="".
When using resourceVersionMatch=Exact and limit is unset, clients must
verify that the collection's .metadata.resourceVersion matches
the requested resourceVersion, and handle the case where it does not. For
example, the client might fall back to a request with limit set.
Semantics for watch
For watch, the semantics of resource version are:
watch:
resourceVersion for watch
resourceVersion unset
resourceVersion="0"
resourceVersion="{value other than 0}"
Get State and Start at Most Recent
Get State and Start at Any
Start at Exact
The meaning of those watch semantics are:
Get State and Start at Any
Caution: Watches initialized this way may return arbitrarily stale
data. Please review this semantic before using it, and favor the other semantics
where possible.
Start a watch at any resource version; the most recent resource version
available is preferred, but not required. Any starting resource version is
allowed. It is possible for the watch to start at a much older resource
version that the client has previously observed, particularly in high availability
configurations, due to partitions or stale caches. Clients that cannot tolerate
this apparent rewinding should not start a watch with this semantic. To
establish initial state, the watch begins with synthetic "Added" events for
all resource instances that exist at the starting resource version. All following
watch events are for all changes that occurred after the resource version the
watch started at.
Get State and Start at Most Recent
Start a watch at the most recent resource version, which must be consistent
(in detail: served from etcd via a quorum read). To establish initial state,
the watch begins with synthetic "Added" events of all resources instances
that exist at the starting resource version. All following watch events are for
all changes that occurred after the resource version the watch started at.
Start at Exact
Start a watch at an exact resource version. The watch events are for all changes
after the provided resource version. Unlike "Get State and Start at Most Recent"
and "Get State and Start at Any", the watch is not started with synthetic
"Added" events for the provided resource version. The client is assumed to already
have the initial state at the starting resource version since the client provided
the resource version.
"410 Gone" responses
Servers are not required to serve all older resource versions and may return a HTTP
410 (Gone) status code if a client requests a resourceVersion older than the
server has retained. Clients must be able to tolerate 410 (Gone) responses. See
Efficient detection of changes for details on
how to handle 410 (Gone) responses when watching resources.
If you request a resourceVersion outside the applicable limit then, depending
on whether a request is served from cache or not, the API server may reply with a
410 Gone HTTP response.
Unavailable resource versions
Servers are not required to serve unrecognized resource versions. If you request
list or get for a resource version that the API server does not recognize,
then the API server may either:
wait briefly for the resource version to become available, then timeout with a
504 (Gateway Timeout) if the provided resource versions does not become available
in a reasonable amount of time;
respond with a Retry-After response header indicating how many seconds a client
should wait before retrying the request.
If you request a resource version that an API server does not recognize, the
kube-apiserver additionally identifies its error responses with a "Too large resource
version" message.
If you make a watch request for an unrecognized resource version, the API server
may wait indefinitely (until the request timeout) for the resource version to become
available.
6.2.2 - Server-Side Apply
FEATURE STATE:Kubernetes v1.22 [stable]
Introduction
Server Side Apply helps users and controllers manage their resources through
declarative configurations. Clients can create and modify their
objects
declaratively by sending their fully specified intent.
A fully specified intent is a partial object that only includes the fields and
values for which the user has an opinion. That intent either creates a new
object or is combined, by the server, with the existing object.
The system supports multiple appliers collaborating on a single object.
Changes to an object's fields are tracked through a "field management"
mechanism. When a field's value changes, ownership moves from its current
manager to the manager making the change. When trying to apply an object,
fields that have a different value and are owned by another manager will
result in a conflict. This is done in order to signal that the
operation might undo another collaborator's changes. Conflicts can be forced,
in which case the value will be overridden, and the ownership will be
transferred.
If you remove a field from a configuration and apply the configuration, server
side apply checks if there are any other field managers that also own the
field. If the field is not owned by any other field managers, it is either
deleted from the live object or reset to its default value, if it has one. The
same rule applies to associative list or map items.
Server side apply is meant both as a replacement for the original kubectl apply and as a simpler mechanism for controllers to enact their changes.
If you have Server Side Apply enabled, the control plane tracks managed fields
for all newly created objects.
Field Management
Compared to the last-applied annotation managed by kubectl, Server Side
Apply uses a more declarative approach, which tracks a user's field management,
rather than a user's last applied state. This means that as a side effect of
using Server Side Apply, information about which field manager manages each
field in an object also becomes available.
For a user to manage a field, in the Server Side Apply sense, means that the
user relies on and expects the value of the field not to change. The user who
last made an assertion about the value of a field will be recorded as the
current field manager. This can be done either by changing the value with
POST, PUT, or non-apply PATCH, or by including the field in a config sent
to the Server Side Apply endpoint. When using Server-Side Apply, trying to
change a field which is managed by someone else will result in a rejected
request (if not forced, see Conflicts).
When two or more appliers set a field to the same value, they share ownership of
that field. Any subsequent attempt to change the value of the shared field, by any of
the appliers, results in a conflict. Shared field owners may give up ownership
of a field by removing it from their configuration.
Field management is stored in amanagedFields field that is part of an object's
metadata.
A simple example of an object created by Server Side Apply could look like this:
apiVersion:v1kind:ConfigMapmetadata:name:test-cmnamespace:defaultlabels:test-label:testmanagedFields:- manager:kubectloperation:ApplyapiVersion:v1time:"2010-10-10T0:00:00Z"fieldsType:FieldsV1fieldsV1:f:metadata:f:labels:f:test-label:{}f:data:f:key:{}data:key:some value
The above object contains a single manager in metadata.managedFields. The
manager consists of basic information about the managing entity itself, like
operation type, API version, and the fields managed by it.
Note: This field is managed by the API server and should not be changed by
the user.
Nevertheless it is possible to change metadata.managedFields through an
Update operation. Doing so is highly discouraged, but might be a reasonable
option to try if, for example, the managedFields get into an inconsistent
state (which clearly should not happen).
The format of the managedFields is described in the
API.
Conflicts
A conflict is a special status error that occurs when an Apply operation tries
to change a field, which another user also claims to manage. This prevents an
applier from unintentionally overwriting the value set by another user. When
this occurs, the applier has 3 options to resolve the conflicts:
Overwrite value, become sole manager: If overwriting the value was
intentional (or if the applier is an automated process like a controller) the
applier should set the force query parameter to true (in kubectl, it can be done by
using the --force-conflicts flag with the apply command) and make the request
again. This forces the operation to succeed, changes the value of the field,
and removes the field from all other managers' entries in managedFields.
Don't overwrite value, give up management claim: If the applier doesn't
care about the value of the field anymore, they can remove it from their
config and make the request again. This leaves the value unchanged, and causes
the field to be removed from the applier's entry in managedFields.
Don't overwrite value, become shared manager: If the applier still cares
about the value of the field, but doesn't want to overwrite it, they can
change the value of the field in their config to match the value of the object
on the server, and make the request again. This leaves the value unchanged,
and causes the field's management to be shared by the applier and all other
field managers that already claimed to manage it.
Managers
Managers identify distinct workflows that are modifying the object (especially
useful on conflicts!), and can be specified through the fieldManager query
parameter as part of a modifying request. It is required for the apply endpoint,
though kubectl will default it to kubectl. For other updates, its default is
computed from the user-agent.
Apply and Update
The two operation types considered by this feature are Apply (PATCH with
content type application/apply-patch+yaml) and Update (all other operations
which modify the object). Both operations update the managedFields, but behave
a little differently.
Note:
Whether you are submitting JSON data or YAML data, use
application/apply-patch+yaml as the Content-Type header value.
All JSON documents are valid YAML.
For instance, only the apply operation fails on conflicts while update does
not. Also, apply operations are required to identify themselves by providing a
fieldManager query parameter, while the query parameter is optional for update
operations. Finally, when using the apply operation you cannot have
managedFields in the object that is being applied.
An example object with multiple managers could look like this:
apiVersion:v1kind:ConfigMapmetadata:name:test-cmnamespace:defaultlabels:test-label:testmanagedFields:- manager:kubectloperation:ApplyapiVersion:v1fields:f:metadata:f:labels:f:test-label:{}- manager:kube-controller-manageroperation:UpdateapiVersion:v1time:'2019-03-30T16:00:00.000Z'fields:f:data:f:key:{}data:key:new value
In this example, a second operation was run as an Update by the manager called
kube-controller-manager. The update changed a value in the data field which
caused the field's management to change to the kube-controller-manager.
If this update would have been an Apply operation, the operation
would have failed due to conflicting ownership.
Merge strategy
The merging strategy, implemented with Server Side Apply, provides a generally
more stable object lifecycle. Server Side Apply tries to merge fields based on
the actor who manages them instead of overruling based on values. This way
multiple actors can update the same object without causing unexpected interference.
When a user sends a "fully-specified intent" object to the Server Side Apply
endpoint, the server merges it with the live object favoring the value in the
applied config if it is specified in both places. If the set of items present in
the applied config is not a superset of the items applied by the same user last
time, each missing item not managed by any other appliers is removed. For
more information about how an object's schema is used to make decisions when
merging, see
sigs.k8s.io/structured-merge-diff.
A number of markers were added in Kubernetes 1.16 and 1.17, to allow API
developers to describe the merge strategy supported by lists, maps, and
structs. These markers can be applied to objects of the respective type,
in Go files or in the OpenAPI schema definition of the
CRD:
Golang marker
OpenAPI extension
Accepted values
Description
Introduced in
//+listType
x-kubernetes-list-type
atomic/set/map
Applicable to lists. set applies to lists that include only scalar elements. These elements must be unique. map applies to lists of nested types only. The key values (see listMapKey) must be unique in the list. atomic can apply to any list. If configured as atomic, the entire list is replaced during merge. At any point in time, a single manager owns the list. If set or map, different managers can manage entries separately.
1.16
//+listMapKey
x-kubernetes-list-map-keys
List of field names, e.g. ["port", "protocol"]
Only applicable when +listType=map. A list of field names whose values uniquely identify entries in the list. While there can be multiple keys, listMapKey is singular because keys need to be specified individually in the Go type. The key fields must be scalars.
1.16
//+mapType
x-kubernetes-map-type
atomic/granular
Applicable to maps. atomic means that the map can only be entirely replaced by a single manager. granular means that the map supports separate managers updating individual fields.
1.17
//+structType
x-kubernetes-map-type
atomic/granular
Applicable to structs; otherwise same usage and OpenAPI annotation as //+mapType.
1.17
If listType is missing, the API server interprets a
patchMergeStrategy=merge marker as a listType=map and the
corresponding patchMergeKey marker as a listMapKey.
The atomic list type is recursive.
These markers are specified as comments and don't have to be repeated as
field tags.
Compatibility across topology changes
On rare occurrences, a CRD or built-in type author may want to change the
specific topology of a field in their resource without incrementing its
version. Changing the topology of types, by upgrading the cluster or
updating the CRD, has different consequences when updating existing
objects. There are two categories of changes: when a field goes from
map/set/granular to atomic and the other way around.
When the listType, mapType, or structType changes from
map/set/granular to atomic, the whole list, map, or struct of
existing objects will end-up being owned by actors who owned an element
of these types. This means that any further change to these objects
would cause a conflict.
When a list, map, or struct changes from atomic to
map/set/granular, the API server won't be able to infer the new
ownership of these fields. Because of that, no conflict will be produced
when objects have these fields updated. For that reason, it is not
recommended to change a type from atomic to map/set/granular.
Before spec.data gets changed from atomic to granular,
manager-one owns the field spec.data, and all the fields within it
(key1 and key2). When the CRD gets changed to make spec.datagranular, manager-one continues to own the top-level field
spec.data (meaning no other managers can delete the map called data
without a conflict), but it no longer owns key1 and key2, so another
manager can then modify or delete those fields without conflict.
Custom Resources
By default, Server Side Apply treats custom resources as unstructured data. All
keys are treated the same as struct fields, and all lists are considered atomic.
If the Custom Resource Definition defines a
schema
that contains annotations as defined in the previous "Merge Strategy"
section, these annotations will be used when merging objects of this
type.
Using Server-Side Apply in a controller
As a developer of a controller, you can use server-side apply as a way to
simplify the update logic of your controller. The main differences with a
read-modify-write and/or patch are the following:
the applied object must contain all the fields that the controller cares about.
there is no way to remove fields that haven't been applied by the controller
before (controller can still send a PATCH/UPDATE for these use-cases).
the object doesn't have to be read beforehand, resourceVersion doesn't have
to be specified.
It is strongly recommended for controllers to always "force" conflicts, since they
might not be able to resolve or act on these conflicts.
Transferring Ownership
In addition to the concurrency controls provided by conflict resolution,
Server Side Apply provides ways to perform coordinated
field ownership transfers from users to controllers.
This is best explained by example. Let's look at how to safely transfer
ownership of the replicas field from a user to a controller while enabling
automatic horizontal scaling for a Deployment, using the HorizontalPodAutoscaler
resource and its accompanying controller.
Say a user has defined deployment with replicas set to the desired value:
Now, the user would like to remove replicas from their configuration, so they
don't accidentally fight with the HPA controller. However, there is a race: it
might take some time before HPA feels the need to adjust replicas, and if
the user removes replicas before the HPA writes to the field and becomes
its owner, then apiserver will set replicas to 1, its default value. This
is not what the user wants to happen, even temporarily.
There are two solutions:
(basic) Leave replicas in the configuration; when HPA eventually writes to that
field, the system gives the user a conflict over it. At that point, it is safe
to remove from the configuration.
(more advanced) If, however, the user doesn't want to wait, for example
because they want to keep the cluster legible to coworkers, then they can take
the following steps to make it safe to remove replicas from their
configuration:
First, the user defines a new configuration containing only the replicas field:
If the apply results in a conflict with the HPA controller, then do nothing. The
conflict indicates the controller has claimed the field earlier in the
process than it sometimes does.
At this point the user may remove the replicas field from their configuration.
Note that whenever the HPA controller sets the replicas field to a new value,
the temporary field manager will no longer own any fields and will be
automatically deleted. No clean up is required.
Transferring Ownership Between Users
Users can transfer ownership of a field between each other by setting the field
to the same value in both of their applied configs, causing them to share
ownership of the field. Once the users share ownership of the field, one of them
can remove the field from their applied configuration to give up ownership and
complete the transfer to the other user.
Comparison with Client Side Apply
A consequence of the conflict detection and resolution implemented by Server
Side Apply is that an applier always has up to date field values in their local
state. If they don't, they get a conflict the next time they apply. Any of the
three options to resolve conflicts results in the applied configuration being an
up to date subset of the object on the server's fields.
This is different from Client Side Apply, where outdated values which have been
overwritten by other users are left in an applier's local config. These values
only become accurate when the user updates that specific field, if ever, and an
applier has no way of knowing whether their next apply will overwrite other
users' changes.
Another difference is that an applier using Client Side Apply is unable to
change the API version they are using, but Server Side Apply supports this use
case.
Upgrading from client-side apply to server-side apply
Client-side apply users who manage a resource with kubectl apply can start
using server-side apply with the following flag.
kubectl apply --server-side [--dry-run=server]
By default, field management of the object transfers from client-side apply to
kubectl server-side apply without encountering conflicts.
Caution:
Keep the last-applied-configuration annotation up to date.
The annotation infers client-side apply's managed fields.
Any fields not managed by client-side apply raise conflicts.
For example, if you used kubectl scale to update the replicas field after
client-side apply, then this field is not owned by client-side apply and
creates conflicts on kubectl apply --server-side.
This behavior applies to server-side apply with the kubectl field manager.
As an exception, you can opt-out of this behavior by specifying a different,
non-default field manager, as seen in the following example. The default field
manager for kubectl server-side apply is kubectl.
Downgrading from server-side apply to client-side apply
If you manage a resource with kubectl apply --server-side,
you can downgrade to client-side apply directly with kubectl apply.
Downgrading works because kubectl server-side apply keeps the
last-applied-configuration annotation up-to-date if you use
kubectl apply.
This behavior applies to server-side apply with the kubectl field manager.
As an exception, you can opt-out of this behavior by specifying a different,
non-default field manager, as seen in the following example. The default field
manager for kubectl server-side apply is kubectl.
With the Server Side Apply feature enabled, the PATCH endpoint accepts the
additional application/apply-patch+yaml content type. Users of Server Side
Apply can send partially specified objects as YAML to this endpoint. When
applying a configuration, one should always include all the fields that they
have an opinion about.
Clearing ManagedFields
It is possible to strip all managedFields from an object by overwriting them
using MergePatch, StrategicMergePatch, JSONPatch, or Update, so every
non-apply operation. This can be done by overwriting the managedFields field
with an empty entry. Two examples are:
This will overwrite the managedFields with a list containing a single empty
entry that then results in the managedFields being stripped entirely from the
object. Note that setting the managedFields to an empty list will not
reset the field. This is on purpose, so managedFields never get stripped by
clients not aware of the field.
In cases where the reset operation is combined with changes to other fields
than the managedFields, this will result in the managedFields being reset
first and the other changes being processed afterwards. As a result the
applier takes ownership of any fields updated in the same request.
Caution: Server Side Apply does not correctly track ownership on
sub-resources that don't receive the resource object type. If you are
using Server Side Apply with such a sub-resource, the changed fields
won't be tracked.
6.2.3 - Client Libraries
This page contains an overview of the client libraries for using the Kubernetes
API from various programming languages.
To write applications using the Kubernetes REST API,
you do not need to implement the API calls and request/response types yourself.
You can use a client library for the programming language you are using.
Client libraries often handle common tasks such as authentication for you.
Most client libraries can discover and use the Kubernetes Service Account to
authenticate if the API client is running inside the Kubernetes cluster, or can
understand the kubeconfig file
format to read the credentials and the API Server address.
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
The following Kubernetes API client libraries are provided and maintained by
their authors, not the Kubernetes team.
This document details the deprecation policy for various facets of the system.
Kubernetes is a large system with many components and many contributors. As
with any such software, the feature set naturally evolves over time, and
sometimes a feature may need to be removed. This could include an API, a flag,
or even an entire feature. To avoid breaking existing users, Kubernetes follows
a deprecation policy for aspects of the system that are slated to be removed.
Deprecating parts of the API
Since Kubernetes is an API-driven system, the API has evolved over time to
reflect the evolving understanding of the problem space. The Kubernetes API is
actually a set of APIs, called "API groups", and each API group is
independently versioned. API versions fall
into 3 main tracks, each of which has different policies for deprecation:
Example
Track
v1
GA (generally available, stable)
v1beta1
Beta (pre-release)
v1alpha1
Alpha (experimental)
A given release of Kubernetes can support any number of API groups and any
number of versions of each.
The following rules govern the deprecation of elements of the API. This
includes:
REST resources (aka API objects)
Fields of REST resources
Annotations on REST resources, including "beta" annotations but not
including "alpha" annotations.
Enumerated or constant values
Component config structures
These rules are enforced between official releases, not between
arbitrary commits to master or release branches.
Rule #1: API elements may only be removed by incrementing the version of the
API group.
Once an API element has been added to an API group at a particular version, it
can not be removed from that version or have its behavior significantly
changed, regardless of track.
Note: For historical reasons, there are 2 "monolithic" API groups - "core" (no
group name) and "extensions". Resources will incrementally be moved from these
legacy API groups into more domain-specific API groups.
Rule #2: API objects must be able to round-trip between API versions in a given
release without information loss, with the exception of whole REST resources
that do not exist in some versions.
For example, an object can be written as v1 and then read back as v2 and
converted to v1, and the resulting v1 resource will be identical to the
original. The representation in v2 might be different from v1, but the system
knows how to convert between them in both directions. Additionally, any new
field added in v2 must be able to round-trip to v1 and back, which means v1
might have to add an equivalent field or represent it as an annotation.
Rule #3: An API version in a given track may not be deprecated in favor of a less stable API version.
GA API versions can replace beta and alpha API versions.
Beta API versions can replace earlier beta and alpha API versions, but may not replace GA API versions.
Alpha API versions can replace earlier alpha API versions, but may not replace GA or beta API versions.
Rule #4a: minimum API lifetime is determined by the API stability level
GA API versions may be marked as deprecated, but must not be removed within a major version of Kubernetes
Beta API versions must be supported for 9 months or 3 releases (whichever is longer) after deprecation
Alpha API versions may be removed in any release without prior deprecation notice
Note: There are no current plans for a major version revision of Kubernetes that removes GA APIs.
Note: Until #52185 is
resolved, no API versions that have been persisted to storage may be removed.
Serving REST endpoints for those versions may be disabled (subject to the
deprecation timelines in this document), but the API server must remain capable
of decoding/converting previously persisted data from storage.
Rule #4b: The "preferred" API version and the "storage version" for a given
group may not advance until after a release has been made that supports both the
new version and the previous version
Users must be able to upgrade to a new release of Kubernetes and then roll back
to a previous release, without converting anything to the new API version or
suffering breakages (unless they explicitly used features only available in the
newer version). This is particularly evident in the stored representation of
objects.
All of this is best illustrated by examples. Imagine a Kubernetes release,
version X, which introduces a new API group. A new Kubernetes release is made
every approximately 4 months (3 per year). The following table describes which
API versions are supported in a series of subsequent releases.
Consider a hypothetical REST resource named Widget, which was present in API v1
in the above timeline, and which needs to be deprecated. We document and
announce the
deprecation in sync with release X+1. The Widget resource still exists in API
version v1 (deprecated) but not in v2alpha1. The Widget resource continues to
exist and function in releases up to and including X+8. Only in release X+9,
when API v1 has aged out, does the Widget resource cease to exist, and the
behavior get removed.
Starting in Kubernetes v1.19, making an API request to a deprecated REST API endpoint:
Returns a Warning header (as defined in RFC7234, Section 5.5) in the API response.
Adds a "k8s.io/deprecated":"true" annotation to the audit event recorded for the request.
Sets an apiserver_requested_deprecated_apis gauge metric to 1 in the kube-apiserver
process. The metric has labels for group, version, resource, subresource that can be joined
to the apiserver_request_total metric, and a removed_release label that indicates the
Kubernetes release in which the API will no longer be served. The following Prometheus query
returns information about requests made to deprecated APIs which will be removed in v1.22:
As with whole REST resources, an individual field which was present in API v1
must exist and function until API v1 is removed. Unlike whole resources, the
v2 APIs may choose a different representation for the field, as long as it can
be round-tripped. For example a v1 field named "magnitude" which was
deprecated might be named "deprecatedMagnitude" in API v2. When v1 is
eventually removed, the deprecated field can be removed from v2.
Enumerated or constant values
As with whole REST resources and fields thereof, a constant value which was
supported in API v1 must exist and function until API v1 is removed.
Component config structures
Component configs are versioned and managed similar to REST resources.
Future work
Over time, Kubernetes will introduce more fine-grained API versions, at which
point these rules will be adjusted as needed.
Deprecating a flag or CLI
The Kubernetes system is comprised of several different programs cooperating.
Sometimes, a Kubernetes release might remove flags or CLI commands
(collectively "CLI elements") in these programs. The individual programs
naturally sort into two main groups - user-facing and admin-facing programs,
which vary slightly in their deprecation policies. Unless a flag is explicitly
prefixed or documented as "alpha" or "beta", it is considered GA.
CLI elements are effectively part of the API to the system, but since they are
not versioned in the same way as the REST API, the rules for deprecation are as
follows:
Rule #5a: CLI elements of user-facing components (e.g. kubectl) must function
after their announced deprecation for no less than:
GA: 12 months or 2 releases (whichever is longer)
Beta: 3 months or 1 release (whichever is longer)
Alpha: 0 releases
Rule #5b: CLI elements of admin-facing components (e.g. kubelet) must function
after their announced deprecation for no less than:
GA: 6 months or 1 release (whichever is longer)
Beta: 3 months or 1 release (whichever is longer)
Alpha: 0 releases
Rule #6: Deprecated CLI elements must emit warnings (optionally disable)
when used.
Deprecating a feature or behavior
Occasionally a Kubernetes release needs to deprecate some feature or behavior
of the system that is not controlled by the API or CLI. In this case, the
rules for deprecation are as follows:
Rule #7: Deprecated behaviors must function for no less than 1 year after their
announced deprecation.
This does not imply that all changes to the system are governed by this policy.
This applies only to significant, user-visible behaviors which impact the
correctness of applications running on Kubernetes or that impact the
administration of Kubernetes clusters, and which are being removed entirely.
An exception to the above rule is feature gates. Feature gates are key=value
pairs that allow for users to enable/disable experimental features.
Feature gates are intended to cover the development life cycle of a feature - they
are not intended to be long-term APIs. As such, they are expected to be deprecated
and removed after a feature becomes GA or is dropped.
As a feature moves through the stages, the associated feature gate evolves.
The feature life cycle matched to its corresponding feature gate is:
Alpha: the feature gate is disabled by default and can be enabled by the user.
Beta: the feature gate is enabled by default and can be disabled by the user.
GA: the feature gate is deprecated (see "Deprecation") and becomes
non-operational.
GA, deprecation window complete: the feature gate is removed and calls to it are
no longer accepted.
Deprecation
Features can be removed at any point in the life cycle prior to GA. When features are
removed prior to GA, their associated feature gates are also deprecated.
When an invocation tries to disable a non-operational feature gate, the call fails in order
to avoid unsupported scenarios that might otherwise run silently.
In some cases, removing pre-GA features requires considerable time. Feature gates can remain
operational until their associated feature is fully removed, at which point the feature gate
itself can be deprecated.
When removing a feature gate for a GA feature also requires considerable time, calls to
feature gates may remain operational if the feature gate has no effect on the feature,
and if the feature gate causes no errors.
Features intended to be disabled by users should include a mechanism for disabling the
feature in the associated feature gate.
Versioning for feature gates is different from the previously discussed components,
therefore the rules for deprecation are as follows:
Rule #8: Feature gates must be deprecated when the corresponding feature they control
transitions a lifecycle stage as follows. Feature gates must function for no less than:
Beta feature to GA: 6 months or 2 releases (whichever is longer)
Beta feature to EOL: 3 months or 1 release (whichever is longer)
Alpha feature to EOL: 0 releases
Rule #9: Deprecated feature gates must respond with a warning when used. When a feature gate
is deprecated it must be documented in both in the release notes and the corresponding CLI help.
Both warnings and documentation must indicate whether a feature gate is non-operational.
Deprecating a metric
Each component of the Kubernetes control-plane exposes metrics (usually the
/metrics endpoint), which are typically ingested by cluster administrators.
Not all metrics are the same: some metrics are commonly used as SLIs or used
to determine SLOs, these tend to have greater import. Other metrics are more
experimental in nature or are used primarily in the Kubernetes development
process.
Accordingly, metrics fall under two stability classes (ALPHA and STABLE);
this impacts removal of a metric during a Kubernetes release. These classes
are determined by the perceived importance of the metric. The rules for
deprecating and removing a metric are as follows:
Rule #9a: Metrics, for the corresponding stability class, must function for no less than:
STABLE: 4 releases or 12 months (whichever is longer)
ALPHA: 0 releases
Rule #9b: Metrics, after their announced deprecation, must function for no less than:
STABLE: 3 releases or 9 months (whichever is longer)
ALPHA: 0 releases
Deprecated metrics will have their description text prefixed with a deprecation notice
string '(Deprecated from x.y)' and a warning log will be emitted during metric
registration. Like their stable undeprecated counterparts, deprecated metrics will
be automatically registered to the metrics endpoint and therefore visible.
On a subsequent release (when the metric's deprecatedVersion is equal to
current_kubernetes_version - 3)), a deprecated metric will become a hidden metric.
Unlike their deprecated counterparts, hidden metrics will no longer be
automatically registered to the metrics endpoint (hence hidden). However, they
can be explicitly enabled through a command line flag on the binary
(--show-hidden-metrics-for-version=). This provides cluster admins an
escape hatch to properly migrate off of a deprecated metric, if they were not
able to react to the earlier deprecation warnings. Hidden metrics should be
deleted after one release.
Exceptions
No policy can cover every possible situation. This policy is a living
document, and will evolve over time. In practice, there will be situations
that do not fit neatly into this policy, or for which this policy becomes a
serious impediment. Such situations should be discussed with SIGs and project
leaders to find the best solutions for those specific cases, always bearing in
mind that Kubernetes is committed to being a stable system that, as much as
possible, never breaks users. Exceptions will always be announced in all
relevant release notes.
6.2.5 - Deprecated API Migration Guide
As the Kubernetes API evolves, APIs are periodically reorganized or upgraded.
When APIs evolve, the old API is deprecated and eventually removed.
This page contains information you need to know when migrating from
deprecated API versions to newer and more stable API versions.
Removed APIs by release
v1.27
The v1.27 release will stop serving the following deprecated API versions:
CSIStorageCapacity
The storage.k8s.io/v1beta1 API version of CSIStorageCapacity will no longer be served in v1.27.
Migrate manifests and API clients to use the storage.k8s.io/v1 API version, available since v1.24.
All existing persisted objects are accessible via the new API
No notable changes
v1.26
The v1.26 release will stop serving the following deprecated API versions:
Flow control resources
The flowcontrol.apiserver.k8s.io/v1beta1 API version of FlowSchema and PriorityLevelConfiguration will no longer be served in v1.26.
Migrate manifests and API clients to use the flowcontrol.apiserver.k8s.io/v1beta2 API version, available since v1.23.
All existing persisted objects are accessible via the new API
No notable changes
HorizontalPodAutoscaler
The autoscaling/v2beta2 API version of HorizontalPodAutoscaler will no longer be served in v1.26.
Migrate manifests and API clients to use the autoscaling/v2 API version, available since v1.23.
All existing persisted objects are accessible via the new API
v1.25
The v1.25 release will stop serving the following deprecated API versions:
CronJob
The batch/v1beta1 API version of CronJob will no longer be served in v1.25.
Migrate manifests and API clients to use the batch/v1 API version, available since v1.21.
All existing persisted objects are accessible via the new API
No notable changes
EndpointSlice
The discovery.k8s.io/v1beta1 API version of EndpointSlice will no longer be served in v1.25.
Migrate manifests and API clients to use the discovery.k8s.io/v1 API version, available since v1.21.
All existing persisted objects are accessible via the new API
Notable changes in discovery.k8s.io/v1:
use per Endpoint nodeName field instead of deprecated topology["kubernetes.io/hostname"] field
use per Endpoint zone field instead of deprecated topology["topology.kubernetes.io/zone"] field
topology is replaced with the deprecatedTopology field which is not writable in v1
Event
The events.k8s.io/v1beta1 API version of Event will no longer be served in v1.25.
Migrate manifests and API clients to use the events.k8s.io/v1 API version, available since v1.19.
All existing persisted objects are accessible via the new API
Notable changes in events.k8s.io/v1:
type is limited to Normal and Warning
involvedObject is renamed to regarding
action, reason, reportingController, and reportingInstance are required when creating new events.k8s.io/v1 Events
use eventTime instead of the deprecated firstTimestamp field (which is renamed to deprecatedFirstTimestamp and not permitted in new events.k8s.io/v1 Events)
use series.lastObservedTime instead of the deprecated lastTimestamp field (which is renamed to deprecatedLastTimestamp and not permitted in new events.k8s.io/v1 Events)
use series.count instead of the deprecated count field (which is renamed to deprecatedCount and not permitted in new events.k8s.io/v1 Events)
use reportingController instead of the deprecated source.component field (which is renamed to deprecatedSource.component and not permitted in new events.k8s.io/v1 Events)
use reportingInstance instead of the deprecated source.host field (which is renamed to deprecatedSource.host and not permitted in new events.k8s.io/v1 Events)
HorizontalPodAutoscaler
The autoscaling/v2beta1 API version of HorizontalPodAutoscaler will no longer be served in v1.25.
Migrate manifests and API clients to use the autoscaling/v2 API version, available since v1.23.
All existing persisted objects are accessible via the new API
PodDisruptionBudget
The policy/v1beta1 API version of PodDisruptionBudget will no longer be served in v1.25.
Migrate manifests and API clients to use the policy/v1 API version, available since v1.21.
All existing persisted objects are accessible via the new API
Notable changes in policy/v1:
an empty spec.selector ({}) written to a policy/v1 PodDisruptionBudget selects all pods in the namespace (in policy/v1beta1 an empty spec.selector selected no pods). An unset spec.selector selects no pods in either API version.
PodSecurityPolicy
PodSecurityPolicy in the policy/v1beta1 API version will no longer be served in v1.25, and the PodSecurityPolicy admission controller will be removed.
RuntimeClass in the node.k8s.io/v1beta1 API version will no longer be served in v1.25.
Migrate manifests and API clients to use the node.k8s.io/v1 API version, available since v1.20.
All existing persisted objects are accessible via the new API
No notable changes
v1.22
The v1.22 release stopped serving the following deprecated API versions:
Webhook resources
The admissionregistration.k8s.io/v1beta1 API version of MutatingWebhookConfiguration and ValidatingWebhookConfiguration is no longer served as of v1.22.
Migrate manifests and API clients to use the admissionregistration.k8s.io/v1 API version, available since v1.16.
All existing persisted objects are accessible via the new APIs
Notable changes:
webhooks[*].failurePolicy default changed from Ignore to Fail for v1
webhooks[*].matchPolicy default changed from Exact to Equivalent for v1
webhooks[*].timeoutSeconds default changed from 30s to 10s for v1
webhooks[*].sideEffects default value is removed, and the field made required, and only None and NoneOnDryRun are permitted for v1
webhooks[*].admissionReviewVersions default value is removed and the field made required for v1 (supported versions for AdmissionReview are v1 and v1beta1)
webhooks[*].name must be unique in the list for objects created via admissionregistration.k8s.io/v1
CustomResourceDefinition
The apiextensions.k8s.io/v1beta1 API version of CustomResourceDefinition is no longer served as of v1.22.
Migrate manifests and API clients to use the apiextensions.k8s.io/v1 API version, available since v1.16.
All existing persisted objects are accessible via the new API
Notable changes:
spec.scope is no longer defaulted to Namespaced and must be explicitly specified
spec.version is removed in v1; use spec.versions instead
spec.validation is removed in v1; use spec.versions[*].schema instead
spec.subresources is removed in v1; use spec.versions[*].subresources instead
spec.additionalPrinterColumns is removed in v1; use spec.versions[*].additionalPrinterColumns instead
spec.conversion.webhookClientConfig is moved to spec.conversion.webhook.clientConfig in v1
spec.conversion.conversionReviewVersions is moved to spec.conversion.webhook.conversionReviewVersions in v1
spec.versions[*].schema.openAPIV3Schema is now required when creating v1 CustomResourceDefinition objects, and must be a structural schema
spec.preserveUnknownFields: true is disallowed when creating v1 CustomResourceDefinition objects; it must be specified within schema definitions as x-kubernetes-preserve-unknown-fields: true
In additionalPrinterColumns items, the JSONPath field was renamed to jsonPath in v1 (fixes #66531)
APIService
The apiregistration.k8s.io/v1beta1 API version of APIService is no longer served as of v1.22.
Migrate manifests and API clients to use the apiregistration.k8s.io/v1 API version, available since v1.10.
All existing persisted objects are accessible via the new API
No notable changes
TokenReview
The authentication.k8s.io/v1beta1 API version of TokenReview is no longer served as of v1.22.
Migrate manifests and API clients to use the authentication.k8s.io/v1 API version, available since v1.6.
No notable changes
SubjectAccessReview resources
The authorization.k8s.io/v1beta1 API version of LocalSubjectAccessReview, SelfSubjectAccessReview, SubjectAccessReview, and SelfSubjectRulesReview is no longer served as of v1.22.
Migrate manifests and API clients to use the authorization.k8s.io/v1 API version, available since v1.6.
Notable changes:
spec.group was renamed to spec.groups in v1 (fixes #32709)
CertificateSigningRequest
The certificates.k8s.io/v1beta1 API version of CertificateSigningRequest is no longer served as of v1.22.
Migrate manifests and API clients to use the certificates.k8s.io/v1 API version, available since v1.19.
All existing persisted objects are accessible via the new API
Notable changes in certificates.k8s.io/v1:
For API clients requesting certificates:
spec.signerName is now required (see known Kubernetes signers), and requests for kubernetes.io/legacy-unknown are not allowed to be created via the certificates.k8s.io/v1 API
spec.usages is now required, may not contain duplicate values, and must only contain known usages
For API clients approving or signing certificates:
status.conditions may not contain duplicate types
status.conditions[*].status is now required
status.certificate must be PEM-encoded, and contain only CERTIFICATE blocks
Lease
The coordination.k8s.io/v1beta1 API version of Lease is no longer served as of v1.22.
Migrate manifests and API clients to use the coordination.k8s.io/v1 API version, available since v1.14.
All existing persisted objects are accessible via the new API
No notable changes
Ingress
The extensions/v1beta1 and networking.k8s.io/v1beta1 API versions of Ingress is no longer served as of v1.22.
Migrate manifests and API clients to use the networking.k8s.io/v1 API version, available since v1.19.
All existing persisted objects are accessible via the new API
Notable changes:
spec.backend is renamed to spec.defaultBackend
The backend serviceName field is renamed to service.name
Numeric backend servicePort fields are renamed to service.port.number
String backend servicePort fields are renamed to service.port.name
pathType is now required for each specified path. Options are Prefix, Exact, and ImplementationSpecific. To match the undefined v1beta1 behavior, use ImplementationSpecific.
IngressClass
The networking.k8s.io/v1beta1 API version of IngressClass is no longer served as of v1.22.
Migrate manifests and API clients to use the networking.k8s.io/v1 API version, available since v1.19.
All existing persisted objects are accessible via the new API
No notable changes
RBAC resources
The rbac.authorization.k8s.io/v1beta1 API version of ClusterRole, ClusterRoleBinding, Role, and RoleBinding is no longer served as of v1.22.
Migrate manifests and API clients to use the rbac.authorization.k8s.io/v1 API version, available since v1.8.
All existing persisted objects are accessible via the new APIs
No notable changes
PriorityClass
The scheduling.k8s.io/v1beta1 API version of PriorityClass is no longer served as of v1.22.
Migrate manifests and API clients to use the scheduling.k8s.io/v1 API version, available since v1.14.
All existing persisted objects are accessible via the new API
No notable changes
Storage resources
The storage.k8s.io/v1beta1 API version of CSIDriver, CSINode, StorageClass, and VolumeAttachment is no longer served as of v1.22.
Migrate manifests and API clients to use the storage.k8s.io/v1 API version
CSIDriver is available in storage.k8s.io/v1 since v1.19.
CSINode is available in storage.k8s.io/v1 since v1.17
StorageClass is available in storage.k8s.io/v1 since v1.6
VolumeAttachment is available in storage.k8s.io/v1 v1.13
All existing persisted objects are accessible via the new APIs
No notable changes
v1.16
The v1.16 release stopped serving the following deprecated API versions:
NetworkPolicy
The extensions/v1beta1 API version of NetworkPolicy is no longer served as of v1.16.
Migrate manifests and API clients to use the networking.k8s.io/v1 API version, available since v1.8.
All existing persisted objects are accessible via the new API
DaemonSet
The extensions/v1beta1 and apps/v1beta2 API versions of DaemonSet are no longer served as of v1.16.
Migrate manifests and API clients to use the apps/v1 API version, available since v1.9.
All existing persisted objects are accessible via the new API
Notable changes:
spec.templateGeneration is removed
spec.selector is now required and immutable after creation; use the existing template labels as the selector for seamless upgrades
spec.updateStrategy.type now defaults to RollingUpdate (the default in extensions/v1beta1 was OnDelete)
Deployment
The extensions/v1beta1, apps/v1beta1, and apps/v1beta2 API versions of Deployment are no longer served as of v1.16.
Migrate manifests and API clients to use the apps/v1 API version, available since v1.9.
All existing persisted objects are accessible via the new API
Notable changes:
spec.rollbackTo is removed
spec.selector is now required and immutable after creation; use the existing template labels as the selector for seamless upgrades
spec.progressDeadlineSeconds now defaults to 600 seconds (the default in extensions/v1beta1 was no deadline)
spec.revisionHistoryLimit now defaults to 10 (the default in apps/v1beta1 was 2, the default in extensions/v1beta1 was to retain all)
maxSurge and maxUnavailable now default to 25% (the default in extensions/v1beta1 was 1)
StatefulSet
The apps/v1beta1 and apps/v1beta2 API versions of StatefulSet are no longer served as of v1.16.
Migrate manifests and API clients to use the apps/v1 API version, available since v1.9.
All existing persisted objects are accessible via the new API
Notable changes:
spec.selector is now required and immutable after creation; use the existing template labels as the selector for seamless upgrades
spec.updateStrategy.type now defaults to RollingUpdate (the default in apps/v1beta1 was OnDelete)
ReplicaSet
The extensions/v1beta1, apps/v1beta1, and apps/v1beta2 API versions of ReplicaSet are no longer served as of v1.16.
Migrate manifests and API clients to use the apps/v1 API version, available since v1.9.
All existing persisted objects are accessible via the new API
Notable changes:
spec.selector is now required and immutable after creation; use the existing template labels as the selector for seamless upgrades
PodSecurityPolicy
The extensions/v1beta1 API version of PodSecurityPolicy is no longer served as of v1.16.
Migrate manifests and API client to use the policy/v1beta1 API version, available since v1.10.
Note that the policy/v1beta1 API version of PodSecurityPolicy will be removed in v1.25.
What to do
Test with deprecated APIs disabled
You can test your clusters by starting an API server with specific API versions disabled
to simulate upcoming removals. Add the following flag to the API server startup arguments:
Note that this may use non-ideal default values. To learn more about a specific
resource, check the Kubernetes API reference.
6.2.6 - Kubernetes API health endpoints
The Kubernetes API server provides API endpoints to indicate the current status of the API server.
This page describes these API endpoints and explains how you can use them.
API endpoints for health
The Kubernetes API server provides 3 API endpoints (healthz, livez and readyz) to indicate the current status of the API server.
The healthz endpoint is deprecated (since Kubernetes v1.16), and you should use the more specific livez and readyz endpoints instead.
The livez endpoint can be used with the --livez-grace-periodflag to specify the startup duration.
For a graceful shutdown you can specify the --shutdown-delay-durationflag with the /readyz endpoint.
Machines that check the healthz/livez/readyz of the API server should rely on the HTTP status code.
A status code 200 indicates the API server is healthy/live/ready, depending on the called endpoint.
The more verbose options shown below are intended to be used by human operators to debug their cluster or understand the state of the API server.
The following examples will show how you can interact with the health API endpoints.
For all endpoints, you can use the verbose parameter to print out the checks and their status.
This can be useful for a human operator to debug the current status of the API server, it is not intended to be consumed by a machine:
curl -k https://localhost:6443/livez?verbose
or from a remote host with authentication:
kubectl get --raw='/readyz?verbose'
The output will look like this:
[+]ping ok
[+]log ok
[+]etcd ok
[+]poststarthook/start-kube-apiserver-admission-initializer ok
[+]poststarthook/generic-apiserver-start-informers ok
[+]poststarthook/start-apiextensions-informers ok
[+]poststarthook/start-apiextensions-controllers ok
[+]poststarthook/crd-informer-synced ok
[+]poststarthook/bootstrap-controller ok
[+]poststarthook/rbac/bootstrap-roles ok
[+]poststarthook/scheduling/bootstrap-system-priority-classes ok
[+]poststarthook/start-cluster-authentication-info-controller ok
[+]poststarthook/start-kube-aggregator-informers ok
[+]poststarthook/apiservice-registration-controller ok
[+]poststarthook/apiservice-status-available-controller ok
[+]poststarthook/kube-apiserver-autoregistration ok
[+]autoregister-completion ok
[+]poststarthook/apiservice-openapi-controller ok
healthz check passed
The Kubernetes API server also supports to exclude specific checks.
The query parameters can also be combined like in this example:
[+]ping ok
[+]log ok
[+]etcd excluded: ok
[+]poststarthook/start-kube-apiserver-admission-initializer ok
[+]poststarthook/generic-apiserver-start-informers ok
[+]poststarthook/start-apiextensions-informers ok
[+]poststarthook/start-apiextensions-controllers ok
[+]poststarthook/crd-informer-synced ok
[+]poststarthook/bootstrap-controller ok
[+]poststarthook/rbac/bootstrap-roles ok
[+]poststarthook/scheduling/bootstrap-system-priority-classes ok
[+]poststarthook/start-cluster-authentication-info-controller ok
[+]poststarthook/start-kube-aggregator-informers ok
[+]poststarthook/apiservice-registration-controller ok
[+]poststarthook/apiservice-status-available-controller ok
[+]poststarthook/kube-apiserver-autoregistration ok
[+]autoregister-completion ok
[+]poststarthook/apiservice-openapi-controller ok
[+]shutdown ok
healthz check passed
Individual health checks
FEATURE STATE:Kubernetes v1.25 [alpha]
Each individual health check exposes an HTTP endpoint and can be checked individually.
The schema for the individual health checks is /livez/<healthcheck-name> where livez and readyz and be used to indicate if you want to check the liveness or the readiness of the API server.
The <healthcheck-name> path can be discovered using the verbose flag from above and take the path between [+] and ok.
These individual health checks should not be consumed by machines but can be helpful for a human operator to debug a system:
All Kubernetes clusters have two categories of users: service accounts managed
by Kubernetes, and normal users.
It is assumed that a cluster-independent service manages normal users in the following ways:
an administrator distributing private keys
a user store like Keystone or Google Accounts
a file with a list of usernames and passwords
In this regard, Kubernetes does not have objects which represent normal user
accounts. Normal users cannot be added to a cluster through an API call.
Even though a normal user cannot be added via an API call, any user that
presents a valid certificate signed by the cluster's certificate authority
(CA) is considered authenticated. In this configuration, Kubernetes determines
the username from the common name field in the 'subject' of the cert (e.g.,
"/CN=bob"). From there, the role based access control (RBAC) sub-system would
determine whether the user is authorized to perform a specific operation on a
resource. For more details, refer to the normal users topic in
certificate request
for more details about this.
In contrast, service accounts are users managed by the Kubernetes API. They are
bound to specific namespaces, and created automatically by the API server or
manually through API calls. Service accounts are tied to a set of credentials
stored as Secrets, which are mounted into pods allowing in-cluster processes
to talk to the Kubernetes API.
API requests are tied to either a normal user or a service account, or are treated
as anonymous requests. This means every process inside or outside the cluster, from
a human user typing kubectl on a workstation, to kubelets on nodes, to members
of the control plane, must authenticate when making requests to the API server,
or be treated as an anonymous user.
Authentication strategies
Kubernetes uses client certificates, bearer tokens, or an authenticating proxy to
authenticate API requests through authentication plugins. As HTTP requests are
made to the API server, plugins attempt to associate the following attributes
with the request:
Username: a string which identifies the end user. Common values might be kube-admin or jane@example.com.
UID: a string which identifies the end user and attempts to be more consistent and unique than username.
Groups: a set of strings, each of which indicates the user's membership in a named logical collection of users. Common values might be system:masters or devops-team.
Extra fields: a map of strings to list of strings which holds additional information authorizers may find useful.
All values are opaque to the authentication system and only hold significance
when interpreted by an authorizer.
You can enable multiple authentication methods at once. You should usually use at least two methods:
service account tokens for service accounts
at least one other method for user authentication.
When multiple authenticator modules are enabled, the first module
to successfully authenticate the request short-circuits evaluation.
The API server does not guarantee the order authenticators run in.
The system:authenticated group is included in the list of groups for all authenticated users.
Integrations with other authentication protocols (LDAP, SAML, Kerberos, alternate x509 schemes, etc)
can be accomplished using an authenticating proxy or the
authentication webhook.
X509 Client Certs
Client certificate authentication is enabled by passing the --client-ca-file=SOMEFILE
option to API server. The referenced file must contain one or more certificate authorities
to use to validate client certificates presented to the API server. If a client certificate
is presented and verified, the common name of the subject is used as the user name for the
request. As of Kubernetes 1.4, client certificates can also indicate a user's group memberships
using the certificate's organization fields. To include multiple group memberships for a user,
include multiple organization fields in the certificate.
For example, using the openssl command line tool to generate a certificate signing request:
The API server reads bearer tokens from a file when given the --token-auth-file=SOMEFILE option on the command line. Currently, tokens last indefinitely, and the token list cannot be
changed without restarting the API server.
The token file is a csv file with a minimum of 3 columns: token, user name, user uid,
followed by optional group names.
Note:
If you have more than one group the column must be double quoted e.g.
token,user,uid,"group1,group2,group3"
Putting a Bearer Token in a Request
When using bearer token authentication from an http client, the API
server expects an Authorization header with a value of Bearer <token>. The bearer token must be a character sequence that can be
put in an HTTP header value using no more than the encoding and
quoting facilities of HTTP. For example: if the bearer token is
31ada4fd-adec-460c-809a-9e56ceb75269 then it would appear in an HTTP
header as shown below.
To allow for streamlined bootstrapping for new clusters, Kubernetes includes a
dynamically-managed Bearer token type called a Bootstrap Token. These tokens
are stored as Secrets in the kube-system namespace, where they can be
dynamically managed and created. Controller Manager contains a TokenCleaner
controller that deletes bootstrap tokens as they expire.
The tokens are of the form [a-z0-9]{6}.[a-z0-9]{16}. The first component is a
Token ID and the second component is the Token Secret. You specify the token
in an HTTP header as follows:
Authorization: Bearer 781292.db7bc3a58fc5f07e
You must enable the Bootstrap Token Authenticator with the
--enable-bootstrap-token-auth flag on the API Server. You must enable
the TokenCleaner controller via the --controllers flag on the Controller
Manager. This is done with something like --controllers=*,tokencleaner.
kubeadm will do this for you if you are using it to bootstrap a cluster.
The authenticator authenticates as system:bootstrap:<Token ID>. It is
included in the system:bootstrappers group. The naming and groups are
intentionally limited to discourage users from using these tokens past
bootstrapping. The user names and group can be used (and are used by kubeadm)
to craft the appropriate authorization policies to support bootstrapping a
cluster.
Please see Bootstrap Tokens for in depth
documentation on the Bootstrap Token authenticator and controllers along with
how to manage these tokens with kubeadm.
Service Account Tokens
A service account is an automatically enabled authenticator that uses signed
bearer tokens to verify requests. The plugin takes two optional flags:
--service-account-key-file File containing PEM-encoded x509 RSA or ECDSA
private or public keys, used to verify ServiceAccount tokens. The specified file
can contain multiple keys, and the flag can be specified multiple times with
different files. If unspecified, --tls-private-key-file is used.
--service-account-lookup If enabled, tokens which are deleted from the API will be revoked.
Service accounts are usually created automatically by the API server and
associated with pods running in the cluster through the ServiceAccountAdmission Controller. Bearer tokens are
mounted into pods at well-known locations, and allow in-cluster processes to
talk to the API server. Accounts may be explicitly associated with pods using the
serviceAccountName field of a PodSpec.
Note:serviceAccountName is usually omitted because this is done automatically.
apiVersion:apps/v1# this apiVersion is relevant as of Kubernetes 1.9kind:Deploymentmetadata:name:nginx-deploymentnamespace:defaultspec:replicas:3template:metadata:# ...spec:serviceAccountName:bob-the-botcontainers:- name:nginximage:nginx:1.14.2
Service account bearer tokens are perfectly valid to use outside the cluster and
can be used to create identities for long standing jobs that wish to talk to the
Kubernetes API. To manually create a service account, use the kubectl create serviceaccount (NAME) command. This creates a service account in the current
namespace and an associated secret.
The created secret holds the public CA of the API server and a signed JSON Web
Token (JWT).
kubectl get secret jenkins-token-1yvwg -o yaml
apiVersion:v1data:ca.crt:(APISERVER'S CA BASE64 ENCODED)namespace:ZGVmYXVsdA==token:(BEARER TOKEN BASE64 ENCODED)kind:Secretmetadata:# ...type:kubernetes.io/service-account-token
Note: Values are base64 encoded because secrets are always base64 encoded.
The signed JWT can be used as a bearer token to authenticate as the given service
account. See above for how the token is included
in a request. Normally these secrets are mounted into pods for in-cluster access to
the API server, but can be used from outside the cluster as well.
Service accounts authenticate with the username system:serviceaccount:(NAMESPACE):(SERVICEACCOUNT),
and are assigned to the groups system:serviceaccounts and system:serviceaccounts:(NAMESPACE).
WARNING: Because service account tokens are stored in secrets, any user with
read access to those secrets can authenticate as the service account. Be cautious
when granting permissions to service accounts and read capabilities for secrets.
OpenID Connect Tokens
OpenID Connect is a flavor of OAuth2 supported by
some OAuth2 providers, notably Azure Active Directory, Salesforce, and Google.
The protocol's main extension of OAuth2 is an additional field returned with
the access token called an ID Token.
This token is a JSON Web Token (JWT) with well known fields, such as a user's
email, signed by the server.
To identify the user, the authenticator uses the id_token (not the access_token)
from the OAuth2 token response
as a bearer token. See above for how the token
is included in a request.
sequenceDiagram
participant user as User
participant idp as Identity Provider
participant kube as Kubectl
participant api as API Server
user ->> idp: 1. Login to IdP
activate idp
idp -->> user: 2. Provide access_token, id_token, and refresh_token
deactivate idp
activate user
user ->> kube: 3. Call Kubectl with --token being the id_token OR add tokens to .kube/config
deactivate user
activate kube
kube ->> api: 4. Authorization: Bearer...
deactivate kube
activate api
api ->> api: 5. Is JWT signature valid?
api ->> api: 6. Has the JWT expired? (iat+exp)
api ->> api: 7. User authorized?
api -->> kube: 8. Authorized: Perform action and return result
deactivate api
activate kube
kube --x user: 9. Return result
deactivate kube
Login to your identity provider
Your identity provider will provide you with an access_token, id_token and a refresh_token
When using kubectl, use your id_token with the --token flag or add it directly to your kubeconfig
kubectl sends your id_token in a header called Authorization to the API server
The API server will make sure the JWT signature is valid by checking against the certificate named in the configuration
Check to make sure the id_token hasn't expired
Make sure the user is authorized
Once authorized the API server returns a response to kubectl
kubectl provides feedback to the user
Since all of the data needed to validate who you are is in the id_token, Kubernetes doesn't need to
"phone home" to the identity provider. In a model where every request is stateless this provides a very scalable solution for authentication. It does offer a few challenges:
Kubernetes has no "web interface" to trigger the authentication process. There is no browser or interface to collect credentials which is why you need to authenticate to your identity provider first.
The id_token can't be revoked, it's like a certificate so it should be short-lived (only a few minutes) so it can be very annoying to have to get a new token every few minutes.
To authenticate to the Kubernetes dashboard, you must use the kubectl proxy command or a reverse proxy that injects the id_token.
Configuring the API Server
To enable the plugin, configure the following flags on the API server:
Parameter
Description
Example
Required
--oidc-issuer-url
URL of the provider which allows the API server to discover public signing keys. Only URLs which use the https:// scheme are accepted. This is typically the provider's discovery URL without a path, for example "https://accounts.google.com" or "https://login.salesforce.com". This URL should point to the level below .well-known/openid-configuration
If the discovery URL is https://accounts.google.com/.well-known/openid-configuration, the value should be https://accounts.google.com
Yes
--oidc-client-id
A client id that all tokens must be issued for.
kubernetes
Yes
--oidc-username-claim
JWT claim to use as the user name. By default sub, which is expected to be a unique identifier of the end user. Admins can choose other claims, such as email or name, depending on their provider. However, claims other than email will be prefixed with the issuer URL to prevent naming clashes with other plugins.
sub
No
--oidc-username-prefix
Prefix prepended to username claims to prevent clashes with existing names (such as system: users). For example, the value oidc: will create usernames like oidc:jane.doe. If this flag isn't provided and --oidc-username-claim is a value other than email the prefix defaults to ( Issuer URL )# where ( Issuer URL ) is the value of --oidc-issuer-url. The value - can be used to disable all prefixing.
oidc:
No
--oidc-groups-claim
JWT claim to use as the user's group. If the claim is present it must be an array of strings.
groups
No
--oidc-groups-prefix
Prefix prepended to group claims to prevent clashes with existing names (such as system: groups). For example, the value oidc: will create group names like oidc:engineering and oidc:infra.
oidc:
No
--oidc-required-claim
A key=value pair that describes a required claim in the ID Token. If set, the claim is verified to be present in the ID Token with a matching value. Repeat this flag to specify multiple claims.
claim=value
No
--oidc-ca-file
The path to the certificate for the CA that signed your identity provider's web certificate. Defaults to the host's root CAs.
/etc/kubernetes/ssl/kc-ca.pem
No
Importantly, the API server is not an OAuth2 client, rather it can only be
configured to trust a single issuer. This allows the use of public providers,
such as Google, without trusting credentials issued to third parties. Admins who
wish to utilize multiple OAuth clients should explore providers which support the
azp (authorized party) claim, a mechanism for allowing one client to issue
tokens on behalf of another.
Kubernetes does not provide an OpenID Connect Identity Provider.
You can use an existing public OpenID Connect Identity Provider (such as Google, or
others).
Or, you can run your own Identity Provider, such as dex,
Keycloak,
CloudFoundry UAA, or
Tremolo Security's OpenUnison.
For an identity provider to work with Kubernetes it must:
Have a CA signed certificate (even if the CA is not a commercial CA or is self signed)
A note about requirement #3 above, requiring a CA signed certificate. If you deploy your own identity provider (as opposed to one of the cloud providers like Google or Microsoft) you MUST have your identity provider's web server certificate signed by a certificate with the CA flag set to TRUE, even if it is self signed. This is due to GoLang's TLS client implementation being very strict to the standards around certificate validation. If you don't have a CA handy, you can use this script from the Dex team to create a simple CA and a signed certificate and key pair.
Or you can use this similar script that generates SHA256 certs with a longer life and larger key size.
The first option is to use the kubectl oidc authenticator, which sets the id_token as a bearer token for all requests and refreshes the token once it expires. After you've logged into your provider, use kubectl to add your id_token, refresh_token, client_id, and client_secret to configure the plugin.
Providers that don't return an id_token as part of their refresh token response aren't supported by this plugin and should use "Option 2" below.
kubectl config set-credentials USER_NAME \
--auth-provider=oidc \
--auth-provider-arg=idp-issuer-url=( issuer url )\
--auth-provider-arg=client-id=( your client id )\
--auth-provider-arg=client-secret=( your client secret )\
--auth-provider-arg=refresh-token=( your refresh token )\
--auth-provider-arg=idp-certificate-authority=( path to your ca certificate )\
--auth-provider-arg=id-token=( your id_token )
As an example, running the below command after authenticating to your identity provider:
Once your id_token expires, kubectl will attempt to refresh your id_token using your refresh_token and client_secret storing the new values for the refresh_token and id_token in your .kube/config.
Option 2 - Use the --token Option
The kubectl command lets you pass in a token using the --token option. Copy and paste the id_token into this option:
kubectl --token=eyJhbGciOiJSUzI1NiJ9.eyJpc3MiOiJodHRwczovL21sYi50cmVtb2xvLmxhbjo4MDQzL2F1dGgvaWRwL29pZGMiLCJhdWQiOiJrdWJlcm5ldGVzIiwiZXhwIjoxNDc0NTk2NjY5LCJqdGkiOiI2RDUzNXoxUEpFNjJOR3QxaWVyYm9RIiwiaWF0IjoxNDc0NTk2MzY5LCJuYmYiOjE0NzQ1OTYyNDksInN1YiI6Im13aW5kdSIsInVzZXJfcm9sZSI6WyJ1c2VycyIsIm5ldy1uYW1lc3BhY2Utdmlld2VyIl0sImVtYWlsIjoibXdpbmR1QG5vbW9yZWplZGkuY29tIn0.f2As579n9VNoaKzoF-dOQGmXkFKf1FMyNV0-va_B63jn-_n9LGSCca_6IVMP8pO-Zb4KvRqGyTP0r3HkHxYy5c81AnIh8ijarruczl-TK_yF5akjSTHFZD-0gRzlevBDiH8Q79NAr-ky0P4iIXS8lY9Vnjch5MF74Zx0c3alKJHJUnnpjIACByfF2SCaYzbWFMUNat-K1PaUk5-ujMBG7yYnr95xD-63n8CO8teGUAAEMx6zRjzfhnhbzX-ajwZLGwGUBT4WqjMs70-6a7_8gZmLZb2az1cZynkFRj2BaCkVT3A2RrjeEwZEtGXlMqKJ1_I2ulrOVsYx01_yD35-rw get nodes
Webhook Token Authentication
Webhook authentication is a hook for verifying bearer tokens.
--authentication-token-webhook-config-file a configuration file describing how to access the remote webhook service.
--authentication-token-webhook-cache-ttl how long to cache authentication decisions. Defaults to two minutes.
--authentication-token-webhook-version determines whether to use authentication.k8s.io/v1beta1 or authentication.k8s.io/v1TokenReview objects to send/receive information from the webhook. Defaults to v1beta1.
The configuration file uses the kubeconfig
file format. Within the file, clusters refers to the remote service and
users refers to the API server webhook. An example would be:
# Kubernetes API versionapiVersion:v1# kind of the API objectkind:Config# clusters refers to the remote service.clusters:- name:name-of-remote-authn-servicecluster:certificate-authority:/path/to/ca.pem # CA for verifying the remote service.server:https://authn.example.com/authenticate# URL of remote service to query. 'https' recommended for production.# users refers to the API server's webhook configuration.users:- name:name-of-api-serveruser:client-certificate:/path/to/cert.pem# cert for the webhook plugin to useclient-key:/path/to/key.pem # key matching the cert# kubeconfig files require a context. Provide one for the API server.current-context:webhookcontexts:- context:cluster:name-of-remote-authn-serviceuser:name-of-api-servername:webhook
When a client attempts to authenticate with the API server using a bearer token as discussed above,
the authentication webhook POSTs a JSON-serialized TokenReview object containing the token to the remote service.
Note that webhook API objects are subject to the same versioning compatibility rules as other Kubernetes API objects.
Implementers should check the apiVersion field of the request to ensure correct deserialization,
and must respond with a TokenReview object of the same version as the request.
Note: The Kubernetes API server defaults to sending authentication.k8s.io/v1beta1 token reviews for backwards compatibility.
To opt into receiving authentication.k8s.io/v1 token reviews, the API server must be started with --authentication-token-webhook-version=v1.
{"apiVersion": "authentication.k8s.io/v1","kind": "TokenReview","spec": {# Opaque bearer token sent to the API server"token": "014fbff9a07c...",# Optional list of the audience identifiers for the server the token was presented to.# Audience-aware token authenticators (for example, OIDC token authenticators) # should verify the token was intended for at least one of the audiences in this list,# and return the intersection of this list and the valid audiences for the token in the response status.# This ensures the token is valid to authenticate to the server it was presented to.# If no audiences are provided, the token should be validated to authenticate to the Kubernetes API server."audiences": ["https://myserver.example.com","https://myserver.internal.example.com"]}}
{"apiVersion": "authentication.k8s.io/v1beta1","kind": "TokenReview","spec": {# Opaque bearer token sent to the API server"token": "014fbff9a07c...",# Optional list of the audience identifiers for the server the token was presented to.# Audience-aware token authenticators (for example, OIDC token authenticators) # should verify the token was intended for at least one of the audiences in this list,# and return the intersection of this list and the valid audiences for the token in the response status.# This ensures the token is valid to authenticate to the server it was presented to.# If no audiences are provided, the token should be validated to authenticate to the Kubernetes API server."audiences": ["https://myserver.example.com","https://myserver.internal.example.com"]}}
The remote service is expected to fill the status field of the request to indicate the success of the login.
The response body's spec field is ignored and may be omitted.
The remote service must return a response using the same TokenReview API version that it received.
A successful validation of the bearer token would return:
{"apiVersion": "authentication.k8s.io/v1","kind": "TokenReview","status": {"authenticated": true,"user": {# Required"username": "janedoe@example.com",# Optional"uid": "42",# Optional group memberships"groups": ["developers","qa"],# Optional additional information provided by the authenticator.# This should not contain confidential data, as it can be recorded in logs# or API objects, and is made available to admission webhooks."extra": {"extrafield1": ["extravalue1","extravalue2"]}},# Optional list audience-aware token authenticators can return,# containing the audiences from the `spec.audiences` list for which the provided token was valid.# If this is omitted, the token is considered to be valid to authenticate to the Kubernetes API server."audiences": ["https://myserver.example.com"]}}
{"apiVersion": "authentication.k8s.io/v1beta1","kind": "TokenReview","status": {"authenticated": true,"user": {# Required"username": "janedoe@example.com",# Optional"uid": "42",# Optional group memberships"groups": ["developers","qa"],# Optional additional information provided by the authenticator.# This should not contain confidential data, as it can be recorded in logs# or API objects, and is made available to admission webhooks."extra": {"extrafield1": ["extravalue1","extravalue2"]}},# Optional list audience-aware token authenticators can return,# containing the audiences from the `spec.audiences` list for which the provided token was valid.# If this is omitted, the token is considered to be valid to authenticate to the Kubernetes API server."audiences": ["https://myserver.example.com"]}}
{"apiVersion": "authentication.k8s.io/v1","kind": "TokenReview","status": {"authenticated": false,# Optionally include details about why authentication failed.# If no error is provided, the API will return a generic Unauthorized message.# The error field is ignored when authenticated=true."error": "Credentials are expired"}}
{"apiVersion": "authentication.k8s.io/v1beta1","kind": "TokenReview","status": {"authenticated": false,# Optionally include details about why authentication failed.# If no error is provided, the API will return a generic Unauthorized message.# The error field is ignored when authenticated=true."error": "Credentials are expired"}}
Authenticating Proxy
The API server can be configured to identify users from request header values, such as X-Remote-User.
It is designed for use in combination with an authenticating proxy, which sets the request header value.
--requestheader-username-headers Required, case-insensitive. Header names to check, in order, for the user identity. The first header containing a value is used as the username.
--requestheader-group-headers 1.6+. Optional, case-insensitive. "X-Remote-Group" is suggested. Header names to check, in order, for the user's groups. All values in all specified headers are used as group names.
--requestheader-extra-headers-prefix 1.6+. Optional, case-insensitive. "X-Remote-Extra-" is suggested. Header prefixes to look for to determine extra information about the user (typically used by the configured authorization plugin). Any headers beginning with any of the specified prefixes have the prefix removed. The remainder of the header name is lowercased and percent-decoded and becomes the extra key, and the header value is the extra value.
Note: Prior to 1.11.3 (and 1.10.7, 1.9.11), the extra key could only contain characters which were legal in HTTP header labels.
In order to prevent header spoofing, the authenticating proxy is required to present a valid client
certificate to the API server for validation against the specified CA before the request headers are
checked. WARNING: do not reuse a CA that is used in a different context unless you understand
the risks and the mechanisms to protect the CA's usage.
--requestheader-client-ca-file Required. PEM-encoded certificate bundle. A valid client certificate must be presented and validated against the certificate authorities in the specified file before the request headers are checked for user names.
--requestheader-allowed-names Optional. List of Common Name values (CNs). If set, a valid client certificate with a CN in the specified list must be presented before the request headers are checked for user names. If empty, any CN is allowed.
Anonymous requests
When enabled, requests that are not rejected by other configured authentication methods are
treated as anonymous requests, and given a username of system:anonymous and a group of
system:unauthenticated.
For example, on a server with token authentication configured, and anonymous access enabled,
a request providing an invalid bearer token would receive a 401 Unauthorized error.
A request providing no bearer token would be treated as an anonymous request.
In 1.5.1-1.5.x, anonymous access is disabled by default, and can be enabled by
passing the --anonymous-auth=true option to the API server.
In 1.6+, anonymous access is enabled by default if an authorization mode other than AlwaysAllow
is used, and can be disabled by passing the --anonymous-auth=false option to the API server.
Starting in 1.6, the ABAC and RBAC authorizers require explicit authorization of the
system:anonymous user or the system:unauthenticated group, so legacy policy rules
that grant access to the * user or * group do not include anonymous users.
User impersonation
A user can act as another user through impersonation headers. These let requests
manually override the user info a request authenticates as. For example, an admin
could use this feature to debug an authorization policy by temporarily
impersonating another user and seeing if a request was denied.
Impersonation requests first authenticate as the requesting user, then switch
to the impersonated user info.
A user makes an API call with their credentials and impersonation headers.
API server authenticates the user.
API server ensures the authenticated users have impersonation privileges.
Request user info is replaced with impersonation values.
Request is evaluated, authorization acts on impersonated user info.
The following HTTP headers can be used to performing an impersonation request:
Impersonate-User: The username to act as.
Impersonate-Group: A group name to act as. Can be provided multiple times to set multiple groups. Optional. Requires "Impersonate-User".
Impersonate-Extra-( extra name ): A dynamic header used to associate extra fields with the user. Optional. Requires "Impersonate-User". In order to be preserved consistently, ( extra name ) must be lower-case, and any characters which aren't legal in HTTP header labels MUST be utf8 and percent-encoded.
Impersonate-Uid: A unique identifier that represents the user being impersonated. Optional. Requires "Impersonate-User". Kubernetes does not impose any format requirements on this string.
Note: Prior to 1.11.3 (and 1.10.7, 1.9.11), ( extra name ) could only contain characters which were legal in HTTP header labels.
Note:Impersonate-Uid is only available in versions 1.22.0 and higher.
An example of the impersonation headers used when impersonating a user with groups:
Note:kubectl cannot impersonate extra fields or UIDs.
To impersonate a user, group, user identifier (UID) or extra fields, the impersonating user must
have the ability to perform the "impersonate" verb on the kind of attribute
being impersonated ("user", "group", "uid", etc.). For clusters that enable the RBAC
authorization plugin, the following ClusterRole encompasses the rules needed to
set user and group impersonation headers:
For impersonation, extra fields and impersonated UIDs are both under the "authentication.k8s.io" apiGroup.
Extra fields are evaluated as sub-resources of the resource "userextras". To
allow a user to use impersonation headers for the extra field "scopes" and
for UIDs, a user should be granted the following role:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:scopes-and-uid-impersonatorrules:# Can set "Impersonate-Extra-scopes" header and the "Impersonate-Uid" header.- apiGroups:["authentication.k8s.io"]resources:["userextras/scopes","uids"]verbs:["impersonate"]
The values of impersonation headers can also be restricted by limiting the set
of resourceNames a resource can take.
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:limited-impersonatorrules:# Can impersonate the user "jane.doe@example.com"- apiGroups:[""]resources:["users"]verbs:["impersonate"]resourceNames:["jane.doe@example.com"]# Can impersonate the groups "developers" and "admins"- apiGroups:[""]resources:["groups"]verbs:["impersonate"]resourceNames:["developers","admins"]# Can impersonate the extras field "scopes" with the values "view" and "development"- apiGroups:["authentication.k8s.io"]resources:["userextras/scopes"]verbs:["impersonate"]resourceNames:["view","development"]# Can impersonate the uid "06f6ce97-e2c5-4ab8-7ba5-7654dd08d52b"- apiGroups:["authentication.k8s.io"]resources:["uids"]verbs:["impersonate"]resourceNames:["06f6ce97-e2c5-4ab8-7ba5-7654dd08d52b"]
Note: Impersonating a user or group allows you to perform any action as if you were that user or group;
for that reason, impersonation is not namespace scoped.
If you want to allow impersonation using Kubernetes RBAC,
this requires using a ClusterRole and a ClusterRoleBinding,
not a Role and RoleBinding.
client-go credential plugins
FEATURE STATE:Kubernetes v1.22 [stable]
k8s.io/client-go and tools using it such as kubectl and kubelet are able to execute an
external command to receive user credentials.
This feature is intended for client side integrations with authentication protocols not natively
supported by k8s.io/client-go (LDAP, Kerberos, OAuth2, SAML, etc.). The plugin implements the
protocol specific logic, then returns opaque credentials to use. Almost all credential plugin
use cases require a server side component with support for the webhook token authenticator
to interpret the credential format produced by the client plugin.
Example use case
In a hypothetical use case, an organization would run an external service that exchanges LDAP credentials
for user specific, signed tokens. The service would also be capable of responding to webhook token
authenticator requests to validate the tokens. Users would be required
to install a credential plugin on their workstation.
To authenticate against the API:
The user issues a kubectl command.
Credential plugin prompts the user for LDAP credentials, exchanges credentials with external service for a token.
Credential plugin returns token to client-go, which uses it as a bearer token against the API server.
apiVersion:v1kind:Configusers:- name:my-useruser:exec:# Command to execute. Required.command:"example-client-go-exec-plugin"# API version to use when decoding the ExecCredentials resource. Required.## The API version returned by the plugin MUST match the version listed here.## To integrate with tools that support multiple versions (such as client.authentication.k8s.io/v1beta1),# set an environment variable, pass an argument to the tool that indicates which version the exec plugin expects,# or read the version from the ExecCredential object in the KUBERNETES_EXEC_INFO environment variable.apiVersion:"client.authentication.k8s.io/v1"# Environment variables to set when executing the plugin. Optional.env:- name:"FOO"value:"bar"# Arguments to pass when executing the plugin. Optional.args:- "arg1"- "arg2"# Text shown to the user when the executable doesn't seem to be present. Optional.installHint:| example-client-go-exec-plugin is required to authenticate
to the current cluster. It can be installed:
On macOS: brew install example-client-go-exec-plugin
On Ubuntu: apt-get install example-client-go-exec-plugin
On Fedora: dnf install example-client-go-exec-plugin
...# Whether or not to provide cluster information, which could potentially contain# very large CA data, to this exec plugin as a part of the KUBERNETES_EXEC_INFO# environment variable.provideClusterInfo:true# The contract between the exec plugin and the standard input I/O stream. If the# contract cannot be satisfied, this plugin will not be run and an error will be# returned. Valid values are "Never" (this exec plugin never uses standard input),# "IfAvailable" (this exec plugin wants to use standard input if it is available),# or "Always" (this exec plugin requires standard input to function). Required.interactiveMode:Neverclusters:- name:my-clustercluster:server:"https://172.17.4.100:6443"certificate-authority:"/etc/kubernetes/ca.pem"extensions:- name:client.authentication.k8s.io/exec# reserved extension name for per cluster exec configextension:arbitrary:configthis:can be provided via the KUBERNETES_EXEC_INFO environment variable upon setting provideClusterInfoyou:["can","put","anything","here"]contexts:- name:my-clustercontext:cluster:my-clusteruser:my-usercurrent-context:my-cluster
apiVersion:v1kind:Configusers:- name:my-useruser:exec:# Command to execute. Required.command:"example-client-go-exec-plugin"# API version to use when decoding the ExecCredentials resource. Required.## The API version returned by the plugin MUST match the version listed here.## To integrate with tools that support multiple versions (such as client.authentication.k8s.io/v1),# set an environment variable, pass an argument to the tool that indicates which version the exec plugin expects,# or read the version from the ExecCredential object in the KUBERNETES_EXEC_INFO environment variable.apiVersion:"client.authentication.k8s.io/v1beta1"# Environment variables to set when executing the plugin. Optional.env:- name:"FOO"value:"bar"# Arguments to pass when executing the plugin. Optional.args:- "arg1"- "arg2"# Text shown to the user when the executable doesn't seem to be present. Optional.installHint:| example-client-go-exec-plugin is required to authenticate
to the current cluster. It can be installed:
On macOS: brew install example-client-go-exec-plugin
On Ubuntu: apt-get install example-client-go-exec-plugin
On Fedora: dnf install example-client-go-exec-plugin
...# Whether or not to provide cluster information, which could potentially contain# very large CA data, to this exec plugin as a part of the KUBERNETES_EXEC_INFO# environment variable.provideClusterInfo:true# The contract between the exec plugin and the standard input I/O stream. If the# contract cannot be satisfied, this plugin will not be run and an error will be# returned. Valid values are "Never" (this exec plugin never uses standard input),# "IfAvailable" (this exec plugin wants to use standard input if it is available),# or "Always" (this exec plugin requires standard input to function). Optional.# Defaults to "IfAvailable".interactiveMode:Neverclusters:- name:my-clustercluster:server:"https://172.17.4.100:6443"certificate-authority:"/etc/kubernetes/ca.pem"extensions:- name:client.authentication.k8s.io/exec# reserved extension name for per cluster exec configextension:arbitrary:configthis:can be provided via the KUBERNETES_EXEC_INFO environment variable upon setting provideClusterInfoyou:["can","put","anything","here"]contexts:- name:my-clustercontext:cluster:my-clusteruser:my-usercurrent-context:my-cluster
Relative command paths are interpreted as relative to the directory of the config file. If
KUBECONFIG is set to /home/jane/kubeconfig and the exec command is ./bin/example-client-go-exec-plugin,
the binary /home/jane/bin/example-client-go-exec-plugin is executed.
- name:my-useruser:exec:# Path relative to the directory of the kubeconfigcommand:"./bin/example-client-go-exec-plugin"apiVersion:"client.authentication.k8s.io/v1"interactiveMode:Never
Input and output formats
The executed command prints an ExecCredential object to stdout. k8s.io/client-go
authenticates against the Kubernetes API using the returned credentials in the status.
The executed command is passed an ExecCredential object as input via the KUBERNETES_EXEC_INFO
environment variable. This input contains helpful information like the expected API version
of the returned ExecCredential object and whether or not the plugin can use stdin to interact
with the user.
When run from an interactive session (i.e., a terminal), stdin can be exposed directly
to the plugin. Plugins should use the spec.interactive field of the input
ExecCredential object from the KUBERNETES_EXEC_INFO environment variable in order to
determine if stdin has been provided. A plugin's stdin requirements (i.e., whether
stdin is optional, strictly required, or never used in order for the plugin
to run successfully) is declared via the user.exec.interactiveMode field in the
kubeconfig (see table
below for valid values). The user.exec.interactiveMode field is optional in client.authentication.k8s.io/v1beta1
and required in client.authentication.k8s.io/v1.
interactiveMode values
interactiveMode Value
Meaning
Never
This exec plugin never needs to use standard input, and therefore the exec plugin will be run regardless of whether standard input is available for user input.
IfAvailable
This exec plugin would like to use standard input if it is available, but can still operate if standard input is not available. Therefore, the exec plugin will be run regardless of whether stdin is available for user input. If standard input is available for user input, then it will be provided to this exec plugin.
Always
This exec plugin requires standard input in order to run, and therefore the exec plugin will only be run if standard input is available for user input. If standard input is not available for user input, then the exec plugin will not be run and an error will be returned by the exec plugin runner.
To use bearer token credentials, the plugin returns a token in the status of the
ExecCredential
Alternatively, a PEM-encoded client certificate and key can be returned to use TLS client auth.
If the plugin returns a different certificate and key on a subsequent call, k8s.io/client-go
will close existing connections with the server to force a new TLS handshake.
If specified, clientKeyData and clientCertificateData must both must be present.
clientCertificateData may contain additional intermediate certificates to send to the server.
Optionally, the response can include the expiry of the credential formatted as a
RFC 3339 timestamp.
Presence or absence of an expiry has the following impact:
If an expiry is included, the bearer token and TLS credentials are cached until
the expiry time is reached, or if the server responds with a 401 HTTP status code,
or when the process exits.
If an expiry is omitted, the bearer token and TLS credentials are cached until
the server responds with a 401 HTTP status code or until the process exits.
To enable the exec plugin to obtain cluster-specific information, set provideClusterInfo on the user.exec
field in the kubeconfig.
The plugin will then be supplied this cluster-specific information in the KUBERNETES_EXEC_INFO environment variable.
Information from this environment variable can be used to perform cluster-specific
credential acquisition logic.
The following ExecCredential manifest describes a cluster information sample.
Bootstrap tokens are a simple bearer token that is meant to be used when
creating new clusters or joining new nodes to an existing cluster. It was built
to support kubeadm, but can be used in other contexts
for users that wish to start clusters without kubeadm. It is also built to
work, via RBAC policy, with the
Kubelet TLS Bootstrapping system.
Bootstrap Tokens Overview
Bootstrap Tokens are defined with a specific type
(bootstrap.kubernetes.io/token) of secrets that lives in the kube-system
namespace. These Secrets are then read by the Bootstrap Authenticator in the
API Server. Expired tokens are removed with the TokenCleaner controller in the
Controller Manager. The tokens are also used to create a signature for a
specific ConfigMap used in a "discovery" process through a BootstrapSigner
controller.
Token Format
Bootstrap Tokens take the form of abcdef.0123456789abcdef. More formally,
they must match the regular expression [a-z0-9]{6}\.[a-z0-9]{16}.
The first part of the token is the "Token ID" and is considered public
information. It is used when referring to a token without leaking the secret
part used for authentication. The second part is the "Token Secret" and should
only be shared with trusted parties.
Enabling Bootstrap Token Authentication
The Bootstrap Token authenticator can be enabled using the following flag on the
API server:
--enable-bootstrap-token-auth
When enabled, bootstrapping tokens can be used as bearer token credentials to
authenticate requests against the API server.
Authorization: Bearer 07401b.f395accd246ae52d
Tokens authenticate as the username system:bootstrap:<token id> and are members
of the group system:bootstrappers. Additional groups may be specified in the
token's Secret.
Expired tokens can be deleted automatically by enabling the tokencleaner
controller on the controller manager.
--controllers=*,tokencleaner
Bootstrap Token Secret Format
Each valid token is backed by a secret in the kube-system namespace. You can
find the full design doc
here.
Here is what the secret looks like.
apiVersion:v1kind:Secretmetadata:# Name MUST be of form "bootstrap-token-<token id>"name:bootstrap-token-07401bnamespace:kube-system# Type MUST be 'bootstrap.kubernetes.io/token'type:bootstrap.kubernetes.io/tokenstringData:# Human readable description. Optional.description:"The default bootstrap token generated by 'kubeadm init'."# Token ID and secret. Required.token-id:07401btoken-secret:f395accd246ae52d# Expiration. Optional.expiration:2017-03-10T03:22:11Z# Allowed usages.usage-bootstrap-authentication:"true"usage-bootstrap-signing:"true"# Extra groups to authenticate the token as. Must start with "system:bootstrappers:"auth-extra-groups:system:bootstrappers:worker,system:bootstrappers:ingress
The type of the secret must be bootstrap.kubernetes.io/token and the name must
be bootstrap-token-<token id>. It must also exist in the kube-system
namespace.
The usage-bootstrap-* members indicate what this secret is intended to be used
for. A value must be set to true to be enabled.
usage-bootstrap-authentication indicates that the token can be used to
authenticate to the API server as a bearer token.
usage-bootstrap-signing indicates that the token may be used to sign the
cluster-info ConfigMap as described below.
The expiration field controls the expiry of the token. Expired tokens are
rejected when used for authentication and ignored during ConfigMap signing.
The expiry value is encoded as an absolute UTC time using RFC3339. Enable the
tokencleaner controller to automatically delete expired tokens.
Token Management with kubeadm
You can use the kubeadm tool to manage tokens on a running cluster. See the
kubeadm token docs for details.
ConfigMap Signing
In addition to authentication, the tokens can be used to sign a ConfigMap. This
is used early in a cluster bootstrap process before the client trusts the API
server. The signed ConfigMap can be authenticated by the shared token.
Enable ConfigMap signing by enabling the bootstrapsigner controller on the
Controller Manager.
--controllers=*,bootstrapsigner
The ConfigMap that is signed is cluster-info in the kube-public namespace.
The typical flow is that a client reads this ConfigMap while unauthenticated and
ignoring TLS errors. It then validates the payload of the ConfigMap by looking
at a signature embedded in the ConfigMap.
The kubeconfig member of the ConfigMap is a config file with only the cluster
information filled out. The key thing being communicated here is the
certificate-authority-data. This may be expanded in the future.
The signature is a JWS signature using the "detached" mode. To validate the
signature, the user should encode the kubeconfig payload according to JWS
rules (base64 encoded while discarding any trailing =). That encoded payload
is then used to form a whole JWS by inserting it between the 2 dots. You can
verify the JWS using the HS256 scheme (HMAC-SHA256) with the full token (e.g.
07401b.f395accd246ae52d) as the shared secret. Users must verify that HS256
is used.
Warning: Any party with a bootstrapping token can create a valid signature for that
token. When using ConfigMap signing it's discouraged to share the same token with
many clients, since a compromised client can potentially man-in-the middle another
client relying on the signature to bootstrap TLS trust.
The Certificates API enables automation of
X.509 credential provisioning by providing
a programmatic interface for clients of the Kubernetes API to request and obtain
X.509 certificates from a Certificate Authority (CA).
A CertificateSigningRequest (CSR) resource is used to request that a certificate be signed
by a denoted signer, after which the request may be approved or denied before
finally being signed.
Request signing process
The CertificateSigningRequest resource type allows a client to ask for an X.509 certificate
be issued, based on a signing request.
The CertificateSigningRequest object includes a PEM-encoded PKCS#10 signing request in
the spec.request field. The CertificateSigningRequest denotes the signer (the
recipient that the request is being made to) using the spec.signerName field.
Note that spec.signerName is a required key after API version certificates.k8s.io/v1.
In Kubernetes v1.22 and later, clients may optionally set the spec.expirationSeconds
field to request a particular lifetime for the issued certificate. The minimum valid
value for this field is 600, i.e. ten minutes.
Once created, a CertificateSigningRequest must be approved before it can be signed.
Depending on the signer selected, a CertificateSigningRequest may be automatically approved
by a controller.
Otherwise, a CertificateSigningRequest must be manually approved either via the REST API (or client-go)
or by running kubectl certificate approve. Likewise, a CertificateSigningRequest may also be denied,
which tells the configured signer that it must not sign the request.
For certificates that have been approved, the next step is signing. The relevant signing controller
first validates that the signing conditions are met and then creates a certificate.
The signing controller then updates the CertificateSigningRequest, storing the new certificate into
the status.certificate field of the existing CertificateSigningRequest object. The
status.certificate field is either empty or contains a X.509 certificate, encoded in PEM format.
The CertificateSigningRequest status.certificate field is empty until the signer does this.
Once the status.certificate field has been populated, the request has been completed and clients can now
fetch the signed certificate PEM data from the CertificateSigningRequest resource.
The signers can instead deny certificate signing if the approval conditions are not met.
In order to reduce the number of old CertificateSigningRequest resources left in a cluster, a garbage collection
controller runs periodically. The garbage collection removes CertificateSigningRequests that have not changed
state for some duration:
Approved requests: automatically deleted after 1 hour
Denied requests: automatically deleted after 1 hour
Failed requests: automatically deleted after 1 hour
Pending requests: automatically deleted after 24 hours
All requests: automatically deleted after the issued certificate has expired
Signers
Custom signerNames can also be specified. All signers should provide information about how they work so that clients can predict what will happen to their CSRs.
This includes:
Trust distribution: how trust (CA bundles) are distributed.
Permitted subjects: any restrictions on and behavior when a disallowed subject is requested.
Permitted x509 extensions: including IP subjectAltNames, DNS subjectAltNames, Email subjectAltNames, URI subjectAltNames etc, and behavior when a disallowed extension is requested.
Permitted key usages / extended key usages: any restrictions on and behavior when usages different than the signer-determined usages are specified in the CSR.
Expiration/certificate lifetime: whether it is fixed by the signer, configurable by the admin, determined by the CSR spec.expirationSeconds field, etc
and the behavior when the signer-determined expiration is different from the CSR spec.expirationSeconds field.
CA bit allowed/disallowed: and behavior if a CSR contains a request a for a CA certificate when the signer does not permit it.
Commonly, the status.certificate field contains a single PEM-encoded X.509
certificate once the CSR is approved and the certificate is issued. Some
signers store multiple certificates into the status.certificate field. In
that case, the documentation for the signer should specify the meaning of
additional certificates; for example, this might be the certificate plus
intermediates to be presented during TLS handshakes.
The PKCS#10 signing request format does not have a standard mechanism to specify a
certificate expiration or lifetime. The expiration or lifetime therefore has to be set
through the spec.expirationSeconds field of the CSR object. The built-in signers
use the ClusterSigningDuration configuration option, which defaults to 1 year,
(the --cluster-signing-duration command-line flag of the kube-controller-manager)
as the default when no spec.expirationSeconds is specified. When spec.expirationSeconds
is specified, the minimum of spec.expirationSeconds and ClusterSigningDuration is
used.
Note: The spec.expirationSeconds field was added in Kubernetes v1.22. Earlier versions of Kubernetes do not honor this field.
Kubernetes API servers prior to v1.22 will silently drop this field when the object is created.
Kubernetes signers
Kubernetes provides built-in signers that each have a well-known signerName:
kubernetes.io/kube-apiserver-client: signs certificates that will be honored as client certificates by the API server.
Never auto-approved by kube-controller-manager.
Trust distribution: signed certificates must be honored as client certificates by the API server. The CA bundle is not distributed by any other means.
Permitted subjects - no subject restrictions, but approvers and signers may choose not to approve or sign.
Certain subjects like cluster-admin level users or groups vary between distributions and installations,
but deserve additional scrutiny before approval and signing.
The CertificateSubjectRestriction admission plugin is enabled by default to restrict system:masters,
but it is often not the only cluster-admin subject in a cluster.
Permitted x509 extensions - honors subjectAltName and key usage extensions and discards other extensions.
Permitted key usages - must include ["client auth"]. Must not include key usages beyond ["digital signature", "key encipherment", "client auth"].
Expiration/certificate lifetime - for the kube-controller-manager implementation of this signer, set to the minimum
of the --cluster-signing-duration option or, if specified, the spec.expirationSeconds field of the CSR object.
CA bit allowed/disallowed - not allowed.
kubernetes.io/kube-apiserver-client-kubelet: signs client certificates that will be honored as client certificates by the
API server.
May be auto-approved by kube-controller-manager.
Trust distribution: signed certificates must be honored as client certificates by the API server. The CA bundle
is not distributed by any other means.
Permitted subjects - organizations are exactly ["system:nodes"], common name starts with "system:node:".
Permitted x509 extensions - honors key usage extensions, forbids subjectAltName extensions and drops other extensions.
Expiration/certificate lifetime - for the kube-controller-manager implementation of this signer, set to the minimum
of the --cluster-signing-duration option or, if specified, the spec.expirationSeconds field of the CSR object.
CA bit allowed/disallowed - not allowed.
kubernetes.io/kubelet-serving: signs serving certificates that are honored as a valid kubelet serving certificate
by the API server, but has no other guarantees.
Never auto-approved by kube-controller-manager.
Trust distribution: signed certificates must be honored by the API server as valid to terminate connections to a kubelet.
The CA bundle is not distributed by any other means.
Permitted subjects - organizations are exactly ["system:nodes"], common name starts with "system:node:".
Permitted x509 extensions - honors key usage and DNSName/IPAddress subjectAltName extensions, forbids EmailAddress and
URI subjectAltName extensions, drops other extensions. At least one DNS or IP subjectAltName must be present.
Expiration/certificate lifetime - for the kube-controller-manager implementation of this signer, set to the minimum
of the --cluster-signing-duration option or, if specified, the spec.expirationSeconds field of the CSR object.
CA bit allowed/disallowed - not allowed.
kubernetes.io/legacy-unknown: has no guarantees for trust at all. Some third-party distributions of Kubernetes
may honor client certificates signed by it. The stable CertificateSigningRequest API (version certificates.k8s.io/v1 and later)
does not allow to set the signerName as kubernetes.io/legacy-unknown.
Never auto-approved by kube-controller-manager.
Trust distribution: None. There is no standard trust or distribution for this signer in a Kubernetes cluster.
Permitted subjects - any
Permitted x509 extensions - honors subjectAltName and key usage extensions and discards other extensions.
Permitted key usages - any
Expiration/certificate lifetime - for the kube-controller-manager implementation of this signer, set to the minimum
of the --cluster-signing-duration option or, if specified, the spec.expirationSeconds field of the CSR object.
CA bit allowed/disallowed - not allowed.
Note: Failures for all of these are only reported in kube-controller-manager logs.
Note: The spec.expirationSeconds field was added in Kubernetes v1.22. Earlier versions of Kubernetes do not honor this field.
Kubernetes API servers prior to v1.22 will silently drop this field when the object is created.
Distribution of trust happens out of band for these signers. Any trust outside of those described above are strictly
coincidental. For instance, some distributions may honor kubernetes.io/legacy-unknown as client certificates for the
kube-apiserver, but this is not a standard.
None of these usages are related to ServiceAccount token secrets .data[ca.crt] in any way. That CA bundle is only
guaranteed to verify a connection to the API server using the default service (kubernetes.default.svc).
Authorization
To allow creating a CertificateSigningRequest and retrieving any CertificateSigningRequest:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:csr-approverrules:- apiGroups:- certificates.k8s.ioresources:- certificatesigningrequestsverbs:- get- list- watch- apiGroups:- certificates.k8s.ioresources:- certificatesigningrequests/approvalverbs:- update- apiGroups:- certificates.k8s.ioresources:- signersresourceNames:- example.com/my-signer-name# example.com/* can be used to authorize for all signers in the 'example.com' domainverbs:- approve
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:csr-signerrules:- apiGroups:- certificates.k8s.ioresources:- certificatesigningrequestsverbs:- get- list- watch- apiGroups:- certificates.k8s.ioresources:- certificatesigningrequests/statusverbs:- update- apiGroups:- certificates.k8s.ioresources:- signersresourceNames:- example.com/my-signer-name# example.com/* can be used to authorize for all signers in the 'example.com' domainverbs:- sign
Normal user
A few steps are required in order to get a normal user to be able to
authenticate and invoke an API. First, this user must have a certificate issued
by the Kubernetes cluster, and then present that certificate to the Kubernetes API.
Create private key
The following scripts show how to generate PKI private key and CSR. It is
important to set CN and O attribute of the CSR. CN is the name of the user and
O is the group that this user will belong to. You can refer to
RBAC for standard groups.
The kube-controller-manager ships with a built-in approver for certificates with
a signerName of kubernetes.io/kube-apiserver-client-kubelet that delegates various
permissions on CSRs for node credentials to authorization.
The kube-controller-manager POSTs SubjectAccessReview resources to the API server
in order to check authorization for certificate approval.
Approval or rejection using kubectl
A Kubernetes administrator (with appropriate permissions) can manually approve
(or deny) CertificateSigningRequests by using the kubectl certificate approve and kubectl certificate deny commands.
Users of the REST API can approve CSRs by submitting an UPDATE request to the approval
subresource of the CSR to be approved. For example, you could write an
operator that watches for a particular
kind of CSR and then sends an UPDATE to approve them.
When you make an approval or rejection request, set either the Approved or Denied
status condition based on the state you determine:
For Approved CSRs:
apiVersion:certificates.k8s.io/v1kind:CertificateSigningRequest...status:conditions:- lastUpdateTime:"2020-02-08T11:37:35Z"lastTransitionTime:"2020-02-08T11:37:35Z"message:Approved by my custom approver controllerreason:ApprovedByMyPolicy# You can set this to any stringtype:Approved
For Denied CSRs:
apiVersion:certificates.k8s.io/v1kind:CertificateSigningRequest...status:conditions:- lastUpdateTime:"2020-02-08T11:37:35Z"lastTransitionTime:"2020-02-08T11:37:35Z"message:Denied by my custom approver controllerreason:DeniedByMyPolicy# You can set this to any stringtype:Denied
It's usual to set status.conditions.reason to a machine-friendly reason
code using TitleCase; this is a convention but you can set it to anything
you like. If you want to add a note for human consumption, use the
status.conditions.message field.
Signing
Control plane signer
The Kubernetes control plane implements each of the
Kubernetes signers,
as part of the kube-controller-manager.
Note: Prior to Kubernetes v1.18, the kube-controller-manager would sign any CSRs that
were marked as approved.
Note: The spec.expirationSeconds field was added in Kubernetes v1.22. Earlier versions of Kubernetes do not honor this field.
Kubernetes API servers prior to v1.22 will silently drop this field when the object is created.
API-based signers
Users of the REST API can sign CSRs by submitting an UPDATE request to the status
subresource of the CSR to be signed.
As part of this request, the status.certificate field should be set to contain the
signed certificate. This field contains one or more PEM-encoded certificates.
All PEM blocks must have the "CERTIFICATE" label, contain no headers,
and the encoded data must be a BER-encoded ASN.1 Certificate structure
as described in section 4 of RFC5280.
Non-PEM content may appear before or after the CERTIFICATE PEM blocks and is unvalidated,
to allow for explanatory text as described in section 5.2 of RFC7468.
When encoded in JSON or YAML, this field is base-64 encoded.
A CertificateSigningRequest containing the example certificate above would look like this:
View the source code for the kube-controller-manager built in signer
View the source code for the kube-controller-manager built in approver
For details of X.509 itself, refer to RFC 5280 section 3.1
For information on the syntax of PKCS#10 certificate signing requests, refer to RFC 2986
6.3.4 - Using Admission Controllers
This page provides an overview of Admission Controllers.
What are they?
An admission controller is a piece of code that intercepts requests to the
Kubernetes API server prior to persistence of the object, but after the request
is authenticated and authorized. The controllers consist of the
list below, are compiled into the
kube-apiserver binary, and may only be configured by the cluster
administrator. In that list, there are two special controllers:
MutatingAdmissionWebhook and ValidatingAdmissionWebhook. These execute the
mutating and validating (respectively)
admission control webhooks
which are configured in the API.
Admission controllers may be "validating", "mutating", or both. Mutating
controllers may modify related objects to the requests they admit; validating controllers may not.
Admission controllers limit requests to create, delete, modify objects or connect to proxy. They do not limit requests to read objects.
The admission control process proceeds in two phases. In the first phase,
mutating admission controllers are run. In the second phase, validating
admission controllers are run. Note again that some of the controllers are
both.
If any of the controllers in either phase reject the request, the entire
request is rejected immediately and an error is returned to the end-user.
Finally, in addition to sometimes mutating the object in question, admission
controllers may sometimes have side effects, that is, mutate related
resources as part of request processing. Incrementing quota usage is the
canonical example of why this is necessary. Any such side-effect needs a
corresponding reclamation or reconciliation process, as a given admission
controller does not know for sure that a given request will pass all of the
other admission controllers.
Why do I need them?
Many advanced features in Kubernetes require an admission controller to be enabled in order
to properly support the feature. As a result, a Kubernetes API server that is not properly
configured with the right set of admission controllers is an incomplete server and will not
support all the features you expect.
How do I turn on an admission controller?
The Kubernetes API server flag enable-admission-plugins takes a comma-delimited list of admission control plugins to invoke prior to modifying objects in the cluster.
For example, the following command line enables the NamespaceLifecycle and the LimitRanger
admission control plugins:
Note: Depending on the way your Kubernetes cluster is deployed and how the API server is
started, you may need to apply the settings in different ways. For example, you may
have to modify the systemd unit file if the API server is deployed as a systemd
service, you may modify the manifest file for the API server if Kubernetes is deployed
in a self-hosted way.
How do I turn off an admission controller?
The Kubernetes API server flag disable-admission-plugins takes a comma-delimited list of admission control plugins to be disabled, even if they are in the list of plugins enabled by default.
This admission controller allows all pods into the cluster. It is deprecated because
its behavior is the same as if there were no admission controller at all.
AlwaysDeny
FEATURE STATE:Kubernetes v1.13 [deprecated]
Rejects all requests. AlwaysDeny is DEPRECATED as it has no real meaning.
AlwaysPullImages
This admission controller modifies every new Pod to force the image pull policy to Always. This is useful in a
multitenant cluster so that users can be assured that their private images can only be used by those
who have the credentials to pull them. Without this admission controller, once an image has been pulled to a
node, any pod from any user can use it by knowing the image's name (assuming the Pod is
scheduled onto the right node), without any authorization check against the image. When this admission controller
is enabled, images are always pulled prior to starting containers, which means valid credentials are
required.
CertificateApproval
This admission controller observes requests to 'approve' CertificateSigningRequest resources and performs additional
authorization checks to ensure the approving user has permission to approve certificate requests with the
spec.signerName requested on the CertificateSigningRequest resource.
See Certificate Signing Requests for more
information on the permissions required to perform different actions on CertificateSigningRequest resources.
CertificateSigning
This admission controller observes updates to the status.certificate field of CertificateSigningRequest resources
and performs an additional authorization checks to ensure the signing user has permission to sign certificate
requests with the spec.signerName requested on the CertificateSigningRequest resource.
See Certificate Signing Requests for more
information on the permissions required to perform different actions on CertificateSigningRequest resources.
CertificateSubjectRestriction
This admission controller observes creation of CertificateSigningRequest resources that have a spec.signerName
of kubernetes.io/kube-apiserver-client. It rejects any request that specifies a 'group' (or 'organization attribute')
of system:masters.
DefaultIngressClass
This admission controller observes creation of Ingress objects that do not request any specific
ingress class and automatically adds a default ingress class to them. This way, users that do not
request any special ingress class do not need to care about them at all and they will get the
default one.
This admission controller does not do anything when no default ingress class is configured. When more than one ingress
class is marked as default, it rejects any creation of Ingress with an error and an administrator
must revisit their IngressClass objects and mark only one as default (with the annotation
"ingressclass.kubernetes.io/is-default-class"). This admission controller ignores any Ingress
updates; it acts only on creation.
See the ingress documentation for more about ingress
classes and how to mark one as default.
DefaultStorageClass
This admission controller observes creation of PersistentVolumeClaim objects that do not request any specific storage class
and automatically adds a default storage class to them.
This way, users that do not request any special storage class do not need to care about them at all and they
will get the default one.
This admission controller does not do anything when no default storage class is configured. When more than one storage
class is marked as default, it rejects any creation of PersistentVolumeClaim with an error and an administrator
must revisit their StorageClass objects and mark only one as default.
This admission controller ignores any PersistentVolumeClaim updates; it acts only on creation.
See persistent volume documentation about persistent volume claims and
storage classes and how to mark a storage class as default.
DefaultTolerationSeconds
This admission controller sets the default forgiveness toleration for pods to tolerate
the taints notready:NoExecute and unreachable:NoExecute based on the k8s-apiserver input parameters
default-not-ready-toleration-seconds and default-unreachable-toleration-seconds if the pods don't already
have toleration for taints node.kubernetes.io/not-ready:NoExecute or
node.kubernetes.io/unreachable:NoExecute.
The default value for default-not-ready-toleration-seconds and default-unreachable-toleration-seconds is 5 minutes.
DenyServiceExternalIPs
This admission controller rejects all net-new usage of the Service field externalIPs. This
feature is very powerful (allows network traffic interception) and not well
controlled by policy. When enabled, users of the cluster may not create new
Services which use externalIPs and may not add new values to externalIPs on
existing Service objects. Existing uses of externalIPs are not affected,
and users may remove values from externalIPs on existing Service objects.
Most users do not need this feature at all, and cluster admins should consider disabling it.
Clusters that do need to use this feature should consider using some custom policy to manage usage
of it.
This admission controller is disabled by default.
EventRateLimit
FEATURE STATE:Kubernetes v1.13 [alpha]
This admission controller mitigates the problem where the API server gets flooded by
event requests. The cluster admin can specify event rate limits by:
Enabling the EventRateLimit admission controller;
Referencing an EventRateLimit configuration file from the file provided to the API
server's command line flag --admission-control-config-file:
This plug-in facilitates creation of dedicated nodes with extended resources.
If operators want to create dedicated nodes with extended resources (like GPUs, FPGAs etc.), they are expected to
taint the node with the extended resource
name as the key. This admission controller, if enabled, automatically
adds tolerations for such taints to pods requesting extended resources, so users don't have to manually
add these tolerations.
This admission controller is diabled by default.
ImagePolicyWebhook
The ImagePolicyWebhook admission controller allows a backend webhook to make admission decisions.
This admission controller is disabled by default.
Configuration File Format
ImagePolicyWebhook uses a configuration file to set options for the behavior of the backend.
This file may be json or yaml and has the following format:
imagePolicy:kubeConfigFile:/path/to/kubeconfig/for/backend# time in s to cache approvalallowTTL:50# time in s to cache denialdenyTTL:50# time in ms to wait between retriesretryBackoff:500# determines behavior if the webhook backend failsdefaultAllow:true
Reference the ImagePolicyWebhook configuration file from the file provided to the API server's command line flag --admission-control-config-file:
The ImagePolicyWebhook config file must reference a
kubeconfig
formatted file which sets up the connection to the backend.
It is required that the backend communicate over TLS.
The kubeconfig file's cluster field must point to the remote service, and the user field
must contain the returned authorizer.
# clusters refers to the remote service.clusters:- name:name-of-remote-imagepolicy-servicecluster:certificate-authority:/path/to/ca.pem # CA for verifying the remote service.server:https://images.example.com/policy# URL of remote service to query. Must use 'https'.# users refers to the API server's webhook configuration.users:- name:name-of-api-serveruser:client-certificate:/path/to/cert.pem# cert for the webhook admission controller to useclient-key:/path/to/key.pem # key matching the cert
For additional HTTP configuration, refer to the
kubeconfig documentation.
Request payloads
When faced with an admission decision, the API Server POSTs a JSON serialized
imagepolicy.k8s.io/v1alpha1ImageReview object describing the action.
This object contains fields describing the containers being admitted, as well as
any pod annotations that match *.image-policy.k8s.io/*.
{{ note }}
The webhook API objects are subject to the same versioning compatibility rules
as other Kubernetes API objects. Implementers should be aware of looser compatibility
promises for alpha objects and check the apiVersion field of the request to
ensure correct deserialization.
Additionally, the API Server must enable the imagepolicy.k8s.io/v1alpha1 API extensions
group (--runtime-config=imagepolicy.k8s.io/v1alpha1=true).
{{ /note }}
The remote service is expected to fill the ImageReviewStatus field of the request and
respond to either allow or disallow access. The response body's spec field is ignored and
may be omitted. A permissive response would return:
All annotations on a Pod that match *.image-policy.k8s.io/* are sent to the webhook.
Sending annotations allows users who are aware of the image policy backend to
send extra information to it, and for different backends implementations to
accept different information.
Examples of information you might put here are:
request to "break glass" to override a policy, in case of emergency.
a ticket number from a ticket system that documents the break-glass request
provide a hint to the policy server as to the imageID of the image being provided, to save it a lookup
In any case, the annotations are provided by the user and are not validated by Kubernetes in any way.
LimitPodHardAntiAffinityTopology
This admission controller denies any pod that defines AntiAffinity topology key other than
kubernetes.io/hostname in requiredDuringSchedulingRequiredDuringExecution.
This admission controller is disabled by default.
LimitRanger
This admission controller will observe the incoming request and ensure that it does not violate
any of the constraints enumerated in the LimitRange object in a Namespace. If you are using
LimitRange objects in your Kubernetes deployment, you MUST use this admission controller to
enforce those constraints. LimitRanger can also be used to apply default resource requests to Pods
that don't specify any; currently, the default LimitRanger applies a 0.1 CPU requirement to all
Pods in the default namespace.
This admission controller calls any mutating webhooks which match the request. Matching
webhooks are called in serial; each one may modify the object if it desires.
This admission controller (as implied by the name) only runs in the mutating phase.
If a webhook called by this has side effects (for example, decrementing quota) it
must have a reconciliation system, as it is not guaranteed that subsequent
webhooks or validating admission controllers will permit the request to finish.
If you disable the MutatingAdmissionWebhook, you must also disable the
MutatingWebhookConfiguration object in the admissionregistration.k8s.io/v1
group/version via the --runtime-config flag, both are on by default.
Use caution when authoring and installing mutating webhooks
Users may be confused when the objects they try to create are different from
what they get back.
Built in control loops may break when the objects they try to create are
different when read back.
Setting originally unset fields is less likely to cause problems than
overwriting fields set in the original request. Avoid doing the latter.
Future changes to control loops for built-in resources or third-party resources
may break webhooks that work well today. Even when the webhook installation API
is finalized, not all possible webhook behaviors will be guaranteed to be supported
indefinitely.
NamespaceAutoProvision
This admission controller examines all incoming requests on namespaced resources and checks
if the referenced namespace does exist.
It creates a namespace if it cannot be found.
This admission controller is useful in deployments that do not want to restrict creation of
a namespace prior to its usage.
NamespaceExists
This admission controller checks all requests on namespaced resources other than Namespace itself.
If the namespace referenced from a request doesn't exist, the request is rejected.
NamespaceLifecycle
This admission controller enforces that a Namespace that is undergoing termination cannot have
new objects created in it, and ensures that requests in a non-existent Namespace are rejected.
This admission controller also prevents deletion of three system reserved namespaces default,
kube-system, kube-public.
A Namespace deletion kicks off a sequence of operations that remove all objects (pods, services,
etc.) in that namespace. In order to enforce integrity of that process, we strongly recommend
running this admission controller.
NodeRestriction
This admission controller limits the Node and Pod objects a kubelet can modify. In order to be limited by this admission controller,
kubelets must use credentials in the system:nodes group, with a username in the form system:node:<nodeName>.
Such kubelets will only be allowed to modify their own Node API object, and only modify Pod API objects that are bound to their node.
kubelets are not allowed to update or remove taints from their Node API object.
The NodeRestriction admission plugin prevents kubelets from deleting their Node API object,
and enforces kubelet modification of labels under the kubernetes.io/ or k8s.io/ prefixes as follows:
Prevents kubelets from adding/removing/updating labels with a node-restriction.kubernetes.io/ prefix.
This label prefix is reserved for administrators to label their Node objects for workload isolation purposes,
and kubelets will not be allowed to modify labels with that prefix.
Allows kubelets to add/remove/update these labels and label prefixes:
Use of any other labels under the kubernetes.io or k8s.io prefixes by kubelets is reserved,
and may be disallowed or allowed by the NodeRestriction admission plugin in the future.
Future versions may add additional restrictions to ensure kubelets have the minimal set of
permissions required to operate correctly.
OwnerReferencesPermissionEnforcement
This admission controller protects the access to the metadata.ownerReferences of an object
so that only users with "delete" permission to the object can change it.
This admission controller also protects the access to metadata.ownerReferences[x].blockOwnerDeletion
of an object, so that only users with "update" permission to the finalizers
subresource of the referenced owner can change it.
PersistentVolumeClaimResize
FEATURE STATE:Kubernetes v1.24 [stable]
This admission controller implements additional validations for checking incoming
PersistentVolumeClaim resize requests.
Enabling the PersistentVolumeClaimResize admission controller is recommended.
This admission controller prevents resizing of all claims by default unless a claim's StorageClass
explicitly enables resizing by setting allowVolumeExpansion to true.
For example: all PersistentVolumeClaims created from the following StorageClass support volume expansion:
This admission controller automatically attaches region or zone labels to PersistentVolumes
as defined by the cloud provider (for example, GCE or AWS).
It helps ensure the Pods and the PersistentVolumes mounted are in the same
region and/or zone.
If the admission controller doesn't support automatic labelling your PersistentVolumes, you
may need to add the labels manually to prevent pods from mounting volumes from
a different zone. PersistentVolumeLabel is DEPRECATED and labeling persistent volumes has been taken over by
the cloud-controller-manager.
This admission controller is disabled by default.
PodNodeSelector
FEATURE STATE:Kubernetes v1.5 [alpha]
This admission controller defaults and limits what node selectors may be used within a namespace
by reading a namespace annotation and a global configuration.
This admission controller is disabled by default.
Configuration file format
PodNodeSelector uses a configuration file to set options for the behavior of the backend.
Note that the configuration file format will move to a versioned file in a future release.
This file may be json or yaml and has the following format:
This admission controller has the following behavior:
If the Namespace has an annotation with a key scheduler.alpha.kubernetes.io/node-selector,
use its value as the node selector.
If the namespace lacks such an annotation, use the clusterDefaultNodeSelector defined in the
PodNodeSelector plugin configuration file as the node selector.
Evaluate the pod's node selector against the namespace node selector for conflicts. Conflicts
result in rejection.
Evaluate the pod's node selector against the namespace-specific allowed selector defined the
plugin configuration file. Conflicts result in rejection.
Note: PodNodeSelector allows forcing pods to run on specifically labeled nodes. Also see the PodTolerationRestriction
admission plugin, which allows preventing pods from running on specifically tainted nodes.
PodSecurity
FEATURE STATE:Kubernetes v1.23 [beta]
This is the replacement for the deprecated PodSecurityPolicy admission controller
defined in the next section. This admission controller acts on creation and modification of the pod and
determines if it should be admitted based on the requested security context and the
Pod Security Standards.
This admission controller acts on creation and modification of the pod and determines if it should be admitted
based on the requested security context and the available Pod Security Policies.
The PodTolerationRestriction admission controller verifies any conflict between tolerations of a
pod and the tolerations of its namespace.
It rejects the pod request if there is a conflict.
It then merges the tolerations annotated on the namespace into the tolerations of the pod.
The resulting tolerations are checked against a list of allowed tolerations annotated to the namespace.
If the check succeeds, the pod request is admitted otherwise it is rejected.
If the namespace of the pod does not have any associated default tolerations or allowed
tolerations annotated, the cluster-level default tolerations or cluster-level list of allowed tolerations are used
instead if they are specified.
Tolerations to a namespace are assigned via the scheduler.alpha.kubernetes.io/defaultTolerations annotation key.
The list of allowed tolerations can be added via the scheduler.alpha.kubernetes.io/tolerationsWhitelist annotation key.
The priority admission controller uses the priorityClassName field and populates the integer
value of the priority.
If the priority class is not found, the Pod is rejected.
ResourceQuota
This admission controller will observe the incoming request and ensure that it does not violate
any of the constraints enumerated in the ResourceQuota object in a Namespace. If you are
using ResourceQuota objects in your Kubernetes deployment, you MUST use this admission
controller to enforce quota constraints.
If you define a RuntimeClass with Pod overhead
configured, this admission controller checks incoming Pods.
When enabled, this admission controller rejects any Pod create requests
that have the overhead already set.
For Pods that have a RuntimeClass configured and selected in their .spec,
this admission controller sets .spec.overhead in the Pod based on the value
defined in the corresponding RuntimeClass.
This admission controller implements automation for
serviceAccounts.
We strongly recommend using this admission controller if you intend to make use of Kubernetes
ServiceAccount objects.
StorageObjectInUseProtection
The StorageObjectInUseProtection plugin adds the kubernetes.io/pvc-protection or kubernetes.io/pv-protection
finalizers to newly created Persistent Volume Claims (PVCs) or Persistent Volumes (PV).
In case a user deletes a PVC or PV the PVC or PV is not removed until the finalizer is removed
from the PVC or PV by PVC or PV Protection Controller.
Refer to the
Storage Object in Use Protection
for more detailed information.
TaintNodesByCondition
This admission controller taints newly created
Nodes as NotReady and NoSchedule. That tainting avoids a race condition that could cause Pods
to be scheduled on new Nodes before their taints were updated to accurately reflect their reported
conditions.
ValidatingAdmissionWebhook
This admission controller calls any validating webhooks which match the request. Matching
webhooks are called in parallel; if any of them rejects the request, the request
fails. This admission controller only runs in the validation phase; the webhooks it calls may not
mutate the object, as opposed to the webhooks called by the MutatingAdmissionWebhook admission controller.
If a webhook called by this has side effects (for example, decrementing quota) it
must have a reconciliation system, as it is not guaranteed that subsequent
webhooks or other validating admission controllers will permit the request to finish.
If you disable the ValidatingAdmissionWebhook, you must also disable the
ValidatingWebhookConfiguration object in the admissionregistration.k8s.io/v1
group/version via the --runtime-config flag.
Is there a recommended set of admission controllers to use?
Yes. The recommended admission controllers are enabled by default
(shown here),
so you do not need to explicitly specify them.
You can enable additional admission controllers beyond the default set using the
--enable-admission-plugins flag (order doesn't matter).
6.3.5 - Dynamic Admission Control
In addition to compiled-in admission plugins,
admission plugins can be developed as extensions and run as webhooks configured at runtime.
This page describes how to build, configure, use, and monitor admission webhooks.
What are admission webhooks?
Admission webhooks are HTTP callbacks that receive admission requests and do
something with them. You can define two types of admission webhooks,
validating admission webhook
and
mutating admission webhook.
Mutating admission webhooks are invoked first, and can modify objects sent to the API server to enforce custom defaults.
After all object modifications are complete, and after the incoming object is validated by the API server,
validating admission webhooks are invoked and can reject requests to enforce custom policies.
Note: Admission webhooks that need to guarantee they see the final state of the object in order to enforce policy
should use a validating admission webhook, since objects can be modified after being seen by mutating webhooks.
Experimenting with admission webhooks
Admission webhooks are essentially part of the cluster control-plane. You should
write and deploy them with great caution. Please read the
user guides
for instructions if you intend to write/deploy production-grade admission webhooks.
In the following, we describe how to quickly experiment with admission webhooks.
Prerequisites
Ensure that MutatingAdmissionWebhook and ValidatingAdmissionWebhook
admission controllers are enabled.
Here
is a recommended set of admission controllers to enable in general.
Ensure that the admissionregistration.k8s.io/v1 API is enabled.
Write an admission webhook server
Please refer to the implementation of the admission webhook server
that is validated in a Kubernetes e2e test. The webhook handles the
AdmissionReview request sent by the API servers, and sends back its decision
as an AdmissionReview object in the same version it received.
See the webhook request section for details on the data sent to webhooks.
See the webhook response section for the data expected from webhooks.
The example admission webhook server leaves the ClientAuth field
empty,
which defaults to NoClientCert. This means that the webhook server does not
authenticate the identity of the clients, supposedly API servers. If you need
mutual TLS or other ways to authenticate the clients, see
how to authenticate API servers.
Deploy the admission webhook service
The webhook server in the e2e test is deployed in the Kubernetes cluster, via
the deployment API.
The test also creates a service
as the front-end of the webhook server. See
code.
You may also deploy your webhooks outside of the cluster. You will need to update
your webhook configurations accordingly.
The following is an example ValidatingWebhookConfiguration, a mutating webhook configuration is similar.
See the webhook configuration section for details about each config field.
Note: You must replace the <CA_BUNDLE> in the above example by a valid CA bundle
which is a PEM-encoded CA bundle for validating the webhook's server certificate.
The scope field specifies if only cluster-scoped resources ("Cluster") or namespace-scoped
resources ("Namespaced") will match this rule. "∗" means that there are no scope restrictions.
Note: When using clientConfig.service, the server cert must be valid for
<svc_name>.<svc_namespace>.svc.
Note: Default timeout for a webhook call is 10 seconds,
You can set the timeout and it is encouraged to use a short timeout for webhooks.
If the webhook call times out, the request is handled according to the webhook's
failure policy.
When an API server receives a request that matches one of the rules, the
API server sends an admissionReview request to webhook as specified in the
clientConfig.
After you create the webhook configuration, the system will take a few seconds
to honor the new configuration.
Authenticate API servers
If your admission webhooks require authentication, you can configure the
API servers to use basic auth, bearer token, or a cert to authenticate itself to
the webhooks. There are three steps to complete the configuration.
When starting the API server, specify the location of the admission control
configuration file via the --admission-control-config-file flag.
In the admission control configuration file, specify where the
MutatingAdmissionWebhook controller and ValidatingAdmissionWebhook controller
should read the credentials. The credentials are stored in kubeConfig files
(yes, the same schema that's used by kubectl), so the field name is
kubeConfigFile. Here is an example admission control configuration file:
# Deprecated in v1.17 in favor of apiserver.config.k8s.io/v1apiVersion:apiserver.k8s.io/v1alpha1kind:AdmissionConfigurationplugins:- name:ValidatingAdmissionWebhookconfiguration:# Deprecated in v1.17 in favor of apiserver.config.k8s.io/v1, kind=WebhookAdmissionConfigurationapiVersion:apiserver.config.k8s.io/v1alpha1kind:WebhookAdmissionkubeConfigFile:"<path-to-kubeconfig-file>"- name:MutatingAdmissionWebhookconfiguration:# Deprecated in v1.17 in favor of apiserver.config.k8s.io/v1, kind=WebhookAdmissionConfigurationapiVersion:apiserver.config.k8s.io/v1alpha1kind:WebhookAdmissionkubeConfigFile:"<path-to-kubeconfig-file>"
apiVersion:v1kind:Configusers:# name should be set to the DNS name of the service or the host (including port) of the URL the webhook is configured to speak to.# If a non-443 port is used for services, it must be included in the name when configuring 1.16+ API servers.## For a webhook configured to speak to a service on the default port (443), specify the DNS name of the service:# - name: webhook1.ns1.svc# user: ...## For a webhook configured to speak to a service on non-default port (e.g. 8443), specify the DNS name and port of the service in 1.16+:# - name: webhook1.ns1.svc:8443# user: ...# and optionally create a second stanza using only the DNS name of the service for compatibility with 1.15 API servers:# - name: webhook1.ns1.svc# user: ...## For webhooks configured to speak to a URL, match the host (and port) specified in the webhook's URL. Examples:# A webhook with `url: https://www.example.com`:# - name: www.example.com# user: ...## A webhook with `url: https://www.example.com:443`:# - name: www.example.com:443# user: ...## A webhook with `url: https://www.example.com:8443`:# - name: www.example.com:8443# user: ...#- name:'webhook1.ns1.svc'user:client-certificate-data:"<pem encoded certificate>"client-key-data:"<pem encoded key>"# The `name` supports using * to wildcard-match prefixing segments.- name:'*.webhook-company.org'user:password:"<password>"username:"<name>"# '*' is the default match.- name:'*'user:token:"<token>"
Of course you need to set up the webhook server to handle these authentication requests.
Webhook request and response
Request
Webhooks are sent as POST requests, with Content-Type: application/json,
with an AdmissionReview API object in the admission.k8s.io API group
serialized to JSON as the body.
Webhooks can specify what versions of AdmissionReview objects they accept
with the admissionReviewVersions field in their configuration:
admissionReviewVersions is a required field when creating webhook configurations.
Webhooks are required to support at least one AdmissionReview
version understood by the current and previous API server.
API servers send the first AdmissionReview version in the admissionReviewVersions list they support.
If none of the versions in the list are supported by the API server, the configuration will not be allowed to be created.
If an API server encounters a webhook configuration that was previously created and does not support any of the AdmissionReview
versions the API server knows how to send, attempts to call to the webhook will fail and be subject to the failure policy.
This example shows the data contained in an AdmissionReview object
for a request to update the scale subresource of an apps/v1Deployment:
apiVersion:admission.k8s.io/v1kind:AdmissionReviewrequest:# Random uid uniquely identifying this admission calluid:705ab4f5-6393-11e8-b7cc-42010a800002# Fully-qualified group/version/kind of the incoming objectkind:group:autoscalingversion:v1kind:Scale# Fully-qualified group/version/kind of the resource being modifiedresource:group:appsversion:v1resource:deployments# subresource, if the request is to a subresourcesubResource:scale# Fully-qualified group/version/kind of the incoming object in the original request to the API server.# This only differs from `kind` if the webhook specified `matchPolicy: Equivalent` and the# original request to the API server was converted to a version the webhook registered for.requestKind:group:autoscalingversion:v1kind:Scale# Fully-qualified group/version/kind of the resource being modified in the original request to the API server.# This only differs from `resource` if the webhook specified `matchPolicy: Equivalent` and the# original request to the API server was converted to a version the webhook registered for.requestResource:group:appsversion:v1resource:deployments# subresource, if the request is to a subresource# This only differs from `subResource` if the webhook specified `matchPolicy: Equivalent` and the# original request to the API server was converted to a version the webhook registered for.requestSubResource:scale# Name of the resource being modifiedname:my-deployment# Namespace of the resource being modified, if the resource is namespaced (or is a Namespace object)namespace:my-namespace# operation can be CREATE, UPDATE, DELETE, or CONNECToperation:UPDATEuserInfo:# Username of the authenticated user making the request to the API serverusername:admin# UID of the authenticated user making the request to the API serveruid:014fbff9a07c# Group memberships of the authenticated user making the request to the API servergroups:- system:authenticated- my-admin-group# Arbitrary extra info associated with the user making the request to the API server.# This is populated by the API server authentication layer and should be included# if any SubjectAccessReview checks are performed by the webhook.extra:some-key:- some-value1- some-value2# object is the new object being admitted.# It is null for DELETE operations.object:apiVersion:autoscaling/v1kind:Scale# oldObject is the existing object.# It is null for CREATE and CONNECT operations.oldObject:apiVersion:autoscaling/v1kind:Scale# options contains the options for the operation being admitted, like meta.k8s.io/v1 CreateOptions, UpdateOptions, or DeleteOptions.# It is null for CONNECT operations.options:apiVersion:meta.k8s.io/v1kind:UpdateOptions# dryRun indicates the API request is running in dry run mode and will not be persisted.# Webhooks with side effects should avoid actuating those side effects when dryRun is true.# See http://k8s.io/docs/reference/using-api/api-concepts/#make-a-dry-run-request for more details.dryRun:False
Response
Webhooks respond with a 200 HTTP status code, Content-Type: application/json,
and a body containing an AdmissionReview object (in the same version they were sent),
with the response stanza populated, serialized to JSON.
At a minimum, the response stanza must contain the following fields:
uid, copied from the request.uid sent to the webhook
allowed, either set to true or false
Example of a minimal response from a webhook to allow a request:
When rejecting a request, the webhook can customize the http code and message returned to the user
using the status field. The specified status object is returned to the user.
See the API documentation
for details about the status type.
Example of a response to forbid a request, customizing the HTTP status code and message presented to the user:
{
"apiVersion": "admission.k8s.io/v1",
"kind": "AdmissionReview",
"response": {
"uid": "<value from request.uid>",
"allowed": false,
"status": {
"code": 403,
"message": "You cannot do this because it is Tuesday and your name starts with A" }
}
}
When allowing a request, a mutating admission webhook may optionally modify the incoming object as well.
This is done using the patch and patchType fields in the response.
The only currently supported patchType is JSONPatch.
See JSON patch documentation for more details.
For patchType: JSONPatch, the patch field contains a base64-encoded array of JSON patch operations.
As an example, a single patch operation that would set spec.replicas would be
[{"op": "add", "path": "/spec/replicas", "value": 3}]
Base64-encoded, this would be W3sib3AiOiAiYWRkIiwgInBhdGgiOiAiL3NwZWMvcmVwbGljYXMiLCAidmFsdWUiOiAzfV0=
Admission webhooks can optionally return warning messages that are returned to the requesting client
in HTTP Warning headers with a warning code of 299. Warnings can be sent with allowed or rejected admission responses.
If you're implementing a webhook that returns a warning:
Don't include a "Warning:" prefix in the message
Use warning messages to describe problems the client making the API request should correct or be aware of
Limit warnings to 120 characters if possible
Caution: Individual warning messages over 256 characters may be truncated by the API server before being returned to clients.
If more than 4096 characters of warning messages are added (from all sources), additional warning messages are ignored.
{
"apiVersion": "admission.k8s.io/v1",
"kind": "AdmissionReview",
"response": {
"uid": "<value from request.uid>",
"allowed": true,
"warnings": [
"duplicate envvar entries specified with name MY_ENV",
"memory request less than 4MB specified for container mycontainer, which will not start successfully" ]
}
}
Webhook configuration
To register admission webhooks, create MutatingWebhookConfiguration or ValidatingWebhookConfiguration API objects.
The name of a MutatingWebhookConfiguration or a ValidatingWebhookConfiguration object must be a valid
DNS subdomain name.
Each configuration can contain one or more webhooks.
If multiple webhooks are specified in a single configuration, each must be given a unique name.
This is required in order to make resulting audit logs and metrics easier to match up to active
configurations.
Each webhook defines the following things.
Matching requests: rules
Each webhook must specify a list of rules used to determine if a request to the API server should be sent to the webhook.
Each rule specifies one or more operations, apiGroups, apiVersions, and resources, and a resource scope:
operations lists one or more operations to match. Can be "CREATE", "UPDATE", "DELETE", "CONNECT",
or "*" to match all.
apiGroups lists one or more API groups to match. "" is the core API group. "*" matches all API groups.
apiVersions lists one or more API versions to match. "*" matches all API versions.
resources lists one or more resources to match.
"*" matches all resources, but not subresources.
"*/*" matches all resources and subresources.
"pods/*" matches all subresources of pods.
"*/status" matches all status subresources.
scope specifies a scope to match. Valid values are "Cluster", "Namespaced", and "*".
Subresources match the scope of their parent resource. Default is "*".
"Cluster" means that only cluster-scoped resources will match this rule (Namespace API objects are cluster-scoped).
"Namespaced" means that only namespaced resources will match this rule.
"*" means that there are no scope restrictions.
If an incoming request matches one of the specified operations, groups, versions,
resources, and scope for any of a webhook's rules, the request is sent to the webhook.
Here are other examples of rules that could be used to specify which resources should be intercepted.
Match CREATE or UPDATE requests to apps/v1 and apps/v1beta1deployments and replicasets:
Webhooks may optionally limit which requests are intercepted based on the labels of the
objects they would be sent, by specifying an objectSelector. If specified, the objectSelector
is evaluated against both the object and oldObject that would be sent to the webhook,
and is considered to match if either object matches the selector.
A null object (oldObject in the case of create, or newObject in the case of delete),
or an object that cannot have labels (like a DeploymentRollback or a PodProxyOptions object)
is not considered to match.
Use the object selector only if the webhook is opt-in, because end users may skip
the admission webhook by setting the labels.
This example shows a mutating webhook that would match a CREATE of any resource with the label foo: bar:
See labels concept
for more examples of label selectors.
Matching requests: namespaceSelector
Webhooks may optionally limit which requests for namespaced resources are intercepted,
based on the labels of the containing namespace, by specifying a namespaceSelector.
The namespaceSelector decides whether to run the webhook on a request for a namespaced resource
(or a Namespace object), based on whether the namespace's labels match the selector.
If the object itself is a namespace, the matching is performed on object.metadata.labels.
If the object is a cluster scoped resource other than a Namespace, namespaceSelector has no effect.
This example shows a mutating webhook that matches a CREATE of any namespaced resource inside a namespace
that does not have a "runlevel" label of "0" or "1":
This example shows a validating webhook that matches a CREATE of any namespaced resource inside
a namespace that is associated with the "environment" of "prod" or "staging":
See labels concept
for more examples of label selectors.
Matching requests: matchPolicy
API servers can make objects available via multiple API groups or versions.
For example, if a webhook only specified a rule for some API groups/versions
(like apiGroups:["apps"], apiVersions:["v1","v1beta1"]),
and a request was made to modify the resource via another API group/version (like extensions/v1beta1),
the request would not be sent to the webhook.
The matchPolicy lets a webhook define how its rules are used to match incoming requests.
Allowed values are Exact or Equivalent.
Exact means a request should be intercepted only if it exactly matches a specified rule.
Equivalent means a request should be intercepted if modifies a resource listed in rules,
even via another API group or version.
In the example given above, the webhook that only registered for apps/v1 could use matchPolicy:
matchPolicy: Exact would mean the extensions/v1beta1 request would not be sent to the webhook
matchPolicy: Equivalent means the extensions/v1beta1 request would be sent to the webhook
(with the objects converted to a version the webhook had specified: apps/v1)
Specifying Equivalent is recommended, and ensures that webhooks continue to intercept the
resources they expect when upgrades enable new versions of the resource in the API server.
When a resource stops being served by the API server, it is no longer considered equivalent to
other versions of that resource that are still served.
For example, extensions/v1beta1 deployments were first deprecated and then removed (in Kubernetes v1.16).
Since that removal, a webhook with a apiGroups:["extensions"], apiVersions:["v1beta1"], resources:["deployments"] rule
does not intercept deployments created via apps/v1 APIs. For that reason, webhooks should prefer registering
for stable versions of resources.
This example shows a validating webhook that intercepts modifications to deployments (no matter the API group or version),
and is always sent an apps/v1Deployment object:
The matchPolicy for an admission webhooks defaults to Equivalent.
Contacting the webhook
Once the API server has determined a request should be sent to a webhook,
it needs to know how to contact the webhook. This is specified in the clientConfig
stanza of the webhook configuration.
Webhooks can either be called via a URL or a service reference,
and can optionally include a custom CA bundle to use to verify the TLS connection.
URL
url gives the location of the webhook, in standard URL form
(scheme://host:port/path).
The host should not refer to a service running in the cluster; use
a service reference by specifying the service field instead.
The host might be resolved via external DNS in some API servers
(e.g., kube-apiserver cannot resolve in-cluster DNS as that would
be a layering violation). host may also be an IP address.
Please note that using localhost or 127.0.0.1 as a host is
risky unless you take great care to run this webhook on all hosts
which run an API server which might need to make calls to this
webhook. Such installations are likely to be non-portable or not readily
run in a new cluster.
The scheme must be "https"; the URL must begin with "https://".
Attempting to use a user or basic auth (for example user:password@) is not allowed.
Fragments (#...) and query parameters (?...) are also not allowed.
Here is an example of a mutating webhook configured to call a URL
(and expects the TLS certificate to be verified using system trust roots, so does not specify a caBundle):
The service stanza inside clientConfig is a reference to the service for this webhook.
If the webhook is running within the cluster, then you should use service instead of url.
The service namespace and name are required. The port is optional and defaults to 443.
The path is optional and defaults to "/".
Here is an example of a mutating webhook configured to call a service on port "1234"
at the subpath "/my-path", and to verify the TLS connection against the ServerName
my-service-name.my-service-namespace.svc using a custom CA bundle:
Note: You must replace the <CA_BUNDLE> in the above example by a valid CA bundle
which is a PEM-encoded CA bundle for validating the webhook's server certificate.
Side effects
Webhooks typically operate only on the content of the AdmissionReview sent to them.
Some webhooks, however, make out-of-band changes as part of processing admission requests.
Webhooks that make out-of-band changes ("side effects") must also have a reconciliation mechanism
(like a controller) that periodically determines the actual state of the world, and adjusts
the out-of-band data modified by the admission webhook to reflect reality.
This is because a call to an admission webhook does not guarantee the admitted object will be persisted as is, or at all.
Later webhooks can modify the content of the object, a conflict could be encountered while writing to storage,
or the server could power off before persisting the object.
Additionally, webhooks with side effects must skip those side-effects when dryRun: true admission requests are handled.
A webhook must explicitly indicate that it will not have side-effects when run with dryRun,
or the dry-run request will not be sent to the webhook and the API request will fail instead.
Webhooks indicate whether they have side effects using the sideEffects field in the webhook configuration:
None: calling the webhook will have no side effects.
NoneOnDryRun: calling the webhook will possibly have side effects, but if a request with
dryRun: true is sent to the webhook, the webhook will suppress the side effects (the webhook
is dryRun-aware).
Here is an example of a validating webhook indicating it has no side effects on dryRun: true requests:
Because webhooks add to API request latency, they should evaluate as quickly as possible.
timeoutSeconds allows configuring how long the API server should wait for a webhook to respond
before treating the call as a failure.
If the timeout expires before the webhook responds, the webhook call will be ignored or
the API call will be rejected based on the failure policy.
The timeout value must be between 1 and 30 seconds.
Here is an example of a validating webhook with a custom timeout of 2 seconds:
The timeout for an admission webhook defaults to 10 seconds.
Reinvocation policy
A single ordering of mutating admissions plugins (including webhooks) does not work for all cases
(see https://issue.k8s.io/64333 as an example). A mutating webhook can add a new sub-structure
to the object (like adding a container to a pod), and other mutating plugins which have already
run may have opinions on those new structures (like setting an imagePullPolicy on all containers).
To allow mutating admission plugins to observe changes made by other plugins,
built-in mutating admission plugins are re-run if a mutating webhook modifies an object,
and mutating webhooks can specify a reinvocationPolicy to control whether they are reinvoked as well.
reinvocationPolicy may be set to Never or IfNeeded. It defaults to Never.
Never: the webhook must not be called more than once in a single admission evaluation.
IfNeeded: the webhook may be called again as part of the admission evaluation if the object
being admitted is modified by other admission plugins after the initial webhook call.
The important elements to note are:
The number of additional invocations is not guaranteed to be exactly one.
If additional invocations result in further modifications to the object, webhooks are not
guaranteed to be invoked again.
Webhooks that use this option may be reordered to minimize the number of additional invocations.
To validate an object after all mutations are guaranteed complete, use a validating admission
webhook instead (recommended for webhooks with side-effects).
Here is an example of a mutating webhook opting into being re-invoked if later admission plugins
modify the object:
Mutating webhooks must be idempotent, able to successfully process an object they have already admitted
and potentially modified. This is true for all mutating admission webhooks, since any change they can make
in an object could already exist in the user-provided object, but it is essential for webhooks that opt into reinvocation.
Failure policy
failurePolicy defines how unrecognized errors and timeout errors from the admission webhook
are handled. Allowed values are Ignore or Fail.
Ignore means that an error calling the webhook is ignored and the API request is allowed to continue.
Fail means that an error calling the webhook causes the admission to fail and the API request to be rejected.
Here is a mutating webhook configured to reject an API request if errors are encountered calling the admission webhook:
The default failurePolicy for an admission webhooks is Fail.
Monitoring admission webhooks
The API server provides ways to monitor admission webhook behaviors. These
monitoring mechanisms help cluster admins to answer questions like:
Which mutating webhook mutated the object in a API request?
What change did the mutating webhook applied to the object?
Which webhooks are frequently rejecting API requests? What's the reason for a rejection?
Mutating webhook auditing annotations
Sometimes it's useful to know which mutating webhook mutated the object in a API request, and what change did the
webhook apply.
The Kubernetes API server performs auditing on each
mutating webhook invocation. Each invocation generates an auditing annotation
capturing if a request object is mutated by the invocation, and optionally generates an annotation
capturing the applied patch from the webhook admission response. The annotations are set in the
audit event for given request on given stage of its execution, which is then pre-processed
according to a certain policy and written to a backend.
The audit level of a event determines which annotations get recorded:
At Metadata audit level or higher, an annotation with key
mutation.webhook.admission.k8s.io/round_{round idx}_index_{order idx} gets logged with JSON
payload indicating a webhook gets invoked for given request and whether it mutated the object or not.
For example, the following annotation gets recorded for a webhook being reinvoked. The webhook is
ordered the third in the mutating webhook chain, and didn't mutated the request object during the
invocation.
# the audit event recorded{"kind": "Event","apiVersion": "audit.k8s.io/v1","annotations": {"mutation.webhook.admission.k8s.io/round_1_index_2": "{\"configuration\":\"my-mutating-webhook-configuration.example.com\",\"webhook\":\"my-webhook.example.com\",\"mutated\": false}"# other annotations...}# other fields...}
# the annotation value deserialized{"configuration": "my-mutating-webhook-configuration.example.com","webhook": "my-webhook.example.com","mutated": false}
The following annotation gets recorded for a webhook being invoked in the first round. The webhook
is ordered the first in the mutating webhook chain, and mutated the request object during the
invocation.
# the audit event recorded{"kind": "Event","apiVersion": "audit.k8s.io/v1","annotations": {"mutation.webhook.admission.k8s.io/round_0_index_0": "{\"configuration\":\"my-mutating-webhook-configuration.example.com\",\"webhook\":\"my-webhook-always-mutate.example.com\",\"mutated\": true}"# other annotations...}# other fields...}
# the annotation value deserialized{"configuration": "my-mutating-webhook-configuration.example.com","webhook": "my-webhook-always-mutate.example.com","mutated": true}
At Request audit level or higher, an annotation with key
patch.webhook.admission.k8s.io/round_{round idx}_index_{order idx} gets logged with JSON payload indicating
a webhook gets invoked for given request and what patch gets applied to the request object.
For example, the following annotation gets recorded for a webhook being reinvoked. The webhook is ordered the fourth in the
mutating webhook chain, and responded with a JSON patch which got applied to the request object.
# the audit event recorded{"kind": "Event","apiVersion": "audit.k8s.io/v1","annotations": {"patch.webhook.admission.k8s.io/round_1_index_3": "{\"configuration\":\"my-other-mutating-webhook-configuration.example.com\",\"webhook\":\"my-webhook-always-mutate.example.com\",\"patch\":[{\"op\":\"add\",\"path\":\"/data/mutation-stage\",\"value\":\"yes\"}],\"patchType\":\"JSONPatch\"}"# other annotations...}# other fields...}
# the annotation value deserialized{"configuration": "my-other-mutating-webhook-configuration.example.com","webhook": "my-webhook-always-mutate.example.com","patchType": "JSONPatch","patch": [{"op": "add","path": "/data/mutation-stage","value": "yes"}]}
Admission webhook metrics
The API server exposes Prometheus metrics from the /metrics endpoint, which can be used for monitoring and
diagnosing API server status. The following metrics record status related to admission webhooks.
API server admission webhook rejection count
Sometimes it's useful to know which admission webhooks are frequently rejecting API requests, and the
reason for a rejection.
The API server exposes a Prometheus counter metric recording admission webhook rejections. The
metrics are labelled to identify the causes of webhook rejection(s):
name: the name of the webhook that rejected a request.
operation: the operation type of the request, can be one of CREATE,
UPDATE, DELETE and CONNECT.
type: the admission webhook type, can be one of admit and validating.
error_type: identifies if an error occurred during the webhook invocation
that caused the rejection. Its value can be one of:
calling_webhook_error: unrecognized errors or timeout errors from the admission webhook happened and the
webhook's Failure policy is set to Fail.
no_error: no error occurred. The webhook rejected the request with allowed: false in the admission
response. The metrics label rejection_code records the .status.code set in the admission response.
apiserver_internal_error: an API server internal error happened.
rejection_code: the HTTP status code set in the admission response when a
webhook rejected a request.
Example of the rejection count metrics:
# HELP apiserver_admission_webhook_rejection_count [ALPHA] Admission webhook rejection count, identified by name and broken out for each admission type (validating or admit) and operation. Additional labels specify an error type (calling_webhook_error or apiserver_internal_error if an error occurred; no_error otherwise) and optionally a non-zero rejection code if the webhook rejects the request with an HTTP status code (honored by the apiserver when the code is greater or equal to 400). Codes greater than 600 are truncated to 600, to keep the metrics cardinality bounded.
# TYPE apiserver_admission_webhook_rejection_count counter
apiserver_admission_webhook_rejection_count{error_type="calling_webhook_error",name="always-timeout-webhook.example.com",operation="CREATE",rejection_code="0",type="validating"} 1
apiserver_admission_webhook_rejection_count{error_type="calling_webhook_error",name="invalid-admission-response-webhook.example.com",operation="CREATE",rejection_code="0",type="validating"} 1
apiserver_admission_webhook_rejection_count{error_type="no_error",name="deny-unwanted-configmap-data.example.com",operation="CREATE",rejection_code="400",type="validating"} 13
Best practices and warnings
Idempotence
An idempotent mutating admission webhook is able to successfully process an object it has already admitted
and potentially modified. The admission can be applied multiple times without changing the result beyond
the initial application.
Example of idempotent mutating admission webhooks:
For a CREATE pod request, set the field .spec.securityContext.runAsNonRoot of the
pod to true, to enforce security best practices.
For a CREATE pod request, if the field .spec.containers[].resources.limits
of a container is not set, set default resource limits.
For a CREATE pod request, inject a sidecar container with name foo-sidecar if no container
with the name foo-sidecar already exists.
In the cases above, the webhook can be safely reinvoked, or admit an object that already has the fields set.
Example of non-idempotent mutating admission webhooks:
For a CREATE pod request, inject a sidecar container with name foo-sidecar
suffixed with the current timestamp (e.g. foo-sidecar-19700101-000000).
For a CREATE/UPDATE pod request, reject if the pod has label "env" set,
otherwise add an "env": "prod" label to the pod.
For a CREATE pod request, blindly append a sidecar container named
foo-sidecar without looking to see if there is already a foo-sidecar
container in the pod.
In the first case above, reinvoking the webhook can result in the same sidecar being injected multiple times to a pod, each time
with a different container name. Similarly the webhook can inject duplicated containers if the sidecar already exists in
a user-provided pod.
In the second case above, reinvoking the webhook will result in the webhook failing on its own output.
In the third case above, reinvoking the webhook will result in duplicated containers in the pod spec, which makes
the request invalid and rejected by the API server.
Intercepting all versions of an object
It is recommended that admission webhooks should always intercept all versions of an object by setting .webhooks[].matchPolicy
to Equivalent. It is also recommended that admission webhooks should prefer registering for stable versions of resources.
Failure to intercept all versions of an object can result in admission policies not being enforced for requests in certain
versions. See Matching requests: matchPolicy for examples.
Availability
It is recommended that admission webhooks should evaluate as quickly as possible (typically in
milliseconds), since they add to API request latency.
It is encouraged to use a small timeout for webhooks. See Timeouts for more detail.
It is recommended that admission webhooks should leverage some format of load-balancing, to
provide high availability and performance benefits. If a webhook is running within the cluster,
you can run multiple webhook backends behind a service to leverage the load-balancing that service
supports.
Guaranteeing the final state of the object is seen
Admission webhooks that need to guarantee they see the final state of the object in order to enforce policy
should use a validating admission webhook, since objects can be modified after being seen by mutating webhooks.
For example, a mutating admission webhook is configured to inject a sidecar container with name
"foo-sidecar" on every CREATE pod request. If the sidecar must be present, a validating
admisson webhook should also be configured to intercept CREATE pod requests, and validate that a
container with name "foo-sidecar" with the expected configuration exists in the to-be-created
object.
Avoiding deadlocks in self-hosted webhooks
A webhook running inside the cluster might cause deadlocks for its own deployment if it is configured
to intercept resources required to start its own pods.
For example, a mutating admission webhook is configured to admit CREATE pod requests only if a certain label is set in the
pod (e.g. "env": "prod"). The webhook server runs in a deployment which doesn't set the "env" label.
When a node that runs the webhook server pods
becomes unhealthy, the webhook deployment will try to reschedule the pods to another node. However the requests will
get rejected by the existing webhook server since the "env" label is unset, and the migration cannot happen.
It is recommended to exclude the namespace where your webhook is running with a
namespaceSelector.
Side effects
It is recommended that admission webhooks should avoid side effects if possible, which means the webhooks operate only on the
content of the AdmissionReview sent to them, and do not make out-of-band changes. The .webhooks[].sideEffects field should
be set to None if a webhook doesn't have any side effect.
If side effects are required during the admission evaluation, they must be suppressed when processing an
AdmissionReview object with dryRun set to true, and the .webhooks[].sideEffects field should be
set to NoneOnDryRun. See Side effects for more detail.
Avoiding operating on the kube-system namespace
The kube-system namespace contains objects created by the Kubernetes system,
e.g. service accounts for the control plane components, pods like kube-dns.
Accidentally mutating or rejecting requests in the kube-system namespace may
cause the control plane components to stop functioning or introduce unknown behavior.
If your admission webhooks don't intend to modify the behavior of the Kubernetes control
plane, exclude the kube-system namespace from being intercepted using a
namespaceSelector.
Support for authorization and user accounts is planned but incomplete. Sometimes
incomplete features are referred to in order to better describe service accounts.
User accounts versus service accounts
Kubernetes distinguishes between the concept of a user account and a service account
for a number of reasons:
User accounts are for humans. Service accounts are for processes, which run
in pods.
User accounts are intended to be global. Names must be unique across all
namespaces of a cluster. Service accounts are namespaced.
Typically, a cluster's user accounts might be synced from a corporate
database, where new user account creation requires special privileges and is
tied to complex business processes. Service account creation is intended to be
more lightweight, allowing cluster users to create service accounts for
specific tasks by following the principle of least privilege.
Auditing considerations for humans and service accounts may differ.
A config bundle for a complex system may include definition of various service
accounts for components of that system. Because service accounts can be created
without many constraints and have namespaced names, such config is portable.
Service account automation
Three separate components cooperate to implement the automation around service accounts:
A ServiceAccount admission controller
A Token controller
A ServiceAccount controller
ServiceAccount Admission Controller
The modification of pods is implemented via a plugin
called an Admission Controller.
It is part of the API server.
It acts synchronously to modify pods as they are created or updated. When this plugin is active
(and it is by default on most distributions), then it does the following when a pod is created or modified:
If the pod does not have a ServiceAccount set, it sets the ServiceAccount to default.
It ensures that the ServiceAccount referenced by the pod exists, and otherwise rejects it.
It adds a volume to the pod which contains a token for API access if neither the ServiceAccount automountServiceAccountToken nor the Pod's automountServiceAccountToken is set to false.
It adds a volumeSource to each container of the pod mounted at /var/run/secrets/kubernetes.io/serviceaccount, if the previous step has created a volume for ServiceAccount token.
If the pod does not contain any imagePullSecrets, then imagePullSecrets of the ServiceAccount are added to the pod.
Bound Service Account Token Volume
FEATURE STATE:Kubernetes v1.22 [stable]
The ServiceAccount admission controller will add the following projected volume instead of a Secret-based volume for the non-expiring service account token created by Token Controller.
A ServiceAccountToken acquired from kube-apiserver via TokenRequest API. It will expire after 1 hour by default or when the pod is deleted. It is bound to the pod and has kube-apiserver as the audience.
A ConfigMap containing a CA bundle used for verifying connections to the kube-apiserver. This feature depends on the RootCAConfigMap feature gate, which publishes a "kube-root-ca.crt" ConfigMap to every namespace. RootCAConfigMap feature gate is graduated to GA in 1.21 and default to true. (This flag will be removed from --feature-gate arg in 1.22)
A DownwardAPI that references the namespace of the pod.
TokenController runs as part of kube-controller-manager. It acts asynchronously. It:
watches ServiceAccount creation and creates a corresponding
ServiceAccount token Secret to allow API access.
watches ServiceAccount deletion and deletes all corresponding ServiceAccount
token Secrets.
watches ServiceAccount token Secret addition, and ensures the referenced
ServiceAccount exists, and adds a token to the Secret if needed.
watches Secret deletion and removes a reference from the corresponding
ServiceAccount if needed.
You must pass a service account private key file to the token controller in
the kube-controller-manager using the --service-account-private-key-file
flag. The private key is used to sign generated service account tokens.
Similarly, you must pass the corresponding public key to the kube-apiserver
using the --service-account-key-file flag. The public key will be used to
verify the tokens during authentication.
To create additional API tokens
A controller loop ensures a Secret with an API token exists for each
ServiceAccount. To create additional API tokens for a ServiceAccount, create a
Secret of type kubernetes.io/service-account-token with an annotation
referencing the ServiceAccount, and the controller will update it with a
generated token:
Below is a sample configuration for such a Secret:
To delete/invalidate a ServiceAccount token Secret
kubectl delete secret mysecretname
ServiceAccount controller
A ServiceAccount controller manages the ServiceAccounts inside namespaces, and
ensures a ServiceAccount named "default" exists in every active namespace.
6.3.7 - Authorization Overview
Learn more about Kubernetes authorization, including details about creating
policies using the supported authorization modules.
In Kubernetes, you must be authenticated (logged in) before your request can be
authorized (granted permission to access). For information about authentication,
see Controlling Access to the Kubernetes API.
Kubernetes expects attributes that are common to REST API requests. This means
that Kubernetes authorization works with existing organization-wide or
cloud-provider-wide access control systems which may handle other APIs besides
the Kubernetes API.
Determine Whether a Request is Allowed or Denied
Kubernetes authorizes API requests using the API server. It evaluates all of the
request attributes against all policies and allows or denies the request. All
parts of an API request must be allowed by some policy in order to proceed. This
means that permissions are denied by default.
(Although Kubernetes uses the API server, access controls and policies that
depend on specific fields of specific kinds of objects are handled by Admission
Controllers.)
When multiple authorization modules are configured, each is checked in sequence.
If any authorizer approves or denies a request, that decision is immediately
returned and no other authorizer is consulted. If all modules have no opinion on
the request, then the request is denied. A deny returns an HTTP status code 403.
Review Your Request Attributes
Kubernetes reviews only the following API request attributes:
user - The user string provided during authentication.
group - The list of group names to which the authenticated user belongs.
extra - A map of arbitrary string keys to string values, provided by the authentication layer.
API - Indicates whether the request is for an API resource.
Request path - Path to miscellaneous non-resource endpoints like /api or /healthz.
API request verb - API verbs like get, list, create, update, patch, watch, delete, and deletecollection are used for resource requests. To determine the request verb for a resource API endpoint, see Determine the request verb.
HTTP request verb - Lowercased HTTP methods like get, post, put, and delete are used for non-resource requests.
Resource - The ID or name of the resource that is being accessed (for resource requests only) -- For resource requests using get, update, patch, and delete verbs, you must provide the resource name.
Subresource - The subresource that is being accessed (for resource requests only).
Namespace - The namespace of the object that is being accessed (for namespaced resource requests only).
API group - The API Group being accessed (for resource requests only). An empty string designates the coreAPI group.
Determine the Request Verb
Non-resource requests
Requests to endpoints other than /api/v1/... or /apis/<group>/<version>/...
are considered "non-resource requests", and use the lower-cased HTTP method of the request as the verb.
For example, a GET request to endpoints like /api or /healthz would use get as the verb.
Resource requests
To determine the request verb for a resource API endpoint, review the HTTP verb
used and whether or not the request acts on an individual resource or a
collection of resources:
HTTP verb
request verb
POST
create
GET, HEAD
get (for individual resources), list (for collections, including full object content), watch (for watching an individual resource or collection of resources)
Caution: The get, list and watch verbs can all return the full details of a resource. In terms of the returned data they are equivalent. For example, list on secrets will still reveal the data attributes of any returned resources.
Kubernetes sometimes checks authorization for additional permissions using specialized verbs. For example:
impersonate verb on users, groups, and serviceaccounts in the core API group, and the userextras in the authentication.k8s.io API group.
Authorization Modes
The Kubernetes API server may authorize a request using one of several authorization modes:
Node - A special-purpose authorization mode that grants permissions to kubelets based on the pods they are scheduled to run. To learn more about using the Node authorization mode, see Node Authorization.
ABAC - Attribute-based access control (ABAC) defines an access control paradigm whereby access rights are granted to users through the use of policies which combine attributes together. The policies can use any type of attributes (user attributes, resource attributes, object, environment attributes, etc). To learn more about using the ABAC mode, see ABAC Mode.
RBAC - Role-based access control (RBAC) is a method of regulating access to computer or network resources based on the roles of individual users within an enterprise. In this context, access is the ability of an individual user to perform a specific task, such as view, create, or modify a file. To learn more about using the RBAC mode, see RBAC Mode
When specified RBAC (Role-Based Access Control) uses the rbac.authorization.k8s.io API group to drive authorization decisions, allowing admins to dynamically configure permission policies through the Kubernetes API.
To enable RBAC, start the apiserver with --authorization-mode=RBAC.
Webhook - A WebHook is an HTTP callback: an HTTP POST that occurs when something happens; a simple event-notification via HTTP POST. A web application implementing WebHooks will POST a message to a URL when certain things happen. To learn more about using the Webhook mode, see Webhook Mode.
Checking API Access
kubectl provides the auth can-i subcommand for quickly querying the API authorization layer.
The command uses the SelfSubjectAccessReview API to determine if the current user can perform
a given action, and works regardless of the authorization mode used.
kubectl auth can-i create deployments --namespace dev
SelfSubjectAccessReview is part of the authorization.k8s.io API group, which
exposes the API server authorization to external services. Other resources in
this group include:
SubjectAccessReview - Access review for any user, not only the current one. Useful for delegating authorization decisions to the API server. For example, the kubelet and extension API servers use this to determine user access to their own APIs.
LocalSubjectAccessReview - Like SubjectAccessReview but restricted to a specific namespace.
SelfSubjectRulesReview - A review which returns the set of actions a user can perform within a namespace. Useful for users to quickly summarize their own access, or for UIs to hide/show actions.
These APIs can be queried by creating normal Kubernetes resources, where the response "status"
field of the returned object is the result of the query.
You must include a flag in your policy to indicate which authorization module
your policies include:
The following flags can be used:
--authorization-mode=ABAC Attribute-Based Access Control (ABAC) mode allows you to configure policies using local files.
--authorization-mode=RBAC Role-based access control (RBAC) mode allows you to create and store policies using the Kubernetes API.
--authorization-mode=Webhook WebHook is an HTTP callback mode that allows you to manage authorization using a remote REST endpoint.
--authorization-mode=Node Node authorization is a special-purpose authorization mode that specifically authorizes API requests made by kubelets.
--authorization-mode=AlwaysDeny This flag blocks all requests. Use this flag only for testing.
--authorization-mode=AlwaysAllow This flag allows all requests. Use this flag only if you do not require authorization for your API requests.
You can choose more than one authorization module. Modules are checked in order
so an earlier module has higher priority to allow or deny a request.
Privilege escalation via workload creation or edits
Users who can create/edit pods in a namespace, either directly or through a controller
such as an operator, could escalate their privileges in that namespace.
Caution: System administrators, use care when granting access to create or edit workloads.
Details of how these can be misused are documented in escalation paths
Escalation paths
Mounting arbitrary secrets in that namespace
Can be used to access secrets meant for other workloads
Can be used to obtain a more privileged service account's service account token
Using arbitrary Service Accounts in that namespace
Can perform Kubernetes API actions as another workload (impersonation)
Can perform any privileged actions that Service Account has
Mounting configmaps meant for other workloads in that namespace
Can be used to obtain information meant for other workloads, such as DB host names.
Mounting volumes meant for other workloads in that namespace
Can be used to obtain information meant for other workloads, and change it.
Caution: System administrators should be cautious when deploying CRDs that
change the above areas. These may open privilege escalations paths.
This should be considered when deciding on your RBAC controls.
Role-based access control (RBAC) is a method of regulating access to computer or
network resources based on the roles of individual users within your organization.
RBAC authorization uses the rbac.authorization.k8s.ioAPI group to drive authorization
decisions, allowing you to dynamically configure policies through the Kubernetes API.
To enable RBAC, start the API server
with the --authorization-mode flag set to a comma-separated list that includes RBAC;
for example:
The RBAC API declares four kinds of Kubernetes object: Role, ClusterRole,
RoleBinding and ClusterRoleBinding. You can
describe objects,
or amend them, using tools such as kubectl, just like any other Kubernetes object.
Caution: These objects, by design, impose access restrictions. If you are making changes
to a cluster as you learn, see
privilege escalation prevention and bootstrapping
to understand how those restrictions can prevent you making some changes.
Role and ClusterRole
An RBAC Role or ClusterRole contains rules that represent a set of permissions.
Permissions are purely additive (there are no "deny" rules).
A Role always sets permissions within a particular namespace;
when you create a Role, you have to specify the namespace it belongs in.
ClusterRole, by contrast, is a non-namespaced resource. The resources have different names (Role
and ClusterRole) because a Kubernetes object always has to be either namespaced or not namespaced;
it can't be both.
ClusterRoles have several uses. You can use a ClusterRole to:
define permissions on namespaced resources and be granted access within individual namespace(s)
define permissions on namespaced resources and be granted access across all namespaces
define permissions on cluster-scoped resources
If you want to define a role within a namespace, use a Role; if you want to define
a role cluster-wide, use a ClusterRole.
Role example
Here's an example Role in the "default" namespace that can be used to grant read access to
pods:
apiVersion:rbac.authorization.k8s.io/v1kind:Rolemetadata:namespace:defaultname:pod-readerrules:- apiGroups:[""]# "" indicates the core API groupresources:["pods"]verbs:["get","watch","list"]
ClusterRole example
A ClusterRole can be used to grant the same permissions as a Role.
Because ClusterRoles are cluster-scoped, you can also use them to grant access to:
namespaced resources (like Pods), across all namespaces
For example: you can use a ClusterRole to allow a particular user to run
kubectl get pods --all-namespaces
Here is an example of a ClusterRole that can be used to grant read access to
secrets in any particular namespace,
or across all namespaces (depending on how it is bound):
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:# "namespace" omitted since ClusterRoles are not namespacedname:secret-readerrules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing Secret# objects is "secrets"resources:["secrets"]verbs:["get","watch","list"]
The name of a Role or a ClusterRole object must be a valid
path segment name.
RoleBinding and ClusterRoleBinding
A role binding grants the permissions defined in a role to a user or set of users.
It holds a list of subjects (users, groups, or service accounts), and a reference to the
role being granted.
A RoleBinding grants permissions within a specific namespace whereas a ClusterRoleBinding
grants that access cluster-wide.
A RoleBinding may reference any Role in the same namespace. Alternatively, a RoleBinding
can reference a ClusterRole and bind that ClusterRole to the namespace of the RoleBinding.
If you want to bind a ClusterRole to all the namespaces in your cluster, you use a
ClusterRoleBinding.
The name of a RoleBinding or ClusterRoleBinding object must be a valid
path segment name.
RoleBinding examples
Here is an example of a RoleBinding that grants the "pod-reader" Role to the user "jane"
within the "default" namespace.
This allows "jane" to read pods in the "default" namespace.
apiVersion:rbac.authorization.k8s.io/v1# This role binding allows "jane" to read pods in the "default" namespace.# You need to already have a Role named "pod-reader" in that namespace.kind:RoleBindingmetadata:name:read-podsnamespace:defaultsubjects:# You can specify more than one "subject"- kind:Username:jane# "name" is case sensitiveapiGroup:rbac.authorization.k8s.ioroleRef:# "roleRef" specifies the binding to a Role / ClusterRolekind:Role#this must be Role or ClusterRolename:pod-reader# this must match the name of the Role or ClusterRole you wish to bind toapiGroup:rbac.authorization.k8s.io
A RoleBinding can also reference a ClusterRole to grant the permissions defined in that
ClusterRole to resources inside the RoleBinding's namespace. This kind of reference
lets you define a set of common roles across your cluster, then reuse them within
multiple namespaces.
For instance, even though the following RoleBinding refers to a ClusterRole,
"dave" (the subject, case sensitive) will only be able to read Secrets in the "development"
namespace, because the RoleBinding's namespace (in its metadata) is "development".
apiVersion:rbac.authorization.k8s.io/v1# This role binding allows "dave" to read secrets in the "development" namespace.# You need to already have a ClusterRole named "secret-reader".kind:RoleBindingmetadata:name:read-secrets## The namespace of the RoleBinding determines where the permissions are granted.# This only grants permissions within the "development" namespace.namespace:developmentsubjects:- kind:Username:dave# Name is case sensitiveapiGroup:rbac.authorization.k8s.ioroleRef:kind:ClusterRolename:secret-readerapiGroup:rbac.authorization.k8s.io
ClusterRoleBinding example
To grant permissions across a whole cluster, you can use a ClusterRoleBinding.
The following ClusterRoleBinding allows any user in the group "manager" to read
secrets in any namespace.
apiVersion:rbac.authorization.k8s.io/v1# This cluster role binding allows anyone in the "manager" group to read secrets in any namespace.kind:ClusterRoleBindingmetadata:name:read-secrets-globalsubjects:- kind:Groupname:manager# Name is case sensitiveapiGroup:rbac.authorization.k8s.ioroleRef:kind:ClusterRolename:secret-readerapiGroup:rbac.authorization.k8s.io
After you create a binding, you cannot change the Role or ClusterRole that it refers to.
If you try to change a binding's roleRef, you get a validation error. If you do want
to change the roleRef for a binding, you need to remove the binding object and create
a replacement.
There are two reasons for this restriction:
Making roleRef immutable allows granting someone update permission on an existing binding
object, so that they can manage the list of subjects, without being able to change
the role that is granted to those subjects.
A binding to a different role is a fundamentally different binding.
Requiring a binding to be deleted/recreated in order to change the roleRef
ensures the full list of subjects in the binding is intended to be granted
the new role (as opposed to enabling or accidentally modifying only the roleRef
without verifying all of the existing subjects should be given the new role's
permissions).
The kubectl auth reconcile command-line utility creates or updates a manifest file containing RBAC objects,
and handles deleting and recreating binding objects if required to change the role they refer to.
See command usage and examples for more information.
Referring to resources
In the Kubernetes API, most resources are represented and accessed using a string representation of
their object name, such as pods for a Pod. RBAC refers to resources using exactly the same
name that appears in the URL for the relevant API endpoint.
Some Kubernetes APIs involve a
subresource, such as the logs for a Pod. A request for a Pod's logs looks like:
GET /api/v1/namespaces/{namespace}/pods/{name}/log
In this case, pods is the namespaced resource for Pod resources, and log is a
subresource of pods. To represent this in an RBAC role, use a slash (/) to
delimit the resource and subresource. To allow a subject to read pods and
also access the log subresource for each of those Pods, you write:
You can also refer to resources by name for certain requests through the resourceNames list.
When specified, requests can be restricted to individual instances of a resource.
Here is an example that restricts its subject to only get or update a
ConfigMap named my-configmap:
apiVersion:rbac.authorization.k8s.io/v1kind:Rolemetadata:namespace:defaultname:configmap-updaterrules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing ConfigMap# objects is "configmaps"resources:["configmaps"]resourceNames:["my-configmap"]verbs:["update","get"]
Note: You cannot restrict create or deletecollection requests by their resource name.
For create, this limitation is because the name of the new object may not be known at authorization time.
If you restrict list or watch by resourceName, clients must include a metadata.name field selector in their list or watch request that matches the specified resourceName in order to be authorized.
For example, kubectl get configmaps --field-selector=metadata.name=my-configmap
Aggregated ClusterRoles
You can aggregate several ClusterRoles into one combined ClusterRole.
A controller, running as part of the cluster control plane, watches for ClusterRole
objects with an aggregationRule set. The aggregationRule defines a label
selector that the controller
uses to match other ClusterRole objects that should be combined into the rules
field of this one.
Here is an example aggregated ClusterRole:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:monitoringaggregationRule:clusterRoleSelectors:- matchLabels:rbac.example.com/aggregate-to-monitoring:"true"rules:[]# The control plane automatically fills in the rules
If you create a new ClusterRole that matches the label selector of an existing aggregated ClusterRole,
that change triggers adding the new rules into the aggregated ClusterRole.
Here is an example that adds rules to the "monitoring" ClusterRole, by creating another
ClusterRole labeled rbac.example.com/aggregate-to-monitoring: true.
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:monitoring-endpointslabels:rbac.example.com/aggregate-to-monitoring:"true"# When you create the "monitoring-endpoints" ClusterRole,# the rules below will be added to the "monitoring" ClusterRole.rules:- apiGroups:[""]resources:["services","endpoints","pods"]verbs:["get","list","watch"]
The default user-facing roles use ClusterRole aggregation. This lets you,
as a cluster administrator, include rules for custom resources, such as those served by
CustomResourceDefinitions
or aggregated API servers, to extend the default roles.
For example: the following ClusterRoles let the "admin" and "edit" default roles manage the custom resource
named CronTab, whereas the "view" role can perform only read actions on CronTab resources.
You can assume that CronTab objects are named "crontabs" in URLs as seen by the API server.
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:aggregate-cron-tabs-editlabels:# Add these permissions to the "admin" and "edit" default roles.rbac.authorization.k8s.io/aggregate-to-admin:"true"rbac.authorization.k8s.io/aggregate-to-edit:"true"rules:- apiGroups:["stable.example.com"]resources:["crontabs"]verbs:["get","list","watch","create","update","patch","delete"]---kind:ClusterRoleapiVersion:rbac.authorization.k8s.io/v1metadata:name:aggregate-cron-tabs-viewlabels:# Add these permissions to the "view" default role.rbac.authorization.k8s.io/aggregate-to-view:"true"rules:- apiGroups:["stable.example.com"]resources:["crontabs"]verbs:["get","list","watch"]
Role examples
The following examples are excerpts from Role or ClusterRole objects, showing only
the rules section.
Allow reading "pods" resources in the core
API Group:
rules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing Pod# objects is "pods"resources:["pods"]verbs:["get","list","watch"]
Allow reading/writing Deployments (at the HTTP level: objects with "deployments"
in the resource part of their URL) in the "apps" API groups:
rules:- apiGroups:["apps"]## at the HTTP level, the name of the resource for accessing Deployment# objects is "deployments"resources:["deployments"]verbs:["get","list","watch","create","update","patch","delete"]
Allow reading Pods in the core API group, as well as reading or writing Job
resources in the "batch" API group:
rules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing Pod# objects is "pods"resources:["pods"]verbs:["get","list","watch"]- apiGroups:["batch"]## at the HTTP level, the name of the resource for accessing Job# objects is "jobs"resources:["jobs"]verbs:["get","list","watch","create","update","patch","delete"]
Allow reading a ConfigMap named "my-config" (must be bound with a
RoleBinding to limit to a single ConfigMap in a single namespace):
rules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing ConfigMap# objects is "configmaps"resources:["configmaps"]resourceNames:["my-config"]verbs:["get"]
Allow reading the resource "nodes" in the core group (because a
Node is cluster-scoped, this must be in a ClusterRole bound with a
ClusterRoleBinding to be effective):
rules:- apiGroups:[""]## at the HTTP level, the name of the resource for accessing Node# objects is "nodes"resources:["nodes"]verbs:["get","list","watch"]
Allow GET and POST requests to the non-resource endpoint /healthz and
all subpaths (must be in a ClusterRole bound with a ClusterRoleBinding
to be effective):
rules:- nonResourceURLs:["/healthz","/healthz/*"]# '*' in a nonResourceURL is a suffix glob matchverbs:["get","post"]
Referring to subjects
A RoleBinding or ClusterRoleBinding binds a role to subjects.
Subjects can be groups, users or
ServiceAccounts.
Kubernetes represents usernames as strings.
These can be: plain names, such as "alice"; email-style names, like "bob@example.com";
or numeric user IDs represented as a string. It is up to you as a cluster administrator
to configure the authentication modules
so that authentication produces usernames in the format you want.
Caution: The prefix system: is reserved for Kubernetes system use, so you should ensure
that you don't have users or groups with names that start with system: by
accident.
Other than this special prefix, the RBAC authorization system does not require any format
for usernames.
In Kubernetes, Authenticator modules provide group information.
Groups, like users, are represented as strings, and that string has no format requirements,
other than that the prefix system: is reserved.
ServiceAccounts have names prefixed
with system:serviceaccount:, and belong to groups that have names prefixed with system:serviceaccounts:.
Note:
system:serviceaccount: (singular) is the prefix for service account usernames.
system:serviceaccounts: (plural) is the prefix for service account groups.
RoleBinding examples
The following examples are RoleBinding excerpts that only
show the subjects section.
API servers create a set of default ClusterRole and ClusterRoleBinding objects.
Many of these are system: prefixed, which indicates that the resource is directly
managed by the cluster control plane.
All of the default ClusterRoles and ClusterRoleBindings are labeled with kubernetes.io/bootstrapping=rbac-defaults.
Caution: Take care when modifying ClusterRoles and ClusterRoleBindings with names
that have a system: prefix.
Modifications to these resources can result in non-functional clusters.
Auto-reconciliation
At each start-up, the API server updates default cluster roles with any missing permissions,
and updates default cluster role bindings with any missing subjects.
This allows the cluster to repair accidental modifications, and helps to keep roles and role bindings
up-to-date as permissions and subjects change in new Kubernetes releases.
To opt out of this reconciliation, set the rbac.authorization.kubernetes.io/autoupdate
annotation on a default cluster role or rolebinding to false.
Be aware that missing default permissions and subjects can result in non-functional clusters.
Auto-reconciliation is enabled by default if the RBAC authorizer is active.
API discovery roles
Default role bindings authorize unauthenticated and authenticated users to read API information that is deemed safe to be publicly accessible (including CustomResourceDefinitions). To disable anonymous unauthenticated access, add --anonymous-auth=false to the API server configuration.
To view the configuration of these roles via kubectl run:
kubectl get clusterroles system:discovery -o yaml
Note: If you edit that ClusterRole, your changes will be overwritten on API server restart
via auto-reconciliation. To avoid that overwriting,
either do not manually edit the role, or disable auto-reconciliation.
Kubernetes RBAC API discovery roles
Default ClusterRole
Default ClusterRoleBinding
Description
system:basic-user
system:authenticated group
Allows a user read-only access to basic information about themselves. Prior to v1.14, this role was also bound to system:unauthenticated by default.
system:discovery
system:authenticated group
Allows read-only access to API discovery endpoints needed to discover and negotiate an API level. Prior to v1.14, this role was also bound to system:unauthenticated by default.
system:public-info-viewer
system:authenticated and system:unauthenticated groups
Allows read-only access to non-sensitive information about the cluster. Introduced in Kubernetes v1.14.
User-facing roles
Some of the default ClusterRoles are not system: prefixed. These are intended to be user-facing roles.
They include super-user roles (cluster-admin), roles intended to be granted cluster-wide
using ClusterRoleBindings, and roles intended to be granted within particular
namespaces using RoleBindings (admin, edit, view).
User-facing ClusterRoles use ClusterRole aggregation to allow admins to include
rules for custom resources on these ClusterRoles. To add rules to the admin, edit, or view roles, create
a ClusterRole with one or more of the following labels:
Allows super-user access to perform any action on any resource.
When used in a ClusterRoleBinding, it gives full control over every resource in the cluster and in all namespaces.
When used in a RoleBinding, it gives full control over every resource in the role binding's namespace, including the namespace itself.
admin
None
Allows admin access, intended to be granted within a namespace using a RoleBinding.
If used in a RoleBinding, allows read/write access to most resources in a namespace,
including the ability to create roles and role bindings within the namespace.
This role does not allow write access to resource quota or to the namespace itself.
This role also does not allow write access to Endpoints in clusters created
using Kubernetes v1.22+. More information is available in the
"Write Access for Endpoints" section.
edit
None
Allows read/write access to most objects in a namespace.
This role does not allow viewing or modifying roles or role bindings.
However, this role allows accessing Secrets and running Pods as any ServiceAccount in
the namespace, so it can be used to gain the API access levels of any ServiceAccount in
the namespace. This role also does not allow write access to Endpoints in
clusters created using Kubernetes v1.22+. More information is available in the
"Write Access for Endpoints" section.
view
None
Allows read-only access to see most objects in a namespace.
It does not allow viewing roles or role bindings.
This role does not allow viewing Secrets, since reading
the contents of Secrets enables access to ServiceAccount credentials
in the namespace, which would allow API access as any ServiceAccount
in the namespace (a form of privilege escalation).
Core component roles
Default ClusterRole
Default ClusterRoleBinding
Description
system:kube-scheduler
system:kube-scheduler user
Allows access to the resources required by the scheduler component.
system:volume-scheduler
system:kube-scheduler user
Allows access to the volume resources required by the kube-scheduler component.
system:kube-controller-manager
system:kube-controller-manager user
Allows access to the resources required by the controller manager component.
The permissions required by individual controllers are detailed in the controller roles.
system:node
None
Allows access to resources required by the kubelet, including read access to all secrets, and write access to all pod status objects.
You should use the Node authorizer and NodeRestriction admission plugin instead of the system:node role, and allow granting API access to kubelets based on the Pods scheduled to run on them.
The system:node role only exists for compatibility with Kubernetes clusters upgraded from versions prior to v1.8.
system:node-proxier
system:kube-proxy user
Allows access to the resources required by the kube-proxy component.
Other component roles
Default ClusterRole
Default ClusterRoleBinding
Description
system:auth-delegator
None
Allows delegated authentication and authorization checks.
This is commonly used by add-on API servers for unified authentication and authorization.
Allows read access to control-plane monitoring endpoints (i.e. kube-apiserver liveness and readiness endpoints (/healthz, /livez, /readyz), the individual health-check endpoints (/healthz/*, /livez/*, /readyz/*), and /metrics). Note that individual health check endpoints and the metric endpoint may expose sensitive information.
Roles for built-in controllers
The Kubernetes controller manager runs
controllers that are built in to the Kubernetes
control plane.
When invoked with --use-service-account-credentials, kube-controller-manager starts each controller
using a separate service account.
Corresponding roles exist for each built-in controller, prefixed with system:controller:.
If the controller manager is not started with --use-service-account-credentials, it runs all control loops
using its own credential, which must be granted all the relevant roles.
These roles include:
The RBAC API prevents users from escalating privileges by editing roles or role bindings.
Because this is enforced at the API level, it applies even when the RBAC authorizer is not in use.
Restrictions on role creation or update
You can only create/update a role if at least one of the following things is true:
You already have all the permissions contained in the role, at the same scope as the object being modified
(cluster-wide for a ClusterRole, within the same namespace or cluster-wide for a Role).
You are granted explicit permission to perform the escalate verb on the roles or clusterroles resource in the rbac.authorization.k8s.io API group.
For example, if user-1 does not have the ability to list Secrets cluster-wide, they cannot create a ClusterRole
containing that permission. To allow a user to create/update roles:
Grant them a role that allows them to create/update Role or ClusterRole objects, as desired.
Grant them permission to include specific permissions in the roles they create/update:
implicitly, by giving them those permissions (if they attempt to create or modify a Role or ClusterRole with permissions they themselves have not been granted, the API request will be forbidden)
or explicitly allow specifying any permission in a Role or ClusterRole by giving them permission to perform the escalate verb on roles or clusterroles resources in the rbac.authorization.k8s.io API group
Restrictions on role binding creation or update
You can only create/update a role binding if you already have all the permissions contained in the referenced role
(at the same scope as the role binding) or if you have been authorized to perform the bind verb on the referenced role.
For example, if user-1 does not have the ability to list Secrets cluster-wide, they cannot create a ClusterRoleBinding
to a role that grants that permission. To allow a user to create/update role bindings:
Grant them a role that allows them to create/update RoleBinding or ClusterRoleBinding objects, as desired.
Grant them permissions needed to bind a particular role:
implicitly, by giving them the permissions contained in the role.
explicitly, by giving them permission to perform the bind verb on the particular Role (or ClusterRole).
For example, this ClusterRole and RoleBinding would allow user-1 to grant other users the admin, edit, and view roles in the namespace user-1-namespace:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:name:role-grantorrules:- apiGroups:["rbac.authorization.k8s.io"]resources:["rolebindings"]verbs:["create"]- apiGroups:["rbac.authorization.k8s.io"]resources:["clusterroles"]verbs:["bind"]# omit resourceNames to allow binding any ClusterRoleresourceNames:["admin","edit","view"]---apiVersion:rbac.authorization.k8s.io/v1kind:RoleBindingmetadata:name:role-grantor-bindingnamespace:user-1-namespaceroleRef:apiGroup:rbac.authorization.k8s.iokind:ClusterRolename:role-grantorsubjects:- apiGroup:rbac.authorization.k8s.iokind:Username:user-1
When bootstrapping the first roles and role bindings, it is necessary for the initial user to grant permissions they do not yet have.
To bootstrap initial roles and role bindings:
Use a credential with the "system:masters" group, which is bound to the "cluster-admin" super-user role by the default bindings.
Command-line utilities
kubectl create role
Creates a Role object defining permissions within a single namespace. Examples:
Create a Role named "pod-reader" that allows users to perform get, watch and list on pods:
kubectl create role pod-reader --verb=get --verb=list --verb=watch --resource=pods
Create a Role named "pod-reader" with resourceNames specified:
kubectl create role pod-reader --verb=get --resource=pods --resource-name=readablepod --resource-name=anotherpod
Create a Role named "foo" with apiGroups specified:
kubectl create role foo --verb=get,list,watch --resource=replicasets.apps
Create a Role named "foo" with subresource permissions:
kubectl create role foo --verb=get,list,watch --resource=pods,pods/status
Create a Role named "my-component-lease-holder" with permissions to get/update a resource with a specific name:
kubectl create role my-component-lease-holder --verb=get,list,watch,update --resource=lease --resource-name=my-component
kubectl create clusterrole
Creates a ClusterRole. Examples:
Create a ClusterRole named "pod-reader" that allows user to perform get, watch and list on pods:
Default RBAC policies grant scoped permissions to control-plane components, nodes,
and controllers, but grant no permissions to service accounts outside the kube-system namespace
(beyond discovery permissions given to all authenticated users).
This allows you to grant particular roles to particular ServiceAccounts as needed.
Fine-grained role bindings provide greater security, but require more effort to administrate.
Broader grants can give unnecessary (and potentially escalating) API access to
ServiceAccounts, but are easier to administrate.
In order from most secure to least secure, the approaches are:
Grant a role to an application-specific service account (best practice)
This requires the application to specify a serviceAccountName in its pod spec,
and for the service account to be created (via the API, application manifest, kubectl create serviceaccount, etc.).
For example, grant read-only permission within "my-namespace" to the "my-sa" service account:
Many add-ons run as the
"default" service account in the kube-system namespace.
To allow those add-ons to run with super-user access, grant cluster-admin
permissions to the "default" service account in the kube-system namespace.
Caution: Enabling this means the kube-system namespace contains Secrets
that grant super-user access to your cluster's API.
Grant a role to all service accounts in a namespace
If you want all applications in a namespace to have a role, no matter what service account they use,
you can grant a role to the service account group for that namespace.
For example, grant read-only permission within "my-namespace" to all service accounts in that namespace:
Grant super-user access to all service accounts cluster-wide (strongly discouraged)
If you don't care about partitioning permissions at all, you can grant super-user access to all service accounts.
Warning: This allows any application full access to your cluster, and also grants
any user with read access to Secrets (or the ability to create any pod)
full access to your cluster.
Kubernetes clusters created before Kubernetes v1.22 include write access to
Endpoints in the aggregated "edit" and "admin" roles. As a mitigation for
CVE-2021-25740, this
access is not part of the aggregated roles in clusters that you create using
Kubernetes v1.22 or later.
Existing clusters that have been upgraded to Kubernetes v1.22 will not be
subject to this change. The CVE
announcement includes
guidance for restricting this access in existing clusters.
If you want new clusters to retain this level of access in the aggregated roles,
you can create the following ClusterRole:
apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRolemetadata:annotations:kubernetes.io/description:|- Add endpoints write permissions to the edit and admin roles. This was
removed by default in 1.22 because of CVE-2021-25740. See
https://issue.k8s.io/103675. This can allow writers to direct LoadBalancer
or Ingress implementations to expose backend IPs that would not otherwise
be accessible, and can circumvent network policies or security controls
intended to prevent/isolate access to those backends.labels:rbac.authorization.k8s.io/aggregate-to-edit:"true"name:custom:aggregate-to-edit:endpoints# you can change this if you wishrules:- apiGroups:[""]resources:["endpoints"]verbs:["create","delete","deletecollection","patch","update"]
Upgrading from ABAC
Clusters that originally ran older Kubernetes versions often used
permissive ABAC policies, including granting full API access to all
service accounts.
Default RBAC policies grant scoped permissions to control-plane components, nodes,
and controllers, but grant no permissions to service accounts outside the kube-system namespace
(beyond discovery permissions given to all authenticated users).
While far more secure, this can be disruptive to existing workloads expecting to automatically receive API permissions.
Here are two approaches for managing this transition:
Parallel authorizers
Run both the RBAC and ABAC authorizers, and specify a policy file that contains
the legacy ABAC policy:
To explain that first command line option in detail: if earlier authorizers, such as Node,
deny a request, then the RBAC authorizer attempts to authorize the API request. If RBAC
also denies that API request, the ABAC authorizer is then run. This means that any request
allowed by either the RBAC or ABAC policies is allowed.
When the kube-apiserver is run with a log level of 5 or higher for the RBAC component
(--vmodule=rbac*=5 or --v=5), you can see RBAC denials in the API server log
(prefixed with RBAC).
You can use that information to determine which roles need to be granted to which users, groups, or service accounts.
Once you have granted roles to service accounts and workloads
are running with no RBAC denial messages in the server logs, you can remove the ABAC authorizer.
Permissive RBAC permissions
You can replicate a permissive ABAC policy using RBAC role bindings.
Warning:
The following policy allows ALL service accounts to act as cluster administrators.
Any application running in a container receives service account credentials automatically,
and could perform any action against the API, including viewing secrets and modifying permissions.
This is not a recommended policy.
After you have transitioned to use RBAC, you should adjust the access controls
for your cluster to ensure that these meet your information security needs.
6.3.9 - Using ABAC Authorization
Attribute-based access control (ABAC) defines an access control paradigm whereby access rights are granted to users through the use of policies which combine attributes together.
Policy File Format
To enable ABAC mode, specify --authorization-policy-file=SOME_FILENAME and --authorization-mode=ABAC on startup.
The file format is one JSON object per line. There
should be no enclosing list or map, only one map per line.
Each line is a "policy object", where each such object is a map with the following
properties:
Versioning properties:
apiVersion, type string; valid values are "abac.authorization.kubernetes.io/v1beta1". Allows versioning and conversion of the policy format.
kind, type string: valid values are "Policy". Allows versioning and conversion of the policy format.
spec property set to a map with the following properties:
Subject-matching properties:
user, type string; the user-string from --token-auth-file. If you specify user, it must match the username of the authenticated user.
group, type string; if you specify group, it must match one of the groups of the authenticated user. system:authenticated matches all authenticated requests. system:unauthenticated matches all unauthenticated requests.
Resource-matching properties:
apiGroup, type string; an API group.
Ex: apps, networking.k8s.io
Wildcard: * matches all API groups.
namespace, type string; a namespace.
Ex: kube-system
Wildcard: * matches all resource requests.
resource, type string; a resource type
Ex: pods, deployments
Wildcard: * matches all resource requests.
Non-resource-matching properties:
nonResourcePath, type string; non-resource request paths.
Ex: /version or /apis
Wildcard:
* matches all non-resource requests.
/foo/* matches all subpaths of /foo/.
readonly, type boolean, when true, means that the Resource-matching policy only applies to get, list, and watch operations, Non-resource-matching policy only applies to get operation.
Note:
An unset property is the same as a property set to the zero value for its type
(e.g. empty string, 0, false). However, unset should be preferred for
readability.
In the future, policies may be expressed in a JSON format, and managed via a
REST interface.
Authorization Algorithm
A request has attributes which correspond to the properties of a policy object.
When a request is received, the attributes are determined. Unknown attributes
are set to the zero value of its type (e.g. empty string, 0, false).
A property set to "*" will match any value of the corresponding attribute.
The tuple of attributes is checked for a match against every policy in the
policy file. If at least one line matches the request attributes, then the
request is authorized (but may fail later validation).
To permit any authenticated user to do something, write a policy with the
group property set to "system:authenticated".
To permit any unauthenticated user to do something, write a policy with the
group property set to "system:unauthenticated".
To permit a user to do anything, write a policy with the apiGroup, namespace,
resource, and nonResourcePath properties set to "*".
Kubectl
Kubectl uses the /api and /apis endpoints of api-server to discover
served resource types, and validates objects sent to the API by create/update
operations using schema information located at /openapi/v2.
When using ABAC authorization, those special resources have to be explicitly
exposed via the nonResourcePath property in a policy (see examples below):
/api, /api/*, /apis, and /apis/* for API version negotiation.
/version for retrieving the server version via kubectl version.
/swaggerapi/* for create/update operations.
To inspect the HTTP calls involved in a specific kubectl operation you can turn
up the verbosity:
Creating a new namespace leads to the creation of a new service account in the following format:
system:serviceaccount:<namespace>:default
For example, if you wanted to grant the default service account (in the kube-system namespace) full
privilege to the API using ABAC, you would add this line to your policy file:
the ability to create TokenReviews and SubjectAccessReviews for delegated authentication/authorization checks
In future releases, the node authorizer may add or remove permissions to ensure kubelets
have the minimal set of permissions required to operate correctly.
In order to be authorized by the Node authorizer, kubelets must use a credential that identifies them as
being in the system:nodes group, with a username of system:node:<nodeName>.
This group and user name format match the identity created for each kubelet as part of
kubelet TLS bootstrapping.
The value of <nodeName>must match precisely the name of the node as registered by the kubelet. By default, this is the host name as provided by hostname, or overridden via the kubelet option--hostname-override. However, when using the --cloud-provider kubelet option, the specific hostname may be determined by the cloud provider, ignoring the local hostname and the --hostname-override option.
For specifics about how the kubelet determines the hostname, see the kubelet options reference.
To enable the Node authorizer, start the apiserver with --authorization-mode=Node.
To limit the API objects kubelets are able to write, enable the NodeRestriction admission plugin by starting the apiserver with --enable-admission-plugins=...,NodeRestriction,...
Migration considerations
Kubelets outside the system:nodes group
Kubelets outside the system:nodes group would not be authorized by the Node authorization mode,
and would need to continue to be authorized via whatever mechanism currently authorizes them.
The node admission plugin would not restrict requests from these kubelets.
Kubelets with undifferentiated usernames
In some deployments, kubelets have credentials that place them in the system:nodes group,
but do not identify the particular node they are associated with,
because they do not have a username in the system:node:... format.
These kubelets would not be authorized by the Node authorization mode,
and would need to continue to be authorized via whatever mechanism currently authorizes them.
The NodeRestriction admission plugin would ignore requests from these kubelets,
since the default node identifier implementation would not consider that a node identity.
Upgrades from previous versions using RBAC
Upgraded pre-1.7 clusters using RBAC will continue functioning as-is because the system:nodes group binding will already exist.
If a cluster admin wishes to start using the Node authorizer and NodeRestriction admission plugin
to limit node access to the API, that can be done non-disruptively:
Enable the Node authorization mode (--authorization-mode=Node,RBAC) and the NodeRestriction admission plugin
Ensure all kubelets' credentials conform to the group/username requirements
Audit apiserver logs to ensure the Node authorizer is not rejecting requests from kubelets (no persistent NODE DENY messages logged)
Delete the system:node cluster role binding
RBAC Node Permissions
In 1.6, the system:node cluster role was automatically bound to the system:nodes group when using the RBAC Authorization mode.
In 1.7, the automatic binding of the system:nodes group to the system:node role is deprecated
because the node authorizer accomplishes the same purpose with the benefit of additional restrictions
on secret and configmap access. If the Node and RBAC authorization modes are both enabled,
the automatic binding of the system:nodes group to the system:node role is not created in 1.7.
In 1.8, the binding will not be created at all.
When using RBAC, the system:node cluster role will continue to be created,
for compatibility with deployment methods that bind other users or groups to that role.
6.3.11 - Mapping PodSecurityPolicies to Pod Security Standards
The tables below enumerate the configuration parameters on
PodSecurityPolicy objects, whether the field mutates
and/or validates pods, and how the configuration values map to the
Pod Security Standards.
For each applicable parameter, the allowed values for the
Baseline and
Restricted profiles are listed.
Anything outside the allowed values for those profiles would fall under the
Privileged profile. "No opinion"
means all values are allowed under all Pod Security Standards.
Baseline: "runtime/default,"(Trailing comma to allow unset)
Restricted: "runtime/default"(No trailing comma)
localhost/* values are also permitted for both Baseline & Restricted.
6.3.12 - Webhook Mode
A WebHook is an HTTP callback: an HTTP POST that occurs when something happens; a simple event-notification via HTTP POST. A web application implementing WebHooks will POST a message to a URL when certain things happen.
When specified, mode Webhook causes Kubernetes to query an outside REST
service when determining user privileges.
Configuration File Format
Mode Webhook requires a file for HTTP configuration, specify by the
--authorization-webhook-config-file=SOME_FILENAME flag.
The configuration file uses the kubeconfig
file format. Within the file "users" refers to the API Server webhook and
"clusters" refers to the remote service.
A configuration example which uses HTTPS client auth:
# Kubernetes API versionapiVersion:v1# kind of the API objectkind:Config# clusters refers to the remote service.clusters:- name:name-of-remote-authz-servicecluster:# CA for verifying the remote service.certificate-authority:/path/to/ca.pem# URL of remote service to query. Must use 'https'. May not include parameters.server:https://authz.example.com/authorize# users refers to the API Server's webhook configuration.users:- name:name-of-api-serveruser:client-certificate:/path/to/cert.pem# cert for the webhook plugin to useclient-key:/path/to/key.pem # key matching the cert# kubeconfig files require a context. Provide one for the API Server.current-context:webhookcontexts:- context:cluster:name-of-remote-authz-serviceuser:name-of-api-servername:webhook
Request Payloads
When faced with an authorization decision, the API Server POSTs a JSON-
serialized authorization.k8s.io/v1beta1SubjectAccessReview object describing the
action. This object contains fields describing the user attempting to make the
request, and either details about the resource being accessed or requests
attributes.
Note that webhook API objects are subject to the same versioning compatibility rules
as other Kubernetes API objects. Implementers should be aware of looser
compatibility promises for beta objects and check the "apiVersion" field of the
request to ensure correct deserialization. Additionally, the API Server must
enable the authorization.k8s.io/v1beta1 API extensions group (--runtime-config=authorization.k8s.io/v1beta1=true).
The remote service is expected to fill the status field of
the request and respond to either allow or disallow access. The response body's
spec field is ignored and may be omitted. A permissive response would return:
The first method is preferred in most cases, and indicates the authorization
webhook does not allow, or has "no opinion" about the request, but if other
authorizers are configured, they are given a chance to allow the request.
If there are no other authorizers, or none of them allow the request, the
request is forbidden. The webhook would return:
{
"apiVersion": "authorization.k8s.io/v1beta1",
"kind": "SubjectAccessReview",
"status": {
"allowed": false,
"reason": "user does not have read access to the namespace" }
}
The second method denies immediately, short-circuiting evaluation by other
configured authorizers. This should only be used by webhooks that have
detailed knowledge of the full authorizer configuration of the cluster.
The webhook would return:
{
"apiVersion": "authorization.k8s.io/v1beta1",
"kind": "SubjectAccessReview",
"status": {
"allowed": false,
"denied": true,
"reason": "user does not have read access to the namespace" }
}
Non-resource paths include: /api, /apis, /metrics,
/logs, /debug, /healthz, /livez, /openapi/v2, /readyz, and
/version. Clients require access to /api, /api/*, /apis, /apis/*,
and /version to discover what resources and versions are present on the server.
Access to other non-resource paths can be disallowed without restricting access
to the REST api.
For further documentation refer to the authorization.v1beta1 API objects and
webhook.go.
6.3.13 - Kubelet authentication/authorization
Overview
A kubelet's HTTPS endpoint exposes APIs which give access to data of varying sensitivity,
and allow you to perform operations with varying levels of power on the node and within containers.
This document describes how to authenticate and authorize access to the kubelet's HTTPS endpoint.
Kubelet authentication
By default, requests to the kubelet's HTTPS endpoint that are not rejected by other configured
authentication methods are treated as anonymous requests, and given a username of system:anonymous
and a group of system:unauthenticated.
To disable anonymous access and send 401 Unauthorized responses to unauthenticated requests:
start the kubelet with the --anonymous-auth=false flag
To enable X509 client certificate authentication to the kubelet's HTTPS endpoint:
start the kubelet with the --client-ca-file flag, providing a CA bundle to verify client certificates with
start the apiserver with --kubelet-client-certificate and --kubelet-client-key flags
To enable API bearer tokens (including service account tokens) to be used to authenticate to the kubelet's HTTPS endpoint:
ensure the authentication.k8s.io/v1beta1 API group is enabled in the API server
start the kubelet with the --authentication-token-webhook and --kubeconfig flags
the kubelet calls the TokenReview API on the configured API server to determine user information from bearer tokens
Kubelet authorization
Any request that is successfully authenticated (including an anonymous request) is then authorized. The default authorization mode is AlwaysAllow, which allows all requests.
There are many possible reasons to subdivide access to the kubelet API:
anonymous auth is enabled, but anonymous users' ability to call the kubelet API should be limited
bearer token auth is enabled, but arbitrary API users' (like service accounts) ability to call the kubelet API should be limited
client certificate auth is enabled, but only some of the client certificates signed by the configured CA should be allowed to use the kubelet API
To subdivide access to the kubelet API, delegate authorization to the API server:
ensure the authorization.k8s.io/v1beta1 API group is enabled in the API server
start the kubelet with the --authorization-mode=Webhook and the --kubeconfig flags
the kubelet calls the SubjectAccessReview API on the configured API server to determine whether each request is authorized
The kubelet authorizes API requests using the same request attributes approach as the apiserver.
The verb is determined from the incoming request's HTTP verb:
HTTP verb
request verb
POST
create
GET, HEAD
get
PUT
update
PATCH
patch
DELETE
delete
The resource and subresource is determined from the incoming request's path:
Kubelet API
resource
subresource
/stats/*
nodes
stats
/metrics/*
nodes
metrics
/logs/*
nodes
log
/spec/*
nodes
spec
all others
nodes
proxy
The namespace and API group attributes are always an empty string, and
the resource name is always the name of the kubelet's Node API object.
When running in this mode, ensure the user identified by the --kubelet-client-certificate and --kubelet-client-key
flags passed to the apiserver is authorized for the following attributes:
verb=*, resource=nodes, subresource=proxy
verb=*, resource=nodes, subresource=stats
verb=*, resource=nodes, subresource=log
verb=*, resource=nodes, subresource=spec
verb=*, resource=nodes, subresource=metrics
6.3.14 - TLS bootstrapping
In a Kubernetes cluster, the components on the worker nodes - kubelet and kube-proxy - need to communicate with Kubernetes control plane components, specifically kube-apiserver.
In order to ensure that communication is kept private, not interfered with, and ensure that each component of the cluster is talking to another trusted component, we strongly
recommend using client TLS certificates on nodes.
The normal process of bootstrapping these components, especially worker nodes that need certificates so they can communicate safely with kube-apiserver,
can be a challenging process as it is often outside of the scope of Kubernetes and requires significant additional work.
This in turn, can make it challenging to initialize or scale a cluster.
In order to simplify the process, beginning in version 1.4, Kubernetes introduced a certificate request and signing API. The proposal can be
found here.
This document describes the process of node initialization, how to set up TLS client certificate bootstrapping for
kubelets, and how it works.
Initialization Process
When a worker node starts up, the kubelet does the following:
Look for its kubeconfig file
Retrieve the URL of the API server and credentials, normally a TLS key and signed certificate from the kubeconfig file
Attempt to communicate with the API server using the credentials.
Assuming that the kube-apiserver successfully validates the kubelet's credentials, it will treat the kubelet as a valid node, and begin to assign pods to it.
Note that the above process depends upon:
Existence of a key and certificate on the local host in the kubeconfig
The certificate having been signed by a Certificate Authority (CA) trusted by the kube-apiserver
All of the following are responsibilities of whoever sets up and manages the cluster:
Creating the CA key and certificate
Distributing the CA certificate to the control plane nodes, where kube-apiserver is running
Creating a key and certificate for each kubelet; strongly recommended to have a unique one, with a unique CN, for each kubelet
Signing the kubelet certificate using the CA key
Distributing the kubelet key and signed certificate to the specific node on which the kubelet is running
The TLS Bootstrapping described in this document is intended to simplify, and partially or even completely automate, steps 3 onwards, as these are the most common when initializing or scaling
a cluster.
Bootstrap Initialization
In the bootstrap initialization process, the following occurs:
kubelet begins
kubelet sees that it does not have a kubeconfig file
kubelet searches for and finds a bootstrap-kubeconfig file
kubelet reads its bootstrap file, retrieving the URL of the API server and a limited usage "token"
kubelet connects to the API server, authenticates using the token
kubelet now has limited credentials to create and retrieve a certificate signing request (CSR)
kubelet creates a CSR for itself with the signerName set to kubernetes.io/kube-apiserver-client-kubelet
CSR is approved in one of two ways:
If configured, kube-controller-manager automatically approves the CSR
If configured, an outside process, possibly a person, approves the CSR using the Kubernetes API or via kubectl
Certificate is created for the kubelet
Certificate is issued to the kubelet
kubelet retrieves the certificate
kubelet creates a proper kubeconfig with the key and signed certificate
kubelet begins normal operation
Optional: if configured, kubelet automatically requests renewal of the certificate when it is close to expiry
The renewed certificate is approved and issued, either automatically or manually, depending on configuration.
The rest of this document describes the necessary steps to configure TLS Bootstrapping, and its limitations.
Configuration
To configure for TLS bootstrapping and optional automatic approval, you must configure options on the following components:
kube-apiserver
kube-controller-manager
kubelet
in-cluster resources: ClusterRoleBinding and potentially ClusterRole
In addition, you need your Kubernetes Certificate Authority (CA).
Certificate Authority
As without bootstrapping, you will need a Certificate Authority (CA) key and certificate. As without bootstrapping, these will be used
to sign the kubelet certificate. As before, it is your responsibility to distribute them to control plane nodes.
For the purposes of this document, we will assume these have been distributed to control plane nodes at /var/lib/kubernetes/ca.pem (certificate) and /var/lib/kubernetes/ca-key.pem (key).
We will refer to these as "Kubernetes CA certificate and key".
All Kubernetes components that use these certificates - kubelet, kube-apiserver, kube-controller-manager - assume the key and certificate to be PEM-encoded.
kube-apiserver configuration
The kube-apiserver has several requirements to enable TLS bootstrapping:
Recognizing CA that signs the client certificate
Authenticating the bootstrapping kubelet to the system:bootstrappers group
Authorize the bootstrapping kubelet to create a certificate signing request (CSR)
Recognizing client certificates
This is normal for all client certificate authentication.
If not already set, add the --client-ca-file=FILENAME flag to the kube-apiserver command to enable
client certificate authentication, referencing a certificate authority bundle
containing the signing certificate, for example
--client-ca-file=/var/lib/kubernetes/ca.pem.
Initial bootstrap authentication
In order for the bootstrapping kubelet to connect to kube-apiserver and request a certificate, it must first authenticate to the server.
You can use any authenticator that can authenticate the kubelet.
While any authentication strategy can be used for the kubelet's initial
bootstrap credentials, the following two authenticators are recommended for ease
of provisioning.
Bootstrap tokens are a simpler and more easily managed method to authenticate kubelets, and do not require any additional flags when starting kube-apiserver.
Whichever method you choose, the requirement is that the kubelet be able to authenticate as a user with the rights to:
create and retrieve CSRs
be automatically approved to request node client certificates, if automatic approval is enabled.
A kubelet authenticating using bootstrap tokens is authenticated as a user in the group system:bootstrappers, which is the standard method to use.
As this feature matures, you
should ensure tokens are bound to a Role Based Access Control (RBAC) policy
which limits requests (using the bootstrap token) strictly to client
requests related to certificate provisioning. With RBAC in place, scoping the
tokens to a group allows for great flexibility. For example, you could disable a
particular bootstrap group's access when you are done provisioning the nodes.
Bootstrap tokens
Bootstrap tokens are described in detail here. These are tokens that are stored as secrets in the Kubernetes cluster,
and then issued to the individual kubelet. You can use a single token for an entire cluster, or issue one per worker node.
The process is two-fold:
Create a Kubernetes secret with the token ID, secret and scope(s).
Issue the token to the kubelet
From the kubelet's perspective, one token is like another and has no special meaning.
From the kube-apiserver's perspective, however, the bootstrap token is special. Due to its type, namespace and name, kube-apiserver recognizes it as a special token,
and grants anyone authenticating with that token special bootstrap rights, notably treating them as a member of the system:bootstrappers group. This fulfills a basic requirement
for TLS bootstrapping.
The details for creating the secret are available here.
If you want to use bootstrap tokens, you must enable it on kube-apiserver with the flag:
--enable-bootstrap-token-auth=true
Token authentication file
kube-apiserver has the ability to accept tokens as authentication.
These tokens are arbitrary but should represent at least 128 bits of entropy derived
from a secure random number generator (such as /dev/urandom on most modern Linux
systems). There are multiple ways you can generate a token. For example:
head -c 16 /dev/urandom | od -An -t x | tr -d ' '
will generate tokens that look like 02b50b05283e98dd0fd71db496ef01e8.
The token file should look like the following example, where the first three
values can be anything and the quoted group name should be as depicted:
Add the --token-auth-file=FILENAME flag to the kube-apiserver command (in your
systemd unit file perhaps) to enable the token file. See docs
here for
further details.
Authorize kubelet to create CSR
Now that the bootstrapping node is authenticated as part of the
system:bootstrappers group, it needs to be authorized to create a
certificate signing request (CSR) as well as retrieve it when done.
Fortunately, Kubernetes ships with a ClusterRole with precisely these (and
only these) permissions, system:node-bootstrapper.
To do this, you only need to create a ClusterRoleBinding that binds the system:bootstrappers group to the cluster role system:node-bootstrapper.
# enable bootstrapping nodes to create CSRapiVersion:rbac.authorization.k8s.io/v1kind:ClusterRoleBindingmetadata:name:create-csrs-for-bootstrappingsubjects:- kind:Groupname:system:bootstrappersapiGroup:rbac.authorization.k8s.ioroleRef:kind:ClusterRolename:system:node-bootstrapperapiGroup:rbac.authorization.k8s.io
kube-controller-manager configuration
While the apiserver receives the requests for certificates from the kubelet and authenticates those requests,
the controller-manager is responsible for issuing actual signed certificates.
The controller-manager performs this function via a certificate-issuing control loop.
This takes the form of a
cfssl local signer using
assets on disk. Currently, all certificates issued have one year validity and a
default set of key usages.
In order for the controller-manager to sign certificates, it needs the following:
access to the "Kubernetes CA key and certificate" that you created and distributed
enabling CSR signing
Access to key and certificate
As described earlier, you need to create a Kubernetes CA key and certificate, and distribute it to the control plane nodes.
These will be used by the controller-manager to sign the kubelet certificates.
Since these signed certificates will, in turn, be used by the kubelet to authenticate as a regular kubelet to kube-apiserver, it is important that the CA
provided to the controller-manager at this stage also be trusted by kube-apiserver for authentication. This is provided to kube-apiserver
with the flag --client-ca-file=FILENAME (for example, --client-ca-file=/var/lib/kubernetes/ca.pem), as described in the kube-apiserver configuration section.
To provide the Kubernetes CA key and certificate to kube-controller-manager, use the following flags:
The validity duration of signed certificates can be configured with flag:
--cluster-signing-duration
Approval
In order to approve CSRs, you need to tell the controller-manager that it is acceptable to approve them. This is done by granting
RBAC permissions to the correct group.
There are two distinct sets of permissions:
nodeclient: If a node is creating a new certificate for a node, then it does not have a certificate yet. It is authenticating using one of the tokens listed above, and thus is part of the group system:bootstrappers.
selfnodeclient: If a node is renewing its certificate, then it already has a certificate (by definition), which it uses continuously to authenticate as part of the group system:nodes.
To enable the kubelet to request and receive a new certificate, create a ClusterRoleBinding that binds the group in which the bootstrapping node is a member system:bootstrappers to the ClusterRole that grants it permission, system:certificates.k8s.io:certificatesigningrequests:nodeclient:
# Approve all CSRs for the group "system:bootstrappers"apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRoleBindingmetadata:name:auto-approve-csrs-for-groupsubjects:- kind:Groupname:system:bootstrappersapiGroup:rbac.authorization.k8s.ioroleRef:kind:ClusterRolename:system:certificates.k8s.io:certificatesigningrequests:nodeclientapiGroup:rbac.authorization.k8s.io
To enable the kubelet to renew its own client certificate, create a ClusterRoleBinding that binds the group in which the fully functioning node is a member system:nodes to the ClusterRole that
grants it permission, system:certificates.k8s.io:certificatesigningrequests:selfnodeclient:
# Approve renewal CSRs for the group "system:nodes"apiVersion:rbac.authorization.k8s.io/v1kind:ClusterRoleBindingmetadata:name:auto-approve-renewals-for-nodessubjects:- kind:Groupname:system:nodesapiGroup:rbac.authorization.k8s.ioroleRef:kind:ClusterRolename:system:certificates.k8s.io:certificatesigningrequests:selfnodeclientapiGroup:rbac.authorization.k8s.io
The csrapproving controller that ships as part of
kube-controller-manager and is enabled
by default. The controller uses the
SubjectAccessReview API to
determine if a given user is authorized to request a CSR, then approves based on
the authorization outcome. To prevent conflicts with other approvers, the
builtin approver doesn't explicitly deny CSRs. It only ignores unauthorized
requests. The controller also prunes expired certificates as part of garbage
collection.
kubelet configuration
Finally, with the control plane nodes properly set up and all of the necessary authentication and authorization in place, we can configure the kubelet.
The kubelet requires the following configuration to bootstrap:
A path to store the key and certificate it generates (optional, can use default)
A path to a kubeconfig file that does not yet exist; it will place the bootstrapped config file here
A path to a bootstrap kubeconfig file to provide the URL for the server and bootstrap credentials, e.g. a bootstrap token
Optional: instructions to rotate certificates
The bootstrap kubeconfig should be in a path available to the kubelet, for example /var/lib/kubelet/bootstrap-kubeconfig.
Its format is identical to a normal kubeconfig file. A sample file might look as follows:
certificate-authority: path to a CA file, used to validate the server certificate presented by kube-apiserver
server: URL to kube-apiserver
token: the token to use
The format of the token does not matter, as long as it matches what kube-apiserver expects. In the above example, we used a bootstrap token.
As stated earlier, any valid authentication method can be used, not only tokens.
Because the bootstrap kubeconfigis a standard kubeconfig, you can use kubectl to generate it. To create the above example file:
When starting the kubelet, if the file specified via --kubeconfig does not
exist, the bootstrap kubeconfig specified via --bootstrap-kubeconfig is used
to request a client certificate from the API server. On approval of the
certificate request and receipt back by the kubelet, a kubeconfig file
referencing the generated key and obtained certificate is written to the path
specified by --kubeconfig. The certificate and key file will be placed in the
directory specified by --cert-dir.
Client and Serving Certificates
All of the above relate to kubelet client certificates, specifically, the certificates a kubelet
uses to authenticate to kube-apiserver.
A kubelet also can use serving certificates. The kubelet itself exposes an https endpoint for certain features.
To secure these, the kubelet can do one of:
use provided key and certificate, via the --tls-private-key-file and --tls-cert-file flags
create self-signed key and certificate, if a key and certificate are not provided
request serving certificates from the cluster server, via the CSR API
The client certificate provided by TLS bootstrapping is signed, by default, for client auth only, and thus cannot
be used as serving certificates, or server auth.
However, you can enable its server certificate, at least partially, via certificate rotation.
Certificate Rotation
Kubernetes v1.8 and higher kubelet implements beta features for enabling
rotation of its client and/or serving certificates. These can be enabled through
the respective RotateKubeletClientCertificate and
RotateKubeletServerCertificate feature flags on the kubelet and are enabled by
default.
RotateKubeletClientCertificate causes the kubelet to rotate its client
certificates by creating new CSRs as its existing credentials expire. To enable
this feature pass the following flag to the kubelet:
--rotate-certificates
RotateKubeletServerCertificate causes the kubelet both to request a serving
certificate after bootstrapping its client credentials and to rotate that
certificate. To enable this feature pass the following flag to the kubelet:
--rotate-server-certificates
Note:
The CSR approving controllers implemented in core Kubernetes do not
approve node serving certificates for security
reasons. To use
RotateKubeletServerCertificate operators need to run a custom approving
controller, or manually approve the serving certificate requests.
A deployment-specific approval process for kubelet serving certificates should typically only approve CSRs which:
are requested by nodes (ensure the spec.username field is of the form
system:node:<nodeName> and spec.groups contains system:nodes)
request usages for a serving certificate (ensure spec.usages contains server auth,
optionally contains digital signature and key encipherment, and contains no other usages)
only have IP and DNS subjectAltNames that belong to the requesting node,
and have no URI and Email subjectAltNames (parse the x509 Certificate Signing Request
in spec.request to verify subjectAltNames)
Other authenticating components
All of TLS bootstrapping described in this document relates to the kubelet. However,
other components may need to communicate directly with kube-apiserver. Notable is kube-proxy, which
is part of the Kubernetes control plane and runs on every node, but may also include other components such as monitoring or networking.
Like the kubelet, these other components also require a method of authenticating to kube-apiserver.
You have several options for generating these credentials:
The old way: Create and distribute certificates the same way you did for kubelet before TLS bootstrapping
DaemonSet: Since the kubelet itself is loaded on each node, and is sufficient to start base services, you can run kube-proxy and other node-specific services not as a standalone process, but rather as a daemonset in the kube-system namespace. Since it will be in-cluster, you can give it a proper service account with appropriate permissions to perform its activities. This may be the simplest way to configure such services.
kubectl approval
CSRs can be approved outside of the approval flows builtin to the controller
manager.
The signing controller does not immediately sign all certificate requests.
Instead, it waits until they have been flagged with an "Approved" status by an
appropriately-privileged user. This flow is intended to allow for automated
approval handled by an external approval controller or the approval controller
implemented in the core controller-manager. However cluster administrators can
also manually approve certificate requests using kubectl. An administrator can
list CSRs with kubectl get csr and describe one in detail with kubectl describe csr <name>. An administrator can approve or deny a CSR with kubectl certificate approve <name> and kubectl certificate deny <name>.
6.4 - Well-Known Labels, Annotations and Taints
Kubernetes reserves all labels and annotations in the kubernetes.io namespace.
This document serves both as a reference to the values and as a coordination point for assigning values.
Labels, annotations and taints used on API objects
The Kubelet populates this with runtime.GOARCH as defined by Go. This can be handy if you are mixing arm and x86 nodes.
kubernetes.io/os
Example: kubernetes.io/os: "linux"
Used on: Node
The Kubelet populates this with runtime.GOOS as defined by Go. This can be handy if you are mixing operating systems in your cluster (for example: mixing Linux and Windows nodes).
The Kubernetes API server (part of the control plane)
sets this label on all namespaces. The label value is set
to the name of the namespace. You can't change this label's value.
This is useful if you want to target a specific namespace with a label
selector.
beta.kubernetes.io/arch (deprecated)
This label has been deprecated. Please use kubernetes.io/arch instead.
beta.kubernetes.io/os (deprecated)
This label has been deprecated. Please use kubernetes.io/os instead.
The Kubelet populates this label with the hostname. Note that the hostname can be changed from the "actual" hostname by passing the --hostname-override flag to the kubelet.
This label is also used as part of the topology hierarchy. See topology.kubernetes.io/zone for more information.
The value for this annotation must be true to take effect. This annotation indicates that pods running as this service account may only reference Secret API objects specified in the service account's secrets field.
This annotation is used to set Pod Deletion Cost
which allows users to influence ReplicaSet downscaling order. The annotation parses into an int32 type.
This annotation controls whether a DaemonSet pod should be evicted by a ClusterAutoscaler.
This annotation needs to be specified on DaemonSet pods in a DaemonSet manifest.
When this annotation is set to "true", the ClusterAutoscaler is allowed to evict a DaemonSet Pod,
even if other rules would normally prevent that. To disallow the ClusterAutoscaler from evicting DaemonSet pods,
you can set this annotation to "false" for important DaemonSet pods.
If this annotation is not set, then the Cluster Autoscaler follows its overall behaviour (i.e evict the DaemonSets based on its configuration).
Note: This annotation only impacts DaemonSet pods.
kubernetes.io/ingress-bandwidth
Note: Ingress traffic shaping annotation is an experimental feature.
If you want to enable traffic shaping support, you must add the bandwidth plugin to your CNI configuration file (default /etc/cni/net.d) and
ensure that the binary is included in your CNI bin dir (default /opt/cni/bin).
Example: kubernetes.io/ingress-bandwidth: 10M
Used on: Pod
You can apply quality-of-service traffic shaping to a pod and effectively limit its available bandwidth.
Ingress traffic (to the pod) is handled by shaping queued packets to effectively handle data.
To limit the bandwidth on a pod, write an object definition JSON file and specify the data traffic
speed using kubernetes.io/ingress-bandwidth annotation. The unit used for specifying ingress
rate is bits per second, as a Quantity.
For example, 10M means 10 megabits per second.
kubernetes.io/egress-bandwidth
Note: Egress traffic shaping annotation is an experimental feature.
If you want to enable traffic shaping support, you must add the bandwidth plugin to your CNI configuration file (default /etc/cni/net.d) and
ensure that the binary is included in your CNI bin dir (default /opt/cni/bin).
Example: kubernetes.io/egress-bandwidth: 10M
Used on: Pod
Egress traffic (from the pod) is handled by policing, which simply drops packets in excess of the configured rate.
The limits you place on a pod do not affect the bandwidth of other pods.
To limit the bandwidth on a pod, write an object definition JSON file and specify the data traffic
speed using kubernetes.io/egress-bandwidth annotation. The unit used for specifying egress
rate is bits per second, as a Quantity.
For example, 10M means 10 megabits per second.
The Kubelet populates this with the instance type as defined by the cloudprovider.
This will be set only if you are using a cloudprovider. This setting is handy
if you want to target certain workloads to certain instance types, but typically you want
to rely on the Kubernetes scheduler to perform resource-based scheduling. You should aim to schedule based on properties rather than on instance types (for example: require a GPU, instead of requiring a g2.2xlarge).
When a StatefulSet controller creates a Pod for the StatefulSet, the control plane
sets this label on that Pod. The value of the label is the name of the Pod being created.
See Pod Name Label in the
StatefulSet topic for more details.
On Node: The kubelet or the external cloud-controller-manager populates this with the information as provided by the cloudprovider. This will be set only if you are using a cloudprovider. However, you should consider setting this on nodes if it makes sense in your topology.
On PersistentVolume: topology-aware volume provisioners will automatically set node affinity constraints on PersistentVolumes.
A zone represents a logical failure domain. It is common for Kubernetes clusters to span multiple zones for increased availability. While the exact definition of a zone is left to infrastructure implementations, common properties of a zone include very low network latency within a zone, no-cost network traffic within a zone, and failure independence from other zones. For example, nodes within a zone might share a network switch, but nodes in different zones should not.
A region represents a larger domain, made up of one or more zones. It is uncommon for Kubernetes clusters to span multiple regions, While the exact definition of a zone or region is left to infrastructure implementations, common properties of a region include higher network latency between them than within them, non-zero cost for network traffic between them, and failure independence from other zones or regions. For example, nodes within a region might share power infrastructure (e.g. a UPS or generator), but nodes in different regions typically would not.
Kubernetes makes a few assumptions about the structure of zones and regions:
regions and zones are hierarchical: zones are strict subsets of regions and no zone can be in 2 regions
zone names are unique across regions; for example region "africa-east-1" might be comprised of zones "africa-east-1a" and "africa-east-1b"
It should be safe to assume that topology labels do not change. Even though labels are strictly mutable, consumers of them can assume that a given node is not going to be moved between zones without being destroyed and recreated.
Kubernetes can use this information in various ways. For example, the scheduler automatically tries to spread the Pods in a ReplicaSet across nodes in a single-zone cluster (to reduce the impact of node failures, see kubernetes.io/hostname). With multiple-zone clusters, this spreading behavior also applies to zones (to reduce the impact of zone failures). This is achieved via SelectorSpreadPriority.
SelectorSpreadPriority is a best effort placement. If the zones in your cluster are heterogeneous (for example: different numbers of nodes, different types of nodes, or different pod resource requirements), this placement might prevent equal spreading of your Pods across zones. If desired, you can use homogenous zones (same number and types of nodes) to reduce the probability of unequal spreading.
The scheduler (through the VolumeZonePredicate predicate) also will ensure that Pods, that claim a given volume, are only placed into the same zone as that volume. Volumes cannot be attached across zones.
If PersistentVolumeLabel does not support automatic labeling of your PersistentVolumes, you should consider
adding the labels manually (or adding support for PersistentVolumeLabel). With PersistentVolumeLabel, the scheduler prevents Pods from mounting volumes in a different zone. If your infrastructure doesn't have this constraint, you don't need to add the zone labels to the volumes at all.
This annotation records the unique ID of the
ServiceAccount that the token (stored in the Secret of type kubernetes.io/service-account-token) represents.
The label is used to indicate the controller or entity that manages an EndpointSlice. This label aims to enable different EndpointSlice objects to be managed by different controllers or entities within the same cluster.
The label can be set to "true" on an Endpoints resource to indicate that the EndpointSliceMirroring controller should not mirror this resource with EndpointSlices.
The annotation is used to run Windows containers with Hyper-V isolation. To use Hyper-V isolation feature and create a Hyper-V isolated container, the kubelet should be started with feature gates HyperVContainer=true and the Pod should include the annotation experimental.windows.kubernetes.io/isolation-type: hyperv.
Note: You can only set this annotation on Pods that have a single container.
Starting from v1.20, this annotation is deprecated. Experimental Hyper-V support was removed in 1.21.
When a single IngressClass resource has this annotation set to "true", new Ingress resource without a class specified will be assigned this default class.
kubernetes.io/ingress.class (deprecated)
Note: Starting in v1.18, this annotation is deprecated in favor of spec.ingressClassName.
When a single StorageClass resource has this annotation set to "true", new PersistentVolumeClaim
resource without a class specified will be assigned this default class.
The kubelet can set this annotation on a Node to denote its configured IPv4 address.
When kubelet is started with the --cloud-provider flag set to any value (includes both external and legacy in-tree cloud providers), it sets this annotation on the Node to denote an IP address set from the command line flag (--node-ip). This IP is verified with the cloud provider as valid by the cloud-controller-manager.
The value of the annotation is the container name that is default for this Pod. For example, kubectl logs or kubectl exec without -c or --container flag will use this default container.
In Kubernetes clusters v1.22 (or later), the Endpoints controller adds this annotation to an Endpoints resource if it has more than 1000 endpoints. The annotation indicates that the Endpoints resource is over capacity and the number of endpoints has been truncated to 1000.
batch.kubernetes.io/job-tracking
Example: batch.kubernetes.io/job-tracking: ""
Used on: Jobs
The presence of this annotation on a Job indicates that the control plane is
tracking the Job status using finalizers.
You should not manually add or remove this annotation.
The kubelet detects memory pressure based on memory.available and allocatableMemory.available observed on a Node. The observed values are then compared to the corresponding thresholds that can be set on the kubelet to determine if the Node condition and taint should be added/removed.
The kubelet detects disk pressure based on imagefs.available, imagefs.inodesFree, nodefs.available and nodefs.inodesFree(Linux only) observed on a Node. The observed values are then compared to the corresponding thresholds that can be set on the kubelet to determine if the Node condition and taint should be added/removed.
This is initially set by the kubelet when the cloud provider used indicates a requirement for additional network configuration. Only when the route on the cloud is configured properly will the taint be removed by the cloud provider.
The kubelet checks D-value of the size of /proc/sys/kernel/pid_max and the PIDs consumed by Kubernetes on a node to get the number of available PIDs that referred to as the pid.available metric. The metric is then compared to the corresponding threshold that can be set on the kubelet to determine if the node condition and taint should be added/removed.
A user can manually add the taint to a Node marking it out-of-service. If the NodeOutOfServiceVolumeDetachfeature gate is enabled on
kube-controller-manager, and a Node is marked out-of-service with this taint, the pods on the node will be forcefully deleted if there are no matching tolerations on it and volume detach operations for the pods terminating on the node will happen immediately. This allows the Pods on the out-of-service node to recover quickly on a different node.
Sets this taint on a node to mark it as unusable, when kubelet is started with the "external" cloud provider, until a controller from the cloud-controller-manager initializes this node, and then removes the taint.
If a Node is in a cloud provider specified shutdown state, the Node gets tainted accordingly with node.cloudprovider.kubernetes.io/shutdown and the taint effect of NoSchedule.
Value must be one of privileged, baseline, or restricted which correspond to
Pod Security Standard levels. Specifically,
the enforce label prohibits the creation of any Pod in the labeled Namespace which does not meet
the requirements outlined in the indicated level.
Value must be latest or a valid Kubernetes version in the format v<MAJOR>.<MINOR>.
This determines the version of the Pod Security Standard
policies to apply when validating a submitted Pod.
Value must be one of privileged, baseline, or restricted which correspond to
Pod Security Standard levels. Specifically,
the audit label does not prevent the creation of a Pod in the labeled Namespace which does not meet
the requirements outlined in the indicated level, but adds an audit annotation to that Pod.
Value must be latest or a valid Kubernetes version in the format v<MAJOR>.<MINOR>.
This determines the version of the Pod Security Standard
policies to apply when validating a submitted Pod.
Value must be one of privileged, baseline, or restricted which correspond to
Pod Security Standard levels. Specifically,
the warn label does not prevent the creation of a Pod in the labeled Namespace which does not meet the
requirements outlined in the indicated level, but returns a warning to the user after doing so.
Note that warnings are also displayed when creating or updating objects that contain Pod templates,
such as Deployments, Jobs, StatefulSets, etc.
Value must be latest or a valid Kubernetes version in the format v<MAJOR>.<MINOR>.
This determines the version of the Pod Security Standard
policies to apply when validating a submitted Pod. Note that warnings are also displayed when creating
or updating objects that contain Pod templates, such as Deployments, Jobs, StatefulSets, etc.
This annotation was only relevant if you were using PodSecurityPolicies.
Kubernetes v1.25 does not support the PodSecurityPolicy API.
When the PodSecurityPolicy admission controller admitted a Pod, the admission controller
modified the Pod to have this annotation.
The value of the annotation was the name of the PodSecurityPolicy that was used for validation.
This annotation has been deprecated since Kubernetes v1.19 and will become non-functional in a future release.
please use the corresponding pod or container securityContext.seccompProfile field instead.
To specify security settings for a Pod, include the securityContext field in the Pod specification.
The securityContext field within a Pod's .spec defines pod-level security attributes.
When you specify the security context for a Pod,
the settings you specify apply to all containers in that Pod.
This annotation has been deprecated since Kubernetes v1.19 and will become non-functional in a future release.
please use the corresponding pod or container securityContext.seccompProfile field instead.
The tutorial Restrict a Container's Syscalls with seccomp takes
you through the steps you follow to apply a seccomp profile to a Pod or to one of
its containers. That tutorial covers the supported mechanism for configuring seccomp in Kubernetes,
based on setting securityContext within the Pod's .spec.
Value can either be true or false.
This determines whether a user can modify the mode of the source volume when a
PersistentVolumeClaim is being
created from a VolumeSnapshot.
Annotation that kubeadm uses to preserve the CRI socket information given to kubeadm at init/join time for later use.
kubeadm annotates the Node object with this information. The annotation remains "alpha", since ideally this should
be a field in KubeletConfiguration instead.
Annotation that kubeadm places on locally managed etcd pods to keep track of a list of URLs where etcd clients
should connect to. This is used mainly for etcd cluster health check purposes.
Annotation that kubeadm places on locally managed kube-apiserver pods to keep track of the exposed advertise
address/port endpoint for that API server instance.
Annotation that kubeadm places on ConfigMaps that it manages for configuring components. It contains a hash (SHA-256)
used to determine if the user has applied settings different from the kubeadm defaults for a particular component.
node-role.kubernetes.io/control-plane
Used on: Node
Label that kubeadm applies on the control plane nodes that it manages.
Taint that kubeadm applies on control plane nodes to allow only critical workloads to schedule on them.
6.4.1 - Audit Annotations
This page serves as a reference for the audit annotations of the kubernetes.io
namespace. These annotations apply to Event object from API group
audit.k8s.io.
Note: The following annotations are not used within the Kubernetes API. When you
enable auditing in your cluster,
audit event data is written using Event from API group audit.k8s.io.
The annotations apply to audit events. Audit events are different from objects in the
Event API (API group
events.k8s.io).
Value must be one of user, namespace, or runtimeClass which correspond to
Pod Security Exemption
dimensions. This annotation indicates on which dimension was based the exemption
from the PodSecurity enforcement.
Value must be privileged:<version>, baseline:<version>,
restricted:<version> which correspond to Pod Security
Standard levels accompanied by
a version which must be latest or a valid Kubernetes version in the format
v<MAJOR>.<MINOR>. This annotations informs about the enforcement level that
allowed or denied the pod during PodSecurity admission.
Example: pod-security.kubernetes.io/audit-violations: would violate PodSecurity "restricted:latest": allowPrivilegeEscalation != false (container "example" must set securityContext.allowPrivilegeEscalation=false), ...
Value details an audit policy violation, it contains the
Pod Security Standard level
that was transgressed as well as the specific policies on the fields that were
violated from the PodSecurity enforcement.
Example: missing-san.invalid-cert.kubernetes.io/example-svc.example-namespace.svc: "relies on a legacy Common Name field instead of the SAN extension for subject validation"
Used by Kubernetes version v1.24 and later
This annotation indicates a webhook or aggregated API server
is using an invalid certificate that is missing subjectAltNames.
Support for these certificates was disabled by default in Kubernetes 1.19,
and removed in Kubernetes 1.23.
Requests to endpoints using these certificates will fail.
Services using these certificates should replace them as soon as possible
to avoid disruption when running in Kubernetes 1.23+ environments.
Example: insecure-sha1.invalid-cert.kubernetes.io/example-svc.example-namespace.svc: "uses an insecure SHA-1 signature"
Used by Kubernetes version v1.24 and later
This annotation indicates a webhook or aggregated API server
is using an insecure certificate signed with a SHA-1 hash.
Support for these insecure certificates is disabled by default in Kubernetes 1.24,
and will be removed in a future release.
Services using these certificates should replace them as soon as possible,
to ensure connections are secured properly and to avoid disruption in future releases.
Kubernetes' API is the application that serves Kubernetes functionality through a RESTful interface and stores the state of the cluster.
Kubernetes resources and "records of intent" are all stored as API objects, and modified via RESTful calls to the API. The API allows configuration to be managed in a declarative way. Users can interact with the Kubernetes API directly, or via tools like kubectl. The core Kubernetes API is flexible and can also be extended to support custom resources.
6.5.1 - Workload Resources
6.5.1.1 - Pod
Pod is a collection of containers that can run on a host.
apiVersion: v1
import "k8s.io/api/core/v1"
Pod
Pod is a collection of containers that can run on a host. This resource is created by clients and scheduled onto hosts.
List of containers belonging to the pod. Containers cannot currently be added or removed. There must be at least one container in a Pod. Cannot be updated.
List of initialization containers belonging to the pod. Init containers are executed in order prior to containers being started. If any init container fails, the pod is considered to have failed and is handled according to its restartPolicy. The name for an init container or normal container must be unique among all containers. Init containers may not have Lifecycle actions, Readiness probes, Liveness probes, or Startup probes. The resourceRequirements of an init container are taken into account during scheduling by finding the highest request/limit for each resource type, and then using the max of of that value or the sum of the normal containers. Limits are applied to init containers in a similar fashion. Init containers cannot currently be added or removed. Cannot be updated. More info: https://kubernetes.io/docs/concepts/workloads/pods/init-containers/
List of ephemeral containers run in this pod. Ephemeral containers may be run in an existing pod to perform user-initiated actions such as debugging. This list cannot be specified when creating a pod, and it cannot be modified by updating the pod spec. In order to add an ephemeral container to an existing pod, use the pod's ephemeralcontainers subresource.
EnableServiceLinks indicates whether information about services should be injected into pod's environment variables, matching the syntax of Docker links. Optional: Defaults to true.
os (PodOS)
Specifies the OS of the containers in the pod. Some pod and container fields are restricted if this is set.
If the OS field is set to linux, the following fields must be unset: -securityContext.windowsOptions
If the OS field is set to windows, following fields must be unset: - spec.hostPID - spec.hostIPC - spec.hostUsers - spec.securityContext.seLinuxOptions - spec.securityContext.seccompProfile - spec.securityContext.fsGroup - spec.securityContext.fsGroupChangePolicy - spec.securityContext.sysctls - spec.shareProcessNamespace - spec.securityContext.runAsUser - spec.securityContext.runAsGroup - spec.securityContext.supplementalGroups - spec.containers[].securityContext.seLinuxOptions - spec.containers[].securityContext.seccompProfile - spec.containers[].securityContext.capabilities - spec.containers[].securityContext.readOnlyRootFilesystem - spec.containers[].securityContext.privileged - spec.containers[].securityContext.allowPrivilegeEscalation - spec.containers[].securityContext.procMount - spec.containers[].securityContext.runAsUser - spec.containers[*].securityContext.runAsGroup
NodeName is a request to schedule this pod onto a specific node. If it is non-empty, the scheduler simply schedules this pod onto that node, assuming that it fits resource requirements.
Describes pod anti-affinity scheduling rules (e.g. avoid putting this pod in the same node, zone, etc. as some other pod(s)).
tolerations ([]Toleration)
If specified, the pod's tolerations.
The pod this Toleration is attached to tolerates any taint that matches the triple <key,value,effect> using the matching operator .
tolerations.key (string)
Key is the taint key that the toleration applies to. Empty means match all taint keys. If the key is empty, operator must be Exists; this combination means to match all values and all keys.
tolerations.operator (string)
Operator represents a key's relationship to the value. Valid operators are Exists and Equal. Defaults to Equal. Exists is equivalent to wildcard for value, so that a pod can tolerate all taints of a particular category.
tolerations.value (string)
Value is the taint value the toleration matches to. If the operator is Exists, the value should be empty, otherwise just a regular string.
tolerations.effect (string)
Effect indicates the taint effect to match. Empty means match all taint effects. When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
tolerations.tolerationSeconds (int64)
TolerationSeconds represents the period of time the toleration (which must be of effect NoExecute, otherwise this field is ignored) tolerates the taint. By default, it is not set, which means tolerate the taint forever (do not evict). Zero and negative values will be treated as 0 (evict immediately) by the system.
schedulerName (string)
If specified, the pod will be dispatched by specified scheduler. If not specified, the pod will be dispatched by default scheduler.
runtimeClassName (string)
RuntimeClassName refers to a RuntimeClass object in the node.k8s.io group, which should be used to run this pod. If no RuntimeClass resource matches the named class, the pod will not be run. If unset or empty, the "legacy" RuntimeClass will be used, which is an implicit class with an empty definition that uses the default runtime handler. More info: https://git.k8s.io/enhancements/keps/sig-node/585-runtime-class
priorityClassName (string)
If specified, indicates the pod's priority. "system-node-critical" and "system-cluster-critical" are two special keywords which indicate the highest priorities with the former being the highest priority. Any other name must be defined by creating a PriorityClass object with that name. If not specified, the pod priority will be default or zero if there is no default.
priority (int32)
The priority value. Various system components use this field to find the priority of the pod. When Priority Admission Controller is enabled, it prevents users from setting this field. The admission controller populates this field from PriorityClassName. The higher the value, the higher the priority.
preemptionPolicy (string)
PreemptionPolicy is the Policy for preempting pods with lower priority. One of Never, PreemptLowerPriority. Defaults to PreemptLowerPriority if unset.
Map: unique values on keys topologyKey, whenUnsatisfiable will be kept during a merge
TopologySpreadConstraints describes how a group of pods ought to spread across topology domains. Scheduler will schedule pods in a way which abides by the constraints. All topologySpreadConstraints are ANDed.
TopologySpreadConstraint specifies how to spread matching pods among the given topology.
MaxSkew describes the degree to which pods may be unevenly distributed. When whenUnsatisfiable=DoNotSchedule, it is the maximum permitted difference between the number of matching pods in the target topology and the global minimum. The global minimum is the minimum number of matching pods in an eligible domain or zero if the number of eligible domains is less than MinDomains. For example, in a 3-zone cluster, MaxSkew is set to 1, and pods with the same labelSelector spread as 2/2/1: In this case, the global minimum is 1. | zone1 | zone2 | zone3 | | P P | P P | P | - if MaxSkew is 1, incoming pod can only be scheduled to zone3 to become 2/2/2; scheduling it onto zone1(zone2) would make the ActualSkew(3-1) on zone1(zone2) violate MaxSkew(1). - if MaxSkew is 2, incoming pod can be scheduled onto any zone. When whenUnsatisfiable=ScheduleAnyway, it is used to give higher precedence to topologies that satisfy it. It's a required field. Default value is 1 and 0 is not allowed.
TopologyKey is the key of node labels. Nodes that have a label with this key and identical values are considered to be in the same topology. We consider each <key, value> as a "bucket", and try to put balanced number of pods into each bucket. We define a domain as a particular instance of a topology. Also, we define an eligible domain as a domain whose nodes meet the requirements of nodeAffinityPolicy and nodeTaintsPolicy. e.g. If TopologyKey is "kubernetes.io/hostname", each Node is a domain of that topology. And, if TopologyKey is "topology.kubernetes.io/zone", each zone is a domain of that topology. It's a required field.
WhenUnsatisfiable indicates how to deal with a pod if it doesn't satisfy the spread constraint. - DoNotSchedule (default) tells the scheduler not to schedule it. - ScheduleAnyway tells the scheduler to schedule the pod in any location,
but giving higher precedence to topologies that would help reduce the
skew.
A constraint is considered "Unsatisfiable" for an incoming pod if and only if every possible node assignment for that pod would violate "MaxSkew" on some topology. For example, in a 3-zone cluster, MaxSkew is set to 1, and pods with the same labelSelector spread as 3/1/1: | zone1 | zone2 | zone3 | | P P P | P | P | If WhenUnsatisfiable is set to DoNotSchedule, incoming pod can only be scheduled to zone2(zone3) to become 3/2/1(3/1/2) as ActualSkew(2-1) on zone2(zone3) satisfies MaxSkew(1). In other words, the cluster can still be imbalanced, but scheduler won't make it more imbalanced. It's a required field.
LabelSelector is used to find matching pods. Pods that match this label selector are counted to determine the number of pods in their corresponding topology domain.
MatchLabelKeys is a set of pod label keys to select the pods over which spreading will be calculated. The keys are used to lookup values from the incoming pod labels, those key-value labels are ANDed with labelSelector to select the group of existing pods over which spreading will be calculated for the incoming pod. Keys that don't exist in the incoming pod labels will be ignored. A null or empty list means only match against labelSelector.
topologySpreadConstraints.minDomains (int32)
MinDomains indicates a minimum number of eligible domains. When the number of eligible domains with matching topology keys is less than minDomains, Pod Topology Spread treats "global minimum" as 0, and then the calculation of Skew is performed. And when the number of eligible domains with matching topology keys equals or greater than minDomains, this value has no effect on scheduling. As a result, when the number of eligible domains is less than minDomains, scheduler won't schedule more than maxSkew Pods to those domains. If value is nil, the constraint behaves as if MinDomains is equal to 1. Valid values are integers greater than 0. When value is not nil, WhenUnsatisfiable must be DoNotSchedule.
For example, in a 3-zone cluster, MaxSkew is set to 2, MinDomains is set to 5 and pods with the same labelSelector spread as 2/2/2: | zone1 | zone2 | zone3 | | P P | P P | P P | The number of domains is less than 5(MinDomains), so "global minimum" is treated as 0. In this situation, new pod with the same labelSelector cannot be scheduled, because computed skew will be 3(3 - 0) if new Pod is scheduled to any of the three zones, it will violate MaxSkew.
This is a beta field and requires the MinDomainsInPodTopologySpread feature gate to be enabled (enabled by default).
NodeAffinityPolicy indicates how we will treat Pod's nodeAffinity/nodeSelector when calculating pod topology spread skew. Options are: - Honor: only nodes matching nodeAffinity/nodeSelector are included in the calculations. - Ignore: nodeAffinity/nodeSelector are ignored. All nodes are included in the calculations.
If this value is nil, the behavior is equivalent to the Honor policy. This is a alpha-level feature enabled by the NodeInclusionPolicyInPodTopologySpread feature flag.
NodeTaintsPolicy indicates how we will treat node taints when calculating pod topology spread skew. Options are: - Honor: nodes without taints, along with tainted nodes for which the incoming pod has a toleration, are included. - Ignore: node taints are ignored. All nodes are included.
If this value is nil, the behavior is equivalent to the Ignore policy. This is a alpha-level feature enabled by the NodeInclusionPolicyInPodTopologySpread feature flag.
Overhead represents the resource overhead associated with running a pod for a given RuntimeClass. This field will be autopopulated at admission time by the RuntimeClass admission controller. If the RuntimeClass admission controller is enabled, overhead must not be set in Pod create requests. The RuntimeClass admission controller will reject Pod create requests which have the overhead already set. If RuntimeClass is configured and selected in the PodSpec, Overhead will be set to the value defined in the corresponding RuntimeClass, otherwise it will remain unset and treated as zero. More info: https://git.k8s.io/enhancements/keps/sig-node/688-pod-overhead/README.md
Optional duration in seconds the pod needs to terminate gracefully. May be decreased in delete request. Value must be non-negative integer. The value zero indicates stop immediately via the kill signal (no opportunity to shut down). If this value is nil, the default grace period will be used instead. The grace period is the duration in seconds after the processes running in the pod are sent a termination signal and the time when the processes are forcibly halted with a kill signal. Set this value longer than the expected cleanup time for your process. Defaults to 30 seconds.
activeDeadlineSeconds (int64)
Optional duration in seconds the pod may be active on the node relative to StartTime before the system will actively try to mark it failed and kill associated containers. Value must be a positive integer.
PodReadinessGate contains the reference to a pod condition
readinessGates.conditionType (string), required
ConditionType refers to a condition in the pod's condition list with matching type.
Hostname and Name resolution
hostname (string)
Specifies the hostname of the Pod If not specified, the pod's hostname will be set to a system-defined value.
setHostnameAsFQDN (boolean)
If true the pod's hostname will be configured as the pod's FQDN, rather than the leaf name (the default). In Linux containers, this means setting the FQDN in the hostname field of the kernel (the nodename field of struct utsname). In Windows containers, this means setting the registry value of hostname for the registry key HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters to FQDN. If a pod does not have FQDN, this has no effect. Default to false.
subdomain (string)
If specified, the fully qualified Pod hostname will be "<hostname>.<subdomain>.<pod namespace>.svc.<cluster domain>". If not specified, the pod will not have a domainname at all.
hostAliases ([]HostAlias)
Patch strategy: merge on key ip
HostAliases is an optional list of hosts and IPs that will be injected into the pod's hosts file if specified. This is only valid for non-hostNetwork pods.
HostAlias holds the mapping between IP and hostnames that will be injected as an entry in the pod's hosts file.
hostAliases.hostnames ([]string)
Hostnames for the above IP address.
hostAliases.ip (string)
IP address of the host file entry.
dnsConfig (PodDNSConfig)
Specifies the DNS parameters of a pod. Parameters specified here will be merged to the generated DNS configuration based on DNSPolicy.
PodDNSConfig defines the DNS parameters of a pod in addition to those generated from DNSPolicy.
dnsConfig.nameservers ([]string)
A list of DNS name server IP addresses. This will be appended to the base nameservers generated from DNSPolicy. Duplicated nameservers will be removed.
dnsConfig.options ([]PodDNSConfigOption)
A list of DNS resolver options. This will be merged with the base options generated from DNSPolicy. Duplicated entries will be removed. Resolution options given in Options will override those that appear in the base DNSPolicy.
PodDNSConfigOption defines DNS resolver options of a pod.
dnsConfig.options.name (string)
Required.
dnsConfig.options.value (string)
dnsConfig.searches ([]string)
A list of DNS search domains for host-name lookup. This will be appended to the base search paths generated from DNSPolicy. Duplicated search paths will be removed.
dnsPolicy (string)
Set DNS policy for the pod. Defaults to "ClusterFirst". Valid values are 'ClusterFirstWithHostNet', 'ClusterFirst', 'Default' or 'None'. DNS parameters given in DNSConfig will be merged with the policy selected with DNSPolicy. To have DNS options set along with hostNetwork, you have to specify DNS policy explicitly to 'ClusterFirstWithHostNet'.
Hosts namespaces
hostNetwork (boolean)
Host networking requested for this pod. Use the host's network namespace. If this option is set, the ports that will be used must be specified. Default to false.
hostPID (boolean)
Use the host's pid namespace. Optional: Default to false.
hostIPC (boolean)
Use the host's ipc namespace. Optional: Default to false.
shareProcessNamespace (boolean)
Share a single process namespace between all of the containers in a pod. When this is set containers will be able to view and signal processes from other containers in the same pod, and the first process in each container will not be assigned PID 1. HostPID and ShareProcessNamespace cannot both be set. Optional: Default to false.
AutomountServiceAccountToken indicates whether a service account token should be automatically mounted.
Security context
securityContext (PodSecurityContext)
SecurityContext holds pod-level security attributes and common container settings. Optional: Defaults to empty. See type description for default values of each field.
PodSecurityContext holds pod-level security attributes and common container settings. Some fields are also present in container.securityContext. Field values of container.securityContext take precedence over field values of PodSecurityContext.
securityContext.runAsUser (int64)
The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
securityContext.runAsNonRoot (boolean)
Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
securityContext.runAsGroup (int64)
The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
securityContext.supplementalGroups ([]int64)
A list of groups applied to the first process run in each container, in addition to the container's primary GID. If unspecified, no groups will be added to any container. Note that this field cannot be set when spec.os.name is windows.
securityContext.fsGroup (int64)
A special supplemental group that applies to all containers in a pod. Some volume types allow the Kubelet to change the ownership of that volume to be owned by the pod:
The owning GID will be the FSGroup 2. The setgid bit is set (new files created in the volume will be owned by FSGroup) 3. The permission bits are OR'd with rw-rw----
If unset, the Kubelet will not modify the ownership and permissions of any volume. Note that this field cannot be set when spec.os.name is windows.
securityContext.fsGroupChangePolicy (string)
fsGroupChangePolicy defines behavior of changing ownership and permission of the volume before being exposed inside Pod. This field will only apply to volume types which support fsGroup based ownership(and permissions). It will have no effect on ephemeral volume types such as: secret, configmaps and emptydir. Valid values are "OnRootMismatch" and "Always". If not specified, "Always" is used. Note that this field cannot be set when spec.os.name is windows.
securityContext.seccompProfile (SeccompProfile)
The seccomp options to use by the containers in this pod. Note that this field cannot be set when spec.os.name is windows.
SeccompProfile defines a pod/container's seccomp profile settings. Only one profile source may be set.
type indicates which kind of seccomp profile will be applied. Valid options are:
Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied.
localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must only be set if type is "Localhost".
securityContext.seLinuxOptions (SELinuxOptions)
The SELinux context to be applied to all containers. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in SecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence for that container. Note that this field cannot be set when spec.os.name is windows.
SELinuxOptions are the labels to be applied to the container
securityContext.seLinuxOptions.level (string)
Level is SELinux level label that applies to the container.
securityContext.seLinuxOptions.role (string)
Role is a SELinux role label that applies to the container.
securityContext.seLinuxOptions.type (string)
Type is a SELinux type label that applies to the container.
securityContext.seLinuxOptions.user (string)
User is a SELinux user label that applies to the container.
securityContext.sysctls ([]Sysctl)
Sysctls hold a list of namespaced sysctls used for the pod. Pods with unsupported sysctls (by the container runtime) might fail to launch. Note that this field cannot be set when spec.os.name is windows.
The Windows specific settings applied to all containers. If unspecified, the options within a container's SecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
WindowsSecurityContextOptions contain Windows-specific options and credentials.
GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
HostProcess determines if a container should be run as a 'Host Process' container. This field is alpha-level and will only be honored by components that enable the WindowsHostProcessContainers feature flag. Setting this field without the feature flag will result in errors when validating the Pod. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
Alpha level
hostUsers (boolean)
Use the host's user namespace. Optional: Default to true. If set to true or not present, the pod will be run in the host user namespace, useful for when the pod needs a feature only available to the host user namespace, such as loading a kernel module with CAP_SYS_MODULE. When set to false, a new userns is created for the pod. Setting false is useful for mitigating container breakout vulnerabilities even allowing users to run their containers as root without actually having root privileges on the host. This field is alpha-level and is only honored by servers that enable the UserNamespacesSupport feature.
Deprecated
serviceAccount (string)
DeprecatedServiceAccount is a depreciated alias for ServiceAccountName. Deprecated: Use serviceAccountName instead.
Container
A single application container that you want to run within a pod.
name (string), required
Name of the container specified as a DNS_LABEL. Each container in a pod must have a unique name (DNS_LABEL). Cannot be updated.
Image
image (string)
Container image name. More info: https://kubernetes.io/docs/concepts/containers/images This field is optional to allow higher level config management to default or override container images in workload controllers like Deployments and StatefulSets.
Entrypoint array. Not executed within a shell. The container image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Cannot be updated. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
args ([]string)
Arguments to the entrypoint. The container image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Cannot be updated. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
workingDir (string)
Container's working directory. If not specified, the container runtime's default will be used, which might be configured in the container image. Cannot be updated.
Ports
ports ([]ContainerPort)
Patch strategy: merge on key containerPort
Map: unique values on keys containerPort, protocol will be kept during a merge
List of ports to expose from the container. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network. Modifying this array with strategic merge patch may corrupt the data. For more information See https://github.com/kubernetes/kubernetes/issues/108255. Cannot be updated.
ContainerPort represents a network port in a single container.
ports.containerPort (int32), required
Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
ports.hostIP (string)
What host IP to bind the external port to.
ports.hostPort (int32)
Number of port to expose on the host. If specified, this must be a valid port number, 0 < x < 65536. If HostNetwork is specified, this must match ContainerPort. Most containers do not need this.
ports.name (string)
If specified, this must be an IANA_SVC_NAME and unique within the pod. Each named port in a pod must have a unique name. Name for the port that can be referred to by services.
ports.protocol (string)
Protocol for port. Must be UDP, TCP, or SCTP. Defaults to "TCP".
Environment variables
env ([]EnvVar)
Patch strategy: merge on key name
List of environment variables to set in the container. Cannot be updated.
EnvVar represents an environment variable present in a Container.
env.name (string), required
Name of the environment variable. Must be a C_IDENTIFIER.
env.value (string)
Variable references $(VAR_NAME) are expanded using the previously defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
env.valueFrom (EnvVarSource)
Source for the environment variable's value. Cannot be used if value is not empty.
EnvVarSource represents a source for the value of an EnvVar.
Selects a field of the pod: supports metadata.name, metadata.namespace, metadata.labels['\<KEY>'], metadata.annotations['\<KEY>'], spec.nodeName, spec.serviceAccountName, status.hostIP, status.podIP, status.podIPs.
Selects a resource of the container: only resources limits and requests (limits.cpu, limits.memory, limits.ephemeral-storage, requests.cpu, requests.memory and requests.ephemeral-storage) are currently supported.
env.valueFrom.secretKeyRef (SecretKeySelector)
Selects a key of a secret in the pod's namespace
SecretKeySelector selects a key of a Secret.
env.valueFrom.secretKeyRef.key (string), required
The key of the secret to select from. Must be a valid secret key.
Specify whether the Secret or its key must be defined
envFrom ([]EnvFromSource)
List of sources to populate environment variables in the container. The keys defined within a source must be a C_IDENTIFIER. All invalid keys will be reported as an event when the container is starting. When a key exists in multiple sources, the value associated with the last source will take precedence. Values defined by an Env with a duplicate key will take precedence. Cannot be updated.
EnvFromSource represents the source of a set of ConfigMaps
envFrom.configMapRef (ConfigMapEnvSource)
The ConfigMap to select from
*ConfigMapEnvSource selects a ConfigMap to populate the environment variables with.
The contents of the target ConfigMap's Data field will represent the key-value pairs as environment variables.*
Pod volumes to mount into the container's filesystem. Cannot be updated.
VolumeMount describes a mounting of a Volume within a container.
volumeMounts.mountPath (string), required
Path within the container at which the volume should be mounted. Must not contain ':'.
volumeMounts.name (string), required
This must match the Name of a Volume.
volumeMounts.mountPropagation (string)
mountPropagation determines how mounts are propagated from the host to container and the other way around. When not set, MountPropagationNone is used. This field is beta in 1.10.
volumeMounts.readOnly (boolean)
Mounted read-only if true, read-write otherwise (false or unspecified). Defaults to false.
volumeMounts.subPath (string)
Path within the volume from which the container's volume should be mounted. Defaults to "" (volume's root).
volumeMounts.subPathExpr (string)
Expanded path within the volume from which the container's volume should be mounted. Behaves similarly to SubPath but environment variable references $(VAR_NAME) are expanded using the container's environment. Defaults to "" (volume's root). SubPathExpr and SubPath are mutually exclusive.
volumeDevices ([]VolumeDevice)
Patch strategy: merge on key devicePath
volumeDevices is the list of block devices to be used by the container.
volumeDevice describes a mapping of a raw block device within a container.
volumeDevices.devicePath (string), required
devicePath is the path inside of the container that the device will be mapped to.
volumeDevices.name (string), required
name must match the name of a persistentVolumeClaim in the pod
Actions that the management system should take in response to container lifecycle events. Cannot be updated.
Lifecycle describes actions that the management system should take in response to container lifecycle events. For the PostStart and PreStop lifecycle handlers, management of the container blocks until the action is complete, unless the container process fails, in which case the handler is aborted.
PreStop is called immediately before a container is terminated due to an API request or management event such as liveness/startup probe failure, preemption, resource contention, etc. The handler is not called if the container crashes or exits. The Pod's termination grace period countdown begins before the PreStop hook is executed. Regardless of the outcome of the handler, the container will eventually terminate within the Pod's termination grace period (unless delayed by finalizers). Other management of the container blocks until the hook completes or until the termination grace period is reached. More info: https://kubernetes.io/docs/concepts/containers/container-lifecycle-hooks/#container-hooks
terminationMessagePath (string)
Optional: Path at which the file to which the container's termination message will be written is mounted into the container's filesystem. Message written is intended to be brief final status, such as an assertion failure message. Will be truncated by the node if greater than 4096 bytes. The total message length across all containers will be limited to 12kb. Defaults to /dev/termination-log. Cannot be updated.
terminationMessagePolicy (string)
Indicate how the termination message should be populated. File will use the contents of terminationMessagePath to populate the container status message on both success and failure. FallbackToLogsOnError will use the last chunk of container log output if the termination message file is empty and the container exited with an error. The log output is limited to 2048 bytes or 80 lines, whichever is smaller. Defaults to File. Cannot be updated.
StartupProbe indicates that the Pod has successfully initialized. If specified, no other probes are executed until this completes successfully. If this probe fails, the Pod will be restarted, just as if the livenessProbe failed. This can be used to provide different probe parameters at the beginning of a Pod's lifecycle, when it might take a long time to load data or warm a cache, than during steady-state operation. This cannot be updated. More info: https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle#container-probes
SecurityContext holds security configuration that will be applied to a container. Some fields are present in both SecurityContext and PodSecurityContext. When both are set, the values in SecurityContext take precedence.
securityContext.runAsUser (int64)
The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
securityContext.runAsNonRoot (boolean)
Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
securityContext.runAsGroup (int64)
The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
securityContext.readOnlyRootFilesystem (boolean)
Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
securityContext.procMount (string)
procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
securityContext.privileged (boolean)
Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.
securityContext.capabilities (Capabilities)
The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
Adds and removes POSIX capabilities from running containers.
securityContext.capabilities.add ([]string)
Added capabilities
securityContext.capabilities.drop ([]string)
Removed capabilities
securityContext.seccompProfile (SeccompProfile)
The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
SeccompProfile defines a pod/container's seccomp profile settings. Only one profile source may be set.
type indicates which kind of seccomp profile will be applied. Valid options are:
Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied.
localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must only be set if type is "Localhost".
securityContext.seLinuxOptions (SELinuxOptions)
The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
SELinuxOptions are the labels to be applied to the container
securityContext.seLinuxOptions.level (string)
Level is SELinux level label that applies to the container.
securityContext.seLinuxOptions.role (string)
Role is a SELinux role label that applies to the container.
securityContext.seLinuxOptions.type (string)
Type is a SELinux type label that applies to the container.
securityContext.seLinuxOptions.user (string)
User is a SELinux user label that applies to the container.
The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
WindowsSecurityContextOptions contain Windows-specific options and credentials.
GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
HostProcess determines if a container should be run as a 'Host Process' container. This field is alpha-level and will only be honored by components that enable the WindowsHostProcessContainers feature flag. Setting this field without the feature flag will result in errors when validating the Pod. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
Debugging
stdin (boolean)
Whether this container should allocate a buffer for stdin in the container runtime. If this is not set, reads from stdin in the container will always result in EOF. Default is false.
stdinOnce (boolean)
Whether the container runtime should close the stdin channel after it has been opened by a single attach. When stdin is true the stdin stream will remain open across multiple attach sessions. If stdinOnce is set to true, stdin is opened on container start, is empty until the first client attaches to stdin, and then remains open and accepts data until the client disconnects, at which time stdin is closed and remains closed until the container is restarted. If this flag is false, a container processes that reads from stdin will never receive an EOF. Default is false
tty (boolean)
Whether this container should allocate a TTY for itself, also requires 'stdin' to be true. Default is false.
EphemeralContainer
An EphemeralContainer is a temporary container that you may add to an existing Pod for user-initiated activities such as debugging. Ephemeral containers have no resource or scheduling guarantees, and they will not be restarted when they exit or when a Pod is removed or restarted. The kubelet may evict a Pod if an ephemeral container causes the Pod to exceed its resource allocation.
To add an ephemeral container, use the ephemeralcontainers subresource of an existing Pod. Ephemeral containers may not be removed or restarted.
name (string), required
Name of the ephemeral container specified as a DNS_LABEL. This name must be unique among all containers, init containers and ephemeral containers.
targetContainerName (string)
If set, the name of the container from PodSpec that this ephemeral container targets. The ephemeral container will be run in the namespaces (IPC, PID, etc) of this container. If not set then the ephemeral container uses the namespaces configured in the Pod spec.
The container runtime must implement support for this feature. If the runtime does not support namespace targeting then the result of setting this field is undefined.
Entrypoint array. Not executed within a shell. The image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Cannot be updated. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
args ([]string)
Arguments to the entrypoint. The image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Cannot be updated. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
workingDir (string)
Container's working directory. If not specified, the container runtime's default will be used, which might be configured in the container image. Cannot be updated.
Environment variables
env ([]EnvVar)
Patch strategy: merge on key name
List of environment variables to set in the container. Cannot be updated.
EnvVar represents an environment variable present in a Container.
env.name (string), required
Name of the environment variable. Must be a C_IDENTIFIER.
env.value (string)
Variable references $(VAR_NAME) are expanded using the previously defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. Double $$ are reduced to a single $, which allows for escaping the $(VAR_NAME) syntax: i.e. "$$(VAR_NAME)" will produce the string literal "$(VAR_NAME)". Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
env.valueFrom (EnvVarSource)
Source for the environment variable's value. Cannot be used if value is not empty.
EnvVarSource represents a source for the value of an EnvVar.
Selects a field of the pod: supports metadata.name, metadata.namespace, metadata.labels['\<KEY>'], metadata.annotations['\<KEY>'], spec.nodeName, spec.serviceAccountName, status.hostIP, status.podIP, status.podIPs.
Selects a resource of the container: only resources limits and requests (limits.cpu, limits.memory, limits.ephemeral-storage, requests.cpu, requests.memory and requests.ephemeral-storage) are currently supported.
env.valueFrom.secretKeyRef (SecretKeySelector)
Selects a key of a secret in the pod's namespace
SecretKeySelector selects a key of a Secret.
env.valueFrom.secretKeyRef.key (string), required
The key of the secret to select from. Must be a valid secret key.
Specify whether the Secret or its key must be defined
envFrom ([]EnvFromSource)
List of sources to populate environment variables in the container. The keys defined within a source must be a C_IDENTIFIER. All invalid keys will be reported as an event when the container is starting. When a key exists in multiple sources, the value associated with the last source will take precedence. Values defined by an Env with a duplicate key will take precedence. Cannot be updated.
EnvFromSource represents the source of a set of ConfigMaps
envFrom.configMapRef (ConfigMapEnvSource)
The ConfigMap to select from
*ConfigMapEnvSource selects a ConfigMap to populate the environment variables with.
The contents of the target ConfigMap's Data field will represent the key-value pairs as environment variables.*
Pod volumes to mount into the container's filesystem. Subpath mounts are not allowed for ephemeral containers. Cannot be updated.
VolumeMount describes a mounting of a Volume within a container.
volumeMounts.mountPath (string), required
Path within the container at which the volume should be mounted. Must not contain ':'.
volumeMounts.name (string), required
This must match the Name of a Volume.
volumeMounts.mountPropagation (string)
mountPropagation determines how mounts are propagated from the host to container and the other way around. When not set, MountPropagationNone is used. This field is beta in 1.10.
volumeMounts.readOnly (boolean)
Mounted read-only if true, read-write otherwise (false or unspecified). Defaults to false.
volumeMounts.subPath (string)
Path within the volume from which the container's volume should be mounted. Defaults to "" (volume's root).
volumeMounts.subPathExpr (string)
Expanded path within the volume from which the container's volume should be mounted. Behaves similarly to SubPath but environment variable references $(VAR_NAME) are expanded using the container's environment. Defaults to "" (volume's root). SubPathExpr and SubPath are mutually exclusive.
volumeDevices ([]VolumeDevice)
Patch strategy: merge on key devicePath
volumeDevices is the list of block devices to be used by the container.
volumeDevice describes a mapping of a raw block device within a container.
volumeDevices.devicePath (string), required
devicePath is the path inside of the container that the device will be mapped to.
volumeDevices.name (string), required
name must match the name of a persistentVolumeClaim in the pod
Lifecycle
terminationMessagePath (string)
Optional: Path at which the file to which the container's termination message will be written is mounted into the container's filesystem. Message written is intended to be brief final status, such as an assertion failure message. Will be truncated by the node if greater than 4096 bytes. The total message length across all containers will be limited to 12kb. Defaults to /dev/termination-log. Cannot be updated.
terminationMessagePolicy (string)
Indicate how the termination message should be populated. File will use the contents of terminationMessagePath to populate the container status message on both success and failure. FallbackToLogsOnError will use the last chunk of container log output if the termination message file is empty and the container exited with an error. The log output is limited to 2048 bytes or 80 lines, whichever is smaller. Defaults to File. Cannot be updated.
Debugging
stdin (boolean)
Whether this container should allocate a buffer for stdin in the container runtime. If this is not set, reads from stdin in the container will always result in EOF. Default is false.
stdinOnce (boolean)
Whether the container runtime should close the stdin channel after it has been opened by a single attach. When stdin is true the stdin stream will remain open across multiple attach sessions. If stdinOnce is set to true, stdin is opened on container start, is empty until the first client attaches to stdin, and then remains open and accepts data until the client disconnects, at which time stdin is closed and remains closed until the container is restarted. If this flag is false, a container processes that reads from stdin will never receive an EOF. Default is false
tty (boolean)
Whether this container should allocate a TTY for itself, also requires 'stdin' to be true. Default is false.
Security context
securityContext (SecurityContext)
Optional: SecurityContext defines the security options the ephemeral container should be run with. If set, the fields of SecurityContext override the equivalent fields of PodSecurityContext.
SecurityContext holds security configuration that will be applied to a container. Some fields are present in both SecurityContext and PodSecurityContext. When both are set, the values in SecurityContext take precedence.
securityContext.runAsUser (int64)
The UID to run the entrypoint of the container process. Defaults to user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
securityContext.runAsNonRoot (boolean)
Indicates that the container must run as a non-root user. If true, the Kubelet will validate the image at runtime to ensure that it does not run as UID 0 (root) and fail to start the container if it does. If unset or false, no such validation will be performed. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
securityContext.runAsGroup (int64)
The GID to run the entrypoint of the container process. Uses runtime default if unset. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
securityContext.readOnlyRootFilesystem (boolean)
Whether this container has a read-only root filesystem. Default is false. Note that this field cannot be set when spec.os.name is windows.
securityContext.procMount (string)
procMount denotes the type of proc mount to use for the containers. The default is DefaultProcMount which uses the container runtime defaults for readonly paths and masked paths. This requires the ProcMountType feature flag to be enabled. Note that this field cannot be set when spec.os.name is windows.
securityContext.privileged (boolean)
Run container in privileged mode. Processes in privileged containers are essentially equivalent to root on the host. Defaults to false. Note that this field cannot be set when spec.os.name is windows.
AllowPrivilegeEscalation controls whether a process can gain more privileges than its parent process. This bool directly controls if the no_new_privs flag will be set on the container process. AllowPrivilegeEscalation is true always when the container is: 1) run as Privileged 2) has CAP_SYS_ADMIN Note that this field cannot be set when spec.os.name is windows.
securityContext.capabilities (Capabilities)
The capabilities to add/drop when running containers. Defaults to the default set of capabilities granted by the container runtime. Note that this field cannot be set when spec.os.name is windows.
Adds and removes POSIX capabilities from running containers.
securityContext.capabilities.add ([]string)
Added capabilities
securityContext.capabilities.drop ([]string)
Removed capabilities
securityContext.seccompProfile (SeccompProfile)
The seccomp options to use by this container. If seccomp options are provided at both the pod & container level, the container options override the pod options. Note that this field cannot be set when spec.os.name is windows.
SeccompProfile defines a pod/container's seccomp profile settings. Only one profile source may be set.
type indicates which kind of seccomp profile will be applied. Valid options are:
Localhost - a profile defined in a file on the node should be used. RuntimeDefault - the container runtime default profile should be used. Unconfined - no profile should be applied.
localhostProfile indicates a profile defined in a file on the node should be used. The profile must be preconfigured on the node to work. Must be a descending path, relative to the kubelet's configured seccomp profile location. Must only be set if type is "Localhost".
securityContext.seLinuxOptions (SELinuxOptions)
The SELinux context to be applied to the container. If unspecified, the container runtime will allocate a random SELinux context for each container. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is windows.
SELinuxOptions are the labels to be applied to the container
securityContext.seLinuxOptions.level (string)
Level is SELinux level label that applies to the container.
securityContext.seLinuxOptions.role (string)
Role is a SELinux role label that applies to the container.
securityContext.seLinuxOptions.type (string)
Type is a SELinux type label that applies to the container.
securityContext.seLinuxOptions.user (string)
User is a SELinux user label that applies to the container.
The Windows specific settings applied to all containers. If unspecified, the options from the PodSecurityContext will be used. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence. Note that this field cannot be set when spec.os.name is linux.
WindowsSecurityContextOptions contain Windows-specific options and credentials.
GMSACredentialSpec is where the GMSA admission webhook (https://github.com/kubernetes-sigs/windows-gmsa) inlines the contents of the GMSA credential spec named by the GMSACredentialSpecName field.
HostProcess determines if a container should be run as a 'Host Process' container. This field is alpha-level and will only be honored by components that enable the WindowsHostProcessContainers feature flag. Setting this field without the feature flag will result in errors when validating the Pod. All of a Pod's containers must have the same effective HostProcess value (it is not allowed to have a mix of HostProcess containers and non-HostProcess containers). In addition, if HostProcess is true then HostNetwork must also be set to true.
The UserName in Windows to run the entrypoint of the container process. Defaults to the user specified in image metadata if unspecified. May also be set in PodSecurityContext. If set in both SecurityContext and PodSecurityContext, the value specified in SecurityContext takes precedence.
Not allowed
ports ([]ContainerPort)
Patch strategy: merge on key containerPort
Map: unique values on keys containerPort, protocol will be kept during a merge
Ports are not allowed for ephemeral containers.
ContainerPort represents a network port in a single container.
ports.containerPort (int32), required
Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
ports.hostIP (string)
What host IP to bind the external port to.
ports.hostPort (int32)
Number of port to expose on the host. If specified, this must be a valid port number, 0 < x < 65536. If HostNetwork is specified, this must match ContainerPort. Most containers do not need this.
ports.name (string)
If specified, this must be an IANA_SVC_NAME and unique within the pod. Each named port in a pod must have a unique name. Name for the port that can be referred to by services.
ports.protocol (string)
Protocol for port. Must be UDP, TCP, or SCTP. Defaults to "TCP".
resources (ResourceRequirements)
Resources are not allowed for ephemeral containers. Ephemeral containers use spare resources already allocated to the pod.
ResourceRequirements describes the compute resource requirements.
Lifecycle is not allowed for ephemeral containers.
Lifecycle describes actions that the management system should take in response to container lifecycle events. For the PostStart and PreStop lifecycle handlers, management of the container blocks until the action is complete, unless the container process fails, in which case the handler is aborted.
PreStop is called immediately before a container is terminated due to an API request or management event such as liveness/startup probe failure, preemption, resource contention, etc. The handler is not called if the container crashes or exits. The Pod's termination grace period countdown begins before the PreStop hook is executed. Regardless of the outcome of the handler, the container will eventually terminate within the Pod's termination grace period (unless delayed by finalizers). Other management of the container blocks until the hook completes or until the termination grace period is reached. More info: https://kubernetes.io/docs/concepts/containers/container-lifecycle-hooks/#container-hooks
LifecycleHandler defines a specific action that should be taken in a lifecycle hook. One and only one of the fields, except TCPSocket must be specified.
exec (ExecAction)
Exec specifies the action to take.
ExecAction describes a "run in container" action.
exec.command ([]string)
Command is the command line to execute inside the container, the working directory for the command is root ('/') in the container's filesystem. The command is simply exec'd, it is not run inside a shell, so traditional shell instructions ('|', etc) won't work. To use a shell, you need to explicitly call out to that shell. Exit status of 0 is treated as live/healthy and non-zero is unhealthy.
httpGet (HTTPGetAction)
HTTPGet specifies the http request to perform.
HTTPGetAction describes an action based on HTTP Get requests.
httpGet.port (IntOrString), required
Name or number of the port to access on the container. Number must be in the range 1 to 65535. Name must be an IANA_SVC_NAME.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
httpGet.host (string)
Host name to connect to, defaults to the pod IP. You probably want to set "Host" in httpHeaders instead.
httpGet.httpHeaders ([]HTTPHeader)
Custom headers to set in the request. HTTP allows repeated headers.
HTTPHeader describes a custom header to be used in HTTP probes
httpGet.httpHeaders.name (string), required
The header field name
httpGet.httpHeaders.value (string), required
The header field value
httpGet.path (string)
Path to access on the HTTP server.
httpGet.scheme (string)
Scheme to use for connecting to the host. Defaults to HTTP.
tcpSocket (TCPSocketAction)
Deprecated. TCPSocket is NOT supported as a LifecycleHandler and kept for the backward compatibility. There are no validation of this field and lifecycle hooks will fail in runtime when tcp handler is specified.
TCPSocketAction describes an action based on opening a socket
tcpSocket.port (IntOrString), required
Number or name of the port to access on the container. Number must be in the range 1 to 65535. Name must be an IANA_SVC_NAME.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
tcpSocket.host (string)
Optional: Host name to connect to, defaults to the pod IP.
NodeAffinity
Node affinity is a group of node affinity scheduling rules.
The scheduler will prefer to schedule pods to nodes that satisfy the affinity expressions specified by this field, but it may choose a node that violates one or more of the expressions. The node that is most preferred is the one with the greatest sum of weights, i.e. for each node that meets all of the scheduling requirements (resource request, requiredDuringScheduling affinity expressions, etc.), compute a sum by iterating through the elements of this field and adding "weight" to the sum if the node matches the corresponding matchExpressions; the node(s) with the highest sum are the most preferred.
An empty preferred scheduling term matches all objects with implicit weight 0 (i.e. it's a no-op). A null preferred scheduling term matches no objects (i.e. is also a no-op).
A node selector term, associated with the corresponding weight.
A null or empty node selector term matches no objects. The requirements of them are ANDed. The TopologySelectorTerm type implements a subset of the NodeSelectorTerm.
If the affinity requirements specified by this field are not met at scheduling time, the pod will not be scheduled onto the node. If the affinity requirements specified by this field cease to be met at some point during pod execution (e.g. due to an update), the system may or may not try to eventually evict the pod from its node.
A node selector represents the union of the results of one or more label queries over a set of nodes; that is, it represents the OR of the selectors represented by the node selector terms.
Required. A list of node selector terms. The terms are ORed.
A null or empty node selector term matches no objects. The requirements of them are ANDed. The TopologySelectorTerm type implements a subset of the NodeSelectorTerm.
The scheduler will prefer to schedule pods to nodes that satisfy the affinity expressions specified by this field, but it may choose a node that violates one or more of the expressions. The node that is most preferred is the one with the greatest sum of weights, i.e. for each node that meets all of the scheduling requirements (resource request, requiredDuringScheduling affinity expressions, etc.), compute a sum by iterating through the elements of this field and adding "weight" to the sum if the node has pods which matches the corresponding podAffinityTerm; the node(s) with the highest sum are the most preferred.
The weights of all of the matched WeightedPodAffinityTerm fields are added per-node to find the most preferred node(s)
Required. A pod affinity term, associated with the corresponding weight.
Defines a set of pods (namely those matching the labelSelector relative to the given namespace(s)) that this pod should be co-located (affinity) or not co-located (anti-affinity) with, where co-located is defined as running on a node whose value of the label with key matches that of any node on which a pod of the set of pods is running
This pod should be co-located (affinity) or not co-located (anti-affinity) with the pods matching the labelSelector in the specified namespaces, where co-located is defined as running on a node whose value of the label with key topologyKey matches that of any node on which any of the selected pods is running. Empty topologyKey is not allowed.
A label query over the set of namespaces that the term applies to. The term is applied to the union of the namespaces selected by this field and the ones listed in the namespaces field. null selector and null or empty namespaces list means "this pod's namespace". An empty selector ({}) matches all namespaces.
namespaces specifies a static list of namespace names that the term applies to. The term is applied to the union of the namespaces listed in this field and the ones selected by namespaceSelector. null or empty namespaces list and null namespaceSelector means "this pod's namespace".
If the affinity requirements specified by this field are not met at scheduling time, the pod will not be scheduled onto the node. If the affinity requirements specified by this field cease to be met at some point during pod execution (e.g. due to a pod label update), the system may or may not try to eventually evict the pod from its node. When there are multiple elements, the lists of nodes corresponding to each podAffinityTerm are intersected, i.e. all terms must be satisfied.
Defines a set of pods (namely those matching the labelSelector relative to the given namespace(s)) that this pod should be co-located (affinity) or not co-located (anti-affinity) with, where co-located is defined as running on a node whose value of the label with key matches that of any node on which a pod of the set of pods is running
This pod should be co-located (affinity) or not co-located (anti-affinity) with the pods matching the labelSelector in the specified namespaces, where co-located is defined as running on a node whose value of the label with key topologyKey matches that of any node on which any of the selected pods is running. Empty topologyKey is not allowed.
A label query over the set of namespaces that the term applies to. The term is applied to the union of the namespaces selected by this field and the ones listed in the namespaces field. null selector and null or empty namespaces list means "this pod's namespace". An empty selector ({}) matches all namespaces.
namespaces specifies a static list of namespace names that the term applies to. The term is applied to the union of the namespaces listed in this field and the ones selected by namespaceSelector. null or empty namespaces list and null namespaceSelector means "this pod's namespace".
PodAntiAffinity
Pod anti affinity is a group of inter pod anti affinity scheduling rules.
The scheduler will prefer to schedule pods to nodes that satisfy the anti-affinity expressions specified by this field, but it may choose a node that violates one or more of the expressions. The node that is most preferred is the one with the greatest sum of weights, i.e. for each node that meets all of the scheduling requirements (resource request, requiredDuringScheduling anti-affinity expressions, etc.), compute a sum by iterating through the elements of this field and adding "weight" to the sum if the node has pods which matches the corresponding podAffinityTerm; the node(s) with the highest sum are the most preferred.
The weights of all of the matched WeightedPodAffinityTerm fields are added per-node to find the most preferred node(s)
Required. A pod affinity term, associated with the corresponding weight.
Defines a set of pods (namely those matching the labelSelector relative to the given namespace(s)) that this pod should be co-located (affinity) or not co-located (anti-affinity) with, where co-located is defined as running on a node whose value of the label with key matches that of any node on which a pod of the set of pods is running
This pod should be co-located (affinity) or not co-located (anti-affinity) with the pods matching the labelSelector in the specified namespaces, where co-located is defined as running on a node whose value of the label with key topologyKey matches that of any node on which any of the selected pods is running. Empty topologyKey is not allowed.
A label query over the set of namespaces that the term applies to. The term is applied to the union of the namespaces selected by this field and the ones listed in the namespaces field. null selector and null or empty namespaces list means "this pod's namespace". An empty selector ({}) matches all namespaces.
namespaces specifies a static list of namespace names that the term applies to. The term is applied to the union of the namespaces listed in this field and the ones selected by namespaceSelector. null or empty namespaces list and null namespaceSelector means "this pod's namespace".
If the anti-affinity requirements specified by this field are not met at scheduling time, the pod will not be scheduled onto the node. If the anti-affinity requirements specified by this field cease to be met at some point during pod execution (e.g. due to a pod label update), the system may or may not try to eventually evict the pod from its node. When there are multiple elements, the lists of nodes corresponding to each podAffinityTerm are intersected, i.e. all terms must be satisfied.
Defines a set of pods (namely those matching the labelSelector relative to the given namespace(s)) that this pod should be co-located (affinity) or not co-located (anti-affinity) with, where co-located is defined as running on a node whose value of the label with key matches that of any node on which a pod of the set of pods is running
This pod should be co-located (affinity) or not co-located (anti-affinity) with the pods matching the labelSelector in the specified namespaces, where co-located is defined as running on a node whose value of the label with key topologyKey matches that of any node on which any of the selected pods is running. Empty topologyKey is not allowed.
A label query over the set of namespaces that the term applies to. The term is applied to the union of the namespaces selected by this field and the ones listed in the namespaces field. null selector and null or empty namespaces list means "this pod's namespace". An empty selector ({}) matches all namespaces.
namespaces specifies a static list of namespace names that the term applies to. The term is applied to the union of the namespaces listed in this field and the ones selected by namespaceSelector. null or empty namespaces list and null namespaceSelector means "this pod's namespace".
Probe
Probe describes a health check to be performed against a container to determine whether it is alive or ready to receive traffic.
exec (ExecAction)
Exec specifies the action to take.
ExecAction describes a "run in container" action.
exec.command ([]string)
Command is the command line to execute inside the container, the working directory for the command is root ('/') in the container's filesystem. The command is simply exec'd, it is not run inside a shell, so traditional shell instructions ('|', etc) won't work. To use a shell, you need to explicitly call out to that shell. Exit status of 0 is treated as live/healthy and non-zero is unhealthy.
httpGet (HTTPGetAction)
HTTPGet specifies the http request to perform.
HTTPGetAction describes an action based on HTTP Get requests.
httpGet.port (IntOrString), required
Name or number of the port to access on the container. Number must be in the range 1 to 65535. Name must be an IANA_SVC_NAME.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
httpGet.host (string)
Host name to connect to, defaults to the pod IP. You probably want to set "Host" in httpHeaders instead.
httpGet.httpHeaders ([]HTTPHeader)
Custom headers to set in the request. HTTP allows repeated headers.
HTTPHeader describes a custom header to be used in HTTP probes
httpGet.httpHeaders.name (string), required
The header field name
httpGet.httpHeaders.value (string), required
The header field value
httpGet.path (string)
Path to access on the HTTP server.
httpGet.scheme (string)
Scheme to use for connecting to the host. Defaults to HTTP.
tcpSocket (TCPSocketAction)
TCPSocket specifies an action involving a TCP port.
TCPSocketAction describes an action based on opening a socket
tcpSocket.port (IntOrString), required
Number or name of the port to access on the container. Number must be in the range 1 to 65535. Name must be an IANA_SVC_NAME.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
tcpSocket.host (string)
Optional: Host name to connect to, defaults to the pod IP.
Optional duration in seconds the pod needs to terminate gracefully upon probe failure. The grace period is the duration in seconds after the processes running in the pod are sent a termination signal and the time when the processes are forcibly halted with a kill signal. Set this value longer than the expected cleanup time for your process. If this value is nil, the pod's terminationGracePeriodSeconds will be used. Otherwise, this value overrides the value provided by the pod spec. Value must be non-negative integer. The value zero indicates stop immediately via the kill signal (no opportunity to shut down). This is a beta field and requires enabling ProbeTerminationGracePeriod feature gate. Minimum value is 1. spec.terminationGracePeriodSeconds is used if unset.
periodSeconds (int32)
How often (in seconds) to perform the probe. Default to 10 seconds. Minimum value is 1.
Minimum consecutive failures for the probe to be considered failed after having succeeded. Defaults to 3. Minimum value is 1.
successThreshold (int32)
Minimum consecutive successes for the probe to be considered successful after having failed. Defaults to 1. Must be 1 for liveness and startup. Minimum value is 1.
grpc (GRPCAction)
GRPC specifies an action involving a GRPC port. This is a beta field and requires enabling GRPCContainerProbe feature gate.
**
grpc.port (int32), required
Port number of the gRPC service. Number must be in the range 1 to 65535.
If this is not specified, the default behavior is defined by gRPC.
PodStatus
PodStatus represents information about the status of a pod. Status may trail the actual state of a system, especially if the node that hosts the pod cannot contact the control plane.
nominatedNodeName (string)
nominatedNodeName is set only when this pod preempts other pods on the node, but it cannot be scheduled right away as preemption victims receive their graceful termination periods. This field does not guarantee that the pod will be scheduled on this node. Scheduler may decide to place the pod elsewhere if other nodes become available sooner. Scheduler may also decide to give the resources on this node to a higher priority pod that is created after preemption. As a result, this field may be different than PodSpec.nodeName when the pod is scheduled.
hostIP (string)
IP address of the host to which the pod is assigned. Empty if not yet scheduled.
startTime (Time)
RFC 3339 date and time at which the object was acknowledged by the Kubelet. This is before the Kubelet pulled the container image(s) for the pod.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
phase (string)
The phase of a Pod is a simple, high-level summary of where the Pod is in its lifecycle. The conditions array, the reason and message fields, and the individual container status arrays contain more detail about the pod's status. There are five possible phase values:
Pending: The pod has been accepted by the Kubernetes system, but one or more of the container images has not been created. This includes time before being scheduled as well as time spent downloading images over the network, which could take a while. Running: The pod has been bound to a node, and all of the containers have been created. At least one container is still running, or is in the process of starting or restarting. Succeeded: All containers in the pod have terminated in success, and will not be restarted. Failed: All containers in the pod have terminated, and at least one container has terminated in failure. The container either exited with non-zero status or was terminated by the system. Unknown: For some reason the state of the pod could not be obtained, typically due to an error in communicating with the host of the pod.
A human readable message indicating details about why the pod is in this condition.
reason (string)
A brief CamelCase message indicating details about why the pod is in this state. e.g. 'Evicted'
podIP (string)
IP address allocated to the pod. Routable at least within the cluster. Empty if not yet allocated.
podIPs ([]PodIP)
Patch strategy: merge on key ip
podIPs holds the IP addresses allocated to the pod. If this field is specified, the 0th entry must match the podIP field. Pods may be allocated at most 1 value for each of IPv4 and IPv6. This list is empty if no IPs have been allocated yet.
*IP address information for entries in the (plural) PodIPs field. Each entry includes:
IP: An IP address allocated to the pod. Routable at least within the cluster.*
podIPs.ip (string)
ip is an IP address (IPv4 or IPv6) assigned to the pod
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastTransitionTime (Time)
Last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
Human-readable message indicating details about last transition.
conditions.reason (string)
Unique, one-word, CamelCase reason for the condition's last transition.
Container's ID in the format '<type>://<container_id>'.
initContainerStatuses.state (ContainerState)
Details about the container's current condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Details about the container's last termination condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
The number of times the container has been restarted.
initContainerStatuses.started (boolean)
Specifies whether the container has passed its startup probe. Initialized as false, becomes true after startupProbe is considered successful. Resets to false when the container is restarted, or if kubelet loses state temporarily. Is always true when no startupProbe is defined.
Container's ID in the format '<type>://<container_id>'.
containerStatuses.state (ContainerState)
Details about the container's current condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
ContainerStateRunning is a running state of a container.
containerStatuses.state.running.startedAt (Time)
Time at which the container was last (re-)started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
ContainerStateWaiting is a waiting state of a container.
containerStatuses.state.waiting.message (string)
Message regarding why the container is not yet running.
containerStatuses.state.waiting.reason (string)
(brief) reason the container is not yet running.
containerStatuses.lastState (ContainerState)
Details about the container's last termination condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Specifies whether the container has passed its readiness probe.
containerStatuses.restartCount (int32), required
The number of times the container has been restarted.
containerStatuses.started (boolean)
Specifies whether the container has passed its startup probe. Initialized as false, becomes true after startupProbe is considered successful. Resets to false when the container is restarted, or if kubelet loses state temporarily. Is always true when no startupProbe is defined.
ephemeralContainerStatuses ([]ContainerStatus)
Status for any ephemeral containers that have run in this pod.
ContainerStatus contains details for the current status of this container.
Container's ID in the format '<type>://<container_id>'.
ephemeralContainerStatuses.state (ContainerState)
Details about the container's current condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Details about the container's last termination condition.
ContainerState holds a possible state of container. Only one of its members may be specified. If none of them is specified, the default one is ContainerStateWaiting.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time at which previous execution of the container started
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
The number of times the container has been restarted.
ephemeralContainerStatuses.started (boolean)
Specifies whether the container has passed its startup probe. Initialized as false, becomes true after startupProbe is considered successful. Resets to false when the container is restarted, or if kubelet loses state temporarily. Is always true when no startupProbe is defined.
The container for which to stream logs. Defaults to only container if there is one container in the pod.
follow (in query): boolean
Follow the log stream of the pod. Defaults to false.
insecureSkipTLSVerifyBackend (in query): boolean
insecureSkipTLSVerifyBackend indicates that the apiserver should not confirm the validity of the serving certificate of the backend it is connecting to. This will make the HTTPS connection between the apiserver and the backend insecure. This means the apiserver cannot verify the log data it is receiving came from the real kubelet. If the kubelet is configured to verify the apiserver's TLS credentials, it does not mean the connection to the real kubelet is vulnerable to a man in the middle attack (e.g. an attacker could not intercept the actual log data coming from the real kubelet).
limitBytes (in query): integer
If set, the number of bytes to read from the server before terminating the log output. This may not display a complete final line of logging, and may return slightly more or slightly less than the specified limit.
Return previous terminated container logs. Defaults to false.
sinceSeconds (in query): integer
A relative time in seconds before the current time from which to show logs. If this value precedes the time a pod was started, only logs since the pod start will be returned. If this value is in the future, no logs will be returned. Only one of sinceSeconds or sinceTime may be specified.
tailLines (in query): integer
If set, the number of lines from the end of the logs to show. If not specified, logs are shown from the creation of the container or sinceSeconds or sinceTime
timestamps (in query): boolean
If true, add an RFC3339 or RFC3339Nano timestamp at the beginning of every line of log output. Defaults to false.
Response
200 (string): OK
401: Unauthorized
get read status of the specified Pod
HTTP Request
GET /api/v1/namespaces/{namespace}/pods/{name}/status
ReplicationControllerSpec is the specification of a replication controller.
selector (map[string]string)
Selector is a label query over pods that should match the Replicas count. If Selector is empty, it is defaulted to the labels present on the Pod template. Label keys and values that must match in order to be controlled by this replication controller, if empty defaulted to labels on Pod template. More info: https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#label-selectors
Minimum number of seconds for which a newly created pod should be ready without any of its container crashing, for it to be considered available. Defaults to 0 (pod will be considered available as soon as it is ready)
ReplicationControllerStatus
ReplicationControllerStatus represents the current status of a replication controller.
The number of available replicas (ready for at least minReadySeconds) for this replication controller.
readyReplicas (int32)
The number of ready replicas for this replication controller.
fullyLabeledReplicas (int32)
The number of pods that have labels matching the labels of the pod template of the replication controller.
conditions ([]ReplicationControllerCondition)
Patch strategy: merge on key type
Represents the latest available observations of a replication controller's current state.
ReplicationControllerCondition describes the state of a replication controller at a certain point.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of replication controller condition.
conditions.lastTransitionTime (Time)
The last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
A human readable message indicating details about the transition.
conditions.reason (string)
The reason for the condition's last transition.
observedGeneration (int64)
ObservedGeneration reflects the generation of the most recently observed replication controller.
ReplicationControllerList
ReplicationControllerList is a collection of replication controllers.
Minimum number of seconds for which a newly created pod should be ready without any of its container crashing, for it to be considered available. Defaults to 0 (pod will be considered available as soon as it is ready)
ReplicaSetStatus
ReplicaSetStatus represents the current status of a ReplicaSet.
The number of available replicas (ready for at least minReadySeconds) for this replica set.
readyReplicas (int32)
readyReplicas is the number of pods targeted by this ReplicaSet with a Ready Condition.
fullyLabeledReplicas (int32)
The number of pods that have labels matching the labels of the pod template of the replicaset.
conditions ([]ReplicaSetCondition)
Patch strategy: merge on key type
Represents the latest available observations of a replica set's current state.
ReplicaSetCondition describes the state of a replica set at a certain point.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of replica set condition.
conditions.lastTransitionTime (Time)
The last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
A human readable message indicating details about the transition.
conditions.reason (string)
The reason for the condition's last transition.
observedGeneration (int64)
ObservedGeneration reflects the generation of the most recently observed ReplicaSet.
Label selector for pods. Existing ReplicaSets whose pods are selected by this will be the ones affected by this deployment. It must match the pod template's labels.
Number of desired pods. This is a pointer to distinguish between explicit zero and not specified. Defaults to 1.
minReadySeconds (int32)
Minimum number of seconds for which a newly created pod should be ready without any of its container crashing, for it to be considered available. Defaults to 0 (pod will be considered available as soon as it is ready)
strategy (DeploymentStrategy)
Patch strategy: retainKeys
The deployment strategy to use to replace existing pods with new ones.
DeploymentStrategy describes how to replace existing pods with new ones.
strategy.type (string)
Type of deployment. Can be "Recreate" or "RollingUpdate". Default is RollingUpdate.
strategy.rollingUpdate (RollingUpdateDeployment)
Rolling update config params. Present only if DeploymentStrategyType = RollingUpdate.
Spec to control the desired behavior of rolling update.
strategy.rollingUpdate.maxSurge (IntOrString)
The maximum number of pods that can be scheduled above the desired number of pods. Value can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%). This can not be 0 if MaxUnavailable is 0. Absolute number is calculated from percentage by rounding up. Defaults to 25%. Example: when this is set to 30%, the new ReplicaSet can be scaled up immediately when the rolling update starts, such that the total number of old and new pods do not exceed 130% of desired pods. Once old pods have been killed, new ReplicaSet can be scaled up further, ensuring that total number of pods running at any time during the update is at most 130% of desired pods.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
The maximum number of pods that can be unavailable during the update. Value can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%). Absolute number is calculated from percentage by rounding down. This can not be 0 if MaxSurge is 0. Defaults to 25%. Example: when this is set to 30%, the old ReplicaSet can be scaled down to 70% of desired pods immediately when the rolling update starts. Once new pods are ready, old ReplicaSet can be scaled down further, followed by scaling up the new ReplicaSet, ensuring that the total number of pods available at all times during the update is at least 70% of desired pods.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
revisionHistoryLimit (int32)
The number of old ReplicaSets to retain to allow rollback. This is a pointer to distinguish between explicit zero and not specified. Defaults to 10.
progressDeadlineSeconds (int32)
The maximum time in seconds for a deployment to make progress before it is considered to be failed. The deployment controller will continue to process failed deployments and a condition with a ProgressDeadlineExceeded reason will be surfaced in the deployment status. Note that progress will not be estimated during the time a deployment is paused. Defaults to 600s.
paused (boolean)
Indicates that the deployment is paused.
DeploymentStatus
DeploymentStatus is the most recently observed status of the Deployment.
replicas (int32)
Total number of non-terminated pods targeted by this deployment (their labels match the selector).
availableReplicas (int32)
Total number of available pods (ready for at least minReadySeconds) targeted by this deployment.
readyReplicas (int32)
readyReplicas is the number of pods targeted by this Deployment with a Ready Condition.
unavailableReplicas (int32)
Total number of unavailable pods targeted by this deployment. This is the total number of pods that are still required for the deployment to have 100% available capacity. They may either be pods that are running but not yet available or pods that still have not been created.
updatedReplicas (int32)
Total number of non-terminated pods targeted by this deployment that have the desired template spec.
collisionCount (int32)
Count of hash collisions for the Deployment. The Deployment controller uses this field as a collision avoidance mechanism when it needs to create the name for the newest ReplicaSet.
conditions ([]DeploymentCondition)
Patch strategy: merge on key type
Represents the latest available observations of a deployment's current state.
DeploymentCondition describes the state of a deployment at a certain point.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of deployment condition.
conditions.lastTransitionTime (Time)
Last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastUpdateTime (Time)
The last time this condition was updated.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
A human readable message indicating details about the transition.
conditions.reason (string)
The reason for the condition's last transition.
observedGeneration (int64)
The generation observed by the deployment controller.
Status is the current status of Pods in this StatefulSet. This data may be out of date by some window of time.
StatefulSetSpec
A StatefulSetSpec is the specification of a StatefulSet.
serviceName (string), required
serviceName is the name of the service that governs this StatefulSet. This service must exist before the StatefulSet, and is responsible for the network identity of the set. Pods get DNS/hostnames that follow the pattern: pod-specific-string.serviceName.default.svc.cluster.local where "pod-specific-string" is managed by the StatefulSet controller.
template is the object that describes the pod that will be created if insufficient replicas are detected. Each pod stamped out by the StatefulSet will fulfill this Template, but have a unique identity from the rest of the StatefulSet.
replicas (int32)
replicas is the desired number of replicas of the given Template. These are replicas in the sense that they are instantiations of the same Template, but individual replicas also have a consistent identity. If unspecified, defaults to 1.
updateStrategy (StatefulSetUpdateStrategy)
updateStrategy indicates the StatefulSetUpdateStrategy that will be employed to update Pods in the StatefulSet when a revision is made to Template.
StatefulSetUpdateStrategy indicates the strategy that the StatefulSet controller will use to perform updates. It includes any additional parameters necessary to perform the update for the indicated strategy.
updateStrategy.type (string)
Type indicates the type of the StatefulSetUpdateStrategy. Default is RollingUpdate.
The maximum number of pods that can be unavailable during the update. Value can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%). Absolute number is calculated from percentage by rounding up. This can not be 0. Defaults to 1. This field is alpha-level and is only honored by servers that enable the MaxUnavailableStatefulSet feature. The field applies to all pods in the range 0 to Replicas-1. That means if there is any unavailable pod in the range 0 to Replicas-1, it will be counted towards MaxUnavailable.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
updateStrategy.rollingUpdate.partition (int32)
Partition indicates the ordinal at which the StatefulSet should be partitioned for updates. During a rolling update, all pods from ordinal Replicas-1 to Partition are updated. All pods from ordinal Partition-1 to 0 remain untouched. This is helpful in being able to do a canary based deployment. The default value is 0.
podManagementPolicy (string)
podManagementPolicy controls how pods are created during initial scale up, when replacing pods on nodes, or when scaling down. The default policy is OrderedReady, where pods are created in increasing order (pod-0, then pod-1, etc) and the controller will wait until each pod is ready before continuing. When scaling down, the pods are removed in the opposite order. The alternative policy is Parallel which will create pods in parallel to match the desired scale without waiting, and on scale down will delete all pods at once.
revisionHistoryLimit (int32)
revisionHistoryLimit is the maximum number of revisions that will be maintained in the StatefulSet's revision history. The revision history consists of all revisions not represented by a currently applied StatefulSetSpec version. The default value is 10.
volumeClaimTemplates is a list of claims that pods are allowed to reference. The StatefulSet controller is responsible for mapping network identities to claims in a way that maintains the identity of a pod. Every claim in this list must have at least one matching (by name) volumeMount in one container in the template. A claim in this list takes precedence over any volumes in the template, with the same name.
minReadySeconds (int32)
Minimum number of seconds for which a newly created pod should be ready without any of its container crashing for it to be considered available. Defaults to 0 (pod will be considered available as soon as it is ready)
persistentVolumeClaimRetentionPolicy describes the lifecycle of persistent volume claims created from volumeClaimTemplates. By default, all persistent volume claims are created as needed and retained until manually deleted. This policy allows the lifecycle to be altered, for example by deleting persistent volume claims when their stateful set is deleted, or when their pod is scaled down. This requires the StatefulSetAutoDeletePVC feature gate to be enabled, which is alpha. +optional
StatefulSetPersistentVolumeClaimRetentionPolicy describes the policy used for PVCs created from the StatefulSet VolumeClaimTemplates.
WhenDeleted specifies what happens to PVCs created from StatefulSet VolumeClaimTemplates when the StatefulSet is deleted. The default policy of Retain causes PVCs to not be affected by StatefulSet deletion. The Delete policy causes those PVCs to be deleted.
WhenScaled specifies what happens to PVCs created from StatefulSet VolumeClaimTemplates when the StatefulSet is scaled down. The default policy of Retain causes PVCs to not be affected by a scaledown. The Delete policy causes the associated PVCs for any excess pods above the replica count to be deleted.
StatefulSetStatus
StatefulSetStatus represents the current state of a StatefulSet.
replicas (int32), required
replicas is the number of Pods created by the StatefulSet controller.
readyReplicas (int32)
readyReplicas is the number of pods created for this StatefulSet with a Ready Condition.
currentReplicas (int32)
currentReplicas is the number of Pods created by the StatefulSet controller from the StatefulSet version indicated by currentRevision.
updatedReplicas (int32)
updatedReplicas is the number of Pods created by the StatefulSet controller from the StatefulSet version indicated by updateRevision.
availableReplicas (int32)
Total number of available pods (ready for at least minReadySeconds) targeted by this statefulset.
collisionCount (int32)
collisionCount is the count of hash collisions for the StatefulSet. The StatefulSet controller uses this field as a collision avoidance mechanism when it needs to create the name for the newest ControllerRevision.
conditions ([]StatefulSetCondition)
Patch strategy: merge on key type
Represents the latest available observations of a statefulset's current state.
StatefulSetCondition describes the state of a statefulset at a certain point.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of statefulset condition.
conditions.lastTransitionTime (Time)
Last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
A human readable message indicating details about the transition.
conditions.reason (string)
The reason for the condition's last transition.
currentRevision (string)
currentRevision, if not empty, indicates the version of the StatefulSet used to generate Pods in the sequence [0,currentReplicas).
updateRevision (string)
updateRevision, if not empty, indicates the version of the StatefulSet used to generate Pods in the sequence [replicas-updatedReplicas,replicas)
observedGeneration (int64)
observedGeneration is the most recent generation observed for this StatefulSet. It corresponds to the StatefulSet's generation, which is updated on mutation by the API Server.
ControllerRevision implements an immutable snapshot of state data.
apiVersion: apps/v1
import "k8s.io/api/apps/v1"
ControllerRevision
ControllerRevision implements an immutable snapshot of state data. Clients are responsible for serializing and deserializing the objects that contain their internal state. Once a ControllerRevision has been successfully created, it can not be updated. The API Server will fail validation of all requests that attempt to mutate the Data field. ControllerRevisions may, however, be deleted. Note that, due to its use by both the DaemonSet and StatefulSet controllers for update and rollback, this object is beta. However, it may be subject to name and representation changes in future releases, and clients should not depend on its stability. It is primarily for internal use by controllers.
Revision indicates the revision of the state represented by Data.
data (RawExtension)
Data is the serialized representation of the state.
*RawExtension is used to hold extensions in external versions.
To use this, make a field which has RawExtension as its type in your external, versioned struct, and Object in your internal struct. You also need to register your various plugin types.
// Internal package:
type MyAPIObject struct {
runtime.TypeMeta json:",inline"
MyPlugin runtime.Object json:"myPlugin"
}
type PluginA struct {
AOption string json:"aOption"
}
// External package:
type MyAPIObject struct {
runtime.TypeMeta json:",inline"
MyPlugin runtime.RawExtension json:"myPlugin"
}
type PluginA struct {
AOption string json:"aOption"
}
// On the wire, the JSON will look something like this:
So what happens? Decode first uses json or yaml to unmarshal the serialized data into your external MyAPIObject. That causes the raw JSON to be stored, but not unpacked. The next step is to copy (using pkg/conversion) into the internal struct. The runtime package's DefaultScheme has conversion functions installed which will unpack the JSON stored in RawExtension, turning it into the correct object type, and storing it in the Object. (TODO: In the case where the object is of an unknown type, a runtime.Unknown object will be created and stored.)*
ControllerRevisionList
ControllerRevisionList is a resource containing a list of ControllerRevision objects.
The minimum number of seconds for which a newly created DaemonSet pod should be ready without any of its container crashing, for it to be considered available. Defaults to 0 (pod will be considered available as soon as it is ready).
updateStrategy (DaemonSetUpdateStrategy)
An update strategy to replace existing DaemonSet pods with new pods.
DaemonSetUpdateStrategy is a struct used to control the update strategy for a DaemonSet.
updateStrategy.type (string)
Type of daemon set update. Can be "RollingUpdate" or "OnDelete". Default is RollingUpdate.
The maximum number of nodes with an existing available DaemonSet pod that can have an updated DaemonSet pod during during an update. Value can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%). This can not be 0 if MaxUnavailable is 0. Absolute number is calculated from percentage by rounding up to a minimum of 1. Default value is 0. Example: when this is set to 30%, at most 30% of the total number of nodes that should be running the daemon pod (i.e. status.desiredNumberScheduled) can have their a new pod created before the old pod is marked as deleted. The update starts by launching new pods on 30% of nodes. Once an updated pod is available (Ready for at least minReadySeconds) the old DaemonSet pod on that node is marked deleted. If the old pod becomes unavailable for any reason (Ready transitions to false, is evicted, or is drained) an updated pod is immediatedly created on that node without considering surge limits. Allowing surge implies the possibility that the resources consumed by the daemonset on any given node can double if the readiness check fails, and so resource intensive daemonsets should take into account that they may cause evictions during disruption.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
The maximum number of DaemonSet pods that can be unavailable during the update. Value can be an absolute number (ex: 5) or a percentage of total number of DaemonSet pods at the start of the update (ex: 10%). Absolute number is calculated from percentage by rounding up. This cannot be 0 if MaxSurge is 0 Default value is 1. Example: when this is set to 30%, at most 30% of the total number of nodes that should be running the daemon pod (i.e. status.desiredNumberScheduled) can have their pods stopped for an update at any given time. The update starts by stopping at most 30% of those DaemonSet pods and then brings up new DaemonSet pods in their place. Once the new pods are available, it then proceeds onto other DaemonSet pods, thus ensuring that at least 70% of original number of DaemonSet pods are available at all times during the update.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
revisionHistoryLimit (int32)
The number of old history to retain to allow rollback. This is a pointer to distinguish between explicit zero and not specified. Defaults to 10.
DaemonSetStatus
DaemonSetStatus represents the current status of a daemon set.
numberReady (int32), required
numberReady is the number of nodes that should be running the daemon pod and have one or more of the daemon pod running with a Ready Condition.
numberAvailable (int32)
The number of nodes that should be running the daemon pod and have one or more of the daemon pod running and available (ready for at least spec.minReadySeconds)
numberUnavailable (int32)
The number of nodes that should be running the daemon pod and have none of the daemon pod running and available (ready for at least spec.minReadySeconds)
The total number of nodes that are running updated daemon pod
collisionCount (int32)
Count of hash collisions for the DaemonSet. The DaemonSet controller uses this field as a collision avoidance mechanism when it needs to create the name for the newest ControllerRevision.
conditions ([]DaemonSetCondition)
Patch strategy: merge on key type
Represents the latest available observations of a DaemonSet's current state.
DaemonSetCondition describes the state of a DaemonSet at a certain point.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of DaemonSet condition.
conditions.lastTransitionTime (Time)
Last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
A human readable message indicating details about the transition.
conditions.reason (string)
The reason for the condition's last transition.
observedGeneration (int64)
The most recent generation observed by the daemon set controller.
Specifies the maximum desired number of pods the job should run at any given time. The actual number of pods running in steady state will be less than this number when ((.spec.completions - .status.successful) < .spec.parallelism), i.e. when the work left to do is less than max parallelism. More info: https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/
Lifecycle
completions (int32)
Specifies the desired number of successfully finished pods the job should be run with. Setting to nil means that the success of any pod signals the success of all pods, and allows parallelism to have any positive value. Setting to 1 means that parallelism is limited to 1 and the success of that pod signals the success of the job. More info: https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/
completionMode (string)
CompletionMode specifies how Pod completions are tracked. It can be NonIndexed (default) or Indexed.
NonIndexed means that the Job is considered complete when there have been .spec.completions successfully completed Pods. Each Pod completion is homologous to each other.
Indexed means that the Pods of a Job get an associated completion index from 0 to (.spec.completions - 1), available in the annotation batch.kubernetes.io/job-completion-index. The Job is considered complete when there is one successfully completed Pod for each index. When value is Indexed, .spec.completions must be specified and .spec.parallelism must be less than or equal to 10^5. In addition, The Pod name takes the form $(job-name)-$(index)-$(random-string), the Pod hostname takes the form $(job-name)-$(index).
More completion modes can be added in the future. If the Job controller observes a mode that it doesn't recognize, which is possible during upgrades due to version skew, the controller skips updates for the Job.
backoffLimit (int32)
Specifies the number of retries before marking this job failed. Defaults to 6
activeDeadlineSeconds (int64)
Specifies the duration in seconds relative to the startTime that the job may be continuously active before the system tries to terminate it; value must be positive integer. If a Job is suspended (at creation or through an update), this timer will effectively be stopped and reset when the Job is resumed again.
ttlSecondsAfterFinished (int32)
ttlSecondsAfterFinished limits the lifetime of a Job that has finished execution (either Complete or Failed). If this field is set, ttlSecondsAfterFinished after the Job finishes, it is eligible to be automatically deleted. When the Job is being deleted, its lifecycle guarantees (e.g. finalizers) will be honored. If this field is unset, the Job won't be automatically deleted. If this field is set to zero, the Job becomes eligible to be deleted immediately after it finishes.
suspend (boolean)
Suspend specifies whether the Job controller should create Pods or not. If a Job is created with suspend set to true, no Pods are created by the Job controller. If a Job is suspended after creation (i.e. the flag goes from false to true), the Job controller will delete all active Pods associated with this Job. Users must design their workload to gracefully handle this. Suspending a Job will reset the StartTime field of the Job, effectively resetting the ActiveDeadlineSeconds timer too. Defaults to false.
manualSelector controls generation of pod labels and pod selectors. Leave manualSelector unset unless you are certain what you are doing. When false or unset, the system pick labels unique to this job and appends those labels to the pod template. When true, the user is responsible for picking unique labels and specifying the selector. Failure to pick a unique label may cause this and other jobs to not function correctly. However, You may see manualSelector=true in jobs that were created with the old extensions/v1beta1 API. More info: https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/#specifying-your-own-pod-selector
Alpha level
podFailurePolicy (PodFailurePolicy)
Specifies the policy of handling failed pods. In particular, it allows to specify the set of actions and conditions which need to be satisfied to take the associated action. If empty, the default behaviour applies - the counter of failed pods, represented by the jobs's .status.failed field, is incremented and it is checked against the backoffLimit. This field cannot be used in combination with restartPolicy=OnFailure.
This field is alpha-level. To use this field, you must enable the JobPodFailurePolicy feature gate (disabled by default).
PodFailurePolicy describes how failed pods influence the backoffLimit.
A list of pod failure policy rules. The rules are evaluated in order. Once a rule matches a Pod failure, the remaining of the rules are ignored. When no rule matches the Pod failure, the default handling applies - the counter of pod failures is incremented and it is checked against the backoffLimit. At most 20 elements are allowed.
PodFailurePolicyRule describes how a pod failure is handled when the requirements are met. One of OnExitCodes and onPodConditions, but not both, can be used in each rule.
podFailurePolicy.rules.action (string), required
Specifies the action taken on a pod failure when the requirements are satisfied. Possible values are: - FailJob: indicates that the pod's job is marked as Failed and all
running pods are terminated.
Ignore: indicates that the counter towards the .backoffLimit is not
incremented and a replacement pod is created.
Count: indicates that the pod is handled in the default way - the
counter towards the .backoffLimit is incremented.
Additional values are considered to be added in the future. Clients should react to an unknown action by skipping the rule.
Represents the requirement on the pod conditions. The requirement is represented as a list of pod condition patterns. The requirement is satisfied if at least one pattern matches an actual pod condition. At most 20 elements are allowed.
PodFailurePolicyOnPodConditionsPattern describes a pattern for matching an actual pod condition type.
Specifies the required Pod condition status. To match a pod condition it is required that the specified status equals the pod condition status. Defaults to True.
Represents the requirement on the container exit codes.
PodFailurePolicyOnExitCodesRequirement describes the requirement for handling a failed pod based on its container exit codes. In particular, it lookups the .state.terminated.exitCode for each app container and init container status, represented by the .status.containerStatuses and .status.initContainerStatuses fields in the Pod status, respectively. Containers completed with success (exit code 0) are excluded from the requirement check.
Represents the relationship between the container exit code(s) and the specified values. Containers completed with success (exit code 0) are excluded from the requirement check. Possible values are: - In: the requirement is satisfied if at least one container exit code
(might be multiple if there are multiple containers not restricted
by the 'containerName' field) is in the set of specified values.
NotIn: the requirement is satisfied if at least one container exit code
(might be multiple if there are multiple containers not restricted
by the 'containerName' field) is not in the set of specified values.
Additional values are considered to be added in the future. Clients should react to an unknown operator by assuming the requirement is not satisfied.
Specifies the set of values. Each returned container exit code (might be multiple in case of multiple containers) is checked against this set of values with respect to the operator. The list of values must be ordered and must not contain duplicates. Value '0' cannot be used for the In operator. At least one element is required. At most 255 elements are allowed.
Restricts the check for exit codes to the container with the specified name. When null, the rule applies to all containers. When specified, it should match one the container or initContainer names in the pod template.
JobStatus
JobStatus represents the current state of a Job.
startTime (Time)
Represents time when the job controller started processing a job. When a Job is created in the suspended state, this field is not set until the first time it is resumed. This field is reset every time a Job is resumed from suspension. It is represented in RFC3339 form and is in UTC.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
completionTime (Time)
Represents time when the job was completed. It is not guaranteed to be set in happens-before order across separate operations. It is represented in RFC3339 form and is in UTC. The completion time is only set when the job finishes successfully.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
active (int32)
The number of pending and running pods.
failed (int32)
The number of pods which reached phase Failed.
succeeded (int32)
The number of pods which reached phase Succeeded.
completedIndexes (string)
CompletedIndexes holds the completed indexes when .spec.completionMode = "Indexed" in a text format. The indexes are represented as decimal integers separated by commas. The numbers are listed in increasing order. Three or more consecutive numbers are compressed and represented by the first and last element of the series, separated by a hyphen. For example, if the completed indexes are 1, 3, 4, 5 and 7, they are represented as "1,3-5,7".
conditions ([]JobCondition)
Patch strategy: merge on key type
Atomic: will be replaced during a merge
The latest available observations of an object's current state. When a Job fails, one of the conditions will have type "Failed" and status true. When a Job is suspended, one of the conditions will have type "Suspended" and status true; when the Job is resumed, the status of this condition will become false. When a Job is completed, one of the conditions will have type "Complete" and status true. More info: https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/
JobCondition describes current state of a job.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of job condition, Complete or Failed.
conditions.lastProbeTime (Time)
Last time the condition was checked.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastTransitionTime (Time)
Last time the condition transit from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
Human readable message indicating details about last transition.
conditions.reason (string)
(brief) reason for the condition's last transition.
uncountedTerminatedPods (UncountedTerminatedPods)
UncountedTerminatedPods holds the UIDs of Pods that have terminated but the job controller hasn't yet accounted for in the status counters.
The job controller creates pods with a finalizer. When a pod terminates (succeeded or failed), the controller does three steps to account for it in the job status: (1) Add the pod UID to the arrays in this field. (2) Remove the pod finalizer. (3) Remove the pod UID from the arrays while increasing the corresponding
counter.
This field is beta-level. The job controller only makes use of this field when the feature gate JobTrackingWithFinalizers is enabled (enabled by default). Old jobs might not be tracked using this field, in which case the field remains null.
UncountedTerminatedPods holds UIDs of Pods that have terminated but haven't been accounted in Job status counters.
uncountedTerminatedPods.failed ([]string)
Set: unique values will be kept during a merge
Failed holds UIDs of failed Pods.
uncountedTerminatedPods.succeeded ([]string)
Set: unique values will be kept during a merge
Succeeded holds UIDs of succeeded Pods.
Beta level
ready (int32)
The number of pods which have a Ready condition.
This field is beta-level. The job controller populates the field when the feature gate JobReadyPods is enabled (enabled by default).
The time zone name for the given schedule, see https://en.wikipedia.org/wiki/List_of_tz_database_time_zones. If not specified, this will default to the time zone of the kube-controller-manager process. The set of valid time zone names and the time zone offset is loaded from the system-wide time zone database by the API server during CronJob validation and the controller manager during execution. If no system-wide time zone database can be found a bundled version of the database is used instead. If the time zone name becomes invalid during the lifetime of a CronJob or due to a change in host configuration, the controller will stop creating new new Jobs and will create a system event with the reason UnknownTimeZone. More information can be found in https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/#time-zones This is beta field and must be enabled via the CronJobTimeZone feature gate.
concurrencyPolicy (string)
Specifies how to treat concurrent executions of a Job. Valid values are: - "Allow" (default): allows CronJobs to run concurrently; - "Forbid": forbids concurrent runs, skipping next run if previous run hasn't finished yet; - "Replace": cancels currently running job and replaces it with a new one
startingDeadlineSeconds (int64)
Optional deadline in seconds for starting the job if it misses scheduled time for any reason. Missed jobs executions will be counted as failed ones.
suspend (boolean)
This flag tells the controller to suspend subsequent executions, it does not apply to already started executions. Defaults to false.
successfulJobsHistoryLimit (int32)
The number of successful finished jobs to retain. Value must be non-negative integer. Defaults to 3.
failedJobsHistoryLimit (int32)
The number of failed finished jobs to retain. Value must be non-negative integer. Defaults to 1.
CronJobStatus
CronJobStatus represents the current state of a cron job.
Information when was the last time the job was successfully scheduled.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
lastSuccessfulTime (Time)
Information when was the last time the job successfully completed.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
reference to scaled resource; horizontal pod autoscaler will learn the current resource consumption and will set the desired number of pods by using its Scale subresource.
CrossVersionObjectReference contains enough information to let you identify the referred resource.
minReplicas is the lower limit for the number of replicas to which the autoscaler can scale down. It defaults to 1 pod. minReplicas is allowed to be 0 if the alpha feature gate HPAScaleToZero is enabled and at least one Object or External metric is configured. Scaling is active as long as at least one metric value is available.
targetCPUUtilizationPercentage (int32)
target average CPU utilization (represented as a percentage of requested CPU) over all the pods; if not specified the default autoscaling policy will be used.
HorizontalPodAutoscalerStatus
current status of a horizontal pod autoscaler
currentReplicas (int32), required
current number of replicas of pods managed by this autoscaler.
desiredReplicas (int32), required
desired number of replicas of pods managed by this autoscaler.
currentCPUUtilizationPercentage (int32)
current average CPU utilization over all pods, represented as a percentage of requested CPU, e.g. 70 means that an average pod is using now 70% of its requested CPU.
lastScaleTime (Time)
last time the HorizontalPodAutoscaler scaled the number of pods; used by the autoscaler to control how often the number of pods is changed.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
observedGeneration (int64)
most recent generation observed by this autoscaler.
HorizontalPodAutoscaler is the configuration for a horizontal pod autoscaler, which automatically manages the replica count of any resource implementing the scale subresource based on the metrics specified.
apiVersion: autoscaling/v2
import "k8s.io/api/autoscaling/v2"
HorizontalPodAutoscaler
HorizontalPodAutoscaler is the configuration for a horizontal pod autoscaler, which automatically manages the replica count of any resource implementing the scale subresource based on the metrics specified.
scaleTargetRef points to the target resource to scale, and is used to the pods for which metrics should be collected, as well as to actually change the replica count.
CrossVersionObjectReference contains enough information to let you identify the referred resource.
minReplicas is the lower limit for the number of replicas to which the autoscaler can scale down. It defaults to 1 pod. minReplicas is allowed to be 0 if the alpha feature gate HPAScaleToZero is enabled and at least one Object or External metric is configured. Scaling is active as long as at least one metric value is available.
behavior (HorizontalPodAutoscalerBehavior)
behavior configures the scaling behavior of the target in both Up and Down directions (scaleUp and scaleDown fields respectively). If not set, the default HPAScalingRules for scale up and scale down are used.
HorizontalPodAutoscalerBehavior configures the scaling behavior of the target in both Up and Down directions (scaleUp and scaleDown fields respectively).
behavior.scaleDown (HPAScalingRules)
scaleDown is scaling policy for scaling Down. If not set, the default value is to allow to scale down to minReplicas pods, with a 300 second stabilization window (i.e., the highest recommendation for the last 300sec is used).
HPAScalingRules configures the scaling behavior for one direction. These Rules are applied after calculating DesiredReplicas from metrics for the HPA. They can limit the scaling velocity by specifying scaling policies. They can prevent flapping by specifying the stabilization window, so that the number of replicas is not set instantly, instead, the safest value from the stabilization window is chosen.
behavior.scaleDown.policies ([]HPAScalingPolicy)
Atomic: will be replaced during a merge
policies is a list of potential scaling polices which can be used during scaling. At least one policy must be specified, otherwise the HPAScalingRules will be discarded as invalid
HPAScalingPolicy is a single policy which must hold true for a specified past interval.
PeriodSeconds specifies the window of time for which the policy should hold true. PeriodSeconds must be greater than zero and less than or equal to 1800 (30 min).
behavior.scaleDown.selectPolicy (string)
selectPolicy is used to specify which policy should be used. If not set, the default value Max is used.
StabilizationWindowSeconds is the number of seconds for which past recommendations should be considered while scaling up or scaling down. StabilizationWindowSeconds must be greater than or equal to zero and less than or equal to 3600 (one hour). If not set, use the default values: - For scale up: 0 (i.e. no stabilization is done). - For scale down: 300 (i.e. the stabilization window is 300 seconds long).
behavior.scaleUp (HPAScalingRules)
scaleUp is scaling policy for scaling Up. If not set, the default value is the higher of:
increase no more than 4 pods per 60 seconds
double the number of pods per 60 seconds
No stabilization is used.
HPAScalingRules configures the scaling behavior for one direction. These Rules are applied after calculating DesiredReplicas from metrics for the HPA. They can limit the scaling velocity by specifying scaling policies. They can prevent flapping by specifying the stabilization window, so that the number of replicas is not set instantly, instead, the safest value from the stabilization window is chosen.
behavior.scaleUp.policies ([]HPAScalingPolicy)
Atomic: will be replaced during a merge
policies is a list of potential scaling polices which can be used during scaling. At least one policy must be specified, otherwise the HPAScalingRules will be discarded as invalid
HPAScalingPolicy is a single policy which must hold true for a specified past interval.
behavior.scaleUp.policies.type (string), required
Type is used to specify the scaling policy.
behavior.scaleUp.policies.value (int32), required
Value contains the amount of change which is permitted by the policy. It must be greater than zero
PeriodSeconds specifies the window of time for which the policy should hold true. PeriodSeconds must be greater than zero and less than or equal to 1800 (30 min).
behavior.scaleUp.selectPolicy (string)
selectPolicy is used to specify which policy should be used. If not set, the default value Max is used.
StabilizationWindowSeconds is the number of seconds for which past recommendations should be considered while scaling up or scaling down. StabilizationWindowSeconds must be greater than or equal to zero and less than or equal to 3600 (one hour). If not set, use the default values: - For scale up: 0 (i.e. no stabilization is done). - For scale down: 300 (i.e. the stabilization window is 300 seconds long).
metrics ([]MetricSpec)
Atomic: will be replaced during a merge
metrics contains the specifications for which to use to calculate the desired replica count (the maximum replica count across all metrics will be used). The desired replica count is calculated multiplying the ratio between the target value and the current value by the current number of pods. Ergo, metrics used must decrease as the pod count is increased, and vice-versa. See the individual metric source types for more information about how each type of metric must respond. If not set, the default metric will be set to 80% average CPU utilization.
MetricSpec specifies how to scale based on a single metric (only type and one other matching field should be set at once).
metrics.type (string), required
type is the type of metric source. It should be one of "ContainerResource", "External", "Object", "Pods" or "Resource", each mapping to a matching field in the object. Note: "ContainerResource" type is available on when the feature-gate HPAContainerMetrics is enabled
containerResource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing a single container in each pod of the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. This is an alpha feature and can be enabled by the HPAContainerMetrics feature flag.
ContainerResourceMetricSource indicates how to scale on a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). The values will be averaged together before being compared to the target. Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. Only one "target" type should be set.
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.external (ExternalMetricSource)
external refers to a global metric that is not associated with any Kubernetes object. It allows autoscaling based on information coming from components running outside of cluster (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
ExternalMetricSource indicates how to scale on a metric not associated with any Kubernetes object (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.external.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.external.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.object.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.object.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
metrics.object.target.averageUtilization (int32)
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.pods (PodsMetricSource)
pods refers to a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
PodsMetricSource indicates how to scale on a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
metrics.pods.metric (MetricIdentifier), required
metric identifies the target metric by name and selector
MetricIdentifier defines the name and optionally selector for a metric
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.pods.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.pods.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
metrics.pods.target.averageUtilization (int32)
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.resource (ResourceMetricSource)
resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ResourceMetricSource indicates how to scale on a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). The values will be averaged together before being compared to the target. Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. Only one "target" type should be set.
metrics.resource.name (string), required
name is the name of the resource in question.
metrics.resource.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.resource.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
HorizontalPodAutoscalerStatus
HorizontalPodAutoscalerStatus describes the current status of a horizontal pod autoscaler.
desiredReplicas (int32), required
desiredReplicas is the desired number of replicas of pods managed by this autoscaler, as last calculated by the autoscaler.
conditions ([]HorizontalPodAutoscalerCondition)
Patch strategy: merge on key type
Map: unique values on key type will be kept during a merge
conditions is the set of conditions required for this autoscaler to scale its target, and indicates whether or not those conditions are met.
HorizontalPodAutoscalerCondition describes the state of a HorizontalPodAutoscaler at a certain point.
conditions.status (string), required
status is the status of the condition (True, False, Unknown)
conditions.type (string), required
type describes the current condition
conditions.lastTransitionTime (Time)
lastTransitionTime is the last time the condition transitioned from one status to another
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is a human-readable explanation containing details about the transition
conditions.reason (string)
reason is the reason for the condition's last transition.
currentMetrics ([]MetricStatus)
Atomic: will be replaced during a merge
currentMetrics is the last read state of the metrics used by this autoscaler.
MetricStatus describes the last-read state of a single metric.
currentMetrics.type (string), required
type is the type of metric source. It will be one of "ContainerResource", "External", "Object", "Pods" or "Resource", each corresponds to a matching field in the object. Note: "ContainerResource" type is available on when the feature-gate HPAContainerMetrics is enabled
container resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing a single container in each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ContainerResourceMetricStatus indicates the current value of a resource metric known to Kubernetes, as specified in requests and limits, describing a single container in each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
external refers to a global metric that is not associated with any Kubernetes object. It allows autoscaling based on information coming from components running outside of cluster (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
ExternalMetricStatus indicates the current value of a global metric not associated with any Kubernetes object.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.object (ObjectMetricStatus)
object refers to a metric describing a single kubernetes object (for example, hits-per-second on an Ingress object).
ObjectMetricStatus indicates the current value of a metric describing a kubernetes object (for example, hits-per-second on an Ingress object).
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.pods (PodsMetricStatus)
pods refers to a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
PodsMetricStatus indicates the current value of a metric describing each pod in the current scale target (for example, transactions-processed-per-second).
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.resource (ResourceMetricStatus)
resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ResourceMetricStatus indicates the current value of a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
value is the current value of the metric (as a quantity).
currentMetrics.resource.name (string), required
Name is the name of the resource in question.
currentReplicas (int32)
currentReplicas is current number of replicas of pods managed by this autoscaler, as last seen by the autoscaler.
lastScaleTime (Time)
lastScaleTime is the last time the HorizontalPodAutoscaler scaled the number of pods, used by the autoscaler to control how often the number of pods is changed.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
observedGeneration (int64)
observedGeneration is the most recent generation observed by this autoscaler.
HorizontalPodAutoscalerList
HorizontalPodAutoscalerList is a list of horizontal pod autoscaler objects.
HorizontalPodAutoscaler is the configuration for a horizontal pod autoscaler, which automatically manages the replica count of any resource implementing the scale subresource based on the metrics specified.
apiVersion: autoscaling/v2beta2
import "k8s.io/api/autoscaling/v2beta2"
HorizontalPodAutoscaler
HorizontalPodAutoscaler is the configuration for a horizontal pod autoscaler, which automatically manages the replica count of any resource implementing the scale subresource based on the metrics specified.
scaleTargetRef points to the target resource to scale, and is used to the pods for which metrics should be collected, as well as to actually change the replica count.
CrossVersionObjectReference contains enough information to let you identify the referred resource.
minReplicas is the lower limit for the number of replicas to which the autoscaler can scale down. It defaults to 1 pod. minReplicas is allowed to be 0 if the alpha feature gate HPAScaleToZero is enabled and at least one Object or External metric is configured. Scaling is active as long as at least one metric value is available.
behavior (HorizontalPodAutoscalerBehavior)
behavior configures the scaling behavior of the target in both Up and Down directions (scaleUp and scaleDown fields respectively). If not set, the default HPAScalingRules for scale up and scale down are used.
HorizontalPodAutoscalerBehavior configures the scaling behavior of the target in both Up and Down directions (scaleUp and scaleDown fields respectively).
behavior.scaleDown (HPAScalingRules)
scaleDown is scaling policy for scaling Down. If not set, the default value is to allow to scale down to minReplicas pods, with a 300 second stabilization window (i.e., the highest recommendation for the last 300sec is used).
HPAScalingRules configures the scaling behavior for one direction. These Rules are applied after calculating DesiredReplicas from metrics for the HPA. They can limit the scaling velocity by specifying scaling policies. They can prevent flapping by specifying the stabilization window, so that the number of replicas is not set instantly, instead, the safest value from the stabilization window is chosen.
behavior.scaleDown.policies ([]HPAScalingPolicy)
policies is a list of potential scaling polices which can be used during scaling. At least one policy must be specified, otherwise the HPAScalingRules will be discarded as invalid
HPAScalingPolicy is a single policy which must hold true for a specified past interval.
PeriodSeconds specifies the window of time for which the policy should hold true. PeriodSeconds must be greater than zero and less than or equal to 1800 (30 min).
behavior.scaleDown.selectPolicy (string)
selectPolicy is used to specify which policy should be used. If not set, the default value MaxPolicySelect is used.
StabilizationWindowSeconds is the number of seconds for which past recommendations should be considered while scaling up or scaling down. StabilizationWindowSeconds must be greater than or equal to zero and less than or equal to 3600 (one hour). If not set, use the default values: - For scale up: 0 (i.e. no stabilization is done). - For scale down: 300 (i.e. the stabilization window is 300 seconds long).
behavior.scaleUp (HPAScalingRules)
scaleUp is scaling policy for scaling Up. If not set, the default value is the higher of:
increase no more than 4 pods per 60 seconds
double the number of pods per 60 seconds
No stabilization is used.
HPAScalingRules configures the scaling behavior for one direction. These Rules are applied after calculating DesiredReplicas from metrics for the HPA. They can limit the scaling velocity by specifying scaling policies. They can prevent flapping by specifying the stabilization window, so that the number of replicas is not set instantly, instead, the safest value from the stabilization window is chosen.
behavior.scaleUp.policies ([]HPAScalingPolicy)
policies is a list of potential scaling polices which can be used during scaling. At least one policy must be specified, otherwise the HPAScalingRules will be discarded as invalid
HPAScalingPolicy is a single policy which must hold true for a specified past interval.
behavior.scaleUp.policies.type (string), required
Type is used to specify the scaling policy.
behavior.scaleUp.policies.value (int32), required
Value contains the amount of change which is permitted by the policy. It must be greater than zero
PeriodSeconds specifies the window of time for which the policy should hold true. PeriodSeconds must be greater than zero and less than or equal to 1800 (30 min).
behavior.scaleUp.selectPolicy (string)
selectPolicy is used to specify which policy should be used. If not set, the default value MaxPolicySelect is used.
StabilizationWindowSeconds is the number of seconds for which past recommendations should be considered while scaling up or scaling down. StabilizationWindowSeconds must be greater than or equal to zero and less than or equal to 3600 (one hour). If not set, use the default values: - For scale up: 0 (i.e. no stabilization is done). - For scale down: 300 (i.e. the stabilization window is 300 seconds long).
metrics ([]MetricSpec)
metrics contains the specifications for which to use to calculate the desired replica count (the maximum replica count across all metrics will be used). The desired replica count is calculated multiplying the ratio between the target value and the current value by the current number of pods. Ergo, metrics used must decrease as the pod count is increased, and vice-versa. See the individual metric source types for more information about how each type of metric must respond. If not set, the default metric will be set to 80% average CPU utilization.
MetricSpec specifies how to scale based on a single metric (only type and one other matching field should be set at once).
metrics.type (string), required
type is the type of metric source. It should be one of "ContainerResource", "External", "Object", "Pods" or "Resource", each mapping to a matching field in the object. Note: "ContainerResource" type is available on when the feature-gate HPAContainerMetrics is enabled
container resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing a single container in each pod of the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. This is an alpha feature and can be enabled by the HPAContainerMetrics feature flag.
ContainerResourceMetricSource indicates how to scale on a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). The values will be averaged together before being compared to the target. Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. Only one "target" type should be set.
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.external (ExternalMetricSource)
external refers to a global metric that is not associated with any Kubernetes object. It allows autoscaling based on information coming from components running outside of cluster (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
ExternalMetricSource indicates how to scale on a metric not associated with any Kubernetes object (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.external.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.external.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.object.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.object.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
metrics.object.target.averageUtilization (int32)
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.pods (PodsMetricSource)
pods refers to a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
PodsMetricSource indicates how to scale on a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
metrics.pods.metric (MetricIdentifier), required
metric identifies the target metric by name and selector
MetricIdentifier defines the name and optionally selector for a metric
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
metrics.pods.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.pods.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
metrics.pods.target.averageUtilization (int32)
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
metrics.resource (ResourceMetricSource)
resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ResourceMetricSource indicates how to scale on a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). The values will be averaged together before being compared to the target. Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source. Only one "target" type should be set.
metrics.resource.name (string), required
name is the name of the resource in question.
metrics.resource.target (MetricTarget), required
target specifies the target value for the given metric
MetricTarget defines the target value, average value, or average utilization of a specific metric
metrics.resource.target.type (string), required
type represents whether the metric type is Utilization, Value, or AverageValue
averageUtilization is the target value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods. Currently only valid for Resource metric source type
value is the target value of the metric (as a quantity).
HorizontalPodAutoscalerStatus
HorizontalPodAutoscalerStatus describes the current status of a horizontal pod autoscaler.
currentReplicas (int32), required
currentReplicas is current number of replicas of pods managed by this autoscaler, as last seen by the autoscaler.
desiredReplicas (int32), required
desiredReplicas is the desired number of replicas of pods managed by this autoscaler, as last calculated by the autoscaler.
conditions ([]HorizontalPodAutoscalerCondition)
conditions is the set of conditions required for this autoscaler to scale its target, and indicates whether or not those conditions are met.
HorizontalPodAutoscalerCondition describes the state of a HorizontalPodAutoscaler at a certain point.
conditions.status (string), required
status is the status of the condition (True, False, Unknown)
conditions.type (string), required
type describes the current condition
conditions.lastTransitionTime (Time)
lastTransitionTime is the last time the condition transitioned from one status to another
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is a human-readable explanation containing details about the transition
conditions.reason (string)
reason is the reason for the condition's last transition.
currentMetrics ([]MetricStatus)
currentMetrics is the last read state of the metrics used by this autoscaler.
MetricStatus describes the last-read state of a single metric.
currentMetrics.type (string), required
type is the type of metric source. It will be one of "ContainerResource", "External", "Object", "Pods" or "Resource", each corresponds to a matching field in the object. Note: "ContainerResource" type is available on when the feature-gate HPAContainerMetrics is enabled
container resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing a single container in each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ContainerResourceMetricStatus indicates the current value of a resource metric known to Kubernetes, as specified in requests and limits, describing a single container in each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
external refers to a global metric that is not associated with any Kubernetes object. It allows autoscaling based on information coming from components running outside of cluster (for example length of queue in cloud messaging service, or QPS from loadbalancer running outside of cluster).
ExternalMetricStatus indicates the current value of a global metric not associated with any Kubernetes object.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.object (ObjectMetricStatus)
object refers to a metric describing a single kubernetes object (for example, hits-per-second on an Ingress object).
ObjectMetricStatus indicates the current value of a metric describing a kubernetes object (for example, hits-per-second on an Ingress object).
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.pods (PodsMetricStatus)
pods refers to a metric describing each pod in the current scale target (for example, transactions-processed-per-second). The values will be averaged together before being compared to the target value.
PodsMetricStatus indicates the current value of a metric describing each pod in the current scale target (for example, transactions-processed-per-second).
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
selector is the string-encoded form of a standard kubernetes label selector for the given metric When set, it is passed as an additional parameter to the metrics server for more specific metrics scoping. When unset, just the metricName will be used to gather metrics.
currentMetrics.resource (ResourceMetricStatus)
resource refers to a resource metric (such as those specified in requests and limits) known to Kubernetes describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
ResourceMetricStatus indicates the current value of a resource metric known to Kubernetes, as specified in requests and limits, describing each pod in the current scale target (e.g. CPU or memory). Such metrics are built in to Kubernetes, and have special scaling options on top of those available to normal per-pod metrics using the "pods" source.
currentAverageUtilization is the current value of the average of the resource metric across all relevant pods, represented as a percentage of the requested value of the resource for the pods.
value is the current value of the metric (as a quantity).
currentMetrics.resource.name (string), required
Name is the name of the resource in question.
lastScaleTime (Time)
lastScaleTime is the last time the HorizontalPodAutoscaler scaled the number of pods, used by the autoscaler to control how often the number of pods is changed.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
observedGeneration (int64)
observedGeneration is the most recent generation observed by this autoscaler.
HorizontalPodAutoscalerList
HorizontalPodAutoscalerList is a list of horizontal pod autoscaler objects.
The value of this priority class. This is the actual priority that pods receive when they have the name of this class in their pod spec.
description (string)
description is an arbitrary string that usually provides guidelines on when this priority class should be used.
globalDefault (boolean)
globalDefault specifies whether this PriorityClass should be considered as the default priority for pods that do not have any priority class. Only one PriorityClass can be marked as globalDefault. However, if more than one PriorityClasses exists with their globalDefault field set to true, the smallest value of such global default PriorityClasses will be used as the default priority.
preemptionPolicy (string)
PreemptionPolicy is the Policy for preempting pods with lower priority. One of Never, PreemptLowerPriority. Defaults to PreemptLowerPriority if unset.
PriorityClassList
PriorityClassList is a collection of priority classes.
Service is a named abstraction of software service (for example, mysql) consisting of local port (for example 3306) that the proxy listens on, and the selector that determines which pods will answer requests sent through the proxy.
apiVersion: v1
import "k8s.io/api/core/v1"
Service
Service is a named abstraction of software service (for example, mysql) consisting of local port (for example 3306) that the proxy listens on, and the selector that determines which pods will answer requests sent through the proxy.
ServiceSpec describes the attributes that a user creates on a service.
selector (map[string]string)
Route service traffic to pods with label keys and values matching this selector. If empty or not present, the service is assumed to have an external process managing its endpoints, which Kubernetes will not modify. Only applies to types ClusterIP, NodePort, and LoadBalancer. Ignored if type is ExternalName. More info: https://kubernetes.io/docs/concepts/services-networking/service/
ports ([]ServicePort)
Patch strategy: merge on key port
Map: unique values on keys port, protocol will be kept during a merge
ServicePort contains information on service's port.
ports.port (int32), required
The port that will be exposed by this service.
ports.targetPort (IntOrString)
Number or name of the port to access on the pods targeted by the service. Number must be in the range 1 to 65535. Name must be an IANA_SVC_NAME. If this is a string, it will be looked up as a named port in the target Pod's container ports. If this is not specified, the value of the 'port' field is used (an identity map). This field is ignored for services with clusterIP=None, and should be omitted or set equal to the 'port' field. More info: https://kubernetes.io/docs/concepts/services-networking/service/#defining-a-service
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
ports.protocol (string)
The IP protocol for this port. Supports "TCP", "UDP", and "SCTP". Default is TCP.
ports.name (string)
The name of this port within the service. This must be a DNS_LABEL. All ports within a ServiceSpec must have unique names. When considering the endpoints for a Service, this must match the 'name' field in the EndpointPort. Optional if only one ServicePort is defined on this service.
ports.nodePort (int32)
The port on each node on which this service is exposed when type is NodePort or LoadBalancer. Usually assigned by the system. If a value is specified, in-range, and not in use it will be used, otherwise the operation will fail. If not specified, a port will be allocated if this Service requires one. If this field is specified when creating a Service which does not need it, creation will fail. This field will be wiped when updating a Service to no longer need it (e.g. changing type from NodePort to ClusterIP). More info: https://kubernetes.io/docs/concepts/services-networking/service/#type-nodeport
ports.appProtocol (string)
The application protocol for this port. This field follows standard Kubernetes label syntax. Un-prefixed names are reserved for IANA standard service names (as per RFC-6335 and https://www.iana.org/assignments/service-names). Non-standard protocols should use prefixed names such as mycompany.com/my-custom-protocol.
type (string)
type determines how the Service is exposed. Defaults to ClusterIP. Valid options are ExternalName, ClusterIP, NodePort, and LoadBalancer. "ClusterIP" allocates a cluster-internal IP address for load-balancing to endpoints. Endpoints are determined by the selector or if that is not specified, by manual construction of an Endpoints object or EndpointSlice objects. If clusterIP is "None", no virtual IP is allocated and the endpoints are published as a set of endpoints rather than a virtual IP. "NodePort" builds on ClusterIP and allocates a port on every node which routes to the same endpoints as the clusterIP. "LoadBalancer" builds on NodePort and creates an external load-balancer (if supported in the current cloud) which routes to the same endpoints as the clusterIP. "ExternalName" aliases this service to the specified externalName. Several other fields do not apply to ExternalName services. More info: https://kubernetes.io/docs/concepts/services-networking/service/#publishing-services-service-types
ipFamilies ([]string)
Atomic: will be replaced during a merge
IPFamilies is a list of IP families (e.g. IPv4, IPv6) assigned to this service. This field is usually assigned automatically based on cluster configuration and the ipFamilyPolicy field. If this field is specified manually, the requested family is available in the cluster, and ipFamilyPolicy allows it, it will be used; otherwise creation of the service will fail. This field is conditionally mutable: it allows for adding or removing a secondary IP family, but it does not allow changing the primary IP family of the Service. Valid values are "IPv4" and "IPv6". This field only applies to Services of types ClusterIP, NodePort, and LoadBalancer, and does apply to "headless" services. This field will be wiped when updating a Service to type ExternalName.
This field may hold a maximum of two entries (dual-stack families, in either order). These families must correspond to the values of the clusterIPs field, if specified. Both clusterIPs and ipFamilies are governed by the ipFamilyPolicy field.
ipFamilyPolicy (string)
IPFamilyPolicy represents the dual-stack-ness requested or required by this Service. If there is no value provided, then this field will be set to SingleStack. Services can be "SingleStack" (a single IP family), "PreferDualStack" (two IP families on dual-stack configured clusters or a single IP family on single-stack clusters), or "RequireDualStack" (two IP families on dual-stack configured clusters, otherwise fail). The ipFamilies and clusterIPs fields depend on the value of this field. This field will be wiped when updating a service to type ExternalName.
clusterIP (string)
clusterIP is the IP address of the service and is usually assigned randomly. If an address is specified manually, is in-range (as per system configuration), and is not in use, it will be allocated to the service; otherwise creation of the service will fail. This field may not be changed through updates unless the type field is also being changed to ExternalName (which requires this field to be blank) or the type field is being changed from ExternalName (in which case this field may optionally be specified, as describe above). Valid values are "None", empty string (""), or a valid IP address. Setting this to "None" makes a "headless service" (no virtual IP), which is useful when direct endpoint connections are preferred and proxying is not required. Only applies to types ClusterIP, NodePort, and LoadBalancer. If this field is specified when creating a Service of type ExternalName, creation will fail. This field will be wiped when updating a Service to type ExternalName. More info: https://kubernetes.io/docs/concepts/services-networking/service/#virtual-ips-and-service-proxies
clusterIPs ([]string)
Atomic: will be replaced during a merge
ClusterIPs is a list of IP addresses assigned to this service, and are usually assigned randomly. If an address is specified manually, is in-range (as per system configuration), and is not in use, it will be allocated to the service; otherwise creation of the service will fail. This field may not be changed through updates unless the type field is also being changed to ExternalName (which requires this field to be empty) or the type field is being changed from ExternalName (in which case this field may optionally be specified, as describe above). Valid values are "None", empty string (""), or a valid IP address. Setting this to "None" makes a "headless service" (no virtual IP), which is useful when direct endpoint connections are preferred and proxying is not required. Only applies to types ClusterIP, NodePort, and LoadBalancer. If this field is specified when creating a Service of type ExternalName, creation will fail. This field will be wiped when updating a Service to type ExternalName. If this field is not specified, it will be initialized from the clusterIP field. If this field is specified, clients must ensure that clusterIPs[0] and clusterIP have the same value.
externalIPs is a list of IP addresses for which nodes in the cluster will also accept traffic for this service. These IPs are not managed by Kubernetes. The user is responsible for ensuring that traffic arrives at a node with this IP. A common example is external load-balancers that are not part of the Kubernetes system.
Only applies to Service Type: LoadBalancer. This feature depends on whether the underlying cloud-provider supports specifying the loadBalancerIP when a load balancer is created. This field will be ignored if the cloud-provider does not support the feature. Deprecated: This field was under-specified and its meaning varies across implementations, and it cannot support dual-stack. As of Kubernetes v1.24, users are encouraged to use implementation-specific annotations when available. This field may be removed in a future API version.
loadBalancerClass is the class of the load balancer implementation this Service belongs to. If specified, the value of this field must be a label-style identifier, with an optional prefix, e.g. "internal-vip" or "example.com/internal-vip". Unprefixed names are reserved for end-users. This field can only be set when the Service type is 'LoadBalancer'. If not set, the default load balancer implementation is used, today this is typically done through the cloud provider integration, but should apply for any default implementation. If set, it is assumed that a load balancer implementation is watching for Services with a matching class. Any default load balancer implementation (e.g. cloud providers) should ignore Services that set this field. This field can only be set when creating or updating a Service to type 'LoadBalancer'. Once set, it can not be changed. This field will be wiped when a service is updated to a non 'LoadBalancer' type.
externalName (string)
externalName is the external reference that discovery mechanisms will return as an alias for this service (e.g. a DNS CNAME record). No proxying will be involved. Must be a lowercase RFC-1123 hostname (https://tools.ietf.org/html/rfc1123) and requires type to be "ExternalName".
externalTrafficPolicy (string)
externalTrafficPolicy describes how nodes distribute service traffic they receive on one of the Service's "externally-facing" addresses (NodePorts, ExternalIPs, and LoadBalancer IPs). If set to "Local", the proxy will configure the service in a way that assumes that external load balancers will take care of balancing the service traffic between nodes, and so each node will deliver traffic only to the node-local endpoints of the service, without masquerading the client source IP. (Traffic mistakenly sent to a node with no endpoints will be dropped.) The default value, "Cluster", uses the standard behavior of routing to all endpoints evenly (possibly modified by topology and other features). Note that traffic sent to an External IP or LoadBalancer IP from within the cluster will always get "Cluster" semantics, but clients sending to a NodePort from within the cluster may need to take traffic policy into account when picking a node.
internalTrafficPolicy (string)
InternalTrafficPolicy describes how nodes distribute service traffic they receive on the ClusterIP. If set to "Local", the proxy will assume that pods only want to talk to endpoints of the service on the same node as the pod, dropping the traffic if there are no local endpoints. The default value, "Cluster", uses the standard behavior of routing to all endpoints evenly (possibly modified by topology and other features).
healthCheckNodePort (int32)
healthCheckNodePort specifies the healthcheck nodePort for the service. This only applies when type is set to LoadBalancer and externalTrafficPolicy is set to Local. If a value is specified, is in-range, and is not in use, it will be used. If not specified, a value will be automatically allocated. External systems (e.g. load-balancers) can use this port to determine if a given node holds endpoints for this service or not. If this field is specified when creating a Service which does not need it, creation will fail. This field will be wiped when updating a Service to no longer need it (e.g. changing type). This field cannot be updated once set.
publishNotReadyAddresses (boolean)
publishNotReadyAddresses indicates that any agent which deals with endpoints for this Service should disregard any indications of ready/not-ready. The primary use case for setting this field is for a StatefulSet's Headless Service to propagate SRV DNS records for its Pods for the purpose of peer discovery. The Kubernetes controllers that generate Endpoints and EndpointSlice resources for Services interpret this to mean that all endpoints are considered "ready" even if the Pods themselves are not. Agents which consume only Kubernetes generated endpoints through the Endpoints or EndpointSlice resources can safely assume this behavior.
sessionAffinityConfig (SessionAffinityConfig)
sessionAffinityConfig contains the configurations of session affinity.
SessionAffinityConfig represents the configurations of session affinity.
sessionAffinityConfig.clientIP (ClientIPConfig)
clientIP contains the configurations of Client IP based session affinity.
ClientIPConfig represents the configurations of Client IP based session affinity.
timeoutSeconds specifies the seconds of ClientIP type session sticky time. The value must be >0 && <=86400(for 1 day) if ServiceAffinity == "ClientIP". Default value is 10800(for 3 hours).
allocateLoadBalancerNodePorts (boolean)
allocateLoadBalancerNodePorts defines if NodePorts will be automatically allocated for services with type LoadBalancer. Default is "true". It may be set to "false" if the cluster load-balancer does not rely on NodePorts. If the caller requests specific NodePorts (by specifying a value), those requests will be respected, regardless of this field. This field may only be set for services with type LoadBalancer and will be cleared if the type is changed to any other type.
ServiceStatus
ServiceStatus represents the current status of a service.
conditions ([]Condition)
Patch strategy: merge on key type
Map: unique values on key type will be kept during a merge
Current service state
Condition contains details for one aspect of the current state of this API Resource.
conditions.lastTransitionTime (Time), required
lastTransitionTime is the last time the condition transitioned from one status to another. This should be when the underlying condition changed. If that is not known, then using the time when the API field changed is acceptable.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string), required
message is a human readable message indicating details about the transition. This may be an empty string.
conditions.reason (string), required
reason contains a programmatic identifier indicating the reason for the condition's last transition. Producers of specific condition types may define expected values and meanings for this field, and whether the values are considered a guaranteed API. The value should be a CamelCase string. This field may not be empty.
conditions.status (string), required
status of the condition, one of True, False, Unknown.
conditions.type (string), required
type of condition in CamelCase or in foo.example.com/CamelCase.
conditions.observedGeneration (int64)
observedGeneration represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.conditions[x].observedGeneration is 9, the condition is out of date with respect to the current state of the instance.
loadBalancer (LoadBalancerStatus)
LoadBalancer contains the current status of the load-balancer, if one is present.
LoadBalancerStatus represents the status of a load-balancer.
loadBalancer.ingress ([]LoadBalancerIngress)
Ingress is a list containing ingress points for the load-balancer. Traffic intended for the service should be sent to these ingress points.
LoadBalancerIngress represents the status of a load-balancer ingress point: traffic intended for the service should be sent to an ingress point.
loadBalancer.ingress.hostname (string)
Hostname is set for load-balancer ingress points that are DNS based (typically AWS load-balancers)
loadBalancer.ingress.ip (string)
IP is set for load-balancer ingress points that are IP based (typically GCE or OpenStack load-balancers)
loadBalancer.ingress.ports ([]PortStatus)
Atomic: will be replaced during a merge
Ports is a list of records of service ports If used, every port defined in the service should have an entry in it
**
loadBalancer.ingress.ports.port (int32), required
Port is the port number of the service port of which status is recorded here
Protocol is the protocol of the service port of which status is recorded here The supported values are: "TCP", "UDP", "SCTP"
loadBalancer.ingress.ports.error (string)
Error is to record the problem with the service port The format of the error shall comply with the following rules: - built-in error values shall be specified in this file and those shall use
CamelCase names
cloud provider specific error values must have names that comply with the
format foo.example.com/CamelCase.
The set of all endpoints is the union of all subsets. Addresses are placed into subsets according to the IPs they share. A single address with multiple ports, some of which are ready and some of which are not (because they come from different containers) will result in the address being displayed in different subsets for the different ports. No address will appear in both Addresses and NotReadyAddresses in the same subset. Sets of addresses and ports that comprise a service.
*EndpointSubset is a group of addresses with a common set of ports. The expanded set of endpoints is the Cartesian product of Addresses x Ports. For example, given:
IP addresses which offer the related ports that are marked as ready. These endpoints should be considered safe for load balancers and clients to utilize.
EndpointAddress is a tuple that describes single IP address.
subsets.addresses.ip (string), required
The IP of this endpoint. May not be loopback (127.0.0.0/8), link-local (169.254.0.0/16), or link-local multicast ((224.0.0.0/24). IPv6 is also accepted but not fully supported on all platforms. Also, certain kubernetes components, like kube-proxy, are not IPv6 ready.
subsets.addresses.hostname (string)
The Hostname of this endpoint
subsets.addresses.nodeName (string)
Optional: Node hosting this endpoint. This can be used to determine endpoints local to a node.
IP addresses which offer the related ports but are not currently marked as ready because they have not yet finished starting, have recently failed a readiness check, or have recently failed a liveness check.
EndpointAddress is a tuple that describes single IP address.
subsets.notReadyAddresses.ip (string), required
The IP of this endpoint. May not be loopback (127.0.0.0/8), link-local (169.254.0.0/16), or link-local multicast ((224.0.0.0/24). IPv6 is also accepted but not fully supported on all platforms. Also, certain kubernetes components, like kube-proxy, are not IPv6 ready.
subsets.notReadyAddresses.hostname (string)
The Hostname of this endpoint
subsets.notReadyAddresses.nodeName (string)
Optional: Node hosting this endpoint. This can be used to determine endpoints local to a node.
Port numbers available on the related IP addresses.
EndpointPort is a tuple that describes a single port.
subsets.ports.port (int32), required
The port number of the endpoint.
subsets.ports.protocol (string)
The IP protocol for this port. Must be UDP, TCP, or SCTP. Default is TCP.
subsets.ports.name (string)
The name of this port. This must match the 'name' field in the corresponding ServicePort. Must be a DNS_LABEL. Optional only if one port is defined.
subsets.ports.appProtocol (string)
The application protocol for this port. This field follows standard Kubernetes label syntax. Un-prefixed names are reserved for IANA standard service names (as per RFC-6335 and https://www.iana.org/assignments/service-names). Non-standard protocols should use prefixed names such as mycompany.com/my-custom-protocol.
EndpointSlice represents a subset of the endpoints that implement a service.
apiVersion: discovery.k8s.io/v1
import "k8s.io/api/discovery/v1"
EndpointSlice
EndpointSlice represents a subset of the endpoints that implement a service. For a given service there may be multiple EndpointSlice objects, selected by labels, which must be joined to produce the full set of endpoints.
addressType specifies the type of address carried by this EndpointSlice. All addresses in this slice must be the same type. This field is immutable after creation. The following address types are currently supported: * IPv4: Represents an IPv4 Address. * IPv6: Represents an IPv6 Address. * FQDN: Represents a Fully Qualified Domain Name.
endpoints ([]Endpoint), required
Atomic: will be replaced during a merge
endpoints is a list of unique endpoints in this slice. Each slice may include a maximum of 1000 endpoints.
Endpoint represents a single logical "backend" implementing a service.
endpoints.addresses ([]string), required
Set: unique values will be kept during a merge
addresses of this endpoint. The contents of this field are interpreted according to the corresponding EndpointSlice addressType field. Consumers must handle different types of addresses in the context of their own capabilities. This must contain at least one address but no more than 100. These are all assumed to be fungible and clients may choose to only use the first element. Refer to: https://issue.k8s.io/106267
endpoints.conditions (EndpointConditions)
conditions contains information about the current status of the endpoint.
EndpointConditions represents the current condition of an endpoint.
endpoints.conditions.ready (boolean)
ready indicates that this endpoint is prepared to receive traffic, according to whatever system is managing the endpoint. A nil value indicates an unknown state. In most cases consumers should interpret this unknown state as ready. For compatibility reasons, ready should never be "true" for terminating endpoints.
endpoints.conditions.serving (boolean)
serving is identical to ready except that it is set regardless of the terminating state of endpoints. This condition should be set to true for a ready endpoint that is terminating. If nil, consumers should defer to the ready condition. This field can be enabled with the EndpointSliceTerminatingCondition feature gate.
endpoints.conditions.terminating (boolean)
terminating indicates that this endpoint is terminating. A nil value indicates an unknown state. Consumers should interpret this unknown state to mean that the endpoint is not terminating. This field can be enabled with the EndpointSliceTerminatingCondition feature gate.
endpoints.deprecatedTopology (map[string]string)
deprecatedTopology contains topology information part of the v1beta1 API. This field is deprecated, and will be removed when the v1beta1 API is removed (no sooner than kubernetes v1.24). While this field can hold values, it is not writable through the v1 API, and any attempts to write to it will be silently ignored. Topology information can be found in the zone and nodeName fields instead.
endpoints.hints (EndpointHints)
hints contains information associated with how an endpoint should be consumed.
EndpointHints provides hints describing how an endpoint should be consumed.
endpoints.hints.forZones ([]ForZone)
Atomic: will be replaced during a merge
forZones indicates the zone(s) this endpoint should be consumed by to enable topology aware routing.
ForZone provides information about which zones should consume this endpoint.
endpoints.hints.forZones.name (string), required
name represents the name of the zone.
endpoints.hostname (string)
hostname of this endpoint. This field may be used by consumers of endpoints to distinguish endpoints from each other (e.g. in DNS names). Multiple endpoints which use the same hostname should be considered fungible (e.g. multiple A values in DNS). Must be lowercase and pass DNS Label (RFC 1123) validation.
endpoints.nodeName (string)
nodeName represents the name of the Node hosting this endpoint. This can be used to determine endpoints local to a Node.
targetRef is a reference to a Kubernetes object that represents this endpoint.
endpoints.zone (string)
zone is the name of the Zone this endpoint exists in.
ports ([]EndpointPort)
Atomic: will be replaced during a merge
ports specifies the list of network ports exposed by each endpoint in this slice. Each port must have a unique name. When ports is empty, it indicates that there are no defined ports. When a port is defined with a nil port value, it indicates "all ports". Each slice may include a maximum of 100 ports.
EndpointPort represents a Port used by an EndpointSlice
ports.port (int32)
The port number of the endpoint. If this is not specified, ports are not restricted and must be interpreted in the context of the specific consumer.
ports.protocol (string)
The IP protocol for this port. Must be UDP, TCP, or SCTP. Default is TCP.
ports.name (string)
The name of this port. All ports in an EndpointSlice must have a unique name. If the EndpointSlice is dervied from a Kubernetes service, this corresponds to the Service.ports[].name. Name must either be an empty string or pass DNS_LABEL validation: * must be no more than 63 characters long. * must consist of lower case alphanumeric characters or '-'. * must start and end with an alphanumeric character. Default is empty string.
ports.appProtocol (string)
The application protocol for this port. This field follows standard Kubernetes label syntax. Un-prefixed names are reserved for IANA standard service names (as per RFC-6335 and https://www.iana.org/assignments/service-names). Non-standard protocols should use prefixed names such as mycompany.com/my-custom-protocol.
EndpointSliceList
EndpointSliceList represents a list of endpoint slices
Ingress is a collection of rules that allow inbound connections to reach the endpoints defined by a backend.
apiVersion: networking.k8s.io/v1
import "k8s.io/api/networking/v1"
Ingress
Ingress is a collection of rules that allow inbound connections to reach the endpoints defined by a backend. An Ingress can be configured to give services externally-reachable urls, load balance traffic, terminate SSL, offer name based virtual hosting etc.
DefaultBackend is the backend that should handle requests that don't match any rule. If Rules are not specified, DefaultBackend must be specified. If DefaultBackend is not set, the handling of requests that do not match any of the rules will be up to the Ingress controller.
ingressClassName (string)
IngressClassName is the name of an IngressClass cluster resource. Ingress controller implementations use this field to know whether they should be serving this Ingress resource, by a transitive connection (controller -> IngressClass -> Ingress resource). Although the kubernetes.io/ingress.class annotation (simple constant name) was never formally defined, it was widely supported by Ingress controllers to create a direct binding between Ingress controller and Ingress resources. Newly created Ingress resources should prefer using the field. However, even though the annotation is officially deprecated, for backwards compatibility reasons, ingress controllers should still honor that annotation if present.
rules ([]IngressRule)
Atomic: will be replaced during a merge
A list of host rules used to configure the Ingress. If unspecified, or no rule matches, all traffic is sent to the default backend.
IngressRule represents the rules mapping the paths under a specified host to the related backend services. Incoming requests are first evaluated for a host match, then routed to the backend associated with the matching IngressRuleValue.
rules.host (string)
Host is the fully qualified domain name of a network host, as defined by RFC 3986. Note the following deviations from the "host" part of the URI as defined in RFC 3986: 1. IPs are not allowed. Currently an IngressRuleValue can only apply to
the IP in the Spec of the parent Ingress.
2. The : delimiter is not respected because ports are not allowed.
Currently the port of an Ingress is implicitly :80 for http and
:443 for https.
Both these may change in the future. Incoming requests are matched against the host before the IngressRuleValue. If the host is unspecified, the Ingress routes all traffic based on the specified IngressRuleValue.
Host can be "precise" which is a domain name without the terminating dot of a network host (e.g. "foo.bar.com") or "wildcard", which is a domain name prefixed with a single wildcard label (e.g. ".foo.com"). The wildcard character '' must appear by itself as the first DNS label and matches only a single label. You cannot have a wildcard label by itself (e.g. Host == "*"). Requests will be matched against the Host field in the following way: 1. If Host is precise, the request matches this rule if the http host header is equal to Host. 2. If Host is a wildcard, then the request matches this rule if the http host header is to equal to the suffix (removing the first label) of the wildcard rule.
rules.http (HTTPIngressRuleValue)
HTTPIngressRuleValue is a list of http selectors pointing to backends. In the example: http:///? -> backend where where parts of the url correspond to RFC 3986, this resource will be used to match against everything after the last '/' and before the first '?' or '#'.
rules.http.paths ([]HTTPIngressPath), required
Atomic: will be replaced during a merge
A collection of paths that map requests to backends.
HTTPIngressPath associates a path with a backend. Incoming urls matching the path are forwarded to the backend.
Backend defines the referenced service endpoint to which the traffic will be forwarded to.
rules.http.paths.pathType (string), required
PathType determines the interpretation of the Path matching. PathType can be one of the following values: * Exact: Matches the URL path exactly. * Prefix: Matches based on a URL path prefix split by '/'. Matching is
done on a path element by element basis. A path element refers is the
list of labels in the path split by the '/' separator. A request is a
match for path p if every p is an element-wise prefix of p of the
request path. Note that if the last element of the path is a substring
of the last element in request path, it is not a match (e.g. /foo/bar
matches /foo/bar/baz, but does not match /foo/barbaz).
ImplementationSpecific: Interpretation of the Path matching is up to
the IngressClass. Implementations can treat this as a separate PathType
or treat it identically to Prefix or Exact path types.
Implementations are required to support all path types.
rules.http.paths.path (string)
Path is matched against the path of an incoming request. Currently it can contain characters disallowed from the conventional "path" part of a URL as defined by RFC 3986. Paths must begin with a '/' and must be present when using PathType with value "Exact" or "Prefix".
tls ([]IngressTLS)
Atomic: will be replaced during a merge
TLS configuration. Currently the Ingress only supports a single TLS port, 443. If multiple members of this list specify different hosts, they will be multiplexed on the same port according to the hostname specified through the SNI TLS extension, if the ingress controller fulfilling the ingress supports SNI.
IngressTLS describes the transport layer security associated with an Ingress.
tls.hosts ([]string)
Atomic: will be replaced during a merge
Hosts are a list of hosts included in the TLS certificate. The values in this list must match the name/s used in the tlsSecret. Defaults to the wildcard host setting for the loadbalancer controller fulfilling this Ingress, if left unspecified.
tls.secretName (string)
SecretName is the name of the secret used to terminate TLS traffic on port 443. Field is left optional to allow TLS routing based on SNI hostname alone. If the SNI host in a listener conflicts with the "Host" header field used by an IngressRule, the SNI host is used for termination and value of the Host header is used for routing.
IngressBackend
IngressBackend describes all endpoints for a given service and port.
Resource is an ObjectRef to another Kubernetes resource in the namespace of the Ingress object. If resource is specified, a service.Name and service.Port must not be specified. This is a mutually exclusive setting with "Service".
service (IngressServiceBackend)
Service references a Service as a Backend. This is a mutually exclusive setting with "Resource".
IngressServiceBackend references a Kubernetes Service as a Backend.
service.name (string), required
Name is the referenced service. The service must exist in the same namespace as the Ingress object.
service.port (ServiceBackendPort)
Port of the referenced service. A port name or port number is required for a IngressServiceBackend.
ServiceBackendPort is the service port being referenced.
service.port.name (string)
Name is the name of the port on the Service. This is a mutually exclusive setting with "Number".
service.port.number (int32)
Number is the numerical port number (e.g. 80) on the Service. This is a mutually exclusive setting with "Name".
IngressStatus
IngressStatus describe the current state of the Ingress.
loadBalancer (LoadBalancerStatus)
LoadBalancer contains the current status of the load-balancer.
LoadBalancerStatus represents the status of a load-balancer.
loadBalancer.ingress ([]LoadBalancerIngress)
Ingress is a list containing ingress points for the load-balancer. Traffic intended for the service should be sent to these ingress points.
LoadBalancerIngress represents the status of a load-balancer ingress point: traffic intended for the service should be sent to an ingress point.
loadBalancer.ingress.hostname (string)
Hostname is set for load-balancer ingress points that are DNS based (typically AWS load-balancers)
loadBalancer.ingress.ip (string)
IP is set for load-balancer ingress points that are IP based (typically GCE or OpenStack load-balancers)
loadBalancer.ingress.ports ([]PortStatus)
Atomic: will be replaced during a merge
Ports is a list of records of service ports If used, every port defined in the service should have an entry in it
**
loadBalancer.ingress.ports.port (int32), required
Port is the port number of the service port of which status is recorded here
Protocol is the protocol of the service port of which status is recorded here The supported values are: "TCP", "UDP", "SCTP"
loadBalancer.ingress.ports.error (string)
Error is to record the problem with the service port The format of the error shall comply with the following rules: - built-in error values shall be specified in this file and those shall use
CamelCase names
cloud provider specific error values must have names that comply with the
format foo.example.com/CamelCase.
IngressClass represents the class of the Ingress, referenced by the Ingress Spec.
apiVersion: networking.k8s.io/v1
import "k8s.io/api/networking/v1"
IngressClass
IngressClass represents the class of the Ingress, referenced by the Ingress Spec. The ingressclass.kubernetes.io/is-default-class annotation can be used to indicate that an IngressClass should be considered default. When a single IngressClass resource has this annotation set to true, new Ingress resources without a class specified will be assigned this default class.
IngressClassSpec provides information about the class of an Ingress.
controller (string)
Controller refers to the name of the controller that should handle this class. This allows for different "flavors" that are controlled by the same controller. For example, you may have different Parameters for the same implementing controller. This should be specified as a domain-prefixed path no more than 250 characters in length, e.g. "acme.io/ingress-controller". This field is immutable.
parameters (IngressClassParametersReference)
Parameters is a link to a custom resource containing additional configuration for the controller. This is optional if the controller does not require extra parameters.
IngressClassParametersReference identifies an API object. This can be used to specify a cluster or namespace-scoped resource.
parameters.kind (string), required
Kind is the type of resource being referenced.
parameters.name (string), required
Name is the name of resource being referenced.
parameters.apiGroup (string)
APIGroup is the group for the resource being referenced. If APIGroup is not specified, the specified Kind must be in the core API group. For any other third-party types, APIGroup is required.
parameters.namespace (string)
Namespace is the namespace of the resource being referenced. This field is required when scope is set to "Namespace" and must be unset when scope is set to "Cluster".
parameters.scope (string)
Scope represents if this refers to a cluster or namespace scoped resource. This may be set to "Cluster" (default) or "Namespace".
IngressClassList
IngressClassList is a collection of IngressClasses.
BinaryData contains the binary data. Each key must consist of alphanumeric characters, '-', '_' or '.'. BinaryData can contain byte sequences that are not in the UTF-8 range. The keys stored in BinaryData must not overlap with the ones in the Data field, this is enforced during validation process. Using this field will require 1.10+ apiserver and kubelet.
data (map[string]string)
Data contains the configuration data. Each key must consist of alphanumeric characters, '-', '_' or '.'. Values with non-UTF-8 byte sequences must use the BinaryData field. The keys stored in Data must not overlap with the keys in the BinaryData field, this is enforced during validation process.
immutable (boolean)
Immutable, if set to true, ensures that data stored in the ConfigMap cannot be updated (only object metadata can be modified). If not set to true, the field can be modified at any time. Defaulted to nil.
ConfigMapList
ConfigMapList is a resource containing a list of ConfigMap objects.
Data contains the secret data. Each key must consist of alphanumeric characters, '-', '_' or '.'. The serialized form of the secret data is a base64 encoded string, representing the arbitrary (possibly non-string) data value here. Described in https://tools.ietf.org/html/rfc4648#section-4
immutable (boolean)
Immutable, if set to true, ensures that data stored in the Secret cannot be updated (only object metadata can be modified). If not set to true, the field can be modified at any time. Defaulted to nil.
stringData (map[string]string)
stringData allows specifying non-binary secret data in string form. It is provided as a write-only input field for convenience. All keys and values are merged into the data field on write, overwriting any existing values. The stringData field is never output when reading from the API.
PersistentVolumeClaimVolumeSource references the user's PVC in the same namespace. This volume finds the bound PV and mounts that volume for the pod. A PersistentVolumeClaimVolumeSource is, essentially, a wrapper around another type of volume that is owned by someone else (the system).
readOnly Will force the ReadOnly setting in VolumeMounts. Default false.
Projections
configMap (ConfigMapVolumeSource)
configMap represents a configMap that should populate this volume
*Adapts a ConfigMap into a volume.
The contents of the target ConfigMap's Data field will be presented in a volume as files using the keys in the Data field as the file names, unless the items element is populated with specific mappings of keys to paths. ConfigMap volumes support ownership management and SELinux relabeling.*
optional specify whether the ConfigMap or its keys must be defined
configMap.defaultMode (int32)
defaultMode is optional: mode bits used to set permissions on created files by default. Must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. Defaults to 0644. Directories within the path are not affected by this setting. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
items if unspecified, each key-value pair in the Data field of the referenced ConfigMap will be projected into the volume as a file whose name is the key and content is the value. If specified, the listed keys will be projected into the specified paths, and unlisted keys will not be present. If a key is specified which is not present in the ConfigMap, the volume setup will error unless it is marked optional. Paths must be relative and may not contain the '..' path or start with '..'.
The contents of the target Secret's Data field will be presented in a volume as files using the keys in the Data field as the file names. Secret volumes support ownership management and SELinux relabeling.*
optional field specify whether the Secret or its keys must be defined
secret.defaultMode (int32)
defaultMode is Optional: mode bits used to set permissions on created files by default. Must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. Defaults to 0644. Directories within the path are not affected by this setting. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
items If unspecified, each key-value pair in the Data field of the referenced Secret will be projected into the volume as a file whose name is the key and content is the value. If specified, the listed keys will be projected into the specified paths, and unlisted keys will not be present. If a key is specified which is not present in the Secret, the volume setup will error unless it is marked optional. Paths must be relative and may not contain the '..' path or start with '..'.
downwardAPI (DownwardAPIVolumeSource)
downwardAPI represents downward API about the pod that should populate this volume
DownwardAPIVolumeSource represents a volume containing downward API info. Downward API volumes support ownership management and SELinux relabeling.
downwardAPI.defaultMode (int32)
Optional: mode bits to use on created files by default. Must be a Optional: mode bits used to set permissions on created files by default. Must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. Defaults to 0644. Directories within the path are not affected by this setting. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
projected items for all in one resources secrets, configmaps, and downward API
Represents a projected volume source
projected.defaultMode (int32)
defaultMode are the mode bits used to set permissions on created files by default. Must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. Directories within the path are not affected by this setting. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
projected.sources ([]VolumeProjection)
sources is the list of volume projections
Projection that may be projected along with other supported volume types
projected.sources.configMap (ConfigMapProjection)
configMap information about the configMap data to project
*Adapts a ConfigMap into a projected volume.
The contents of the target ConfigMap's Data field will be presented in a projected volume as files using the keys in the Data field as the file names, unless the items element is populated with specific mappings of keys to paths. Note that this is identical to a configmap volume source without the default mode.*
items if unspecified, each key-value pair in the Data field of the referenced ConfigMap will be projected into the volume as a file whose name is the key and content is the value. If specified, the listed keys will be projected into the specified paths, and unlisted keys will not be present. If a key is specified which is not present in the ConfigMap, the volume setup will error unless it is marked optional. Paths must be relative and may not contain the '..' path or start with '..'.
downwardAPI information about the downwardAPI data to project
Represents downward API info for projecting into a projected volume. Note that this is identical to a downwardAPI volume source without the default mode.
secret information about the secret data to project
*Adapts a secret into a projected volume.
The contents of the target Secret's Data field will be presented in a projected volume as files using the keys in the Data field as the file names. Note that this is identical to a secret volume source without the default mode.*
items if unspecified, each key-value pair in the Data field of the referenced Secret will be projected into the volume as a file whose name is the key and content is the value. If specified, the listed keys will be projected into the specified paths, and unlisted keys will not be present. If a key is specified which is not present in the Secret, the volume setup will error unless it is marked optional. Paths must be relative and may not contain the '..' path or start with '..'.
serviceAccountToken is information about the serviceAccountToken data to project
ServiceAccountTokenProjection represents a projected service account token volume. This projection can be used to insert a service account token into the pods runtime filesystem for use against APIs (Kubernetes API Server or otherwise).
audience is the intended audience of the token. A recipient of a token must identify itself with an identifier specified in the audience of the token, and otherwise should reject the token. The audience defaults to the identifier of the apiserver.
expirationSeconds is the requested duration of validity of the service account token. As the token approaches expiration, the kubelet volume plugin will proactively rotate the service account token. The kubelet will start trying to rotate the token if the token is older than 80 percent of its time to live or if the token is older than 24 hours.Defaults to 1 hour and must be at least 10 minutes.
Represents an empty directory for a pod. Empty directory volumes support ownership management and SELinux relabeling.
emptyDir.medium (string)
medium represents what type of storage medium should back this directory. The default is "" which means to use the node's default medium. Must be an empty string (default) or Memory. More info: https://kubernetes.io/docs/concepts/storage/volumes#emptydir
sizeLimit is the total amount of local storage required for this EmptyDir volume. The size limit is also applicable for memory medium. The maximum usage on memory medium EmptyDir would be the minimum value between the SizeLimit specified here and the sum of memory limits of all containers in a pod. The default is nil which means that the limit is undefined. More info: http://kubernetes.io/docs/user-guide/volumes#emptydir
hostPath (HostPathVolumeSource)
hostPath represents a pre-existing file or directory on the host machine that is directly exposed to the container. This is generally used for system agents or other privileged things that are allowed to see the host machine. Most containers will NOT need this. More info: https://kubernetes.io/docs/concepts/storage/volumes#hostpath
Represents a host path mapped into a pod. Host path volumes do not support ownership management or SELinux relabeling.
An AWS EBS disk must exist before mounting to a container. The disk must also be in the same AWS zone as the kubelet. An AWS EBS disk can only be mounted as read/write once. AWS EBS volumes support ownership management and SELinux relabeling.*
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#awselasticblockstore
awsElasticBlockStore.partition (int32)
partition is the partition in the volume that you want to mount. If omitted, the default is to mount by volume name. Examples: For volume /dev/sda1, you specify the partition as "1". Similarly, the volume partition for /dev/sda is "0" (or you can leave the property empty).
azureDisk represents an Azure Data Disk mount on the host and bind mount to the pod.
AzureDisk represents an Azure Data Disk mount on the host and bind mount to the pod.
azureDisk.diskName (string), required
diskName is the Name of the data disk in the blob storage
azureDisk.diskURI (string), required
diskURI is the URI of data disk in the blob storage
azureDisk.cachingMode (string)
cachingMode is the Host Caching mode: None, Read Only, Read Write.
azureDisk.fsType (string)
fsType is Filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
azureDisk.kind (string)
kind expected values are Shared: multiple blob disks per storage account Dedicated: single blob disk per storage account Managed: azure managed data disk (only in managed availability set). defaults to shared
azureDisk.readOnly (boolean)
readOnly Defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
azureFile (AzureFileVolumeSource)
azureFile represents an Azure File Service mount on the host and bind mount to the pod.
AzureFile represents an Azure File Service mount on the host and bind mount to the pod.
azureFile.secretName (string), required
secretName is the name of secret that contains Azure Storage Account Name and Key
azureFile.shareName (string), required
shareName is the azure share Name
azureFile.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
cephfs (CephFSVolumeSource)
cephFS represents a Ceph FS mount on the host that shares a pod's lifetime
Represents a Ceph Filesystem mount that lasts the lifetime of a pod Cephfs volumes do not support ownership management or SELinux relabeling.
Represents a cinder volume resource in Openstack. A Cinder volume must exist before mounting to a container. The volume must also be in the same region as the kubelet. Cinder volumes support ownership management and SELinux relabeling.
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://examples.k8s.io/mysql-cinder-pd/README.md
secretRef is optional: points to a secret object containing parameters used to connect to OpenStack.
csi (CSIVolumeSource)
csi (Container Storage Interface) represents ephemeral storage that is handled by certain external CSI drivers (Beta feature).
Represents a source location of a volume to mount, managed by an external CSI driver
csi.driver (string), required
driver is the name of the CSI driver that handles this volume. Consult with your admin for the correct name as registered in the cluster.
csi.fsType (string)
fsType to mount. Ex. "ext4", "xfs", "ntfs". If not provided, the empty value is passed to the associated CSI driver which will determine the default filesystem to apply.
nodePublishSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI NodePublishVolume and NodeUnpublishVolume calls. This field is optional, and may be empty if no secret is required. If the secret object contains more than one secret, all secret references are passed.
csi.readOnly (boolean)
readOnly specifies a read-only configuration for the volume. Defaults to false (read/write).
csi.volumeAttributes (map[string]string)
volumeAttributes stores driver-specific properties that are passed to the CSI driver. Consult your driver's documentation for supported values.
ephemeral (EphemeralVolumeSource)
ephemeral represents a volume that is handled by a cluster storage driver. The volume's lifecycle is tied to the pod that defines it - it will be created before the pod starts, and deleted when the pod is removed.
Use this if: a) the volume is only needed while the pod runs, b) features of normal volumes like restoring from snapshot or capacity
tracking are needed,
c) the storage driver is specified through a storage class, and d) the storage driver supports dynamic volume provisioning through
a PersistentVolumeClaim (see EphemeralVolumeSource for more
information on the connection between this volume type
and PersistentVolumeClaim).
Use PersistentVolumeClaim or one of the vendor-specific APIs for volumes that persist for longer than the lifecycle of an individual pod.
Use CSI for light-weight local ephemeral volumes if the CSI driver is meant to be used that way - see the documentation of the driver for more information.
A pod can use both types of ephemeral volumes and persistent volumes at the same time.
Represents an ephemeral volume that is handled by a normal storage driver.
Will be used to create a stand-alone PVC to provision the volume. The pod in which this EphemeralVolumeSource is embedded will be the owner of the PVC, i.e. the PVC will be deleted together with the pod. The name of the PVC will be \<pod name>-\<volume name> where \<volume name> is the name from the PodSpec.Volumes array entry. Pod validation will reject the pod if the concatenated name is not valid for a PVC (for example, too long).
An existing PVC with that name that is not owned by the pod will not be used for the pod to avoid using an unrelated volume by mistake. Starting the pod is then blocked until the unrelated PVC is removed. If such a pre-created PVC is meant to be used by the pod, the PVC has to updated with an owner reference to the pod once the pod exists. Normally this should not be necessary, but it may be useful when manually reconstructing a broken cluster.
This field is read-only and no changes will be made by Kubernetes to the PVC after it has been created.
Required, must not be nil.
PersistentVolumeClaimTemplate is used to produce PersistentVolumeClaim objects as part of an EphemeralVolumeSource.
The specification for the PersistentVolumeClaim. The entire content is copied unchanged into the PVC that gets created from this template. The same fields as in a PersistentVolumeClaim are also valid here.
May contain labels and annotations that will be copied into the PVC when creating it. No other fields are allowed and will be rejected during validation.
fc (FCVolumeSource)
fc represents a Fibre Channel resource that is attached to a kubelet's host machine and then exposed to the pod.
Represents a Fibre Channel volume. Fibre Channel volumes can only be mounted as read/write once. Fibre Channel volumes support ownership management and SELinux relabeling.
fc.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
fc.lun (int32)
lun is Optional: FC target lun number
fc.readOnly (boolean)
readOnly is Optional: Defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
fc.targetWWNs ([]string)
targetWWNs is Optional: FC target worldwide names (WWNs)
fc.wwids ([]string)
wwids Optional: FC volume world wide identifiers (wwids) Either wwids or combination of targetWWNs and lun must be set, but not both simultaneously.
flexVolume (FlexVolumeSource)
flexVolume represents a generic volume resource that is provisioned/attached using an exec based plugin.
FlexVolume represents a generic volume resource that is provisioned/attached using an exec based plugin.
flexVolume.driver (string), required
driver is the name of the driver to use for this volume.
flexVolume.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". The default filesystem depends on FlexVolume script.
flexVolume.options (map[string]string)
options is Optional: this field holds extra command options if any.
flexVolume.readOnly (boolean)
readOnly is Optional: defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
secretRef is Optional: secretRef is reference to the secret object containing sensitive information to pass to the plugin scripts. This may be empty if no secret object is specified. If the secret object contains more than one secret, all secrets are passed to the plugin scripts.
flocker (FlockerVolumeSource)
flocker represents a Flocker volume attached to a kubelet's host machine. This depends on the Flocker control service being running
Represents a Flocker volume mounted by the Flocker agent. One and only one of datasetName and datasetUUID should be set. Flocker volumes do not support ownership management or SELinux relabeling.
flocker.datasetName (string)
datasetName is Name of the dataset stored as metadata -> name on the dataset for Flocker should be considered as deprecated
flocker.datasetUUID (string)
datasetUUID is the UUID of the dataset. This is unique identifier of a Flocker dataset
*Represents a Persistent Disk resource in Google Compute Engine.
A GCE PD must exist before mounting to a container. The disk must also be in the same GCE project and zone as the kubelet. A GCE PD can only be mounted as read/write once or read-only many times. GCE PDs support ownership management and SELinux relabeling.*
fsType is filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#gcepersistentdisk
gcePersistentDisk.partition (int32)
partition is the partition in the volume that you want to mount. If omitted, the default is to mount by volume name. Examples: For volume /dev/sda1, you specify the partition as "1". Similarly, the volume partition for /dev/sda is "0" (or you can leave the property empty). More info: https://kubernetes.io/docs/concepts/storage/volumes#gcepersistentdisk
Represents an ISCSI disk. ISCSI volumes can only be mounted as read/write once. ISCSI volumes support ownership management and SELinux relabeling.
iscsi.iqn (string), required
iqn is the target iSCSI Qualified Name.
iscsi.lun (int32), required
lun represents iSCSI Target Lun number.
iscsi.targetPortal (string), required
targetPortal is iSCSI Target Portal. The Portal is either an IP or ip_addr:port if the port is other than default (typically TCP ports 860 and 3260).
iscsi.chapAuthDiscovery (boolean)
chapAuthDiscovery defines whether support iSCSI Discovery CHAP authentication
iscsi.chapAuthSession (boolean)
chapAuthSession defines whether support iSCSI Session CHAP authentication
iscsi.fsType (string)
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#iscsi
iscsi.initiatorName (string)
initiatorName is the custom iSCSI Initiator Name. If initiatorName is specified with iscsiInterface simultaneously, new iSCSI interface <target portal>:<volume name> will be created for the connection.
iscsi.iscsiInterface (string)
iscsiInterface is the interface Name that uses an iSCSI transport. Defaults to 'default' (tcp).
iscsi.portals ([]string)
portals is the iSCSI Target Portal List. The portal is either an IP or ip_addr:port if the port is other than default (typically TCP ports 860 and 3260).
iscsi.readOnly (boolean)
readOnly here will force the ReadOnly setting in VolumeMounts. Defaults to false.
photonPersistentDisk represents a PhotonController persistent disk attached and mounted on kubelets host machine
Represents a Photon Controller persistent disk resource.
photonPersistentDisk.pdID (string), required
pdID is the ID that identifies Photon Controller persistent disk
photonPersistentDisk.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
portworxVolume (PortworxVolumeSource)
portworxVolume represents a portworx volume attached and mounted on kubelets host machine
PortworxVolumeSource represents a Portworx volume resource.
portworxVolume.volumeID (string), required
volumeID uniquely identifies a Portworx volume
portworxVolume.fsType (string)
fSType represents the filesystem type to mount Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs". Implicitly inferred to be "ext4" if unspecified.
portworxVolume.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
quobyte (QuobyteVolumeSource)
quobyte represents a Quobyte mount on the host that shares a pod's lifetime
Represents a Quobyte mount that lasts the lifetime of a pod. Quobyte volumes do not support ownership management or SELinux relabeling.
quobyte.registry (string), required
registry represents a single or multiple Quobyte Registry services specified as a string as host:port pair (multiple entries are separated with commas) which acts as the central registry for volumes
quobyte.volume (string), required
volume is a string that references an already created Quobyte volume by name.
quobyte.group (string)
group to map volume access to Default is no group
quobyte.readOnly (boolean)
readOnly here will force the Quobyte volume to be mounted with read-only permissions. Defaults to false.
quobyte.tenant (string)
tenant owning the given Quobyte volume in the Backend Used with dynamically provisioned Quobyte volumes, value is set by the plugin
quobyte.user (string)
user to map volume access to Defaults to serivceaccount user
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#rbd
secretRef references to the secret for ScaleIO user and other sensitive information. If this is not provided, Login operation will fail.
scaleIO.system (string), required
system is the name of the storage system as configured in ScaleIO.
scaleIO.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Default is "xfs".
scaleIO.protectionDomain (string)
protectionDomain is the name of the ScaleIO Protection Domain for the configured storage.
scaleIO.readOnly (boolean)
readOnly Defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
scaleIO.sslEnabled (boolean)
sslEnabled Flag enable/disable SSL communication with Gateway, default false
scaleIO.storageMode (string)
storageMode indicates whether the storage for a volume should be ThickProvisioned or ThinProvisioned. Default is ThinProvisioned.
scaleIO.storagePool (string)
storagePool is the ScaleIO Storage Pool associated with the protection domain.
scaleIO.volumeName (string)
volumeName is the name of a volume already created in the ScaleIO system that is associated with this volume source.
storageos (StorageOSVolumeSource)
storageOS represents a StorageOS volume attached and mounted on Kubernetes nodes.
Represents a StorageOS persistent volume resource.
storageos.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
storageos.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
secretRef specifies the secret to use for obtaining the StorageOS API credentials. If not specified, default values will be attempted.
storageos.volumeName (string)
volumeName is the human-readable name of the StorageOS volume. Volume names are only unique within a namespace.
storageos.volumeNamespace (string)
volumeNamespace specifies the scope of the volume within StorageOS. If no namespace is specified then the Pod's namespace will be used. This allows the Kubernetes name scoping to be mirrored within StorageOS for tighter integration. Set VolumeName to any name to override the default behaviour. Set to "default" if you are not using namespaces within StorageOS. Namespaces that do not pre-exist within StorageOS will be created.
vsphereVolume (VsphereVirtualDiskVolumeSource)
vsphereVolume represents a vSphere volume attached and mounted on kubelets host machine
Represents a vSphere volume resource.
vsphereVolume.volumePath (string), required
volumePath is the path that identifies vSphere volume vmdk
vsphereVolume.fsType (string)
fsType is filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
vsphereVolume.storagePolicyID (string)
storagePolicyID is the storage Policy Based Management (SPBM) profile ID associated with the StoragePolicyName.
vsphereVolume.storagePolicyName (string)
storagePolicyName is the storage Policy Based Management (SPBM) profile name.
Deprecated
gitRepo (GitRepoVolumeSource)
gitRepo represents a git repository at a particular revision. DEPRECATED: GitRepo is deprecated. To provision a container with a git repo, mount an EmptyDir into an InitContainer that clones the repo using git, then mount the EmptyDir into the Pod's container.
*Represents a volume that is populated with the contents of a git repository. Git repo volumes do not support ownership management. Git repo volumes support SELinux relabeling.
DEPRECATED: GitRepo is deprecated. To provision a container with a git repo, mount an EmptyDir into an InitContainer that clones the repo using git, then mount the EmptyDir into the Pod's container.*
gitRepo.repository (string), required
repository is the URL
gitRepo.directory (string)
directory is the target directory name. Must not contain or start with '..'. If '.' is supplied, the volume directory will be the git repository. Otherwise, if specified, the volume will contain the git repository in the subdirectory with the given name.
gitRepo.revision (string)
revision is the commit hash for the specified revision.
DownwardAPIVolumeFile
DownwardAPIVolumeFile represents information to create the file containing the pod field
path (string), required
Required: Path is the relative path name of the file to be created. Must not be absolute or contain the '..' path. Must be utf-8 encoded. The first item of the relative path must not start with '..'
Required: Selects a field of the pod: only annotations, labels, name and namespace are supported.
mode (int32)
Optional: mode bits used to set permissions on this file, must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. If not specified, the volume defaultMode will be used. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
Selects a resource of the container: only resources limits and requests (limits.cpu, limits.memory, requests.cpu and requests.memory) are currently supported.
KeyToPath
Maps a string key to a path within a volume.
key (string), required
key is the key to project.
path (string), required
path is the relative path of the file to map the key to. May not be an absolute path. May not contain the path element '..'. May not start with the string '..'.
mode (int32)
mode is Optional: mode bits used to set permissions on this file. Must be an octal value between 0000 and 0777 or a decimal value between 0 and 511. YAML accepts both octal and decimal values, JSON requires decimal values for mode bits. If not specified, the volume defaultMode will be used. This might be in conflict with other options that affect the file mode, like fsGroup, and the result can be other mode bits set.
6.5.3.4 - PersistentVolumeClaim
PersistentVolumeClaim is a user's request for and claim to a persistent volume.
apiVersion: v1
import "k8s.io/api/core/v1"
PersistentVolumeClaim
PersistentVolumeClaim is a user's request for and claim to a persistent volume
selector is a label query over volumes to consider for binding.
resources (ResourceRequirements)
resources represents the minimum resources the volume should have. If RecoverVolumeExpansionFailure feature is enabled users are allowed to specify resource requirements that are lower than previous value but must still be higher than capacity recorded in the status field of the claim. More info: https://kubernetes.io/docs/concepts/storage/persistent-volumes#resources
ResourceRequirements describes the compute resource requirements.
dataSource field can be used to specify either: * An existing VolumeSnapshot object (snapshot.storage.k8s.io/VolumeSnapshot) * An existing PVC (PersistentVolumeClaim) If the provisioner or an external controller can support the specified data source, it will create a new volume based on the contents of the specified data source. If the AnyVolumeDataSource feature gate is enabled, this field will always have the same contents as the DataSourceRef field.
dataSourceRef specifies the object from which to populate the volume with data, if a non-empty volume is desired. This may be any local object from a non-empty API group (non core object) or a PersistentVolumeClaim object. When this field is specified, volume binding will only succeed if the type of the specified object matches some installed volume populator or dynamic provisioner. This field will replace the functionality of the DataSource field and as such if both fields are non-empty, they must have the same value. For backwards compatibility, both fields (DataSource and DataSourceRef) will be set to the same value automatically if one of them is empty and the other is non-empty. There are two important differences between DataSource and DataSourceRef: * While DataSource only allows two specific types of objects, DataSourceRef
allows any non-core object, as well as PersistentVolumeClaim objects.
While DataSource ignores disallowed values (dropping them), DataSourceRef
preserves all values, and generates an error if a disallowed value is
specified.
(Beta) Using this field requires the AnyVolumeDataSource feature gate to be enabled.
PersistentVolumeClaimStatus
PersistentVolumeClaimStatus is the current status of a persistent volume claim.
allocatedResources is the storage resource within AllocatedResources tracks the capacity allocated to a PVC. It may be larger than the actual capacity when a volume expansion operation is requested. For storage quota, the larger value from allocatedResources and PVC.spec.resources is used. If allocatedResources is not set, PVC.spec.resources alone is used for quota calculation. If a volume expansion capacity request is lowered, allocatedResources is only lowered if there are no expansion operations in progress and if the actual volume capacity is equal or lower than the requested capacity. This is an alpha field and requires enabling RecoverVolumeExpansionFailure feature.
capacity represents the actual resources of the underlying volume.
conditions ([]PersistentVolumeClaimCondition)
Patch strategy: merge on key type
conditions is the current Condition of persistent volume claim. If underlying persistent volume is being resized then the Condition will be set to 'ResizeStarted'.
PersistentVolumeClaimCondition contails details about state of pvc
conditions.status (string), required
conditions.type (string), required
conditions.lastProbeTime (Time)
lastProbeTime is the time we probed the condition.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastTransitionTime (Time)
lastTransitionTime is the time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is the human-readable message indicating details about last transition.
conditions.reason (string)
reason is a unique, this should be a short, machine understandable string that gives the reason for condition's last transition. If it reports "ResizeStarted" that means the underlying persistent volume is being resized.
phase (string)
phase represents the current phase of PersistentVolumeClaim.
resizeStatus (string)
resizeStatus stores status of resize operation. ResizeStatus is not set by default but when expansion is complete resizeStatus is set to empty string by resize controller or kubelet. This is an alpha field and requires enabling RecoverVolumeExpansionFailure feature.
PersistentVolumeClaimList
PersistentVolumeClaimList is a list of PersistentVolumeClaim items.
nodeAffinity defines constraints that limit what nodes this volume can be accessed from. This field influences the scheduling of pods that use this volume.
VolumeNodeAffinity defines constraints that limit what nodes this volume can be accessed from.
nodeAffinity.required (NodeSelector)
required specifies hard node constraints that must be met.
A node selector represents the union of the results of one or more label queries over a set of nodes; that is, it represents the OR of the selectors represented by the node selector terms.
Required. A list of node selector terms. The terms are ORed.
A null or empty node selector term matches no objects. The requirements of them are ANDed. The TopologySelectorTerm type implements a subset of the NodeSelectorTerm.
A list of node selector requirements by node's fields.
persistentVolumeReclaimPolicy (string)
persistentVolumeReclaimPolicy defines what happens to a persistent volume when released from its claim. Valid options are Retain (default for manually created PersistentVolumes), Delete (default for dynamically provisioned PersistentVolumes), and Recycle (deprecated). Recycle must be supported by the volume plugin underlying this PersistentVolume. More info: https://kubernetes.io/docs/concepts/storage/persistent-volumes#reclaiming
storageClassName (string)
storageClassName is the name of StorageClass to which this persistent volume belongs. Empty value means that this volume does not belong to any StorageClass.
volumeMode (string)
volumeMode defines if a volume is intended to be used with a formatted filesystem or to remain in raw block state. Value of Filesystem is implied when not included in spec.
Local
hostPath (HostPathVolumeSource)
hostPath represents a directory on the host. Provisioned by a developer or tester. This is useful for single-node development and testing only! On-host storage is not supported in any way and WILL NOT WORK in a multi-node cluster. More info: https://kubernetes.io/docs/concepts/storage/volumes#hostpath
Represents a host path mapped into a pod. Host path volumes do not support ownership management or SELinux relabeling.
local represents directly-attached storage with node affinity
Local represents directly-attached storage with node affinity (Beta feature)
local.path (string), required
path of the full path to the volume on the node. It can be either a directory or block device (disk, partition, ...).
local.fsType (string)
fsType is the filesystem type to mount. It applies only when the Path is a block device. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". The default value is to auto-select a filesystem if unspecified.
An AWS EBS disk must exist before mounting to a container. The disk must also be in the same AWS zone as the kubelet. An AWS EBS disk can only be mounted as read/write once. AWS EBS volumes support ownership management and SELinux relabeling.*
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#awselasticblockstore
awsElasticBlockStore.partition (int32)
partition is the partition in the volume that you want to mount. If omitted, the default is to mount by volume name. Examples: For volume /dev/sda1, you specify the partition as "1". Similarly, the volume partition for /dev/sda is "0" (or you can leave the property empty).
azureDisk represents an Azure Data Disk mount on the host and bind mount to the pod.
AzureDisk represents an Azure Data Disk mount on the host and bind mount to the pod.
azureDisk.diskName (string), required
diskName is the Name of the data disk in the blob storage
azureDisk.diskURI (string), required
diskURI is the URI of data disk in the blob storage
azureDisk.cachingMode (string)
cachingMode is the Host Caching mode: None, Read Only, Read Write.
azureDisk.fsType (string)
fsType is Filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
azureDisk.kind (string)
kind expected values are Shared: multiple blob disks per storage account Dedicated: single blob disk per storage account Managed: azure managed data disk (only in managed availability set). defaults to shared
azureDisk.readOnly (boolean)
readOnly Defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
azureFile (AzureFilePersistentVolumeSource)
azureFile represents an Azure File Service mount on the host and bind mount to the pod.
AzureFile represents an Azure File Service mount on the host and bind mount to the pod.
azureFile.secretName (string), required
secretName is the name of secret that contains Azure Storage Account Name and Key
azureFile.shareName (string), required
shareName is the azure Share Name
azureFile.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
azureFile.secretNamespace (string)
secretNamespace is the namespace of the secret that contains Azure Storage Account Name and Key default is the same as the Pod
cephfs (CephFSPersistentVolumeSource)
cephFS represents a Ceph FS mount on the host that shares a pod's lifetime
Represents a Ceph Filesystem mount that lasts the lifetime of a pod Cephfs volumes do not support ownership management or SELinux relabeling.
Represents a cinder volume resource in Openstack. A Cinder volume must exist before mounting to a container. The volume must also be in the same region as the kubelet. Cinder volumes support ownership management and SELinux relabeling.
fsType Filesystem type to mount. Must be a filesystem type supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://examples.k8s.io/mysql-cinder-pd/README.md
secretRef is Optional: points to a secret object containing parameters used to connect to OpenStack.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
cinder.secretRef.name (string)
name is unique within a namespace to reference a secret resource.
cinder.secretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi (CSIPersistentVolumeSource)
csi represents storage that is handled by an external CSI driver (Beta feature).
Represents storage that is managed by an external CSI volume driver (Beta feature)
csi.driver (string), required
driver is the name of the driver to use for this volume. Required.
csi.volumeHandle (string), required
volumeHandle is the unique volume name returned by the CSI volume plugin’s CreateVolume to refer to the volume on all subsequent calls. Required.
csi.controllerExpandSecretRef (SecretReference)
controllerExpandSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI ControllerExpandVolume call. This is an beta field and requires enabling ExpandCSIVolumes feature gate. This field is optional, and may be empty if no secret is required. If the secret object contains more than one secret, all secrets are passed.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
csi.controllerExpandSecretRef.name (string)
name is unique within a namespace to reference a secret resource.
csi.controllerExpandSecretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi.controllerPublishSecretRef (SecretReference)
controllerPublishSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI ControllerPublishVolume and ControllerUnpublishVolume calls. This field is optional, and may be empty if no secret is required. If the secret object contains more than one secret, all secrets are passed.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
csi.controllerPublishSecretRef.name (string)
name is unique within a namespace to reference a secret resource.
csi.controllerPublishSecretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi.fsType (string)
fsType to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs".
csi.nodeExpandSecretRef (SecretReference)
nodeExpandSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI NodeExpandVolume call. This is an alpha field and requires enabling CSINodeExpandSecret feature gate. This field is optional, may be omitted if no secret is required. If the secret object contains more than one secret, all secrets are passed.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
csi.nodeExpandSecretRef.name (string)
name is unique within a namespace to reference a secret resource.
csi.nodeExpandSecretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi.nodePublishSecretRef (SecretReference)
nodePublishSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI NodePublishVolume and NodeUnpublishVolume calls. This field is optional, and may be empty if no secret is required. If the secret object contains more than one secret, all secrets are passed.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
csi.nodePublishSecretRef.name (string)
name is unique within a namespace to reference a secret resource.
csi.nodePublishSecretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi.nodeStageSecretRef (SecretReference)
nodeStageSecretRef is a reference to the secret object containing sensitive information to pass to the CSI driver to complete the CSI NodeStageVolume and NodeStageVolume and NodeUnstageVolume calls. This field is optional, and may be empty if no secret is required. If the secret object contains more than one secret, all secrets are passed.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
csi.nodeStageSecretRef.name (string)
name is unique within a namespace to reference a secret resource.
csi.nodeStageSecretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
csi.readOnly (boolean)
readOnly value to pass to ControllerPublishVolumeRequest. Defaults to false (read/write).
csi.volumeAttributes (map[string]string)
volumeAttributes of the volume to publish.
fc (FCVolumeSource)
fc represents a Fibre Channel resource that is attached to a kubelet's host machine and then exposed to the pod.
Represents a Fibre Channel volume. Fibre Channel volumes can only be mounted as read/write once. Fibre Channel volumes support ownership management and SELinux relabeling.
fc.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
fc.lun (int32)
lun is Optional: FC target lun number
fc.readOnly (boolean)
readOnly is Optional: Defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
fc.targetWWNs ([]string)
targetWWNs is Optional: FC target worldwide names (WWNs)
fc.wwids ([]string)
wwids Optional: FC volume world wide identifiers (wwids) Either wwids or combination of targetWWNs and lun must be set, but not both simultaneously.
flexVolume (FlexPersistentVolumeSource)
flexVolume represents a generic volume resource that is provisioned/attached using an exec based plugin.
FlexPersistentVolumeSource represents a generic persistent volume resource that is provisioned/attached using an exec based plugin.
flexVolume.driver (string), required
driver is the name of the driver to use for this volume.
flexVolume.fsType (string)
fsType is the Filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". The default filesystem depends on FlexVolume script.
flexVolume.options (map[string]string)
options is Optional: this field holds extra command options if any.
flexVolume.readOnly (boolean)
readOnly is Optional: defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
flexVolume.secretRef (SecretReference)
secretRef is Optional: SecretRef is reference to the secret object containing sensitive information to pass to the plugin scripts. This may be empty if no secret object is specified. If the secret object contains more than one secret, all secrets are passed to the plugin scripts.
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
flexVolume.secretRef.name (string)
name is unique within a namespace to reference a secret resource.
flexVolume.secretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
flocker (FlockerVolumeSource)
flocker represents a Flocker volume attached to a kubelet's host machine and exposed to the pod for its usage. This depends on the Flocker control service being running
Represents a Flocker volume mounted by the Flocker agent. One and only one of datasetName and datasetUUID should be set. Flocker volumes do not support ownership management or SELinux relabeling.
flocker.datasetName (string)
datasetName is Name of the dataset stored as metadata -> name on the dataset for Flocker should be considered as deprecated
flocker.datasetUUID (string)
datasetUUID is the UUID of the dataset. This is unique identifier of a Flocker dataset
*Represents a Persistent Disk resource in Google Compute Engine.
A GCE PD must exist before mounting to a container. The disk must also be in the same GCE project and zone as the kubelet. A GCE PD can only be mounted as read/write once or read-only many times. GCE PDs support ownership management and SELinux relabeling.*
fsType is filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#gcepersistentdisk
gcePersistentDisk.partition (int32)
partition is the partition in the volume that you want to mount. If omitted, the default is to mount by volume name. Examples: For volume /dev/sda1, you specify the partition as "1". Similarly, the volume partition for /dev/sda is "0" (or you can leave the property empty). More info: https://kubernetes.io/docs/concepts/storage/volumes#gcepersistentdisk
iscsi represents an ISCSI Disk resource that is attached to a kubelet's host machine and then exposed to the pod. Provisioned by an admin.
ISCSIPersistentVolumeSource represents an ISCSI disk. ISCSI volumes can only be mounted as read/write once. ISCSI volumes support ownership management and SELinux relabeling.
iscsi.iqn (string), required
iqn is Target iSCSI Qualified Name.
iscsi.lun (int32), required
lun is iSCSI Target Lun number.
iscsi.targetPortal (string), required
targetPortal is iSCSI Target Portal. The Portal is either an IP or ip_addr:port if the port is other than default (typically TCP ports 860 and 3260).
iscsi.chapAuthDiscovery (boolean)
chapAuthDiscovery defines whether support iSCSI Discovery CHAP authentication
iscsi.chapAuthSession (boolean)
chapAuthSession defines whether support iSCSI Session CHAP authentication
iscsi.fsType (string)
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#iscsi
iscsi.initiatorName (string)
initiatorName is the custom iSCSI Initiator Name. If initiatorName is specified with iscsiInterface simultaneously, new iSCSI interface <target portal>:<volume name> will be created for the connection.
iscsi.iscsiInterface (string)
iscsiInterface is the interface Name that uses an iSCSI transport. Defaults to 'default' (tcp).
iscsi.portals ([]string)
portals is the iSCSI Target Portal List. The Portal is either an IP or ip_addr:port if the port is other than default (typically TCP ports 860 and 3260).
iscsi.readOnly (boolean)
readOnly here will force the ReadOnly setting in VolumeMounts. Defaults to false.
iscsi.secretRef (SecretReference)
secretRef is the CHAP Secret for iSCSI target and initiator authentication
SecretReference represents a Secret Reference. It has enough information to retrieve secret in any namespace
iscsi.secretRef.name (string)
name is unique within a namespace to reference a secret resource.
iscsi.secretRef.namespace (string)
namespace defines the space within which the secret name must be unique.
photonPersistentDisk represents a PhotonController persistent disk attached and mounted on kubelets host machine
Represents a Photon Controller persistent disk resource.
photonPersistentDisk.pdID (string), required
pdID is the ID that identifies Photon Controller persistent disk
photonPersistentDisk.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
portworxVolume (PortworxVolumeSource)
portworxVolume represents a portworx volume attached and mounted on kubelets host machine
PortworxVolumeSource represents a Portworx volume resource.
portworxVolume.volumeID (string), required
volumeID uniquely identifies a Portworx volume
portworxVolume.fsType (string)
fSType represents the filesystem type to mount Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs". Implicitly inferred to be "ext4" if unspecified.
portworxVolume.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
quobyte (QuobyteVolumeSource)
quobyte represents a Quobyte mount on the host that shares a pod's lifetime
Represents a Quobyte mount that lasts the lifetime of a pod. Quobyte volumes do not support ownership management or SELinux relabeling.
quobyte.registry (string), required
registry represents a single or multiple Quobyte Registry services specified as a string as host:port pair (multiple entries are separated with commas) which acts as the central registry for volumes
quobyte.volume (string), required
volume is a string that references an already created Quobyte volume by name.
quobyte.group (string)
group to map volume access to Default is no group
quobyte.readOnly (boolean)
readOnly here will force the Quobyte volume to be mounted with read-only permissions. Defaults to false.
quobyte.tenant (string)
tenant owning the given Quobyte volume in the Backend Used with dynamically provisioned Quobyte volumes, value is set by the plugin
quobyte.user (string)
user to map volume access to Defaults to serivceaccount user
fsType is the filesystem type of the volume that you want to mount. Tip: Ensure that the filesystem type is supported by the host operating system. Examples: "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified. More info: https://kubernetes.io/docs/concepts/storage/volumes#rbd
Represents a StorageOS persistent volume resource.
storageos.fsType (string)
fsType is the filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
storageos.readOnly (boolean)
readOnly defaults to false (read/write). ReadOnly here will force the ReadOnly setting in VolumeMounts.
secretRef specifies the secret to use for obtaining the StorageOS API credentials. If not specified, default values will be attempted.
storageos.volumeName (string)
volumeName is the human-readable name of the StorageOS volume. Volume names are only unique within a namespace.
storageos.volumeNamespace (string)
volumeNamespace specifies the scope of the volume within StorageOS. If no namespace is specified then the Pod's namespace will be used. This allows the Kubernetes name scoping to be mirrored within StorageOS for tighter integration. Set VolumeName to any name to override the default behaviour. Set to "default" if you are not using namespaces within StorageOS. Namespaces that do not pre-exist within StorageOS will be created.
vsphereVolume (VsphereVirtualDiskVolumeSource)
vsphereVolume represents a vSphere volume attached and mounted on kubelets host machine
Represents a vSphere volume resource.
vsphereVolume.volumePath (string), required
volumePath is the path that identifies vSphere volume vmdk
vsphereVolume.fsType (string)
fsType is filesystem type to mount. Must be a filesystem type supported by the host operating system. Ex. "ext4", "xfs", "ntfs". Implicitly inferred to be "ext4" if unspecified.
vsphereVolume.storagePolicyID (string)
storagePolicyID is the storage Policy Based Management (SPBM) profile ID associated with the StoragePolicyName.
vsphereVolume.storagePolicyName (string)
storagePolicyName is the storage Policy Based Management (SPBM) profile name.
PersistentVolumeStatus
PersistentVolumeStatus is the current status of a persistent volume.
message (string)
message is a human-readable message indicating details about why the volume is in this state.
Provisioner indicates the type of the provisioner.
allowVolumeExpansion (boolean)
AllowVolumeExpansion shows whether the storage class allow volume expand
allowedTopologies ([]TopologySelectorTerm)
Atomic: will be replaced during a merge
Restrict the node topologies where volumes can be dynamically provisioned. Each volume plugin defines its own supported topology specifications. An empty TopologySelectorTerm list means there is no topology restriction. This field is only honored by servers that enable the VolumeScheduling feature.
A topology selector term represents the result of label queries. A null or empty topology selector term matches no objects. The requirements of them are ANDed. It provides a subset of functionality as NodeSelectorTerm. This is an alpha feature and may change in the future.
An array of string values. One value must match the label to be selected. Each entry in Values is ORed.
mountOptions ([]string)
Dynamically provisioned PersistentVolumes of this storage class are created with these mountOptions, e.g. ["ro", "soft"]. Not validated - mount of the PVs will simply fail if one is invalid.
parameters (map[string]string)
Parameters holds the parameters for the provisioner that should create volumes of this storage class.
reclaimPolicy (string)
Dynamically provisioned PersistentVolumes of this storage class are created with this reclaimPolicy. Defaults to Delete.
volumeBindingMode (string)
VolumeBindingMode indicates how PersistentVolumeClaims should be provisioned and bound. When unset, VolumeBindingImmediate is used. This field is only honored by servers that enable the VolumeScheduling feature.
StorageClassList
StorageClassList is a collection of storage classes.
Status of the VolumeAttachment request. Populated by the entity completing the attach or detach operation, i.e. the external-attacher.
VolumeAttachmentSpec
VolumeAttachmentSpec is the specification of a VolumeAttachment request.
attacher (string), required
Attacher indicates the name of the volume driver that MUST handle this request. This is the name returned by GetPluginName().
nodeName (string), required
The node that the volume should be attached to.
source (VolumeAttachmentSource), required
Source represents the volume that should be attached.
VolumeAttachmentSource represents a volume that should be attached. Right now only PersistenVolumes can be attached via external attacher, in future we may allow also inline volumes in pods. Exactly one member can be set.
inlineVolumeSpec contains all the information necessary to attach a persistent volume defined by a pod's inline VolumeSource. This field is populated only for the CSIMigration feature. It contains translated fields from a pod's inline VolumeSource to a PersistentVolumeSpec. This field is beta-level and is only honored by servers that enabled the CSIMigration feature.
source.persistentVolumeName (string)
Name of the persistent volume to attach.
VolumeAttachmentStatus
VolumeAttachmentStatus is the status of a VolumeAttachment request.
attached (boolean), required
Indicates the volume is successfully attached. This field must only be set by the entity completing the attach operation, i.e. the external-attacher.
attachError (VolumeError)
The last error encountered during attach operation, if any. This field must only be set by the entity completing the attach operation, i.e. the external-attacher.
VolumeError captures an error encountered during a volume operation.
attachError.message (string)
String detailing the error encountered during Attach or Detach operation. This string may be logged, so it should not contain sensitive information.
attachError.time (Time)
Time the error was encountered.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
attachmentMetadata (map[string]string)
Upon successful attach, this field is populated with any information returned by the attach operation that must be passed into subsequent WaitForAttach or Mount calls. This field must only be set by the entity completing the attach operation, i.e. the external-attacher.
detachError (VolumeError)
The last error encountered during detach operation, if any. This field must only be set by the entity completing the detach operation, i.e. the external-attacher.
VolumeError captures an error encountered during a volume operation.
detachError.message (string)
String detailing the error encountered during Attach or Detach operation. This string may be logged, so it should not contain sensitive information.
detachError.time (Time)
Time the error was encountered.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
VolumeAttachmentList
VolumeAttachmentList is a collection of VolumeAttachment objects.
CSIDriver captures information about a Container Storage Interface (CSI) volume driver deployed on the cluster.
apiVersion: storage.k8s.io/v1
import "k8s.io/api/storage/v1"
CSIDriver
CSIDriver captures information about a Container Storage Interface (CSI) volume driver deployed on the cluster. Kubernetes attach detach controller uses this object to determine whether attach is required. Kubelet uses this object to determine whether pod information needs to be passed on mount. CSIDriver objects are non-namespaced.
Standard object metadata. metadata.Name indicates the name of the CSI driver that this object refers to; it MUST be the same name returned by the CSI GetPluginName() call for that driver. The driver name must be 63 characters or less, beginning and ending with an alphanumeric character ([a-z0-9A-Z]) with dashes (-), dots (.), and alphanumerics between. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#metadata
CSIDriverSpec is the specification of a CSIDriver.
attachRequired (boolean)
attachRequired indicates this CSI volume driver requires an attach operation (because it implements the CSI ControllerPublishVolume() method), and that the Kubernetes attach detach controller should call the attach volume interface which checks the volumeattachment status and waits until the volume is attached before proceeding to mounting. The CSI external-attacher coordinates with CSI volume driver and updates the volumeattachment status when the attach operation is complete. If the CSIDriverRegistry feature gate is enabled and the value is specified to false, the attach operation will be skipped. Otherwise the attach operation will be called.
This field is immutable.
fsGroupPolicy (string)
Defines if the underlying volume supports changing ownership and permission of the volume before being mounted. Refer to the specific FSGroupPolicy values for additional details.
This field is immutable.
Defaults to ReadWriteOnceWithFSType, which will examine each volume to determine if Kubernetes should modify ownership and permissions of the volume. With the default policy the defined fsGroup will only be applied if a fstype is defined and the volume's access mode contains ReadWriteOnce.
podInfoOnMount (boolean)
If set to true, podInfoOnMount indicates this CSI volume driver requires additional pod information (like podName, podUID, etc.) during mount operations. If set to false, pod information will not be passed on mount. Default is false. The CSI driver specifies podInfoOnMount as part of driver deployment. If true, Kubelet will pass pod information as VolumeContext in the CSI NodePublishVolume() calls. The CSI driver is responsible for parsing and validating the information passed in as VolumeContext. The following VolumeConext will be passed if podInfoOnMount is set to true. This list might grow, but the prefix will be used. "csi.storage.k8s.io/pod.name": pod.Name "csi.storage.k8s.io/pod.namespace": pod.Namespace "csi.storage.k8s.io/pod.uid": string(pod.UID) "csi.storage.k8s.io/ephemeral": "true" if the volume is an ephemeral inline volume
defined by a CSIVolumeSource, otherwise "false"
"csi.storage.k8s.io/ephemeral" is a new feature in Kubernetes 1.16. It is only required for drivers which support both the "Persistent" and "Ephemeral" VolumeLifecycleMode. Other drivers can leave pod info disabled and/or ignore this field. As Kubernetes 1.15 doesn't support this field, drivers can only support one mode when deployed on such a cluster and the deployment determines which mode that is, for example via a command line parameter of the driver.
This field is immutable.
requiresRepublish (boolean)
RequiresRepublish indicates the CSI driver wants NodePublishVolume being periodically called to reflect any possible change in the mounted volume. This field defaults to false.
Note: After a successful initial NodePublishVolume call, subsequent calls to NodePublishVolume should only update the contents of the volume. New mount points will not be seen by a running container.
seLinuxMount (boolean)
SELinuxMount specifies if the CSI driver supports "-o context" mount option.
When "true", the CSI driver must ensure that all volumes provided by this CSI driver can be mounted separately with different -o context options. This is typical for storage backends that provide volumes as filesystems on block devices or as independent shared volumes. Kubernetes will call NodeStage / NodePublish with "-o context=xyz" mount option when mounting a ReadWriteOncePod volume used in Pod that has explicitly set SELinux context. In the future, it may be expanded to other volume AccessModes. In any case, Kubernetes will ensure that the volume is mounted only with a single SELinux context.
When "false", Kubernetes won't pass any special SELinux mount options to the driver. This is typical for volumes that represent subdirectories of a bigger shared filesystem.
Default is "false".
storageCapacity (boolean)
If set to true, storageCapacity indicates that the CSI volume driver wants pod scheduling to consider the storage capacity that the driver deployment will report by creating CSIStorageCapacity objects with capacity information.
The check can be enabled immediately when deploying a driver. In that case, provisioning new volumes with late binding will pause until the driver deployment has published some suitable CSIStorageCapacity object.
Alternatively, the driver can be deployed with the field unset or false and it can be flipped later when storage capacity information has been published.
This field was immutable in Kubernetes <= 1.22 and now is mutable.
tokenRequests ([]TokenRequest)
Atomic: will be replaced during a merge
TokenRequests indicates the CSI driver needs pods' service account tokens it is mounting volume for to do necessary authentication. Kubelet will pass the tokens in VolumeContext in the CSI NodePublishVolume calls. The CSI driver should parse and validate the following VolumeContext: "csi.storage.k8s.io/serviceAccount.tokens": {
"<audience>": {
"token": <token>,
"expirationTimestamp": <expiration timestamp in RFC3339>,
},
...
}
Note: Audience in each TokenRequest should be different and at most one token is empty string. To receive a new token after expiry, RequiresRepublish can be used to trigger NodePublishVolume periodically.
TokenRequest contains parameters of a service account token.
tokenRequests.audience (string), required
Audience is the intended audience of the token in "TokenRequestSpec". It will default to the audiences of kube apiserver.
tokenRequests.expirationSeconds (int64)
ExpirationSeconds is the duration of validity of the token in "TokenRequestSpec". It has the same default value of "ExpirationSeconds" in "TokenRequestSpec".
volumeLifecycleModes ([]string)
Set: unique values will be kept during a merge
volumeLifecycleModes defines what kind of volumes this CSI volume driver supports. The default if the list is empty is "Persistent", which is the usage defined by the CSI specification and implemented in Kubernetes via the usual PV/PVC mechanism. The other mode is "Ephemeral". In this mode, volumes are defined inline inside the pod spec with CSIVolumeSource and their lifecycle is tied to the lifecycle of that pod. A driver has to be aware of this because it is only going to get a NodePublishVolume call for such a volume. For more information about implementing this mode, see https://kubernetes-csi.github.io/docs/ephemeral-local-volumes.html A driver can support one or more of these modes and more modes may be added in the future. This field is beta.
This field is immutable.
CSIDriverList
CSIDriverList is a collection of CSIDriver objects.
CSINode holds information about all CSI drivers installed on a node.
apiVersion: storage.k8s.io/v1
import "k8s.io/api/storage/v1"
CSINode
CSINode holds information about all CSI drivers installed on a node. CSI drivers do not need to create the CSINode object directly. As long as they use the node-driver-registrar sidecar container, the kubelet will automatically populate the CSINode object for the CSI driver as part of kubelet plugin registration. CSINode has the same name as a node. If the object is missing, it means either there are no CSI Drivers available on the node, or the Kubelet version is low enough that it doesn't create this object. CSINode has an OwnerReference that points to the corresponding node object.
CSINodeSpec holds information about the specification of all CSI drivers installed on a node
drivers ([]CSINodeDriver), required
Patch strategy: merge on key name
drivers is a list of information of all CSI Drivers existing on a node. If all drivers in the list are uninstalled, this can become empty.
CSINodeDriver holds information about the specification of one CSI driver installed on a node
drivers.name (string), required
This is the name of the CSI driver that this object refers to. This MUST be the same name returned by the CSI GetPluginName() call for that driver.
drivers.nodeID (string), required
nodeID of the node from the driver point of view. This field enables Kubernetes to communicate with storage systems that do not share the same nomenclature for nodes. For example, Kubernetes may refer to a given node as "node1", but the storage system may refer to the same node as "nodeA". When Kubernetes issues a command to the storage system to attach a volume to a specific node, it can use this field to refer to the node name using the ID that the storage system will understand, e.g. "nodeA" instead of "node1". This field is required.
drivers.allocatable (VolumeNodeResources)
allocatable represents the volume resources of a node that are available for scheduling. This field is beta.
VolumeNodeResources is a set of resource limits for scheduling of volumes.
drivers.allocatable.count (int32)
Maximum number of unique volumes managed by the CSI driver that can be used on a node. A volume that is both attached and mounted on a node is considered to be used once, not twice. The same rule applies for a unique volume that is shared among multiple pods on the same node. If this field is not specified, then the supported number of volumes on this node is unbounded.
drivers.topologyKeys ([]string)
topologyKeys is the list of keys supported by the driver. When a driver is initialized on a cluster, it provides a set of topology keys that it understands (e.g. "company.com/zone", "company.com/region"). When a driver is initialized on a node, it provides the same topology keys along with values. Kubelet will expose these topology keys as labels on its own node object. When Kubernetes does topology aware provisioning, it can use this list to determine which labels it should retrieve from the node object and pass back to the driver. It is possible for different nodes to use different topology keys. This can be empty if driver does not support topology.
CSIStorageCapacity stores the result of one CSI GetCapacity call.
apiVersion: storage.k8s.io/v1
import "k8s.io/api/storage/v1"
CSIStorageCapacity
CSIStorageCapacity stores the result of one CSI GetCapacity call. For a given StorageClass, this describes the available capacity in a particular topology segment. This can be used when considering where to instantiate new PersistentVolumes.
For example this can express things like: - StorageClass "standard" has "1234 GiB" available in "topology.kubernetes.io/zone=us-east1" - StorageClass "localssd" has "10 GiB" available in "kubernetes.io/hostname=knode-abc123"
The following three cases all imply that no capacity is available for a certain combination: - no object exists with suitable topology and storage class name - such an object exists, but the capacity is unset - such an object exists, but the capacity is zero
The producer of these objects can decide which approach is more suitable.
They are consumed by the kube-scheduler when a CSI driver opts into capacity-aware scheduling with CSIDriverSpec.StorageCapacity. The scheduler compares the MaximumVolumeSize against the requested size of pending volumes to filter out unsuitable nodes. If MaximumVolumeSize is unset, it falls back to a comparison against the less precise Capacity. If that is also unset, the scheduler assumes that capacity is insufficient and tries some other node.
Standard object's metadata. The name has no particular meaning. It must be be a DNS subdomain (dots allowed, 253 characters). To ensure that there are no conflicts with other CSI drivers on the cluster, the recommendation is to use csisc-<uuid>, a generated name, or a reverse-domain name which ends with the unique CSI driver name.
The name of the StorageClass that the reported capacity applies to. It must meet the same requirements as the name of a StorageClass object (non-empty, DNS subdomain). If that object no longer exists, the CSIStorageCapacity object is obsolete and should be removed by its creator. This field is immutable.
Capacity is the value reported by the CSI driver in its GetCapacityResponse for a GetCapacityRequest with topology and parameters that match the previous fields.
The semantic is currently (CSI spec 1.2) defined as: The available capacity, in bytes, of the storage that can be used to provision volumes. If not set, that information is currently unavailable.
MaximumVolumeSize is the value reported by the CSI driver in its GetCapacityResponse for a GetCapacityRequest with topology and parameters that match the previous fields.
This is defined since CSI spec 1.4.0 as the largest size that may be used in a CreateVolumeRequest.capacity_range.required_bytes field to create a volume with the same parameters as those in GetCapacityRequest. The corresponding value in the Kubernetes API is ResourceRequirements.Requests in a volume claim.
NodeTopology defines which nodes have access to the storage for which capacity was reported. If not set, the storage is not accessible from any node in the cluster. If empty, the storage is accessible from all nodes. This field is immutable.
CSIStorageCapacityList
CSIStorageCapacityList is a collection of CSIStorageCapacity objects.
ServiceAccount binds together: * a name, understood by users, and perhaps by peripheral systems, for an identity * a principal that can be authenticated and authorized * a set of secrets.
apiVersion: v1
import "k8s.io/api/core/v1"
ServiceAccount
ServiceAccount binds together: * a name, understood by users, and perhaps by peripheral systems, for an identity * a principal that can be authenticated and authorized * a set of secrets
AutomountServiceAccountToken indicates whether pods running as this service account should have an API token automatically mounted. Can be overridden at the pod level.
ImagePullSecrets is a list of references to secrets in the same namespace to use for pulling any images in pods that reference this ServiceAccount. ImagePullSecrets are distinct from Secrets because Secrets can be mounted in the pod, but ImagePullSecrets are only accessed by the kubelet. More info: https://kubernetes.io/docs/concepts/containers/images/#specifying-imagepullsecrets-on-a-pod
Secrets is a list of the secrets in the same namespace that pods running using this ServiceAccount are allowed to use. Pods are only limited to this list if this service account has a "kubernetes.io/enforce-mountable-secrets" annotation set to "true". This field should not be used to find auto-generated service account token secrets for use outside of pods. Instead, tokens can be requested directly using the TokenRequest API, or service account token secrets can be manually created. More info: https://kubernetes.io/docs/concepts/configuration/secret
ServiceAccountList
ServiceAccountList is a list of ServiceAccount objects
Status is filled in by the server and indicates whether the token can be authenticated.
TokenRequestSpec
TokenRequestSpec contains client provided parameters of a token request.
audiences ([]string), required
Audiences are the intendend audiences of the token. A recipient of a token must identify themself with an identifier in the list of audiences of the token, and otherwise should reject the token. A token issued for multiple audiences may be used to authenticate against any of the audiences listed but implies a high degree of trust between the target audiences.
boundObjectRef (BoundObjectReference)
BoundObjectRef is a reference to an object that the token will be bound to. The token will only be valid for as long as the bound object exists. NOTE: The API server's TokenReview endpoint will validate the BoundObjectRef, but other audiences may not. Keep ExpirationSeconds small if you want prompt revocation.
BoundObjectReference is a reference to an object that a token is bound to.
boundObjectRef.apiVersion (string)
API version of the referent.
boundObjectRef.kind (string)
Kind of the referent. Valid kinds are 'Pod' and 'Secret'.
boundObjectRef.name (string)
Name of the referent.
boundObjectRef.uid (string)
UID of the referent.
expirationSeconds (int64)
ExpirationSeconds is the requested duration of validity of the request. The token issuer may return a token with a different validity duration so a client needs to check the 'expiration' field in a response.
TokenRequestStatus
TokenRequestStatus is the result of a token request.
expirationTimestamp (Time), required
ExpirationTimestamp is the time of expiration of the returned token.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
token (string), required
Token is the opaque bearer token.
Operations
create create token of a ServiceAccount
HTTP Request
POST /api/v1/namespaces/{namespace}/serviceaccounts/{name}/token
TokenReview attempts to authenticate a token to a known user.
apiVersion: authentication.k8s.io/v1
import "k8s.io/api/authentication/v1"
TokenReview
TokenReview attempts to authenticate a token to a known user. Note: TokenReview requests may be cached by the webhook token authenticator plugin in the kube-apiserver.
Status is filled in by the server and indicates whether the request can be authenticated.
TokenReviewSpec
TokenReviewSpec is a description of the token authentication request.
audiences ([]string)
Audiences is a list of the identifiers that the resource server presented with the token identifies as. Audience-aware token authenticators will verify that the token was intended for at least one of the audiences in this list. If no audiences are provided, the audience will default to the audience of the Kubernetes apiserver.
token (string)
Token is the opaque bearer token.
TokenReviewStatus
TokenReviewStatus is the result of the token authentication request.
audiences ([]string)
Audiences are audience identifiers chosen by the authenticator that are compatible with both the TokenReview and token. An identifier is any identifier in the intersection of the TokenReviewSpec audiences and the token's audiences. A client of the TokenReview API that sets the spec.audiences field should validate that a compatible audience identifier is returned in the status.audiences field to ensure that the TokenReview server is audience aware. If a TokenReview returns an empty status.audience field where status.authenticated is "true", the token is valid against the audience of the Kubernetes API server.
authenticated (boolean)
Authenticated indicates that the token was associated with a known user.
error (string)
Error indicates that the token couldn't be checked
user (UserInfo)
User is the UserInfo associated with the provided token.
UserInfo holds the information about the user needed to implement the user.Info interface.
user.extra (map[string][]string)
Any additional information provided by the authenticator.
user.groups ([]string)
The names of groups this user is a part of.
user.uid (string)
A unique value that identifies this user across time. If this user is deleted and another user by the same name is added, they will have different UIDs.
user.username (string)
The name that uniquely identifies this user among all active users.
CertificateSigningRequest objects provide a mechanism to obtain x509 certificates by submitting a certificate signing request, and having it asynchronously approved and issued.
apiVersion: certificates.k8s.io/v1
import "k8s.io/api/certificates/v1"
CertificateSigningRequest
CertificateSigningRequest objects provide a mechanism to obtain x509 certificates by submitting a certificate signing request, and having it asynchronously approved and issued.
Kubelets use this API to obtain:
client certificates to authenticate to kube-apiserver (with the "kubernetes.io/kube-apiserver-client-kubelet" signerName).
serving certificates for TLS endpoints kube-apiserver can connect to securely (with the "kubernetes.io/kubelet-serving" signerName).
This API can be used to request client certificates to authenticate to kube-apiserver (with the "kubernetes.io/kube-apiserver-client" signerName), or to obtain certificates from custom non-Kubernetes signers.
spec contains the certificate request, and is immutable after creation. Only the request, signerName, expirationSeconds, and usages fields can be set on creation. Other fields are derived by Kubernetes and cannot be modified by users.
status contains information about whether the request is approved or denied, and the certificate issued by the signer, or the failure condition indicating signer failure.
CertificateSigningRequestSpec
CertificateSigningRequestSpec contains the certificate request.
request ([]byte), required
Atomic: will be replaced during a merge
request contains an x509 certificate signing request encoded in a "CERTIFICATE REQUEST" PEM block. When serialized as JSON or YAML, the data is additionally base64-encoded.
signerName (string), required
signerName indicates the requested signer, and is a qualified name.
List/watch requests for CertificateSigningRequests can filter on this field using a "spec.signerName=NAME" fieldSelector.
Well-known Kubernetes signers are:
"kubernetes.io/kube-apiserver-client": issues client certificates that can be used to authenticate to kube-apiserver.
Requests for this signer are never auto-approved by kube-controller-manager, can be issued by the "csrsigning" controller in kube-controller-manager.
"kubernetes.io/kube-apiserver-client-kubelet": issues client certificates that kubelets use to authenticate to kube-apiserver.
Requests for this signer can be auto-approved by the "csrapproving" controller in kube-controller-manager, and can be issued by the "csrsigning" controller in kube-controller-manager.
"kubernetes.io/kubelet-serving" issues serving certificates that kubelets use to serve TLS endpoints, which kube-apiserver can connect to securely.
Requests for this signer are never auto-approved by kube-controller-manager, and can be issued by the "csrsigning" controller in kube-controller-manager.
Custom signerNames can also be specified. The signer defines:
Trust distribution: how trust (CA bundles) are distributed.
Permitted subjects: and behavior when a disallowed subject is requested.
Required, permitted, or forbidden x509 extensions in the request (including whether subjectAltNames are allowed, which types, restrictions on allowed values) and behavior when a disallowed extension is requested.
Required, permitted, or forbidden key usages / extended key usages.
Expiration/certificate lifetime: whether it is fixed by the signer, configurable by the admin.
Whether or not requests for CA certificates are allowed.
expirationSeconds (int32)
expirationSeconds is the requested duration of validity of the issued certificate. The certificate signer may issue a certificate with a different validity duration so a client must check the delta between the notBefore and and notAfter fields in the issued certificate to determine the actual duration.
The v1.22+ in-tree implementations of the well-known Kubernetes signers will honor this field as long as the requested duration is not greater than the maximum duration they will honor per the --cluster-signing-duration CLI flag to the Kubernetes controller manager.
Certificate signers may not honor this field for various reasons:
Old signer that is unaware of the field (such as the in-tree
implementations prior to v1.22)
Signer whose configured maximum is shorter than the requested duration
Signer whose configured minimum is longer than the requested duration
The minimum valid value for expirationSeconds is 600, i.e. 10 minutes.
extra (map[string][]string)
extra contains extra attributes of the user that created the CertificateSigningRequest. Populated by the API server on creation and immutable.
groups ([]string)
Atomic: will be replaced during a merge
groups contains group membership of the user that created the CertificateSigningRequest. Populated by the API server on creation and immutable.
uid (string)
uid contains the uid of the user that created the CertificateSigningRequest. Populated by the API server on creation and immutable.
usages ([]string)
Atomic: will be replaced during a merge
usages specifies a set of key usages requested in the issued certificate.
Requests for TLS client certificates typically request: "digital signature", "key encipherment", "client auth".
Requests for TLS serving certificates typically request: "key encipherment", "digital signature", "server auth".
username contains the name of the user that created the CertificateSigningRequest. Populated by the API server on creation and immutable.
CertificateSigningRequestStatus
CertificateSigningRequestStatus contains conditions used to indicate approved/denied/failed status of the request, and the issued certificate.
certificate ([]byte)
Atomic: will be replaced during a merge
certificate is populated with an issued certificate by the signer after an Approved condition is present. This field is set via the /status subresource. Once populated, this field is immutable.
If the certificate signing request is denied, a condition of type "Denied" is added and this field remains empty. If the signer cannot issue the certificate, a condition of type "Failed" is added and this field remains empty.
Validation requirements:
certificate must contain one or more PEM blocks.
All PEM blocks must have the "CERTIFICATE" label, contain no headers, and the encoded data
must be a BER-encoded ASN.1 Certificate structure as described in section 4 of RFC5280.
Non-PEM content may appear before or after the "CERTIFICATE" PEM blocks and is unvalidated,
to allow for explanatory text as described in section 5.2 of RFC7468.
If more than one PEM block is present, and the definition of the requested spec.signerName does not indicate otherwise, the first block is the issued certificate, and subsequent blocks should be treated as intermediate certificates and presented in TLS handshakes.
The certificate is encoded in PEM format.
When serialized as JSON or YAML, the data is additionally base64-encoded, so it consists of:
Map: unique values on key type will be kept during a merge
conditions applied to the request. Known conditions are "Approved", "Denied", and "Failed".
CertificateSigningRequestCondition describes a condition of a CertificateSigningRequest object
conditions.status (string), required
status of the condition, one of True, False, Unknown. Approved, Denied, and Failed conditions may not be "False" or "Unknown".
conditions.type (string), required
type of the condition. Known conditions are "Approved", "Denied", and "Failed".
An "Approved" condition is added via the /approval subresource, indicating the request was approved and should be issued by the signer.
A "Denied" condition is added via the /approval subresource, indicating the request was denied and should not be issued by the signer.
A "Failed" condition is added via the /status subresource, indicating the signer failed to issue the certificate.
Approved and Denied conditions are mutually exclusive. Approved, Denied, and Failed conditions cannot be removed once added.
Only one condition of a given type is allowed.
conditions.lastTransitionTime (Time)
lastTransitionTime is the time the condition last transitioned from one status to another. If unset, when a new condition type is added or an existing condition's status is changed, the server defaults this to the current time.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastUpdateTime (Time)
lastUpdateTime is the time of the last update to this condition
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message contains a human readable message with details about the request state
conditions.reason (string)
reason indicates a brief reason for the request state
CertificateSigningRequestList
CertificateSigningRequestList is a collection of CertificateSigningRequest objects
LocalSubjectAccessReview checks whether or not a user or group can perform an action in a given namespace.
apiVersion: authorization.k8s.io/v1
import "k8s.io/api/authorization/v1"
LocalSubjectAccessReview
LocalSubjectAccessReview checks whether or not a user or group can perform an action in a given namespace. Having a namespace scoped resource makes it much easier to grant namespace scoped policy that includes permissions checking.
Spec holds information about the request being evaluated. spec.namespace must be equal to the namespace you made the request against. If empty, it is defaulted.
SelfSubjectAccessReview checks whether or the current user can perform an action.
apiVersion: authorization.k8s.io/v1
import "k8s.io/api/authorization/v1"
SelfSubjectAccessReview
SelfSubjectAccessReview checks whether or the current user can perform an action. Not filling in a spec.namespace means "in all namespaces". Self is a special case, because users should always be able to check whether they can perform an action
Status is filled in by the server and indicates whether the request is allowed or not
SelfSubjectAccessReviewSpec
SelfSubjectAccessReviewSpec is a description of the access request. Exactly one of ResourceAuthorizationAttributes and NonResourceAuthorizationAttributes must be set
nonResourceAttributes (NonResourceAttributes)
NonResourceAttributes describes information for a non-resource access request
NonResourceAttributes includes the authorization attributes available for non-resource requests to the Authorizer interface
nonResourceAttributes.path (string)
Path is the URL path of the request
nonResourceAttributes.verb (string)
Verb is the standard HTTP verb
resourceAttributes (ResourceAttributes)
ResourceAuthorizationAttributes describes information for a resource access request
ResourceAttributes includes the authorization attributes available for resource requests to the Authorizer interface
resourceAttributes.group (string)
Group is the API Group of the Resource. "*" means all.
resourceAttributes.name (string)
Name is the name of the resource being requested for a "get" or deleted for a "delete". "" (empty) means all.
resourceAttributes.namespace (string)
Namespace is the namespace of the action being requested. Currently, there is no distinction between no namespace and all namespaces "" (empty) is defaulted for LocalSubjectAccessReviews "" (empty) is empty for cluster-scoped resources "" (empty) means "all" for namespace scoped resources from a SubjectAccessReview or SelfSubjectAccessReview
resourceAttributes.resource (string)
Resource is one of the existing resource types. "*" means all.
resourceAttributes.subresource (string)
Subresource is one of the existing resource types. "" means none.
resourceAttributes.verb (string)
Verb is a kubernetes resource API verb, like: get, list, watch, create, update, delete, proxy. "*" means all.
resourceAttributes.version (string)
Version is the API Version of the Resource. "*" means all.
Operations
create create a SelfSubjectAccessReview
HTTP Request
POST /apis/authorization.k8s.io/v1/selfsubjectaccessreviews
SelfSubjectRulesReview enumerates the set of actions the current user can perform within a namespace.
apiVersion: authorization.k8s.io/v1
import "k8s.io/api/authorization/v1"
SelfSubjectRulesReview
SelfSubjectRulesReview enumerates the set of actions the current user can perform within a namespace. The returned list of actions may be incomplete depending on the server's authorization mode, and any errors experienced during the evaluation. SelfSubjectRulesReview should be used by UIs to show/hide actions, or to quickly let an end user reason about their permissions. It should NOT Be used by external systems to drive authorization decisions as this raises confused deputy, cache lifetime/revocation, and correctness concerns. SubjectAccessReview, and LocalAccessReview are the correct way to defer authorization decisions to the API server.
Spec holds information about the request being evaluated.
status (SubjectRulesReviewStatus)
Status is filled in by the server and indicates the set of actions a user can perform.
SubjectRulesReviewStatus contains the result of a rules check. This check can be incomplete depending on the set of authorizers the server is configured with and any errors experienced during evaluation. Because authorization rules are additive, if a rule appears in a list it's safe to assume the subject has that permission, even if that list is incomplete.
status.incomplete (boolean), required
Incomplete is true when the rules returned by this call are incomplete. This is most commonly encountered when an authorizer, such as an external authorizer, doesn't support rules evaluation.
NonResourceRules is the list of actions the subject is allowed to perform on non-resources. The list ordering isn't significant, may contain duplicates, and possibly be incomplete.
NonResourceRule holds information that describes a rule for the non-resource
NonResourceURLs is a set of partial urls that a user should have access to. s are allowed, but only as the full, final step in the path. "" means all.
status.resourceRules ([]ResourceRule), required
ResourceRules is the list of actions the subject is allowed to perform on resources. The list ordering isn't significant, may contain duplicates, and possibly be incomplete.
ResourceRule is the list of actions the subject is allowed to perform on resources. The list ordering isn't significant, may contain duplicates, and possibly be incomplete.
status.resourceRules.verbs ([]string), required
Verb is a list of kubernetes resource API verbs, like: get, list, watch, create, update, delete, proxy. "*" means all.
status.resourceRules.apiGroups ([]string)
APIGroups is the name of the APIGroup that contains the resources. If multiple API groups are specified, any action requested against one of the enumerated resources in any API group will be allowed. "*" means all.
status.resourceRules.resourceNames ([]string)
ResourceNames is an optional white list of names that the rule applies to. An empty set means that everything is allowed. "*" means all.
status.resourceRules.resources ([]string)
Resources is a list of resources this rule applies to. "" means all in the specified apiGroups.
"/foo" represents the subresource 'foo' for all resources in the specified apiGroups.
status.evaluationError (string)
EvaluationError can appear in combination with Rules. It indicates an error occurred during rule evaluation, such as an authorizer that doesn't support rule evaluation, and that ResourceRules and/or NonResourceRules may be incomplete.
SelfSubjectRulesReviewSpec
SelfSubjectRulesReviewSpec defines the specification for SelfSubjectRulesReview.
namespace (string)
Namespace to evaluate rules for. Required.
Operations
create create a SelfSubjectRulesReview
HTTP Request
POST /apis/authorization.k8s.io/v1/selfsubjectrulesreviews
Status is filled in by the server and indicates whether the request is allowed or not
SubjectAccessReviewSpec
SubjectAccessReviewSpec is a description of the access request. Exactly one of ResourceAuthorizationAttributes and NonResourceAuthorizationAttributes must be set
extra (map[string][]string)
Extra corresponds to the user.Info.GetExtra() method from the authenticator. Since that is input to the authorizer it needs a reflection here.
groups ([]string)
Groups is the groups you're testing for.
nonResourceAttributes (NonResourceAttributes)
NonResourceAttributes describes information for a non-resource access request
NonResourceAttributes includes the authorization attributes available for non-resource requests to the Authorizer interface
nonResourceAttributes.path (string)
Path is the URL path of the request
nonResourceAttributes.verb (string)
Verb is the standard HTTP verb
resourceAttributes (ResourceAttributes)
ResourceAuthorizationAttributes describes information for a resource access request
ResourceAttributes includes the authorization attributes available for resource requests to the Authorizer interface
resourceAttributes.group (string)
Group is the API Group of the Resource. "*" means all.
resourceAttributes.name (string)
Name is the name of the resource being requested for a "get" or deleted for a "delete". "" (empty) means all.
resourceAttributes.namespace (string)
Namespace is the namespace of the action being requested. Currently, there is no distinction between no namespace and all namespaces "" (empty) is defaulted for LocalSubjectAccessReviews "" (empty) is empty for cluster-scoped resources "" (empty) means "all" for namespace scoped resources from a SubjectAccessReview or SelfSubjectAccessReview
resourceAttributes.resource (string)
Resource is one of the existing resource types. "*" means all.
resourceAttributes.subresource (string)
Subresource is one of the existing resource types. "" means none.
resourceAttributes.verb (string)
Verb is a kubernetes resource API verb, like: get, list, watch, create, update, delete, proxy. "*" means all.
resourceAttributes.version (string)
Version is the API Version of the Resource. "*" means all.
uid (string)
UID information about the requesting user.
user (string)
User is the user you're testing for. If you specify "User" but not "Groups", then is it interpreted as "What if User were not a member of any groups
SubjectAccessReviewStatus
SubjectAccessReviewStatus
allowed (boolean), required
Allowed is required. True if the action would be allowed, false otherwise.
denied (boolean)
Denied is optional. True if the action would be denied, otherwise false. If both allowed is false and denied is false, then the authorizer has no opinion on whether to authorize the action. Denied may not be true if Allowed is true.
evaluationError (string)
EvaluationError is an indication that some error occurred during the authorization check. It is entirely possible to get an error and be able to continue determine authorization status in spite of it. For instance, RBAC can be missing a role, but enough roles are still present and bound to reason about the request.
reason (string)
Reason is optional. It indicates why a request was allowed or denied.
Operations
create create a SubjectAccessReview
HTTP Request
POST /apis/authorization.k8s.io/v1/subjectaccessreviews
AggregationRule is an optional field that describes how to build the Rules for this ClusterRole. If AggregationRule is set, then the Rules are controller managed and direct changes to Rules will be stomped by the controller.
AggregationRule describes how to locate ClusterRoles to aggregate into the ClusterRole
ClusterRoleSelectors holds a list of selectors which will be used to find ClusterRoles and create the rules. If any of the selectors match, then the ClusterRole's permissions will be added
rules ([]PolicyRule)
Rules holds all the PolicyRules for this ClusterRole
PolicyRule holds information that describes a policy rule, but does not contain information about who the rule applies to or which namespace the rule applies to.
rules.apiGroups ([]string)
APIGroups is the name of the APIGroup that contains the resources. If multiple API groups are specified, any action requested against one of the enumerated resources in any API group will be allowed. "" represents the core API group and "*" represents all API groups.
rules.resources ([]string)
Resources is a list of resources this rule applies to. '*' represents all resources.
rules.verbs ([]string), required
Verbs is a list of Verbs that apply to ALL the ResourceKinds contained in this rule. '*' represents all verbs.
rules.resourceNames ([]string)
ResourceNames is an optional white list of names that the rule applies to. An empty set means that everything is allowed.
rules.nonResourceURLs ([]string)
NonResourceURLs is a set of partial urls that a user should have access to. *s are allowed, but only as the full, final step in the path Since non-resource URLs are not namespaced, this field is only applicable for ClusterRoles referenced from a ClusterRoleBinding. Rules can either apply to API resources (such as "pods" or "secrets") or non-resource URL paths (such as "/api"), but not both.
ClusterRoleBinding references a ClusterRole, but not contain it.
apiVersion: rbac.authorization.k8s.io/v1
import "k8s.io/api/rbac/v1"
ClusterRoleBinding
ClusterRoleBinding references a ClusterRole, but not contain it. It can reference a ClusterRole in the global namespace, and adds who information via Subject.
RoleRef can only reference a ClusterRole in the global namespace. If the RoleRef cannot be resolved, the Authorizer must return an error.
RoleRef contains information that points to the role being used
roleRef.apiGroup (string), required
APIGroup is the group for the resource being referenced
roleRef.kind (string), required
Kind is the type of resource being referenced
roleRef.name (string), required
Name is the name of resource being referenced
subjects ([]Subject)
Subjects holds references to the objects the role applies to.
Subject contains a reference to the object or user identities a role binding applies to. This can either hold a direct API object reference, or a value for non-objects such as user and group names.
subjects.kind (string), required
Kind of object being referenced. Values defined by this API group are "User", "Group", and "ServiceAccount". If the Authorizer does not recognized the kind value, the Authorizer should report an error.
subjects.name (string), required
Name of the object being referenced.
subjects.apiGroup (string)
APIGroup holds the API group of the referenced subject. Defaults to "" for ServiceAccount subjects. Defaults to "rbac.authorization.k8s.io" for User and Group subjects.
subjects.namespace (string)
Namespace of the referenced object. If the object kind is non-namespace, such as "User" or "Group", and this value is not empty the Authorizer should report an error.
ClusterRoleBindingList
ClusterRoleBindingList is a collection of ClusterRoleBindings
PolicyRule holds information that describes a policy rule, but does not contain information about who the rule applies to or which namespace the rule applies to.
rules.apiGroups ([]string)
APIGroups is the name of the APIGroup that contains the resources. If multiple API groups are specified, any action requested against one of the enumerated resources in any API group will be allowed. "" represents the core API group and "*" represents all API groups.
rules.resources ([]string)
Resources is a list of resources this rule applies to. '*' represents all resources.
rules.verbs ([]string), required
Verbs is a list of Verbs that apply to ALL the ResourceKinds contained in this rule. '*' represents all verbs.
rules.resourceNames ([]string)
ResourceNames is an optional white list of names that the rule applies to. An empty set means that everything is allowed.
rules.nonResourceURLs ([]string)
NonResourceURLs is a set of partial urls that a user should have access to. *s are allowed, but only as the full, final step in the path Since non-resource URLs are not namespaced, this field is only applicable for ClusterRoles referenced from a ClusterRoleBinding. Rules can either apply to API resources (such as "pods" or "secrets") or non-resource URL paths (such as "/api"), but not both.
RoleBinding references a role, but does not contain it.
apiVersion: rbac.authorization.k8s.io/v1
import "k8s.io/api/rbac/v1"
RoleBinding
RoleBinding references a role, but does not contain it. It can reference a Role in the same namespace or a ClusterRole in the global namespace. It adds who information via Subjects and namespace information by which namespace it exists in. RoleBindings in a given namespace only have effect in that namespace.
RoleRef can reference a Role in the current namespace or a ClusterRole in the global namespace. If the RoleRef cannot be resolved, the Authorizer must return an error.
RoleRef contains information that points to the role being used
roleRef.apiGroup (string), required
APIGroup is the group for the resource being referenced
roleRef.kind (string), required
Kind is the type of resource being referenced
roleRef.name (string), required
Name is the name of resource being referenced
subjects ([]Subject)
Subjects holds references to the objects the role applies to.
Subject contains a reference to the object or user identities a role binding applies to. This can either hold a direct API object reference, or a value for non-objects such as user and group names.
subjects.kind (string), required
Kind of object being referenced. Values defined by this API group are "User", "Group", and "ServiceAccount". If the Authorizer does not recognized the kind value, the Authorizer should report an error.
subjects.name (string), required
Name of the object being referenced.
subjects.apiGroup (string)
APIGroup holds the API group of the referenced subject. Defaults to "" for ServiceAccount subjects. Defaults to "rbac.authorization.k8s.io" for User and Group subjects.
subjects.namespace (string)
Namespace of the referenced object. If the object kind is non-namespace, such as "User" or "Group", and this value is not empty the Authorizer should report an error.
MaxLimitRequestRatio if specified, the named resource must have a request and limit that are both non-zero where limit divided by request is less than or equal to the enumerated value; this represents the max burst for the named resource.
scopeSelector is also a collection of filters like scopes that must match each object tracked by a quota but expressed using ScopeSelectorOperator in combination with possible values. For a resource to match, both scopes AND scopeSelector (if specified in spec), must be matched.
A scope selector represents the AND of the selectors represented by the scoped-resource selector requirements.
The name of the scope that the selector applies to.
scopeSelector.matchExpressions.values ([]string)
An array of string values. If the operator is In or NotIn, the values array must be non-empty. If the operator is Exists or DoesNotExist, the values array must be empty. This array is replaced during a strategic merge patch.
scopes ([]string)
A collection of filters that must match each object tracked by a quota. If not specified, the quota matches all objects.
ResourceQuotaStatus
ResourceQuotaStatus defines the enforced hard limits and observed use.
Selects the pods to which this NetworkPolicy object applies. The array of ingress rules is applied to any pods selected by this field. Multiple network policies can select the same set of pods. In this case, the ingress rules for each are combined additively. This field is NOT optional and follows standard label selector semantics. An empty podSelector matches all pods in this namespace.
policyTypes ([]string)
List of rule types that the NetworkPolicy relates to. Valid options are ["Ingress"], ["Egress"], or ["Ingress", "Egress"]. If this field is not specified, it will default based on the existence of Ingress or Egress rules; policies that contain an Egress section are assumed to affect Egress, and all policies (whether or not they contain an Ingress section) are assumed to affect Ingress. If you want to write an egress-only policy, you must explicitly specify policyTypes [ "Egress" ]. Likewise, if you want to write a policy that specifies that no egress is allowed, you must specify a policyTypes value that include "Egress" (since such a policy would not include an Egress section and would otherwise default to just [ "Ingress" ]). This field is beta-level in 1.8
ingress ([]NetworkPolicyIngressRule)
List of ingress rules to be applied to the selected pods. Traffic is allowed to a pod if there are no NetworkPolicies selecting the pod (and cluster policy otherwise allows the traffic), OR if the traffic source is the pod's local node, OR if the traffic matches at least one ingress rule across all of the NetworkPolicy objects whose podSelector matches the pod. If this field is empty then this NetworkPolicy does not allow any traffic (and serves solely to ensure that the pods it selects are isolated by default)
NetworkPolicyIngressRule describes a particular set of traffic that is allowed to the pods matched by a NetworkPolicySpec's podSelector. The traffic must match both ports and from.
ingress.from ([]NetworkPolicyPeer)
List of sources which should be able to access the pods selected for this rule. Items in this list are combined using a logical OR operation. If this field is empty or missing, this rule matches all sources (traffic not restricted by source). If this field is present and contains at least one item, this rule allows traffic only if the traffic matches at least one item in the from list.
NetworkPolicyPeer describes a peer to allow traffic to/from. Only certain combinations of fields are allowed
ingress.from.ipBlock (IPBlock)
IPBlock defines policy on a particular IPBlock. If this field is set then neither of the other fields can be.
IPBlock describes a particular CIDR (Ex. "192.168.1.1/24","2001:db9::/64") that is allowed to the pods matched by a NetworkPolicySpec's podSelector. The except entry describes CIDRs that should not be included within this rule.
ingress.from.ipBlock.cidr (string), required
CIDR is a string representing the IP Block Valid examples are "192.168.1.1/24" or "2001:db9::/64"
ingress.from.ipBlock.except ([]string)
Except is a slice of CIDRs that should not be included within an IP Block Valid examples are "192.168.1.1/24" or "2001:db9::/64" Except values will be rejected if they are outside the CIDR range
Selects Namespaces using cluster-scoped labels. This field follows standard label selector semantics; if present but empty, it selects all namespaces.
If PodSelector is also set, then the NetworkPolicyPeer as a whole selects the Pods matching PodSelector in the Namespaces selected by NamespaceSelector. Otherwise it selects all Pods in the Namespaces selected by NamespaceSelector.
This is a label selector which selects Pods. This field follows standard label selector semantics; if present but empty, it selects all pods.
If NamespaceSelector is also set, then the NetworkPolicyPeer as a whole selects the Pods matching PodSelector in the Namespaces selected by NamespaceSelector. Otherwise it selects the Pods matching PodSelector in the policy's own Namespace.
ingress.ports ([]NetworkPolicyPort)
List of ports which should be made accessible on the pods selected for this rule. Each item in this list is combined using a logical OR. If this field is empty or missing, this rule matches all ports (traffic not restricted by port). If this field is present and contains at least one item, then this rule allows traffic only if the traffic matches at least one port in the list.
NetworkPolicyPort describes a port to allow traffic on
ingress.ports.port (IntOrString)
The port on the given protocol. This can either be a numerical or named port on a pod. If this field is not provided, this matches all port names and numbers. If present, only traffic on the specified protocol AND port will be matched.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
ingress.ports.endPort (int32)
If set, indicates that the range of ports from port to endPort, inclusive, should be allowed by the policy. This field cannot be defined if the port field is not defined or if the port field is defined as a named (string) port. The endPort must be equal or greater than port.
ingress.ports.protocol (string)
The protocol (TCP, UDP, or SCTP) which traffic must match. If not specified, this field defaults to TCP.
egress ([]NetworkPolicyEgressRule)
List of egress rules to be applied to the selected pods. Outgoing traffic is allowed if there are no NetworkPolicies selecting the pod (and cluster policy otherwise allows the traffic), OR if the traffic matches at least one egress rule across all of the NetworkPolicy objects whose podSelector matches the pod. If this field is empty then this NetworkPolicy limits all outgoing traffic (and serves solely to ensure that the pods it selects are isolated by default). This field is beta-level in 1.8
NetworkPolicyEgressRule describes a particular set of traffic that is allowed out of pods matched by a NetworkPolicySpec's podSelector. The traffic must match both ports and to. This type is beta-level in 1.8
egress.to ([]NetworkPolicyPeer)
List of destinations for outgoing traffic of pods selected for this rule. Items in this list are combined using a logical OR operation. If this field is empty or missing, this rule matches all destinations (traffic not restricted by destination). If this field is present and contains at least one item, this rule allows traffic only if the traffic matches at least one item in the to list.
NetworkPolicyPeer describes a peer to allow traffic to/from. Only certain combinations of fields are allowed
egress.to.ipBlock (IPBlock)
IPBlock defines policy on a particular IPBlock. If this field is set then neither of the other fields can be.
IPBlock describes a particular CIDR (Ex. "192.168.1.1/24","2001:db9::/64") that is allowed to the pods matched by a NetworkPolicySpec's podSelector. The except entry describes CIDRs that should not be included within this rule.
egress.to.ipBlock.cidr (string), required
CIDR is a string representing the IP Block Valid examples are "192.168.1.1/24" or "2001:db9::/64"
egress.to.ipBlock.except ([]string)
Except is a slice of CIDRs that should not be included within an IP Block Valid examples are "192.168.1.1/24" or "2001:db9::/64" Except values will be rejected if they are outside the CIDR range
Selects Namespaces using cluster-scoped labels. This field follows standard label selector semantics; if present but empty, it selects all namespaces.
If PodSelector is also set, then the NetworkPolicyPeer as a whole selects the Pods matching PodSelector in the Namespaces selected by NamespaceSelector. Otherwise it selects all Pods in the Namespaces selected by NamespaceSelector.
This is a label selector which selects Pods. This field follows standard label selector semantics; if present but empty, it selects all pods.
If NamespaceSelector is also set, then the NetworkPolicyPeer as a whole selects the Pods matching PodSelector in the Namespaces selected by NamespaceSelector. Otherwise it selects the Pods matching PodSelector in the policy's own Namespace.
egress.ports ([]NetworkPolicyPort)
List of destination ports for outgoing traffic. Each item in this list is combined using a logical OR. If this field is empty or missing, this rule matches all ports (traffic not restricted by port). If this field is present and contains at least one item, then this rule allows traffic only if the traffic matches at least one port in the list.
NetworkPolicyPort describes a port to allow traffic on
egress.ports.port (IntOrString)
The port on the given protocol. This can either be a numerical or named port on a pod. If this field is not provided, this matches all port names and numbers. If present, only traffic on the specified protocol AND port will be matched.
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
egress.ports.endPort (int32)
If set, indicates that the range of ports from port to endPort, inclusive, should be allowed by the policy. This field cannot be defined if the port field is not defined or if the port field is defined as a named (string) port. The endPort must be equal or greater than port.
egress.ports.protocol (string)
The protocol (TCP, UDP, or SCTP) which traffic must match. If not specified, this field defaults to TCP.
NetworkPolicyStatus
NetworkPolicyStatus describe the current state of the NetworkPolicy.
conditions ([]Condition)
Patch strategy: merge on key type
Map: unique values on key type will be kept during a merge
Conditions holds an array of metav1.Condition that describe the state of the NetworkPolicy. Current service state
Condition contains details for one aspect of the current state of this API Resource.
conditions.lastTransitionTime (Time), required
lastTransitionTime is the last time the condition transitioned from one status to another. This should be when the underlying condition changed. If that is not known, then using the time when the API field changed is acceptable.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string), required
message is a human readable message indicating details about the transition. This may be an empty string.
conditions.reason (string), required
reason contains a programmatic identifier indicating the reason for the condition's last transition. Producers of specific condition types may define expected values and meanings for this field, and whether the values are considered a guaranteed API. The value should be a CamelCase string. This field may not be empty.
conditions.status (string), required
status of the condition, one of True, False, Unknown.
conditions.type (string), required
type of condition in CamelCase or in foo.example.com/CamelCase.
conditions.observedGeneration (int64)
observedGeneration represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.conditions[x].observedGeneration is 9, the condition is out of date with respect to the current state of the instance.
NetworkPolicyList
NetworkPolicyList is a list of NetworkPolicy objects.
Most recently observed status of the PodDisruptionBudget.
PodDisruptionBudgetSpec
PodDisruptionBudgetSpec is a description of a PodDisruptionBudget.
maxUnavailable (IntOrString)
An eviction is allowed if at most "maxUnavailable" pods selected by "selector" are unavailable after the eviction, i.e. even in absence of the evicted pod. For example, one can prevent all voluntary evictions by specifying 0. This is a mutually exclusive setting with "minAvailable".
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
minAvailable (IntOrString)
An eviction is allowed if at least "minAvailable" pods selected by "selector" will still be available after the eviction, i.e. even in the absence of the evicted pod. So for example you can prevent all voluntary evictions by specifying "100%".
IntOrString is a type that can hold an int32 or a string. When used in JSON or YAML marshalling and unmarshalling, it produces or consumes the inner type. This allows you to have, for example, a JSON field that can accept a name or number.
Label query over pods whose evictions are managed by the disruption budget. A null selector will match no pods, while an empty ({}) selector will select all pods within the namespace.
PodDisruptionBudgetStatus
PodDisruptionBudgetStatus represents information about the status of a PodDisruptionBudget. Status may trail the actual state of a system.
currentHealthy (int32), required
current number of healthy pods
desiredHealthy (int32), required
minimum desired number of healthy pods
disruptionsAllowed (int32), required
Number of pod disruptions that are currently allowed.
expectedPods (int32), required
total number of pods counted by this disruption budget
conditions ([]Condition)
Patch strategy: merge on key type
Map: unique values on key type will be kept during a merge
Conditions contain conditions for PDB. The disruption controller sets the DisruptionAllowed condition. The following are known values for the reason field (additional reasons could be added in the future): - SyncFailed: The controller encountered an error and wasn't able to compute
the number of allowed disruptions. Therefore no disruptions are
allowed and the status of the condition will be False.
InsufficientPods: The number of pods are either at or below the number
required by the PodDisruptionBudget. No disruptions are
allowed and the status of the condition will be False.
SufficientPods: There are more pods than required by the PodDisruptionBudget.
The condition will be True, and the number of allowed
disruptions are provided by the disruptionsAllowed property.
Condition contains details for one aspect of the current state of this API Resource.
conditions.lastTransitionTime (Time), required
lastTransitionTime is the last time the condition transitioned from one status to another. This should be when the underlying condition changed. If that is not known, then using the time when the API field changed is acceptable.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string), required
message is a human readable message indicating details about the transition. This may be an empty string.
conditions.reason (string), required
reason contains a programmatic identifier indicating the reason for the condition's last transition. Producers of specific condition types may define expected values and meanings for this field, and whether the values are considered a guaranteed API. The value should be a CamelCase string. This field may not be empty.
conditions.status (string), required
status of the condition, one of True, False, Unknown.
conditions.type (string), required
type of condition in CamelCase or in foo.example.com/CamelCase.
conditions.observedGeneration (int64)
observedGeneration represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.conditions[x].observedGeneration is 9, the condition is out of date with respect to the current state of the instance.
disruptedPods (map[string]Time)
DisruptedPods contains information about pods whose eviction was processed by the API server eviction subresource handler but has not yet been observed by the PodDisruptionBudget controller. A pod will be in this map from the time when the API server processed the eviction request to the time when the pod is seen by PDB controller as having been marked for deletion (or after a timeout). The key in the map is the name of the pod and the value is the time when the API server processed the eviction request. If the deletion didn't occur and a pod is still there it will be removed from the list automatically by PodDisruptionBudget controller after some time. If everything goes smooth this map should be empty for the most of the time. Large number of entries in the map may indicate problems with pod deletions.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
observedGeneration (int64)
Most recent generation observed when updating this PDB status. DisruptionsAllowed and other status information is valid only if observedGeneration equals to PDB's object generation.
PodDisruptionBudgetList
PodDisruptionBudgetList is a collection of PodDisruptionBudgets.
status indicates the actual state of the CustomResourceDefinition
CustomResourceDefinitionSpec
CustomResourceDefinitionSpec describes how a user wants their resource to appear
group (string), required
group is the API group of the defined custom resource. The custom resources are served under /apis/\<group>/.... Must match the name of the CustomResourceDefinition (in the form \<names.plural>.\<group>).
names (CustomResourceDefinitionNames), required
names specify the resource and kind names for the custom resource.
CustomResourceDefinitionNames indicates the names to serve this CustomResourceDefinition
names.kind (string), required
kind is the serialized kind of the resource. It is normally CamelCase and singular. Custom resource instances will use this value as the kind attribute in API calls.
names.plural (string), required
plural is the plural name of the resource to serve. The custom resources are served under /apis/\<group>/\<version>/.../\<plural>. Must match the name of the CustomResourceDefinition (in the form \<names.plural>.\<group>). Must be all lowercase.
names.categories ([]string)
categories is a list of grouped resources this custom resource belongs to (e.g. 'all'). This is published in API discovery documents, and used by clients to support invocations like kubectl get all.
names.listKind (string)
listKind is the serialized kind of the list for this resource. Defaults to "kindList".
names.shortNames ([]string)
shortNames are short names for the resource, exposed in API discovery documents, and used by clients to support invocations like kubectl get \<shortname>. It must be all lowercase.
names.singular (string)
singular is the singular name of the resource. It must be all lowercase. Defaults to lowercased kind.
scope (string), required
scope indicates whether the defined custom resource is cluster- or namespace-scoped. Allowed values are Cluster and Namespaced.
versions is the list of all API versions of the defined custom resource. Version names are used to compute the order in which served versions are listed in API discovery. If the version string is "kube-like", it will sort above non "kube-like" version strings, which are ordered lexicographically. "Kube-like" versions start with a "v", then are followed by a number (the major version), then optionally the string "alpha" or "beta" and another number (the minor version). These are sorted first by GA > beta > alpha (where GA is a version with no suffix such as beta or alpha), and then by comparing major version, then minor version. An example sorted list of versions: v10, v2, v1, v11beta2, v10beta3, v3beta1, v12alpha1, v11alpha2, foo1, foo10.
CustomResourceDefinitionVersion describes a version for CRD.
versions.name (string), required
name is the version name, e.g. “v1”, “v2beta1”, etc. The custom resources are served under this version at /apis/\<group>/\<version>/... if served is true.
versions.served (boolean), required
served is a flag enabling/disabling this version from being served via REST APIs
versions.storage (boolean), required
storage indicates this version should be used when persisting custom resources to storage. There must be exactly one version with storage=true.
priority is an integer defining the relative importance of this column compared to others. Lower numbers are considered higher priority. Columns that may be omitted in limited space scenarios should be given a priority greater than 0.
versions.deprecated (boolean)
deprecated indicates this version of the custom resource API is deprecated. When set to true, API requests to this version receive a warning header in the server response. Defaults to false.
versions.deprecationWarning (string)
deprecationWarning overrides the default warning returned to API clients. May only be set when deprecated is true. The default warning indicates this version is deprecated and recommends use of the newest served version of equal or greater stability, if one exists.
versions.schema (CustomResourceValidation)
schema describes the schema used for validation, pruning, and defaulting of this version of the custom resource.
CustomResourceValidation is a list of validation methods for CustomResources.
specReplicasPath defines the JSON path inside of a custom resource that corresponds to Scale spec.replicas. Only JSON paths without the array notation are allowed. Must be a JSON Path under .spec. If there is no value under the given path in the custom resource, the /scale subresource will return an error on GET.
statusReplicasPath defines the JSON path inside of a custom resource that corresponds to Scale status.replicas. Only JSON paths without the array notation are allowed. Must be a JSON Path under .status. If there is no value under the given path in the custom resource, the status.replicas value in the /scale subresource will default to 0.
labelSelectorPath defines the JSON path inside of a custom resource that corresponds to Scale status.selector. Only JSON paths without the array notation are allowed. Must be a JSON Path under .status or .spec. Must be set to work with HorizontalPodAutoscaler. The field pointed by this JSON path must be a string field (not a complex selector struct) which contains a serialized label selector in string form. More info: https://kubernetes.io/docs/tasks/access-kubernetes-api/custom-resources/custom-resource-definitions#scale-subresource If there is no value under the given path in the custom resource, the status.selector value in the /scale subresource will default to the empty string.
status indicates the custom resource should serve a /status subresource. When enabled: 1. requests to the custom resource primary endpoint ignore changes to the status stanza of the object. 2. requests to the custom resource /status subresource ignore changes to anything other than the status stanza of the object.
CustomResourceSubresourceStatus defines how to serve the status subresource for CustomResources. Status is represented by the .status JSON path inside of a CustomResource. When set, * exposes a /status subresource for the custom resource * PUT requests to the /status subresource take a custom resource object, and ignore changes to anything except the status stanza * PUT/POST/PATCH requests to the custom resource ignore changes to the status stanza
conversion (CustomResourceConversion)
conversion defines conversion settings for the CRD.
CustomResourceConversion describes how to convert different versions of a CR.
conversion.strategy (string), required
strategy specifies how custom resources are converted between versions. Allowed values are: - None: The converter only change the apiVersion and would not touch any other field in the custom resource. - Webhook: API Server will call to an external webhook to do the conversion. Additional information
is needed for this option. This requires spec.preserveUnknownFields to be false, and spec.conversion.webhook to be set.
conversion.webhook (WebhookConversion)
webhook describes how to call the conversion webhook. Required when strategy is set to Webhook.
WebhookConversion describes how to call a conversion webhook
conversionReviewVersions is an ordered list of preferred ConversionReview versions the Webhook expects. The API server will use the first version in the list which it supports. If none of the versions specified in this list are supported by API server, conversion will fail for the custom resource. If a persisted Webhook configuration specifies allowed versions and does not include any versions known to the API Server, calls to the webhook will fail.
clientConfig is the instructions for how to call the webhook if strategy is Webhook.
WebhookClientConfig contains the information to make a TLS connection with the webhook.
conversion.webhook.clientConfig.caBundle ([]byte)
caBundle is a PEM encoded CA bundle which will be used to validate the webhook's server certificate. If unspecified, system trust roots on the apiserver are used.
port is an optional service port at which the webhook will be contacted. port should be a valid port number (1-65535, inclusive). Defaults to 443 for backward compatibility.
conversion.webhook.clientConfig.url (string)
url gives the location of the webhook, in standard URL form (scheme://host:port/path). Exactly one of url or service must be specified.
The host should not refer to a service running in the cluster; use the service field instead. The host might be resolved via external DNS in some apiservers (e.g., kube-apiserver cannot resolve in-cluster DNS as that would be a layering violation). host may also be an IP address.
Please note that using localhost or 127.0.0.1 as a host is risky unless you take great care to run this webhook on all hosts which run an apiserver which might need to make calls to this webhook. Such installs are likely to be non-portable, i.e., not easy to turn up in a new cluster.
The scheme must be "https"; the URL must begin with "https://".
A path is optional, and if present may be any string permissible in a URL. You may use the path to pass an arbitrary string to the webhook, for example, a cluster identifier.
Attempting to use a user or basic auth e.g. "user:password@" is not allowed. Fragments ("#...") and query parameters ("?...") are not allowed, either.
default is a default value for undefined object fields. Defaulting is a beta feature under the CustomResourceDefaulting feature gate. Defaulting requires spec.preserveUnknownFields to be false.
JSON represents any valid JSON value. These types are supported: bool, int64, float64, string, []interface{}, map[string]interface{} and nil.
JSONSchemaPropsOrStringArray represents a JSONSchemaProps or a string array.
description (string)
enum ([]JSON)
JSON represents any valid JSON value. These types are supported: bool, int64, float64, string, []interface{}, map[string]interface{} and nil.
example (JSON)
JSON represents any valid JSON value. These types are supported: bool, int64, float64, string, []interface{}, map[string]interface{} and nil.
exclusiveMaximum (boolean)
exclusiveMinimum (boolean)
externalDocs (ExternalDocumentation)
ExternalDocumentation allows referencing an external resource for extended documentation.
externalDocs.description (string)
externalDocs.url (string)
format (string)
format is an OpenAPI v3 format string. Unknown formats are ignored. The following formats are validated:
bsonobjectid: a bson object ID, i.e. a 24 characters hex string - uri: an URI as parsed by Golang net/url.ParseRequestURI - email: an email address as parsed by Golang net/mail.ParseAddress - hostname: a valid representation for an Internet host name, as defined by RFC 1034, section 3.1 [RFC1034]. - ipv4: an IPv4 IP as parsed by Golang net.ParseIP - ipv6: an IPv6 IP as parsed by Golang net.ParseIP - cidr: a CIDR as parsed by Golang net.ParseCIDR - mac: a MAC address as parsed by Golang net.ParseMAC - uuid: an UUID that allows uppercase defined by the regex (?i)^[0-9a-f]{8}-?[0-9a-f]{4}-?[0-9a-f]{4}-?[0-9a-f]{4}-?[0-9a-f]{12}$ - uuid3: an UUID3 that allows uppercase defined by the regex (?i)^[0-9a-f]{8}-?[0-9a-f]{4}-?3[0-9a-f]{3}-?[0-9a-f]{4}-?[0-9a-f]{12}$ - uuid4: an UUID4 that allows uppercase defined by the regex (?i)^[0-9a-f]{8}-?[0-9a-f]{4}-?4[0-9a-f]{3}-?[89ab][0-9a-f]{3}-?[0-9a-f]{12}$ - uuid5: an UUID5 that allows uppercase defined by the regex (?i)^[0-9a-f]{8}-?[0-9a-f]{4}-?5[0-9a-f]{3}-?[89ab][0-9a-f]{3}-?[0-9a-f]{12}$ - isbn: an ISBN10 or ISBN13 number string like "0321751043" or "978-0321751041" - isbn10: an ISBN10 number string like "0321751043" - isbn13: an ISBN13 number string like "978-0321751041" - creditcard: a credit card number defined by the regex ^(?:4[0-9]{12}(?:[0-9]{3})?|5[1-5][0-9]{14}|6(?:011|5[0-9][0-9])[0-9]{12}|3[47][0-9]{13}|3(?:0[0-5]|[68][0-9])[0-9]{11}|(?:2131|1800|35\d{3})\d{11})$ with any non digit characters mixed in - ssn: a U.S. social security number following the regex ^\d{3}[- ]?\d{2}[- ]?\d{4}$ - hexcolor: an hexadecimal color code like "#FFFFFF: following the regex ^#?([0-9a-fA-F]{3}|[0-9a-fA-F]{6})$ - rgbcolor: an RGB color code like rgb like "rgb(255,255,2559" - byte: base64 encoded binary data - password: any kind of string - date: a date string like "2006-01-02" as defined by full-date in RFC3339 - duration: a duration string like "22 ns" as parsed by Golang time.ParseDuration or compatible with Scala duration format - datetime: a date time string like "2014-12-15T19:30:20.000Z" as defined by date-time in RFC3339.
id (string)
items (JSONSchemaPropsOrArray)
JSONSchemaPropsOrArray represents a value that can either be a JSONSchemaProps or an array of JSONSchemaProps. Mainly here for serialization purposes.
x-kubernetes-embedded-resource defines that the value is an embedded Kubernetes runtime.Object, with TypeMeta and ObjectMeta. The type must be object. It is allowed to further restrict the embedded object. kind, apiVersion and metadata are validated automatically. x-kubernetes-preserve-unknown-fields is allowed to be true, but does not have to be if the object is fully specified (up to kind, apiVersion, metadata).
x-kubernetes-int-or-string (boolean)
x-kubernetes-int-or-string specifies that this value is either an integer or a string. If this is true, an empty type is allowed and type as child of anyOf is permitted if following one of the following patterns:
anyOf:
type: integer
type: string
allOf:
anyOf:
type: integer
type: string
... zero or more
x-kubernetes-list-map-keys ([]string)
x-kubernetes-list-map-keys annotates an array with the x-kubernetes-list-type map by specifying the keys used as the index of the map.
This tag MUST only be used on lists that have the "x-kubernetes-list-type" extension set to "map". Also, the values specified for this attribute must be a scalar typed field of the child structure (no nesting is supported).
The properties specified must either be required or have a default value, to ensure those properties are present for all list items.
x-kubernetes-list-type (string)
x-kubernetes-list-type annotates an array to further describe its topology. This extension must only be used on lists and may have 3 possible values:
atomic: the list is treated as a single entity, like a scalar.
Atomic lists will be entirely replaced when updated. This extension
may be used on any type of list (struct, scalar, ...).
set:
Sets are lists that must not have multiple items with the same value. Each
value must be a scalar, an object with x-kubernetes-map-type atomic or an
array with x-kubernetes-list-type atomic.
map:
These lists are like maps in that their elements have a non-index key
used to identify them. Order is preserved upon merge. The map tag
must only be used on a list with elements of type object.
Defaults to atomic for arrays.
x-kubernetes-map-type (string)
x-kubernetes-map-type annotates an object to further describe its topology. This extension must only be used when type is object and may have 2 possible values:
granular:
These maps are actual maps (key-value pairs) and each fields are independent
from each other (they can each be manipulated by separate actors). This is
the default behaviour for all maps.
atomic: the list is treated as a single entity, like a scalar.
Atomic maps will be entirely replaced when updated.
x-kubernetes-preserve-unknown-fields (boolean)
x-kubernetes-preserve-unknown-fields stops the API server decoding step from pruning fields which are not specified in the validation schema. This affects fields recursively, but switches back to normal pruning behaviour if nested properties or additionalProperties are specified in the schema. This can either be true or undefined. False is forbidden.
x-kubernetes-validations ([]ValidationRule)
Patch strategy: merge on key rule
Map: unique values on key rule will be kept during a merge
x-kubernetes-validations describes a list of validation rules written in the CEL expression language. This field is an alpha-level. Using this field requires the feature gate CustomResourceValidationExpressions to be enabled.
ValidationRule describes a validation rule written in the CEL expression language.
x-kubernetes-validations.rule (string), required
Rule represents the expression which will be evaluated by CEL. ref: https://github.com/google/cel-spec The Rule is scoped to the location of the x-kubernetes-validations extension in the schema. The self variable in the CEL expression is bound to the scoped value. Example: - Rule scoped to the root of a resource with a status subresource: {"rule": "self.status.actual <= self.spec.maxDesired"}
If the Rule is scoped to an object with properties, the accessible properties of the object are field selectable via self.field and field presence can be checked via has(self.field). Null valued fields are treated as absent fields in CEL expressions. If the Rule is scoped to an object with additionalProperties (i.e. a map) the value of the map are accessible via self[mapKey], map containment can be checked via mapKey in self and all entries of the map are accessible via CEL macros and functions such as self.all(...). If the Rule is scoped to an array, the elements of the array are accessible via self[i] and also by macros and functions. If the Rule is scoped to a scalar, self is bound to the scalar value. Examples: - Rule scoped to a map of objects: {"rule": "self.components['Widget'].priority < 10"} - Rule scoped to a list of integers: {"rule": "self.values.all(value, value >= 0 && value < 100)"} - Rule scoped to a string value: {"rule": "self.startsWith('kube')"}
The apiVersion, kind, metadata.name and metadata.generateName are always accessible from the root of the object and from any x-kubernetes-embedded-resource annotated objects. No other metadata properties are accessible.
Unknown data preserved in custom resources via x-kubernetes-preserve-unknown-fields is not accessible in CEL expressions. This includes: - Unknown field values that are preserved by object schemas with x-kubernetes-preserve-unknown-fields. - Object properties where the property schema is of an "unknown type". An "unknown type" is recursively defined as:
A schema with no type and x-kubernetes-preserve-unknown-fields set to true
An array where the items schema is of an "unknown type"
An object where the additionalProperties schema is of an "unknown type"
Only property names of the form [a-zA-Z_.-/][a-zA-Z0-9_.-/]* are accessible. Accessible property names are escaped according to the following rules when accessed in the expression: - '' escapes to 'underscores' - '.' escapes to 'dot' - '-' escapes to 'dash' - '/' escapes to 'slash' - Property names that exactly match a CEL RESERVED keyword escape to '{keyword}__'. The keywords are:
"true", "false", "null", "in", "as", "break", "const", "continue", "else", "for", "function", "if",
"import", "let", "loop", "package", "namespace", "return".
Examples:
Rule accessing a property named "namespace": {"rule": "self.namespace > 0"}
Rule accessing a property named "x-prop": {"rule": "self.x__dash__prop > 0"}
Rule accessing a property named "redact__d": {"rule": "self.redact__underscores__d > 0"}
Equality on arrays with x-kubernetes-list-type of 'set' or 'map' ignores element order, i.e. [1, 2] == [2, 1]. Concatenation on arrays with x-kubernetes-list-type use the semantics of the list type:
'set': X + Y performs a union where the array positions of all elements in X are preserved and
non-intersecting elements in Y are appended, retaining their partial order.
'map': X + Y performs a merge where the array positions of all keys in X are preserved but the values
are overwritten by values in Y when the key sets of X and Y intersect. Elements in Y with
non-intersecting keys are appended, retaining their partial order.
x-kubernetes-validations.message (string)
Message represents the message displayed when validation fails. The message is required if the Rule contains line breaks. The message must not contain line breaks. If unset, the message is "failed rule: {Rule}". e.g. "must be a URL with the host matching spec.host"
CustomResourceDefinitionStatus
CustomResourceDefinitionStatus indicates the state of the CustomResourceDefinition
acceptedNames (CustomResourceDefinitionNames)
acceptedNames are the names that are actually being used to serve discovery. They may be different than the names in spec.
CustomResourceDefinitionNames indicates the names to serve this CustomResourceDefinition
acceptedNames.kind (string), required
kind is the serialized kind of the resource. It is normally CamelCase and singular. Custom resource instances will use this value as the kind attribute in API calls.
acceptedNames.plural (string), required
plural is the plural name of the resource to serve. The custom resources are served under /apis/\<group>/\<version>/.../\<plural>. Must match the name of the CustomResourceDefinition (in the form \<names.plural>.\<group>). Must be all lowercase.
acceptedNames.categories ([]string)
categories is a list of grouped resources this custom resource belongs to (e.g. 'all'). This is published in API discovery documents, and used by clients to support invocations like kubectl get all.
acceptedNames.listKind (string)
listKind is the serialized kind of the list for this resource. Defaults to "kindList".
acceptedNames.shortNames ([]string)
shortNames are short names for the resource, exposed in API discovery documents, and used by clients to support invocations like kubectl get \<shortname>. It must be all lowercase.
acceptedNames.singular (string)
singular is the singular name of the resource. It must be all lowercase. Defaults to lowercased kind.
conditions ([]CustomResourceDefinitionCondition)
Map: unique values on key type will be kept during a merge
conditions indicate state for particular aspects of a CustomResourceDefinition
CustomResourceDefinitionCondition contains details for the current condition of this pod.
conditions.status (string), required
status is the status of the condition. Can be True, False, Unknown.
conditions.type (string), required
type is the type of the condition. Types include Established, NamesAccepted and Terminating.
conditions.lastTransitionTime (Time)
lastTransitionTime last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is a human-readable message indicating details about last transition.
conditions.reason (string)
reason is a unique, one-word, CamelCase reason for the condition's last transition.
storedVersions ([]string)
storedVersions lists all versions of CustomResources that were ever persisted. Tracking these versions allows a migration path for stored versions in etcd. The field is mutable so a migration controller can finish a migration to another version (ensuring no old objects are left in storage), and then remove the rest of the versions from this list. Versions may not be removed from spec.versions while they exist in this list.
CustomResourceDefinitionList
CustomResourceDefinitionList is a list of CustomResourceDefinition objects.
AdmissionReviewVersions is an ordered list of preferred AdmissionReview versions the Webhook expects. API server will try to use first version in the list which it supports. If none of the versions specified in this list supported by API server, validation will fail for this object. If a persisted webhook configuration specifies allowed versions and does not include any versions known to the API Server, calls to the webhook will fail and be subject to the failure policy.
ClientConfig defines how to communicate with the hook. Required
WebhookClientConfig contains the information to make a TLS connection with the webhook
webhooks.clientConfig.caBundle ([]byte)
caBundle is a PEM encoded CA bundle which will be used to validate the webhook's server certificate. If unspecified, system trust roots on the apiserver are used.
webhooks.clientConfig.service (ServiceReference)
service is a reference to the service for this webhook. Either service or url must be specified.
If the webhook is running within the cluster, then you should use service.
ServiceReference holds a reference to Service.legacy.k8s.io
namespace is the namespace of the service. Required
webhooks.clientConfig.service.path (string)
path is an optional URL path which will be sent in any request to this service.
webhooks.clientConfig.service.port (int32)
If specified, the port on the service that hosting webhook. Default to 443 for backward compatibility. port should be a valid port number (1-65535, inclusive).
webhooks.clientConfig.url (string)
url gives the location of the webhook, in standard URL form (scheme://host:port/path). Exactly one of url or service must be specified.
The host should not refer to a service running in the cluster; use the service field instead. The host might be resolved via external DNS in some apiservers (e.g., kube-apiserver cannot resolve in-cluster DNS as that would be a layering violation). host may also be an IP address.
Please note that using localhost or 127.0.0.1 as a host is risky unless you take great care to run this webhook on all hosts which run an apiserver which might need to make calls to this webhook. Such installs are likely to be non-portable, i.e., not easy to turn up in a new cluster.
The scheme must be "https"; the URL must begin with "https://".
A path is optional, and if present may be any string permissible in a URL. You may use the path to pass an arbitrary string to the webhook, for example, a cluster identifier.
Attempting to use a user or basic auth e.g. "user:password@" is not allowed. Fragments ("#...") and query parameters ("?...") are not allowed, either.
webhooks.name (string), required
The name of the admission webhook. Name should be fully qualified, e.g., imagepolicy.kubernetes.io, where "imagepolicy" is the name of the webhook, and kubernetes.io is the name of the organization. Required.
webhooks.sideEffects (string), required
SideEffects states whether this webhook has side effects. Acceptable values are: None, NoneOnDryRun (webhooks created via v1beta1 may also specify Some or Unknown). Webhooks with side effects MUST implement a reconciliation system, since a request may be rejected by a future step in the admission chain and the side effects therefore need to be undone. Requests with the dryRun attribute will be auto-rejected if they match a webhook with sideEffects == Unknown or Some.
webhooks.failurePolicy (string)
FailurePolicy defines how unrecognized errors from the admission endpoint are handled - allowed values are Ignore or Fail. Defaults to Fail.
webhooks.matchPolicy (string)
matchPolicy defines how the "rules" list is used to match incoming requests. Allowed values are "Exact" or "Equivalent".
Exact: match a request only if it exactly matches a specified rule. For example, if deployments can be modified via apps/v1, apps/v1beta1, and extensions/v1beta1, but "rules" only included apiGroups:["apps"], apiVersions:["v1"], resources: ["deployments"], a request to apps/v1beta1 or extensions/v1beta1 would not be sent to the webhook.
Equivalent: match a request if modifies a resource listed in rules, even via another API group or version. For example, if deployments can be modified via apps/v1, apps/v1beta1, and extensions/v1beta1, and "rules" only included apiGroups:["apps"], apiVersions:["v1"], resources: ["deployments"], a request to apps/v1beta1 or extensions/v1beta1 would be converted to apps/v1 and sent to the webhook.
NamespaceSelector decides whether to run the webhook on an object based on whether the namespace for that object matches the selector. If the object itself is a namespace, the matching is performed on object.metadata.labels. If the object is another cluster scoped resource, it never skips the webhook.
For example, to run the webhook on any objects whose namespace is not associated with "runlevel" of "0" or "1"; you will set the selector as follows: "namespaceSelector": {
"matchExpressions": [
{
"key": "runlevel",
"operator": "NotIn",
"values": [
"0",
"1"
]
}
]
}
If instead you want to only run the webhook on any objects whose namespace is associated with the "environment" of "prod" or "staging"; you will set the selector as follows: "namespaceSelector": {
"matchExpressions": [
{
"key": "environment",
"operator": "In",
"values": [
"prod",
"staging"
]
}
]
}
ObjectSelector decides whether to run the webhook based on if the object has matching labels. objectSelector is evaluated against both the oldObject and newObject that would be sent to the webhook, and is considered to match if either object matches the selector. A null object (oldObject in the case of create, or newObject in the case of delete) or an object that cannot have labels (like a DeploymentRollback or a PodProxyOptions object) is not considered to match. Use the object selector only if the webhook is opt-in, because end users may skip the admission webhook by setting the labels. Default to the empty LabelSelector, which matches everything.
webhooks.reinvocationPolicy (string)
reinvocationPolicy indicates whether this webhook should be called multiple times as part of a single admission evaluation. Allowed values are "Never" and "IfNeeded".
Never: the webhook will not be called more than once in a single admission evaluation.
IfNeeded: the webhook will be called at least one additional time as part of the admission evaluation if the object being admitted is modified by other admission plugins after the initial webhook call. Webhooks that specify this option must be idempotent, able to process objects they previously admitted. Note: * the number of additional invocations is not guaranteed to be exactly one. * if additional invocations result in further modifications to the object, webhooks are not guaranteed to be invoked again. * webhooks that use this option may be reordered to minimize the number of additional invocations. * to validate an object after all mutations are guaranteed complete, use a validating admission webhook instead.
Defaults to "Never".
webhooks.rules ([]RuleWithOperations)
Rules describes what operations on what resources/subresources the webhook cares about. The webhook cares about an operation if it matches any Rule. However, in order to prevent ValidatingAdmissionWebhooks and MutatingAdmissionWebhooks from putting the cluster in a state which cannot be recovered from without completely disabling the plugin, ValidatingAdmissionWebhooks and MutatingAdmissionWebhooks are never called on admission requests for ValidatingWebhookConfiguration and MutatingWebhookConfiguration objects.
RuleWithOperations is a tuple of Operations and Resources. It is recommended to make sure that all the tuple expansions are valid.
webhooks.rules.apiGroups ([]string)
APIGroups is the API groups the resources belong to. '' is all groups. If '' is present, the length of the slice must be one. Required.
webhooks.rules.apiVersions ([]string)
APIVersions is the API versions the resources belong to. '' is all versions. If '' is present, the length of the slice must be one. Required.
webhooks.rules.operations ([]string)
Operations is the operations the admission hook cares about - CREATE, UPDATE, DELETE, CONNECT or * for all of those operations and any future admission operations that are added. If '*' is present, the length of the slice must be one. Required.
webhooks.rules.resources ([]string)
Resources is a list of resources this rule applies to.
For example: 'pods' means pods. 'pods/log' means the log subresource of pods. '' means all resources, but not subresources. 'pods/' means all subresources of pods. '/scale' means all scale subresources. '/*' means all resources and their subresources.
If wildcard is present, the validation rule will ensure resources do not overlap with each other.
Depending on the enclosing object, subresources might not be allowed. Required.
webhooks.rules.scope (string)
scope specifies the scope of this rule. Valid values are "Cluster", "Namespaced", and "" "Cluster" means that only cluster-scoped resources will match this rule. Namespace API objects are cluster-scoped. "Namespaced" means that only namespaced resources will match this rule. "" means that there are no scope restrictions. Subresources match the scope of their parent resource. Default is "*".
webhooks.timeoutSeconds (int32)
TimeoutSeconds specifies the timeout for this webhook. After the timeout passes, the webhook call will be ignored or the API call will fail based on the failure policy. The timeout value must be between 1 and 30 seconds. Default to 10 seconds.
MutatingWebhookConfigurationList
MutatingWebhookConfigurationList is a list of MutatingWebhookConfiguration.
AdmissionReviewVersions is an ordered list of preferred AdmissionReview versions the Webhook expects. API server will try to use first version in the list which it supports. If none of the versions specified in this list supported by API server, validation will fail for this object. If a persisted webhook configuration specifies allowed versions and does not include any versions known to the API Server, calls to the webhook will fail and be subject to the failure policy.
ClientConfig defines how to communicate with the hook. Required
WebhookClientConfig contains the information to make a TLS connection with the webhook
webhooks.clientConfig.caBundle ([]byte)
caBundle is a PEM encoded CA bundle which will be used to validate the webhook's server certificate. If unspecified, system trust roots on the apiserver are used.
webhooks.clientConfig.service (ServiceReference)
service is a reference to the service for this webhook. Either service or url must be specified.
If the webhook is running within the cluster, then you should use service.
ServiceReference holds a reference to Service.legacy.k8s.io
namespace is the namespace of the service. Required
webhooks.clientConfig.service.path (string)
path is an optional URL path which will be sent in any request to this service.
webhooks.clientConfig.service.port (int32)
If specified, the port on the service that hosting webhook. Default to 443 for backward compatibility. port should be a valid port number (1-65535, inclusive).
webhooks.clientConfig.url (string)
url gives the location of the webhook, in standard URL form (scheme://host:port/path). Exactly one of url or service must be specified.
The host should not refer to a service running in the cluster; use the service field instead. The host might be resolved via external DNS in some apiservers (e.g., kube-apiserver cannot resolve in-cluster DNS as that would be a layering violation). host may also be an IP address.
Please note that using localhost or 127.0.0.1 as a host is risky unless you take great care to run this webhook on all hosts which run an apiserver which might need to make calls to this webhook. Such installs are likely to be non-portable, i.e., not easy to turn up in a new cluster.
The scheme must be "https"; the URL must begin with "https://".
A path is optional, and if present may be any string permissible in a URL. You may use the path to pass an arbitrary string to the webhook, for example, a cluster identifier.
Attempting to use a user or basic auth e.g. "user:password@" is not allowed. Fragments ("#...") and query parameters ("?...") are not allowed, either.
webhooks.name (string), required
The name of the admission webhook. Name should be fully qualified, e.g., imagepolicy.kubernetes.io, where "imagepolicy" is the name of the webhook, and kubernetes.io is the name of the organization. Required.
webhooks.sideEffects (string), required
SideEffects states whether this webhook has side effects. Acceptable values are: None, NoneOnDryRun (webhooks created via v1beta1 may also specify Some or Unknown). Webhooks with side effects MUST implement a reconciliation system, since a request may be rejected by a future step in the admission chain and the side effects therefore need to be undone. Requests with the dryRun attribute will be auto-rejected if they match a webhook with sideEffects == Unknown or Some.
webhooks.failurePolicy (string)
FailurePolicy defines how unrecognized errors from the admission endpoint are handled - allowed values are Ignore or Fail. Defaults to Fail.
webhooks.matchPolicy (string)
matchPolicy defines how the "rules" list is used to match incoming requests. Allowed values are "Exact" or "Equivalent".
Exact: match a request only if it exactly matches a specified rule. For example, if deployments can be modified via apps/v1, apps/v1beta1, and extensions/v1beta1, but "rules" only included apiGroups:["apps"], apiVersions:["v1"], resources: ["deployments"], a request to apps/v1beta1 or extensions/v1beta1 would not be sent to the webhook.
Equivalent: match a request if modifies a resource listed in rules, even via another API group or version. For example, if deployments can be modified via apps/v1, apps/v1beta1, and extensions/v1beta1, and "rules" only included apiGroups:["apps"], apiVersions:["v1"], resources: ["deployments"], a request to apps/v1beta1 or extensions/v1beta1 would be converted to apps/v1 and sent to the webhook.
NamespaceSelector decides whether to run the webhook on an object based on whether the namespace for that object matches the selector. If the object itself is a namespace, the matching is performed on object.metadata.labels. If the object is another cluster scoped resource, it never skips the webhook.
For example, to run the webhook on any objects whose namespace is not associated with "runlevel" of "0" or "1"; you will set the selector as follows: "namespaceSelector": {
"matchExpressions": [
{
"key": "runlevel",
"operator": "NotIn",
"values": [
"0",
"1"
]
}
]
}
If instead you want to only run the webhook on any objects whose namespace is associated with the "environment" of "prod" or "staging"; you will set the selector as follows: "namespaceSelector": {
"matchExpressions": [
{
"key": "environment",
"operator": "In",
"values": [
"prod",
"staging"
]
}
]
}
ObjectSelector decides whether to run the webhook based on if the object has matching labels. objectSelector is evaluated against both the oldObject and newObject that would be sent to the webhook, and is considered to match if either object matches the selector. A null object (oldObject in the case of create, or newObject in the case of delete) or an object that cannot have labels (like a DeploymentRollback or a PodProxyOptions object) is not considered to match. Use the object selector only if the webhook is opt-in, because end users may skip the admission webhook by setting the labels. Default to the empty LabelSelector, which matches everything.
webhooks.rules ([]RuleWithOperations)
Rules describes what operations on what resources/subresources the webhook cares about. The webhook cares about an operation if it matches any Rule. However, in order to prevent ValidatingAdmissionWebhooks and MutatingAdmissionWebhooks from putting the cluster in a state which cannot be recovered from without completely disabling the plugin, ValidatingAdmissionWebhooks and MutatingAdmissionWebhooks are never called on admission requests for ValidatingWebhookConfiguration and MutatingWebhookConfiguration objects.
RuleWithOperations is a tuple of Operations and Resources. It is recommended to make sure that all the tuple expansions are valid.
webhooks.rules.apiGroups ([]string)
APIGroups is the API groups the resources belong to. '' is all groups. If '' is present, the length of the slice must be one. Required.
webhooks.rules.apiVersions ([]string)
APIVersions is the API versions the resources belong to. '' is all versions. If '' is present, the length of the slice must be one. Required.
webhooks.rules.operations ([]string)
Operations is the operations the admission hook cares about - CREATE, UPDATE, DELETE, CONNECT or * for all of those operations and any future admission operations that are added. If '*' is present, the length of the slice must be one. Required.
webhooks.rules.resources ([]string)
Resources is a list of resources this rule applies to.
For example: 'pods' means pods. 'pods/log' means the log subresource of pods. '' means all resources, but not subresources. 'pods/' means all subresources of pods. '/scale' means all scale subresources. '/*' means all resources and their subresources.
If wildcard is present, the validation rule will ensure resources do not overlap with each other.
Depending on the enclosing object, subresources might not be allowed. Required.
webhooks.rules.scope (string)
scope specifies the scope of this rule. Valid values are "Cluster", "Namespaced", and "" "Cluster" means that only cluster-scoped resources will match this rule. Namespace API objects are cluster-scoped. "Namespaced" means that only namespaced resources will match this rule. "" means that there are no scope restrictions. Subresources match the scope of their parent resource. Default is "*".
webhooks.timeoutSeconds (int32)
TimeoutSeconds specifies the timeout for this webhook. After the timeout passes, the webhook call will be ignored or the API call will fail based on the failure policy. The timeout value must be between 1 and 30 seconds. Default to 10 seconds.
ValidatingWebhookConfigurationList
ValidatingWebhookConfigurationList is a list of ValidatingWebhookConfiguration.
NodeSpec describes the attributes that a node is created with.
configSource (NodeConfigSource)
Deprecated: Previously used to specify the source of the node's configuration for the DynamicKubeletConfig feature. This feature is removed from Kubelets as of 1.24 and will be fully removed in 1.26.
NodeConfigSource specifies a source of node configuration. Exactly one subfield (excluding metadata) must be non-nil. This API is deprecated since 1.22
PodCIDR represents the pod IP range assigned to the node.
podCIDRs ([]string)
podCIDRs represents the IP ranges assigned to the node for usage by Pods on that node. If this field is specified, the 0th entry must match the podCIDR field. It may contain at most 1 value for each of IPv4 and IPv6.
providerID (string)
ID of the node assigned by the cloud provider in the format: <ProviderName>://<ProviderSpecificNodeID>
taints ([]Taint)
If specified, the node's taints.
The node this Taint is attached to has the "effect" on any pod that does not tolerate the Taint.
taints.effect (string), required
Required. The effect of the taint on pods that do not tolerate the taint. Valid effects are NoSchedule, PreferNoSchedule and NoExecute.
taints.key (string), required
Required. The taint key to be applied to a node.
taints.timeAdded (Time)
TimeAdded represents the time at which the taint was added. It is only written for NoExecute taints.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
NodeStatus is information about the current status of a node.
addresses ([]NodeAddress)
Patch strategy: merge on key type
List of addresses reachable to the node. Queried from cloud provider, if available. More info: https://kubernetes.io/docs/concepts/nodes/node/#addresses Note: This field is declared as mergeable, but the merge key is not sufficiently unique, which can cause data corruption when it is merged. Callers should instead use a full-replacement patch. See http://pr.k8s.io/79391 for an example.
NodeAddress contains information for the node's address.
addresses.address (string), required
The node address.
addresses.type (string), required
Node address type, one of Hostname, ExternalIP or InternalIP.
NodeCondition contains condition information for a node.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of node condition.
conditions.lastHeartbeatTime (Time)
Last time we got an update on a given condition.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.lastTransitionTime (Time)
Last time the condition transit from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
Human readable message indicating details about last transition.
conditions.reason (string)
(brief) reason for the condition's last transition.
config (NodeConfigStatus)
Status of the config assigned to the node via the dynamic Kubelet config feature.
NodeConfigStatus describes the status of the config assigned by Node.Spec.ConfigSource.
config.active (NodeConfigSource)
Active reports the checkpointed config the node is actively using. Active will represent either the current version of the Assigned config, or the current LastKnownGood config, depending on whether attempting to use the Assigned config results in an error.
NodeConfigSource specifies a source of node configuration. Exactly one subfield (excluding metadata) must be non-nil. This API is deprecated since 1.22
Namespace is the metadata.namespace of the referenced ConfigMap. This field is required in all cases.
config.active.configMap.resourceVersion (string)
ResourceVersion is the metadata.ResourceVersion of the referenced ConfigMap. This field is forbidden in Node.Spec, and required in Node.Status.
config.active.configMap.uid (string)
UID is the metadata.UID of the referenced ConfigMap. This field is forbidden in Node.Spec, and required in Node.Status.
config.assigned (NodeConfigSource)
Assigned reports the checkpointed config the node will try to use. When Node.Spec.ConfigSource is updated, the node checkpoints the associated config payload to local disk, along with a record indicating intended config. The node refers to this record to choose its config checkpoint, and reports this record in Assigned. Assigned only updates in the status after the record has been checkpointed to disk. When the Kubelet is restarted, it tries to make the Assigned config the Active config by loading and validating the checkpointed payload identified by Assigned.
NodeConfigSource specifies a source of node configuration. Exactly one subfield (excluding metadata) must be non-nil. This API is deprecated since 1.22
ResourceVersion is the metadata.ResourceVersion of the referenced ConfigMap. This field is forbidden in Node.Spec, and required in Node.Status.
config.assigned.configMap.uid (string)
UID is the metadata.UID of the referenced ConfigMap. This field is forbidden in Node.Spec, and required in Node.Status.
config.error (string)
Error describes any problems reconciling the Spec.ConfigSource to the Active config. Errors may occur, for example, attempting to checkpoint Spec.ConfigSource to the local Assigned record, attempting to checkpoint the payload associated with Spec.ConfigSource, attempting to load or validate the Assigned config, etc. Errors may occur at different points while syncing config. Earlier errors (e.g. download or checkpointing errors) will not result in a rollback to LastKnownGood, and may resolve across Kubelet retries. Later errors (e.g. loading or validating a checkpointed config) will result in a rollback to LastKnownGood. In the latter case, it is usually possible to resolve the error by fixing the config assigned in Spec.ConfigSource. You can find additional information for debugging by searching the error message in the Kubelet log. Error is a human-readable description of the error state; machines can check whether or not Error is empty, but should not rely on the stability of the Error text across Kubelet versions.
config.lastKnownGood (NodeConfigSource)
LastKnownGood reports the checkpointed config the node will fall back to when it encounters an error attempting to use the Assigned config. The Assigned config becomes the LastKnownGood config when the node determines that the Assigned config is stable and correct. This is currently implemented as a 10-minute soak period starting when the local record of Assigned config is updated. If the Assigned config is Active at the end of this period, it becomes the LastKnownGood. Note that if Spec.ConfigSource is reset to nil (use local defaults), the LastKnownGood is also immediately reset to nil, because the local default config is always assumed good. You should not make assumptions about the node's method of determining config stability and correctness, as this may change or become configurable in the future.
NodeConfigSource specifies a source of node configuration. Exactly one subfield (excluding metadata) must be non-nil. This API is deprecated since 1.22
NamespaceStatus is information about the current status of a Namespace.
conditions ([]NamespaceCondition)
Patch strategy: merge on key type
Represents the latest available observations of a namespace's current state.
NamespaceCondition contains details about state of namespace.
conditions.status (string), required
Status of the condition, one of True, False, Unknown.
conditions.type (string), required
Type of namespace controller condition.
conditions.lastTransitionTime (Time)
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
Event is a report of an event somewhere in the cluster.
apiVersion: events.k8s.io/v1
import "k8s.io/api/events/v1"
Event
Event is a report of an event somewhere in the cluster. It generally denotes some state change in the system. Events have a limited retention time and triggers and messages may evolve with time. Event consumers should not rely on the timing of an event with a given Reason reflecting a consistent underlying trigger, or the continued existence of events with that Reason. Events should be treated as informative, best-effort, supplemental data.
eventTime is the time when this Event was first observed. It is required.
MicroTime is version of Time with microsecond level precision.
action (string)
action is what action was taken/failed regarding to the regarding object. It is machine-readable. This field cannot be empty for new Events and it can have at most 128 characters.
deprecatedCount (int32)
deprecatedCount is the deprecated field assuring backward compatibility with core.v1 Event type.
deprecatedFirstTimestamp (Time)
deprecatedFirstTimestamp is the deprecated field assuring backward compatibility with core.v1 Event type.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
deprecatedLastTimestamp (Time)
deprecatedLastTimestamp is the deprecated field assuring backward compatibility with core.v1 Event type.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
deprecatedSource (EventSource)
deprecatedSource is the deprecated field assuring backward compatibility with core.v1 Event type.
EventSource contains information for an event.
deprecatedSource.component (string)
Component from which the event is generated.
deprecatedSource.host (string)
Node name on which the event is generated.
note (string)
note is a human-readable description of the status of this operation. Maximal length of the note is 1kB, but libraries should be prepared to handle values up to 64kB.
reason (string)
reason is why the action was taken. It is human-readable. This field cannot be empty for new Events and it can have at most 128 characters.
regarding contains the object this Event is about. In most cases it's an Object reporting controller implements, e.g. ReplicaSetController implements ReplicaSets and this event is emitted because it acts on some changes in a ReplicaSet object.
related is the optional secondary object for more complex actions. E.g. when regarding object triggers a creation or deletion of related object.
reportingController (string)
reportingController is the name of the controller that emitted this Event, e.g. kubernetes.io/kubelet. This field cannot be empty for new Events.
reportingInstance (string)
reportingInstance is the ID of the controller instance, e.g. kubelet-xyzf. This field cannot be empty for new Events and it can have at most 128 characters.
series (EventSeries)
series is data about the Event series this event represents or nil if it's a singleton Event.
EventSeries contain information on series of events, i.e. thing that was/is happening continuously for some time. How often to update the EventSeries is up to the event reporters. The default event reporter in "k8s.io/client-go/tools/events/event_broadcaster.go" shows how this struct is updated on heartbeats and can guide customized reporter implementations.
series.count (int32), required
count is the number of occurrences in this series up to the last heartbeat time.
series.lastObservedTime (MicroTime), required
lastObservedTime is the time when last Event from the series was seen before last heartbeat.
MicroTime is version of Time with microsecond level precision.
type (string)
type is the type of this event (Normal, Warning), new types could be added in the future. It is machine-readable. This field cannot be empty for new Events.
Status contains derived information about an API server
APIServiceSpec
APIServiceSpec contains information for locating and communicating with a server. Only https is supported, though you are able to disable certificate verification.
groupPriorityMinimum (int32), required
GroupPriorityMininum is the priority this group should have at least. Higher priority means that the group is preferred by clients over lower priority ones. Note that other versions of this group might specify even higher GroupPriorityMininum values such that the whole group gets a higher priority. The primary sort is based on GroupPriorityMinimum, ordered highest number to lowest (20 before 10). The secondary sort is based on the alphabetical comparison of the name of the object. (v1.bar before v1.foo) We'd recommend something like: *.k8s.io (except extensions) at 18000 and PaaSes (OpenShift, Deis) are recommended to be in the 2000s
versionPriority (int32), required
VersionPriority controls the ordering of this API version inside of its group. Must be greater than zero. The primary sort is based on VersionPriority, ordered highest to lowest (20 before 10). Since it's inside of a group, the number can be small, probably in the 10s. In case of equal version priorities, the version string will be used to compute the order inside a group. If the version string is "kube-like", it will sort above non "kube-like" version strings, which are ordered lexicographically. "Kube-like" versions start with a "v", then are followed by a number (the major version), then optionally the string "alpha" or "beta" and another number (the minor version). These are sorted first by GA > beta > alpha (where GA is a version with no suffix such as beta or alpha), and then by comparing major version, then minor version. An example sorted list of versions: v10, v2, v1, v11beta2, v10beta3, v3beta1, v12alpha1, v11alpha2, foo1, foo10.
caBundle ([]byte)
Atomic: will be replaced during a merge
CABundle is a PEM encoded CA bundle which will be used to validate an API server's serving certificate. If unspecified, system trust roots on the apiserver are used.
group (string)
Group is the API group name this server hosts
insecureSkipTLSVerify (boolean)
InsecureSkipTLSVerify disables TLS certificate verification when communicating with this server. This is strongly discouraged. You should use the CABundle instead.
service (ServiceReference)
Service is a reference to the service for this API server. It must communicate on port 443. If the Service is nil, that means the handling for the API groupversion is handled locally on this server. The call will simply delegate to the normal handler chain to be fulfilled.
ServiceReference holds a reference to Service.legacy.k8s.io
service.name (string)
Name is the name of the service
service.namespace (string)
Namespace is the namespace of the service
service.port (int32)
If specified, the port on the service that hosting webhook. Default to 443 for backward compatibility. port should be a valid port number (1-65535, inclusive).
version (string)
Version is the API version this server hosts. For example, "v1"
APIServiceStatus
APIServiceStatus contains derived information about an API server
conditions ([]APIServiceCondition)
Patch strategy: merge on key type
Map: unique values on key type will be kept during a merge
Current service state of apiService.
APIServiceCondition describes the state of an APIService at a particular point
conditions.status (string), required
Status is the status of the condition. Can be True, False, Unknown.
conditions.type (string), required
Type is the type of the condition.
conditions.lastTransitionTime (Time)
Last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
Human-readable message indicating details about last transition.
conditions.reason (string)
Unique, one-word, CamelCase reason for the condition's last transition.
acquireTime is a time when the current lease was acquired.
MicroTime is version of Time with microsecond level precision.
holderIdentity (string)
holderIdentity contains the identity of the holder of a current lease.
leaseDurationSeconds (int32)
leaseDurationSeconds is a duration that candidates for a lease need to wait to force acquire it. This is measure against time of last observed RenewTime.
leaseTransitions (int32)
leaseTransitions is the number of transitions of a lease between holders.
renewTime (MicroTime)
renewTime is a time when the current holder of a lease has last updated the lease.
MicroTime is version of Time with microsecond level precision.
RuntimeClass defines a class of container runtime supported in the cluster.
apiVersion: node.k8s.io/v1
import "k8s.io/api/node/v1"
RuntimeClass
RuntimeClass defines a class of container runtime supported in the cluster. The RuntimeClass is used to determine which container runtime is used to run all containers in a pod. RuntimeClasses are manually defined by a user or cluster provisioner, and referenced in the PodSpec. The Kubelet is responsible for resolving the RuntimeClassName reference before running the pod. For more details, see https://kubernetes.io/docs/concepts/containers/runtime-class/
Handler specifies the underlying runtime and configuration that the CRI implementation will use to handle pods of this class. The possible values are specific to the node & CRI configuration. It is assumed that all handlers are available on every node, and handlers of the same name are equivalent on every node. For example, a handler called "runc" might specify that the runc OCI runtime (using native Linux containers) will be used to run the containers in a pod. The Handler must be lowercase, conform to the DNS Label (RFC 1123) requirements, and is immutable.
PodFixed represents the fixed resource overhead associated with running a pod.
scheduling (Scheduling)
Scheduling holds the scheduling constraints to ensure that pods running with this RuntimeClass are scheduled to nodes that support it. If scheduling is nil, this RuntimeClass is assumed to be supported by all nodes.
Scheduling specifies the scheduling constraints for nodes supporting a RuntimeClass.
scheduling.nodeSelector (map[string]string)
nodeSelector lists labels that must be present on nodes that support this RuntimeClass. Pods using this RuntimeClass can only be scheduled to a node matched by this selector. The RuntimeClass nodeSelector is merged with a pod's existing nodeSelector. Any conflicts will cause the pod to be rejected in admission.
scheduling.tolerations ([]Toleration)
Atomic: will be replaced during a merge
tolerations are appended (excluding duplicates) to pods running with this RuntimeClass during admission, effectively unioning the set of nodes tolerated by the pod and the RuntimeClass.
The pod this Toleration is attached to tolerates any taint that matches the triple <key,value,effect> using the matching operator .
scheduling.tolerations.key (string)
Key is the taint key that the toleration applies to. Empty means match all taint keys. If the key is empty, operator must be Exists; this combination means to match all values and all keys.
scheduling.tolerations.operator (string)
Operator represents a key's relationship to the value. Valid operators are Exists and Equal. Defaults to Equal. Exists is equivalent to wildcard for value, so that a pod can tolerate all taints of a particular category.
scheduling.tolerations.value (string)
Value is the taint value the toleration matches to. If the operator is Exists, the value should be empty, otherwise just a regular string.
scheduling.tolerations.effect (string)
Effect indicates the taint effect to match. Empty means match all taint effects. When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
scheduling.tolerations.tolerationSeconds (int64)
TolerationSeconds represents the period of time the toleration (which must be of effect NoExecute, otherwise this field is ignored) tolerates the taint. By default, it is not set, which means tolerate the taint forever (do not evict). Zero and negative values will be treated as 0 (evict immediately) by the system.
RuntimeClassList
RuntimeClassList is a list of RuntimeClass objects.
FlowSchema defines the schema of a group of flows.
apiVersion: flowcontrol.apiserver.k8s.io/v1beta2
import "k8s.io/api/flowcontrol/v1beta2"
FlowSchema
FlowSchema defines the schema of a group of flows. Note that a flow is made up of a set of inbound API requests with similar attributes and is identified by a pair of strings: the name of the FlowSchema and a "flow distinguisher".
priorityLevelConfiguration should reference a PriorityLevelConfiguration in the cluster. If the reference cannot be resolved, the FlowSchema will be ignored and marked as invalid in its status. Required.
PriorityLevelConfigurationReference contains information that points to the "request-priority" being used.
name is the name of the priority level configuration being referenced Required.
distinguisherMethod (FlowDistinguisherMethod)
distinguisherMethod defines how to compute the flow distinguisher for requests that match this schema. nil specifies that the distinguisher is disabled and thus will always be the empty string.
FlowDistinguisherMethod specifies the method of a flow distinguisher.
distinguisherMethod.type (string), required
type is the type of flow distinguisher method The supported types are "ByUser" and "ByNamespace". Required.
matchingPrecedence (int32)
matchingPrecedence is used to choose among the FlowSchemas that match a given request. The chosen FlowSchema is among those with the numerically lowest (which we take to be logically highest) MatchingPrecedence. Each MatchingPrecedence value must be ranged in [1,10000]. Note that if the precedence is not specified, it will be set to 1000 as default.
rules ([]PolicyRulesWithSubjects)
Atomic: will be replaced during a merge
rules describes which requests will match this flow schema. This FlowSchema matches a request if and only if at least one member of rules matches the request. if it is an empty slice, there will be no requests matching the FlowSchema.
PolicyRulesWithSubjects prescribes a test that applies to a request to an apiserver. The test considers the subject making the request, the verb being requested, and the resource to be acted upon. This PolicyRulesWithSubjects matches a request if and only if both (a) at least one member of subjects matches the request and (b) at least one member of resourceRules or nonResourceRules matches the request.
rules.subjects ([]Subject), required
Atomic: will be replaced during a merge
subjects is the list of normal user, serviceaccount, or group that this rule cares about. There must be at least one member in this slice. A slice that includes both the system:authenticated and system:unauthenticated user groups matches every request. Required.
Subject matches the originator of a request, as identified by the request authentication system. There are three ways of matching an originator; by user, group, or service account.
rules.subjects.kind (string), required
kind indicates which one of the other fields is non-empty. Required
rules.subjects.group (GroupSubject)
group matches based on user group name.
GroupSubject holds detailed information for group-kind subject.
namespace is the namespace of matching ServiceAccount objects. Required.
rules.subjects.user (UserSubject)
user matches based on username.
UserSubject holds detailed information for user-kind subject.
rules.subjects.user.name (string), required
name is the username that matches, or "*" to match all usernames. Required.
rules.nonResourceRules ([]NonResourcePolicyRule)
Atomic: will be replaced during a merge
nonResourceRules is a list of NonResourcePolicyRules that identify matching requests according to their verb and the target non-resource URL.
NonResourcePolicyRule is a predicate that matches non-resource requests according to their verb and the target non-resource URL. A NonResourcePolicyRule matches a request if and only if both (a) at least one member of verbs matches the request and (b) at least one member of nonResourceURLs matches the request.
nonResourceURLs is a set of url prefixes that a user should have access to and may not be empty. For example:
"/healthz" is legal
"/hea*" is illegal
"/hea" is legal but matches nothing
"/hea/*" also matches nothing
"/healthz/" matches all per-component health checks.
"" matches all non-resource urls. if it is present, it must be the only entry. Required.
rules.nonResourceRules.verbs ([]string), required
Set: unique values will be kept during a merge
verbs is a list of matching verbs and may not be empty. "*" matches all verbs. If it is present, it must be the only entry. Required.
rules.resourceRules ([]ResourcePolicyRule)
Atomic: will be replaced during a merge
resourceRules is a slice of ResourcePolicyRules that identify matching requests according to their verb and the target resource. At least one of resourceRules and nonResourceRules has to be non-empty.
ResourcePolicyRule is a predicate that matches some resource requests, testing the request's verb and the target resource. A ResourcePolicyRule matches a resource request if and only if: (a) at least one member of verbs matches the request, (b) at least one member of apiGroups matches the request, (c) at least one member of resources matches the request, and (d) either (d1) the request does not specify a namespace (i.e., Namespace=="") and clusterScope is true or (d2) the request specifies a namespace and least one member of namespaces matches the request's namespace.
resources is a list of matching resources (i.e., lowercase and plural) with, if desired, subresource. For example, [ "services", "nodes/status" ]. This list may not be empty. "*" matches all resources and, if present, must be the only entry. Required.
rules.resourceRules.verbs ([]string), required
Set: unique values will be kept during a merge
verbs is a list of matching verbs and may not be empty. "*" matches all verbs and, if present, must be the only entry. Required.
rules.resourceRules.clusterScope (boolean)
clusterScope indicates whether to match requests that do not specify a namespace (which happens either because the resource is not namespaced or the request targets all namespaces). If this field is omitted or false then the namespaces field must contain a non-empty list.
rules.resourceRules.namespaces ([]string)
Set: unique values will be kept during a merge
namespaces is a list of target namespaces that restricts matches. A request that specifies a target namespace matches only if either (a) this list contains that target namespace or (b) this list contains "". Note that "" matches any specified namespace but does not match a request that does not specify a namespace (see the clusterScope field for that). This list may be empty, but only if clusterScope is true.
FlowSchemaStatus
FlowSchemaStatus represents the current state of a FlowSchema.
conditions ([]FlowSchemaCondition)
Map: unique values on key type will be kept during a merge
conditions is a list of the current states of FlowSchema.
FlowSchemaCondition describes conditions for a FlowSchema.
conditions.lastTransitionTime (Time)
lastTransitionTime is the last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is a human-readable message indicating details about last transition.
conditions.reason (string)
reason is a unique, one-word, CamelCase reason for the condition's last transition.
conditions.status (string)
status is the status of the condition. Can be True, False, Unknown. Required.
PriorityLevelConfigurationSpec specifies the configuration of a priority level.
type (string), required
type indicates whether this priority level is subject to limitation on request execution. A value of "Exempt" means that requests of this priority level are not subject to a limit (and thus are never queued) and do not detract from the capacity made available to other priority levels. A value of "Limited" means that (a) requests of this priority level are subject to limits and (b) some of the server's limited capacity is made available exclusively to this priority level. Required.
limited (LimitedPriorityLevelConfiguration)
limited specifies how requests are handled for a Limited priority level. This field must be non-empty if and only if type is "Limited".
*LimitedPriorityLevelConfiguration specifies how to handle requests that are subject to limits. It addresses two issues:
How are requests for this priority level limited?
What should be done with requests that exceed the limit?*
limited.assuredConcurrencyShares (int32)
assuredConcurrencyShares (ACS) configures the execution limit, which is a limit on the number of requests of this priority level that may be exeucting at a given time. ACS must be a positive number. The server's concurrency limit (SCL) is divided among the concurrency-controlled priority levels in proportion to their assured concurrency shares. This produces the assured concurrency value (ACV) --- the number of requests that may be executing at a time --- for each such priority level:
bigger numbers of ACS mean more reserved concurrent requests (at the expense of every other PL). This field has a default value of 30.
limited.limitResponse (LimitResponse)
limitResponse indicates what to do with requests that can not be executed right now
LimitResponse defines how to handle requests that can not be executed right now.
limited.limitResponse.type (string), required
type is "Queue" or "Reject". "Queue" means that requests that can not be executed upon arrival are held in a queue until they can be executed or a queuing limit is reached. "Reject" means that requests that can not be executed upon arrival are rejected. Required.
queuing holds the configuration parameters for queuing. This field may be non-empty only if type is "Queue".
QueuingConfiguration holds the configuration parameters for queuing
limited.limitResponse.queuing.handSize (int32)
handSize is a small positive number that configures the shuffle sharding of requests into queues. When enqueuing a request at this priority level the request's flow identifier (a string pair) is hashed and the hash value is used to shuffle the list of queues and deal a hand of the size specified here. The request is put into one of the shortest queues in that hand. handSize must be no larger than queues, and should be significantly smaller (so that a few heavy flows do not saturate most of the queues). See the user-facing documentation for more extensive guidance on setting this field. This field has a default value of 8.
queueLengthLimit is the maximum number of requests allowed to be waiting in a given queue of this priority level at a time; excess requests are rejected. This value must be positive. If not specified, it will be defaulted to 50.
limited.limitResponse.queuing.queues (int32)
queues is the number of queues for this priority level. The queues exist independently at each apiserver. The value must be positive. Setting it to 1 effectively precludes shufflesharding and thus makes the distinguisher method of associated flow schemas irrelevant. This field has a default value of 64.
PriorityLevelConfigurationStatus
PriorityLevelConfigurationStatus represents the current state of a "request-priority".
Map: unique values on key type will be kept during a merge
conditions is the current state of "request-priority".
PriorityLevelConfigurationCondition defines the condition of priority level.
conditions.lastTransitionTime (Time)
lastTransitionTime is the last time the condition transitioned from one status to another.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
conditions.message (string)
message is a human-readable message indicating details about last transition.
conditions.reason (string)
reason is a unique, one-word, CamelCase reason for the condition's last transition.
conditions.status (string)
status is the status of the condition. Can be True, False, Unknown. Required.
conditions.type (string)
type is the type of the condition. Required.
PriorityLevelConfigurationList
PriorityLevelConfigurationList is a list of PriorityLevelConfiguration objects.
Binding ties one object to another; for example, a pod is bound to a node by a scheduler.
apiVersion: v1
import "k8s.io/api/core/v1"
Binding
Binding ties one object to another; for example, a pod is bound to a node by a scheduler. Deprecated in 1.7, please use the bindings subresource of pods instead.
ClusterCIDR represents a single configuration for per-Node Pod CIDR allocations when the MultiCIDRRangeAllocator is enabled (see the config for kube-controller-manager).
apiVersion: networking.k8s.io/v1alpha1
import "k8s.io/api/networking/v1alpha1"
ClusterCIDR
ClusterCIDR represents a single configuration for per-Node Pod CIDR allocations when the MultiCIDRRangeAllocator is enabled (see the config for kube-controller-manager). A cluster may have any number of ClusterCIDR resources, all of which will be considered when allocating a CIDR for a Node. A ClusterCIDR is eligible to be used for a given Node when the node selector matches the node in question and has free CIDRs to allocate. In case of multiple matching ClusterCIDR resources, the allocator will attempt to break ties using internal heuristics, but any ClusterCIDR whose node selector matches the Node may be used.
ClusterCIDRSpec defines the desired state of ClusterCIDR.
perNodeHostBits (int32), required
PerNodeHostBits defines the number of host bits to be configured per node. A subnet mask determines how much of the address is used for network bits and host bits. For example an IPv4 address of 192.168.0.0/24, splits the address into 24 bits for the network portion and 8 bits for the host portion. To allocate 256 IPs, set this field to 8 (a /24 mask for IPv4 or a /120 for IPv6). Minimum value is 4 (16 IPs). This field is immutable.
ipv4 (string)
IPv4 defines an IPv4 IP block in CIDR notation(e.g. "10.0.0.0/8"). At least one of IPv4 and IPv6 must be specified. This field is immutable.
ipv6 (string)
IPv6 defines an IPv6 IP block in CIDR notation(e.g. "fd12:3456:789a:1::/64"). At least one of IPv4 and IPv6 must be specified. This field is immutable.
nodeSelector (NodeSelector)
NodeSelector defines which nodes the config is applicable to. An empty or nil NodeSelector selects all nodes. This field is immutable.
A node selector represents the union of the results of one or more label queries over a set of nodes; that is, it represents the OR of the selectors represented by the node selector terms.
Required. A list of node selector terms. The terms are ORed.
A null or empty node selector term matches no objects. The requirements of them are ANDed. The TopologySelectorTerm type implements a subset of the NodeSelectorTerm.
When present, indicates that modifications should not be persisted. An invalid or unrecognized dryRun directive will result in an error response and no further processing of the request. Valid values are: - All: all dry run stages will be processed
gracePeriodSeconds (int64)
The duration in seconds before the object should be deleted. Value must be non-negative integer. The value zero indicates delete immediately. If this value is nil, the default grace period for the specified type will be used. Defaults to a per object value if not specified. zero means delete immediately.
Deprecated: please use the PropagationPolicy, this field will be deprecated in 1.7. Should the dependent objects be orphaned. If true/false, the "orphan" finalizer will be added to/removed from the object's finalizers list. Either this field or PropagationPolicy may be set, but not both.
preconditions (Preconditions)
Must be fulfilled before a deletion is carried out. If not possible, a 409 Conflict status will be returned.
Preconditions must be fulfilled before an operation (update, delete, etc.) is carried out.
preconditions.resourceVersion (string)
Specifies the target ResourceVersion
preconditions.uid (string)
Specifies the target UID.
propagationPolicy (string)
Whether and how garbage collection will be performed. Either this field or OrphanDependents may be set, but not both. The default policy is decided by the existing finalizer set in the metadata.finalizers and the resource-specific default policy. Acceptable values are: 'Orphan' - orphan the dependents; 'Background' - allow the garbage collector to delete the dependents in the background; 'Foreground' - a cascading policy that deletes all dependents in the foreground.
6.5.9.2 - LabelSelector
A label selector is a label query over a set of resources.
import "k8s.io/apimachinery/pkg/apis/meta/v1"
A label selector is a label query over a set of resources. The result of matchLabels and matchExpressions are ANDed. An empty label selector matches all objects. A null label selector matches no objects.
matchExpressions ([]LabelSelectorRequirement)
matchExpressions is a list of label selector requirements. The requirements are ANDed.
A label selector requirement is a selector that contains values, a key, and an operator that relates the key and values.
matchExpressions.key (string), required
Patch strategy: merge on key key
key is the label key that the selector applies to.
matchExpressions.operator (string), required
operator represents a key's relationship to a set of values. Valid operators are In, NotIn, Exists and DoesNotExist.
matchExpressions.values ([]string)
values is an array of string values. If the operator is In or NotIn, the values array must be non-empty. If the operator is Exists or DoesNotExist, the values array must be empty. This array is replaced during a strategic merge patch.
matchLabels (map[string]string)
matchLabels is a map of {key,value} pairs. A single {key,value} in the matchLabels map is equivalent to an element of matchExpressions, whose key field is "key", the operator is "In", and the values array contains only "value". The requirements are ANDed.
6.5.9.3 - ListMeta
ListMeta describes metadata that synthetic resources must have, including lists and various status objects.
import "k8s.io/apimachinery/pkg/apis/meta/v1"
ListMeta describes metadata that synthetic resources must have, including lists and various status objects. A resource may have only one of {ObjectMeta, ListMeta}.
continue (string)
continue may be set if the user set a limit on the number of items returned, and indicates that the server has more data available. The value is opaque and may be used to issue another request to the endpoint that served this list to retrieve the next set of available objects. Continuing a consistent list may not be possible if the server configuration has changed or more than a few minutes have passed. The resourceVersion field returned when using this continue value will be identical to the value in the first response, unless you have received this token from an error message.
remainingItemCount (int64)
remainingItemCount is the number of subsequent items in the list which are not included in this list response. If the list request contained label or field selectors, then the number of remaining items is unknown and the field will be left unset and omitted during serialization. If the list is complete (either because it is not chunking or because this is the last chunk), then there are no more remaining items and this field will be left unset and omitted during serialization. Servers older than v1.15 do not set this field. The intended use of the remainingItemCount is estimating the size of a collection. Clients should not rely on the remainingItemCount to be set or to be exact.
A node selector requirement is a selector that contains values, a key, and an operator that relates the key and values.
import "k8s.io/api/core/v1"
A node selector requirement is a selector that contains values, a key, and an operator that relates the key and values.
key (string), required
The label key that the selector applies to.
operator (string), required
Represents a key's relationship to a set of values. Valid operators are In, NotIn, Exists, DoesNotExist. Gt, and Lt.
values ([]string)
An array of string values. If the operator is In or NotIn, the values array must be non-empty. If the operator is Exists or DoesNotExist, the values array must be empty. If the operator is Gt or Lt, the values array must have a single element, which will be interpreted as an integer. This array is replaced during a strategic merge patch.
6.5.9.6 - ObjectFieldSelector
ObjectFieldSelector selects an APIVersioned field of an object.
import "k8s.io/api/core/v1"
ObjectFieldSelector selects an APIVersioned field of an object.
fieldPath (string), required
Path of the field to select in the specified API version.
apiVersion (string)
Version of the schema the FieldPath is written in terms of, defaults to "v1".
6.5.9.7 - ObjectMeta
ObjectMeta is metadata that all persisted resources must have, which includes all objects users must create.
import "k8s.io/apimachinery/pkg/apis/meta/v1"
ObjectMeta is metadata that all persisted resources must have, which includes all objects users must create.
name (string)
Name must be unique within a namespace. Is required when creating resources, although some resources may allow a client to request the generation of an appropriate name automatically. Name is primarily intended for creation idempotence and configuration definition. Cannot be updated. More info: http://kubernetes.io/docs/user-guide/identifiers#names
generateName (string)
GenerateName is an optional prefix, used by the server, to generate a unique name ONLY IF the Name field has not been provided. If this field is used, the name returned to the client will be different than the name passed. This value will also be combined with a unique suffix. The provided value has the same validation rules as the Name field, and may be truncated by the length of the suffix required to make the value unique on the server.
If this field is specified and the generated name exists, the server will return a 409.
Namespace defines the space within which each name must be unique. An empty namespace is equivalent to the "default" namespace, but "default" is the canonical representation. Not all objects are required to be scoped to a namespace - the value of this field for those objects will be empty.
Map of string keys and values that can be used to organize and categorize (scope and select) objects. May match selectors of replication controllers and services. More info: http://kubernetes.io/docs/user-guide/labels
annotations (map[string]string)
Annotations is an unstructured key value map stored with a resource that may be set by external tools to store and retrieve arbitrary metadata. They are not queryable and should be preserved when modifying objects. More info: http://kubernetes.io/docs/user-guide/annotations
System
finalizers ([]string)
Must be empty before the object is deleted from the registry. Each entry is an identifier for the responsible component that will remove the entry from the list. If the deletionTimestamp of the object is non-nil, entries in this list can only be removed. Finalizers may be processed and removed in any order. Order is NOT enforced because it introduces significant risk of stuck finalizers. finalizers is a shared field, any actor with permission can reorder it. If the finalizer list is processed in order, then this can lead to a situation in which the component responsible for the first finalizer in the list is waiting for a signal (field value, external system, or other) produced by a component responsible for a finalizer later in the list, resulting in a deadlock. Without enforced ordering finalizers are free to order amongst themselves and are not vulnerable to ordering changes in the list.
managedFields ([]ManagedFieldsEntry)
ManagedFields maps workflow-id and version to the set of fields that are managed by that workflow. This is mostly for internal housekeeping, and users typically shouldn't need to set or understand this field. A workflow can be the user's name, a controller's name, or the name of a specific apply path like "ci-cd". The set of fields is always in the version that the workflow used when modifying the object.
ManagedFieldsEntry is a workflow-id, a FieldSet and the group version of the resource that the fieldset applies to.
managedFields.apiVersion (string)
APIVersion defines the version of this resource that this field set applies to. The format is "group/version" just like the top-level APIVersion field. It is necessary to track the version of a field set because it cannot be automatically converted.
managedFields.fieldsType (string)
FieldsType is the discriminator for the different fields format and version. There is currently only one possible value: "FieldsV1"
managedFields.fieldsV1 (FieldsV1)
FieldsV1 holds the first JSON version format as described in the "FieldsV1" type.
*FieldsV1 stores a set of fields in a data structure like a Trie, in JSON format.
Each key is either a '.' representing the field itself, and will always map to an empty set, or a string representing a sub-field or item. The string will follow one of these four formats: 'f:', where is the name of a field in a struct, or key in a map 'v:', where is the exact json formatted value of a list item 'i:', where is position of a item in a list 'k:', where is a map of a list item's key fields to their unique values If a key maps to an empty Fields value, the field that key represents is part of the set.
The exact format is defined in sigs.k8s.io/structured-merge-diff*
managedFields.manager (string)
Manager is an identifier of the workflow managing these fields.
managedFields.operation (string)
Operation is the type of operation which lead to this ManagedFieldsEntry being created. The only valid values for this field are 'Apply' and 'Update'.
managedFields.subresource (string)
Subresource is the name of the subresource used to update that object, or empty string if the object was updated through the main resource. The value of this field is used to distinguish between managers, even if they share the same name. For example, a status update will be distinct from a regular update using the same manager name. Note that the APIVersion field is not related to the Subresource field and it always corresponds to the version of the main resource.
managedFields.time (Time)
Time is the timestamp of when the ManagedFields entry was added. The timestamp will also be updated if a field is added, the manager changes any of the owned fields value or removes a field. The timestamp does not update when a field is removed from the entry because another manager took it over.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
ownerReferences ([]OwnerReference)
Patch strategy: merge on key uid
List of objects depended by this object. If ALL objects in the list have been deleted, this object will be garbage collected. If this object is managed by a controller, then an entry in this list will point to this controller, with the controller field set to true. There cannot be more than one managing controller.
OwnerReference contains enough information to let you identify an owning object. An owning object must be in the same namespace as the dependent, or be cluster-scoped, so there is no namespace field.
If true, AND if the owner has the "foregroundDeletion" finalizer, then the owner cannot be deleted from the key-value store until this reference is removed. See https://kubernetes.io/docs/concepts/architecture/garbage-collection/#foreground-deletion for how the garbage collector interacts with this field and enforces the foreground deletion. Defaults to false. To set this field, a user needs "delete" permission of the owner, otherwise 422 (Unprocessable Entity) will be returned.
ownerReferences.controller (boolean)
If true, this reference points to the managing controller.
Read-only
creationTimestamp (Time)
CreationTimestamp is a timestamp representing the server time when this object was created. It is not guaranteed to be set in happens-before order across separate operations. Clients may not set this value. It is represented in RFC3339 form and is in UTC.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
deletionGracePeriodSeconds (int64)
Number of seconds allowed for this object to gracefully terminate before it will be removed from the system. Only set when deletionTimestamp is also set. May only be shortened. Read-only.
deletionTimestamp (Time)
DeletionTimestamp is RFC 3339 date and time at which this resource will be deleted. This field is set by the server when a graceful deletion is requested by the user, and is not directly settable by a client. The resource is expected to be deleted (no longer visible from resource lists, and not reachable by name) after the time in this field, once the finalizers list is empty. As long as the finalizers list contains items, deletion is blocked. Once the deletionTimestamp is set, this value may not be unset or be set further into the future, although it may be shortened or the resource may be deleted prior to this time. For example, a user may request that a pod is deleted in 30 seconds. The Kubelet will react by sending a graceful termination signal to the containers in the pod. After that 30 seconds, the Kubelet will send a hard termination signal (SIGKILL) to the container and after cleanup, remove the pod from the API. In the presence of network partitions, this object may still exist after this timestamp, until an administrator or automated process can determine the resource is fully terminated. If not set, graceful deletion of the object has not been requested.
Time is a wrapper around time.Time which supports correct marshaling to YAML and JSON. Wrappers are provided for many of the factory methods that the time package offers.
generation (int64)
A sequence number representing a specific generation of the desired state. Populated by the system. Read-only.
resourceVersion (string)
An opaque value that represents the internal version of this object that can be used by clients to determine when objects have changed. May be used for optimistic concurrency, change detection, and the watch operation on a resource or set of resources. Clients must treat these values as opaque and passed unmodified back to the server. They may only be valid for a particular resource or set of resources.
Deprecated: selfLink is a legacy read-only field that is no longer populated by the system.
uid (string)
UID is the unique in time and space value for this object. It is typically generated by the server on successful creation of a resource and is not allowed to change on PUT operations.
ObjectReference contains enough information to let you inspect or modify the referred object.
import "k8s.io/api/core/v1"
ObjectReference contains enough information to let you inspect or modify the referred object.
apiVersion (string)
API version of the referent.
fieldPath (string)
If referring to a piece of an object instead of an entire object, this string should contain a valid JSON/Go field access statement, such as desiredState.manifest.containers[2]. For example, if the object reference is to a container within a pod, this would take on a value like: "spec.containers{name}" (where "name" refers to the name of the container that triggered the event) or if no container name is specified "spec.containers[2]" (container with index 2 in this pod). This syntax is chosen only to have some well-defined way of referencing a part of an object.
Patch is provided to give a concrete name and type to the Kubernetes PATCH request body.
import "k8s.io/apimachinery/pkg/apis/meta/v1"
Patch is provided to give a concrete name and type to the Kubernetes PATCH request body.
6.5.9.10 - Quantity
Quantity is a fixed-point representation of a number.
import "k8s.io/apimachinery/pkg/api/resource"
Quantity is a fixed-point representation of a number. It provides convenient marshaling/unmarshaling in JSON and YAML, in addition to String() and AsInt64() accessors.
The serialization format is:
(Note that \<suffix> may be empty, from the "" case in \<decimalSI>.)
\<digit> ::= 0 | 1 | ... | 9 \<digits> ::= \<digit> | \<digit>\<digits> \<number> ::= \<digits> | \<digits>.\<digits> | \<digits>. | .\<digits> \<sign> ::= "+" | "-" \<signedNumber> ::= \<number> | \<sign>\<number> \<suffix> ::= \<binarySI> | \<decimalExponent> | \<decimalSI> \<binarySI> ::= Ki | Mi | Gi | Ti | Pi | Ei
(International System of units; See: http://physics.nist.gov/cuu/Units/binary.html)
\<decimalSI> ::= m | "" | k | M | G | T | P | E
(Note that 1024 = 1Ki but 1000 = 1k; I didn't choose the capitalization.)
\<decimalExponent> ::= "e" \<signedNumber> | "E" \<signedNumber> ```
No matter which of the three exponent forms is used, no quantity may represent a number greater than 2^63-1 in magnitude, nor may it have more than 3 decimal places. Numbers larger or more precise will be capped or rounded up. (E.g.: 0.1m will rounded up to 1m.) This may be extended in the future if we require larger or smaller quantities.
When a Quantity is parsed from a string, it will remember the type of suffix it had, and will use the same type again when it is serialized.
Before serializing, Quantity will be put in "canonical form". This means that Exponent/suffix will be adjusted up or down (with a corresponding increase or decrease in Mantissa) such that:
- No precision is lost - No fractional digits will be emitted - The exponent (or suffix) is as large as possible.
The sign will be omitted unless the number is negative.
Examples:
- 1.5 will be serialized as "1500m" - 1.5Gi will be serialized as "1536Mi"
Note that the quantity will NEVER be internally represented by a floating point number. That is the whole point of this exercise.
Non-canonical values will still parse as long as they are well formed, but will be re-emitted in their canonical form. (So always use canonical form, or don't diff.)
This format is intended to make it difficult to use these numbers without writing some sort of special handling code in the hopes that that will cause implementors to also use a fixed point implementation.
<hr>
6.5.9.11 - ResourceFieldSelector
ResourceFieldSelector represents container resources (cpu, memory) and their output format.
import "k8s.io/api/core/v1"
ResourceFieldSelector represents container resources (cpu, memory) and their output format
resource (string), required
Required: resource to select
containerName (string)
Container name: required for volumes, optional for env vars
Suggested HTTP return code for this status, 0 if not set.
details (StatusDetails)
Extended data associated with the reason. Each reason may define its own extended details. This field is optional and the data returned is not guaranteed to conform to any schema except that defined by the reason type.
StatusDetails is a set of additional properties that MAY be set by the server to provide additional information about a response. The Reason field of a Status object defines what attributes will be set. Clients must ignore fields that do not match the defined type of each attribute, and should assume that any attribute may be empty, invalid, or under defined.
details.causes ([]StatusCause)
The Causes array includes more details associated with the StatusReason failure. Not all StatusReasons may provide detailed causes.
StatusCause provides more information about an api.Status failure, including cases when multiple errors are encountered.
details.causes.field (string)
The field of the resource that has caused this error, as named by its JSON serialization. May include dot and postfix notation for nested attributes. Arrays are zero-indexed. Fields may appear more than once in an array of causes due to fields having multiple errors. Optional.
Examples:
"name" - the field "name" on the current resource
"items[0].name" - the field "name" on the first array entry in "items"
details.causes.message (string)
A human-readable description of the cause of the error. This field may be presented as-is to a reader.
details.causes.reason (string)
A machine-readable description of the cause of the error. If this value is empty there is no information available.
details.group (string)
The group attribute of the resource associated with the status StatusReason.
The name attribute of the resource associated with the status StatusReason (when there is a single name which can be described).
details.retryAfterSeconds (int32)
If specified, the time in seconds before the operation should be retried. Some errors may indicate the client must take an alternate action - for those errors this field may indicate how long to wait before taking the alternate action.
A machine-readable description of why this operation is in the "Failure" status. If this value is empty there is no information available. A Reason clarifies an HTTP status code but does not override it.
TypedLocalObjectReference contains enough information to let you locate the typed referenced object inside the same namespace.
import "k8s.io/api/core/v1"
TypedLocalObjectReference contains enough information to let you locate the typed referenced object inside the same namespace.
kind (string), required
Kind is the type of resource being referenced
name (string), required
Name is the name of resource being referenced
apiGroup (string)
APIGroup is the group for the resource being referenced. If APIGroup is not specified, the specified Kind must be in the core API group. For any other third-party types, APIGroup is required.
6.5.10 - Common Parameters
allowWatchBookmarks
allowWatchBookmarks requests watch events with type "BOOKMARK". Servers that do not implement bookmarks may ignore this flag and bookmarks are sent at the server's discretion. Clients should not assume bookmarks are returned at any specific interval, nor may they assume the server will send any BOOKMARK event during a session. If this is not a watch, this field is ignored.
continue
The continue option should be set when retrieving more results from the server. Since this value is server defined, clients may only use the continue value from a previous query result with identical query parameters (except for the value of continue) and the server may reject a continue value it does not recognize. If the specified continue value is no longer valid whether due to expiration (generally five to fifteen minutes) or a configuration change on the server, the server will respond with a 410 ResourceExpired error together with a continue token. If the client needs a consistent list, it must restart their list without the continue field. Otherwise, the client may send another list request with the token received with the 410 error, the server will respond with a list starting from the next key, but from the latest snapshot, which is inconsistent from the previous list results - objects that are created, modified, or deleted after the first list request will be included in the response, as long as their keys are after the "next key".
This field is not supported when watch is true. Clients may start a watch from the last resourceVersion value returned by the server and not miss any modifications.
dryRun
When present, indicates that modifications should not be persisted. An invalid or unrecognized dryRun directive will result in an error response and no further processing of the request. Valid values are: - All: all dry run stages will be processed
fieldManager
fieldManager is a name associated with the actor or entity that is making these changes. The value must be less than or 128 characters long, and only contain printable characters, as defined by https://golang.org/pkg/unicode/#IsPrint.
fieldSelector
A selector to restrict the list of returned objects by their fields. Defaults to everything.
fieldValidation
fieldValidation instructs the server on how to handle objects in the request (POST/PUT/PATCH) containing unknown or duplicate fields, provided that the ServerSideFieldValidation feature gate is also enabled. Valid values are: - Ignore: This will ignore any unknown fields that are silently dropped from the object, and will ignore all but the last duplicate field that the decoder encounters. This is the default behavior prior to v1.23 and is the default behavior when the ServerSideFieldValidation feature gate is disabled. - Warn: This will send a warning via the standard warning response header for each unknown field that is dropped from the object, and for each duplicate field that is encountered. The request will still succeed if there are no other errors, and will only persist the last of any duplicate fields. This is the default when the ServerSideFieldValidation feature gate is enabled. - Strict: This will fail the request with a BadRequest error if any unknown fields would be dropped from the object, or if any duplicate fields are present. The error returned from the server will contain all unknown and duplicate fields encountered.
force
Force is going to "force" Apply requests. It means user will re-acquire conflicting fields owned by other people. Force flag must be unset for non-apply patch requests.
gracePeriodSeconds
The duration in seconds before the object should be deleted. Value must be non-negative integer. The value zero indicates delete immediately. If this value is nil, the default grace period for the specified type will be used. Defaults to a per object value if not specified. zero means delete immediately.
labelSelector
A selector to restrict the list of returned objects by their labels. Defaults to everything.
limit
limit is a maximum number of responses to return for a list call. If more items exist, the server will set the continue field on the list metadata to a value that can be used with the same initial query to retrieve the next set of results. Setting a limit may return fewer than the requested amount of items (up to zero items) in the event all requested objects are filtered out and clients should only use the presence of the continue field to determine whether more results are available. Servers may choose not to support the limit argument and will return all of the available results. If limit is specified and the continue field is empty, clients may assume that no more results are available. This field is not supported if watch is true.
The server guarantees that the objects returned when using continue will be identical to issuing a single list call without a limit - that is, no objects created, modified, or deleted after the first request is issued will be included in any subsequent continued requests. This is sometimes referred to as a consistent snapshot, and ensures that a client that is using limit to receive smaller chunks of a very large result can ensure they see all possible objects. If objects are updated during a chunked list the version of the object that was present at the time the first list result was calculated is returned.
namespace
object name and auth scope, such as for teams and projects
pretty
If 'true', then the output is pretty printed.
propagationPolicy
Whether and how garbage collection will be performed. Either this field or OrphanDependents may be set, but not both. The default policy is decided by the existing finalizer set in the metadata.finalizers and the resource-specific default policy. Acceptable values are: 'Orphan' - orphan the dependents; 'Background' - allow the garbage collector to delete the dependents in the background; 'Foreground' - a cascading policy that deletes all dependents in the foreground.
We're extremely grateful for security researchers and users that report vulnerabilities to the Kubernetes Open Source Community. All reports are thoroughly investigated by a set of community volunteers.
To make a report, submit your vulnerability to the Kubernetes bug bounty program. This allows triage and handling of the vulnerability with standardized response times.
You may encrypt your email to this list using the GPG keys of the Security Response Committee members. Encryption using GPG is NOT required to make a disclosure.
When Should I Report a Vulnerability?
You think you discovered a potential security vulnerability in Kubernetes
You are unsure how a vulnerability affects Kubernetes
You think you discovered a vulnerability in another project that Kubernetes depends on
For projects with their own vulnerability reporting and disclosure process, please report it directly there
When Should I NOT Report a Vulnerability?
You need help tuning Kubernetes components for security
You need help applying security related updates
Your issue is not security related
Security Vulnerability Response
Each report is acknowledged and analyzed by Security Response Committee members within 3 working days. This will set off the Security Release Process.
Any vulnerability information shared with Security Response Committee stays within Kubernetes project and will not be disseminated to other projects unless it is necessary to get the issue fixed.
As the security issue moves from triage, to identified fix, to release planning we will keep the reporter updated.
Public Disclosure Timing
A public disclosure date is negotiated by the Kubernetes Security Response Committee and the bug submitter. We prefer to fully disclose the bug as soon as possible once a user mitigation is available. It is reasonable to delay disclosure when the bug or the fix is not yet fully understood, the solution is not well-tested, or for vendor coordination. The timeframe for disclosure is from immediate (especially if it's already publicly known) to a few weeks. For a vulnerability with a straightforward mitigation, we expect report date to disclosure date to be on the order of 7 days. The Kubernetes Security Response Committee holds the final say when setting a disclosure date.
The Kubernetes project publishes a programmatically accessible
JSON Feed of
published security issues. You can access it by executing the following command:
This feed is auto-refreshing with a noticeable but small lag (minutes to hours)
from the time a CVE is announced to the time it is accessible in this feed.
The source of truth of this feed is a set of GitHub Issues, filtered by a controlled and
restricted label official-cve-feed. The raw data is stored in a Google Cloud
Bucket which is writable only by a small number of trusted members of the
Community.
6.7 - Node Reference Information
6.7.1 - Kubelet Checkpoint API
FEATURE STATE:Kubernetes v1.25 [alpha]
Checkpointing a container is the functionality to create a stateful copy of a
running container. Once you have a stateful copy of a container, you could
move it to a different computer for debugging or similar purposes.
If you move the checkpointed container data to a computer that's able to restore
it, that restored container continues to run at exactly the same
point it was checkpointed. You can also inspect the saved data, provided that you
have suitable tools for doing so.
Creating a checkpoint of a container might have security implications. Typically
a checkpoint contains all memory pages of all processes in the checkpointed
container. This means that everything that used to be in memory is now available
on the local disk. This includes all private data and possibly keys used for
encryption. The underlying CRI implementations (the container runtime on that node)
should create the checkpoint archive to be only accessible by the root user. It
is still important to remember if the checkpoint archive is transferred to another
system all memory pages will be readable by the owner of the checkpoint archive.
Operations
post checkpoint the specified container
Tell the kubelet to checkpoint a specific container from the specified Pod.
The kubelet will request a checkpoint from the underlying
CRI implementation. In the checkpoint
request the kubelet will specify the name of the checkpoint archive as
checkpoint-<podFullName>-<containerName>-<timestamp>.tar and also request to
store the checkpoint archive in the checkpoints directory below its root
directory (as defined by --root-dir). This defaults to
/var/lib/kubelet/checkpoints.
The checkpoint archive is in tar format, and could be listed using an implementation of
tar. The contents of the
archive depend on the underlying CRI implementation (the container runtime on that node).
Timeout in seconds to wait until the checkpoint creation is finished.
If zero or no timeout is specfied the default CRI timeout value will be used. Checkpoint
creation time depends directly on the used memory of the container.
The more memory a container uses the more time is required to create
the corresponding checkpoint.
Response
200: OK
401: Unauthorized
404: Not Found (if the CheckpointContainer feature gate is disabled)
404: Not Found (if the specified namespace, pod or container cannot be found)
500: Internal Server Error (if the CRI implementation encounter an error during checkpointing (see error message for further details))
500: Internal Server Error (if the CRI implementation does not implement the checkpoint CRI API (see error message for further details))
6.7.2 - Articles on dockershim Removal and on Using CRI-compatible Runtimes
This is a list of articles and other pages that are either
about the Kubernetes' deprecation and removal of dockershim,
or about using CRI-compatible container runtimes, in connection
with that removal.
When running Kubernetes in an environment with strict network boundaries, such
as on-premises datacenter with physical network firewalls or Virtual
Networks in Public Cloud, it is useful to be aware of the ports and protocols
used by Kubernetes components.
Control plane
Protocol
Direction
Port Range
Purpose
Used By
TCP
Inbound
6443
Kubernetes API server
All
TCP
Inbound
2379-2380
etcd server client API
kube-apiserver, etcd
TCP
Inbound
10250
Kubelet API
Self, Control plane
TCP
Inbound
10259
kube-scheduler
Self
TCP
Inbound
10257
kube-controller-manager
Self
Although etcd ports are included in control plane section, you can also host your own
etcd cluster externally or on custom ports.
All default port numbers can be overridden. When custom ports are used those
ports need to be open instead of defaults mentioned here.
One common example is API server port that is sometimes switched
to 443. Alternatively, the default port is kept as is and API server is put
behind a load balancer that listens on 443 and routes the requests to API server
on the default port.
6.9 - Setup tools
6.9.1 - Kubeadm
Kubeadm is a tool built to provide kubeadm init and kubeadm join as best-practice "fast paths" for creating Kubernetes clusters.
kubeadm performs the actions necessary to get a minimum viable cluster up and running. By design, it cares only about bootstrapping, not about provisioning machines. Likewise, installing various nice-to-have addons, like the Kubernetes Dashboard, monitoring solutions, and cloud-specific addons, is not in scope.
Instead, we expect higher-level and more tailored tooling to be built on top of kubeadm, and ideally, using kubeadm as the basis of all deployments will make it easier to create conformant clusters.
kubeadm alpha to preview a set of features made available for gathering feedback from the community
6.9.1.1 - Kubeadm Generated
6.9.1.1.1 -
kubeadm: easily bootstrap a secure Kubernetes cluster
Synopsis
┌──────────────────────────────────────────────────────────┐
│ KUBEADM │
│ Easily bootstrap a secure Kubernetes cluster │
│ │
│ Please give us feedback at: │
│ https://github.com/kubernetes/kubeadm/issues │
└──────────────────────────────────────────────────────────┘
Example usage:
Create a two-machine cluster with one control-plane node
(which controls the cluster), and one worker node
(where your workloads, like Pods and Deployments run).
┌──────────────────────────────────────────────────────────┐
│ On the first machine: │
├──────────────────────────────────────────────────────────┤
│ control-plane# kubeadm init │
└──────────────────────────────────────────────────────────┘
┌──────────────────────────────────────────────────────────┐
│ On the second machine: │
├──────────────────────────────────────────────────────────┤
│ worker# kubeadm join <arguments-returned-from-init> │
└──────────────────────────────────────────────────────────┘
You can then repeat the second step on as many other machines as you like.
Options
-h, --help
help for kubeadm
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.2 -
Commands related to handling kubernetes certificates
Synopsis
Commands related to handling kubernetes certificates
Options
-h, --help
help for certs
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.3 -
Generate certificate keys
Synopsis
This command will print out a secure randomly-generated certificate key that can be used with
the "init" command.
You can also use "kubeadm init --upload-certs" without specifying a certificate key and it will
generate and print one for you.
kubeadm certs certificate-key [flags]
Options
-h, --help
help for certificate-key
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.4 -
Check certificates expiration for a Kubernetes cluster
Synopsis
Checks expiration for the certificates in the local PKI managed by kubeadm.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.5 -
Generate keys and certificate signing requests
Synopsis
Generates keys and certificate signing requests (CSRs) for all the certificates required to run the control plane. This command also generates partial kubeconfig files with private key data in the "users > user > client-key-data" field, and for each kubeconfig file an accompanying ".csr" file is created.
This command is designed for use in Kubeadm External CA Mode. It generates CSRs which you can then submit to your external certificate authority for signing.
The PEM encoded signed certificates should then be saved alongside the key files, using ".crt" as the file extension, or in the case of kubeconfig files, the PEM encoded signed certificate should be base64 encoded and added to the kubeconfig file in the "users > user > client-certificate-data" field.
kubeadm certs generate-csr [flags]
Examples
# The following command will generate keys and CSRs for all control-plane certificates and kubeconfig files:
kubeadm certs generate-csr --kubeconfig-dir /tmp/etc-k8s --cert-dir /tmp/etc-k8s/pki
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.6 -
Renew certificates for a Kubernetes cluster
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm certs renew [flags]
Options
-h, --help
help for renew
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.7 -
Renew the certificate embedded in the kubeconfig file for the admin to use and for kubeadm itself
Synopsis
Renew the certificate embedded in the kubeconfig file for the admin to use and for kubeadm itself.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.8 -
Renew all available certificates
Synopsis
Renew all known certificates necessary to run the control plane. Renewals are run unconditionally, regardless of expiration date. Renewals can also be run individually for more control.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.9 -
Renew the certificate the apiserver uses to access etcd
Synopsis
Renew the certificate the apiserver uses to access etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.10 -
Renew the certificate for the API server to connect to kubelet
Synopsis
Renew the certificate for the API server to connect to kubelet.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.11 -
Renew the certificate for serving the Kubernetes API
Synopsis
Renew the certificate for serving the Kubernetes API.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.12 -
Renew the certificate embedded in the kubeconfig file for the controller manager to use
Synopsis
Renew the certificate embedded in the kubeconfig file for the controller manager to use.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.13 -
Renew the certificate for liveness probes to healthcheck etcd
Synopsis
Renew the certificate for liveness probes to healthcheck etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.14 -
Renew the certificate for etcd nodes to communicate with each other
Synopsis
Renew the certificate for etcd nodes to communicate with each other.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.15 -
Renew the certificate for serving etcd
Synopsis
Renew the certificate for serving etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.16 -
Renew the certificate for the front proxy client
Synopsis
Renew the certificate for the front proxy client.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.17 -
Renew the certificate embedded in the kubeconfig file for the scheduler manager to use
Synopsis
Renew the certificate embedded in the kubeconfig file for the scheduler manager to use.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.18 -
Output shell completion code for the specified shell (bash or zsh)
Synopsis
Output shell completion code for the specified shell (bash or zsh).
The shell code must be evaluated to provide interactive
completion of kubeadm commands. This can be done by sourcing it from
the .bash_profile.
Note: this requires the bash-completion framework.
To install it on Mac use homebrew:
$ brew install bash-completion
Once installed, bash_completion must be evaluated. This can be done by adding the
following line to the .bash_profile
$ source $(brew --prefix)/etc/bash_completion
If bash-completion is not installed on Linux, please install the 'bash-completion' package
via your distribution's package manager.
Note for zsh users: [1] zsh completions are only supported in versions of zsh >= 5.2
kubeadm completion SHELL [flags]
Examples
# Install bash completion on a Mac using homebrew
brew install bash-completion
printf "\n# Bash completion support\nsource $(brew --prefix)/etc/bash_completion\n" >> $HOME/.bash_profile
source $HOME/.bash_profile
# Load the kubeadm completion code for bash into the current shell
source <(kubeadm completion bash)
# Write bash completion code to a file and source it from .bash_profile
kubeadm completion bash > ~/.kube/kubeadm_completion.bash.inc
printf "\n# Kubeadm shell completion\nsource '$HOME/.kube/kubeadm_completion.bash.inc'\n" >> $HOME/.bash_profile
source $HOME/.bash_profile
# Load the kubeadm completion code for zsh[1] into the current shell
source <(kubeadm completion zsh)
Options
-h, --help
help for completion
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.19 -
Manage configuration for a kubeadm cluster persisted in a ConfigMap in the cluster
Synopsis
There is a ConfigMap in the kube-system namespace called "kubeadm-config" that kubeadm uses to store internal configuration about the
cluster. kubeadm CLI v1.8.0+ automatically creates this ConfigMap with the config used with 'kubeadm init', but if you
initialized your cluster using kubeadm v1.7.x or lower, you must use the 'config upload' command to create this
ConfigMap. This is required so that 'kubeadm upgrade' can configure your upgraded cluster correctly.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.21 -
Print a list of images kubeadm will use. The configuration file is used in case any images or image repositories are customized
Synopsis
Print a list of images kubeadm will use. The configuration file is used in case any images or image repositories are customized
kubeadm config images list [flags]
Options
--allow-missing-template-keys Default: true
If true, ignore any errors in templates when a field or map key is missing in the template. Only applies to golang and jsonpath output formats.
--config string
Path to a kubeadm configuration file.
-o, --experimental-output string Default: "text"
Output format. One of: text|json|yaml|go-template|go-template-file|template|templatefile|jsonpath|jsonpath-as-json|jsonpath-file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.22 -
Pull images used by kubeadm
Synopsis
Pull images used by kubeadm
kubeadm config images pull [flags]
Options
--config string
Path to a kubeadm configuration file.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.23 -
Read an older version of the kubeadm configuration API types from a file, and output the similar config object for the newer version
Synopsis
This command lets you convert configuration objects of older versions to the latest supported version,
locally in the CLI tool without ever touching anything in the cluster.
In this version of kubeadm, the following API versions are supported:
kubeadm.k8s.io/v1beta3
Further, kubeadm can only write out config of version "kubeadm.k8s.io/v1beta3", but read both types.
So regardless of what version you pass to the --old-config parameter here, the API object will be
read, deserialized, defaulted, converted, validated, and re-serialized when written to stdout or
--new-config if specified.
In other words, the output of this command is what kubeadm actually would read internally if you
submitted this file to "kubeadm init"
kubeadm config migrate [flags]
Options
-h, --help
help for migrate
--new-config string
Path to the resulting equivalent kubeadm config file using the new API version. Optional, if not specified output will be sent to STDOUT.
--old-config string
Path to the kubeadm config file that is using an old API version and should be converted. This flag is mandatory.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.25 -
Print default init configuration, that can be used for 'kubeadm init'
Synopsis
This command prints objects such as the default init configuration that is used for 'kubeadm init'.
Note that sensitive values like the Bootstrap Token fields are replaced with placeholder values like "abcdef.0123456789abcdef" in order to pass validation but
not perform the real computation for creating a token.
kubeadm config print init-defaults [flags]
Options
--component-configs strings
A comma-separated list for component config API objects to print the default values for. Available values: [KubeProxyConfiguration KubeletConfiguration]. If this flag is not set, no component configs will be printed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.26 -
Print default join configuration, that can be used for 'kubeadm join'
Synopsis
This command prints objects such as the default join configuration that is used for 'kubeadm join'.
Note that sensitive values like the Bootstrap Token fields are replaced with placeholder values like "abcdef.0123456789abcdef" in order to pass validation but
not perform the real computation for creating a token.
kubeadm config print join-defaults [flags]
Options
--component-configs strings
A comma-separated list for component config API objects to print the default values for. Available values: [KubeProxyConfiguration KubeletConfiguration]. If this flag is not set, no component configs will be printed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.27 -
Run this command in order to set up the Kubernetes control plane
Synopsis
Run this command in order to set up the Kubernetes control plane
The "init" command executes the following phases:
preflight Run pre-flight checks
certs Certificate generation
/ca Generate the self-signed Kubernetes CA to provision identities for other Kubernetes components
/apiserver Generate the certificate for serving the Kubernetes API
/apiserver-kubelet-client Generate the certificate for the API server to connect to kubelet
/front-proxy-ca Generate the self-signed CA to provision identities for front proxy
/front-proxy-client Generate the certificate for the front proxy client
/etcd-ca Generate the self-signed CA to provision identities for etcd
/etcd-server Generate the certificate for serving etcd
/etcd-peer Generate the certificate for etcd nodes to communicate with each other
/etcd-healthcheck-client Generate the certificate for liveness probes to healthcheck etcd
/apiserver-etcd-client Generate the certificate the apiserver uses to access etcd
/sa Generate a private key for signing service account tokens along with its public key
kubeconfig Generate all kubeconfig files necessary to establish the control plane and the admin kubeconfig file
/admin Generate a kubeconfig file for the admin to use and for kubeadm itself
/kubelet Generate a kubeconfig file for the kubelet to use *only* for cluster bootstrapping purposes
/controller-manager Generate a kubeconfig file for the controller manager to use
/scheduler Generate a kubeconfig file for the scheduler to use
kubelet-start Write kubelet settings and (re)start the kubelet
control-plane Generate all static Pod manifest files necessary to establish the control plane
/apiserver Generates the kube-apiserver static Pod manifest
/controller-manager Generates the kube-controller-manager static Pod manifest
/scheduler Generates the kube-scheduler static Pod manifest
etcd Generate static Pod manifest file for local etcd
/local Generate the static Pod manifest file for a local, single-node local etcd instance
upload-config Upload the kubeadm and kubelet configuration to a ConfigMap
/kubeadm Upload the kubeadm ClusterConfiguration to a ConfigMap
/kubelet Upload the kubelet component config to a ConfigMap
upload-certs Upload certificates to kubeadm-certs
mark-control-plane Mark a node as a control-plane
bootstrap-token Generates bootstrap tokens used to join a node to a cluster
kubelet-finalize Updates settings relevant to the kubelet after TLS bootstrap
/experimental-cert-rotation Enable kubelet client certificate rotation
addon Install required addons for passing conformance tests
/coredns Install the CoreDNS addon to a Kubernetes cluster
/kube-proxy Install the kube-proxy addon to a Kubernetes cluster
kubeadm init [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--apiserver-cert-extra-sans strings
Optional extra Subject Alternative Names (SANs) to use for the API Server serving certificate. Can be both IP addresses and DNS names.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--certificate-key string
Key used to encrypt the control-plane certificates in the kubeadm-certs Secret.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-h, --help
help for init
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Use alternative domain for services, e.g. "myorg.internal".
--skip-certificate-key-print
Don't print the key used to encrypt the control-plane certificates.
--skip-phases strings
List of phases to be skipped
--skip-token-print
Skip printing of the default bootstrap token generated by 'kubeadm init'.
--token string
The token to use for establishing bidirectional trust between nodes and control-plane nodes. The format is [a-z0-9]{6}.[a-z0-9]{16} - e.g. abcdef.0123456789abcdef
--token-ttl duration Default: 24h0m0s
The duration before the token is automatically deleted (e.g. 1s, 2m, 3h). If set to '0', the token will never expire
--upload-certs
Upload control-plane certificates to the kubeadm-certs Secret.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.28 -
Use this command to invoke single phase of the init workflow
Synopsis
Use this command to invoke single phase of the init workflow
Options
-h, --help
help for phase
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.29 -
Install required addons for passing conformance tests
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase addon [flags]
Options
-h, --help
help for addon
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.30 -
Install all the addons
Synopsis
Install all the addons
kubeadm init phase addon all [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Use alternative domain for services, e.g. "myorg.internal".
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.31 -
Install the CoreDNS addon to a Kubernetes cluster
Synopsis
Install the CoreDNS addon components via the API server. Please note that although the DNS server is deployed, it will not be scheduled until CNI is installed.
kubeadm init phase addon coredns [flags]
Options
--config string
Path to a kubeadm configuration file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.33 -
Generates bootstrap tokens used to join a node to a cluster
Synopsis
Bootstrap tokens are used for establishing bidirectional trust between a node joining the cluster and a control-plane node.
This command makes all the configurations required to make bootstrap tokens works and then creates an initial token.
kubeadm init phase bootstrap-token [flags]
Examples
# Make all the bootstrap token configurations and create an initial token, functionally
# equivalent to what generated by kubeadm init.
kubeadm init phase bootstrap-token
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-token-print
Skip printing of the default bootstrap token generated by 'kubeadm init'.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.34 -
Certificate generation
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase certs [flags]
Options
-h, --help
help for certs
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.35 -
Generate all certificates
Synopsis
Generate all certificates
kubeadm init phase certs all [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-cert-extra-sans strings
Optional extra Subject Alternative Names (SANs) to use for the API Server serving certificate. Can be both IP addresses and DNS names.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
-h, --help
help for all
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for apiserver-etcd-client
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.37 -
Generate the certificate for the API server to connect to kubelet
Synopsis
Generate the certificate for the API server to connect to kubelet, and save them into apiserver-kubelet-client.crt and apiserver-kubelet-client.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
Use alternative domain for services, e.g. "myorg.internal".
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.39 -
Generate the self-signed Kubernetes CA to provision identities for other Kubernetes components
Synopsis
Generate the self-signed Kubernetes CA to provision identities for other Kubernetes components, and save them into ca.crt and ca.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs ca [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for ca
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.40 -
Generate the self-signed CA to provision identities for etcd
Synopsis
Generate the self-signed CA to provision identities for etcd, and save them into etcd/ca.crt and etcd/ca.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs etcd-ca [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for etcd-ca
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.41 -
Generate the certificate for liveness probes to healthcheck etcd
Synopsis
Generate the certificate for liveness probes to healthcheck etcd, and save them into etcd/healthcheck-client.crt and etcd/healthcheck-client.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for front-proxy-client
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.46 -
Generate a private key for signing service account tokens along with its public key
Synopsis
Generate the private key for signing service account tokens along with its public key, and save them into sa.key and sa.pub files. If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs sa [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
-h, --help
help for sa
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.47 -
Generate all static Pod manifest files necessary to establish the control plane
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase control-plane [flags]
Options
-h, --help
help for control-plane
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.48 -
Generate all static Pod manifest files
Synopsis
Generate all static Pod manifest files
kubeadm init phase control-plane all [flags]
Examples
# Generates all static Pod manifest files for control plane components,
# functionally equivalent to what is generated by kubeadm init.
kubeadm init phase control-plane all
# Generates all static Pod manifest files using options read from a configuration file.
kubeadm init phase control-plane all --config config.yaml
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
A set of extra flags to pass to the Controller Manager or override default ones in form of <flagname>=<value>
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
A set of extra flags to pass to the API Server or override default ones in form of <flagname>=<value>
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.50 -
Generates the kube-controller-manager static Pod manifest
Synopsis
Generates the kube-controller-manager static Pod manifest
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
A set of extra flags to pass to the Scheduler or override default ones in form of <flagname>=<value>
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.52 -
Generate static Pod manifest file for local etcd
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase etcd [flags]
Options
-h, --help
help for etcd
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.53 -
Generate the static Pod manifest file for a local, single-node local etcd instance
Synopsis
Generate the static Pod manifest file for a local, single-node local etcd instance
kubeadm init phase etcd local [flags]
Examples
# Generates the static Pod manifest file for etcd, functionally
# equivalent to what is generated by kubeadm init.
kubeadm init phase etcd local
# Generates the static Pod manifest file for etcd using options
# read from a configuration file.
kubeadm init phase etcd local --config config.yaml
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
Choose a container registry to pull control plane images from
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.54 -
Generate all kubeconfig files necessary to establish the control plane and the admin kubeconfig file
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase kubeconfig [flags]
Options
-h, --help
help for kubeconfig
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.55 -
Generate a kubeconfig file for the admin to use and for kubeadm itself
Synopsis
Generate the kubeconfig file for the admin and for kubeadm itself, and save it to admin.conf file.
kubeadm init phase kubeconfig admin [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.58 -
Generate a kubeconfig file for the kubelet to use only for cluster bootstrapping purposes
Synopsis
Generate the kubeconfig file for the kubelet to use and save it to kubelet.conf file.
Please note that this should only be used for cluster bootstrapping purposes. After your control plane is up, you should request all kubelet credentials from the CSR API.
kubeadm init phase kubeconfig kubelet [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for experimental-cert-rotation
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.63 -
Write kubelet settings and (re)start the kubelet
Synopsis
Write a file with KubeletConfiguration and an environment file with node specific kubelet settings, and then (re)start kubelet.
kubeadm init phase kubelet-start [flags]
Examples
# Writes a dynamic environment file with kubelet flags from a InitConfiguration file.
kubeadm init phase kubelet-start --config config.yaml
Options
--config string
Path to a kubeadm configuration file.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
-h, --help
help for kubelet-start
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.64 -
Mark a node as a control-plane
Synopsis
Mark a node as a control-plane
kubeadm init phase mark-control-plane [flags]
Examples
# Applies control-plane label and taint to the current node, functionally equivalent to what executed by kubeadm init.
kubeadm init phase mark-control-plane --config config.yaml
# Applies control-plane label and taint to a specific node
kubeadm init phase mark-control-plane --node-name myNode
Options
--config string
Path to a kubeadm configuration file.
-h, --help
help for mark-control-plane
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.65 -
Run pre-flight checks
Synopsis
Run pre-flight checks for kubeadm init.
kubeadm init phase preflight [flags]
Examples
# Run pre-flight checks for kubeadm init using a config file.
kubeadm init phase preflight --config kubeadm-config.yaml
Options
--config string
Path to a kubeadm configuration file.
-h, --help
help for preflight
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.66 -
Upload certificates to kubeadm-certs
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase upload-certs [flags]
Options
--certificate-key string
Key used to encrypt the control-plane certificates in the kubeadm-certs Secret.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-certificate-key-print
Don't print the key used to encrypt the control-plane certificates.
--upload-certs
Upload control-plane certificates to the kubeadm-certs Secret.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.67 -
Upload the kubeadm and kubelet configuration to a ConfigMap
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase upload-config [flags]
Options
-h, --help
help for upload-config
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.69 -
Upload the kubeadm ClusterConfiguration to a ConfigMap
Synopsis
Upload the kubeadm ClusterConfiguration to a ConfigMap called kubeadm-config in the kube-system namespace. This enables correct configuration of system components and a seamless user experience when upgrading.
Alternatively, you can use kubeadm config.
kubeadm init phase upload-config kubeadm [flags]
Examples
# upload the configuration of your cluster
kubeadm init phase upload-config --config=myConfig.yaml
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.70 -
Upload the kubelet component config to a ConfigMap
Synopsis
Upload the kubelet configuration extracted from the kubeadm InitConfiguration object to a kubelet-config ConfigMap in the cluster
kubeadm init phase upload-config kubelet [flags]
Examples
# Upload the kubelet configuration from the kubeadm Config file to a ConfigMap in the cluster.
kubeadm init phase upload-config kubelet --config kubeadm.yaml
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.71 -
Run this on any machine you wish to join an existing cluster
Synopsis
When joining a kubeadm initialized cluster, we need to establish
bidirectional trust. This is split into discovery (having the Node
trust the Kubernetes Control Plane) and TLS bootstrap (having the
Kubernetes Control Plane trust the Node).
There are 2 main schemes for discovery. The first is to use a shared
token along with the IP address of the API server. The second is to
provide a file - a subset of the standard kubeconfig file. This file
can be a local file or downloaded via an HTTPS URL. The forms are
kubeadm join --discovery-token abcdef.1234567890abcdef 1.2.3.4:6443,
kubeadm join --discovery-file path/to/file.conf, or kubeadm join
--discovery-file https://url/file.conf. Only one form can be used. If
the discovery information is loaded from a URL, HTTPS must be used.
Also, in that case the host installed CA bundle is used to verify
the connection.
If you use a shared token for discovery, you should also pass the
--discovery-token-ca-cert-hash flag to validate the public key of the
root certificate authority (CA) presented by the Kubernetes Control Plane.
The value of this flag is specified as "<hash-type>:<hex-encoded-value>",
where the supported hash type is "sha256". The hash is calculated over
the bytes of the Subject Public Key Info (SPKI) object (as in RFC7469).
This value is available in the output of "kubeadm init" or can be
calculated using standard tools. The --discovery-token-ca-cert-hash flag
may be repeated multiple times to allow more than one public key.
If you cannot know the CA public key hash ahead of time, you can pass
the --discovery-token-unsafe-skip-ca-verification flag to disable this
verification. This weakens the kubeadm security model since other nodes
can potentially impersonate the Kubernetes Control Plane.
The TLS bootstrap mechanism is also driven via a shared token. This is
used to temporarily authenticate with the Kubernetes Control Plane to submit a
certificate signing request (CSR) for a locally created key pair. By
default, kubeadm will set up the Kubernetes Control Plane to automatically
approve these signing requests. This token is passed in with the
--tls-bootstrap-token abcdef.1234567890abcdef flag.
Often times the same token is used for both parts. In this case, the
--token flag can be used instead of specifying each token individually.
The "join [api-server-endpoint]" command executes the following phases:
preflight Run join pre-flight checks
control-plane-prepare Prepare the machine for serving a control plane
/download-certs [EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
/certs Generate the certificates for the new control plane components
/kubeconfig Generate the kubeconfig for the new control plane components
/control-plane Generate the manifests for the new control plane components
kubelet-start Write kubelet settings, certificates and (re)start the kubelet
control-plane-join Join a machine as a control plane instance
/etcd Add a new local etcd member
/update-status Register the new control-plane node into the ClusterStatus maintained in the kubeadm-config ConfigMap (DEPRECATED)
/mark-control-plane Mark a node as a control-plane
kubeadm join [api-server-endpoint] [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
--dry-run
Don't apply any changes; just output what would be done.
-h, --help
help for join
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--skip-phases strings
List of phases to be skipped
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.72 -
Use this command to invoke single phase of the join workflow
Synopsis
Use this command to invoke single phase of the join workflow
Options
-h, --help
help for phase
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.73 -
Join a machine as a control plane instance
Synopsis
Join a machine as a control plane instance
kubeadm join phase control-plane-join [flags]
Examples
# Joins a machine as a control plane instance
kubeadm join phase control-plane-join all
Options
-h, --help
help for control-plane-join
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.74 -
Join a machine as a control plane instance
Synopsis
Join a machine as a control plane instance
kubeadm join phase control-plane-join all [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for all
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for etcd
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for update-status
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.78 -
Prepare the machine for serving a control plane
Synopsis
Prepare the machine for serving a control plane
kubeadm join phase control-plane-prepare [flags]
Examples
# Prepares the machine for serving a control plane
kubeadm join phase control-plane-prepare all
Options
-h, --help
help for control-plane-prepare
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.79 -
Prepare the machine for serving a control plane
Synopsis
Prepare the machine for serving a control plane
kubeadm join phase control-plane-prepare all [api-server-endpoint] [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for all
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.80 -
Generate the certificates for the new control plane components
Synopsis
Generate the certificates for the new control plane components
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for certs
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.81 -
Generate the manifests for the new control plane components
Synopsis
Generate the manifests for the new control plane components
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for control-plane
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.82 -
[EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
Synopsis
[EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for kubelet-start
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
# Run join pre-flight checks using a config file.
kubeadm join phase preflight --config kubeadm-config.yaml
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for preflight
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.86 -
Kubeconfig file utilities
Synopsis
Kubeconfig file utilities.
Options
-h, --help
help for kubeconfig
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.87 -
Output a kubeconfig file for an additional user
Synopsis
Output a kubeconfig file for an additional user.
kubeadm kubeconfig user [flags]
Examples
# Output a kubeconfig file for an additional user named foo using a kubeadm config file bar
kubeadm kubeconfig user --client-name=foo --config=bar
Options
--client-name string
The name of user. It will be used as the CN if client certificates are created
--config string
Path to a kubeadm configuration file.
-h, --help
help for user
--org strings
The orgnizations of the client certificate. It will be used as the O if client certificates are created
--token string
The token that should be used as the authentication mechanism for this kubeconfig, instead of client certificates
--validity-period duration Default: 8760h0m0s
The validity period of the client certificate. It is an offset from the current time.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.88 -
Performs a best effort revert of changes made to this host by 'kubeadm init' or 'kubeadm join'
Synopsis
Performs a best effort revert of changes made to this host by 'kubeadm init' or 'kubeadm join'
The "reset" command executes the following phases:
preflight Run reset pre-flight checks
remove-etcd-member Remove a local etcd member.
cleanup-node Run cleanup node.
kubeadm reset [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path to the directory where the certificates are stored. If specified, clean this directory.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--dry-run
Don't apply any changes; just output what would be done.
-f, --force
Reset the node without prompting for confirmation.
-h, --help
help for reset
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-phases strings
List of phases to be skipped
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.89 -
Use this command to invoke single phase of the reset workflow
Synopsis
Use this command to invoke single phase of the reset workflow
Options
-h, --help
help for phase
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.90 -
Run cleanup node.
Synopsis
Run cleanup node.
kubeadm reset phase cleanup-node [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path to the directory where the certificates are stored. If specified, clean this directory.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
-h, --help
help for cleanup-node
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.91 -
Run reset pre-flight checks
Synopsis
Run pre-flight checks for kubeadm reset.
kubeadm reset phase preflight [flags]
Options
-f, --force
Reset the node without prompting for confirmation.
-h, --help
help for preflight
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.92 -
Remove a local etcd member.
Synopsis
Remove a local etcd member for a control plane node.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.93 -
Manage bootstrap tokens
Synopsis
This command manages bootstrap tokens. It is optional and needed only for advanced use cases.
In short, bootstrap tokens are used for establishing bidirectional trust between a client and a server.
A bootstrap token can be used when a client (for example a node that is about to join the cluster) needs
to trust the server it is talking to. Then a bootstrap token with the "signing" usage can be used.
bootstrap tokens can also function as a way to allow short-lived authentication to the API Server
(the token serves as a way for the API Server to trust the client), for example for doing the TLS Bootstrap.
What is a bootstrap token more exactly?
It is a Secret in the kube-system namespace of type "bootstrap.kubernetes.io/token".
A bootstrap token must be of the form "[a-z0-9]{6}.[a-z0-9]{16}". The former part is the public token ID,
while the latter is the Token Secret and it must be kept private at all circumstances!
The name of the Secret must be named "bootstrap-token-(token-id)".
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.94 -
Create bootstrap tokens on the server
Synopsis
This command will create a bootstrap token for you.
You can specify the usages for this token, the "time to live" and an optional human friendly description.
The [token] is the actual token to write.
This should be a securely generated random token of the form "[a-z0-9]{6}.[a-z0-9]{16}".
If no [token] is given, kubeadm will generate a random token instead.
kubeadm token create [token]
Options
--certificate-key string
When used together with '--print-join-command', print the full 'kubeadm join' flag needed to join the cluster as a control-plane. To create a new certificate key you must use 'kubeadm init phase upload-certs --upload-certs'.
--config string
Path to a kubeadm configuration file.
--description string
A human friendly description of how this token is used.
Describes the ways in which this token can be used. You can pass --usages multiple times or provide a comma separated list of options. Valid options: [signing,authentication]
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.95 -
Delete bootstrap tokens on the server
Synopsis
This command will delete a list of bootstrap tokens for you.
The [token-value] is the full Token of the form "[a-z0-9]{6}.[a-z0-9]{16}" or the
Token ID of the form "[a-z0-9]{6}" to delete.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.96 -
Generate and print a bootstrap token, but do not create it on the server
Synopsis
This command will print out a randomly-generated bootstrap token that can be used with
the "init" and "join" commands.
You don't have to use this command in order to generate a token. You can do so
yourself as long as it is in the format "[a-z0-9]{6}.[a-z0-9]{16}". This
command is provided for convenience to generate tokens in the given format.
You can also use "kubeadm init" without specifying a token and it will
generate and print one for you.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.97 -
List bootstrap tokens on the server
Synopsis
This command will list all bootstrap tokens for you.
kubeadm token list [flags]
Options
--allow-missing-template-keys Default: true
If true, ignore any errors in templates when a field or map key is missing in the template. Only applies to golang and jsonpath output formats.
-o, --experimental-output string Default: "text"
Output format. One of: text|json|yaml|go-template|go-template-file|template|templatefile|jsonpath|jsonpath-as-json|jsonpath-file.
-h, --help
help for list
--show-managed-fields
If true, keep the managedFields when printing objects in JSON or YAML format.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.98 -
Upgrade your cluster smoothly to a newer version with this command
Synopsis
Upgrade your cluster smoothly to a newer version with this command
kubeadm upgrade [flags]
Options
-h, --help
help for upgrade
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.99 -
Upgrade your Kubernetes cluster to the specified version
Synopsis
Upgrade your Kubernetes cluster to the specified version
kubeadm upgrade apply [version]
Options
--allow-experimental-upgrades
Show unstable versions of Kubernetes as an upgrade alternative and allow upgrading to an alpha/beta/release candidate versions of Kubernetes.
--allow-release-candidate-upgrades
Show release candidate versions of Kubernetes as an upgrade alternative and allow upgrading to a release candidate versions of Kubernetes.
--certificate-renewal Default: true
Perform the renewal of certificates used by component changed during upgrades.
--config string
Path to a kubeadm configuration file.
--dry-run
Do not change any state, just output what actions would be performed.
--etcd-upgrade Default: true
Perform the upgrade of etcd.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-f, --force
Force upgrading although some requirements might not be met. This also implies non-interactive mode.
-h, --help
help for apply
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--print-config
Specifies whether the configuration file that will be used in the upgrade should be printed or not.
-y, --yes
Perform the upgrade and do not prompt for confirmation (non-interactive mode).
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.100 -
Show what differences would be applied to existing static pod manifests. See also: kubeadm upgrade apply --dry-run
Synopsis
Show what differences would be applied to existing static pod manifests. See also: kubeadm upgrade apply --dry-run
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.101 -
Upgrade commands for a node in the cluster
Synopsis
Upgrade commands for a node in the cluster
The "node" command executes the following phases:
preflight Run upgrade node pre-flight checks
control-plane Upgrade the control plane instance deployed on this node, if any
kubelet-config Upgrade the kubelet configuration for this node
kubeadm upgrade node [flags]
Options
--certificate-renewal Default: true
Perform the renewal of certificates used by component changed during upgrades.
--dry-run
Do not change any state, just output the actions that would be performed.
--etcd-upgrade Default: true
Perform the upgrade of etcd.
-h, --help
help for node
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--skip-phases strings
List of phases to be skipped
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.102 -
Use this command to invoke single phase of the node workflow
Synopsis
Use this command to invoke single phase of the node workflow
Options
-h, --help
help for phase
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.103 -
Upgrade the control plane instance deployed on this node, if any
Synopsis
Upgrade the control plane instance deployed on this node, if any
kubeadm upgrade node phase control-plane [flags]
Options
--certificate-renewal Default: true
Perform the renewal of certificates used by component changed during upgrades.
--dry-run
Do not change any state, just output the actions that would be performed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.104 -
Upgrade the kubelet configuration for this node
Synopsis
Download the kubelet configuration from the kubelet-config ConfigMap stored in the cluster
kubeadm upgrade node phase kubelet-config [flags]
Options
--dry-run
Do not change any state, just output the actions that would be performed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.105 -
Run upgrade node pre-flight checks
Synopsis
Run pre-flight checks for kubeadm upgrade node.
kubeadm upgrade node phase preflight [flags]
Options
-h, --help
help for preflight
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.106 -
Check which versions are available to upgrade to and validate whether your current cluster is upgradeable. To skip the internet check, pass in the optional [version] parameter
Synopsis
Check which versions are available to upgrade to and validate whether your current cluster is upgradeable. To skip the internet check, pass in the optional [version] parameter
kubeadm upgrade plan [version] [flags]
Options
--allow-experimental-upgrades
Show unstable versions of Kubernetes as an upgrade alternative and allow upgrading to an alpha/beta/release candidate versions of Kubernetes.
--allow-release-candidate-upgrades
Show release candidate versions of Kubernetes as an upgrade alternative and allow upgrading to a release candidate versions of Kubernetes.
--config string
Path to a kubeadm configuration file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-h, --help
help for plan
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--print-config
Specifies whether the configuration file that will be used in the upgrade should be printed or not.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.107 -
Print the version of kubeadm
Synopsis
Print the version of kubeadm
kubeadm version [flags]
Options
-h, --help
help for version
-o, --output string
Output format; available options are 'yaml', 'json' and 'short'
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.1.108 -
All files in this directory are auto-generated from other repos. Do not edit them manually. You must edit them in their upstream repo.
6.9.1.2 - kubeadm init
This command initializes a Kubernetes control-plane node.
Run this command in order to set up the Kubernetes control plane
Synopsis
Run this command in order to set up the Kubernetes control plane
The "init" command executes the following phases:
preflight Run pre-flight checks
certs Certificate generation
/ca Generate the self-signed Kubernetes CA to provision identities for other Kubernetes components
/apiserver Generate the certificate for serving the Kubernetes API
/apiserver-kubelet-client Generate the certificate for the API server to connect to kubelet
/front-proxy-ca Generate the self-signed CA to provision identities for front proxy
/front-proxy-client Generate the certificate for the front proxy client
/etcd-ca Generate the self-signed CA to provision identities for etcd
/etcd-server Generate the certificate for serving etcd
/etcd-peer Generate the certificate for etcd nodes to communicate with each other
/etcd-healthcheck-client Generate the certificate for liveness probes to healthcheck etcd
/apiserver-etcd-client Generate the certificate the apiserver uses to access etcd
/sa Generate a private key for signing service account tokens along with its public key
kubeconfig Generate all kubeconfig files necessary to establish the control plane and the admin kubeconfig file
/admin Generate a kubeconfig file for the admin to use and for kubeadm itself
/kubelet Generate a kubeconfig file for the kubelet to use *only* for cluster bootstrapping purposes
/controller-manager Generate a kubeconfig file for the controller manager to use
/scheduler Generate a kubeconfig file for the scheduler to use
kubelet-start Write kubelet settings and (re)start the kubelet
control-plane Generate all static Pod manifest files necessary to establish the control plane
/apiserver Generates the kube-apiserver static Pod manifest
/controller-manager Generates the kube-controller-manager static Pod manifest
/scheduler Generates the kube-scheduler static Pod manifest
etcd Generate static Pod manifest file for local etcd
/local Generate the static Pod manifest file for a local, single-node local etcd instance
upload-config Upload the kubeadm and kubelet configuration to a ConfigMap
/kubeadm Upload the kubeadm ClusterConfiguration to a ConfigMap
/kubelet Upload the kubelet component config to a ConfigMap
upload-certs Upload certificates to kubeadm-certs
mark-control-plane Mark a node as a control-plane
bootstrap-token Generates bootstrap tokens used to join a node to a cluster
kubelet-finalize Updates settings relevant to the kubelet after TLS bootstrap
/experimental-cert-rotation Enable kubelet client certificate rotation
addon Install required addons for passing conformance tests
/coredns Install the CoreDNS addon to a Kubernetes cluster
/kube-proxy Install the kube-proxy addon to a Kubernetes cluster
kubeadm init [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--apiserver-cert-extra-sans strings
Optional extra Subject Alternative Names (SANs) to use for the API Server serving certificate. Can be both IP addresses and DNS names.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--certificate-key string
Key used to encrypt the control-plane certificates in the kubeadm-certs Secret.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-h, --help
help for init
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Use alternative domain for services, e.g. "myorg.internal".
--skip-certificate-key-print
Don't print the key used to encrypt the control-plane certificates.
--skip-phases strings
List of phases to be skipped
--skip-token-print
Skip printing of the default bootstrap token generated by 'kubeadm init'.
--token string
The token to use for establishing bidirectional trust between nodes and control-plane nodes. The format is [a-z0-9]{6}.[a-z0-9]{16} - e.g. abcdef.0123456789abcdef
--token-ttl duration Default: 24h0m0s
The duration before the token is automatically deleted (e.g. 1s, 2m, 3h). If set to '0', the token will never expire
--upload-certs
Upload control-plane certificates to the kubeadm-certs Secret.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Init workflow
kubeadm init bootstraps a Kubernetes control-plane node by executing the
following steps:
Runs a series of pre-flight checks to validate the system state
before making changes. Some checks only trigger warnings, others are
considered errors and will exit kubeadm until the problem is corrected or the
user specifies --ignore-preflight-errors=<list-of-errors>.
Generates a self-signed CA to set up identities for each component in the cluster. The user can provide their
own CA cert and/or key by dropping it in the cert directory configured via --cert-dir
(/etc/kubernetes/pki by default).
The APIServer certs will have additional SAN entries for any --apiserver-cert-extra-sans
arguments, lowercased if necessary.
Writes kubeconfig files in /etc/kubernetes/ for the kubelet, the controller-manager and the
scheduler to use to connect to the API server, each with its own identity, as well as an
additional kubeconfig file for administration named admin.conf.
Generates static Pod manifests for the API server,
controller-manager and scheduler. In case an external etcd is not provided,
an additional static Pod manifest is generated for etcd.
Static Pod manifests are written to /etc/kubernetes/manifests; the kubelet
watches this directory for Pods to create on startup.
Once control plane Pods are up and running, the kubeadm init sequence can continue.
Apply labels and taints to the control-plane node so that no additional workloads will
run there.
Generates the token that additional nodes can use to register
themselves with a control-plane in the future. Optionally, the user can provide a
token via --token, as described in the
kubeadm token docs.
Makes all the necessary configurations for allowing node joining with the
Bootstrap Tokens and
TLS Bootstrap
mechanism:
Write a ConfigMap for making available all the information required
for joining, and set up related RBAC access rules.
Installs a DNS server (CoreDNS) and the kube-proxy addon components via the API server.
In Kubernetes version 1.11 and later CoreDNS is the default DNS server.
Please note that although the DNS server is deployed, it will not be scheduled until CNI is installed.
Warning: kube-dns usage with kubeadm is deprecated as of v1.18 and is removed in v1.21.
Using init phases with kubeadm
Kubeadm allows you to create a control-plane node in phases using the kubeadm init phase command.
To view the ordered list of phases and sub-phases you can call kubeadm init --help. The list
will be located at the top of the help screen and each phase will have a description next to it.
Note that by calling kubeadm init all of the phases and sub-phases will be executed in this exact order.
Some phases have unique flags, so if you want to have a look at the list of available options add
--help, for example:
You can also use --help to see the list of sub-phases for a certain parent phase:
sudo kubeadm init phase control-plane --help
kubeadm init also exposes a flag called --skip-phases that can be used to skip certain phases.
The flag accepts a list of phase names and the names can be taken from the above ordered list.
An example:
sudo kubeadm init phase control-plane all --config=configfile.yaml
sudo kubeadm init phase etcd local --config=configfile.yaml
# you can now modify the control plane and etcd manifest filessudo kubeadm init --skip-phases=control-plane,etcd --config=configfile.yaml
What this example would do is write the manifest files for the control plane and etcd in
/etc/kubernetes/manifests based on the configuration in configfile.yaml. This allows you to
modify the files and then skip these phases using --skip-phases. By calling the last command you
will create a control plane node with the custom manifest files.
FEATURE STATE:Kubernetes v1.22 [beta]
Alternatively, you can use the skipPhases field under InitConfiguration.
Using kubeadm init with a configuration file
Caution: The config file is still considered beta and may change in future versions.
It's possible to configure kubeadm init with a configuration file instead of command
line flags, and some more advanced features may only be available as
configuration file options. This file is passed using the --config flag and it must
contain a ClusterConfiguration structure and optionally more structures separated by ---\n
Mixing --config with others flags may not be allowed in some cases.
The default configuration can be printed out using the
kubeadm config print command.
If your configuration is not using the latest version it is recommended that you migrate using
the kubeadm config migrate command.
For more information on the fields and usage of the configuration you can navigate to our
API reference page.
Using kubeadm init with feature gates
Kubeadm supports a set of feature gates that are unique to kubeadm and can only be applied
during cluster creation with kubeadm init. These features can control the behavior
of the cluster. Feature gates are removed after a feature graduates to GA.
To pass a feature gate you can either use the --feature-gates flag for
kubeadm init, or you can add items into the featureGates field when you pass
a configuration file
using --config.
Note: Once a feature gate goes GA its value becomes locked to true by default.
Feature gate descriptions:
PublicKeysECDSA
Can be used to create a cluster that uses ECDSA certificates instead of the default RSA algorithm.
Renewal of existing ECDSA certificates is also supported using kubeadm certs renew, but you cannot
switch between the RSA and ECDSA algorithms on the fly or during upgrades.
RootlessControlPlane
Setting this flag configures the kubeadm deployed control plane component static Pod containers
for kube-apiserver, kube-controller-manager, kube-scheduler and etcd to run as non-root users.
If the flag is not set, those components run as root. You can change the value of this feature gate before
you upgrade to a newer version of Kubernetes.
UnversionedKubeletConfigMap
This flag controls the name of the ConfigMap where kubeadm stores
kubelet configuration data. With this flag not specified or set to true, the ConfigMap is named kubelet-config.
If you set this flag to false, the name of the ConfigMap includes the major and minor version for Kubernetes
(for example: kubelet-config-1.25). Kubeadm ensures that RBAC rules for reading and writing
that ConfigMap are appropriate for the value you set. When kubeadm writes this ConfigMap (during kubeadm init
or kubeadm upgrade apply), kubeadm respects the value of UnversionedKubeletConfigMap. When reading that ConfigMap
(during kubeadm join, kubeadm reset, kubeadm upgrade ...), kubeadm attempts to use unversioned ConfigMap name first;
if that does not succeed, kubeadm falls back to using the legacy (versioned) name for that ConfigMap.
Adding kube-proxy parameters
For information about kube-proxy parameters in the kubeadm configuration see:
For running kubeadm without an Internet connection you have to pre-pull the required control-plane images.
You can list and pull the images using the kubeadm config images sub-command:
kubeadm config images list
kubeadm config images pull
You can pass --config to the above commands with a kubeadm configuration file
to control the kubernetesVersion and imageRepository fields.
All default registry.k8s.io images that kubeadm requires support multiple architectures.
Using custom images
By default, kubeadm pulls images from registry.k8s.io. If the
requested Kubernetes version is a CI label (such as ci/latest)
gcr.io/k8s-staging-ci-images is used.
To provide kubernetesVersion which affects the version of the images.
To provide an alternative imageRepository to be used instead of
registry.k8s.io.
To provide a specific imageRepository and imageTag for etcd or CoreDNS.
Image paths between the default registry.k8s.io and a custom repository specified using
imageRepository may differ for backwards compatibility reasons. For example,
one image might have a subpath at registry.k8s.io/subpath/image, but be defaulted
to my.customrepository.io/image when using a custom repository.
To ensure you push the images to your custom repository in paths that kubeadm
can consume, you must:
Pull images from the defaults paths at registry.k8s.io using kubeadm config images {list|pull}.
Push images to the paths from kubeadm config images list --config=config.yaml,
where config.yaml contains the custom imageRepository, and/or imageTag
for etcd and CoreDNS.
Pass the same config.yaml to kubeadm init.
Custom sandbox (pause) images
To set a custom image for these you need to configure this in your
container runtime
to use the image.
Consult the documentation for your container runtime to find out how to change this setting;
for selected container runtimes, you can also find advice within the
Container Runtimes topic.
Uploading control-plane certificates to the cluster
By adding the flag --upload-certs to kubeadm init you can temporary upload
the control-plane certificates to a Secret in the cluster. Please note that this Secret
will expire automatically after 2 hours. The certificates are encrypted using
a 32byte key that can be specified using --certificate-key. The same key can be used
to download the certificates when additional control-plane nodes are joining, by passing
--control-plane and --certificate-key to kubeadm join.
The following phase command can be used to re-upload the certificates after expiration:
If the flag --certificate-key is not passed to kubeadm init and
kubeadm init phase upload-certs a new key will be generated automatically.
The following command can be used to generate a new key on demand:
kubeadm certs certificate-key
Certificate management with kubeadm
For detailed information on certificate management with kubeadm see
Certificate Management with kubeadm.
The document includes information about using external CA, custom certificates
and certificate renewal.
Managing the kubeadm drop-in file for the kubelet
The kubeadm package ships with a configuration file for running the kubelet by systemd.
Note that the kubeadm CLI never touches this drop-in file. This drop-in file is part of the kubeadm
DEB/RPM package.
By default kubeadm attempts to detect your container runtime. For more details on this detection,
see the kubeadm CRI installation guide.
Setting the node name
By default, kubeadm assigns a node name based on a machine's host address.
You can override this setting with the --node-name flag.
The flag passes the appropriate --hostname-override
value to the kubelet.
Rather than copying the token you obtained from kubeadm init to each node, as
in the basic kubeadm tutorial,
you can parallelize the token distribution for easier automation. To implement this automation,
you must know the IP address that the control-plane node will have after it is started, or use a
DNS name or an address of a load balancer.
Generate a token. This token must have the form <6 character string>.<16 character string>. More formally, it must match the regex:
[a-z0-9]{6}\.[a-z0-9]{16}.
kubeadm can generate a token for you:
kubeadm token generate
Start both the control-plane node and the worker nodes concurrently with this token.
As they come up they should find each other and form the cluster. The same
--token argument can be used on both kubeadm init and kubeadm join.
Similar can be done for --certificate-key when joining additional control-plane
nodes. The key can be generated using:
kubeadm certs certificate-key
Once the cluster is up, you can grab the admin credentials from the control-plane node
at /etc/kubernetes/admin.conf and use that to talk to the cluster.
Note that this style of bootstrap has some relaxed security guarantees because
it does not allow the root CA hash to be validated with
--discovery-token-ca-cert-hash (since it's not generated when the nodes are
provisioned). For details, see the kubeadm join.
kubeadm join to bootstrap a Kubernetes
worker node and join it to the cluster
kubeadm upgrade to upgrade a Kubernetes
cluster to a newer version
kubeadm reset to revert any changes made
to this host by kubeadm init or kubeadm join
6.9.1.3 - kubeadm join
This command initializes a Kubernetes worker node and joins it to the cluster.
Run this on any machine you wish to join an existing cluster
Synopsis
When joining a kubeadm initialized cluster, we need to establish
bidirectional trust. This is split into discovery (having the Node
trust the Kubernetes Control Plane) and TLS bootstrap (having the
Kubernetes Control Plane trust the Node).
There are 2 main schemes for discovery. The first is to use a shared
token along with the IP address of the API server. The second is to
provide a file - a subset of the standard kubeconfig file. This file
can be a local file or downloaded via an HTTPS URL. The forms are
kubeadm join --discovery-token abcdef.1234567890abcdef 1.2.3.4:6443,
kubeadm join --discovery-file path/to/file.conf, or kubeadm join
--discovery-file https://url/file.conf. Only one form can be used. If
the discovery information is loaded from a URL, HTTPS must be used.
Also, in that case the host installed CA bundle is used to verify
the connection.
If you use a shared token for discovery, you should also pass the
--discovery-token-ca-cert-hash flag to validate the public key of the
root certificate authority (CA) presented by the Kubernetes Control Plane.
The value of this flag is specified as "<hash-type>:<hex-encoded-value>",
where the supported hash type is "sha256". The hash is calculated over
the bytes of the Subject Public Key Info (SPKI) object (as in RFC7469).
This value is available in the output of "kubeadm init" or can be
calculated using standard tools. The --discovery-token-ca-cert-hash flag
may be repeated multiple times to allow more than one public key.
If you cannot know the CA public key hash ahead of time, you can pass
the --discovery-token-unsafe-skip-ca-verification flag to disable this
verification. This weakens the kubeadm security model since other nodes
can potentially impersonate the Kubernetes Control Plane.
The TLS bootstrap mechanism is also driven via a shared token. This is
used to temporarily authenticate with the Kubernetes Control Plane to submit a
certificate signing request (CSR) for a locally created key pair. By
default, kubeadm will set up the Kubernetes Control Plane to automatically
approve these signing requests. This token is passed in with the
--tls-bootstrap-token abcdef.1234567890abcdef flag.
Often times the same token is used for both parts. In this case, the
--token flag can be used instead of specifying each token individually.
The "join [api-server-endpoint]" command executes the following phases:
preflight Run join pre-flight checks
control-plane-prepare Prepare the machine for serving a control plane
/download-certs [EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
/certs Generate the certificates for the new control plane components
/kubeconfig Generate the kubeconfig for the new control plane components
/control-plane Generate the manifests for the new control plane components
kubelet-start Write kubelet settings, certificates and (re)start the kubelet
control-plane-join Join a machine as a control plane instance
/etcd Add a new local etcd member
/update-status Register the new control-plane node into the ClusterStatus maintained in the kubeadm-config ConfigMap (DEPRECATED)
/mark-control-plane Mark a node as a control-plane
kubeadm join [api-server-endpoint] [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
--dry-run
Don't apply any changes; just output what would be done.
-h, --help
help for join
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--skip-phases strings
List of phases to be skipped
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
The join workflow
kubeadm join bootstraps a Kubernetes worker node or a control-plane node and adds it to the cluster.
This action consists of the following steps for worker nodes:
kubeadm downloads necessary cluster information from the API server.
By default, it uses the bootstrap token and the CA key hash to verify the
authenticity of that data. The root CA can also be discovered directly via a
file or URL.
Once the cluster information is known, kubelet can start the TLS bootstrapping
process.
The TLS bootstrap uses the shared token to temporarily authenticate
with the Kubernetes API server to submit a certificate signing request (CSR); by
default the control plane signs this CSR request automatically.
Finally, kubeadm configures the local kubelet to connect to the API
server with the definitive identity assigned to the node.
For control-plane nodes additional steps are performed:
Downloading certificates shared among control-plane nodes from the cluster
(if explicitly requested by the user).
Generating control-plane component manifests, certificates and kubeconfig.
Adding new local etcd member.
Using join phases with kubeadm
Kubeadm allows you join a node to the cluster in phases using kubeadm join phase.
To view the ordered list of phases and sub-phases you can call kubeadm join --help. The list will be located
at the top of the help screen and each phase will have a description next to it.
Note that by calling kubeadm join all of the phases and sub-phases will be executed in this exact order.
Some phases have unique flags, so if you want to have a look at the list of available options add --help, for example:
kubeadm join phase kubelet-start --help
Similar to the kubeadm init phase
command, kubeadm join phase allows you to skip a list of phases using the --skip-phases flag.
Alternatively, you can use the skipPhases field in JoinConfiguration.
Discovering what cluster CA to trust
The kubeadm discovery has several options, each with security tradeoffs.
The right method for your environment depends on how you provision nodes and the
security expectations you have about your network and node lifecycles.
Token-based discovery with CA pinning
This is the default mode in kubeadm. In this mode, kubeadm downloads
the cluster configuration (including root CA) and validates it using the token
as well as validating that the root CA public key matches the provided hash and
that the API server certificate is valid under the root CA.
The CA key hash has the format sha256:<hex_encoded_hash>. By default, the hash value is returned in the kubeadm join command printed at the end of kubeadm init or in the output of kubeadm token create --print-join-command. It is in a standard format (see RFC7469) and can also be calculated by 3rd party tools or provisioning systems. For example, using the OpenSSL CLI:
You can also call join for a control-plane node with --certificate-key to copy certificates to this node,
if the kubeadm init command was called with --upload-certs.
Advantages:
Allows bootstrapping nodes to securely discover a root of trust for the
control-plane node even if other worker nodes or the network are compromised.
Convenient to execute manually since all of the information required fits
into a single kubeadm join command.
Disadvantages:
The CA hash is not normally known until the control-plane node has been provisioned,
which can make it more difficult to build automated provisioning tools that
use kubeadm. By generating your CA in beforehand, you may workaround this
limitation.
Token-based discovery without CA pinning
This mode relies only on the symmetric token to sign
(HMAC-SHA256) the discovery information that establishes the root of trust for
the control-plane. To use the mode the joining nodes must skip the hash validation of the
CA public key, using --discovery-token-unsafe-skip-ca-verification. You should consider
using one of the other modes if possible.
Still protects against many network-level attacks.
The token can be generated ahead of time and shared with the control-plane node and
worker nodes, which can then bootstrap in parallel without coordination. This
allows it to be used in many provisioning scenarios.
Disadvantages:
If an attacker is able to steal a bootstrap token via some vulnerability,
they can use that token (along with network-level access) to impersonate the
control-plane node to other bootstrapping nodes. This may or may not be an appropriate
tradeoff in your environment.
File or HTTPS-based discovery
This provides an out-of-band way to establish a root of trust between the control-plane node
and bootstrapping nodes. Consider using this mode if you are building automated provisioning
using kubeadm. The format of the discovery file is a regular Kubernetes
kubeconfig file.
In case the discovery file does not contain credentials, the TLS discovery token will be used.
Allows bootstrapping nodes to securely discover a root of trust for the
control-plane node even if the network or other worker nodes are compromised.
Disadvantages:
Requires that you have some way to carry the discovery information from
the control-plane node to the bootstrapping nodes. If the discovery file contains credentials
you must keep it secret and transfer it over a secure channel. This might be possible with your
cloud provider or provisioning tool.
Securing your installation even more
The defaults for kubeadm may not work for everyone. This section documents how to tighten up a kubeadm installation
at the cost of some usability.
Turning off auto-approval of node client certificates
By default, there is a CSR auto-approver enabled that basically approves any client certificate request
for a kubelet when a Bootstrap Token was used when authenticating. If you don't want the cluster to
automatically approve kubelet client certs, you can turn it off by executing this command:
NAME AGE REQUESTOR CONDITION
node-csr-c69HXe7aYcqkS1bKmH4faEnHAWxn6i2bHZ2mD04jZyQ 1m system:bootstrap:878f07 Approved,Issued
This forces the workflow that kubeadm join will only succeed if kubectl certificate approve has been run.
Turning off public access to the cluster-info ConfigMap
In order to achieve the joining flow using the token as the only piece of validation information, a
ConfigMap with some data needed for validation of the control-plane node's identity is exposed publicly by
default. While there is no private data in this ConfigMap, some users might wish to turn
it off regardless. Doing so will disable the ability to use the --discovery-token flag of the
kubeadm join flow. Here are the steps to do so:
Fetch the cluster-info file from the API Server:
kubectl -n kube-public get cm cluster-info -o yaml | grep "kubeconfig:" -A11 | grep "apiVersion" -A10 | sed "s/ //" | tee cluster-info.yaml
These commands should be run after kubeadm init but before kubeadm join.
Using kubeadm join with a configuration file
Caution: The config file is still considered beta and may change in future versions.
It's possible to configure kubeadm join with a configuration file instead of command
line flags, and some more advanced features may only be available as
configuration file options. This file is passed using the --config flag and it must
contain a JoinConfiguration structure. Mixing --config with others flags may not be
allowed in some cases.
The default configuration can be printed out using the
kubeadm config print command.
If your configuration is not using the latest version it is recommended that you migrate using
the kubeadm config migrate command.
For more information on the fields and usage of the configuration you can navigate to our
API reference.
What's next
kubeadm init to bootstrap a Kubernetes control-plane node
kubeadm reset to revert any changes made to this host by kubeadm init or kubeadm join
6.9.1.4 - kubeadm upgrade
kubeadm upgrade is a user-friendly command that wraps complex upgrading logic
behind one command, with support for both planning an upgrade and actually performing it.
kubeadm upgrade guidance
The steps for performing an upgrade using kubeadm are outlined in this document.
For older versions of kubeadm, please refer to older documentation sets of the Kubernetes website.
You can use kubeadm upgrade diff to see the changes that would be applied to static pod manifests.
In Kubernetes v1.15.0 and later, kubeadm upgrade apply and kubeadm upgrade node will also
automatically renew the kubeadm managed certificates on this node, including those stored in kubeconfig files.
To opt-out, it is possible to pass the flag --certificate-renewal=false. For more details about certificate
renewal see the certificate management documentation.
Note: The commands kubeadm upgrade apply and kubeadm upgrade plan have a legacy --config
flag which makes it possible to reconfigure the cluster, while performing planning or upgrade of that particular
control-plane node. Please be aware that the upgrade workflow was not designed for this scenario and there are
reports of unexpected results.
kubeadm upgrade plan
Check which versions are available to upgrade to and validate whether your current cluster is upgradeable. To skip the internet check, pass in the optional [version] parameter
Synopsis
Check which versions are available to upgrade to and validate whether your current cluster is upgradeable. To skip the internet check, pass in the optional [version] parameter
kubeadm upgrade plan [version] [flags]
Options
--allow-experimental-upgrades
Show unstable versions of Kubernetes as an upgrade alternative and allow upgrading to an alpha/beta/release candidate versions of Kubernetes.
--allow-release-candidate-upgrades
Show release candidate versions of Kubernetes as an upgrade alternative and allow upgrading to a release candidate versions of Kubernetes.
--config string
Path to a kubeadm configuration file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-h, --help
help for plan
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--print-config
Specifies whether the configuration file that will be used in the upgrade should be printed or not.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm upgrade apply
Upgrade your Kubernetes cluster to the specified version
Synopsis
Upgrade your Kubernetes cluster to the specified version
kubeadm upgrade apply [version]
Options
--allow-experimental-upgrades
Show unstable versions of Kubernetes as an upgrade alternative and allow upgrading to an alpha/beta/release candidate versions of Kubernetes.
--allow-release-candidate-upgrades
Show release candidate versions of Kubernetes as an upgrade alternative and allow upgrading to a release candidate versions of Kubernetes.
--certificate-renewal Default: true
Perform the renewal of certificates used by component changed during upgrades.
--config string
Path to a kubeadm configuration file.
--dry-run
Do not change any state, just output what actions would be performed.
--etcd-upgrade Default: true
Perform the upgrade of etcd.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
-f, --force
Force upgrading although some requirements might not be met. This also implies non-interactive mode.
-h, --help
help for apply
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--print-config
Specifies whether the configuration file that will be used in the upgrade should be printed or not.
-y, --yes
Perform the upgrade and do not prompt for confirmation (non-interactive mode).
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm upgrade diff
Show what differences would be applied to existing static pod manifests. See also: kubeadm upgrade apply --dry-run
Synopsis
Show what differences would be applied to existing static pod manifests. See also: kubeadm upgrade apply --dry-run
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm upgrade node
Upgrade commands for a node in the cluster
Synopsis
Upgrade commands for a node in the cluster
The "node" command executes the following phases:
preflight Run upgrade node pre-flight checks
control-plane Upgrade the control plane instance deployed on this node, if any
kubelet-config Upgrade the kubelet configuration for this node
kubeadm upgrade node [flags]
Options
--certificate-renewal Default: true
Perform the renewal of certificates used by component changed during upgrades.
--dry-run
Do not change any state, just output the actions that would be performed.
--etcd-upgrade Default: true
Perform the upgrade of etcd.
-h, --help
help for node
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--skip-phases strings
List of phases to be skipped
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
What's next
kubeadm config if you initialized your cluster using kubeadm v1.7.x or lower, to configure your cluster for kubeadm upgrade
6.9.1.5 - kubeadm config
During kubeadm init, kubeadm uploads the ClusterConfiguration object to your cluster
in a ConfigMap called kubeadm-config in the kube-system namespace. This configuration is then read during
kubeadm join, kubeadm reset and kubeadm upgrade.
You can use kubeadm config print to print the default static configuration that kubeadm
uses for kubeadm init and kubeadm join.
Note: The output of the command is meant to serve as an example. You must manually edit the output
of this command to adapt to your setup. Remove the fields that you are not certain about and kubeadm
will try to default them on runtime by examining the host.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm config print init-defaults
Print default init configuration, that can be used for 'kubeadm init'
Synopsis
This command prints objects such as the default init configuration that is used for 'kubeadm init'.
Note that sensitive values like the Bootstrap Token fields are replaced with placeholder values like "abcdef.0123456789abcdef" in order to pass validation but
not perform the real computation for creating a token.
kubeadm config print init-defaults [flags]
Options
--component-configs strings
A comma-separated list for component config API objects to print the default values for. Available values: [KubeProxyConfiguration KubeletConfiguration]. If this flag is not set, no component configs will be printed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm config print join-defaults
Print default join configuration, that can be used for 'kubeadm join'
Synopsis
This command prints objects such as the default join configuration that is used for 'kubeadm join'.
Note that sensitive values like the Bootstrap Token fields are replaced with placeholder values like "abcdef.0123456789abcdef" in order to pass validation but
not perform the real computation for creating a token.
kubeadm config print join-defaults [flags]
Options
--component-configs strings
A comma-separated list for component config API objects to print the default values for. Available values: [KubeProxyConfiguration KubeletConfiguration]. If this flag is not set, no component configs will be printed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm config migrate
Read an older version of the kubeadm configuration API types from a file, and output the similar config object for the newer version
Synopsis
This command lets you convert configuration objects of older versions to the latest supported version,
locally in the CLI tool without ever touching anything in the cluster.
In this version of kubeadm, the following API versions are supported:
kubeadm.k8s.io/v1beta3
Further, kubeadm can only write out config of version "kubeadm.k8s.io/v1beta3", but read both types.
So regardless of what version you pass to the --old-config parameter here, the API object will be
read, deserialized, defaulted, converted, validated, and re-serialized when written to stdout or
--new-config if specified.
In other words, the output of this command is what kubeadm actually would read internally if you
submitted this file to "kubeadm init"
kubeadm config migrate [flags]
Options
-h, --help
help for migrate
--new-config string
Path to the resulting equivalent kubeadm config file using the new API version. Optional, if not specified output will be sent to STDOUT.
--old-config string
Path to the kubeadm config file that is using an old API version and should be converted. This flag is mandatory.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm config images list
Print a list of images kubeadm will use. The configuration file is used in case any images or image repositories are customized
Synopsis
Print a list of images kubeadm will use. The configuration file is used in case any images or image repositories are customized
kubeadm config images list [flags]
Options
--allow-missing-template-keys Default: true
If true, ignore any errors in templates when a field or map key is missing in the template. Only applies to golang and jsonpath output formats.
--config string
Path to a kubeadm configuration file.
-o, --experimental-output string Default: "text"
Output format. One of: text|json|yaml|go-template|go-template-file|template|templatefile|jsonpath|jsonpath-as-json|jsonpath-file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm config images pull
Pull images used by kubeadm
Synopsis
Pull images used by kubeadm
kubeadm config images pull [flags]
Options
--config string
Path to a kubeadm configuration file.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
What's next
kubeadm upgrade to upgrade a Kubernetes cluster to a newer version
6.9.1.6 - kubeadm reset
Performs a best effort revert of changes made by kubeadm init or kubeadm join.
Performs a best effort revert of changes made to this host by 'kubeadm init' or 'kubeadm join'
Synopsis
Performs a best effort revert of changes made to this host by 'kubeadm init' or 'kubeadm join'
The "reset" command executes the following phases:
preflight Run reset pre-flight checks
remove-etcd-member Remove a local etcd member.
cleanup-node Run cleanup node.
kubeadm reset [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path to the directory where the certificates are stored. If specified, clean this directory.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--dry-run
Don't apply any changes; just output what would be done.
-f, --force
Reset the node without prompting for confirmation.
-h, --help
help for reset
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-phases strings
List of phases to be skipped
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Reset workflow
kubeadm reset is responsible for cleaning up a node local file system from files that were created using
the kubeadm init or kubeadm join commands. For control-plane nodes reset also removes the local stacked
etcd member of this node from the etcd cluster.
kubeadm reset phase can be used to execute the separate phases of the above workflow.
To skip a list of phases you can use the --skip-phases flag, which works in a similar way to
the kubeadm join and kubeadm init phase runners.
External etcd clean up
kubeadm reset will not delete any etcd data if external etcd is used. This means that if you run kubeadm init again using the same etcd endpoints, you will see state from previous clusters.
To wipe etcd data it is recommended you use a client like etcdctl, such as:
kubeadm init to bootstrap a Kubernetes control-plane node
kubeadm join to bootstrap a Kubernetes worker node and join it to the cluster
6.9.1.7 - kubeadm token
Bootstrap tokens are used for establishing bidirectional trust between a node joining
the cluster and a control-plane node, as described in authenticating with bootstrap tokens.
kubeadm init creates an initial token with a 24-hour TTL. The following commands allow you to manage
such a token and also to create and manage new ones.
kubeadm token create
Create bootstrap tokens on the server
Synopsis
This command will create a bootstrap token for you.
You can specify the usages for this token, the "time to live" and an optional human friendly description.
The [token] is the actual token to write.
This should be a securely generated random token of the form "[a-z0-9]{6}.[a-z0-9]{16}".
If no [token] is given, kubeadm will generate a random token instead.
kubeadm token create [token]
Options
--certificate-key string
When used together with '--print-join-command', print the full 'kubeadm join' flag needed to join the cluster as a control-plane. To create a new certificate key you must use 'kubeadm init phase upload-certs --upload-certs'.
--config string
Path to a kubeadm configuration file.
--description string
A human friendly description of how this token is used.
Describes the ways in which this token can be used. You can pass --usages multiple times or provide a comma separated list of options. Valid options: [signing,authentication]
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm token delete
Delete bootstrap tokens on the server
Synopsis
This command will delete a list of bootstrap tokens for you.
The [token-value] is the full Token of the form "[a-z0-9]{6}.[a-z0-9]{16}" or the
Token ID of the form "[a-z0-9]{6}" to delete.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm token generate
Generate and print a bootstrap token, but do not create it on the server
Synopsis
This command will print out a randomly-generated bootstrap token that can be used with
the "init" and "join" commands.
You don't have to use this command in order to generate a token. You can do so
yourself as long as it is in the format "[a-z0-9]{6}.[a-z0-9]{16}". This
command is provided for convenience to generate tokens in the given format.
You can also use "kubeadm init" without specifying a token and it will
generate and print one for you.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm token list
List bootstrap tokens on the server
Synopsis
This command will list all bootstrap tokens for you.
kubeadm token list [flags]
Options
--allow-missing-template-keys Default: true
If true, ignore any errors in templates when a field or map key is missing in the template. Only applies to golang and jsonpath output formats.
-o, --experimental-output string Default: "text"
Output format. One of: text|json|yaml|go-template|go-template-file|template|templatefile|jsonpath|jsonpath-as-json|jsonpath-file.
-h, --help
help for list
--show-managed-fields
If true, keep the managedFields when printing objects in JSON or YAML format.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
What's next
kubeadm join to bootstrap a Kubernetes worker node and join it to the cluster
6.9.1.8 - kubeadm version
This command prints the version of kubeadm.
Print the version of kubeadm
Synopsis
Print the version of kubeadm
kubeadm version [flags]
Options
-h, --help
help for version
-o, --output string
Output format; available options are 'yaml', 'json' and 'short'
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.9 - kubeadm alpha
Caution:kubeadm alpha provides a preview of a set of features made available for gathering feedback
from the community. Please try it out and give us feedback!
Currently there are no experimental commands under kubeadm alpha.
What's next
kubeadm init to bootstrap a Kubernetes control-plane node
kubeadm reset to revert any changes made to this host by kubeadm init or kubeadm join
6.9.1.10 - kubeadm certs
kubeadm certs provides utilities for managing certificates.
For more details on how these commands can be used, see
Certificate Management with kubeadm.
kubeadm certs
A collection of operations for operating Kubernetes certificates.
This command is not meant to be run on its own. See list of available subcommands.
kubeadm certs renew [flags]
Options
-h, --help
help for renew
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew all available certificates
Synopsis
Renew all known certificates necessary to run the control plane. Renewals are run unconditionally, regardless of expiration date. Renewals can also be run individually for more control.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate embedded in the kubeconfig file for the admin to use and for kubeadm itself
Synopsis
Renew the certificate embedded in the kubeconfig file for the admin to use and for kubeadm itself.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate the apiserver uses to access etcd
Synopsis
Renew the certificate the apiserver uses to access etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for the API server to connect to kubelet
Synopsis
Renew the certificate for the API server to connect to kubelet.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for serving the Kubernetes API
Synopsis
Renew the certificate for serving the Kubernetes API.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate embedded in the kubeconfig file for the controller manager to use
Synopsis
Renew the certificate embedded in the kubeconfig file for the controller manager to use.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for liveness probes to healthcheck etcd
Synopsis
Renew the certificate for liveness probes to healthcheck etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for etcd nodes to communicate with each other
Synopsis
Renew the certificate for etcd nodes to communicate with each other.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for serving etcd
Synopsis
Renew the certificate for serving etcd.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate for the front proxy client
Synopsis
Renew the certificate for the front proxy client.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Renew the certificate embedded in the kubeconfig file for the scheduler manager to use
Synopsis
Renew the certificate embedded in the kubeconfig file for the scheduler manager to use.
Renewals run unconditionally, regardless of certificate expiration date; extra attributes such as SANs will be based on the existing file/certificates, there is no need to resupply them.
Renewal by default tries to use the certificate authority in the local PKI managed by kubeadm; as alternative it is possible to use K8s certificate API for certificate renewal, or as a last option, to generate a CSR request.
After renewal, in order to make changes effective, is required to restart control-plane components and eventually re-distribute the renewed certificate in case the file is used elsewhere.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm certs certificate-key
This command can be used to generate a new control-plane certificate key.
The key can be passed as --certificate-key to kubeadm init
and kubeadm join
to enable the automatic copy of certificates when joining additional control-plane nodes.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm certs generate-csr
This command can be used to generate keys and CSRs for all control-plane certificates and kubeconfig files.
The user can then sign the CSRs with a CA of their choice.
Generates keys and certificate signing requests (CSRs) for all the certificates required to run the control plane. This command also generates partial kubeconfig files with private key data in the "users > user > client-key-data" field, and for each kubeconfig file an accompanying ".csr" file is created.
This command is designed for use in Kubeadm External CA Mode. It generates CSRs which you can then submit to your external certificate authority for signing.
The PEM encoded signed certificates should then be saved alongside the key files, using ".crt" as the file extension, or in the case of kubeconfig files, the PEM encoded signed certificate should be base64 encoded and added to the kubeconfig file in the "users > user > client-certificate-data" field.
kubeadm certs generate-csr [flags]
Examples
# The following command will generate keys and CSRs for all control-plane certificates and kubeconfig files:
kubeadm certs generate-csr --kubeconfig-dir /tmp/etc-k8s --cert-dir /tmp/etc-k8s/pki
kubeadm reset to revert any changes made to this host by kubeadm init or kubeadm join
6.9.1.11 - kubeadm init phase
kubeadm init phase enables you to invoke atomic steps of the bootstrap process.
Hence, you can let kubeadm do some of the work and you can fill in the gaps
if you wish to apply customization.
kubeadm init phase is consistent with the kubeadm init workflow,
and behind the scene both use the same code.
kubeadm init phase preflight
Using this command you can execute preflight checks on a control-plane node.
Write a file with KubeletConfiguration and an environment file with node specific kubelet settings, and then (re)start kubelet.
kubeadm init phase kubelet-start [flags]
Examples
# Writes a dynamic environment file with kubelet flags from a InitConfiguration file.
kubeadm init phase kubelet-start --config config.yaml
Options
--config string
Path to a kubeadm configuration file.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
-h, --help
help for kubelet-start
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase certs
Can be used to create all required certificates by kubeadm.
Use alternative domain for services, e.g. "myorg.internal".
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the certificate for the API server to connect to kubelet
Synopsis
Generate the certificate for the API server to connect to kubelet, and save them into apiserver-kubelet-client.crt and apiserver-kubelet-client.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for front-proxy-client
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the self-signed CA to provision identities for etcd
Synopsis
Generate the self-signed CA to provision identities for etcd, and save them into etcd/ca.crt and etcd/ca.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs etcd-ca [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for etcd-ca
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the certificate for serving etcd
Synopsis
Generate the certificate for serving etcd, and save them into etcd/server.crt and etcd/server.key files.
Default SANs are localhost, 127.0.0.1, 127.0.0.1, ::1
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs etcd-server [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for etcd-server
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the certificate for etcd nodes to communicate with each other
Synopsis
Generate the certificate for etcd nodes to communicate with each other, and save them into etcd/peer.crt and etcd/peer.key files.
Default SANs are localhost, 127.0.0.1, 127.0.0.1, ::1
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs etcd-peer [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for etcd-peer
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the certificate for liveness probes to healthcheck etcd
Synopsis
Generate the certificate for liveness probes to healthcheck etcd, and save them into etcd/healthcheck-client.crt and etcd/healthcheck-client.key files.
If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
-h, --help
help for apiserver-etcd-client
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate a private key for signing service account tokens along with its public key
Synopsis
Generate the private key for signing service account tokens along with its public key, and save them into sa.key and sa.pub files. If both files already exist, kubeadm skips the generation step and existing files will be used.
Alpha Disclaimer: this command is currently alpha.
kubeadm init phase certs sa [flags]
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
-h, --help
help for sa
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase kubeconfig
You can create all required kubeconfig files by calling the all subcommand or call them individually.
Choose a specific Kubernetes version for the control plane.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate a kubeconfig file for the kubelet to use only for cluster bootstrapping purposes
Synopsis
Generate the kubeconfig file for the kubelet to use and save it to kubelet.conf file.
Please note that this should only be used for cluster bootstrapping purposes. After your control plane is up, you should request all kubelet credentials from the CSR API.
kubeadm init phase kubeconfig kubelet [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
Generate all static Pod manifest files necessary to establish the control plane
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase control-plane [flags]
Options
-h, --help
help for control-plane
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate all static Pod manifest files
Synopsis
Generate all static Pod manifest files
kubeadm init phase control-plane all [flags]
Examples
# Generates all static Pod manifest files for control plane components,
# functionally equivalent to what is generated by kubeadm init.
kubeadm init phase control-plane all
# Generates all static Pod manifest files using options read from a configuration file.
kubeadm init phase control-plane all --config config.yaml
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
A set of extra flags to pass to the Controller Manager or override default ones in form of <flagname>=<value>
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
A set of extra flags to pass to the API Server or override default ones in form of <flagname>=<value>
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generates the kube-controller-manager static Pod manifest
Synopsis
Generates the kube-controller-manager static Pod manifest
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Choose a container registry to pull control plane images from
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase etcd [flags]
Options
-h, --help
help for etcd
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the static Pod manifest file for a local, single-node local etcd instance
Synopsis
Generate the static Pod manifest file for a local, single-node local etcd instance
kubeadm init phase etcd local [flags]
Examples
# Generates the static Pod manifest file for etcd, functionally
# equivalent to what is generated by kubeadm init.
kubeadm init phase etcd local
# Generates the static Pod manifest file for etcd using options
# read from a configuration file.
kubeadm init phase etcd local --config config.yaml
Options
--cert-dir string Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
Choose a container registry to pull control plane images from
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase upload-config
You can use this command to upload the kubeadm configuration to your cluster.
Alternatively, you can use kubeadm config.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Upload the kubeadm ClusterConfiguration to a ConfigMap
Synopsis
Upload the kubeadm ClusterConfiguration to a ConfigMap called kubeadm-config in the kube-system namespace. This enables correct configuration of system components and a seamless user experience when upgrading.
Alternatively, you can use kubeadm config.
kubeadm init phase upload-config kubeadm [flags]
Examples
# upload the configuration of your cluster
kubeadm init phase upload-config --config=myConfig.yaml
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Upload the kubelet component config to a ConfigMap
Synopsis
Upload the kubelet configuration extracted from the kubeadm InitConfiguration object to a kubelet-config ConfigMap in the cluster
kubeadm init phase upload-config kubelet [flags]
Examples
# Upload the kubelet configuration from the kubeadm Config file to a ConfigMap in the cluster.
kubeadm init phase upload-config kubelet --config kubeadm.yaml
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase upload-certs
Use the following phase to upload control-plane certificates to the cluster.
By default the certs and encryption key expire after two hours.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-certificate-key-print
Don't print the key used to encrypt the control-plane certificates.
--upload-certs
Upload control-plane certificates to the kubeadm-certs Secret.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase mark-control-plane
Use the following phase to label and taint the node as a control plane node.
# Applies control-plane label and taint to the current node, functionally equivalent to what executed by kubeadm init.
kubeadm init phase mark-control-plane --config config.yaml
# Applies control-plane label and taint to a specific node
kubeadm init phase mark-control-plane --node-name myNode
Options
--config string
Path to a kubeadm configuration file.
-h, --help
help for mark-control-plane
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase bootstrap-token
Use the following phase to configure bootstrap tokens.
Generates bootstrap tokens used to join a node to a cluster
Synopsis
Bootstrap tokens are used for establishing bidirectional trust between a node joining the cluster and a control-plane node.
This command makes all the configurations required to make bootstrap tokens works and then creates an initial token.
kubeadm init phase bootstrap-token [flags]
Examples
# Make all the bootstrap token configurations and create an initial token, functionally
# equivalent to what generated by kubeadm init.
kubeadm init phase bootstrap-token
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--skip-token-print
Skip printing of the default bootstrap token generated by 'kubeadm init'.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm init phase kubelet-finalize
Use the following phase to update settings relevant to the kubelet after TLS
bootstrap. You can use the all subcommand to run all kubelet-finalize
phases.
Install required addons for passing conformance tests
Synopsis
This command is not meant to be run on its own. See list of available subcommands.
kubeadm init phase addon [flags]
Options
-h, --help
help for addon
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Install all the addons
Synopsis
Install all the addons
kubeadm init phase addon all [flags]
Options
--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
Port for the API Server to bind to.
--config string
Path to a kubeadm configuration file.
--control-plane-endpoint string
Specify a stable IP address or DNS name for the control plane.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
Use alternative domain for services, e.g. "myorg.internal".
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Install the CoreDNS addon to a Kubernetes cluster
Synopsis
Install the CoreDNS addon components via the API server. Please note that although the DNS server is deployed, it will not be scheduled until CNI is installed.
kubeadm init phase addon coredns [flags]
Options
--config string
Path to a kubeadm configuration file.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are: PublicKeysECDSA=true|false (ALPHA - default=false) RootlessControlPlane=true|false (ALPHA - default=false) UnversionedKubeletConfigMap=true|false (BETA - default=true)
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--service-cidr string Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--kubernetes-version string Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
For more details on each field in the v1beta3 configuration you can navigate to our
API reference pages.
What's next
kubeadm init to bootstrap a Kubernetes control-plane node
kubeadm join phase enables you to invoke atomic steps of the join process.
Hence, you can let kubeadm do some of the work and you can fill in the gaps
if you wish to apply customization.
kubeadm join phase is consistent with the kubeadm join workflow,
and behind the scene both use the same code.
# Run join pre-flight checks using a config file.
kubeadm join phase preflight --config kubeadm-config.yaml
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for preflight
--ignore-preflight-errors strings
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm join phase control-plane-prepare
Using this phase you can prepare a node for serving a control-plane.
# Prepares the machine for serving a control plane
kubeadm join phase control-plane-prepare all
Options
-h, --help
help for control-plane-prepare
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Prepare the machine for serving a control plane
Synopsis
Prepare the machine for serving a control plane
kubeadm join phase control-plane-prepare all [api-server-endpoint] [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--certificate-key string
Use this key to decrypt the certificate secrets uploaded by init.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for all
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
[EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
Synopsis
[EXPERIMENTAL] Download certificates shared among control-plane nodes from the kubeadm-certs Secret
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for certs
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Generate the kubeconfig for the new control plane components
Synopsis
Generate the kubeconfig for the new control plane components
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--apiserver-bind-port int32 Default: 6443
If the node should host a new control plane instance, the port for the API Server to bind to.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for control-plane
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm join phase kubelet-start
Using this phase you can write the kubelet settings, certificates and (re)start the kubelet.
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
--discovery-file string
For file-based discovery, a file or URL from which to load cluster information.
--discovery-token string
For token-based discovery, the token used to validate cluster information fetched from the API server.
--discovery-token-ca-cert-hash strings
For token-based discovery, validate that the root CA public key matches this hash (format: "<type>:<value>").
--discovery-token-unsafe-skip-ca-verification
For token-based discovery, allow joining without --discovery-token-ca-cert-hash pinning.
-h, --help
help for kubelet-start
--node-name string
Specify the node name.
--tls-bootstrap-token string
Specify the token used to temporarily authenticate with the Kubernetes Control Plane while joining the node.
--token string
Use this token for both discovery-token and tls-bootstrap-token when those values are not provided.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm join phase control-plane-join
Using this phase you can join a node as a control-plane instance.
# Joins a machine as a control plane instance
kubeadm join phase control-plane-join all
Options
-h, --help
help for control-plane-join
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Join a machine as a control plane instance
Synopsis
Join a machine as a control plane instance
kubeadm join phase control-plane-join all [flags]
Options
--apiserver-advertise-address string
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for all
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for etcd
--node-name string
Specify the node name.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Register the new control-plane node into the ClusterStatus maintained in the kubeadm-config ConfigMap (DEPRECATED)
Synopsis
Register the new control-plane node into the ClusterStatus maintained in the kubeadm-config ConfigMap (DEPRECATED)
If the node should host a new control plane instance, the IP address the API Server will advertise it's listening on. If not set the default network interface will be used.
--config string
Path to kubeadm config file.
--control-plane
Create a new control plane instance on this node
-h, --help
help for update-status
--node-name string
Specify the node name.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
# Output a kubeconfig file for an additional user named foo using a kubeadm config file bar
kubeadm kubeconfig user --client-name=foo --config=bar
Options
--client-name string
The name of user. It will be used as the CN if client certificates are created
--config string
Path to a kubeadm configuration file.
-h, --help
help for user
--org strings
The orgnizations of the client certificate. It will be used as the O if client certificates are created
--token string
The token that should be used as the authentication mechanism for this kubeconfig, instead of client certificates
--validity-period duration Default: 8760h0m0s
The validity period of the client certificate. It is an offset from the current time.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
6.9.1.14 - kubeadm reset phase
kubeadm reset phase enables you to invoke atomic steps of the node reset process.
Hence, you can let kubeadm do some of the work and you can fill in the gaps
if you wish to apply customization.
kubeadm reset phase is consistent with the kubeadm reset workflow,
and behind the scene both use the same code.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
kubeadm reset phase cleanup-node
Using this phase you can perform cleanup on this node.
The path to the directory where the certificates are stored. If specified, clean this directory.
--cri-socket string
Path to the CRI socket to connect. If empty kubeadm will try to auto-detect this value; use this option only if you have more than one CRI installed or if you have non-standard CRI socket.
-h, --help
help for cleanup-node
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
What's next
kubeadm init to bootstrap a Kubernetes control-plane node
In v1.15.0, kubeadm introduced preliminary support for kubeadm upgrade node phases.
Phases for other kubeadm upgrade sub-commands such as apply, could be added in the
following releases.
kubeadm upgrade node phase
Using this phase you can choose to execute the separate steps of the upgrade of
secondary control-plane or worker nodes. Please note that kubeadm upgrade apply still has to
be called on a primary control-plane node.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
--patches string
Path to a directory that contains files named "target[suffix][+patchtype].extension". For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of "kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can be one of "strategic", "merge" or "json" and they match the patch formats supported by kubectl. The default "patchtype" is "strategic". "extension" must be either "json" or "yaml". "suffix" is an optional string that can be used to determine which patches are applied first alpha-numerically.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
Upgrade the kubelet configuration for this node
Synopsis
Download the kubelet configuration from the kubelet-config ConfigMap stored in the cluster
kubeadm upgrade node phase kubelet-config [flags]
Options
--dry-run
Do not change any state, just output the actions that would be performed.
The kubeconfig file to use when talking to the cluster. If the flag is not set, a set of standard locations can be searched for an existing kubeconfig file.
Options inherited from parent commands
--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.
What's next
kubeadm init to bootstrap a Kubernetes control-plane node
kubeadm init and kubeadm join together provides a nice user experience for creating a
best-practice but bare Kubernetes cluster from scratch.
However, it might not be obvious how kubeadm does that.
This document provides additional details on what happen under the hood, with the aim of sharing
knowledge on Kubernetes cluster best practices.
Core design principles
The cluster that kubeadm init and kubeadm join set up should be:
Secure: It should adopt latest best-practices like:
enforcing RBAC
using the Node Authorizer
using secure communication between the control plane components
using secure communication between the API server and the kubelets
lock-down the kubelet API
locking down access to the API for system components like the kube-proxy and CoreDNS
locking down what a Bootstrap Token can access
User-friendly: The user should not have to run anything more than a couple of commands:
kubeadm init
export KUBECONFIG=/etc/kubernetes/admin.conf
kubectl apply -f <network-of-choice.yaml>
kubeadm join --token <token> <endpoint>:<port>
Extendable:
It should not favor any particular network provider. Configuring the cluster network is out-of-scope
It should provide the possibility to use a config file for customizing various parameters
Constants and well-known values and paths
In order to reduce complexity and to simplify development of higher level tools that build on top of kubeadm, it uses a
limited set of constant values for well-known paths and file names.
The Kubernetes directory /etc/kubernetes is a constant in the application, since it is clearly the given path
in a majority of cases, and the most intuitive location; other constants paths and file names are:
/etc/kubernetes/manifests as the path where kubelet should look for static Pod manifests.
Names of static Pod manifests are:
etcd.yaml
kube-apiserver.yaml
kube-controller-manager.yaml
kube-scheduler.yaml
/etc/kubernetes/ as the path where kubeconfig files with identities for control plane
components are stored. Names of kubeconfig files are:
kubelet.conf (bootstrap-kubelet.conf during TLS bootstrap)
controller-manager.conf
scheduler.conf
admin.conf for the cluster admin and kubeadm itself
Names of certificates and key files :
ca.crt, ca.key for the Kubernetes certificate authority
apiserver.crt, apiserver.key for the API server certificate
apiserver-kubelet-client.crt, apiserver-kubelet-client.key for the client certificate used
by the API server to connect to the kubelets securely
sa.pub, sa.key for the key used by the controller manager when signing ServiceAccount
front-proxy-ca.crt, front-proxy-ca.key for the front proxy certificate authority
front-proxy-client.crt, front-proxy-client.key for the front proxy client
kubeadm init workflow internal design
The kubeadm initinternal workflow
consists of a sequence of atomic work tasks to perform,
as described in kubeadm init.
The kubeadm init phase command allows
users to invoke each task individually, and ultimately offers a reusable and composable
API/toolbox that can be used by other Kubernetes bootstrap tools, by any IT automation tool or by
an advanced user for creating custom clusters.
Preflight checks
Kubeadm executes a set of preflight checks before starting the init, with the aim to verify
preconditions and avoid common cluster startup problems.
The user can skip specific preflight checks or all of them with the --ignore-preflight-errors option.
[warning] If the Kubernetes version to use (specified with the --kubernetes-version flag) is
at least one minor version higher than the kubeadm CLI version.
Kubernetes system requirements:
if running on linux:
[error] if Kernel is older than the minimum required version
[error] if required cgroups subsystem aren't set up
[error] if the CRI endpoint does not answer
[error] if user is not root
[error] if the machine hostname is not a valid DNS subdomain
[warning] if the host name cannot be reached via network lookup
[error] if kubelet version is lower that the minimum kubelet version supported by kubeadm (current minor -1)
[error] if kubelet version is at least one minor higher than the required controlplane version (unsupported version skew)
[warning] if kubelet service does not exist or if it is disabled
[warning] if firewalld is active
[error] if API server bindPort or ports 10250/10251/10252 are used
[Error] if /etc/kubernetes/manifest folder already exists and it is not empty
[Error] if /proc/sys/net/bridge/bridge-nf-call-iptables file does not exist/does not contain 1
[Error] if advertise address is ipv6 and /proc/sys/net/bridge/bridge-nf-call-ip6tables does not exist/does not contain 1.
[Error] if swap is on
[Error] if conntrack, ip, iptables, mount, nsenter commands are not present in the command path
[warning] if ebtables, ethtool, socat, tc, touch, crictl commands are not present in the command path
[warning] if extra arg flags for API server, controller manager, scheduler contains some invalid options
[warning] if connection to https://API.AdvertiseAddress:API.BindPort goes through proxy
[warning] if connection to services subnet goes through proxy (only first address checked)
[warning] if connection to Pods subnet goes through proxy (only first address checked)
If external etcd is provided:
[Error] if etcd version is older than the minimum required version
[Error] if etcd certificates or keys are specified, but not provided
If external etcd is NOT provided (and thus local etcd will be installed):
[Error] if ports 2379 is used
[Error] if Etcd.DataDir folder already exists and it is not empty
Kubeadm generates certificate and private key pairs for different purposes:
A self signed certificate authority for the Kubernetes cluster saved into ca.crt file and
ca.key private key file
A serving certificate for the API server, generated using ca.crt as the CA, and saved into
apiserver.crt file with its private key apiserver.key. This certificate should contain
following alternative names:
The Kubernetes service's internal clusterIP (the first address in the services CIDR, e.g.
10.96.0.1 if service subnet is 10.96.0.0/12)
Kubernetes DNS names, e.g. kubernetes.default.svc.cluster.local if --service-dns-domain
flag value is cluster.local, plus default DNS names kubernetes.default.svc,
kubernetes.default, kubernetes
The node-name
The --apiserver-advertise-address
Additional alternative names specified by the user
A client certificate for the API server to connect to the kubelets securely, generated using
ca.crt as the CA and saved into apiserver-kubelet-client.crt file with its private key
apiserver-kubelet-client.key.
This certificate should be in the system:masters organization
A private key for signing ServiceAccount Tokens saved into sa.key file along with its public key sa.pub
A certificate authority for the front proxy saved into front-proxy-ca.crt file with its key
front-proxy-ca.key
A client cert for the front proxy client, generate using front-proxy-ca.crt as the CA and
saved into front-proxy-client.crt file with its private keyfront-proxy-client.key
Certificates are stored by default in /etc/kubernetes/pki, but this directory is configurable
using the --cert-dir flag.
Please note that:
If a given certificate and private key pair both exist, and its content is evaluated compliant with the above specs, the existing files will
be used and the generation phase for the given certificate skipped. This means the user can, for example, copy an existing CA to
/etc/kubernetes/pki/ca.{crt,key}, and then kubeadm will use those files for signing the rest of the certs.
See also using custom certificates
Only for the CA, it is possible to provide the ca.crt file but not the ca.key file, if all other certificates and kubeconfig files
already are in place kubeadm recognize this condition and activates the ExternalCA , which also implies the csrsignercontroller in
controller-manager won't be started
If kubeadm is running in external CA mode;
all the certificates must be provided by the user, because kubeadm cannot generate them by itself
In case of kubeadm is executed in the --dry-run mode, certificates files are written in a temporary folder
Generate kubeconfig files for control plane components
Kubeadm generates kubeconfig files with identities for control plane components:
A kubeconfig file for the kubelet to use during TLS bootstrap -
/etc/kubernetes/bootstrap-kubelet.conf. Inside this file there is a bootstrap-token or embedded
client certificates for authenticating this node with the cluster.
This client cert should:
Be in the system:nodes organization, as required by the
Node Authorization module
Have the Common Name (CN) system:node:<hostname-lowercased>
A kubeconfig file for controller-manager, /etc/kubernetes/controller-manager.conf; inside this
file is embedded a client certificate with controller-manager identity. This client cert should
have the CN system:kube-controller-manager, as defined by default
RBAC core components roles
A kubeconfig file for scheduler, /etc/kubernetes/scheduler.conf; inside this file is embedded
a client certificate with scheduler identity.
This client cert should have the CN system:kube-scheduler, as defined by default
RBAC core components roles
Additionally, a kubeconfig file for kubeadm itself and the admin is generated and saved into the
/etc/kubernetes/admin.conf file. The "admin" here is defined as the actual person(s) that is
administering the cluster and wants to have full control (root) over the cluster. The
embedded client certificate for admin should be in the system:masters organization, as defined
by default RBAC user facing role bindings.
It should also include a CN. Kubeadm uses the kubernetes-admin CN.
Please note that:
ca.crt certificate is embedded in all the kubeconfig files.
If a given kubeconfig file exists, and its content is evaluated compliant with the above specs,
the existing file will be used and the generation phase for the given kubeconfig skipped
If kubeadm is running in ExternalCA mode,
all the required kubeconfig must be provided by the user as well, because kubeadm cannot
generate any of them by itself
In case of kubeadm is executed in the --dry-run mode, kubeconfig files are written in a temporary folder
Generate static Pod manifests for control plane components
Kubeadm writes static Pod manifest files for control plane components to
/etc/kubernetes/manifests. The kubelet watches this directory for Pods to create on startup.
Static Pod manifest share a set of common properties:
All static Pods are deployed on kube-system namespace
All static Pods get tier:control-plane and component:{component-name} labels
All static Pods use the system-node-critical priority class
hostNetwork: true is set on all static Pods to allow control plane startup before a network is
configured; as a consequence:
The address that the controller-manager and the scheduler use to refer the API server is 127.0.0.1
If using a local etcd server, etcd-servers address will be set to 127.0.0.1:2379
Leader election is enabled for both the controller-manager and the scheduler
Controller-manager and the scheduler will reference kubeconfig files with their respective, unique identities
The static Pod manifest for the API server is affected by following parameters provided by the users:
The apiserver-advertise-address and apiserver-bind-port to bind to; if not provided, those
value defaults to the IP address of the default network interface on the machine and port 6443
The service-cluster-ip-range to use for services
If an external etcd server is specified, the etcd-servers address and related TLS settings
(etcd-cafile, etcd-certfile, etcd-keyfile);
if an external etcd server is not be provided, a local etcd will be used (via host network)
If a cloud provider is specified, the corresponding --cloud-provider is configured, together
with the --cloud-config path if such file exists (this is experimental, alpha and will be
removed in a future version)
Other API server flags that are set unconditionally are:
--insecure-port=0 to avoid insecure connections to the api server
--enable-bootstrap-token-auth=true to enable the BootstrapTokenAuthenticator authentication module.
See TLS Bootstrapping for more details
--allow-privileged to true (required e.g. by kube proxy)
--requestheader-client-ca-file to front-proxy-ca.crt
PersistentVolumeLabel
attaches region or zone labels to PersistentVolumes as defined by the cloud provider (This
admission controller is deprecated and will be removed in a future version.
It is not deployed by kubeadm by default with v1.9 onwards when not explicitly opting into
using gce or aws as cloud providers)
DefaultStorageClass
to enforce default storage class on PersistentVolumeClaim objects
NodeRestriction
to limit what a kubelet can modify (e.g. only pods on this node)
--kubelet-preferred-address-types to InternalIP,ExternalIP,Hostname; this makes kubectl logs and other API server-kubelet communication work in environments where the hostnames of the
nodes aren't resolvable
Flags for using certificates generated in previous steps:
--client-ca-file to ca.crt
--tls-cert-file to apiserver.crt
--tls-private-key-file to apiserver.key
--kubelet-client-certificate to apiserver-kubelet-client.crt
--kubelet-client-key to apiserver-kubelet-client.key
The static Pod manifest for the controller manager is affected by following parameters provided by
the users:
If kubeadm is invoked specifying a --pod-network-cidr, the subnet manager feature required for
some CNI network plugins is enabled by setting:
--allocate-node-cidrs=true
--cluster-cidr and --node-cidr-mask-size flags according to the given CIDR
If a cloud provider is specified, the corresponding --cloud-provider is specified, together
with the --cloud-config path if such configuration file exists (this is experimental, alpha
and will be removed in a future version)
Other flags that are set unconditionally are:
--controllers enabling all the default controllers plus BootstrapSigner and TokenCleaner
controllers for TLS bootstrap. See TLS Bootstrapping
for more details
--use-service-account-credentials to true
Flags for using certificates generated in previous steps:
--root-ca-file to ca.crt
--cluster-signing-cert-file to ca.crt, if External CA mode is disabled, otherwise to ""
--cluster-signing-key-file to ca.key, if External CA mode is disabled, otherwise to ""
--service-account-private-key-file to sa.key
Scheduler
The static Pod manifest for the scheduler is not affected by parameters provided by the users.
Generate static Pod manifest for local etcd
If you specified an external etcd this step will be skipped, otherwise kubeadm generates a
static Pod manifest file for creating a local etcd instance running in a Pod with following attributes:
listen on localhost:2379 and use HostNetwork=true
make a hostPath mount out from the dataDir to the host's filesystem
Any extra flags specified by the user
Please note that:
The etcd container image will be pulled from registry.gcr.io by default. See
using custom images
for customizing the image repository.
If you run kubeadm in --dry-run mode, the etcd static Pod manifest is written
into a temporary folder.
You can directly invoke static Pod manifest generation for local etcd, using the
kubeadm init phase etcd local
command.
Wait for the control plane to come up
kubeadm waits (upto 4m0s) until localhost:6443/healthz (kube-apiserver liveness) returns ok.
However in order to detect deadlock conditions, kubeadm fails fast if localhost:10255/healthz
(kubelet liveness) or localhost:10255/healthz/syncloop (kubelet readiness) don't return ok
within 40s and 60s respectively.
kubeadm relies on the kubelet to pull the control plane images and run them properly as static Pods.
After the control plane is up, kubeadm completes the tasks described in following paragraphs.
Save the kubeadm ClusterConfiguration in a ConfigMap for later reference
kubeadm saves the configuration passed to kubeadm init in a ConfigMap named kubeadm-config
under kube-system namespace.
This will ensure that kubeadm actions executed in future (e.g kubeadm upgrade) will be able to
determine the actual/current cluster state and make new decisions based on that data.
Please note that:
Before saving the ClusterConfiguration, sensitive information like the token is stripped from the configuration
kubeadm init ensures that everything is properly configured for this process, and this includes
following steps as well as setting API server and controller flags as already described in
previous paragraphs.
Please note that:
TLS bootstrapping for nodes can be configured with the command
kubeadm init phase bootstrap-token,
executing all the configuration steps described in following paragraphs;
alternatively, each step can be invoked individually
Create a bootstrap token
kubeadm init create a first bootstrap token, either generated automatically or provided by the
user with the --token flag; as documented in bootstrap token specification, token should be
saved as secrets with name bootstrap-token-<token-id> under kube-system namespace. Please
note that:
The default token created by kubeadm init will be used to validate temporary user during TLS
bootstrap process; those users will be member of
system:bootstrappers:kubeadm:default-node-token group
The token has a limited validity, default 24 hours (the interval may be changed with the —token-ttl flag)
Additional tokens can be created with the kubeadm token
command, that provide as well other useful functions for token management.
Allow joining nodes to call CSR API
Kubeadm ensures that users in system:bootstrappers:kubeadm:default-node-token group are able to
access the certificate signing API.
This is implemented by creating a ClusterRoleBinding named kubeadm:kubelet-bootstrap between the
group above and the default RBAC role system:node-bootstrapper.
Set up auto approval for new bootstrap tokens
Kubeadm ensures that the Bootstrap Token will get its CSR request automatically approved by the
csrapprover controller.
This is implemented by creating ClusterRoleBinding named kubeadm:node-autoapprove-bootstrap
between the system:bootstrappers:kubeadm:default-node-token group and the default role
system:certificates.k8s.io:certificatesigningrequests:nodeclient.
The role system:certificates.k8s.io:certificatesigningrequests:nodeclient should be created as
well, granting POST permission to
/apis/certificates.k8s.io/certificatesigningrequests/nodeclient.
Set up nodes certificate rotation with auto approval
Kubeadm ensures that certificate rotation is enabled for nodes, and that new certificate request
for nodes will get its CSR request automatically approved by the csrapprover controller.
This is implemented by creating ClusterRoleBinding named
kubeadm:node-autoapprove-certificate-rotation between the system:nodes group and the default
role system:certificates.k8s.io:certificatesigningrequests:selfnodeclient.
Create the public cluster-info ConfigMap
This phase creates the cluster-info ConfigMap in the kube-public namespace.
Additionally it creates a Role and a RoleBinding granting access to the ConfigMap for
unauthenticated users (i.e. users in RBAC group system:unauthenticated).
Please note that:
The access to the cluster-info ConfigMap is not rate-limited. This may or may not be a
problem if you expose your cluster's API server to the internet; worst-case scenario here is a
DoS attack where an attacker uses all the in-flight requests the kube-apiserver can handle to
serving the cluster-info ConfigMap.
Install addons
Kubeadm installs the internal DNS server and the kube-proxy addon components via the API server.
Please note that:
A ServiceAccount for kube-proxy is created in the kube-system namespace; then kube-proxy is
deployed as a DaemonSet:
The credentials (ca.crt and token) to the control plane come from the ServiceAccount
The location (URL) of the API server comes from a ConfigMap
The kube-proxy ServiceAccount is bound to the privileges in the system:node-proxier ClusterRole
DNS
The CoreDNS service is named kube-dns. This is done to prevent any interruption
in service when the user is switching the cluster DNS from kube-dns to CoreDNS
the --config method described here.
A ServiceAccount for CoreDNS is created in the kube-system namespace.
The coredns ServiceAccount is bound to the privileges in the system:coredns ClusterRole
In Kubernetes version 1.21, support for using kube-dns with kubeadm was removed.
You can use CoreDNS with kubeadm even when the related Service is named kube-dns.
kubeadm join phases internal design
Similarly to kubeadm init, also kubeadm join internal workflow consists of a sequence of
atomic work tasks to perform.
This is split into discovery (having the Node trust the Kubernetes Master) and TLS bootstrap
(having the Kubernetes Master trust the Node).
kubeadm executes a set of preflight checks before starting the join, with the aim to verify
preconditions and avoid common cluster startup problems.
Please note that:
kubeadm join preflight checks are basically a subset kubeadm init preflight checks
Starting from 1.24, kubeadm uses crictl to communicate to all known CRI endpoints.
Starting from 1.9, kubeadm provides support for joining nodes running on Windows; in that case,
linux specific controls are skipped.
In any case the user can skip specific preflight checks (or eventually all preflight checks)
with the --ignore-preflight-errors option.
Discovery cluster-info
There are 2 main schemes for discovery. The first is to use a shared token along with the IP
address of the API server.
The second is to provide a file (that is a subset of the standard kubeconfig file).
Shared token discovery
If kubeadm join is invoked with --discovery-token, token discovery is used; in this case the
node basically retrieves the cluster CA certificates from the cluster-info ConfigMap in the
kube-public namespace.
In order to prevent "man in the middle" attacks, several steps are taken:
First, the CA certificate is retrieved via insecure connection (this is possible because
kubeadm init granted access to cluster-info users for system:unauthenticated )
Then the CA certificate goes trough following validation steps:
Basic validation: using the token ID against a JWT signature
Pub key validation: using provided --discovery-token-ca-cert-hash. This value is available
in the output of kubeadm init or can be calculated using standard tools (the hash is
calculated over the bytes of the Subject Public Key Info (SPKI) object as in RFC7469). The
--discovery-token-ca-cert-hash flag may be repeated multiple times to allow more than one public key.
As a additional validation, the CA certificate is retrieved via secure connection and then
compared with the CA retrieved initially
Please note that:
Pub key validation can be skipped passing --discovery-token-unsafe-skip-ca-verification flag;
This weakens the kubeadm security model since others can potentially impersonate the Kubernetes Master.
File/https discovery
If kubeadm join is invoked with --discovery-file, file discovery is used; this file can be a
local file or downloaded via an HTTPS URL; in case of HTTPS, the host installed CA bundle is used
to verify the connection.
With file discovery, the cluster CA certificates is provided into the file itself; in fact, the
discovery file is a kubeconfig file with only server and certificate-authority-data attributes
set, as described in kubeadm join
reference doc; when the connection with the cluster is established, kubeadm try to access the
cluster-info ConfigMap, and if available, uses it.
TLS Bootstrap
Once the cluster info are known, the file bootstrap-kubelet.conf is written, thus allowing
kubelet to do TLS Bootstrapping.
The TLS bootstrap mechanism uses the shared token to temporarily authenticate with the Kubernetes
API server to submit a certificate signing request (CSR) for a locally created key pair.
The request is then automatically approved and the operation completes saving ca.crt file and
kubelet.conf file to be used by kubelet for joining the cluster, whilebootstrap-kubelet.conf
is deleted.
Please note that:
The temporary authentication is validated against the token saved during the kubeadm init
process (or with additional tokens created with kubeadm token)
The temporary authentication resolve to a user member of
system:bootstrappers:kubeadm:default-node-token group which was granted access to CSR api
during the kubeadm init process
The automatic CSR approval is managed by the csrapprover controller, according with
configuration done the kubeadm init process
6.10 - Command line tool (kubectl)
Kubernetes provides a command line tool for communicating with a Kubernetes cluster's
control plane,
using the Kubernetes API.
This tool is named kubectl.
For configuration, kubectl looks for a file named config in the $HOME/.kube directory.
You can specify other kubeconfig
files by setting the KUBECONFIG environment variable or by setting the
--kubeconfig flag.
This overview covers kubectl syntax, describes the command operations, and provides common examples.
For details about each command, including all the supported flags and subcommands, see the
kubectl reference documentation.
For installation instructions, see Installing kubectl;
for a quick guide, see the cheat sheet.
If you're used to using the docker command-line tool, kubectl for Docker Users explains some equivalent commands for Kubernetes.
Syntax
Use the following syntax to run kubectl commands from your terminal window:
kubectl [command][TYPE][NAME][flags]
where command, TYPE, NAME, and flags are:
command: Specifies the operation that you want to perform on one or more resources,
for example create, get, describe, delete.
TYPE: Specifies the resource type. Resource types are case-insensitive and
you can specify the singular, plural, or abbreviated forms.
For example, the following commands produce the same output:
kubectl get pod pod1
kubectl get pods pod1
kubectl get po pod1
NAME: Specifies the name of the resource. Names are case-sensitive. If the name is omitted, details for all resources are displayed, for example kubectl get pods.
When performing an operation on multiple resources, you can specify each resource by type and name or specify one or more files:
To specify resources by type and name:
To group resources if they are all the same type: TYPE1 name1 name2 name<#>.
Example: kubectl get pod example-pod1 example-pod2
To specify multiple resource types individually: TYPE1/name1 TYPE1/name2 TYPE2/name3 TYPE<#>/name<#>.
Example: kubectl get pod/example-pod1 replicationcontroller/example-rc1
To specify resources with one or more files: -f file1 -f file2 -f file<#>
Use YAML rather than JSON since YAML tends to be more user-friendly, especially for configuration files.
Example: kubectl get -f ./pod.yaml
flags: Specifies optional flags. For example, you can use the -s or --server flags to specify the address and port of the Kubernetes API server.
Caution: Flags that you specify from the command line override default values and any corresponding environment variables.
If you need help, run kubectl help from the terminal window.
In-cluster authentication and namespace overrides
By default kubectl will first determine if it is running within a pod, and thus in a cluster. It starts by checking for the KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT environment variables and the existence of a service account token file at /var/run/secrets/kubernetes.io/serviceaccount/token. If all three are found in-cluster authentication is assumed.
To maintain backwards compatibility, if the POD_NAMESPACE environment variable is set during in-cluster authentication it will override the default namespace from the service account token. Any manifests or tools relying on namespace defaulting will be affected by this.
POD_NAMESPACE environment variable
If the POD_NAMESPACE environment variable is set, cli operations on namespaced resources will default to the variable value. For example, if the variable is set to seattle, kubectl get pods would return pods in the seattle namespace. This is because pods are a namespaced resource, and no namespace was provided in the command. Review the output of kubectl api-resources to determine if a resource is namespaced.
Explicit use of --namespace <value> overrides this behavior.
How kubectl handles ServiceAccount tokens
If:
there is Kubernetes service account token file mounted at
/var/run/secrets/kubernetes.io/serviceaccount/token, and
the KUBERNETES_SERVICE_HOST environment variable is set, and
the KUBERNETES_SERVICE_PORT environment variable is set, and
you don't explicitly specify a namespace on the kubectl command line
then kubectl assumes it is running in your cluster. The kubectl tool looks up the
namespace of that ServiceAccount (this is the same as the namespace of the Pod)
and acts against that namespace. This is different from what happens outside of a
cluster; when kubectl runs outside a cluster and you don't specify a namespace,
the kubectl command acts against the default namespace.
Operations
The following table includes short descriptions and the general syntax for all of the kubectl operations:
Operation
Syntax
Description
alpha
kubectl alpha SUBCOMMAND [flags]
List the available commands that correspond to alpha features, which are not enabled in Kubernetes clusters by default.
annotate
kubectl annotate (-f FILENAME | TYPE NAME | TYPE/NAME) KEY_1=VAL_1 ... KEY_N=VAL_N [--overwrite] [--all] [--resource-version=version] [flags]
Add or update the annotations of one or more resources.
api-resources
kubectl api-resources [flags]
List the API resources that are available.
api-versions
kubectl api-versions [flags]
List the API versions that are available.
apply
kubectl apply -f FILENAME [flags]
Apply a configuration change to a resource from a file or stdin.
attach
kubectl attach POD -c CONTAINER [-i] [-t] [flags]
Attach to a running container either to view the output stream or interact with the container (stdin).
auth
kubectl auth [flags] [options]
Inspect authorization.
autoscale
kubectl autoscale (-f FILENAME | TYPE NAME | TYPE/NAME) [--min=MINPODS] --max=MAXPODS [--cpu-percent=CPU] [flags]
Automatically scale the set of pods that are managed by a replication controller.
certificate
kubectl certificate SUBCOMMAND [options]
Modify certificate resources.
cluster-info
kubectl cluster-info [flags]
Display endpoint information about the master and services in the cluster.
completion
kubectl completion SHELL [options]
Output shell completion code for the specified shell (bash or zsh).
config
kubectl config SUBCOMMAND [flags]
Modifies kubeconfig files. See the individual subcommands for details.
convert
kubectl convert -f FILENAME [options]
Convert config files between different API versions. Both YAML and JSON formats are accepted. Note - requires kubectl-convert plugin to be installed.
Get documentation of various resources. For instance pods, nodes, services, etc.
expose
kubectl expose (-f FILENAME | TYPE NAME | TYPE/NAME) [--port=port] [--protocol=TCP|UDP] [--target-port=number-or-name] [--name=name] [--external-ip=external-ip-of-service] [--type=type] [flags]
Expose a replication controller, service, or pod as a new Kubernetes service.
get
kubectl get (-f FILENAME | TYPE [NAME | /NAME | -l label]) [--watch] [--sort-by=FIELD] [[-o | --output]=OUTPUT_FORMAT] [flags]
List one or more resources.
kustomize
kubectl kustomize <dir> [flags] [options]
List a set of API resources generated from instructions in a kustomization.yaml file. The argument must be the path to the directory containing the file, or a git repository URL with a path suffix specifying same with respect to the repository root.
label
kubectl label (-f FILENAME | TYPE NAME | TYPE/NAME) KEY_1=VAL_1 ... KEY_N=VAL_N [--overwrite] [--all] [--resource-version=version] [flags]
Add or update the labels of one or more resources.
logs
kubectl logs POD [-c CONTAINER] [--follow] [flags]
Print the logs for a container in a pod.
options
kubectl options
List of global command-line options, which apply to all commands.
patch
kubectl patch (-f FILENAME | TYPE NAME | TYPE/NAME) --patch PATCH [flags]
Update one or more fields of a resource by using the strategic merge patch process.
plugin
kubectl plugin [flags] [options]
Provides utilities for interacting with plugins.
port-forward
kubectl port-forward POD [LOCAL_PORT:]REMOTE_PORT [...[LOCAL_PORT_N:]REMOTE_PORT_N] [flags]
Experimental: Wait for a specific condition on one or many resources.
To learn more about command operations, see the kubectl reference documentation.
Resource types
The following table includes a list of all the supported resource types and their abbreviated aliases.
(This output can be retrieved from kubectl api-resources, and was accurate as of Kubernetes 1.19.1.)
NAME
SHORTNAMES
APIGROUP
NAMESPACED
KIND
bindings
true
Binding
componentstatuses
cs
false
ComponentStatus
configmaps
cm
true
ConfigMap
endpoints
ep
true
Endpoints
events
ev
true
Event
limitranges
limits
true
LimitRange
namespaces
ns
false
Namespace
nodes
no
false
Node
persistentvolumeclaims
pvc
true
PersistentVolumeClaim
persistentvolumes
pv
false
PersistentVolume
pods
po
true
Pod
podtemplates
true
PodTemplate
replicationcontrollers
rc
true
ReplicationController
resourcequotas
quota
true
ResourceQuota
secrets
true
Secret
serviceaccounts
sa
true
ServiceAccount
services
svc
true
Service
mutatingwebhookconfigurations
admissionregistration.k8s.io
false
MutatingWebhookConfiguration
validatingwebhookconfigurations
admissionregistration.k8s.io
false
ValidatingWebhookConfiguration
customresourcedefinitions
crd,crds
apiextensions.k8s.io
false
CustomResourceDefinition
apiservices
apiregistration.k8s.io
false
APIService
controllerrevisions
apps
true
ControllerRevision
daemonsets
ds
apps
true
DaemonSet
deployments
deploy
apps
true
Deployment
replicasets
rs
apps
true
ReplicaSet
statefulsets
sts
apps
true
StatefulSet
tokenreviews
authentication.k8s.io
false
TokenReview
localsubjectaccessreviews
authorization.k8s.io
true
LocalSubjectAccessReview
selfsubjectaccessreviews
authorization.k8s.io
false
SelfSubjectAccessReview
selfsubjectrulesreviews
authorization.k8s.io
false
SelfSubjectRulesReview
subjectaccessreviews
authorization.k8s.io
false
SubjectAccessReview
horizontalpodautoscalers
hpa
autoscaling
true
HorizontalPodAutoscaler
cronjobs
cj
batch
true
CronJob
jobs
batch
true
Job
certificatesigningrequests
csr
certificates.k8s.io
false
CertificateSigningRequest
leases
coordination.k8s.io
true
Lease
endpointslices
discovery.k8s.io
true
EndpointSlice
events
ev
events.k8s.io
true
Event
ingresses
ing
extensions
true
Ingress
flowschemas
flowcontrol.apiserver.k8s.io
false
FlowSchema
prioritylevelconfigurations
flowcontrol.apiserver.k8s.io
false
PriorityLevelConfiguration
ingressclasses
networking.k8s.io
false
IngressClass
ingresses
ing
networking.k8s.io
true
Ingress
networkpolicies
netpol
networking.k8s.io
true
NetworkPolicy
runtimeclasses
node.k8s.io
false
RuntimeClass
poddisruptionbudgets
pdb
policy
true
PodDisruptionBudget
podsecuritypolicies
psp
policy
false
PodSecurityPolicy
clusterrolebindings
rbac.authorization.k8s.io
false
ClusterRoleBinding
clusterroles
rbac.authorization.k8s.io
false
ClusterRole
rolebindings
rbac.authorization.k8s.io
true
RoleBinding
roles
rbac.authorization.k8s.io
true
Role
priorityclasses
pc
scheduling.k8s.io
false
PriorityClass
csidrivers
storage.k8s.io
false
CSIDriver
csinodes
storage.k8s.io
false
CSINode
storageclasses
sc
storage.k8s.io
false
StorageClass
volumeattachments
storage.k8s.io
false
VolumeAttachment
Output options
Use the following sections for information about how you can format or sort the output of certain commands. For details about which commands support the various output options, see the kubectl reference documentation.
Formatting output
The default output format for all kubectl commands is the human readable plain-text format. To output details to your terminal window in a specific format, you can add either the -o or --output flags to a supported kubectl command.
Syntax
kubectl [command][TYPE][NAME] -o <output_format>
Depending on the kubectl operation, the following output formats are supported:
Output format
Description
-o custom-columns=<spec>
Print a table using a comma separated list of custom columns.
-o custom-columns-file=<filename>
Print a table using the custom columns template in the <filename> file.
-o json
Output a JSON formatted API object.
-o jsonpath=<template>
Print the fields defined in a jsonpath expression.
-o jsonpath-file=<filename>
Print the fields defined by the jsonpath expression in the <filename> file.
-o name
Print only the resource name and nothing else.
-o wide
Output in the plain-text format with any additional information. For pods, the node name is included.
-o yaml
Output a YAML formatted API object.
Example
In this example, the following command outputs the details for a single pod as a YAML formatted object:
kubectl get pod web-pod-13je7 -o yaml
Remember: See the kubectl reference documentation
for details about which output format is supported by each command.
Custom columns
To define custom columns and output only the details that you want into a table, you can use the custom-columns option.
You can choose to define the custom columns inline or use a template file: -o custom-columns=<spec> or -o custom-columns-file=<filename>.
Examples
Inline:
kubectl get pods <pod-name> -o custom-columns=NAME:.metadata.name,RSRC:.metadata.resourceVersion
Template file:
kubectl get pods <pod-name> -o custom-columns-file=template.txt
where the template.txt file contains:
NAME RSRC
metadata.name metadata.resourceVersion
The result of running either command is similar to:
NAME RSRC
submit-queue 610995
Server-side columns
kubectl supports receiving specific column information from the server about objects.
This means that for any given resource, the server will return columns and rows relevant to that resource, for the client to print.
This allows for consistent human-readable output across clients used against the same cluster, by having the server encapsulate the details of printing.
This feature is enabled by default. To disable it, add the
--server-print=false flag to the kubectl get command.
Examples
To print information about the status of a pod, use a command like the following:
kubectl get pods <pod-name> --server-print=false
The output is similar to:
NAME AGE
pod-name 1m
Sorting list objects
To output objects to a sorted list in your terminal window, you can add the --sort-by flag to a supported kubectl command. Sort your objects by specifying any numeric or string field with the --sort-by flag. To specify a field, use a jsonpath expression.
Use the following set of examples to help you familiarize yourself with running the commonly used kubectl operations:
kubectl apply - Apply or Update a resource from a file or stdin.
# Create a service using the definition in example-service.yaml.kubectl apply -f example-service.yaml
# Create a replication controller using the definition in example-controller.yaml.kubectl apply -f example-controller.yaml
# Create the objects that are defined in any .yaml, .yml, or .json file within the <directory> directory.kubectl apply -f <directory>
kubectl get - List one or more resources.
# List all pods in plain-text output format.kubectl get pods
# List all pods in plain-text output format and include additional information (such as node name).kubectl get pods -o wide
# List the replication controller with the specified name in plain-text output format. Tip: You can shorten and replace the 'replicationcontroller' resource type with the alias 'rc'.kubectl get replicationcontroller <rc-name>
# List all replication controllers and services together in plain-text output format.kubectl get rc,services
# List all daemon sets in plain-text output format.kubectl get ds
# List all pods running on node server01kubectl get pods --field-selector=spec.nodeName=server01
kubectl describe - Display detailed state of one or more resources, including the uninitialized ones by default.
# Display the details of the node with name <node-name>.kubectl describe nodes <node-name>
# Display the details of the pod with name <pod-name>.kubectl describe pods/<pod-name>
# Display the details of all the pods that are managed by the replication controller named <rc-name>.# Remember: Any pods that are created by the replication controller get prefixed with the name of the replication controller.kubectl describe pods <rc-name>
# Describe all podskubectl describe pods
Note: The kubectl get command is usually used for retrieving one or more
resources of the same resource type. It features a rich set of flags that allows
you to customize the output format using the -o or --output flag, for example.
You can specify the -w or --watch flag to start watching updates to a particular
object. The kubectl describe command is more focused on describing the many
related aspects of a specified resource. It may invoke several API calls to the
API server to build a view for the user. For example, the kubectl describe node
command retrieves not only the information about the node, but also a summary of
the pods running on it, the events generated for the node etc.
kubectl delete - Delete resources either from a file, stdin, or specifying label selectors, names, resource selectors, or resources.
# Delete a pod using the type and name specified in the pod.yaml file.kubectl delete -f pod.yaml
# Delete all the pods and services that have the label '<label-key>=<label-value>'.kubectl delete pods,services -l <label-key>=<label-value>
# Delete all pods, including uninitialized ones.kubectl delete pods --all
kubectl exec - Execute a command against a container in a pod.
# Get output from running 'date' from pod <pod-name>. By default, output is from the first container.kubectl exec <pod-name> -- date
# Get output from running 'date' in container <container-name> of pod <pod-name>.kubectl exec <pod-name> -c <container-name> -- date
# Get an interactive TTY and run /bin/bash from pod <pod-name>. By default, output is from the first container.kubectl exec -ti <pod-name> -- /bin/bash
kubectl logs - Print the logs for a container in a pod.
# Return a snapshot of the logs from pod <pod-name>.kubectl logs <pod-name>
# Start streaming the logs from pod <pod-name>. This is similar to the 'tail -f' Linux command.kubectl logs -f <pod-name>
kubectl diff - View a diff of the proposed updates to a cluster.
# Diff resources included in "pod.json".kubectl diff -f pod.json
# Diff file read from stdin.cat service.yaml | kubectl diff -f -
Examples: Creating and using plugins
Use the following set of examples to help you familiarize yourself with writing and using kubectl plugins:
# create a simple plugin in any language and name the resulting executable file# so that it begins with the prefix "kubectl-"cat ./kubectl-hello
#!/bin/sh
# this plugin prints the words "hello world"echo"hello world"
With a plugin written, let's make it executable:
chmod a+x ./kubectl-hello
# and move it to a location in our PATHsudo mv ./kubectl-hello /usr/local/bin
sudo chown root:root /usr/local/bin
# You have now created and "installed" a kubectl plugin.# You can begin using this plugin by invoking it from kubectl as if it were a regular commandkubectl hello
hello world
# You can "uninstall" a plugin, by removing it from the folder in your# $PATH where you placed itsudo rm /usr/local/bin/kubectl-hello
In order to view all of the plugins that are available to kubectl, use
the kubectl plugin list subcommand:
kubectl plugin list
The output is similar to:
The following kubectl-compatible plugins are available:
/usr/local/bin/kubectl-hello
/usr/local/bin/kubectl-foo
/usr/local/bin/kubectl-bar
kubectl plugin list also warns you about plugins that are not
executable, or that are shadowed by other plugins; for example:
sudo chmod -x /usr/local/bin/kubectl-foo # remove execute permissionkubectl plugin list
The following kubectl-compatible plugins are available:
/usr/local/bin/kubectl-hello
/usr/local/bin/kubectl-foo
- warning: /usr/local/bin/kubectl-foo identified as a plugin, but it is not executable
/usr/local/bin/kubectl-bar
error: one plugin warning was found
You can think of plugins as a means to build more complex functionality on top
of the existing kubectl commands:
cat ./kubectl-whoami
The next few examples assume that you already made kubectl-whoami have
the following contents:
#!/bin/bash
# this plugin makes use of the `kubectl config` command in order to output# information about the current user, based on the currently selected contextkubectl config view --template='{{ range .contexts }}{{ if eq .name "'$(kubectl config current-context)'" }}Current user: {{ printf "%s\n" .context.user }}{{ end }}{{ end }}'
Running the above command gives you an output containing the user for the
current context in your KUBECONFIG file:
# make the file executablesudo chmod +x ./kubectl-whoami
# and move it into your PATHsudo mv ./kubectl-whoami /usr/local/bin
kubectl whoami
Current user: plugins-user
This page contains a list of commonly used kubectl commands and flags.
Kubectl autocomplete
BASH
source <(kubectl completion bash)# setup autocomplete in bash into the current shell, bash-completion package should be installed first.echo"source <(kubectl completion bash)" >> ~/.bashrc # add autocomplete permanently to your bash shell.
You can also use a shorthand alias for kubectl that also works with completion:
aliask=kubectl
complete -o default -F __start_kubectl k
ZSH
source <(kubectl completion zsh)# setup autocomplete in zsh into the current shellecho'[[ $commands[kubectl] ]] && source <(kubectl completion zsh)' >> ~/.zshrc # add autocomplete permanently to your zsh shell
A Note on --all-namespaces
Appending --all-namespaces happens frequently enough where you should be aware of the shorthand for --all-namespaces:
kubectl -A
Kubectl context and configuration
Set which Kubernetes cluster kubectl communicates with and modifies configuration
information. See Authenticating Across Clusters with kubeconfig documentation for
detailed config file information.
kubectl config view # Show Merged kubeconfig settings.# use multiple kubeconfig files at the same time and view merged configKUBECONFIG=~/.kube/config:~/.kube/kubconfig2
kubectl config view
# get the password for the e2e userkubectl config view -o jsonpath='{.users[?(@.name == "e2e")].user.password}'kubectl config view -o jsonpath='{.users[].name}'# display the first userkubectl config view -o jsonpath='{.users[*].name}'# get a list of userskubectl config get-contexts # display list of contextskubectl config current-context # display the current-contextkubectl config use-context my-cluster-name # set the default context to my-cluster-namekubectl config set-cluster my-cluster-name # set a cluster entry in the kubeconfig# configure the URL to a proxy server to use for requests made by this client in the kubeconfigkubectl config set-cluster my-cluster-name --proxy-url=my-proxy-url
# add a new user to your kubeconf that supports basic authkubectl config set-credentials kubeuser/foo.kubernetes.com --username=kubeuser --password=kubepassword
# permanently save the namespace for all subsequent kubectl commands in that context.kubectl config set-context --current --namespace=ggckad-s2
# set a context utilizing a specific username and namespace.kubectl config set-context gce --user=cluster-admin --namespace=foo \
&& kubectl config use-context gce
kubectl config unset users.foo # delete user foo# short alias to set/show context/namespace (only works for bash and bash-compatible shells, current context to be set before using kn to set namespace) aliaskx='f() { [ "$1" ] && kubectl config use-context $1 || kubectl config current-context ; } ; f'aliaskn='f() { [ "$1" ] && kubectl config set-context --current --namespace $1 || kubectl config view --minify | grep namespace | cut -d" " -f6 ; } ; f'
Kubectl apply
apply manages applications through files defining Kubernetes resources. It creates and updates resources in a cluster through running kubectl apply. This is the recommended way of managing Kubernetes applications on production. See Kubectl Book.
Creating objects
Kubernetes manifests can be defined in YAML or JSON. The file extension .yaml,
.yml, and .json can be used.
kubectl apply -f ./my-manifest.yaml # create resource(s)kubectl apply -f ./my1.yaml -f ./my2.yaml # create from multiple fileskubectl apply -f ./dir # create resource(s) in all manifest files in dirkubectl apply -f https://git.io/vPieo # create resource(s) from urlkubectl create deployment nginx --image=nginx # start a single instance of nginx# create a Job which prints "Hello World"kubectl create job hello --image=busybox:1.28 -- echo"Hello World"# create a CronJob that prints "Hello World" every minutekubectl create cronjob hello --image=busybox:1.28 --schedule="*/1 * * * *" -- echo"Hello World"kubectl explain pods # get the documentation for pod manifests# Create multiple YAML objects from stdincat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
name: busybox-sleep
spec:
containers:
- name: busybox
image: busybox:1.28
args:
- sleep
- "1000000"
---
apiVersion: v1
kind: Pod
metadata:
name: busybox-sleep-less
spec:
containers:
- name: busybox
image: busybox:1.28
args:
- sleep
- "1000"
EOF# Create a secret with several keyscat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Secret
metadata:
name: mysecret
type: Opaque
data:
password: $(echo -n "s33msi4" | base64 -w0)
username: $(echo -n "jane" | base64 -w0)
EOF
Viewing, finding resources
# Get commands with basic outputkubectl get services # List all services in the namespacekubectl get pods --all-namespaces # List all pods in all namespaceskubectl get pods -o wide # List all pods in the current namespace, with more detailskubectl get deployment my-dep # List a particular deploymentkubectl get pods # List all pods in the namespacekubectl get pod my-pod -o yaml # Get a pod's YAML# Describe commands with verbose outputkubectl describe nodes my-node
kubectl describe pods my-pod
# List Services Sorted by Namekubectl get services --sort-by=.metadata.name
# List pods Sorted by Restart Countkubectl get pods --sort-by='.status.containerStatuses[0].restartCount'# List PersistentVolumes sorted by capacitykubectl get pv --sort-by=.spec.capacity.storage
# Get the version label of all pods with label app=cassandrakubectl get pods --selector=app=cassandra -o \
jsonpath='{.items[*].metadata.labels.version}'# Retrieve the value of a key with dots, e.g. 'ca.crt'kubectl get configmap myconfig \
-o jsonpath='{.data.ca\.crt}'# Retrieve a base64 encoded value with dashes instead of underscores.kubectl get secret my-secret --template='{{index .data "key-name-with-dashes"}}'# Get all worker nodes (use a selector to exclude results that have a label# named 'node-role.kubernetes.io/control-plane')kubectl get node --selector='!node-role.kubernetes.io/control-plane'# Get all running pods in the namespacekubectl get pods --field-selector=status.phase=Running
# Get ExternalIPs of all nodeskubectl get nodes -o jsonpath='{.items[*].status.addresses[?(@.type=="ExternalIP")].address}'# List Names of Pods that belong to Particular RC# "jq" command useful for transformations that are too complex for jsonpath, it can be found at https://stedolan.github.io/jq/sel=${$(kubectl get rc my-rc --output=json | jq -j '.spec.selector | to_entries | .[] | "\(.key)=\(.value),"')%?}echo$(kubectl get pods --selector=$sel --output=jsonpath={.items..metadata.name})# Show labels for all pods (or any other Kubernetes object that supports labelling)kubectl get pods --show-labels
# Check which nodes are readyJSONPATH='{range .items[*]}{@.metadata.name}:{range @.status.conditions[*]}{@.type}={@.status};{end}{end}'\
&& kubectl get nodes -o jsonpath="$JSONPATH" | grep "Ready=True"# Output decoded secrets without external toolskubectl get secret my-secret -o go-template='{{range $k,$v := .data}}{{"### "}}{{$k}}{{"\n"}}{{$v|base64decode}}{{"\n\n"}}{{end}}'# List all Secrets currently in use by a podkubectl get pods -o json | jq '.items[].spec.containers[].env[]?.valueFrom.secretKeyRef.name' | grep -v null | sort | uniq
# List all containerIDs of initContainer of all pods# Helpful when cleaning up stopped containers, while avoiding removal of initContainers.kubectl get pods --all-namespaces -o jsonpath='{range .items[*].status.initContainerStatuses[*]}{.containerID}{"\n"}{end}' | cut -d/ -f3
# List Events sorted by timestampkubectl get events --sort-by=.metadata.creationTimestamp
# Compares the current state of the cluster against the state that the cluster would be in if the manifest was applied.kubectl diff -f ./my-manifest.yaml
# Produce a period-delimited tree of all keys returned for nodes# Helpful when locating a key within a complex nested JSON structurekubectl get nodes -o json | jq -c 'paths|join(".")'# Produce a period-delimited tree of all keys returned for pods, etckubectl get pods -o json | jq -c 'paths|join(".")'# Produce ENV for all pods, assuming you have a default container for the pods, default namespace and the `env` command is supported.# Helpful when running any supported command across all pods, not just `env`for pod in $(kubectl get po --output=jsonpath={.items..metadata.name}); doecho$pod&& kubectl exec -it $pod -- env; done# Get a deployment's status subresourcekubectl get deployment nginx-deployment --subresource=status
Updating resources
kubectl set image deployment/frontend www=image:v2 # Rolling update "www" containers of "frontend" deployment, updating the imagekubectl rollout history deployment/frontend # Check the history of deployments including the revisionkubectl rollout undo deployment/frontend # Rollback to the previous deploymentkubectl rollout undo deployment/frontend --to-revision=2# Rollback to a specific revisionkubectl rollout status -w deployment/frontend # Watch rolling update status of "frontend" deployment until completionkubectl rollout restart deployment/frontend # Rolling restart of the "frontend" deploymentcat pod.json | kubectl replace -f - # Replace a pod based on the JSON passed into stdin# Force replace, delete and then re-create the resource. Will cause a service outage.kubectl replace --force -f ./pod.json
# Create a service for a replicated nginx, which serves on port 80 and connects to the containers on port 8000kubectl expose rc nginx --port=80 --target-port=8000# Update a single-container pod's image version (tag) to v4kubectl get pod mypod -o yaml | sed 's/\(image: myimage\):.*$/\1:v4/' | kubectl replace -f -
kubectl label pods my-pod new-label=awesome # Add a Labelkubectl annotate pods my-pod icon-url=http://goo.gl/XXBTWq # Add an annotationkubectl autoscale deployment foo --min=2 --max=10# Auto scale a deployment "foo"
Patching resources
# Partially update a nodekubectl patch node k8s-node-1 -p '{"spec":{"unschedulable":true}}'# Update a container's image; spec.containers[*].name is required because it's a merge keykubectl patch pod valid-pod -p '{"spec":{"containers":[{"name":"kubernetes-serve-hostname","image":"new image"}]}}'# Update a container's image using a json patch with positional arrayskubectl patch pod valid-pod --type='json' -p='[{"op": "replace", "path": "/spec/containers/0/image", "value":"new image"}]'# Disable a deployment livenessProbe using a json patch with positional arrayskubectl patch deployment valid-deployment --type json -p='[{"op": "remove", "path": "/spec/template/spec/containers/0/livenessProbe"}]'# Add a new element to a positional arraykubectl patch sa default --type='json' -p='[{"op": "add", "path": "/secrets/1", "value": {"name": "whatever" } }]'# Update a deployment's replica count by patching its scale subresourcekubectl patch deployment nginx-deployment --subresource='scale' --type='merge' -p '{"spec":{"replicas":2}}'
Editing resources
Edit any API resource in your preferred editor.
kubectl edit svc/docker-registry # Edit the service named docker-registryKUBE_EDITOR="nano" kubectl edit svc/docker-registry # Use an alternative editor
Scaling resources
kubectl scale --replicas=3 rs/foo # Scale a replicaset named 'foo' to 3kubectl scale --replicas=3 -f foo.yaml # Scale a resource specified in "foo.yaml" to 3kubectl scale --current-replicas=2 --replicas=3 deployment/mysql # If the deployment named mysql's current size is 2, scale mysql to 3kubectl scale --replicas=5 rc/foo rc/bar rc/baz # Scale multiple replication controllers
Deleting resources
kubectl delete -f ./pod.json # Delete a pod using the type and name specified in pod.jsonkubectl delete pod unwanted --now # Delete a pod with no grace periodkubectl delete pod,service baz foo # Delete pods and services with same names "baz" and "foo"kubectl delete pods,services -l name=myLabel # Delete pods and services with label name=myLabelkubectl -n my-ns delete pod,svc --all # Delete all pods and services in namespace my-ns,# Delete all pods matching the awk pattern1 or pattern2kubectl get pods -n mynamespace --no-headers=true | awk '/pattern1|pattern2/{print $1}' | xargs kubectl delete -n mynamespace pod
Interacting with running Pods
kubectl logs my-pod # dump pod logs (stdout)kubectl logs -l name=myLabel # dump pod logs, with label name=myLabel (stdout)kubectl logs my-pod --previous # dump pod logs (stdout) for a previous instantiation of a containerkubectl logs my-pod -c my-container # dump pod container logs (stdout, multi-container case)kubectl logs -l name=myLabel -c my-container # dump pod logs, with label name=myLabel (stdout)kubectl logs my-pod -c my-container --previous # dump pod container logs (stdout, multi-container case) for a previous instantiation of a containerkubectl logs -f my-pod # stream pod logs (stdout)kubectl logs -f my-pod -c my-container # stream pod container logs (stdout, multi-container case)kubectl logs -f -l name=myLabel --all-containers # stream all pods logs with label name=myLabel (stdout)kubectl run -i --tty busybox --image=busybox:1.28 -- sh # Run pod as interactive shellkubectl run nginx --image=nginx -n mynamespace # Start a single instance of nginx pod in the namespace of mynamespacekubectl run nginx --image=nginx # Run pod nginx and write its spec into a file called pod.yaml--dry-run=client -o yaml > pod.yaml
kubectl attach my-pod -i # Attach to Running Containerkubectl port-forward my-pod 5000:6000 # Listen on port 5000 on the local machine and forward to port 6000 on my-podkubectl exec my-pod -- ls / # Run command in existing pod (1 container case)kubectl exec --stdin --tty my-pod -- /bin/sh # Interactive shell access to a running pod (1 container case)kubectl exec my-pod -c my-container -- ls / # Run command in existing pod (multi-container case)kubectl top pod POD_NAME --containers # Show metrics for a given pod and its containerskubectl top pod POD_NAME --sort-by=cpu # Show metrics for a given pod and sort it by 'cpu' or 'memory'
Copy files and directories to and from containers
kubectl cp /tmp/foo_dir my-pod:/tmp/bar_dir # Copy /tmp/foo_dir local directory to /tmp/bar_dir in a remote pod in the current namespacekubectl cp /tmp/foo my-pod:/tmp/bar -c my-container # Copy /tmp/foo local file to /tmp/bar in a remote pod in a specific containerkubectl cp /tmp/foo my-namespace/my-pod:/tmp/bar # Copy /tmp/foo local file to /tmp/bar in a remote pod in namespace my-namespacekubectl cp my-namespace/my-pod:/tmp/foo /tmp/bar # Copy /tmp/foo from a remote pod to /tmp/bar locally
Note:kubectl cp requires that the 'tar' binary is present in your container image. If 'tar' is not present,kubectl cp will fail.
For advanced use cases, such as symlinks, wildcard expansion or file mode preservation consider using kubectl exec.
tar cf - /tmp/foo | kubectl exec -i -n my-namespace my-pod -- tar xf - -C /tmp/bar # Copy /tmp/foo local file to /tmp/bar in a remote pod in namespace my-namespacekubectl exec -n my-namespace my-pod -- tar cf - /tmp/foo | tar xf - -C /tmp/bar # Copy /tmp/foo from a remote pod to /tmp/bar locally
Interacting with Deployments and Services
kubectl logs deploy/my-deployment # dump Pod logs for a Deployment (single-container case)kubectl logs deploy/my-deployment -c my-container # dump Pod logs for a Deployment (multi-container case)kubectl port-forward svc/my-service 5000# listen on local port 5000 and forward to port 5000 on Service backendkubectl port-forward svc/my-service 5000:my-service-port # listen on local port 5000 and forward to Service target port with name <my-service-port>kubectl port-forward deploy/my-deployment 5000:6000 # listen on local port 5000 and forward to port 6000 on a Pod created by <my-deployment>kubectl exec deploy/my-deployment -- ls # run command in first Pod and first container in Deployment (single- or multi-container cases)
Interacting with Nodes and cluster
kubectl cordon my-node # Mark my-node as unschedulablekubectl drain my-node # Drain my-node in preparation for maintenancekubectl uncordon my-node # Mark my-node as schedulablekubectl top node my-node # Show metrics for a given nodekubectl cluster-info # Display addresses of the master and serviceskubectl cluster-info dump # Dump current cluster state to stdoutkubectl cluster-info dump --output-directory=/path/to/cluster-state # Dump current cluster state to /path/to/cluster-state# View existing taints on which exist on current nodes.kubectl get nodes -o=custom-columns=NodeName:.metadata.name,TaintKey:.spec.taints[*].key,TaintValue:.spec.taints[*].value,TaintEffect:.spec.taints[*].effect
# If a taint with that key and effect already exists, its value is replaced as specified.kubectl taint nodes foo dedicated=special-user:NoSchedule
Resource types
List all supported resource types along with their shortnames, API group, whether they are namespaced, and Kind:
kubectl api-resources
Other operations for exploring API resources:
kubectl api-resources --namespaced=true# All namespaced resourceskubectl api-resources --namespaced=false# All non-namespaced resourceskubectl api-resources -o name # All resources with simple output (only the resource name)kubectl api-resources -o wide # All resources with expanded (aka "wide") outputkubectl api-resources --verbs=list,get # All resources that support the "list" and "get" request verbskubectl api-resources --api-group=extensions # All resources in the "extensions" API group
Formatting output
To output details to your terminal window in a specific format, add the -o (or --output) flag to a supported kubectl command.
Output format
Description
-o=custom-columns=<spec>
Print a table using a comma separated list of custom columns
-o=custom-columns-file=<filename>
Print a table using the custom columns template in the <filename> file
Print the fields defined by the jsonpath expression in the <filename> file
-o=name
Print only the resource name and nothing else
-o=wide
Output in the plain-text format with any additional information, and for pods, the node name is included
-o=yaml
Output a YAML formatted API object
Examples using -o=custom-columns:
# All images running in a clusterkubectl get pods -A -o=custom-columns='DATA:spec.containers[*].image'# All images running in namespace: default, grouped by Podkubectl get pods --namespace default --output=custom-columns="NAME:.metadata.name,IMAGE:.spec.containers[*].image"# All images excluding "registry.k8s.io/coredns:1.6.2"kubectl get pods -A -o=custom-columns='DATA:spec.containers[?(@.image!="registry.k8s.io/coredns:1.6.2")].image'# All fields under metadata regardless of namekubectl get pods -A -o=custom-columns='DATA:metadata.*'
Kubectl verbosity is controlled with the -v or --v flags followed by an integer representing the log level. General Kubernetes logging conventions and the associated log levels are described here.
Verbosity
Description
--v=0
Generally useful for this to always be visible to a cluster operator.
--v=1
A reasonable default log level if you don't want verbosity.
--v=2
Useful steady state information about the service and important log messages that may correlate to significant changes in the system. This is the recommended default log level for most systems.
--v=3
Extended information about changes.
--v=4
Debug level verbosity.
--v=5
Trace level verbosity.
--v=6
Display requested resources.
--v=7
Display HTTP request headers.
--v=8
Display HTTP request contents.
--v=9
Display HTTP request contents without truncation of contents.
CIDRs opened in GCE firewall for L4 LB traffic proxy & health checks
--cluster string
The name of the kubeconfig cluster to use
--context string
The name of the kubeconfig context to use
--default-not-ready-toleration-seconds int Default: 300
Indicates the tolerationSeconds of the toleration for notReady:NoExecute that is added by default to every pod that does not already have such a toleration.
--default-unreachable-toleration-seconds int Default: 300
Indicates the tolerationSeconds of the toleration for unreachable:NoExecute that is added by default to every pod that does not already have such a toleration.
-h, --help
help for kubectl
--insecure-skip-tls-verify
If true, the server's certificate will not be checked for validity. This will make your HTTPS connections insecure
--kubeconfig string
Path to the kubeconfig file to use for CLI requests.
--log-backtrace-at traceLocation Default: :0
when logging hits line file:N, emit a stack trace
--log-dir string
If non-empty, write log files in this directory
--log-file string
If non-empty, use this log file
--log-file-max-size uint Default: 1800
Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited.
--log-flush-frequency duration Default: 5s
Maximum number of seconds between log flushes
--logtostderr Default: true
log to standard error instead of files
--match-server-version
Require server version to match client version
-n, --namespace string
If present, the namespace scope for this CLI request
--one-output
If true, only write logs to their native severity level (vs also writing to each lower severity level
--password string
Password for basic authentication to the API server
--profile string Default: "none"
Name of profile to capture. One of (none|cpu|heap|goroutine|threadcreate|block|mutex)
--profile-output string Default: "profile.pprof"
Name of the file to write the profile to
--request-timeout string Default: "0"
The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests.
-s, --server string
The address and port of the Kubernetes API server
--skip-headers
If true, avoid header prefixes in the log messages
--skip-log-headers
If true, avoid headers when opening log files
--stderrthreshold severity Default: 2
logs at or above this threshold go to stderr
--tls-server-name string
Server name to use for server certificate validation. If it is not provided, the hostname used to contact the server is used
--token string
Bearer token for authentication to the API server
--user string
The name of the kubeconfig user to use
--username string
Username for basic authentication to the API server
-v, --v Level
number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule moduleSpec
comma-separated list of pattern=N settings for file-filtered logging
--warnings-as-errors
Treat warnings received from the server as errors and exit with a non-zero exit code
Environment variables
KUBECONFIG
Path to the kubectl configuration ("kubeconfig") file. Default: "$HOME/.kube/config"
KUBECTL_COMMAND_HEADERS
When set to false, turns off extra HTTP headers detailing invoked kubectl command (Kubernetes version v1.22 or later)
kubectl version - Print the client and server version information
kubectl wait - Experimental: Wait for a specific condition on one or many resources.
6.10.4 - JSONPath Support
Kubectl supports JSONPath template.
JSONPath template is composed of JSONPath expressions enclosed by curly braces {}.
Kubectl uses JSONPath expressions to filter on specific fields in the JSON object and format the output.
In addition to the original JSONPath template syntax, the following functions and syntax are valid:
Use double quotes to quote text inside JSONPath expressions.
Use the range, end operators to iterate lists.
Use negative slice indices to step backwards through a list. Negative indices do not "wrap around" a list and are valid as long as -index + listLength >= 0.
Note:
The $ operator is optional since the expression always starts from the root object by default.
The result object is printed as its String() function.
kubectl get pods -o json
kubectl get pods -o=jsonpath='{@}'kubectl get pods -o=jsonpath='{.items[0]}'kubectl get pods -o=jsonpath='{.items[0].metadata.name}'kubectl get pods -o=jsonpath="{.items[*]['metadata.name', 'status.capacity']}"kubectl get pods -o=jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.status.startTime}{"\n"}{end}'
Note:
On Windows, you must double quote any JSONPath template that contains spaces (not single quote as shown above for bash). This in turn means that you must use a single quote or escaped double quote around any literals in the template. For example:
kubectl get pods -o=jsonpath="{range .items[*]}{.metadata.name}{'\t'}{.status.startTime}{'\n'}{end}"kubectl get pods -o=jsonpath="{range .items[*]}{.metadata.name}{\"\t\"}{.status.startTime}{\"\n\"}{end}"
Note:
JSONPath regular expressions are not supported. If you want to match using regular expressions, you can use a tool such as jq.
# kubectl does not support regular expressions for JSONpath output# The following command does not workkubectl get pods -o jsonpath='{.items[?(@.metadata.name=~/^test$/)].metadata.name}'# The following command achieves the desired resultkubectl get pods -o json | jq -r '.items[] | select(.metadata.name | test("test-")).spec.containers[].image'
6.10.5 - kubectl for Docker Users
You can use the Kubernetes command line tool kubectl to interact with the API Server. Using kubectl is straightforward if you are familiar with the Docker command line tool. However, there are a few differences between the Docker commands and the kubectl commands. The following sections show a Docker sub-command and describe the equivalent kubectl command.
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
55c103fa1296 nginx "nginx -g 'daemon of…" 9 seconds ago Up 9 seconds 0.0.0.0:80->80/tcp nginx-app
kubectl:
# start the pod running nginxkubectl create deployment --image=nginx nginx-app
deployment.apps/nginx-app created
# add env to nginx-appkubectl set env deployment/nginx-app DOMAIN=cluster
deployment.apps/nginx-app env updated
Note:kubectl commands print the type and name of the resource created or mutated, which can then be used in subsequent commands. You can expose a new Service after a Deployment is created.
# expose a port through with a servicekubectl expose deployment nginx-app --port=80 --name=nginx-http
service "nginx-http" exposed
By using kubectl, you can create a Deployment to ensure that N pods are running nginx, where N is the number of replicas stated in the spec and defaults to 1. You can also create a service with a selector that matches the pod labels. For more information, see Use a Service to Access an Application in a Cluster.
By default images run in the background, similar to docker run -d .... To run things in the foreground, use kubectl run to create pod:
kubectl run [-i][--tty] --attach <name> --image=<image>
Unlike docker run ..., if you specify --attach, then you attach stdin, stdout and stderr. You cannot control which streams are attached (docker -a ...).
To detach from the container, you can type the escape sequence Ctrl+P followed by Ctrl+Q.
docker ps
To list what is currently running, see kubectl get.
docker:
docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
14636241935f ubuntu:16.04 "echo test" 5 seconds ago Exited (0) 5 seconds ago cocky_fermi
55c103fa1296 nginx "nginx -g 'daemon of…" About a minute ago Up About a minute 0.0.0.0:80->80/tcp nginx-app
kubectl:
kubectl get po
NAME READY STATUS RESTARTS AGE
nginx-app-8df569cb7-4gd89 1/1 Running 0 3m
ubuntu 0/1 Completed 0 20s
docker attach
To attach a process that is already running in a container, see kubectl attach.
docker:
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
55c103fa1296 nginx "nginx -g 'daemon of…" 5 minutes ago Up 5 minutes 0.0.0.0:80->80/tcp nginx-app
docker attach 55c103fa1296
...
kubectl:
kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-app-5jyvm 1/1 Running 0 10m
kubectl attach -it nginx-app-5jyvm
...
To detach from the container, you can type the escape sequence Ctrl+P followed by Ctrl+Q.
docker exec
To execute a command in a container, see kubectl exec.
docker:
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
55c103fa1296 nginx "nginx -g 'daemon of…" 6 minutes ago Up 6 minutes 0.0.0.0:80->80/tcp nginx-app
docker exec 55c103fa1296 cat /etc/hostname
55c103fa1296
kubectl:
kubectl get po
NAME READY STATUS RESTARTS AGE
nginx-app-5jyvm 1/1 Running 0 10m
There is a slight difference between pods and containers; by default pods do not terminate if their processes exit. Instead the pods restart the process. This is similar to the docker run option --restart=always with one major difference. In docker, the output for each invocation of the process is concatenated, but for Kubernetes, each invocation is separate. To see the output from a previous run in Kubernetes, do this:
To stop and delete a running process, see kubectl delete.
docker:
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
a9ec34d98787 nginx "nginx -g 'daemon of" 22 hours ago Up 22 hours 0.0.0.0:80->80/tcp, 443/tcp nginx-app
docker stop a9ec34d98787
a9ec34d98787
docker rm a9ec34d98787
a9ec34d98787
kubectl:
kubectl get deployment nginx-app
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-app 1/1 1 1 2m
kubectl get po -l app=nginx-app
NAME READY STATUS RESTARTS AGE
nginx-app-2883164633-aklf7 1/1 Running 0 2m
kubectl delete deployment nginx-app
deployment "nginx-app" deleted
kubectl get po -l app=nginx-app
# Return nothing
Note: When you use kubectl, you don't delete the pod directly. You have to first delete the Deployment that owns the pod. If you delete the pod directly, the Deployment recreates the pod.
docker login
There is no direct analog of docker login in kubectl. If you are interested in using Kubernetes with a private registry, see Using a Private Registry.
docker version
To get the version of client and server, see kubectl version.
docker:
docker version
Client version: 1.7.0
Client API version: 1.19
Go version (client): go1.4.2
Git commit (client): 0baf609
OS/Arch (client): linux/amd64
Server version: 1.7.0
Server API version: 1.19
Go version (server): go1.4.2
Git commit (server): 0baf609
OS/Arch (server): linux/amd64
Kubernetes master is running at https://203.0.113.141
KubeDNS is running at https://203.0.113.141/api/v1/namespaces/kube-system/services/kube-dns/proxy
kubernetes-dashboard is running at https://203.0.113.141/api/v1/namespaces/kube-system/services/kubernetes-dashboard/proxy
Grafana is running at https://203.0.113.141/api/v1/namespaces/kube-system/services/monitoring-grafana/proxy
Heapster is running at https://203.0.113.141/api/v1/namespaces/kube-system/services/monitoring-heapster/proxy
InfluxDB is running at https://203.0.113.141/api/v1/namespaces/kube-system/services/monitoring-influxdb/proxy
6.10.6 - kubectl Usage Conventions
Recommended usage conventions for kubectl.
Using kubectl in Reusable Scripts
For a stable output in a script:
Request one of the machine-oriented output forms, such as -o name, -o json, -o yaml, -o go-template, or -o jsonpath.
Fully-qualify the version. For example, jobs.v1.batch/myjob. This will ensure that kubectl does not use its default version that can change over time.
Don't rely on context, preferences, or other implicit states.
Subresources
You can use the --subresource alpha flag for kubectl commands like get, patch,
edit and replace to fetch and update subresources for all resources that
support them. Currently, only the status and scale subresources are supported.
The API contract against a subresource is identical to a full resource. While updating the
status subresource to a new value, keep in mind that the subresource could be potentially
reconciled by a controller to a different value.
Best Practices
kubectl run
For kubectl run to satisfy infrastructure as code:
Tag the image with a version-specific tag and don't move that tag to a new version. For example, use :v1234, v1.2.3, r03062016-1-4, rather than :latest (For more information, see Best Practices for Configuration).
Check in the script for an image that is heavily parameterized.
Switch to configuration files checked into source control for features that are needed, but not expressible via kubectl run flags.
You can use the --dry-run=client flag to preview the object that would be sent to your cluster, without really submitting it.
kubectl apply
You can use kubectl apply to create or update resources. For more information about using kubectl apply to update resources, see Kubectl Book.
6.11 - Component tools
6.11.1 - Feature Gates
This page contains an overview of the various feature gates an administrator
can specify on different Kubernetes components.
See feature stages for an explanation of the stages for a feature.
Overview
Feature gates are a set of key=value pairs that describe Kubernetes features.
You can turn these features on or off using the --feature-gates command line flag
on each Kubernetes component.
Each Kubernetes component lets you enable or disable a set of feature gates that
are relevant to that component.
Use -h flag to see a full set of feature gates for all components.
To set feature gates for a component, such as kubelet, use the --feature-gates
flag assigned to a list of feature pairs:
--feature-gates=...,GracefulNodeShutdown=true
The following tables are a summary of the feature gates that you can set on
different Kubernetes components.
The "Since" column contains the Kubernetes release when a feature is introduced
or its release stage is changed.
The "Until" column, if not empty, contains the last Kubernetes release in which
you can still use a feature gate.
Feature gates for features in Alpha or Beta states
Feature
Default
Stage
Since
Until
APIListChunking
false
Alpha
1.8
1.8
APIListChunking
true
Beta
1.9
APIPriorityAndFairness
false
Alpha
1.18
1.19
APIPriorityAndFairness
true
Beta
1.20
APIResponseCompression
false
Alpha
1.7
1.15
APIResponseCompression
true
Beta
1.16
APIServerIdentity
false
Alpha
1.20
APIServerTracing
false
Alpha
1.22
AllowInsecureBackendProxy
true
Beta
1.17
AnyVolumeDataSource
false
Alpha
1.18
1.23
AnyVolumeDataSource
true
Beta
1.24
AppArmor
true
Beta
1.4
CheckpointContainer
false
Alpha
1.25
CPUManager
false
Alpha
1.8
1.9
CPUManager
true
Beta
1.10
CPUManagerPolicyAlphaOptions
false
Alpha
1.23
CPUManagerPolicyBetaOptions
true
Beta
1.23
CPUManagerPolicyOptions
false
Alpha
1.22
1.22
CPUManagerPolicyOptions
true
Beta
1.23
CSIMigrationAzureFile
false
Alpha
1.15
1.20
CSIMigrationAzureFile
false
Beta
1.21
1.23
CSIMigrationAzureFile
true
Beta
1.24
CSIMigrationPortworx
false
Alpha
1.23
1.24
CSIMigrationPortworx
false
Beta
1.25
CSIMigrationRBD
false
Alpha
1.23
CSIMigrationvSphere
false
Alpha
1.18
1.18
CSIMigrationvSphere
false
Beta
1.19
1.24
CSIMigrationvSphere
true
Beta
1.25
CSINodeExpandSecret
false
Alpha
1.25
CSIVolumeHealth
false
Alpha
1.21
ContextualLogging
false
Alpha
1.24
CustomCPUCFSQuotaPeriod
false
Alpha
1.12
CustomResourceValidationExpressions
false
Alpha
1.23
1.24
CustomResourceValidationExpressions
true
Beta
1.25
DelegateFSGroupToCSIDriver
false
Alpha
1.22
1.22
DelegateFSGroupToCSIDriver
true
Beta
1.23
DevicePlugins
false
Alpha
1.8
1.9
DevicePlugins
true
Beta
1.10
DisableAcceleratorUsageMetrics
false
Alpha
1.19
1.19
DisableAcceleratorUsageMetrics
true
Beta
1.20
DisableCloudProviders
false
Alpha
1.22
DisableKubeletCloudCredentialProviders
false
Alpha
1.23
DownwardAPIHugePages
false
Alpha
1.20
1.20
DownwardAPIHugePages
false
Beta
1.21
1.21
DownwardAPIHugePages
true
Beta
1.22
EndpointSliceTerminatingCondition
false
Alpha
1.20
1.21
EndpointSliceTerminatingCondition
true
Beta
1.22
ExpandedDNSConfig
false
Alpha
1.22
ExperimentalHostUserNamespaceDefaulting
false
Beta
1.5
GracefulNodeShutdown
false
Alpha
1.20
1.20
GracefulNodeShutdown
true
Beta
1.21
GracefulNodeShutdownBasedOnPodPriority
false
Alpha
1.23
1.23
GracefulNodeShutdownBasedOnPodPriority
true
Beta
1.24
GRPCContainerProbe
false
Alpha
1.23
1.23
GRPCContainerProbe
true
Beta
1.24
HonorPVReclaimPolicy
false
Alpha
1.23
HPAContainerMetrics
false
Alpha
1.20
HPAScaleToZero
false
Alpha
1.16
InTreePluginAWSUnregister
false
Alpha
1.21
InTreePluginAzureDiskUnregister
false
Alpha
1.21
InTreePluginAzureFileUnregister
false
Alpha
1.21
InTreePluginGCEUnregister
false
Alpha
1.21
InTreePluginOpenStackUnregister
false
Alpha
1.21
InTreePluginPortworxUnregister
false
Alpha
1.23
InTreePluginRBDUnregister
false
Alpha
1.23
InTreePluginvSphereUnregister
false
Alpha
1.21
JobMutableNodeSchedulingDirectives
true
Beta
1.23
JobReadyPods
false
Alpha
1.23
1.23
JobReadyPods
true
Beta
1.24
JobTrackingWithFinalizers
false
Alpha
1.22
1.22
JobTrackingWithFinalizers
true
Beta
1.23
1.23
JobTrackingWithFinalizers
false
Beta
1.24
KubeletCredentialProviders
false
Alpha
1.20
1.23
KubeletCredentialProviders
true
Beta
1.24
KubeletInUserNamespace
false
Alpha
1.22
KubeletPodResources
false
Alpha
1.13
1.14
KubeletPodResources
true
Beta
1.15
KubeletPodResourcesGetAllocatable
false
Alpha
1.21
1.22
KubeletPodResourcesGetAllocatable
true
Beta
1.23
KubeletTracing
false
Alpha
1.25
LocalStorageCapacityIsolationFSQuotaMonitoring
false
Alpha
1.15
1.24
LocalStorageCapacityIsolationFSQuotaMonitoring
true
Beta
1.25
LogarithmicScaleDown
false
Alpha
1.21
1.21
LogarithmicScaleDown
true
Beta
1.22
MatchLabelKeysInPodTopologySpread
false
Alpha
1.25
MaxUnavailableStatefulSet
false
Alpha
1.24
MemoryManager
false
Alpha
1.21
1.21
MemoryManager
true
Beta
1.22
MemoryQoS
false
Alpha
1.22
MinDomainsInPodTopologySpread
false
Alpha
1.24
1.24
MinDomainsInPodTopologySpread
true
Beta
1.25
MixedProtocolLBService
false
Alpha
1.20
1.23
MixedProtocolLBService
true
Beta
1.24
NetworkPolicyStatus
false
Alpha
1.24
NodeInclusionPolicyInPodTopologySpread
false
Alpha
1.25
NodeSwap
false
Alpha
1.22
NodeOutOfServiceVolumeDetach
false
Alpha
1.24
OpenAPIEnums
false
Alpha
1.23
1.23
OpenAPIEnums
true
Beta
1.24
OpenAPIV3
false
Alpha
1.23
1.23
OpenAPIV3
true
Beta
1.24
PodAndContainerStatsFromCRI
false
Alpha
1.23
PodDeletionCost
false
Alpha
1.21
1.21
PodDeletionCost
true
Beta
1.22
PodHasNetworkCondition
false
Alpha
1.25
PodSecurity
false
Alpha
1.22
1.22
PodSecurity
true
Beta
1.23
ProbeTerminationGracePeriod
false
Alpha
1.21
1.21
ProbeTerminationGracePeriod
false
Beta
1.22
1.24
ProbeTerminationGracePeriod
true
Beta
1.25
ProcMountType
false
Alpha
1.12
ProxyTerminatingEndpoints
false
Alpha
1.22
QOSReserved
false
Alpha
1.11
ReadWriteOncePod
false
Alpha
1.22
RecoverVolumeExpansionFailure
false
Alpha
1.23
RemainingItemCount
false
Alpha
1.15
1.15
RemainingItemCount
true
Beta
1.16
RotateKubeletServerCertificate
false
Alpha
1.7
1.11
RotateKubeletServerCertificate
true
Beta
1.12
SeccompDefault
false
Alpha
1.22
1.24
SeccompDefault
true
Beta
1.25
ServerSideFieldValidation
false
Alpha
1.23
1.24
ServerSideFieldValidation
true
Beta
1.25
ServiceInternalTrafficPolicy
false
Alpha
1.21
1.21
ServiceInternalTrafficPolicy
true
Beta
1.22
ServiceIPStaticSubrange
false
Alpha
1.24
1.24
ServiceIPStaticSubrange
true
Beta
1.25
SizeMemoryBackedVolumes
false
Alpha
1.20
1.21
SizeMemoryBackedVolumes
true
Beta
1.22
StatefulSetAutoDeletePVC
false
Alpha
1.22
StorageVersionAPI
false
Alpha
1.20
StorageVersionHash
false
Alpha
1.14
1.14
StorageVersionHash
true
Beta
1.15
TopologyAwareHints
false
Alpha
1.21
1.22
TopologyAwareHints
false
Beta
1.23
1.23
TopologyAwareHints
true
Beta
1.24
TopologyManager
false
Alpha
1.16
1.17
TopologyManager
true
Beta
1.18
VolumeCapacityPriority
false
Alpha
1.21
-
WinDSR
false
Alpha
1.14
WinOverlay
false
Alpha
1.14
1.19
WinOverlay
true
Beta
1.20
WindowsHostProcessContainers
false
Alpha
1.22
1.22
WindowsHostProcessContainers
true
Beta
1.23
Feature gates for graduated or deprecated features
Feature Gates for Graduated or Deprecated Features
Feature
Default
Stage
Since
Until
Accelerators
false
Alpha
1.6
1.10
Accelerators
-
Deprecated
1.11
-
AdvancedAuditing
false
Alpha
1.7
1.7
AdvancedAuditing
true
Beta
1.8
1.11
AdvancedAuditing
true
GA
1.12
-
AffinityInAnnotations
false
Alpha
1.6
1.7
AffinityInAnnotations
-
Deprecated
1.8
-
AllowExtTrafficLocalEndpoints
false
Beta
1.4
1.6
AllowExtTrafficLocalEndpoints
true
GA
1.7
-
AttachVolumeLimit
false
Alpha
1.11
1.11
AttachVolumeLimit
true
Beta
1.12
1.16
AttachVolumeLimit
true
GA
1.17
-
BalanceAttachedNodeVolumes
false
Alpha
1.11
1.21
BalanceAttachedNodeVolumes
false
Deprecated
1.22
BlockVolume
false
Alpha
1.9
1.12
BlockVolume
true
Beta
1.13
1.17
BlockVolume
true
GA
1.18
-
BoundServiceAccountTokenVolume
false
Alpha
1.13
1.20
BoundServiceAccountTokenVolume
true
Beta
1.21
1.21
BoundServiceAccountTokenVolume
true
GA
1.22
-
ConfigurableFSGroupPolicy
false
Alpha
1.18
1.19
ConfigurableFSGroupPolicy
true
Beta
1.20
1.22
ConfigurableFSGroupPolicy
true
GA
1.23
-
ControllerManagerLeaderMigration
false
Alpha
1.21
1.21
ControllerManagerLeaderMigration
true
Beta
1.22
1.23
ControllerManagerLeaderMigration
true
GA
1.24
-
CRIContainerLogRotation
false
Alpha
1.10
1.10
CRIContainerLogRotation
true
Beta
1.11
1.20
CRIContainerLogRotation
true
GA
1.21
-
CSIBlockVolume
false
Alpha
1.11
1.13
CSIBlockVolume
true
Beta
1.14
1.17
CSIBlockVolume
true
GA
1.18
-
CSIDriverRegistry
false
Alpha
1.12
1.13
CSIDriverRegistry
true
Beta
1.14
1.17
CSIDriverRegistry
true
GA
1.18
-
CSIInlineVolume
false
Alpha
1.15
1.15
CSIInlineVolume
true
Beta
1.16
1.24
CSIInlineVolume
true
GA
1.25
-
CSIMigration
false
Alpha
1.14
1.16
CSIMigration
true
Beta
1.17
1.24
CSIMigration
true
GA
1.25
-
CSIMigrationAWS
false
Alpha
1.14
1.16
CSIMigrationAWS
false
Beta
1.17
1.22
CSIMigrationAWS
true
Beta
1.23
1.24
CSIMigrationAWS
true
GA
1.25
-
CSIMigrationAWSComplete
false
Alpha
1.17
1.20
CSIMigrationAWSComplete
-
Deprecated
1.21
-
CSIMigrationAzureDisk
false
Alpha
1.15
1.18
CSIMigrationAzureDisk
false
Beta
1.19
1.22
CSIMigrationAzureDisk
true
Beta
1.23
1.23
CSIMigrationAzureDisk
true
GA
1.24
CSIMigrationAzureDiskComplete
false
Alpha
1.17
1.20
CSIMigrationAzureDiskComplete
-
Deprecated
1.21
-
CSIMigrationAzureFileComplete
false
Alpha
1.17
1.20
CSIMigrationAzureFileComplete
-
Deprecated
1.21
-
CSIMigrationGCE
false
Alpha
1.14
1.16
CSIMigrationGCE
false
Beta
1.17
1.22
CSIMigrationGCE
true
Beta
1.23
1.24
CSIMigrationGCE
true
GA
1.25
-
CSIMigrationGCEComplete
false
Alpha
1.17
1.20
CSIMigrationGCEComplete
-
Deprecated
1.21
-
CSIMigrationOpenStack
false
Alpha
1.14
1.17
CSIMigrationOpenStack
true
Beta
1.18
1.23
CSIMigrationOpenStack
true
GA
1.24
CSIMigrationOpenStackComplete
false
Alpha
1.17
1.20
CSIMigrationOpenStackComplete
-
Deprecated
1.21
-
CSIMigrationvSphereComplete
false
Beta
1.19
1.21
CSIMigrationvSphereComplete
-
Deprecated
1.22
-
CSINodeInfo
false
Alpha
1.12
1.13
CSINodeInfo
true
Beta
1.14
1.16
CSINodeInfo
true
GA
1.17
-
CSIPersistentVolume
false
Alpha
1.9
1.9
CSIPersistentVolume
true
Beta
1.10
1.12
CSIPersistentVolume
true
GA
1.13
-
CSIServiceAccountToken
false
Alpha
1.20
1.20
CSIServiceAccountToken
true
Beta
1.21
1.21
CSIServiceAccountToken
true
GA
1.22
-
CSIStorageCapacity
false
Alpha
1.19
1.20
CSIStorageCapacity
true
Beta
1.21
1.23
CSIStorageCapacity
true
GA
1.24
-
CSIVolumeFSGroupPolicy
false
Alpha
1.19
1.19
CSIVolumeFSGroupPolicy
true
Beta
1.20
1.22
CSIVolumeFSGroupPolicy
true
GA
1.23
CSRDuration
true
Beta
1.22
1.23
CSRDuration
true
GA
1.24
-
CronJobControllerV2
false
Alpha
1.20
1.20
CronJobControllerV2
true
Beta
1.21
1.21
CronJobControllerV2
true
GA
1.22
-
CronJobTimeZone
false
Alpha
1.24
1.24
CronJobTimeZone
true
Beta
1.25
CustomPodDNS
false
Alpha
1.9
1.9
CustomPodDNS
true
Beta
1.10
1.13
CustomPodDNS
true
GA
1.14
-
CustomResourceDefaulting
false
Alpha
1.15
1.15
CustomResourceDefaulting
true
Beta
1.16
1.16
CustomResourceDefaulting
true
GA
1.17
-
CustomResourcePublishOpenAPI
false
Alpha
1.14
1.14
CustomResourcePublishOpenAPI
true
Beta
1.15
1.15
CustomResourcePublishOpenAPI
true
GA
1.16
-
CustomResourceSubresources
false
Alpha
1.10
1.10
CustomResourceSubresources
true
Beta
1.11
1.15
CustomResourceSubresources
true
GA
1.16
-
CustomResourceValidation
false
Alpha
1.8
1.8
CustomResourceValidation
true
Beta
1.9
1.15
CustomResourceValidation
true
GA
1.16
-
CustomResourceWebhookConversion
false
Alpha
1.13
1.14
CustomResourceWebhookConversion
true
Beta
1.15
1.15
CustomResourceWebhookConversion
true
GA
1.16
-
DaemonSetUpdateSurge
false
Alpha
1.21
1.21
DaemonSetUpdateSurge
true
Beta
1.22
1.24
DaemonSetUpdateSurge
true
GA
1.25
-
DefaultPodTopologySpread
false
Alpha
1.19
1.19
DefaultPodTopologySpread
true
Beta
1.20
1.23
DefaultPodTopologySpread
true
GA
1.24
-
DryRun
false
Alpha
1.12
1.12
DryRun
true
Beta
1.13
1.18
DryRun
true
GA
1.19
-
DynamicAuditing
false
Alpha
1.13
1.18
DynamicAuditing
-
Deprecated
1.19
-
DynamicKubeletConfig
false
Alpha
1.4
1.10
DynamicKubeletConfig
true
Beta
1.11
1.21
DynamicKubeletConfig
false
Deprecated
1.22
-
DynamicProvisioningScheduling
false
Alpha
1.11
1.11
DynamicProvisioningScheduling
-
Deprecated
1.12
-
DynamicVolumeProvisioning
true
Alpha
1.3
1.7
DynamicVolumeProvisioning
true
GA
1.8
-
EfficientWatchResumption
false
Alpha
1.20
1.20
EfficientWatchResumption
true
Beta
1.21
1.23
EfficientWatchResumption
true
GA
1.24
-
EnableAggregatedDiscoveryTimeout
true
Deprecated
1.16
-
EnableEquivalenceClassCache
false
Alpha
1.8
1.14
EnableEquivalenceClassCache
-
Deprecated
1.15
-
EndpointSlice
false
Alpha
1.16
1.16
EndpointSlice
false
Beta
1.17
1.17
EndpointSlice
true
Beta
1.18
1.20
EndpointSlice
true
GA
1.21
-
EndpointSliceNodeName
false
Alpha
1.20
1.20
EndpointSliceNodeName
true
GA
1.21
-
EndpointSliceProxying
false
Alpha
1.18
1.18
EndpointSliceProxying
true
Beta
1.19
1.21
EndpointSliceProxying
true
GA
1.22
-
EphemeralContainers
false
Alpha
1.16
1.22
EphemeralContainers
true
Beta
1.23
1.24
EphemeralContainers
true
GA
1.25
-
EvenPodsSpread
false
Alpha
1.16
1.17
EvenPodsSpread
true
Beta
1.18
1.18
EvenPodsSpread
true
GA
1.19
-
ExecProbeTimeout
true
GA
1.20
-
ExpandCSIVolumes
false
Alpha
1.14
1.15
ExpandCSIVolumes
true
Beta
1.16
1.23
ExpandCSIVolumes
true
GA
1.24
-
ExpandInUsePersistentVolumes
false
Alpha
1.11
1.14
ExpandInUsePersistentVolumes
true
Beta
1.15
1.23
ExpandInUsePersistentVolumes
true
GA
1.24
-
ExpandPersistentVolumes
false
Alpha
1.8
1.10
ExpandPersistentVolumes
true
Beta
1.11
1.23
ExpandPersistentVolumes
true
GA
1.24
-
ExperimentalCriticalPodAnnotation
false
Alpha
1.5
1.12
ExperimentalCriticalPodAnnotation
false
Deprecated
1.13
-
ExternalPolicyForExternalIP
true
GA
1.18
-
GCERegionalPersistentDisk
true
Beta
1.10
1.12
GCERegionalPersistentDisk
true
GA
1.13
-
GenericEphemeralVolume
false
Alpha
1.19
1.20
GenericEphemeralVolume
true
Beta
1.21
1.22
GenericEphemeralVolume
true
GA
1.23
-
HugePageStorageMediumSize
false
Alpha
1.18
1.18
HugePageStorageMediumSize
true
Beta
1.19
1.21
HugePageStorageMediumSize
true
GA
1.22
-
HugePages
false
Alpha
1.8
1.9
HugePages
true
Beta
1.10
1.13
HugePages
true
GA
1.14
-
HyperVContainer
false
Alpha
1.10
1.19
HyperVContainer
false
Deprecated
1.20
-
IdentifyPodOS
false
Alpha
1.23
1.23
IdentifyPodOS
true
Beta
1.24
1.24
IdentifyPodOS
true
GA
1.25
-
IPv6DualStack
false
Alpha
1.15
1.20
IPv6DualStack
true
Beta
1.21
1.22
IPv6DualStack
true
GA
1.23
-
ImmutableEphemeralVolumes
false
Alpha
1.18
1.18
ImmutableEphemeralVolumes
true
Beta
1.19
1.20
ImmutableEphemeralVolumes
true
GA
1.21
IndexedJob
false
Alpha
1.21
1.21
IndexedJob
true
Beta
1.22
1.23
IndexedJob
true
GA
1.24
-
IngressClassNamespacedParams
false
Alpha
1.21
1.21
IngressClassNamespacedParams
true
Beta
1.22
1.22
IngressClassNamespacedParams
true
GA
1.23
-
Initializers
false
Alpha
1.7
1.13
Initializers
-
Deprecated
1.14
-
JobPodFailurePolicy
false
Alpha
1.25
-
KubeletConfigFile
false
Alpha
1.8
1.9
KubeletConfigFile
-
Deprecated
1.10
-
KubeletPluginsWatcher
false
Alpha
1.11
1.11
KubeletPluginsWatcher
true
Beta
1.12
1.12
KubeletPluginsWatcher
true
GA
1.13
-
LegacyNodeRoleBehavior
false
Alpha
1.16
1.18
LegacyNodeRoleBehavior
true
Beta
1.19
1.20
LegacyNodeRoleBehavior
false
GA
1.21
-
LegacyServiceAccountTokenNoAutoGeneration
true
Beta
1.24
LocalStorageCapacityIsolation
false
Alpha
1.7
1.9
LocalStorageCapacityIsolation
true
Beta
1.10
1.24
LocalStorageCapacityIsolation
true
GA
1.25
-
MountContainers
false
Alpha
1.9
1.16
MountContainers
false
Deprecated
1.17
-
MountPropagation
false
Alpha
1.8
1.9
MountPropagation
true
Beta
1.10
1.11
MountPropagation
true
GA
1.12
-
NamespaceDefaultLabelName
true
Beta
1.21
1.21
NamespaceDefaultLabelName
true
GA
1.22
-
NetworkPolicyEndPort
false
Alpha
1.21
1.21
NetworkPolicyEndPort
true
Beta
1.22
1.24
NetworkPolicyEndPort
true
GA
1.25
-
NodeDisruptionExclusion
false
Alpha
1.16
1.18
NodeDisruptionExclusion
true
Beta
1.19
1.20
NodeDisruptionExclusion
true
GA
1.21
-
NodeLease
false
Alpha
1.12
1.13
NodeLease
true
Beta
1.14
1.16
NodeLease
true
GA
1.17
-
NonPreemptingPriority
false
Alpha
1.15
1.18
NonPreemptingPriority
true
Beta
1.19
1.23
NonPreemptingPriority
true
GA
1.24
-
PVCProtection
false
Alpha
1.9
1.9
PVCProtection
-
Deprecated
1.10
-
PersistentLocalVolumes
false
Alpha
1.7
1.9
PersistentLocalVolumes
true
Beta
1.10
1.13
PersistentLocalVolumes
true
GA
1.14
-
PodAffinityNamespaceSelector
false
Alpha
1.21
1.21
PodAffinityNamespaceSelector
true
Beta
1.22
1.23
PodAffinityNamespaceSelector
true
GA
1.24
-
PodDisruptionBudget
false
Alpha
1.3
1.4
PodDisruptionBudget
true
Beta
1.5
1.20
PodDisruptionBudget
true
GA
1.21
-
PodDisruptionConditions
false
Alpha
1.25
-
PodOverhead
false
Alpha
1.16
1.17
PodOverhead
true
Beta
1.18
1.23
PodOverhead
true
GA
1.24
-
PodPriority
false
Alpha
1.8
1.10
PodPriority
true
Beta
1.11
1.13
PodPriority
true
GA
1.14
-
PodReadinessGates
false
Alpha
1.11
1.11
PodReadinessGates
true
Beta
1.12
1.13
PodReadinessGates
true
GA
1.14
-
PodShareProcessNamespace
false
Alpha
1.10
1.11
PodShareProcessNamespace
true
Beta
1.12
1.16
PodShareProcessNamespace
true
GA
1.17
-
PreferNominatedNode
false
Alpha
1.21
1.21
PreferNominatedNode
true
Beta
1.22
1.23
PreferNominatedNode
true
GA
1.24
-
RemoveSelfLink
false
Alpha
1.16
1.19
RemoveSelfLink
true
Beta
1.20
1.23
RemoveSelfLink
true
GA
1.24
-
RequestManagement
false
Alpha
1.15
1.16
RequestManagement
-
Deprecated
1.17
-
ResourceLimitsPriorityFunction
false
Alpha
1.9
1.18
ResourceLimitsPriorityFunction
-
Deprecated
1.19
-
ResourceQuotaScopeSelectors
false
Alpha
1.11
1.11
ResourceQuotaScopeSelectors
true
Beta
1.12
1.16
ResourceQuotaScopeSelectors
true
GA
1.17
-
RootCAConfigMap
false
Alpha
1.13
1.19
RootCAConfigMap
true
Beta
1.20
1.20
RootCAConfigMap
true
GA
1.21
-
RotateKubeletClientCertificate
true
Beta
1.8
1.18
RotateKubeletClientCertificate
true
GA
1.19
-
RunAsGroup
true
Beta
1.14
1.20
RunAsGroup
true
GA
1.21
-
RuntimeClass
false
Alpha
1.12
1.13
RuntimeClass
true
Beta
1.14
1.19
RuntimeClass
true
GA
1.20
-
SCTPSupport
false
Alpha
1.12
1.18
SCTPSupport
true
Beta
1.19
1.19
SCTPSupport
true
GA
1.20
-
ScheduleDaemonSetPods
false
Alpha
1.11
1.11
ScheduleDaemonSetPods
true
Beta
1.12
1.16
ScheduleDaemonSetPods
true
GA
1.17
-
SelectorIndex
false
Alpha
1.18
1.18
SelectorIndex
true
Beta
1.19
1.19
SelectorIndex
true
GA
1.20
-
ServerSideApply
false
Alpha
1.14
1.15
ServerSideApply
true
Beta
1.16
1.21
ServerSideApply
true
GA
1.22
-
ServiceAccountIssuerDiscovery
false
Alpha
1.18
1.19
ServiceAccountIssuerDiscovery
true
Beta
1.20
1.20
ServiceAccountIssuerDiscovery
true
GA
1.21
-
ServiceAppProtocol
false
Alpha
1.18
1.18
ServiceAppProtocol
true
Beta
1.19
1.19
ServiceAppProtocol
true
GA
1.20
-
ServiceLBNodePortControl
false
Alpha
1.20
1.21
ServiceLBNodePortControl
true
Beta
1.22
1.23
ServiceLBNodePortControl
true
GA
1.24
-
ServiceLoadBalancerClass
false
Alpha
1.21
1.21
ServiceLoadBalancerClass
true
Beta
1.22
1.23
ServiceLoadBalancerClass
true
GA
1.24
-
ServiceLoadBalancerFinalizer
false
Alpha
1.15
1.15
ServiceLoadBalancerFinalizer
true
Beta
1.16
1.16
ServiceLoadBalancerFinalizer
true
GA
1.17
-
ServiceNodeExclusion
false
Alpha
1.8
1.18
ServiceNodeExclusion
true
Beta
1.19
1.20
ServiceNodeExclusion
true
GA
1.21
-
ServiceTopology
false
Alpha
1.17
1.19
ServiceTopology
false
Deprecated
1.20
-
SetHostnameAsFQDN
false
Alpha
1.19
1.19
SetHostnameAsFQDN
true
Beta
1.20
1.21
SetHostnameAsFQDN
true
GA
1.22
-
StartupProbe
false
Alpha
1.16
1.17
StartupProbe
true
Beta
1.18
1.19
StartupProbe
true
GA
1.20
-
StatefulSetMinReadySeconds
false
Alpha
1.22
1.22
StatefulSetMinReadySeconds
true
Beta
1.23
1.24
StatefulSetMinReadySeconds
true
GA
1.25
-
StorageObjectInUseProtection
true
Beta
1.10
1.10
StorageObjectInUseProtection
true
GA
1.11
-
StreamingProxyRedirects
false
Beta
1.5
1.5
StreamingProxyRedirects
true
Beta
1.6
1.17
StreamingProxyRedirects
true
Deprecated
1.18
1.21
StreamingProxyRedirects
false
Deprecated
1.22
-
SupportIPVSProxyMode
false
Alpha
1.8
1.8
SupportIPVSProxyMode
false
Beta
1.9
1.9
SupportIPVSProxyMode
true
Beta
1.10
1.10
SupportIPVSProxyMode
true
GA
1.11
-
SupportNodePidsLimit
false
Alpha
1.14
1.14
SupportNodePidsLimit
true
Beta
1.15
1.19
SupportNodePidsLimit
true
GA
1.20
-
SupportPodPidsLimit
false
Alpha
1.10
1.13
SupportPodPidsLimit
true
Beta
1.14
1.19
SupportPodPidsLimit
true
GA
1.20
-
SuspendJob
false
Alpha
1.21
1.21
SuspendJob
true
Beta
1.22
1.23
SuspendJob
true
GA
1.24
-
Sysctls
true
Beta
1.11
1.20
Sysctls
true
GA
1.21
-
TTLAfterFinished
false
Alpha
1.12
1.20
TTLAfterFinished
true
Beta
1.21
1.22
TTLAfterFinished
true
GA
1.23
-
TaintBasedEvictions
false
Alpha
1.6
1.12
TaintBasedEvictions
true
Beta
1.13
1.17
TaintBasedEvictions
true
GA
1.18
-
TaintNodesByCondition
false
Alpha
1.8
1.11
TaintNodesByCondition
true
Beta
1.12
1.16
TaintNodesByCondition
true
GA
1.17
-
TokenRequest
false
Alpha
1.10
1.11
TokenRequest
true
Beta
1.12
1.19
TokenRequest
true
GA
1.20
-
TokenRequestProjection
false
Alpha
1.11
1.11
TokenRequestProjection
true
Beta
1.12
1.19
TokenRequestProjection
true
GA
1.20
-
ValidateProxyRedirects
false
Alpha
1.12
1.13
ValidateProxyRedirects
true
Beta
1.14
1.21
ValidateProxyRedirects
true
Deprecated
1.22
-
VolumePVCDataSource
false
Alpha
1.15
1.15
VolumePVCDataSource
true
Beta
1.16
1.17
VolumePVCDataSource
true
GA
1.18
-
VolumeScheduling
false
Alpha
1.9
1.9
VolumeScheduling
true
Beta
1.10
1.12
VolumeScheduling
true
GA
1.13
-
VolumeSnapshotDataSource
false
Alpha
1.12
1.16
VolumeSnapshotDataSource
true
Beta
1.17
1.19
VolumeSnapshotDataSource
true
GA
1.20
-
VolumeSubpath
true
GA
1.10
-
VolumeSubpathEnvExpansion
false
Alpha
1.14
1.14
VolumeSubpathEnvExpansion
true
Beta
1.15
1.16
VolumeSubpathEnvExpansion
true
GA
1.17
-
WarningHeaders
true
Beta
1.19
1.21
WarningHeaders
true
GA
1.22
-
WatchBookmark
false
Alpha
1.15
1.15
WatchBookmark
true
Beta
1.16
1.16
WatchBookmark
true
GA
1.17
-
WindowsEndpointSliceProxying
false
Alpha
1.19
1.20
WindowsEndpointSliceProxying
true
Beta
1.21
1.21
WindowsEndpointSliceProxying
true
GA
1.22
-
WindowsGMSA
false
Alpha
1.14
1.15
WindowsGMSA
true
Beta
1.16
1.17
WindowsGMSA
true
GA
1.18
-
WindowsRunAsUserName
false
Alpha
1.16
1.16
WindowsRunAsUserName
true
Beta
1.17
1.17
WindowsRunAsUserName
true
GA
1.18
-
Using a feature
Feature stages
A feature can be in Alpha, Beta or GA stage.
An Alpha feature means:
Disabled by default.
Might be buggy. Enabling the feature may expose bugs.
Support for feature may be dropped at any time without notice.
The API may change in incompatible ways in a later software release without notice.
Recommended for use only in short-lived testing clusters, due to increased
risk of bugs and lack of long-term support.
A Beta feature means:
Enabled by default.
The feature is well tested. Enabling the feature is considered safe.
Support for the overall feature will not be dropped, though details may change.
The schema and/or semantics of objects may change in incompatible ways in a
subsequent beta or stable release. When this happens, we will provide instructions
for migrating to the next version. This may require deleting, editing, and
re-creating API objects. The editing process may require some thought.
This may require downtime for applications that rely on the feature.
Recommended for only non-business-critical uses because of potential for
incompatible changes in subsequent releases. If you have multiple clusters
that can be upgraded independently, you may be able to relax this restriction.
Note: Please do try Beta features and give feedback on them!
After they exit beta, it may not be practical for us to make more changes.
A General Availability (GA) feature is also referred to as a stable feature. It means:
The feature is always enabled; you cannot disable it.
The corresponding feature gate is no longer needed.
Stable versions of features will appear in released software for many subsequent versions.
List of feature gates
Each feature gate is designed for enabling/disabling a specific feature:
APIListChunking: Enable the API clients to retrieve (LIST or GET)
resources from API server in chunks.
APIPriorityAndFairness: Enable managing request concurrency with
prioritization and fairness at each server. (Renamed from RequestManagement)
APIResponseCompression: Compress the API responses for LIST or GET requests.
APIServerIdentity: Assign each API server an ID in a cluster.
Accelerators: Provided an early form of plugin to enable Nvidia GPU support when using
Docker Engine; no longer available. See
Device Plugins for
an alternative.
AllowExtTrafficLocalEndpoints: Enable a service to route external requests to node local endpoints.
AllowInsecureBackendProxy: Enable the users to skip TLS verification of
kubelets on Pod log requests.
AnyVolumeDataSource: Enable use of any custom resource as the DataSource of a
PVC.
AppArmor: Enable use of AppArmor mandatory access control for Pods running on Linux nodes.
See AppArmor Tutorial for more details.
AttachVolumeLimit: Enable volume plugins to report limits on number of volumes
that can be attached to a node.
See dynamic volume limits
for more details.
BalanceAttachedNodeVolumes: Include volume count on node to be considered for
balanced resource allocation while scheduling. A node which has closer CPU,
memory utilization, and volume count is favored by the scheduler while making decisions.
BlockVolume: Enable the definition and consumption of raw block devices in Pods.
See Raw Block Volume Support
for more details.
BoundServiceAccountTokenVolume: Migrate ServiceAccount volumes to use a projected volume
consisting of a ServiceAccountTokenVolumeProjection. Cluster admins can use metric
serviceaccount_stale_tokens_total to monitor workloads that are depending on the extended
tokens. If there are no such workloads, turn off extended tokens by starting kube-apiserver with
flag --service-account-extend-token-expiration=false.
Check Bound Service Account Tokens
for more details.
CheckpointContainer: Enables the kubelet checkpoint API.
See Kubelet Checkpoint API for more details.
ControllerManagerLeaderMigration: Enables Leader Migration for
kube-controller-manager and
cloud-controller-manager
which allows a cluster operator to live migrate
controllers from the kube-controller-manager into an external controller-manager
(e.g. the cloud-controller-manager) in an HA cluster without downtime.
CPUManagerPolicyAlphaOptions: This allows fine-tuning of CPUManager policies,
experimental, Alpha-quality options
This feature gate guards a group of CPUManager options whose quality level is alpha.
This feature gate will never graduate to beta or stable.
CPUManagerPolicyBetaOptions: This allows fine-tuning of CPUManager policies,
experimental, Beta-quality options
This feature gate guards a group of CPUManager options whose quality level is beta.
This feature gate will never graduate to stable.
CPUManagerPolicyOptions: Allow fine-tuning of CPUManager policies.
CRIContainerLogRotation: Enable container log rotation for CRI container runtime.
The default max size of a log file is 10MB and the default max number of
log files allowed for a container is 5.
These values can be configured in the kubelet config.
See logging at node level
for more details.
CSIBlockVolume: Enable external CSI volume drivers to support block storage.
See csi raw block volume support
for more details.
CSIDriverRegistry: Enable all logic related to the CSIDriver API object in
csi.storage.k8s.io.
CSIInlineVolume: Enable CSI Inline volumes support for pods.
CSIMigration: Enables shims and translation logic to route volume
operations from in-tree plugins to corresponding pre-installed CSI plugins
CSIMigrationAWS: Enables shims and translation logic to route volume
operations from the AWS-EBS in-tree plugin to EBS CSI plugin. Supports
falling back to in-tree EBS plugin for mount operations to nodes that have
the feature disabled or that do not have EBS CSI plugin installed and
configured. Does not support falling back for provision operations, for those
the CSI plugin must be installed and configured.
CSIMigrationAWSComplete: Stops registering the EBS in-tree plugin in
kubelet and volume controllers and enables shims and translation logic to
route volume operations from the AWS-EBS in-tree plugin to EBS CSI plugin.
Requires CSIMigration and CSIMigrationAWS feature flags enabled and EBS CSI
plugin installed and configured on all nodes in the cluster. This flag has
been deprecated in favor of the InTreePluginAWSUnregister feature flag
which prevents the registration of in-tree EBS plugin.
CSIMigrationAzureDisk: Enables shims and translation logic to route volume
operations from the Azure-Disk in-tree plugin to AzureDisk CSI plugin.
Supports falling back to in-tree AzureDisk plugin for mount operations to
nodes that have the feature disabled or that do not have AzureDisk CSI plugin
installed and configured. Does not support falling back for provision
operations, for those the CSI plugin must be installed and configured.
Requires CSIMigration feature flag enabled.
CSIMigrationAzureDiskComplete: Stops registering the Azure-Disk in-tree
plugin in kubelet and volume controllers and enables shims and translation
logic to route volume operations from the Azure-Disk in-tree plugin to
AzureDisk CSI plugin. Requires CSIMigration and CSIMigrationAzureDisk feature
flags enabled and AzureDisk CSI plugin installed and configured on all nodes
in the cluster. This flag has been deprecated in favor of the
InTreePluginAzureDiskUnregister feature flag which prevents the registration
of in-tree AzureDisk plugin.
CSIMigrationAzureFile: Enables shims and translation logic to route volume
operations from the Azure-File in-tree plugin to AzureFile CSI plugin.
Supports falling back to in-tree AzureFile plugin for mount operations to
nodes that have the feature disabled or that do not have AzureFile CSI plugin
installed and configured. Does not support falling back for provision
operations, for those the CSI plugin must be installed and configured.
Requires CSIMigration feature flag enabled.
CSIMigrationAzureFileComplete: Stops registering the Azure-File in-tree
plugin in kubelet and volume controllers and enables shims and translation
logic to route volume operations from the Azure-File in-tree plugin to
AzureFile CSI plugin. Requires CSIMigration and CSIMigrationAzureFile feature
flags enabled and AzureFile CSI plugin installed and configured on all nodes
in the cluster. This flag has been deprecated in favor of the
InTreePluginAzureFileUnregister feature flag which prevents the registration
of in-tree AzureFile plugin.
CSIMigrationGCE: Enables shims and translation logic to route volume
operations from the GCE-PD in-tree plugin to PD CSI plugin. Supports falling
back to in-tree GCE plugin for mount operations to nodes that have the
feature disabled or that do not have PD CSI plugin installed and configured.
Does not support falling back for provision operations, for those the CSI
plugin must be installed and configured. Requires CSIMigration feature flag
enabled.
CSIMigrationGCEComplete: Stops registering the GCE-PD in-tree plugin in
kubelet and volume controllers and enables shims and translation logic to
route volume operations from the GCE-PD in-tree plugin to PD CSI plugin.
Requires CSIMigration and CSIMigrationGCE feature flags enabled and PD CSI
plugin installed and configured on all nodes in the cluster. This flag has
been deprecated in favor of the InTreePluginGCEUnregister feature flag which
prevents the registration of in-tree GCE PD plugin.
CSIMigrationOpenStack: Enables shims and translation logic to route volume
operations from the Cinder in-tree plugin to Cinder CSI plugin. Supports
falling back to in-tree Cinder plugin for mount operations to nodes that have
the feature disabled or that do not have Cinder CSI plugin installed and
configured. Does not support falling back for provision operations, for those
the CSI plugin must be installed and configured. Requires CSIMigration
feature flag enabled.
CSIMigrationOpenStackComplete: Stops registering the Cinder in-tree plugin in
kubelet and volume controllers and enables shims and translation logic to route
volume operations from the Cinder in-tree plugin to Cinder CSI plugin.
Requires CSIMigration and CSIMigrationOpenStack feature flags enabled and Cinder
CSI plugin installed and configured on all nodes in the cluster. This flag has
been deprecated in favor of the InTreePluginOpenStackUnregister feature flag
which prevents the registration of in-tree openstack cinder plugin.
csiMigrationRBD: Enables shims and translation logic to route volume
operations from the RBD in-tree plugin to Ceph RBD CSI plugin. Requires
CSIMigration and csiMigrationRBD feature flags enabled and Ceph CSI plugin
installed and configured in the cluster. This flag has been deprecated in
favor of the InTreePluginRBDUnregister feature flag which prevents the registration of
in-tree RBD plugin.
CSIMigrationvSphere: Enables shims and translation logic to route volume operations
from the vSphere in-tree plugin to vSphere CSI plugin. Supports falling back
to in-tree vSphere plugin for mount operations to nodes that have the feature
disabled or that do not have vSphere CSI plugin installed and configured.
Does not support falling back for provision operations, for those the CSI
plugin must be installed and configured. Requires CSIMigration feature flag
enabled.
CSIMigrationvSphereComplete: Stops registering the vSphere in-tree plugin in kubelet
and volume controllers and enables shims and translation logic to route volume operations
from the vSphere in-tree plugin to vSphere CSI plugin. Requires CSIMigration and
CSIMigrationvSphere feature flags enabled and vSphere CSI plugin installed and
configured on all nodes in the cluster. This flag has been deprecated in favor
of the InTreePluginvSphereUnregister feature flag which prevents the
registration of in-tree vsphere plugin.
CSIMigrationPortworx: Enables shims and translation logic to route volume operations
from the Portworx in-tree plugin to Portworx CSI plugin.
Requires Portworx CSI driver to be installed and configured in the cluster.
CSINodeInfo: Enable all logic related to the CSINodeInfo API object in csi.storage.k8s.io.
CSINodeExpandSecret: Enable passing secret authentication data to a CSI driver for use
during a NodeExpandVolume CSI operation.
CSIPersistentVolume: Enable discovering and mounting volumes provisioned through a
CSI (Container Storage Interface)
compatible volume plugin.
CSIServiceAccountToken: Enable CSI drivers to receive the pods' service account token
that they mount volumes for. See
Token Requests.
CSIStorageCapacity: Enables CSI drivers to publish storage capacity information
and the Kubernetes scheduler to use that information when scheduling pods. See
Storage Capacity.
Check the csi volume type documentation for more details.
CSIVolumeFSGroupPolicy: Allows CSIDrivers to use the fsGroupPolicy field.
This field controls whether volumes created by a CSIDriver support volume ownership
and permission modifications when these volumes are mounted.
CSIVolumeHealth: Enable support for CSI volume health monitoring on node.
CSRDuration: Allows clients to request a duration for certificates issued
via the Kubernetes CSR API.
ContextualLogging: When you enable this feature gate, Kubernetes components that support
contextual logging add extra detail to log output.
ControllerManagerLeaderMigration: Enables leader migration for
kube-controller-manager and cloud-controller-manager.
CronJobControllerV2: Use an alternative implementation of the
CronJob controller. Otherwise,
version 1 of the same controller is selected.
CronJobTimeZone: Allow the use of the timeZone optional field in CronJobs
CustomCPUCFSQuotaPeriod: Enable nodes to change cpuCFSQuotaPeriod in
kubelet config.
CustomResourceValidationExpressions: Enable expression language validation in CRD
which will validate customer resource based on validation rules written in
the x-kubernetes-validations extension.
CustomPodDNS: Enable customizing the DNS settings for a Pod using its dnsConfig property.
Check Pod's DNS Config
for more details.
CustomResourceDefaulting: Enable CRD support for default values in OpenAPI v3 validation schemas.
CustomResourcePublishOpenAPI: Enables publishing of CRD OpenAPI specs.
CustomResourceSubresources: Enable /status and /scale subresources
on resources created from CustomResourceDefinition.
CustomResourceValidation: Enable schema based validation on resources created from
CustomResourceDefinition.
CustomResourceWebhookConversion: Enable webhook-based conversion
on resources created from CustomResourceDefinition.
DefaultPodTopologySpread: Enables the use of PodTopologySpread scheduling plugin to do
default spreading.
DelegateFSGroupToCSIDriver: If supported by the CSI driver, delegates the
role of applying fsGroup from a Pod's securityContext to the driver by
passing fsGroup through the NodeStageVolume and NodePublishVolume CSI calls.
DevicePlugins: Enable the device-plugins
based resource provisioning on nodes.
DisableCloudProviders: Disables any functionality in kube-apiserver,
kube-controller-manager and kubelet related to the --cloud-provider
component flag.
DisableKubeletCloudCredentialProviders: Disable the in-tree functionality in kubelet
to authenticate to a cloud provider container registry for image pull credentials.
DownwardAPIHugePages: Enables usage of hugepages in
downward API.
DryRun: Enable server-side dry run requests
so that validation, merging, and mutation can be tested without committing.
DynamicAuditing: Used to enable dynamic auditing before v1.19.
DynamicKubeletConfig: Enable the dynamic configuration of kubelet. The
feature is no longer supported outside of supported skew policy. The feature
gate was removed from kubelet in 1.24. See Reconfigure kubelet.
DynamicProvisioningScheduling: Extend the default scheduler to be aware of
volume topology and handle PV provisioning.
This feature is superseded by the VolumeScheduling feature completely in v1.12.
DynamicVolumeProvisioning: Enable the
dynamic provisioning of persistent volumes to Pods.
EfficientWatchResumption: Allows for storage-originated bookmark (progress
notify) events to be delivered to the users. This is only applied to watch
operations.
EnableAggregatedDiscoveryTimeout: Enable the five second
timeout on aggregated discovery calls.
EnableEquivalenceClassCache: Enable the scheduler to cache equivalence of
nodes when scheduling Pods.
EndpointSlice: Enables EndpointSlices for more scalable and extensible
network endpoints. See Enabling EndpointSlices.
EndpointSliceProxying: When enabled, kube-proxy running
on Linux will use EndpointSlices as the primary data source instead of
Endpoints, enabling scalability and performance improvements. See
Enabling Endpoint Slices.
EndpointSliceTerminatingCondition: Enables EndpointSlice terminating and serving
condition fields.
EphemeralContainers: Enable the ability to add
ephemeral containers
to running pods.
ExecProbeTimeout: Ensure kubelet respects exec probe timeouts.
This feature gate exists in case any of your existing workloads depend on a
now-corrected fault where Kubernetes ignored exec probe timeouts. See
readiness probes.
ExpandCSIVolumes: Enable the expanding of CSI volumes.
ExpandedDNSConfig: Enable kubelet and kube-apiserver to allow more DNS
search paths and longer list of DNS search paths. This feature requires container
runtime support(Containerd: v1.5.6 or higher, CRI-O: v1.22 or higher). See
Expanded DNS Configuration.
ExperimentalCriticalPodAnnotation: Enable annotating specific pods as critical
so that their scheduling is guaranteed.
This feature is deprecated by Pod Priority and Preemption as of v1.13.
ExperimentalHostUserNamespaceDefaulting: Enabling the defaulting user
namespace to host. This is for containers that are using other host namespaces,
host mounts, or containers that are privileged or using specific non-namespaced
capabilities (e.g. MKNODE, SYS_MODULE etc.). This should only be enabled
if user namespace remapping is enabled in the Docker daemon.
ExternalPolicyForExternalIP: Fix a bug where ExternalTrafficPolicy is not
applied to Service ExternalIPs.
GCERegionalPersistentDisk: Enable the regional PD feature on GCE.
GenericEphemeralVolume: Enables ephemeral, inline volumes that support all features
of normal volumes (can be provided by third-party storage vendors, storage capacity tracking,
restore from snapshot, etc.).
See Ephemeral Volumes.
GracefulNodeShutdown: Enables support for graceful shutdown in kubelet.
During a system shutdown, kubelet will attempt to detect the shutdown event
and gracefully terminate pods running on the node. See
Graceful Node Shutdown
for more details.
GracefulNodeShutdownBasedOnPodPriority: Enables the kubelet to check Pod priorities
when shutting down a node gracefully.
HonorPVReclaimPolicy: Honor persistent volume reclaim policy when it is Delete irrespective of PV-PVC deletion ordering.
For more details, check the
PersistentVolume deletion protection finalizer
documentation.
HPAContainerMetrics: Enable the HorizontalPodAutoscaler to scale based on
metrics from individual containers in target pods.
HPAScaleToZero: Enables setting minReplicas to 0 for HorizontalPodAutoscaler
resources when using custom or external metrics.
HugePages: Enable the allocation and consumption of pre-allocated
huge pages.
HugePageStorageMediumSize: Enable support for multiple sizes pre-allocated
huge pages.
IdentifyPodOS: Allows the Pod OS field to be specified. This helps in identifying
the OS of the pod authoritatively during the API server admission time.
In Kubernetes 1.25, the allowed values for the pod.spec.os.name
are windows and linux.
ImmutableEphemeralVolumes: Allows for marking individual Secrets and ConfigMaps as
immutable for better safety and performance.
IndexedJob: Allows the Job
controller to manage Pod completions per completion index.
IngressClassNamespacedParams: Allow namespace-scoped parameters reference in
IngressClass resource. This feature adds two fields - Scope and Namespace
to IngressClass.spec.parameters.
Initializers: Allow asynchronous coordination of object creation using the
Initializers admission plugin.
InTreePluginAWSUnregister: Stops registering the aws-ebs in-tree plugin in kubelet
and volume controllers.
InTreePluginAzureDiskUnregister: Stops registering the azuredisk in-tree plugin in kubelet
and volume controllers.
InTreePluginAzureFileUnregister: Stops registering the azurefile in-tree plugin in kubelet
and volume controllers.
InTreePluginGCEUnregister: Stops registering the gce-pd in-tree plugin in kubelet
and volume controllers.
InTreePluginOpenStackUnregister: Stops registering the OpenStack cinder in-tree plugin in kubelet
and volume controllers.
InTreePluginPortworxUnregister: Stops registering the Portworx in-tree plugin in kubelet
and volume controllers.
InTreePluginRBDUnregister: Stops registering the RBD in-tree plugin in kubelet
and volume controllers.
InTreePluginvSphereUnregister: Stops registering the vSphere in-tree plugin in kubelet
and volume controllers.
IPv6DualStack: Enable dual stack
support for IPv6.
JobMutableNodeSchedulingDirectives: Allows updating node scheduling directives in
the pod template of Job.
JobPodFailurePolicy: Allow users to specify handling of pod failures based on container exit codes and pod conditions.
JobReadyPods: Enables tracking the number of Pods that have a Readycondition.
The count of Ready pods is recorded in the
status
of a Job status.
JobTrackingWithFinalizers: Enables tracking Job
completions without relying on Pods remaining in the cluster indefinitely.
The Job controller uses Pod finalizers and a field in the Job status to keep
track of the finished Pods to count towards completion.
KubeletPluginsWatcher: Enable probe-based plugin watcher utility to enable kubelet
to discover plugins such as CSI volume drivers.
KubeletPodResources: Enable the kubelet's pod resources gRPC endpoint. See
Support Device Monitoring
for more details.
KubeletPodResourcesGetAllocatable: Enable the kubelet's pod resources
GetAllocatableResources functionality. This API augments the
resource allocation reporting
with informations about the allocatable resources, enabling clients to properly
track the free compute resources on a node.
LegacyNodeRoleBehavior: When disabled, legacy behavior in service load balancers and
node disruption will ignore the node-role.kubernetes.io/master label in favor of the
feature-specific labels provided by NodeDisruptionExclusion and ServiceNodeExclusion.
KubeletTracing: Add support for distributed tracing in the kubelet.
When enabled, kubelet CRI interface and authenticated http servers are instrumented to generate
OpenTelemetry trace spans.
See Traces for Kubernetes System Components for more details.
LegacyServiceAccountTokenNoAutoGeneration: Stop auto-generation of Secret-based
service account tokens.
LocalStorageCapacityIsolationFSQuotaMonitoring: When LocalStorageCapacityIsolation
is enabled for
local ephemeral storage
and the backing filesystem for emptyDir volumes
supports project quotas and they are enabled, use project quotas to monitor
emptyDir volume storage consumption rather than
filesystem walk for better performance and accuracy.
LogarithmicScaleDown: Enable semi-random selection of pods to evict on controller scaledown
based on logarithmic bucketing of pod timestamps.
MaxUnavailableStatefulSet: Enables setting the maxUnavailable field for the
rolling update strategy
of a StatefulSet. The field specifies the maximum number of Pods
that can be unavailable during the update.
MemoryManager: Allows setting memory affinity for a container based on
NUMA topology.
MemoryQoS: Enable memory protection and usage throttle on pod / container using
cgroup v2 memory controller.
MixedProtocolLBService: Enable using different protocols in the same LoadBalancer type
Service instance.
MountContainers: Enable using utility containers on host as the volume mounter.
MountPropagation: Enable sharing volume mounted by one container to other containers or pods.
For more details, please see mount propagation.
NamespaceDefaultLabelName: Configure the API Server to set an immutable
labelkubernetes.io/metadata.name
on all namespaces, containing the namespace name.
NetworkPolicyEndPort: Enable use of the field endPort in NetworkPolicy objects,
allowing the selection of a port range instead of a single port.
NetworkPolicyStatus: Enable the status subresource for NetworkPolicy objects.
NodeInclusionPolicyInPodTopologySpread: Enable using nodeAffinityPolicy and nodeTaintsPolicy in
Pod topology spread constraints
when calculating pod topology spread skew.
NodeDisruptionExclusion: Enable use of the Node label node.kubernetes.io/exclude-disruption
which prevents nodes from being evacuated during zone failures.
NodeLease: Enable the new Lease API to report node heartbeats, which could be used as a node health signal.
NodeOutOfServiceVolumeDetach: When a Node is marked out-of-service using the
node.kubernetes.io/out-of-service taint, Pods on the node will be forcefully deleted
if they can not tolerate this taint, and the volume detach operations for Pods terminating
on the node will happen immediately. The deleted Pods can recover quickly on different nodes.
NodeSwap: Enable the kubelet to allocate swap memory for Kubernetes workloads on a node.
Must be used with KubeletConfiguration.failSwapOn set to false.
For more details, please see swap memory
NonPreemptingPriority: Enable preemptionPolicy field for PriorityClass and Pod.
OpenAPIEnums: Enables populating "enum" fields of OpenAPI schemas in the
spec returned from the API server.
OpenAPIV3: Enables the API server to publish OpenAPI v3.
PodDeletionCost: Enable the Pod Deletion Cost
feature which allows users to influence ReplicaSet downscaling order.
PersistentLocalVolumes: Enable the usage of local volume type in Pods.
Pod affinity has to be specified if requesting a local volume.
PodAndContainerStatsFromCRI: Configure the kubelet to gather container and
pod stats from the CRI container runtime rather than gathering them from cAdvisor.
PodDisruptionConditions: Enables support for appending a dedicated pod condition indicating that the pod is being deleted due to a disruption.
PodHasNetworkCondition: Enable the kubelet to mark the PodHasNetwork condition on pods.
PodOverhead: Enable the PodOverhead
feature to account for pod overheads.
PodPriority: Enable the descheduling and preemption of Pods based on their
priorities.
PodReadinessGates: Enable the setting of PodReadinessGate field for extending
Pod readiness evaluation. See Pod readiness gate
for more details.
PodSecurity: Enables the PodSecurity admission plugin.
PodShareProcessNamespace: Enable the setting of shareProcessNamespace in a Pod for sharing
a single process namespace between containers running in a pod. More details can be found in
Share Process Namespace between Containers in a Pod.
PreferNominatedNode: This flag tells the scheduler whether the nominated
nodes will be checked first before looping through all the other nodes in
the cluster.
ProcMountType: Enables control over the type proc mounts for containers
by setting the procMount field of a SecurityContext.
ProxyTerminatingEndpoints: Enable the kube-proxy to handle terminating
endpoints when ExternalTrafficPolicy=Local.
PVCProtection: Enable the prevention of a PersistentVolumeClaim (PVC) from
being deleted when it is still used by any Pod.
QOSReserved: Allows resource reservations at the QoS level preventing pods
at lower QoS levels from bursting into resources requested at higher QoS levels
(memory only for now).
ReadWriteOncePod: Enables the usage of ReadWriteOncePod PersistentVolume
access mode.
RecoverVolumeExpansionFailure: Enables users to edit their PVCs to smaller
sizes so as they can recover from previously issued volume expansion failures.
See Recovering from Failure when Expanding Volumes
for more details.
RemainingItemCount: Allow the API servers to show a count of remaining
items in the response to a
chunking list request.
RemoveSelfLink: Sets the .metadata.selfLink field to blank (empty string) for all
objects and collections. This field has been deprecated since the Kubernetes v1.16
release. When this feature is enabled, the .metadata.selfLink field remains part of
the Kubernetes API, but is always unset.
RequestManagement: Enables managing request concurrency with prioritization and fairness
at each API server. Deprecated by APIPriorityAndFairness since 1.17.
ResourceLimitsPriorityFunction: Enable a scheduler priority function that
assigns a lowest possible score of 1 to a node that satisfies at least one of
the input Pod's cpu and memory limits. The intent is to break ties between
nodes with same scores.
RootCAConfigMap: Configure the kube-controller-manager to publish a
ConfigMap named kube-root-ca.crt
to every namespace. This ConfigMap contains a CA bundle used for verifying connections
to the kube-apiserver. See
Bound Service Account Tokens
for more details.
RotateKubeletClientCertificate: Enable the rotation of the client TLS certificate on the kubelet.
See kubelet configuration
for more details.
RotateKubeletServerCertificate: Enable the rotation of the server TLS certificate on the kubelet.
See kubelet configuration
for more details.
RunAsGroup: Enable control over the primary group ID set on the init
processes of containers.
RuntimeClass: Enable the RuntimeClass feature
for selecting container runtime configurations.
ScheduleDaemonSetPods: Enable DaemonSet Pods to be scheduled by the default scheduler
instead of the DaemonSet controller.
SCTPSupport: Enables the SCTPprotocol value in Pod, Service,
Endpoints, EndpointSlice, and NetworkPolicy definitions.
SeccompDefault: Enables the use of RuntimeDefault as the default seccomp profile
for all workloads.
The seccomp profile is specified in the securityContext of a Pod and/or a Container.
SelectorIndex: Allows label and field based indexes in API server watch
cache to accelerate list operations.
SELinuxMountReadWriteOncePod: Allows kubelet to mount volumes for a Pod directly with the
right SELinux label instead of applying the SELinux label recursively on every file on the
volume.
ServerSideFieldValidation: Enables server-side field validation. This means the validation
of resource schema is performed at the API server side rather than the client side
(for example, the kubectl create or kubectl apply command line).
ServiceAccountIssuerDiscovery: Enable OIDC discovery endpoints (issuer and
JWKS URLs) for the service account issuer in the API server. See
Configure Service Accounts for Pods
for more details.
ServiceAppProtocol: Enables the appProtocol field on Services and Endpoints.
ServiceInternalTrafficPolicy: Enables the internalTrafficPolicy field on Services
ServiceLBNodePortControl: Enables the allocateLoadBalancerNodePorts field on Services.
ServiceLoadBalancerFinalizer: Enable finalizer protection for Service load balancers.
ServiceNodeExclusion: Enable the exclusion of nodes from load balancers
created by a cloud provider. A node is eligible for exclusion if labelled with
"node.kubernetes.io/exclude-from-external-load-balancers".
ServiceTopology: Enable service to route traffic based upon the Node
topology of the cluster. See
ServiceTopology
for more details.
ServiceIPStaticSubrange: Enables a strategy for Services ClusterIP allocations, whereby the
ClusterIP range is subdivided. Dynamic allocated ClusterIP addresses will be allocated preferently
from the upper range allowing users to assign static ClusterIPs from the lower range with a low
risk of collision. See
Avoiding collisions
for more details.
SetHostnameAsFQDN: Enable the ability of setting Fully Qualified Domain
Name(FQDN) as the hostname of a pod. See
Pod's setHostnameAsFQDN field.
SizeMemoryBackedVolumes: Enable kubelets to determine the size limit for
memory-backed volumes (mainly emptyDir volumes).
StartupProbe: Enable the
startup
probe in the kubelet.
StatefulSetMinReadySeconds: Allows minReadySeconds to be respected by
the StatefulSet controller.
StorageObjectInUseProtection: Postpone the deletion of PersistentVolume or
PersistentVolumeClaim objects if they are still being used.
StorageVersionHash: Allow API servers to expose the storage version hash in the
discovery.
StreamingProxyRedirects: Instructs the API server to intercept (and follow)
redirects from the backend (kubelet) for streaming requests.
Examples of streaming requests include the exec, attach and port-forward requests.
SupportIPVSProxyMode: Enable providing in-cluster service load balancing using IPVS.
See service proxies for more details.
SupportNodePidsLimit: Enable the support to limiting PIDs on the Node.
The parameter pid=<number> in the --system-reserved and --kube-reserved
options can be specified to ensure that the specified number of process IDs
will be reserved for the system as a whole and for Kubernetes system daemons
respectively.
SupportPodPidsLimit: Enable the support to limiting PIDs in Pods.
SuspendJob: Enable support to suspend and resume Jobs. See
the Jobs docs for
more details.
Sysctls: Enable support for namespaced kernel parameters (sysctls) that can be
set for each pod. See
sysctls for more details.
TTLAfterFinished: Allow a
TTL controller
to clean up resources after they finish execution.
TaintBasedEvictions: Enable evicting pods from nodes based on taints on Nodes
and tolerations on Pods.
See taints and tolerations
for more details.
TaintNodesByCondition: Enable automatic tainting nodes based on
node conditions.
TokenRequest: Enable the TokenRequest endpoint on service account resources.
TokenRequestProjection: Enable the injection of service account tokens into a
Pod through a projected volume.
TopologyAwareHints: Enables topology aware routing based on topology hints
in EndpointSlices. See Topology Aware
Hints for more
details.
TopologyManager: Enable a mechanism to coordinate fine-grained hardware resource
assignments for different components in Kubernetes. See
Control Topology Management Policies on a node.
ValidateProxyRedirects: This flag controls whether the API server should
validate that redirects are only followed to the same host. Only used if the
StreamingProxyRedirects flag is enabled.
VolumeCapacityPriority: Enable support for prioritizing nodes in different
topologies based on available PV capacity.
VolumePVCDataSource: Enable support for specifying an existing PVC as a DataSource.
VolumeScheduling: Enable volume topology aware scheduling and make the
PersistentVolumeClaim (PVC) binding aware of scheduling decisions. It also
enables the usage of local volume
type when used together with the PersistentLocalVolumes feature gate.
VolumeSnapshotDataSource: Enable volume snapshot data source support.
VolumeSubpath: Allow mounting a subpath of a volume in a container.
VolumeSubpathEnvExpansion: Enable subPathExpr field for expanding environment
variables into a subPath.
WarningHeaders: Allow sending warning headers in API responses.
WatchBookmark: Enable support for watch bookmark events.
WinDSR: Allows kube-proxy to create DSR loadbalancers for Windows.
WinOverlay: Allows kube-proxy to run in overlay mode for Windows.
WindowsEndpointSliceProxying: When enabled, kube-proxy running on Windows
will use EndpointSlices as the primary data source instead of Endpoints,
enabling scalability and performance improvements. See
Enabling Endpoint Slices.
WindowsGMSA: Enables passing of GMSA credential specs from pods to container runtimes.
WindowsHostProcessContainers: Enables support for Windows HostProcess containers.
WindowsRunAsUserName : Enable support for running applications in Windows containers
with as a non-default user. See
Configuring RunAsUserName
for more details.
What's next
The deprecation policy for Kubernetes explains
the project's approach to removing features and components.
Since Kubernetes 1.24, new beta APIs are not enabled by default. When enabling a beta
feature, you will also need to enable any associated API resources.
For example, to enable a particular resource like
storage.k8s.io/v1beta1/csistoragecapacities, set --runtime-config=storage.k8s.io/v1beta1/csistoragecapacities.
See API Versioning for more details on the command line flags.
6.11.2 - kubelet
Synopsis
The kubelet is the primary "node agent" that runs on each node. It can
register the node with the apiserver using one of: the hostname; a flag to
override the hostname; or specific logic for a cloud provider.
The kubelet works in terms of a PodSpec. A PodSpec is a YAML or JSON object
that describes a pod. The kubelet takes a set of PodSpecs that are provided
through various mechanisms (primarily through the apiserver) and ensures that
the containers described in those PodSpecs are running and healthy. The
kubelet doesn't manage containers which were not created by Kubernetes.
Other than from a PodSpec from the apiserver, there are three ways that a
container manifest can be provided to the Kubelet.
File: Path passed as a flag on the command line. Files under this path will be
monitored periodically for updates. The monitoring period is 20s by default
and is configurable via a flag.
HTTP endpoint: HTTP endpoint passed as a parameter on the command line. This
endpoint is checked every 20 seconds (also configurable with a flag).
HTTP server: The kubelet can also listen for HTTP and respond to a simple API
(underspec'd currently) to submit a new manifest.
kubelet [flags]
Options
--add-dir-header
If true, adds the file directory to the header of the log messages (DEPRECATED: will be removed in a future release, see here.)
--address string Default: 0.0.0.0
The IP address for the Kubelet to serve on (set to 0.0.0.0 or :: for listening in all interfaces and IP families) (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--allowed-unsafe-sysctls strings
Comma-separated whitelist of unsafe sysctls or unsafe sysctl patterns (ending in *). Use these at your own risk. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--alsologtostderr
Log to standard error as well as files (DEPRECATED: will be removed in a future release, see here.)
--anonymous-auth Default: true
Enables anonymous requests to the Kubelet server. Requests that are not rejected by another authentication method are treated as anonymous requests. Anonymous requests have a username of system:anonymous, and a group name of system:unauthenticated. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--authentication-token-webhook
Use the TokenReview API to determine authentication for bearer tokens. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
The duration to cache responses from the webhook token authenticator. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--authorization-mode string Default: AlwaysAllow
Authorization mode for Kubelet server. Valid options are AlwaysAllow or Webhook. Webhook mode uses the SubjectAccessReview API to determine authorization. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
The duration to cache 'authorized' responses from the webhook authorizer. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
The duration to cache 'unauthorized' responses from the webhook authorizer. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--azure-container-registry-config string
Path to the file containing Azure container registry configuration information.
--bootstrap-kubeconfig string
Path to a kubeconfig file that will be used to get client certificate for kubelet. If the file specified by --kubeconfig does not exist, the bootstrap kubeconfig is used to request a client certificate from the API server. On success, a kubeconfig file referencing the generated client certificate and key is written to the path specified by --kubeconfig. The client certificate and key file will be stored in the directory pointed by --cert-dir.
--cert-dir string Default: /var/lib/kubelet/pki
The directory where the TLS certs are located. If --tls-cert-file and --tls-private-key-file are provided, this flag will be ignored.
--cgroup-driver string Default: cgroupfs
Driver that the kubelet uses to manipulate cgroups on the host. Possible values: cgroupfs, systemd. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cgroup-root string Default: ''
Optional root cgroup to use for pods. This is handled by the container runtime on a best effort basis. Default: '', which means use the container runtime default. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cgroups-per-qos Default: true
Enable creation of QoS cgroup hierarchy, if true top level QoS and pod cgroups are created. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--client-ca-file string
If set, any request presenting a client certificate signed by one of the authorities in the client-ca-file is authenticated with an identity corresponding to the CommonName of the client certificate. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cloud-config string
The path to the cloud provider configuration file. Empty string for no configuration file. (DEPRECATED: will be removed in 1.24 or later, in favor of removing cloud providers code from kubelet.)
--cloud-provider string
The provider for cloud services. Set to empty string for running with no cloud provider. If set, the cloud provider determines the name of the node (consult cloud provider documentation to determine if and how the hostname is used). (DEPRECATED: will be removed in 1.24 or later, in favor of removing cloud provider code from Kubelet.)
--cluster-dns strings
Comma-separated list of DNS server IP address. This value is used for containers DNS server in case of Pods with "dnsPolicy=ClusterFirst". Note: all DNS servers appearing in the list MUST serve the same set of records otherwise name resolution within the cluster may not work correctly. There is no guarantee as to which DNS server may be contacted for name resolution. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cluster-domain string
Domain for this cluster. If set, kubelet will configure all containers to search this domain in addition to the host's search domains (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--config string
The Kubelet will load its initial configuration from this file. The path may be absolute or relative; relative paths start at the Kubelet's current working directory. Omit this flag to use the built-in default configuration values. Command-line flags override configuration from this file.
--container-log-max-files int32 Default: 5
<Warning: Beta feature> Set the maximum number of container log files that can be present for a container. The number must be >= 2. This flag can only be used with --container-runtime=remote. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--container-log-max-size string Default: 10Mi
<Warning: Beta feature> Set the maximum size (e.g. 10Mi) of container log file before it is rotated. This flag can only be used with --container-runtime=remote. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--container-runtime string Default: remote
The container runtime to use. Possible values: docker, remote. (DEPRECATED: will be removed in 1.27 as the only valid value is 'remote')
--container-runtime-endpoint string
The endpoint of remote runtime service. Unix Domain Sockets are supported on Linux, while npipe and tcp endpoints are supported on windows. Examples: unix:///path/to/runtime.sock, npipe:////./pipe/runtime.
--contention-profiling
Enable lock contention profiling, if profiling is enabled (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cpu-cfs-quota Default: true
Enable CPU CFS quota enforcement for containers that specify CPU limits (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cpu-cfs-quota-period duration Default: 100ms
Sets CPU CFS quota period value, cpu.cfs_period_us, defaults to Linux Kernel default. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cpu-manager-policy string Default: none
CPU Manager policy to use. Possible values: none, static. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--cpu-manager-policy-options mapStringString
A set of key=value CPU Manager policy options to use, to fine tune their behaviour. If not supplied, keep the default behaviour. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
<Warning: Alpha feature> CPU Manager reconciliation period. Examples: 10s, or 1m. If not supplied, defaults to node status update frequency. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--enable-controller-attach-detach Default: true
Enables the Attach/Detach controller to manage attachment/detachment of volumes scheduled to this node, and disables kubelet from executing any attach/detach operations. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--enable-debugging-handlers Default: true
Enables server endpoints for log collection and local running of containers and commands. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--enable-server Default: true
Enable the Kubelet's server. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--enforce-node-allocatable strings Default: pods
A comma separated list of levels of node allocatable enforcement to be enforced by kubelet. Acceptable options are none, pods, system-reserved, and kube-reserved. If the latter two options are specified, --system-reserved-cgroup and --kube-reserved-cgroup must also be set, respectively. If none is specified, no additional options should be set. See here for more details. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--event-burst int32 Default: 10
Maximum size of a bursty event records, temporarily allows event records to burst to this number, while still not exceeding --event-qps. The number must be >= 0. If 0 will use default burst (10). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--event-qps int32 Default: 5
QPS to limit event creations. The number must be >= 0. If 0 will use default QPS (5). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
A set of eviction thresholds (e.g. memory.available<1Gi) that if met would trigger a pod eviction. On a Linux node, the default value also includes nodefs.inodesFree<5%. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--eviction-max-pod-grace-period int32
Maximum allowed grace period (in seconds) to use when terminating pods in response to a soft eviction threshold being met. If negative, defer to pod specified value. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--eviction-minimum-reclaim mapStringString
A set of minimum reclaims (e.g. imagefs.available=2Gi) that describes the minimum amount of resource the kubelet will reclaim when performing a pod eviction if that resource is under pressure. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
Duration for which the kubelet has to wait before transitioning out of an eviction pressure condition. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--eviction-soft mapStringString
A set of eviction thresholds (e.g. memory.available<1.5Gi) that if met over a corresponding grace period would trigger a pod eviction. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--eviction-soft-grace-period mapStringString
A set of eviction grace periods (e.g. memory.available=1m30s) that correspond to how long a soft eviction threshold must hold before triggering a pod eviction. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--exit-on-lock-contention
Whether kubelet should exit upon lock-file contention.
When set to true, hard eviction thresholds will be ignored while calculating node allocatable. See here for more details. (DEPRECATED: will be removed in 1.24 or later)
--experimental-kernel-memcg-notification
Use kernelMemcgNotification configuration, this flag will be removed in 1.24 or later. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--experimental-mounter-path string Default: mount
[Experimental] Path of mounter binary. Leave empty to use the default mount. (DEPRECATED: will be removed in 1.24 or later, in favor of using CSI.)
--fail-swap-on Default: true
Makes the Kubelet fail to start if swap is enabled on the node. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--feature-gates <A list of 'key=true/false' pairs>
Duration between checking config files for new data. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--hairpin-mode string Default: promiscuous-bridge
How should the kubelet setup hairpin NAT. This allows endpoints of a Service to load balance back to themselves if they should try to access their own Service. Valid values are promiscuous-bridge, hairpin-veth and none. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--healthz-bind-address string Default: 127.0.0.1
The IP address for the healthz server to serve on (set to 0.0.0.0 or :: for listening in all interfaces and IP families). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--healthz-port int32 Default: 10248
The port of the localhost healthz endpoint (set to 0 to disable). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
-h, --help
help for kubelet
--hostname-override string
If non-empty, will use this string as identification instead of the actual hostname. If --cloud-provider is set, the cloud provider determines the name of the node (consult cloud provider documentation to determine if and how the hostname is used).
--http-check-frequency duration Default: 20s
Duration between checking HTTP for new data. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--image-credential-provider-bin-dir string
The path to the directory where credential provider plugin binaries are located.
--image-credential-provider-config string
The path to the credential provider plugin config file.
--image-gc-high-threshold int32 Default: 85
The percent of disk usage after which image garbage collection is always run. Values must be within the range [0, 100], To disable image garbage collection, set to 100. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--image-gc-low-threshold int32 Default: 80
The percent of disk usage before which image garbage collection is never run. Lowest disk usage to garbage collect to. Values must be within the range [0, 100] and should not be larger than that of --image-gc-high-threshold. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--image-service-endpoint string
[Experimental] The endpoint of remote image service. If not specified, it will be the same with --container-runtime-endpoint by default. Unix Domain Socket are supported on Linux, while npipe and TCP endpoints are supported on Windows. Examples: unix:///var/run/dockershim.sock, npipe:////./pipe/dockershim
--iptables-drop-bit int32 Default: 15
The bit of the fwmark space to mark packets for dropping. Must be within the range [0, 31]. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--iptables-masquerade-bit int32 Default: 14
The bit of the fwmark space to mark packets for SNAT. Must be within the range [0, 31]. Please match this parameter with corresponding parameter in kube-proxy. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--keep-terminated-pod-volumes
Keep terminated pod volumes mounted to the node after the pod terminates. Can be useful for debugging volume related issues. (DEPRECATED: will be removed in a future version)
--kernel-memcg-notification
If enabled, the kubelet will integrate with the kernel memcg notification to determine if memory eviction thresholds are crossed rather than polling. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--kube-api-burst int32 Default: 10
Burst to use while talking with kubernetes API server. The number must be >= 0. If 0 will use default burst (10). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
Content type of requests sent to apiserver. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--kube-api-qps int32 Default: 5
QPS to use while talking with kubernetes API server. The number must be >= 0. If 0 will use default QPS (5). Doesn't cover events and node heartbeat apis which rate limiting is controlled by a different set of flags. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--kube-reserved mapStringString Default: <None>
A set of <resource name>=<resource quantity> (e.g. cpu=200m,memory=500Mi,ephemeral-storage=1Gi,pid='100') pairs that describe resources reserved for kubernetes system components. Currently cpu, memory and local ephemeral-storage for root file system are supported. See here for more detail. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--kube-reserved-cgroup string Default: ''
Absolute name of the top level cgroup that is used to manage kubernetes components for which compute resources were reserved via --kube-reserved flag. Ex. /kube-reserved. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--kubeconfig string
Path to a kubeconfig file, specifying how to connect to the API server. Providing --kubeconfig enables API server mode, omitting --kubeconfig enables standalone mode.
--kubelet-cgroups string
Optional absolute name of cgroups to create and run the Kubelet in. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--lock-file string
<Warning: Alpha feature> The path to file for kubelet to use as a lock file.
--log-backtrace-at <A string of format 'file:line'> Default: ":0"
When logging hits line :, emit a stack trace. (DEPRECATED: will be removed in a future release, see here.)
--log-dir string
If non-empty, write log files in this directory. (DEPRECATED: will be removed in a future release, see here.)
--log-file string
If non-empty, use this log file. (DEPRECATED: will be removed in a future release, see here.)
--log-file-max-size uint Default: 1800
Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited. (DEPRECATED: will be removed in a future release, see here.)
--log-flush-frequency duration Default: 5s
Maximum number of seconds between log flushes.
--log-json-info-buffer-size string Default: '0'
[Experimental] In JSON format with split output streams, the info messages can be buffered for a while to increase performance. The default value of zero bytes disables buffering. The size can be specified as number of bytes (512), multiples of 1000 (1K), multiples of 1024 (2Ki), or powers of those (3M, 4G, 5Mi, 6Gi). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--log-json-split-stream
[Experimental] In JSON format, write error messages to stderr and info messages to stdout. The default is to write a single stream to stdout. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--logging-format string Default: text
Sets the log format. Permitted formats: text, json. Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log-dir, --log-file, --log-file-max-size, --logtostderr, --skip_headers, --skip_log_headers, --stderrthreshold, --log-flush-frequency. Non-default choices are currently alpha and subject to change without warning. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--logtostderr Default: true
log to standard error instead of files. (DEPRECATED: will be removed in a future release, see here.)
--make-iptables-util-chains Default: true
If true, kubelet will ensure iptables utility rules are present on host. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--manifest-url string
URL for accessing additional Pod specifications to run (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--manifest-url-header string
Comma-separated list of HTTP headers to use when accessing the URL provided to --manifest-url. Multiple headers with the same name will be added in the same order provided. This flag can be repeatedly invoked. For example: --manifest-url-header 'a:hello,b:again,c:world' --manifest-url-header 'b:beautiful' (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
The namespace from which the kubernetes master services should be injected into pods. (DEPRECATED: This flag will be removed in a future version.)
--max-open-files int Default: 1000000
Number of files that can be opened by Kubelet process. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--max-pods int32 Default: 110
Number of Pods that can run on this Kubelet. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--maximum-dead-containers int32 Default: -1
Maximum number of old instances of containers to retain globally. Each container takes up some disk space. To disable, set to a negative number. (DEPRECATED: Use --eviction-hard or --eviction-soft instead. Will be removed in a future version.)
Maximum number of old instances to retain per container. Each container takes up some disk space. (DEPRECATED: Use --eviction-hard or --eviction-soft instead. Will be removed in a future version.)
--memory-manager-policy string Default: None
Memory Manager policy to use. Possible values: 'None', 'Static'. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--minimum-container-ttl-duration duration
Minimum age for a finished container before it is garbage collected. Examples: '300ms', '10s' or '2h45m' (DEPRECATED: Use --eviction-hard or --eviction-soft instead. Will be removed in a future version.)
Minimum age for an unused image before it is garbage collected. Examples: '300ms', '10s' or '2h45m'. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--node-ip string
IP address (or comma-separated dual-stack IP addresses) of the node. If unset, kubelet will use the node's default IPv4 address, if any, or its default IPv6 address if it has no IPv4 addresses. You can pass '::' to make it prefer the default IPv6 address rather than the default IPv4 address.
--node-labels mapStringString
<Warning: Alpha feature>Labels to add when registering the node in the cluster. Labels must be key=value pairs separated by ','. Labels in the 'kubernetes.io' namespace must begin with an allowed prefix ('kubelet.kubernetes.io', 'node.kubernetes.io') or be in the specifically allowed set ('beta.kubernetes.io/arch', 'beta.kubernetes.io/instance-type', 'beta.kubernetes.io/os', 'failure-domain.beta.kubernetes.io/region', 'failure-domain.beta.kubernetes.io/zone', 'kubernetes.io/arch', 'kubernetes.io/hostname', 'kubernetes.io/os', 'node.kubernetes.io/instance-type', 'topology.kubernetes.io/region', 'topology.kubernetes.io/zone')
--node-status-max-images int32 Default: 50
The maximum number of images to report in node.status.images. If -1 is specified, no cap will be applied. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
Specifies how often kubelet posts node status to master. Note: be cautious when changing the constant, it must work with nodeMonitorGracePeriod in Node controller. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--one-output
If true, only write logs to their native severity level (vs also writing to each lower severity level). (DEPRECATED: will be removed in a future release, see here.)
--oom-score-adj int32 Default: -999
The oom-score-adj value for kubelet process. Values must be within the range [-1000, 1000]. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--pod-cidr string
The CIDR to use for pod IP addresses, only used in standalone mode. In cluster mode, this is obtained from the master. For IPv6, the maximum number of IP's allocated is 65536 (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
Specified image will not be pruned by the image garbage collector. When container-runtime is set to docker, all containers in each pod will use the network/IPC namespaces from this image. Other CRI implementations have their own configuration to set this image.
--pod-manifest-path string
Path to the directory containing static pod files to run, or the path to a single static pod file. Files starting with dots will be ignored. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--pod-max-pids int Default: -1
Set the maximum number of processes per pod. If -1, the kubelet defaults to the node allocatable PID capacity. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--pods-per-core int32
Number of Pods per core that can run on this kubelet. The total number of pods on this kubelet cannot exceed --max-pods, so --max-pods will be used if this calculation results in a larger number of pods allowed on the kubelet. A value of 0 disables this limit. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--port int32 Default: 10250
The port for the kubelet to serve on. (DEPRECATED: This parameter should be set via the config file specified by the kubelet's --config flag. See kubelet-config-file for more information.)
--protect-kernel-defaults
Default kubelet behaviour for kernel tuning. If set, kubelet errors if any of kernel tunables is different than kubelet defaults. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--provider-id string
Unique identifier for identifying the node in a machine database, i.e cloud provider. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--qos-reserved mapStringString
<Warning: Alpha feature> A set of <resource name>=<percentage> (e.g. memory=50%) pairs that describe how pod resource requests are reserved at the QoS level. Currently only memory is supported. Requires the QOSReserved feature gate to be enabled. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--read-only-port int32 Default: 10255
The read-only port for the kubelet to serve on with no authentication/authorization (set to 0 to disable). (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--register-node Default: true
Register the node with the API server. If --kubeconfig is not provided, this flag is irrelevant, as the Kubelet won't have an API server to register with. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--register-schedulable Default: true
Register the node as schedulable. Won't have any effect if --register-node is false. (DEPRECATED: will be removed in a future version)
--register-with-taints mapStringString
Register the node with the given list of taints (comma separated <key>=<value>:<effect>). No-op if --register-node is false. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--registry-burst int32 Default: 10
Maximum size of a bursty pulls, temporarily allows pulls to burst to this number, while still not exceeding --registry-qps. Only used if --registry-qps is greater than 0. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--registry-qps int32 Default: 5
If > 0, limit registry pull QPS to this value. If 0, unlimited. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--reserved-cpus string
A comma-separated list of CPUs or CPU ranges that are reserved for system and kubernetes usage. This specific list will supersede cpu counts in --system-reserved and --kube-reserved. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--reserved-memory string
A comma-separated list of memory reservations for NUMA nodes. (e.g. --reserved-memory 0:memory=1Gi,hugepages-1M=2Gi --reserved-memory 1:memory=2Gi). The total sum for each memory type should be equal to the sum of --kube-reserved, --system-reserved and --eviction-threshold. See here for more details. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--resolv-conf string Default: /etc/resolv.conf
Resolver configuration file used as the basis for the container DNS resolution configuration. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--root-dir string Default: /var/lib/kubelet
Directory path for managing kubelet files (volume mounts, etc).
--rotate-certificates
<Warning: Beta feature> Auto rotate the kubelet client certificates by requesting new certificates from the kube-apiserver when the certificate expiration approaches. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--rotate-server-certificates
Auto-request and rotate the kubelet serving certificates by requesting new certificates from the kube-apiserver when the certificate expiration approaches. Requires the RotateKubeletServerCertificate feature gate to be enabled, and approval of the submitted CertificateSigningRequest objects. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--runonce
If true, exit after spawning pods from local manifests or remote urls. Exclusive with --enable-server (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--runtime-cgroups string
Optional absolute name of cgroups to create and run the runtime in.
--runtime-request-timeout duration Default: 2m0s
Timeout of all runtime requests except long running request - pull, logs, exec and attach. When timeout exceeded, kubelet will cancel the request, throw out an error and retry later. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--seccomp-default string
<Warning: Alpha feature> Enable the use of RuntimeDefault as the default seccomp profile for all workloads. The SeccompDefault feature gate must be enabled to allow this flag, which is disabled by default.
--serialize-image-pulls Default: true
Pull images one at a time. We recommend *not* changing the default value on nodes that run docker daemon with version < 1.9 or an aufs storage backend. Issue #10959 has more details. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--skip-headers
If true, avoid header prefixes in the log messages. (DEPRECATED: will be removed in a future release, see here.)
--skip-log-headers
If true, avoid headers when opening log files. (DEPRECATED: will be removed in a future release, see here.)
--stderrthreshold int Default: 2
logs at or above this threshold go to stderr. (DEPRECATED: will be removed in a future release, see here.)
Maximum time a streaming connection can be idle before the connection is automatically closed. 0 indicates no timeout. Example: 5m. Note: All connections to the kubelet server have a maximum duration of 4 hours. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--sync-frequency duration Default: 1m0s
Max period between synchronizing running containers and config. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--system-cgroups string
Optional absolute name of cgroups in which to place all non-kernel processes that are not already inside a cgroup under '/'. Empty for no container. Rolling back the flag requires a reboot. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--system-reserved mapStringString Default: <none>
A set of <resource name>=<resource quantity> (e.g. cpu=200m,memory=500Mi,ephemeral-storage=1Gi,pid='100') pairs that describe resources reserved for non-kubernetes components. Currently only cpu and memory are supported. See here for more detail. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--system-reserved-cgroup string Default: ''
Absolute name of the top level cgroup that is used to manage non-kubernetes components for which compute resources were reserved via --system-reserved flag. Ex. /system-reserved. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--tls-cert-file string
File containing x509 Certificate used for serving HTTPS (with intermediate certs, if any, concatenated after server cert). If --tls-cert-file and --tls-private-key-file are not provided, a self-signed certificate and key are generated for the public address and saved to the directory passed to --cert-dir. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--tls-cipher-suites strings
Comma-separated list of cipher suites for the server. If omitted, the default Go cipher suites will be used.
Preferred values:
`TLS_AES_128_GCM_SHA256`, `TLS_AES_256_GCM_SHA384`, `TLS_CHACHA20_POLY1305_SHA256`, `TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA`, `TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256`, `TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA`, `TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384`, `TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305`, `TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256`, `TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA`, `TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256`, `TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA`, `TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384`, `TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305`, `TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256`, `TLS_RSA_WITH_AES_128_CBC_SHA`, `TLS_RSA_WITH_AES_128_GCM_SHA256`, `TLS_RSA_WITH_AES_256_CBC_SHA`, `TLS_RSA_WITH_AES_256_GCM_SHA384`
Insecure values:
`TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256`, `TLS_ECDHE_ECDSA_WITH_RC4_128_SHA`, `TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA`, `TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256`, `TLS_ECDHE_RSA_WITH_RC4_128_SHA`, `TLS_RSA_WITH_3DES_EDE_CBC_SHA`, `TLS_RSA_WITH_AES_128_CBC_SHA256`, `TLS_RSA_WITH_RC4_128_SHA`.
(DEPRECATED: This parameter should be set via the config file specified by the Kubelet's `--config` flag. See kubelet-config-file for more information.)
--tls-min-version string
Minimum TLS version supported. Possible values: VersionTLS10, VersionTLS11, VersionTLS12, VersionTLS13. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--tls-private-key-file string
File containing x509 private key matching --tls-cert-file. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--topology-manager-policy string Default: 'none'
Topology Manager policy to use. Possible values: 'none', 'best-effort', 'restricted', 'single-numa-node'. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
Scope to which topology hints applied. Topology Manager collects hints from Hint Providers and applies them to defined scope to ensure the pod admission. Possible values: 'container', 'pod'. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
-v, --v Level
Number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule <A list of 'pattern=N' strings>
Comma-separated list of pattern=N settings for file-filtered logging
The full path of the directory in which to search for additional third party volume plugins. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
--volume-stats-agg-period duration Default: 1m0s
Specifies interval for kubelet to calculate and cache the volume disk usage for all pods and volumes. To disable volume calculations, set to 0. (DEPRECATED: This parameter should be set via the config file specified by the Kubelet's --config flag. See kubelet-config-file for more information.)
6.11.3 - kube-apiserver
Synopsis
The Kubernetes API server validates and configures data
for the api objects which include pods, services, replicationcontrollers, and
others. The API Server services REST operations and provides the frontend to the
cluster's shared state through which all other components interact.
kube-apiserver [flags]
Options
--admission-control-config-file string
File with admission control configuration.
--advertise-address string
The IP address on which to advertise the apiserver to members of the cluster. This address must be reachable by the rest of the cluster. If blank, the --bind-address will be used. If --bind-address is unspecified, the host's default interface will be used.
--allow-metric-labels stringToString Default: []
The map from metric-label to value allow-list of this label. The key's format is <MetricName>,<LabelName>. The value's format is <allowed_value>,<allowed_value>...e.g. metric1,label1='v1,v2,v3', metric1,label2='v1,v2,v3' metric2,label1='v1,v2,v3'.
--allow-privileged
If true, allow privileged containers. [default=false]
--anonymous-auth Default: true
Enables anonymous requests to the secure port of the API server. Requests that are not rejected by another authentication method are treated as anonymous requests. Anonymous requests have a username of system:anonymous, and a group name of system:unauthenticated.
--api-audiences strings
Identifiers of the API. The service account token authenticator will validate that tokens used against the API are bound to at least one of these audiences. If the --service-account-issuer flag is configured and this flag is not, this field defaults to a single element list containing the issuer URL.
--audit-log-batch-buffer-size int Default: 10000
The size of the buffer to store events before batching and writing. Only used in batch mode.
--audit-log-batch-max-size int Default: 1
The maximum size of a batch. Only used in batch mode.
--audit-log-batch-max-wait duration
The amount of time to wait before force writing the batch that hadn't reached the max size. Only used in batch mode.
--audit-log-batch-throttle-burst int
Maximum number of requests sent at the same moment if ThrottleQPS was not utilized before. Only used in batch mode.
--audit-log-batch-throttle-enable
Whether batching throttling is enabled. Only used in batch mode.
--audit-log-batch-throttle-qps float
Maximum average number of batches per second. Only used in batch mode.
--audit-log-compress
If set, the rotated log files will be compressed using gzip.
--audit-log-format string Default: "json"
Format of saved audits. "legacy" indicates 1-line text format for each event. "json" indicates structured json format. Known formats are legacy,json.
--audit-log-maxage int
The maximum number of days to retain old audit log files based on the timestamp encoded in their filename.
--audit-log-maxbackup int
The maximum number of old audit log files to retain. Setting a value of 0 will mean there's no restriction on the number of files.
--audit-log-maxsize int
The maximum size in megabytes of the audit log file before it gets rotated.
--audit-log-mode string Default: "blocking"
Strategy for sending audit events. Blocking indicates sending events should block server responses. Batch causes the backend to buffer and write events asynchronously. Known modes are batch,blocking,blocking-strict.
--audit-log-path string
If set, all requests coming to the apiserver will be logged to this file. '-' means standard out.
--audit-log-truncate-enabled
Whether event and batch truncating is enabled.
--audit-log-truncate-max-batch-size int Default: 10485760
Maximum size of the batch sent to the underlying backend. Actual serialized size can be several hundreds of bytes greater. If a batch exceeds this limit, it is split into several batches of smaller size.
--audit-log-truncate-max-event-size int Default: 102400
Maximum size of the audit event sent to the underlying backend. If the size of an event is greater than this number, first request and response are removed, and if this doesn't reduce the size enough, event is discarded.
The amount of time to wait before retrying the first failed request.
--audit-webhook-mode string Default: "batch"
Strategy for sending audit events. Blocking indicates sending events should block server responses. Batch causes the backend to buffer and write events asynchronously. Known modes are batch,blocking,blocking-strict.
--audit-webhook-truncate-enabled
Whether event and batch truncating is enabled.
--audit-webhook-truncate-max-batch-size int Default: 10485760
Maximum size of the batch sent to the underlying backend. Actual serialized size can be several hundreds of bytes greater. If a batch exceeds this limit, it is split into several batches of smaller size.
--audit-webhook-truncate-max-event-size int Default: 102400
Maximum size of the audit event sent to the underlying backend. If the size of an event is greater than this number, first request and response are removed, and if this doesn't reduce the size enough, event is discarded.
The duration to cache responses from the webhook token authenticator.
--authentication-token-webhook-config-file string
File with webhook configuration for token authentication in kubeconfig format. The API server will query the remote service to determine authentication for bearer tokens.
The duration to cache 'unauthorized' responses from the webhook authorizer.
--authorization-webhook-config-file string
File with webhook configuration in kubeconfig format, used with --authorization-mode=Webhook. The API server will query the remote service to determine access on the API server's secure port.
The API version of the authorization.k8s.io SubjectAccessReview to send to and expect from the webhook.
--azure-container-registry-config string
Path to the file containing Azure container registry configuration information.
--bind-address string Default: 0.0.0.0
The IP address on which to listen for the --secure-port port. The associated interface(s) must be reachable by the rest of the cluster, and by CLI/web clients. If blank or an unspecified address (0.0.0.0 or ::), all interfaces will be used.
--cert-dir string Default: "/var/run/kubernetes"
The directory where the TLS certs are located. If --tls-cert-file and --tls-private-key-file are provided, this flag will be ignored.
--client-ca-file string
If set, any request presenting a client certificate signed by one of the authorities in the client-ca-file is authenticated with an identity corresponding to the CommonName of the client certificate.
--cloud-config string
The path to the cloud provider configuration file. Empty string for no configuration file.
--cloud-provider string
The provider for cloud services. Empty string for no provider.
CIDRs opened in GCE firewall for L7 LB traffic proxy & health checks
--contention-profiling
Enable lock contention profiling, if profiling is enabled
--cors-allowed-origins strings
List of allowed origins for CORS, comma separated. An allowed origin can be a regular expression to support subdomain matching. If this list is empty CORS will not be enabled.
--default-not-ready-toleration-seconds int Default: 300
Indicates the tolerationSeconds of the toleration for notReady:NoExecute that is added by default to every pod that does not already have such a toleration.
--default-unreachable-toleration-seconds int Default: 300
Indicates the tolerationSeconds of the toleration for unreachable:NoExecute that is added by default to every pod that does not already have such a toleration.
--default-watch-cache-size int Default: 100
Default watch cache size. If zero, watch cache will be disabled for resources that do not have a default watch size set.
--delete-collection-workers int Default: 1
Number of workers spawned for DeleteCollection call. These are used to speed up namespace cleanup.
--disable-admission-plugins strings
admission plugins that should be disabled although they are in the default enabled plugins list (NamespaceLifecycle, LimitRanger, ServiceAccount, TaintNodesByCondition, PodSecurity, Priority, DefaultTolerationSeconds, DefaultStorageClass, StorageObjectInUseProtection, PersistentVolumeClaimResize, RuntimeClass, CertificateApproval, CertificateSigning, CertificateSubjectRestriction, DefaultIngressClass, MutatingAdmissionWebhook, ValidatingAdmissionWebhook, ResourceQuota). Comma-delimited list of admission plugins: AlwaysAdmit, AlwaysDeny, AlwaysPullImages, CertificateApproval, CertificateSigning, CertificateSubjectRestriction, DefaultIngressClass, DefaultStorageClass, DefaultTolerationSeconds, DenyServiceExternalIPs, EventRateLimit, ExtendedResourceToleration, ImagePolicyWebhook, LimitPodHardAntiAffinityTopology, LimitRanger, MutatingAdmissionWebhook, NamespaceAutoProvision, NamespaceExists, NamespaceLifecycle, NodeRestriction, OwnerReferencesPermissionEnforcement, PersistentVolumeClaimResize, PersistentVolumeLabel, PodNodeSelector, PodSecurity, PodSecurityPolicy, PodTolerationRestriction, Priority, ResourceQuota, RuntimeClass, SecurityContextDeny, ServiceAccount, StorageObjectInUseProtection, TaintNodesByCondition, ValidatingAdmissionWebhook. The order of plugins in this flag does not matter.
--disabled-metrics strings
This flag provides an escape hatch for misbehaving metrics. You must provide the fully qualified metric name in order to disable it. Disclaimer: disabling metrics is higher in precedence than showing hidden metrics.
--egress-selector-config-file string
File with apiserver egress selector configuration.
--enable-admission-plugins strings
admission plugins that should be enabled in addition to default enabled ones (NamespaceLifecycle, LimitRanger, ServiceAccount, TaintNodesByCondition, PodSecurity, Priority, DefaultTolerationSeconds, DefaultStorageClass, StorageObjectInUseProtection, PersistentVolumeClaimResize, RuntimeClass, CertificateApproval, CertificateSigning, CertificateSubjectRestriction, DefaultIngressClass, MutatingAdmissionWebhook, ValidatingAdmissionWebhook, ResourceQuota). Comma-delimited list of admission plugins: AlwaysAdmit, AlwaysDeny, AlwaysPullImages, CertificateApproval, CertificateSigning, CertificateSubjectRestriction, DefaultIngressClass, DefaultStorageClass, DefaultTolerationSeconds, DenyServiceExternalIPs, EventRateLimit, ExtendedResourceToleration, ImagePolicyWebhook, LimitPodHardAntiAffinityTopology, LimitRanger, MutatingAdmissionWebhook, NamespaceAutoProvision, NamespaceExists, NamespaceLifecycle, NodeRestriction, OwnerReferencesPermissionEnforcement, PersistentVolumeClaimResize, PersistentVolumeLabel, PodNodeSelector, PodSecurity, PodSecurityPolicy, PodTolerationRestriction, Priority, ResourceQuota, RuntimeClass, SecurityContextDeny, ServiceAccount, StorageObjectInUseProtection, TaintNodesByCondition, ValidatingAdmissionWebhook. The order of plugins in this flag does not matter.
--enable-aggregator-routing
Turns on aggregator routing requests to endpoints IP rather than cluster IP.
--enable-bootstrap-token-auth
Enable to allow secrets of type 'bootstrap.kubernetes.io/token' in the 'kube-system' namespace to be used for TLS bootstrapping authentication.
--enable-garbage-collector Default: true
Enables the generic garbage collector. MUST be synced with the corresponding flag of the kube-controller-manager.
--enable-priority-and-fairness Default: true
If true and the APIPriorityAndFairness feature gate is enabled, replace the max-in-flight handler with an enhanced one that queues and dispatches with priority and fairness
--encryption-provider-config string
The file containing configuration for encryption providers to be used for storing secrets in etcd
The interval of requests to poll etcd and update metric. 0 disables the metric collection
--etcd-healthcheck-timeout duration Default: 2s
The timeout to use when checking etcd health.
--etcd-keyfile string
SSL key file used to secure etcd communication.
--etcd-prefix string Default: "/registry"
The prefix to prepend to all resource paths in etcd.
--etcd-servers strings
List of etcd servers to connect with (scheme://ip:port), comma separated.
--etcd-servers-overrides strings
Per-resource etcd servers overrides, comma separated. The individual override format: group/resource#servers, where servers are URLs, semicolon separated. Note that this applies only to resources compiled into this server binary.
--event-ttl duration Default: 1h0m0s
Amount of time to retain events.
--external-hostname string
The hostname to use when generating externalized URLs for this master (e.g. Swagger API Docs or OpenID Discovery).
To prevent HTTP/2 clients from getting stuck on a single apiserver, randomly close a connection (GOAWAY). The client's other in-flight requests won't be affected, and the client will reconnect, likely landing on a different apiserver after going through the load balancer again. This argument sets the fraction of requests that will be sent a GOAWAY. Clusters with single apiservers, or which don't use a load balancer, should NOT enable this. Min is 0 (off), Max is .02 (1/50 requests); .001 (1/1000) is a recommended starting point.
-h, --help
help for kube-apiserver
--http2-max-streams-per-connection int
The limit that the server gives to clients for the maximum number of streams in an HTTP/2 connection. Zero means to use golang's default.
--identity-lease-duration-seconds int Default: 3600
The duration of kube-apiserver lease in seconds, must be a positive number. (In use when the APIServerIdentity feature gate is enabled.)
--identity-lease-renew-interval-seconds int Default: 10
The interval of kube-apiserver renewing its lease in seconds, must be a positive number. (In use when the APIServerIdentity feature gate is enabled.)
--kubelet-certificate-authority string
Path to a cert file for the certificate authority.
List of the preferred NodeAddressTypes to use for kubelet connections.
--kubelet-timeout duration Default: 5s
Timeout for kubelet operations.
--kubernetes-service-node-port int
If non-zero, the Kubernetes master service (which apiserver creates/maintains) will be of type NodePort, using this as the value of the port. If zero, the Kubernetes master service will be of type ClusterIP.
--lease-reuse-duration-seconds int Default: 60
The time in seconds that each lease is reused. A lower value could avoid large number of objects reusing the same lease. Notice that a too small value may cause performance problems at storage layer.
--livez-grace-period duration
This option represents the maximum amount of time it should take for apiserver to complete its startup sequence and become live. From apiserver's start time to when this amount of time has elapsed, /livez will assume that unfinished post-start hooks will complete successfully and therefore return true.
--log-flush-frequency duration Default: 5s
Maximum number of seconds between log flushes
--logging-format string Default: "text"
Sets the log format. Permitted formats: "text". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log-dir, --log-file, --log-file-max-size, --logtostderr, --one-output, --skip-headers, --skip-log-headers, --stderrthreshold, --vmodule. Non-default choices are currently alpha and subject to change without warning.
DEPRECATED: the namespace from which the Kubernetes master services should be injected into pods.
--max-connection-bytes-per-sec int
If non-zero, throttle each user connection to this number of bytes/sec. Currently only applies to long-running requests.
--max-mutating-requests-inflight int Default: 200
This and --max-requests-inflight are summed to determine the server's total concurrency limit (which must be positive) if --enable-priority-and-fairness is true. Otherwise, this flag limits the maximum number of mutating requests in flight, or a zero value disables the limit completely.
--max-requests-inflight int Default: 400
This and --max-mutating-requests-inflight are summed to determine the server's total concurrency limit (which must be positive) if --enable-priority-and-fairness is true. Otherwise, this flag limits the maximum number of non-mutating requests in flight, or a zero value disables the limit completely.
--min-request-timeout int Default: 1800
An optional field indicating the minimum number of seconds a handler must keep a request open before timing it out. Currently only honored by the watch request handler, which picks a randomized value above this number as the connection timeout, to spread out load.
--oidc-ca-file string
If set, the OpenID server's certificate will be verified by one of the authorities in the oidc-ca-file, otherwise the host's root CA set will be used.
--oidc-client-id string
The client ID for the OpenID Connect client, must be set if oidc-issuer-url is set.
--oidc-groups-claim string
If provided, the name of a custom OpenID Connect claim for specifying user groups. The claim value is expected to be a string or array of strings. This flag is experimental, please see the authentication documentation for further details.
--oidc-groups-prefix string
If provided, all groups will be prefixed with this value to prevent conflicts with other authentication strategies.
--oidc-issuer-url string
The URL of the OpenID issuer, only HTTPS scheme will be accepted. If set, it will be used to verify the OIDC JSON Web Token (JWT).
A key=value pair that describes a required claim in the ID Token. If set, the claim is verified to be present in the ID Token with a matching value. Repeat this flag to specify multiple claims.
--oidc-signing-algs strings Default: "RS256"
Comma-separated list of allowed JOSE asymmetric signing algorithms. JWTs with a supported 'alg' header values are: RS256, RS384, RS512, ES256, ES384, ES512, PS256, PS384, PS512. Values are defined by RFC 7518 https://tools.ietf.org/html/rfc7518#section-3.1.
--oidc-username-claim string Default: "sub"
The OpenID claim to use as the user name. Note that claims other than the default ('sub') is not guaranteed to be unique and immutable. This flag is experimental, please see the authentication documentation for further details.
--oidc-username-prefix string
If provided, all usernames will be prefixed with this value. If not provided, username claims other than 'email' are prefixed by the issuer URL to avoid clashes. To skip any prefixing, provide the value '-'.
--permit-address-sharing
If true, SO_REUSEADDR will be used when binding the port. This allows binding to wildcard IPs like 0.0.0.0 and specific IPs in parallel, and it avoids waiting for the kernel to release sockets in TIME_WAIT state. [default=false]
--permit-port-sharing
If true, SO_REUSEPORT will be used when binding the port, which allows more than one instance to bind on the same address and port. [default=false]
--profiling Default: true
Enable profiling via web interface host:port/debug/pprof/
--proxy-client-cert-file string
Client certificate used to prove the identity of the aggregator or kube-apiserver when it must call out during a request. This includes proxying requests to a user api-server and calling out to webhook admission plugins. It is expected that this cert includes a signature from the CA in the --requestheader-client-ca-file flag. That CA is published in the 'extension-apiserver-authentication' configmap in the kube-system namespace. Components receiving calls from kube-aggregator should use that CA to perform their half of the mutual TLS verification.
--proxy-client-key-file string
Private key for the client certificate used to prove the identity of the aggregator or kube-apiserver when it must call out during a request. This includes proxying requests to a user api-server and calling out to webhook admission plugins.
--request-timeout duration Default: 1m0s
An optional field indicating the duration a handler must keep a request open before timing it out. This is the default request timeout for requests but may be overridden by flags such as --min-request-timeout for specific types of requests.
--requestheader-allowed-names strings
List of client certificate common names to allow to provide usernames in headers specified by --requestheader-username-headers. If empty, any client certificate validated by the authorities in --requestheader-client-ca-file is allowed.
--requestheader-client-ca-file string
Root certificate bundle to use to verify client certificates on incoming requests before trusting usernames in headers specified by --requestheader-username-headers. WARNING: generally do not depend on authorization being already done for incoming requests.
--requestheader-extra-headers-prefix strings
List of request header prefixes to inspect. X-Remote-Extra- is suggested.
--requestheader-group-headers strings
List of request headers to inspect for groups. X-Remote-Group is suggested.
--requestheader-username-headers strings
List of request headers to inspect for usernames. X-Remote-User is common.
A set of key=value pairs that enable or disable built-in APIs. Supported options are: v1=true|false for the core API group <group>/<version>=true|false for a specific API group and version (e.g. apps/v1=true) api/all=true|false controls all API versions api/ga=true|false controls all API versions of the form v[0-9]+ api/beta=true|false controls all API versions of the form v[0-9]+beta[0-9]+ api/alpha=true|false controls all API versions of the form v[0-9]+alpha[0-9]+ api/legacy is deprecated, and will be removed in a future version
--secure-port int Default: 6443
The port on which to serve HTTPS with authentication and authorization. It cannot be switched off with 0.
Turns on projected service account expiration extension during token generation, which helps safe transition from legacy token to bound service account token feature. If this flag is enabled, admission injected tokens would be extended up to 1 year to prevent unexpected failure during transition, ignoring value of service-account-max-token-expiration.
--service-account-issuer strings
Identifier of the service account token issuer. The issuer will assert this identifier in "iss" claim of issued tokens. This value is a string or URI. If this option is not a valid URI per the OpenID Discovery 1.0 spec, the ServiceAccountIssuerDiscovery feature will remain disabled, even if the feature gate is set to true. It is highly recommended that this value comply with the OpenID spec: https://openid.net/specs/openid-connect-discovery-1_0.html. In practice, this means that service-account-issuer must be an https URL. It is also highly recommended that this URL be capable of serving OpenID discovery documents at {service-account-issuer}/.well-known/openid-configuration. When this flag is specified multiple times, the first is used to generate tokens and all are used to determine which issuers are accepted.
--service-account-jwks-uri string
Overrides the URI for the JSON Web Key Set in the discovery doc served at /.well-known/openid-configuration. This flag is useful if the discovery docand key set are served to relying parties from a URL other than the API server's external (as auto-detected or overridden with external-hostname).
--service-account-key-file strings
File containing PEM-encoded x509 RSA or ECDSA private or public keys, used to verify ServiceAccount tokens. The specified file can contain multiple keys, and the flag can be specified multiple times with different files. If unspecified, --tls-private-key-file is used. Must be specified when --service-account-signing-key-file is provided
--service-account-lookup Default: true
If true, validate ServiceAccount tokens exist in etcd as part of authentication.
--service-account-max-token-expiration duration
The maximum validity duration of a token created by the service account token issuer. If an otherwise valid TokenRequest with a validity duration larger than this value is requested, a token will be issued with a validity duration of this value.
--service-account-signing-key-file string
Path to the file that contains the current private key of the service account token issuer. The issuer will sign issued ID tokens with this private key.
--service-cluster-ip-range string
A CIDR notation IP range from which to assign service cluster IPs. This must not overlap with any IP ranges assigned to nodes or pods. Max of two dual-stack CIDRs is allowed.
--service-node-port-range <a string in the form 'N1-N2'> Default: 30000-32767
A port range to reserve for services with NodePort visibility. Example: '30000-32767'. Inclusive at both ends of the range.
--show-hidden-metrics-for-version string
The previous version for which you want to show hidden metrics. Only the previous minor version is meaningful, other values will not be allowed. The format is <major>.<minor>, e.g.: '1.16'. The purpose of this format is make sure you have the opportunity to notice if the next release hides additional metrics, rather than being surprised when they are permanently removed in the release after that.
--shutdown-delay-duration duration
Time to delay the termination. During that time the server keeps serving requests normally. The endpoints /healthz and /livez will return success, but /readyz immediately returns failure. Graceful termination starts after this delay has elapsed. This can be used to allow load balancer to stop sending traffic to this server.
--shutdown-send-retry-after
If true the HTTP Server will continue listening until all non long running request(s) in flight have been drained, during this window all incoming requests will be rejected with a status code 429 and a 'Retry-After' response header, in addition 'Connection: close' response header is set in order to tear down the TCP connection when idle.
--storage-backend string
The storage backend for persistence. Options: 'etcd3' (default).
The media type to use to store objects in storage. Some resources or storage backends may only support a specific media type and will ignore this setting.
--strict-transport-security-directives strings
List of directives for HSTS, comma separated. If this list is empty, then HSTS directives will not be added. Example: 'max-age=31536000,includeSubDomains,preload'
--tls-cert-file string
File containing the default x509 Certificate for HTTPS. (CA cert, if any, concatenated after server cert). If HTTPS serving is enabled, and --tls-cert-file and --tls-private-key-file are not provided, a self-signed certificate and key are generated for the public address and saved to the directory specified by --cert-dir.
--tls-cipher-suites strings
Comma-separated list of cipher suites for the server. If omitted, the default Go cipher suites will be used. Preferred values: TLS_AES_128_GCM_SHA256, TLS_AES_256_GCM_SHA384, TLS_CHACHA20_POLY1305_SHA256, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_256_GCM_SHA384. Insecure values: TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_ECDSA_WITH_RC4_128_SHA, TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_RSA_WITH_RC4_128_SHA, TLS_RSA_WITH_3DES_EDE_CBC_SHA, TLS_RSA_WITH_AES_128_CBC_SHA256, TLS_RSA_WITH_RC4_128_SHA.
--tls-min-version string
Minimum TLS version supported. Possible values: VersionTLS10, VersionTLS11, VersionTLS12, VersionTLS13
--tls-private-key-file string
File containing the default x509 private key matching --tls-cert-file.
--tls-sni-cert-key string
A pair of x509 certificate and private key file paths, optionally suffixed with a list of domain patterns which are fully qualified domain names, possibly with prefixed wildcard segments. The domain patterns also allow IP addresses, but IPs should only be used if the apiserver has visibility to the IP address requested by a client. If no domain patterns are provided, the names of the certificate are extracted. Non-wildcard matches trump over wildcard matches, explicit domain patterns trump over extracted names. For multiple key/certificate pairs, use the --tls-sni-cert-key multiple times. Examples: "example.crt,example.key" or "foo.crt,foo.key:*.foo.com,foo.com".
--token-auth-file string
If set, the file that will be used to secure the secure port of the API server via token authentication.
--tracing-config-file string
File with apiserver tracing configuration.
-v, --v int
number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule pattern=N,...
comma-separated list of pattern=N settings for file-filtered logging (only works for text log format)
--watch-cache Default: true
Enable watch caching in the apiserver
--watch-cache-sizes strings
Watch cache size settings for some resources (pods, nodes, etc.), comma separated. The individual setting format: resource[.group]#size, where resource is lowercase plural (no version), group is omitted for resources of apiVersion v1 (the legacy core API) and included for others, and size is a number. It takes effect when watch-cache is enabled. Some resources (replicationcontrollers, endpoints, nodes, pods, services, apiservices.apiregistration.k8s.io) have system defaults set by heuristics, others default to default-watch-cache-size
6.11.4 - kube-controller-manager
Synopsis
The Kubernetes controller manager is a daemon that embeds
the core control loops shipped with Kubernetes. In applications of robotics and
automation, a control loop is a non-terminating loop that regulates the state of
the system. In Kubernetes, a controller is a control loop that watches the shared
state of the cluster through the apiserver and makes changes attempting to move the
current state towards the desired state. Examples of controllers that ship with
Kubernetes today are the replication controller, endpoints controller, namespace
controller, and serviceaccounts controller.
kube-controller-manager [flags]
Options
--allocate-node-cidrs
Should CIDRs for Pods be allocated and set on the cloud provider.
--allow-metric-labels stringToString Default: []
The map from metric-label to value allow-list of this label. The key's format is <MetricName>,<LabelName>. The value's format is <allowed_value>,<allowed_value>...e.g. metric1,label1='v1,v2,v3', metric1,label2='v1,v2,v3' metric2,label1='v1,v2,v3'.
The reconciler sync wait time between volume attach detach. This duration must be larger than one second, and increasing this value from the default may allow for volumes to be mismatched with pods.
--authentication-kubeconfig string
kubeconfig file pointing at the 'core' kubernetes server with enough rights to create tokenreviews.authentication.k8s.io. This is optional. If empty, all token requests are considered to be anonymous and no client CA is looked up in the cluster.
--authentication-skip-lookup
If false, the authentication-kubeconfig will be used to lookup missing authentication configuration from the cluster.
The duration to cache responses from the webhook token authenticator.
--authentication-tolerate-lookup-failure
If true, failures to look up missing authentication configuration from the cluster are not considered fatal. Note that this can result in authentication that treats all requests as anonymous.
A list of HTTP paths to skip during authorization, i.e. these are authorized without contacting the 'core' kubernetes server.
--authorization-kubeconfig string
kubeconfig file pointing at the 'core' kubernetes server with enough rights to create subjectaccessreviews.authorization.k8s.io. This is optional. If empty, all requests not skipped by authorization are forbidden.
The duration to cache 'unauthorized' responses from the webhook authorizer.
--azure-container-registry-config string
Path to the file containing Azure container registry configuration information.
--bind-address string Default: 0.0.0.0
The IP address on which to listen for the --secure-port port. The associated interface(s) must be reachable by the rest of the cluster, and by CLI/web clients. If blank or an unspecified address (0.0.0.0 or ::), all interfaces will be used.
--cert-dir string
The directory where the TLS certs are located. If --tls-cert-file and --tls-private-key-file are provided, this flag will be ignored.
If set, any request presenting a client certificate signed by one of the authorities in the client-ca-file is authenticated with an identity corresponding to the CommonName of the client certificate.
--cloud-config string
The path to the cloud provider configuration file. Empty string for no configuration file.
--cloud-provider string
The provider for cloud services. Empty string for no provider.
--cluster-cidr string
CIDR Range for Pods in cluster. Requires --allocate-node-cidrs to be true
--cluster-name string Default: "kubernetes"
The instance prefix for the cluster.
--cluster-signing-cert-file string
Filename containing a PEM-encoded X509 CA certificate used to issue cluster-scoped certificates. If specified, no more specific --cluster-signing-* flag may be specified.
The max length of duration signed certificates will be given. Individual CSRs may request shorter certs by setting spec.expirationSeconds.
--cluster-signing-key-file string
Filename containing a PEM-encoded RSA or ECDSA private key used to sign cluster-scoped certificates. If specified, no more specific --cluster-signing-* flag may be specified.
Filename containing a PEM-encoded X509 CA certificate used to issue certificates for the kubernetes.io/kube-apiserver-client signer. If specified, --cluster-signing-{cert,key}-file must not be set.
Filename containing a PEM-encoded RSA or ECDSA private key used to sign certificates for the kubernetes.io/kube-apiserver-client signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--cluster-signing-kubelet-client-cert-file string
Filename containing a PEM-encoded X509 CA certificate used to issue certificates for the kubernetes.io/kube-apiserver-client-kubelet signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--cluster-signing-kubelet-client-key-file string
Filename containing a PEM-encoded RSA or ECDSA private key used to sign certificates for the kubernetes.io/kube-apiserver-client-kubelet signer. If specified, --cluster-signing-{cert,key}-file must not be set.
Filename containing a PEM-encoded X509 CA certificate used to issue certificates for the kubernetes.io/kubelet-serving signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--cluster-signing-kubelet-serving-key-file string
Filename containing a PEM-encoded RSA or ECDSA private key used to sign certificates for the kubernetes.io/kubelet-serving signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--cluster-signing-legacy-unknown-cert-file string
Filename containing a PEM-encoded X509 CA certificate used to issue certificates for the kubernetes.io/legacy-unknown signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--cluster-signing-legacy-unknown-key-file string
Filename containing a PEM-encoded RSA or ECDSA private key used to sign certificates for the kubernetes.io/legacy-unknown signer. If specified, --cluster-signing-{cert,key}-file must not be set.
--concurrent-deployment-syncs int32 Default: 5
The number of deployment objects that are allowed to sync concurrently. Larger number = more responsive deployments, but more CPU (and network) load
--concurrent-endpoint-syncs int32 Default: 5
The number of endpoint syncing operations that will be done concurrently. Larger number = faster endpoint updating, but more CPU (and network) load
The number of ephemeral volume syncing operations that will be done concurrently. Larger number = faster ephemeral volume updating, but more CPU (and network) load
--concurrent-gc-syncs int32 Default: 20
The number of garbage collector workers that are allowed to sync concurrently.
--concurrent-namespace-syncs int32 Default: 10
The number of namespace objects that are allowed to sync concurrently. Larger number = more responsive namespace termination, but more CPU (and network) load
--concurrent-rc-syncs int32 Default: 5
The number of replication controllers that are allowed to sync concurrently. Larger number = more responsive replica management, but more CPU (and network) load
--concurrent-replicaset-syncs int32 Default: 5
The number of replica sets that are allowed to sync concurrently. Larger number = more responsive replica management, but more CPU (and network) load
The number of service endpoint syncing operations that will be done concurrently. Larger number = faster endpoint slice updating, but more CPU (and network) load. Defaults to 5.
--concurrent-service-syncs int32 Default: 1
The number of services that are allowed to sync concurrently. Larger number = more responsive service management, but more CPU (and network) load
The number of service account token objects that are allowed to sync concurrently. Larger number = more responsive token generation, but more CPU (and network) load
--concurrent-statefulset-syncs int32 Default: 5
The number of statefulset objects that are allowed to sync concurrently. Larger number = more responsive statefulsets, but more CPU (and network) load
The number of TTL-after-finished controller workers that are allowed to sync concurrently.
--configure-cloud-routes Default: true
Should CIDRs allocated by allocate-node-cidrs be configured on the cloud provider.
--contention-profiling
Enable lock contention profiling, if profiling is enabled
--controller-start-interval duration
Interval between starting controller managers.
--controllers strings Default: "*"
A list of controllers to enable. '*' enables all on-by-default controllers, 'foo' enables the controller named 'foo', '-foo' disables the controller named 'foo'. All controllers: attachdetach, bootstrapsigner, cloud-node-lifecycle, clusterrole-aggregation, cronjob, csrapproving, csrcleaner, csrsigning, daemonset, deployment, disruption, endpoint, endpointslice, endpointslicemirroring, ephemeral-volume, garbagecollector, horizontalpodautoscaling, job, namespace, nodeipam, nodelifecycle, persistentvolume-binder, persistentvolume-expander, podgc, pv-protection, pvc-protection, replicaset, replicationcontroller, resourcequota, root-ca-cert-publisher, route, service, serviceaccount, serviceaccount-token, statefulset, tokencleaner, ttl, ttl-after-finished Disabled-by-default controllers: bootstrapsigner, tokencleaner
--disable-attach-detach-reconcile-sync
Disable volume attach detach reconciler sync. Disabling this may cause volumes to be mismatched with pods. Use wisely.
--disabled-metrics strings
This flag provides an escape hatch for misbehaving metrics. You must provide the fully qualified metric name in order to disable it. Disclaimer: disabling metrics is higher in precedence than showing hidden metrics.
--enable-dynamic-provisioning Default: true
Enable dynamic provisioning for environments that support it.
--enable-garbage-collector Default: true
Enables the generic garbage collector. MUST be synced with the corresponding flag of the kube-apiserver.
--enable-hostpath-provisioner
Enable HostPath PV provisioning when running without a cloud provider. This allows testing and development of provisioning features. HostPath provisioning is not supported in any way, won't work in a multi-node cluster, and should not be used for anything other than testing or development.
--enable-leader-migration
Whether to enable controller leader migration.
--enable-taint-manager Default: true
WARNING: Beta feature. If set to true enables NoExecute Taints and will evict all not-tolerating Pod running on Nodes tainted with this kind of Taints.
--endpoint-updates-batch-period duration
The length of endpoint updates batching period. Processing of pod changes will be delayed by this duration to join them with potential upcoming updates and reduce the overall number of endpoints updates. Larger number = higher endpoint programming latency, but lower number of endpoints revision generated
--endpointslice-updates-batch-period duration
The length of endpoint slice updates batching period. Processing of pod changes will be delayed by this duration to join them with potential upcoming updates and reduce the overall number of endpoints updates. Larger number = higher endpoint programming latency, but lower number of endpoints revision generated
--external-cloud-volume-plugin string
The plugin to use when cloud provider is set to external. Can be empty, should only be set when cloud-provider is external. Currently used to allow node and volume controllers to work for in tree cloud providers.
QPS to use while talking with kubernetes apiserver.
--kubeconfig string
Path to kubeconfig file with authorization and master location information.
--large-cluster-size-threshold int32 Default: 50
Number of nodes from which NodeController treats the cluster as large for the eviction logic purposes. --secondary-node-eviction-rate is implicitly overridden to 0 for clusters this size or smaller.
--leader-elect Default: true
Start a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.
The duration that non-leader candidates will wait after observing a leadership renewal until attempting to acquire leadership of a led but unrenewed leader slot. This is effectively the maximum duration that a leader can be stopped before it is replaced by another candidate. This is only applicable if leader election is enabled.
The interval between attempts by the acting master to renew a leadership slot before it stops leading. This must be less than or equal to the lease duration. This is only applicable if leader election is enabled.
The namespace of resource object that is used for locking during leader election.
--leader-elect-retry-period duration Default: 2s
The duration the clients should wait between attempting acquisition and renewal of a leadership. This is only applicable if leader election is enabled.
--leader-migration-config string
Path to the config file for controller leader migration, or empty to use the value that reflects default configuration of the controller manager. The config file should be of type LeaderMigrationConfiguration, group controllermanager.config.k8s.io, version v1alpha1.
--log-flush-frequency duration Default: 5s
Maximum number of seconds between log flushes
--logging-format string Default: "text"
Sets the log format. Permitted formats: "text". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log-dir, --log-file, --log-file-max-size, --logtostderr, --one-output, --skip-headers, --skip-log-headers, --stderrthreshold, --vmodule. Non-default choices are currently alpha and subject to change without warning.
--master string
The address of the Kubernetes API server (overrides any value in kubeconfig).
--max-endpoints-per-slice int32 Default: 100
The maximum number of endpoints that will be added to an EndpointSlice. More endpoints per slice will result in less endpoint slices, but larger resources. Defaults to 100.
--min-resync-period duration Default: 12h0m0s
The resync period in reflectors will be random between MinResyncPeriod and 2*MinResyncPeriod.
The number of service endpoint syncing operations that will be done concurrently by the EndpointSliceMirroring controller. Larger number = faster endpoint slice updating, but more CPU (and network) load. Defaults to 5.
The length of EndpointSlice updates batching period for EndpointSliceMirroring controller. Processing of EndpointSlice changes will be delayed by this duration to join them with potential upcoming updates and reduce the overall number of EndpointSlice updates. Larger number = higher endpoint programming latency, but lower number of endpoints revision generated
The maximum number of endpoints that will be added to an EndpointSlice by the EndpointSliceMirroring controller. More endpoints per slice will result in less endpoint slices, but larger resources. Defaults to 100.
--namespace-sync-period duration Default: 5m0s
The period for syncing namespace life-cycle updates
--node-cidr-mask-size int32
Mask size for node cidr in cluster. Default is 24 for IPv4 and 64 for IPv6.
--node-cidr-mask-size-ipv4 int32
Mask size for IPv4 node cidr in dual-stack cluster. Default is 24.
--node-cidr-mask-size-ipv6 int32
Mask size for IPv6 node cidr in dual-stack cluster. Default is 64.
--node-eviction-rate float Default: 0.1
Number of nodes per second on which pods are deleted in case of node failure when a zone is healthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters.
--node-monitor-grace-period duration Default: 40s
Amount of time which we allow running Node to be unresponsive before marking it unhealthy. Must be N times more than kubelet's nodeStatusUpdateFrequency, where N means number of retries allowed for kubelet to post node status.
--node-monitor-period duration Default: 5s
The period for syncing NodeStatus in NodeController.
Amount of time which we allow starting Node to be unresponsive before marking it unhealthy.
--permit-address-sharing
If true, SO_REUSEADDR will be used when binding the port. This allows binding to wildcard IPs like 0.0.0.0 and specific IPs in parallel, and it avoids waiting for the kernel to release sockets in TIME_WAIT state. [default=false]
--permit-port-sharing
If true, SO_REUSEPORT will be used when binding the port, which allows more than one instance to bind on the same address and port. [default=false]
--pod-eviction-timeout duration Default: 5m0s
The grace period for deleting pods on failed nodes.
--profiling Default: true
Enable profiling via web interface host:port/debug/pprof/
The minimum ActiveDeadlineSeconds to use for a HostPath Recycler pod. This is for development and testing only and will not work in a multi-node cluster.
The file path to a pod definition used as a template for HostPath persistent volume recycling. This is for development and testing only and will not work in a multi-node cluster.
--pv-recycler-pod-template-filepath-nfs string
The file path to a pod definition used as a template for NFS persistent volume recycling
the increment of time added per Gi to ActiveDeadlineSeconds for a HostPath scrubber pod. This is for development and testing only and will not work in a multi-node cluster.
--pvclaimbinder-sync-period duration Default: 15s
The period for syncing persistent volumes and persistent volume claims
--requestheader-allowed-names strings
List of client certificate common names to allow to provide usernames in headers specified by --requestheader-username-headers. If empty, any client certificate validated by the authorities in --requestheader-client-ca-file is allowed.
--requestheader-client-ca-file string
Root certificate bundle to use to verify client certificates on incoming requests before trusting usernames in headers specified by --requestheader-username-headers. WARNING: generally do not depend on authorization being already done for incoming requests.
Number of nodes per second on which pods are deleted in case of node failure when a zone is unhealthy (see --unhealthy-zone-threshold for definition of healthy/unhealthy). Zone refers to entire cluster in non-multizone clusters. This value is implicitly overridden to 0 if the cluster size is smaller than --large-cluster-size-threshold.
--secure-port int Default: 10257
The port on which to serve HTTPS with authentication and authorization. If 0, don't serve HTTPS at all.
--service-account-private-key-file string
Filename containing a PEM-encoded private RSA or ECDSA key used to sign service account tokens.
--service-cluster-ip-range string
CIDR Range for Services in cluster. Requires --allocate-node-cidrs to be true
--show-hidden-metrics-for-version string
The previous version for which you want to show hidden metrics. Only the previous minor version is meaningful, other values will not be allowed. The format is <major>.<minor>, e.g.: '1.16'. The purpose of this format is make sure you have the opportunity to notice if the next release hides additional metrics, rather than being surprised when they are permanently removed in the release after that.
Number of terminated pods that can exist before the terminated pod garbage collector starts deleting terminated pods. If <= 0, the terminated pod garbage collector is disabled.
--tls-cert-file string
File containing the default x509 Certificate for HTTPS. (CA cert, if any, concatenated after server cert). If HTTPS serving is enabled, and --tls-cert-file and --tls-private-key-file are not provided, a self-signed certificate and key are generated for the public address and saved to the directory specified by --cert-dir.
--tls-cipher-suites strings
Comma-separated list of cipher suites for the server. If omitted, the default Go cipher suites will be used. Preferred values: TLS_AES_128_GCM_SHA256, TLS_AES_256_GCM_SHA384, TLS_CHACHA20_POLY1305_SHA256, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_256_GCM_SHA384. Insecure values: TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_ECDSA_WITH_RC4_128_SHA, TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_RSA_WITH_RC4_128_SHA, TLS_RSA_WITH_3DES_EDE_CBC_SHA, TLS_RSA_WITH_AES_128_CBC_SHA256, TLS_RSA_WITH_RC4_128_SHA.
--tls-min-version string
Minimum TLS version supported. Possible values: VersionTLS10, VersionTLS11, VersionTLS12, VersionTLS13
--tls-private-key-file string
File containing the default x509 private key matching --tls-cert-file.
--tls-sni-cert-key string
A pair of x509 certificate and private key file paths, optionally suffixed with a list of domain patterns which are fully qualified domain names, possibly with prefixed wildcard segments. The domain patterns also allow IP addresses, but IPs should only be used if the apiserver has visibility to the IP address requested by a client. If no domain patterns are provided, the names of the certificate are extracted. Non-wildcard matches trump over wildcard matches, explicit domain patterns trump over extracted names. For multiple key/certificate pairs, use the --tls-sni-cert-key multiple times. Examples: "example.crt,example.key" or "foo.crt,foo.key:*.foo.com,foo.com".
--unhealthy-zone-threshold float Default: 0.55
Fraction of Nodes in a zone which needs to be not Ready (minimum 3) for zone to be treated as unhealthy.
--use-service-account-credentials
If true, use individual service account credentials for each controller.
-v, --v int
number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule pattern=N,...
comma-separated list of pattern=N settings for file-filtered logging (only works for text log format)
--volume-host-allow-local-loopback Default: true
If false, deny local loopback IPs in addition to any CIDR ranges in --volume-host-cidr-denylist
--volume-host-cidr-denylist strings
A comma-separated list of CIDR ranges to avoid from volume plugins.
6.11.5 - kube-proxy
Synopsis
The Kubernetes network proxy runs on each node. This
reflects services as defined in the Kubernetes API on each node and can do simple
TCP, UDP, and SCTP stream forwarding or round robin TCP, UDP, and SCTP forwarding across a set of backends.
Service cluster IPs and ports are currently found through Docker-links-compatible
environment variables specifying ports opened by the service proxy. There is an optional
addon that provides cluster DNS for these cluster IPs. The user must create a service
with the apiserver API to configure the proxy.
kube-proxy [flags]
Options
--add_dir_header
If true, adds the file directory to the header of the log messages
--alsologtostderr
log to standard error as well as files
--bind-address string Default: 0.0.0.0
The IP address for the proxy server to serve on (set to '0.0.0.0' for all IPv4 interfaces and '::' for all IPv6 interfaces). This parameter is ignored if a config file is specified by --config.
--bind-address-hard-fail
If true kube-proxy will treat failure to bind to a port as fatal and exit
Comma-separated list of files to check for boot-id. Use the first one that exists.
--cleanup
If true cleanup iptables and ipvs rules and exit.
--cluster-cidr string
The CIDR range of pods in the cluster. When configured, traffic sent to a Service cluster IP from outside this range will be masqueraded and traffic sent from pods to an external LoadBalancer IP will be directed to the respective cluster IP instead. For dual-stack clusters, a comma-separated list is accepted with at least one CIDR per IP family (IPv4 and IPv6). This parameter is ignored if a config file is specified by --config.
--config string
The path to the configuration file.
--config-sync-period duration Default: 15m0s
How often configuration from the apiserver is refreshed. Must be greater than 0.
--conntrack-max-per-core int32 Default: 32768
Maximum number of NAT connections to track per CPU core (0 to leave the limit as-is and ignore conntrack-min).
--conntrack-min int32 Default: 131072
Minimum number of conntrack entries to allocate, regardless of conntrack-max-per-core (set conntrack-max-per-core=0 to leave the limit as-is).
The IP address with port for the health check server to serve on (set to '0.0.0.0:10256' for all IPv4 interfaces and '[::]:10256' for all IPv6 interfaces). Set empty to disable. This parameter is ignored if a config file is specified by --config.
-h, --help
help for kube-proxy
--hostname-override string
If non-empty, will use this string as identification instead of the actual hostname.
--iptables-masquerade-bit int32 Default: 14
If using the pure iptables proxy, the bit of the fwmark space to mark packets requiring SNAT with. Must be within the range [0, 31].
--iptables-min-sync-period duration Default: 1s
The minimum interval of how often the iptables rules can be refreshed as endpoints and services change (e.g. '5s', '1m', '2h22m').
--iptables-sync-period duration Default: 30s
The maximum interval of how often iptables rules are refreshed (e.g. '5s', '1m', '2h22m'). Must be greater than 0.
--ipvs-exclude-cidrs strings
A comma-separated list of CIDR's which the ipvs proxier should not touch when cleaning up IPVS rules.
--ipvs-min-sync-period duration
The minimum interval of how often the ipvs rules can be refreshed as endpoints and services change (e.g. '5s', '1m', '2h22m').
--ipvs-scheduler string
The ipvs scheduler type when proxy mode is ipvs
--ipvs-strict-arp
Enable strict ARP by setting arp_ignore to 1 and arp_announce to 2
--ipvs-sync-period duration Default: 30s
The maximum interval of how often ipvs rules are refreshed (e.g. '5s', '1m', '2h22m'). Must be greater than 0.
--ipvs-tcp-timeout duration
The timeout for idle IPVS TCP connections, 0 to leave as-is. (e.g. '5s', '1m', '2h22m').
--ipvs-tcpfin-timeout duration
The timeout for IPVS TCP connections after receiving a FIN packet, 0 to leave as-is. (e.g. '5s', '1m', '2h22m').
--ipvs-udp-timeout duration
The timeout for IPVS UDP packets, 0 to leave as-is. (e.g. '5s', '1m', '2h22m').
--kube-api-burst int32 Default: 10
Burst to use while talking with kubernetes apiserver
The IP address with port for the metrics server to serve on (set to '0.0.0.0:10249' for all IPv4 interfaces and '[::]:10249' for all IPv6 interfaces). Set empty to disable. This parameter is ignored if a config file is specified by --config.
--nodeport-addresses strings
A string slice of values which specify the addresses to use for NodePorts. Values may be valid IP blocks (e.g. 1.2.3.0/24, 1.2.3.4/32). The default empty string slice ([]) means to use all local addresses. This parameter is ignored if a config file is specified by --config.
--one_output
If true, only write logs to their native severity level (vs also writing to each lower severity level)
--oom-score-adj int32 Default: -999
The oom-score-adj value for kube-proxy process. Values must be within the range [-1000, 1000]. This parameter is ignored if a config file is specified by --config.
--pod-bridge-interface string
A bridge interface name in the cluster. Kube-proxy considers traffic as local if originating from an interface which matches the value. This argument should be set if DetectLocalMode is set to BridgeInterface.
--pod-interface-name-prefix string
An interface prefix in the cluster. Kube-proxy considers traffic as local if originating from interfaces that match the given prefix. This argument should be set if DetectLocalMode is set to InterfaceNamePrefix.
--profiling
If true enables profiling via web interface on /debug/pprof handler. This parameter is ignored if a config file is specified by --config.
--proxy-mode ProxyMode
Which proxy mode to use: 'iptables' (Linux-only), 'ipvs' (Linux-only), 'kernelspace' (Windows-only), or 'userspace' (Linux/Windows, deprecated). The default value is 'iptables' on Linux and 'userspace' on Windows.This parameter is ignored if a config file is specified by --config.
--proxy-port-range port-range
Range of host ports (beginPort-endPort, single port or beginPort+offset, inclusive) that may be consumed in order to proxy service traffic. If (unspecified, 0, or 0-0) then ports will be randomly chosen.
--show-hidden-metrics-for-version string
The previous version for which you want to show hidden metrics. Only the previous minor version is meaningful, other values will not be allowed. The format is <major>.<minor>, e.g.: '1.16'. The purpose of this format is make sure you have the opportunity to notice if the next release hides additional metrics, rather than being surprised when they are permanently removed in the release after that.This parameter is ignored if a config file is specified by --config.
--skip_headers
If true, avoid header prefixes in the log messages
--skip_log_headers
If true, avoid headers when opening log files
--stderrthreshold int Default: 2
logs at or above this threshold go to stderr
--udp-timeout duration Default: 250ms
How long an idle UDP connection will be kept open (e.g. '250ms', '2s'). Must be greater than 0. Only applicable for proxy-mode=userspace
-v, --v int
number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule <comma-separated 'pattern=N' settings>
comma-separated list of pattern=N settings for file-filtered logging
--write-config-to string
If set, write the default configuration values to this file and exit.
6.11.6 - kube-scheduler
Synopsis
The Kubernetes scheduler is a control plane process which assigns
Pods to Nodes. The scheduler determines which Nodes are valid placements for
each Pod in the scheduling queue according to constraints and available
resources. The scheduler then ranks each valid Node and binds the Pod to a
suitable Node. Multiple different schedulers may be used within a cluster;
kube-scheduler is the reference implementation.
See scheduling
for more information about scheduling and the kube-scheduler component.
kube-scheduler [flags]
Options
--allow-metric-labels stringToString Default: []
The map from metric-label to value allow-list of this label. The key's format is <MetricName>,<LabelName>. The value's format is <allowed_value>,<allowed_value>...e.g. metric1,label1='v1,v2,v3', metric1,label2='v1,v2,v3' metric2,label1='v1,v2,v3'.
--authentication-kubeconfig string
kubeconfig file pointing at the 'core' kubernetes server with enough rights to create tokenreviews.authentication.k8s.io. This is optional. If empty, all token requests are considered to be anonymous and no client CA is looked up in the cluster.
--authentication-skip-lookup
If false, the authentication-kubeconfig will be used to lookup missing authentication configuration from the cluster.
If true, failures to look up missing authentication configuration from the cluster are not considered fatal. Note that this can result in authentication that treats all requests as anonymous.
A list of HTTP paths to skip during authorization, i.e. these are authorized without contacting the 'core' kubernetes server.
--authorization-kubeconfig string
kubeconfig file pointing at the 'core' kubernetes server with enough rights to create subjectaccessreviews.authorization.k8s.io. This is optional. If empty, all requests not skipped by authorization are forbidden.
The duration to cache 'unauthorized' responses from the webhook authorizer.
--azure-container-registry-config string
Path to the file containing Azure container registry configuration information.
--bind-address string Default: 0.0.0.0
The IP address on which to listen for the --secure-port port. The associated interface(s) must be reachable by the rest of the cluster, and by CLI/web clients. If blank or an unspecified address (0.0.0.0 or ::), all interfaces will be used.
--cert-dir string
The directory where the TLS certs are located. If --tls-cert-file and --tls-private-key-file are provided, this flag will be ignored.
--client-ca-file string
If set, any request presenting a client certificate signed by one of the authorities in the client-ca-file is authenticated with an identity corresponding to the CommonName of the client certificate.
--config string
The path to the configuration file.
--contention-profiling Default: true
DEPRECATED: enable lock contention profiling, if profiling is enabled. This parameter is ignored if a config file is specified in --config.
--disabled-metrics strings
This flag provides an escape hatch for misbehaving metrics. You must provide the fully qualified metric name in order to disable it. Disclaimer: disabling metrics is higher in precedence than showing hidden metrics.
DEPRECATED: content type of requests sent to apiserver. This parameter is ignored if a config file is specified in --config.
--kube-api-qps float Default: 50
DEPRECATED: QPS to use while talking with kubernetes apiserver. This parameter is ignored if a config file is specified in --config.
--kubeconfig string
DEPRECATED: path to kubeconfig file with authorization and master location information. This parameter is ignored if a config file is specified in --config.
--leader-elect Default: true
Start a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.
The duration that non-leader candidates will wait after observing a leadership renewal until attempting to acquire leadership of a led but unrenewed leader slot. This is effectively the maximum duration that a leader can be stopped before it is replaced by another candidate. This is only applicable if leader election is enabled.
The interval between attempts by the acting master to renew a leadership slot before it stops leading. This must be less than or equal to the lease duration. This is only applicable if leader election is enabled.
The namespace of resource object that is used for locking during leader election.
--leader-elect-retry-period duration Default: 2s
The duration the clients should wait between attempting acquisition and renewal of a leadership. This is only applicable if leader election is enabled.
DEPRECATED: define the name of the lock object. Will be removed in favor of leader-elect-resource-name. This parameter is ignored if a config file is specified in --config.
DEPRECATED: define the namespace of the lock object. Will be removed in favor of leader-elect-resource-namespace. This parameter is ignored if a config file is specified in --config.
--log-flush-frequency duration Default: 5s
Maximum number of seconds between log flushes
--logging-format string Default: "text"
Sets the log format. Permitted formats: "text". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log-dir, --log-file, --log-file-max-size, --logtostderr, --one-output, --skip-headers, --skip-log-headers, --stderrthreshold, --vmodule. Non-default choices are currently alpha and subject to change without warning.
--master string
The address of the Kubernetes API server (overrides any value in kubeconfig)
--permit-address-sharing
If true, SO_REUSEADDR will be used when binding the port. This allows binding to wildcard IPs like 0.0.0.0 and specific IPs in parallel, and it avoids waiting for the kernel to release sockets in TIME_WAIT state. [default=false]
--permit-port-sharing
If true, SO_REUSEPORT will be used when binding the port, which allows more than one instance to bind on the same address and port. [default=false]
DEPRECATED: the maximum time a pod can stay in unschedulablePods. If a pod stays in unschedulablePods for longer than this value, the pod will be moved from unschedulablePods to backoffQ or activeQ. This flag is deprecated and will be removed in 1.26
--profiling Default: true
DEPRECATED: enable profiling via web interface host:port/debug/pprof/. This parameter is ignored if a config file is specified in --config.
--requestheader-allowed-names strings
List of client certificate common names to allow to provide usernames in headers specified by --requestheader-username-headers. If empty, any client certificate validated by the authorities in --requestheader-client-ca-file is allowed.
--requestheader-client-ca-file string
Root certificate bundle to use to verify client certificates on incoming requests before trusting usernames in headers specified by --requestheader-username-headers. WARNING: generally do not depend on authorization being already done for incoming requests.
List of request headers to inspect for usernames. X-Remote-User is common.
--secure-port int Default: 10259
The port on which to serve HTTPS with authentication and authorization. If 0, don't serve HTTPS at all.
--show-hidden-metrics-for-version string
The previous version for which you want to show hidden metrics. Only the previous minor version is meaningful, other values will not be allowed. The format is <major>.<minor>, e.g.: '1.16'. The purpose of this format is make sure you have the opportunity to notice if the next release hides additional metrics, rather than being surprised when they are permanently removed in the release after that.
--tls-cert-file string
File containing the default x509 Certificate for HTTPS. (CA cert, if any, concatenated after server cert). If HTTPS serving is enabled, and --tls-cert-file and --tls-private-key-file are not provided, a self-signed certificate and key are generated for the public address and saved to the directory specified by --cert-dir.
--tls-cipher-suites strings
Comma-separated list of cipher suites for the server. If omitted, the default Go cipher suites will be used. Preferred values: TLS_AES_128_GCM_SHA256, TLS_AES_256_GCM_SHA384, TLS_CHACHA20_POLY1305_SHA256, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_256_GCM_SHA384. Insecure values: TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_ECDSA_WITH_RC4_128_SHA, TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_RSA_WITH_RC4_128_SHA, TLS_RSA_WITH_3DES_EDE_CBC_SHA, TLS_RSA_WITH_AES_128_CBC_SHA256, TLS_RSA_WITH_RC4_128_SHA.
--tls-min-version string
Minimum TLS version supported. Possible values: VersionTLS10, VersionTLS11, VersionTLS12, VersionTLS13
--tls-private-key-file string
File containing the default x509 private key matching --tls-cert-file.
--tls-sni-cert-key string
A pair of x509 certificate and private key file paths, optionally suffixed with a list of domain patterns which are fully qualified domain names, possibly with prefixed wildcard segments. The domain patterns also allow IP addresses, but IPs should only be used if the apiserver has visibility to the IP address requested by a client. If no domain patterns are provided, the names of the certificate are extracted. Non-wildcard matches trump over wildcard matches, explicit domain patterns trump over extracted names. For multiple key/certificate pairs, use the --tls-sni-cert-key multiple times. Examples: "example.crt,example.key" or "foo.crt,foo.key:*.foo.com,foo.com".
-v, --v int
number for the log level verbosity
--version version[=true]
Print version information and quit
--vmodule pattern=N,...
comma-separated list of pattern=N settings for file-filtered logging (only works for text log format)
--write-config-to string
If set, write the configuration values to this file and exit.
Cluster contains information to allow an exec plugin to communicate
with the kubernetes cluster being authenticated to.
To ensure that this struct contains everything someone would need to communicate
with a kubernetes cluster (just like they would via a kubeconfig), the fields
should shadow "k8s.io/client-go/tools/clientcmd/api/v1".Cluster, with the exception
of CertificateAuthority, since CA data will always be passed to the plugin as bytes.
Field
Description
server[Required] string
Server is the address of the kubernetes cluster (https://hostname:port).
tls-server-name string
TLSServerName is passed to the server for SNI and is used in the client to
check server certificates against. If ServerName is empty, the hostname
used to contact the server is used.
insecure-skip-tls-verify bool
InsecureSkipTLSVerify skips the validity check for the server's certificate.
This will make your HTTPS connections insecure.
certificate-authority-data []byte
CAData contains PEM-encoded certificate authority certificates.
If empty, system roots should be used.
proxy-url string
ProxyURL is the URL to the proxy to be used for all requests to this
cluster.
Config holds additional config data that is specific to the exec
plugin with regards to the cluster being authenticated to.
This data is sourced from the clientcmd Cluster object's
extensions[client.authentication.k8s.io/exec] field:
clusters:
name: my-cluster
cluster:
...
extensions:
name: client.authentication.k8s.io/exec # reserved extension name for per cluster exec config
extension:
audience: 06e3fbd18de8 # arbitrary config
In some environments, the user config may be exactly the same across many clusters
(i.e. call this exec plugin) minus some details that are specific to each cluster
such as the audience. This field allows the per cluster config to be directly
specified with the cluster info. Using this field to store secret data is not
recommended as one of the prime benefits of exec plugins is that no secrets need
to be stored directly in the kubeconfig.
Cluster contains information to allow an exec plugin to communicate with the
kubernetes cluster being authenticated to. Note that Cluster is non-nil only
when provideClusterInfo is set to true in the exec provider config (i.e.,
ExecConfig.ProvideClusterInfo).
interactive[Required] bool
Interactive declares whether stdin has been passed to this exec plugin.
ExecCredentialStatus holds credentials for the transport to use.
Token and ClientKeyData are sensitive fields. This data should only be
transmitted in-memory between client and exec plugin process. Exec plugin
itself should at least be protected via file permissions.
Cluster contains information to allow an exec plugin to communicate
with the kubernetes cluster being authenticated to.
To ensure that this struct contains everything someone would need to communicate
with a kubernetes cluster (just like they would via a kubeconfig), the fields
should shadow "k8s.io/client-go/tools/clientcmd/api/v1".Cluster, with the exception
of CertificateAuthority, since CA data will always be passed to the plugin as bytes.
Field
Description
server[Required] string
Server is the address of the kubernetes cluster (https://hostname:port).
tls-server-name string
TLSServerName is passed to the server for SNI and is used in the client to
check server certificates against. If ServerName is empty, the hostname
used to contact the server is used.
insecure-skip-tls-verify bool
InsecureSkipTLSVerify skips the validity check for the server's certificate.
This will make your HTTPS connections insecure.
certificate-authority-data []byte
CAData contains PEM-encoded certificate authority certificates.
If empty, system roots should be used.
proxy-url string
ProxyURL is the URL to the proxy to be used for all requests to this
cluster.
Config holds additional config data that is specific to the exec
plugin with regards to the cluster being authenticated to.
This data is sourced from the clientcmd Cluster object's
extensions[client.authentication.k8s.io/exec] field:
clusters:
name: my-cluster
cluster:
...
extensions:
name: client.authentication.k8s.io/exec # reserved extension name for per cluster exec config
extension:
audience: 06e3fbd18de8 # arbitrary config
In some environments, the user config may be exactly the same across many clusters
(i.e. call this exec plugin) minus some details that are specific to each cluster
such as the audience. This field allows the per cluster config to be directly
specified with the cluster info. Using this field to store secret data is not
recommended as one of the prime benefits of exec plugins is that no secrets need
to be stored directly in the kubeconfig.
Cluster contains information to allow an exec plugin to communicate with the
kubernetes cluster being authenticated to. Note that Cluster is non-nil only
when provideClusterInfo is set to true in the exec provider config (i.e.,
ExecConfig.ProvideClusterInfo).
interactive[Required] bool
Interactive declares whether stdin has been passed to this exec plugin.
ExecCredentialStatus holds credentials for the transport to use.
Token and ClientKeyData are sensitive fields. This data should only be
transmitted in-memory between client and exec plugin process. Exec plugin
itself should at least be protected via file permissions.
limits are the limits to place on event queries received.
Limits can be placed on events received server-wide, per namespace,
per user, and per source+object.
At least one limit is required.
type is the type of limit to which this configuration applies
qps[Required] int32
qps is the number of event queries per second that are allowed for this
type of limit. The qps and burst fields are used together to determine if
a particular event query is accepted. The qps determines how many queries
are accepted once the burst amount of queries has been exhausted.
burst[Required] int32
burst is the burst number of event queries that are allowed for this type
of limit. The qps and burst fields are used together to determine if a
particular event query is accepted. The burst determines the maximum size
of the allowance granted for a particular bucket. For example, if the burst
is 10 and the qps is 3, then the admission control will accept 10 queries
before blocking any queries. Every second, 3 more queries will be allowed.
If some of that allowance is not used, then it will roll over to the next
second, until the maximum allowance of 10 is reached.
cacheSize int32
cacheSize is the size of the LRU cache for this type of limit. If a bucket
is evicted from the cache, then the allowance for that bucket is reset. If
more queries are later received for an evicted bucket, then that bucket
will re-enter the cache with a clean slate, giving that bucket a full
allowance of burst queries.
The default cache size is 4096.
If limitType is 'server', then cacheSize is ignored.
Containers is a list of a subset of the information in each container of the Pod being created.
annotations map[string]string
Annotations is a list of key-value pairs extracted from the Pod's annotations.
It only includes keys which match the pattern *.image-policy.k8s.io/*.
It is up to each webhook backend to determine how to interpret these annotations, if at all.
namespace string
Namespace is the namespace the pod is being created in.
ImageReviewStatus is the result of the review for the pod creation request.
Field
Description
allowed[Required] bool
Allowed indicates that all images were allowed to be run.
reason string
Reason should be empty unless Allowed is false in which case it
may contain a short description of what is wrong. Kubernetes
may truncate excessively long errors when displaying to the user.
auditAnnotations map[string]string
AuditAnnotations will be added to the attributes object of the
admission controller request using 'AddAnnotation'. The keys should
be prefix-less (i.e., the admission controller will add an
appropriate prefix).
Source IPs, from where the request originated and intermediate proxies.
The source IPs are listed from (in order):
X-Forwarded-For request header IPs
X-Real-Ip header, if not present in the X-Forwarded-For list
The remote address for the connection, if it doesn't match the last
IP in the list up to here (X-Forwarded-For or X-Real-Ip).
Note: All but the last IP can be arbitrarily set by the client.
userAgent string
UserAgent records the user agent string reported by the client.
Note that the UserAgent is provided by the client, and must not be trusted.
The response status, populated even when the ResponseObject is not a Status type.
For successful responses, this will only include the Code and StatusSuccess.
For non-status type error responses, this will be auto-populated with the error Message.
API object from the request, in JSON format. The RequestObject is recorded as-is in the request
(possibly re-encoded as JSON), prior to version conversion, defaulting, admission or
merging. It is an external versioned object type, and may not be a valid object on its own.
Omitted for non-resource requests. Only logged at Request Level and higher.
API object returned in the response, in JSON. The ResponseObject is recorded after conversion
to the external type, and serialized as JSON. Omitted for non-resource requests. Only logged
at Response Level.
Annotations is an unstructured key value map stored with an audit event that may be set by
plugins invoked in the request serving chain, including authentication, authorization and
admission plugins. Note that these annotations are for the audit event, and do not correspond
to the metadata.annotations of the submitted object. Keys should uniquely identify the informing
component to avoid name collisions (e.g. podsecuritypolicy.admission.k8s.io/policy). Values
should be short. Annotations are included in the Metadata level.
Rules specify the audit Level a request should be recorded at.
A request may match multiple rules, in which case the FIRST matching rule is used.
The default audit level is None, but can be overridden by a catch-all rule at the end of the list.
PolicyRules are strictly ordered.
OmitStages is a list of stages for which no events are created. Note that this can also
be specified per rule in which case the union of both are omitted.
omitManagedFields bool
OmitManagedFields indicates whether to omit the managed fields of the request
and response bodies from being written to the API audit log.
This is used as a global default - a value of 'true' will omit the managed fileds,
otherwise the managed fields will be included in the API audit log.
Note that this can also be specified per rule in which case the value specified
in a rule will override the global default.
GroupResources represents resource kinds in an API group.
Field
Description
group string
Group is the name of the API group that contains the resources.
The empty string represents the core API group.
resources []string
Resources is a list of resources this rule applies to.
For example:
'pods' matches pods.
'pods/log' matches the log subresource of pods.
'' matches all resources and their subresources.
'pods/' matches all subresources of pods.
'*/scale' matches all scale subresources.
If wildcard is present, the validation rule will ensure resources do not
overlap with each other.
An empty list implies all resources and subresources in this API groups apply.
resourceNames []string
ResourceNames is a list of resource instance names that the policy matches.
Using this field requires Resources to be specified.
An empty list implies that every instance of the resource is matched.
The Level that requests matching this rule are recorded at.
users []string
The users (by authenticated user name) this rule applies to.
An empty list implies every user.
userGroups []string
The user groups this rule applies to. A user is considered matching
if it is a member of any of the UserGroups.
An empty list implies every user group.
verbs []string
The verbs that match this rule.
An empty list implies every verb.
Resources that this rule matches. An empty list implies all kinds in all API groups.
namespaces []string
Namespaces that this rule matches.
The empty string "" matches non-namespaced resources.
An empty list implies every namespace.
nonResourceURLs []string
NonResourceURLs is a set of URL paths that should be audited.
s are allowed, but only as the full, final step in the path.
Examples:
"/metrics" - Log requests for apiserver metrics
"/healthz" - Log all health checks
OmitStages is a list of stages for which no events are created. Note that this can also
be specified policy wide in which case the union of both are omitted.
An empty list means no restrictions will apply.
omitManagedFields bool
OmitManagedFields indicates whether to omit the managed fields of the request
and response bodies from being written to the API audit log.
a value of 'true' will drop the managed fields from the API audit log
a value of 'false' indicates that the managed fileds should be included
in the API audit log
Note that the value, if specified, in this rule will override the global default
If a value is not specified then the global default specified in
Policy.OmitManagedFields will stand.
Configuration is an embedded configuration object to be used as the plugin's
configuration. If present, it will be used instead of the path to the configuration file.
6.12.7 - kube-apiserver Configuration (v1alpha1)
Package v1alpha1 is the v1alpha1 version of the API.
connectionServices contains a list of egress selection client configurations
TracingConfiguration
TracingConfiguration provides versioned configuration for tracing clients.
Field
Description
apiVersion string
apiserver.k8s.io/v1alpha1
kind string
TracingConfiguration
endpoint string
Endpoint of the collector that's running on the control-plane node.
The APIServer uses the egressType ControlPlane when sending data to the collector.
The syntax is defined in https://github.com/grpc/grpc/blob/master/doc/naming.md.
Defaults to the otlpgrpc default, localhost:4317
The connection is insecure, and does not support TLS.
samplingRatePerMillion int32
SamplingRatePerMillion is the number of samples to collect per million spans.
Defaults to 0.
Configuration is an embedded configuration object to be used as the plugin's
configuration. If present, it will be used instead of the path to the configuration file.
EgressSelection provides the configuration for a single egress selection client.
Field
Description
name[Required] string
name is the name of the egress selection.
Currently supported values are "controlplane", "master", "etcd" and "cluster"
The "master" egress selector is deprecated in favor of "controlplane"
TLSConfig provides the authentication information to connect to konnectivity server
Only used with TCPTransport
Field
Description
caBundle string
caBundle is the file location of the CA to be used to determine trust with the konnectivity server.
Must be absent/empty if TCPTransport.URL is prefixed with http://
If absent while TCPTransport.URL is prefixed with https://, default to system trust roots.
clientKey string
clientKey is the file location of the client key to be used in mtls handshakes with the konnectivity server.
Must be absent/empty if TCPTransport.URL is prefixed with http://
Must be configured if TCPTransport.URL is prefixed with https://
clientCert string
clientCert is the file location of the client certificate to be used in mtls handshakes with the konnectivity server.
Must be absent/empty if TCPTransport.URL is prefixed with http://
Must be configured if TCPTransport.URL is prefixed with https://
TCP is the TCP configuration for communicating with the konnectivity server via TCP
ProxyProtocol of GRPC is not supported with TCP transport at the moment
Requires at least one of TCP or UDS to be set
UDSTransport provides the information to connect to konnectivity server via UDS
Field
Description
udsName[Required] string
UDSName is the name of the unix domain socket to connect to konnectivity server
This does not use a unix:// prefix. (Eg: /etc/srv/kubernetes/konnectivity-server/konnectivity-server.socket)
KubeProxyConfiguration contains everything necessary to configure the
Kubernetes proxy server.
Field
Description
apiVersion string
kubeproxy.config.k8s.io/v1alpha1
kind string
KubeProxyConfiguration
featureGates[Required] map[string]bool
featureGates is a map of feature names to bools that enable or disable alpha/experimental features.
bindAddress[Required] string
bindAddress is the IP address for the proxy server to serve on (set to 0.0.0.0
for all interfaces)
healthzBindAddress[Required] string
healthzBindAddress is the IP address and port for the health check server to serve on,
defaulting to 0.0.0.0:10256
metricsBindAddress[Required] string
metricsBindAddress is the IP address and port for the metrics server to serve on,
defaulting to 127.0.0.1:10249 (set to 0.0.0.0 for all interfaces)
bindAddressHardFail[Required] bool
bindAddressHardFail, if true, kube-proxy will treat failure to bind to a port as fatal and exit
enableProfiling[Required] bool
enableProfiling enables profiling via web interface on /debug/pprof handler.
Profiling handlers will be handled by metrics server.
clusterCIDR[Required] string
clusterCIDR is the CIDR range of the pods in the cluster. It is used to
bridge traffic coming from outside of the cluster. If not provided,
no off-cluster bridging will be performed.
hostnameOverride[Required] string
hostnameOverride, if non-empty, will be used as the identity instead of the actual hostname.
portRange is the range of host ports (beginPort-endPort, inclusive) that may be consumed
in order to proxy service traffic. If unspecified (0-0) then ports will be randomly chosen.
udpIdleTimeout is how long an idle UDP connection will be kept open (e.g. '250ms', '2s').
Must be greater than 0. Only applicable for proxyMode=userspace.
configSyncPeriod is how often configuration from the apiserver is refreshed. Must be greater
than 0.
nodePortAddresses[Required] []string
nodePortAddresses is the --nodeport-addresses value for kube-proxy process. Values must be valid
IP blocks. These values are as a parameter to select the interfaces where nodeport works.
In case someone would like to expose a service on localhost for local visit and some other interfaces for
particular purpose, a list of IP blocks would do that.
If set it to "127.0.0.0/8", kube-proxy will only select the loopback interface for NodePort.
If set it to a non-zero IP block, kube-proxy will filter that down to just the IPs that applied to the node.
An empty string slice is meant to select all network interfaces.
DetectLocalConfiguration contains optional settings related to DetectLocalMode option
Field
Description
bridgeInterface[Required] string
BridgeInterface is a string argument which represents a single bridge interface name.
Kube-proxy considers traffic as local if originating from this given bridge.
This argument should be set if DetectLocalMode is set to LocalModeBridgeInterface.
interfaceNamePrefix[Required] string
InterfaceNamePrefix is a string argument which represents a single interface prefix name.
Kube-proxy considers traffic as local if originating from one or more interfaces which match
the given prefix. This argument should be set if DetectLocalMode is set to LocalModeInterfaceNamePrefix.
tcpCloseWaitTimeout is how long an idle conntrack entry
in CLOSE_WAIT state will remain in the conntrack
table. (e.g. '60s'). Must be greater than 0 to set.
tcpFinTimeout is the timeout value used for IPVS TCP sessions after receiving a FIN.
The default value is 0, which preserves the current timeout value on the system.
ProxyMode represents modes used by the Kubernetes proxy server.
Currently, three modes of proxy are available in Linux platform: 'userspace' (older, going to be EOL), 'iptables'
(newer, faster), 'ipvs'(newest, better in performance and scalability).
Two modes of proxy are available in Windows platform: 'userspace'(older, stable) and 'kernelspace' (newer, faster).
In Linux platform, if proxy mode is blank, use the best-available proxy (currently iptables, but may change in the
future). If the iptables proxy is selected, regardless of how, but the system's kernel or iptables versions are
insufficient, this always falls back to the userspace proxy. IPVS mode will be enabled when proxy mode is set to 'ipvs',
and the fall back path is firstly iptables and then userspace.
In Windows platform, if proxy mode is blank, use the best-available proxy (currently userspace, but may change in the
future). If winkernel proxy is selected, regardless of how, but the Windows kernel can't support this mode of proxy,
this always falls back to the userspace proxy.
ClientConnectionConfiguration contains details for constructing a client.
Field
Description
kubeconfig[Required] string
kubeconfig is the path to a KubeConfig file.
acceptContentTypes[Required] string
acceptContentTypes defines the Accept header sent by clients when connecting to a server, overriding the
default value of 'application/json'. This field will control all connections to the server used by a particular
client.
contentType[Required] string
contentType is the content type used when sending data to the server from this client.
qps[Required] float32
qps controls the number of queries per second allowed for this connection.
burst[Required] int32
burst allows extra queries to accumulate when a client is exceeding its rate.
JSONOptions contains options for logging format "json".
Field
Description
splitStream[Required] bool
[Experimental] SplitStream redirects error messages to stderr while
info messages go to stdout, with buffering. The default is to write
both to stdout, without buffering.
LeaderElectionConfiguration defines the configuration of leader election
clients for components that can run with leader election enabled.
Field
Description
leaderElect[Required] bool
leaderElect enables a leader election client to gain leadership
before executing the main loop. Enable this when running replicated
components for high availability.
leaseDuration is the duration that non-leader candidates will wait
after observing a leadership renewal until attempting to acquire
leadership of a led but unrenewed leader slot. This is effectively the
maximum duration that a leader can be stopped before it is replaced
by another candidate. This is only applicable if leader election is
enabled.
renewDeadline is the interval between attempts by the acting master to
renew a leadership slot before it stops leading. This must be less
than or equal to the lease duration. This is only applicable if leader
election is enabled.
retryPeriod is the duration the clients should wait between attempting
acquisition and renewal of a leadership. This is only applicable if
leader election is enabled.
resourceLock[Required] string
resourceLock indicates the resource object type that will be used to lock
during leader election cycles.
resourceName[Required] string
resourceName indicates the name of resource object that will be used to lock
during leader election cycles.
resourceNamespace[Required] string
resourceName indicates the namespace of resource object that will be used to lock
during leader election cycles.
Maximum number of nanoseconds (i.e. 1s = 1000000000) between log
flushes. Ignored if the selected logging backend writes log
messages without buffering.
verbosity[Required] uint32
Verbosity is the threshold that determines which log messages are
logged. Default is zero which logs only the most important
messages. Higher values enable additional messages. Error messages
are always logged.
[Experimental] Options holds additional parameters that are specific
to the different logging formats. Only the options for the selected
format get used, but all of them get validated.
VModuleConfiguration
(Alias of []k8s.io/component-base/config/v1alpha1.VModuleItem)
ClientConnectionConfiguration contains details for constructing a client.
Field
Description
kubeconfig[Required] string
kubeconfig is the path to a KubeConfig file.
acceptContentTypes[Required] string
acceptContentTypes defines the Accept header sent by clients when connecting to a server, overriding the
default value of 'application/json'. This field will control all connections to the server used by a particular
client.
contentType[Required] string
contentType is the content type used when sending data to the server from this client.
qps[Required] float32
qps controls the number of queries per second allowed for this connection.
burst[Required] int32
burst allows extra queries to accumulate when a client is exceeding its rate.
JSONOptions contains options for logging format "json".
Field
Description
splitStream[Required] bool
[Experimental] SplitStream redirects error messages to stderr while
info messages go to stdout, with buffering. The default is to write
both to stdout, without buffering.
LeaderElectionConfiguration defines the configuration of leader election
clients for components that can run with leader election enabled.
Field
Description
leaderElect[Required] bool
leaderElect enables a leader election client to gain leadership
before executing the main loop. Enable this when running replicated
components for high availability.
leaseDuration is the duration that non-leader candidates will wait
after observing a leadership renewal until attempting to acquire
leadership of a led but unrenewed leader slot. This is effectively the
maximum duration that a leader can be stopped before it is replaced
by another candidate. This is only applicable if leader election is
enabled.
renewDeadline is the interval between attempts by the acting master to
renew a leadership slot before it stops leading. This must be less
than or equal to the lease duration. This is only applicable if leader
election is enabled.
retryPeriod is the duration the clients should wait between attempting
acquisition and renewal of a leadership. This is only applicable if
leader election is enabled.
resourceLock[Required] string
resourceLock indicates the resource object type that will be used to lock
during leader election cycles.
resourceName[Required] string
resourceName indicates the name of resource object that will be used to lock
during leader election cycles.
resourceNamespace[Required] string
resourceName indicates the namespace of resource object that will be used to lock
during leader election cycles.
Maximum number of nanoseconds (i.e. 1s = 1000000000) between log
flushes. Ignored if the selected logging backend writes log
messages without buffering.
verbosity[Required] uint32
Verbosity is the threshold that determines which log messages are
logged. Default is zero which logs only the most important
messages. Higher values enable additional messages. Error messages
are always logged.
[Experimental] Options holds additional parameters that are specific
to the different logging formats. Only the options for the selected
format get used, but all of them get validated.
VModuleConfiguration
(Alias of []k8s.io/component-base/config/v1alpha1.VModuleItem)
VModuleConfiguration is a collection of individual file names or patterns
and the corresponding verbosity threshold.
DefaultPreemptionArgs
DefaultPreemptionArgs holds arguments used to configure the
DefaultPreemption plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta2
kind string
DefaultPreemptionArgs
minCandidateNodesPercentage[Required] int32
MinCandidateNodesPercentage is the minimum number of candidates to
shortlist when dry running preemption as a percentage of number of nodes.
Must be in the range [0, 100]. Defaults to 10% of the cluster size if
unspecified.
minCandidateNodesAbsolute[Required] int32
MinCandidateNodesAbsolute is the absolute minimum number of candidates to
shortlist. The likely number of candidates enumerated for dry running
preemption is given by the formula:
numCandidates = max(numNodes * minCandidateNodesPercentage, minCandidateNodesAbsolute)
We say "likely" because there are other factors such as PDB violations
that play a role in the number of candidates shortlisted. Must be at least
0 nodes. Defaults to 100 nodes if unspecified.
InterPodAffinityArgs
InterPodAffinityArgs holds arguments used to configure the InterPodAffinity plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta2
kind string
InterPodAffinityArgs
hardPodAffinityWeight[Required] int32
HardPodAffinityWeight is the scoring weight for existing pods with a
matching hard affinity to the incoming pod.
KubeSchedulerConfiguration
KubeSchedulerConfiguration configures a scheduler
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta2
kind string
KubeSchedulerConfiguration
parallelism[Required] int32
Parallelism defines the amount of parallelism in algorithms for scheduling a Pods. Must be greater than 0. Defaults to 16
ClientConnection specifies the kubeconfig file and client connection
settings for the proxy server to use when communicating with the apiserver.
healthzBindAddress[Required] string
Note: Both HealthzBindAddress and MetricsBindAddress fields are deprecated.
Only empty address or port 0 is allowed. Anything else will fail validation.
HealthzBindAddress is the IP address and port for the health check server to serve on.
metricsBindAddress[Required] string
MetricsBindAddress is the IP address and port for the metrics server to serve on.
(Members of DebuggingConfiguration are embedded into this type.)
DebuggingConfiguration holds configuration for Debugging related features
TODO: We might wanna make this a substruct like Debugging componentbaseconfigv1alpha1.DebuggingConfiguration
percentageOfNodesToScore[Required] int32
PercentageOfNodesToScore is the percentage of all nodes that once found feasible
for running a pod, the scheduler stops its search for more feasible nodes in
the cluster. This helps improve scheduler's performance. Scheduler always tries to find
at least "minFeasibleNodesToFind" feasible nodes no matter what the value of this flag is.
Example: if the cluster size is 500 nodes and the value of this flag is 30,
then scheduler stops finding further feasible nodes once it finds 150 feasible ones.
When the value is 0, default percentage (5%--50% based on the size of the cluster) of the
nodes will be scored.
podInitialBackoffSeconds[Required] int64
PodInitialBackoffSeconds is the initial backoff for unschedulable pods.
If specified, it must be greater than 0. If this value is null, the default value (1s)
will be used.
podMaxBackoffSeconds[Required] int64
PodMaxBackoffSeconds is the max backoff for unschedulable pods.
If specified, it must be greater than podInitialBackoffSeconds. If this value is null,
the default value (10s) will be used.
Profiles are scheduling profiles that kube-scheduler supports. Pods can
choose to be scheduled under a particular profile by setting its associated
scheduler name. Pods that don't specify any scheduler name are scheduled
with the "default-scheduler" profile, if present here.
Extenders are the list of scheduler extenders, each holding the values of how to communicate
with the extender. These extenders are shared by all scheduler profiles.
NodeAffinityArgs
NodeAffinityArgs holds arguments to configure the NodeAffinity plugin.
AddedAffinity is applied to all Pods additionally to the NodeAffinity
specified in the PodSpec. That is, Nodes need to satisfy AddedAffinity
AND .spec.NodeAffinity. AddedAffinity is empty by default (all Nodes
match).
When AddedAffinity is used, some Pods with affinity requirements that match
a specific Node (such as Daemonset Pods) might remain unschedulable.
NodeResourcesBalancedAllocationArgs
NodeResourcesBalancedAllocationArgs holds arguments used to configure NodeResourcesBalancedAllocation plugin.
Resources to be managed, the default is "cpu" and "memory" if not specified.
NodeResourcesFitArgs
NodeResourcesFitArgs holds arguments used to configure the NodeResourcesFit plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta2
kind string
NodeResourcesFitArgs
ignoredResources[Required] []string
IgnoredResources is the list of resources that NodeResources fit filter
should ignore. This doesn't apply to scoring.
ignoredResourceGroups[Required] []string
IgnoredResourceGroups defines the list of resource groups that NodeResources fit filter should ignore.
e.g. if group is ["example.com"], it will ignore all resource names that begin
with "example.com", such as "example.com/aaa" and "example.com/bbb".
A resource group name can't contain '/'. This doesn't apply to scoring.
DefaultConstraints defines topology spread constraints to be applied to
Pods that don't define any in pod.spec.topologySpreadConstraints.
.defaultConstraints[*].labelSelectors must be empty, as they are
deduced from the Pod's membership to Services, ReplicationControllers,
ReplicaSets or StatefulSets.
When not empty, .defaultingType must be "List".
DefaultingType determines how .defaultConstraints are deduced. Can be one
of "System" or "List".
"System": Use kubernetes defined constraints that spread Pods among
Nodes and Zones.
"List": Use constraints defined in .defaultConstraints.
Defaults to "System".
VolumeBindingArgs
VolumeBindingArgs holds arguments used to configure the VolumeBinding plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta2
kind string
VolumeBindingArgs
bindTimeoutSeconds[Required] int64
BindTimeoutSeconds is the timeout in seconds in volume binding operation.
Value must be non-negative integer. The value zero indicates no waiting.
If this value is nil, the default value (600) will be used.
Shape specifies the points defining the score function shape, which is
used to score nodes based on the utilization of statically provisioned
PVs. The utilization is calculated by dividing the total requested
storage of the pod by the total capacity of feasible PVs on each node.
Each point contains utilization (ranges from 0 to 100) and its
associated score (ranges from 0 to 10). You can turn the priority by
specifying different scores for different utilization numbers.
The default shape points are:
0 for 0 utilization
10 for 100 utilization
All points must be sorted in increasing order by utilization.
Extender holds the parameters used to communicate with the extender. If a verb is unspecified/empty,
it is assumed that the extender chose not to provide that extension.
Field
Description
urlPrefix[Required] string
URLPrefix at which the extender is available
filterVerb[Required] string
Verb for the filter call, empty if not supported. This verb is appended to the URLPrefix when issuing the filter call to extender.
preemptVerb[Required] string
Verb for the preempt call, empty if not supported. This verb is appended to the URLPrefix when issuing the preempt call to extender.
prioritizeVerb[Required] string
Verb for the prioritize call, empty if not supported. This verb is appended to the URLPrefix when issuing the prioritize call to extender.
weight[Required] int64
The numeric multiplier for the node scores that the prioritize call generates.
The weight should be a positive integer
bindVerb[Required] string
Verb for the bind call, empty if not supported. This verb is appended to the URLPrefix when issuing the bind call to extender.
If this method is implemented by the extender, it is the extender's responsibility to bind the pod to apiserver. Only one extender
can implement this function.
enableHTTPS[Required] bool
EnableHTTPS specifies whether https should be used to communicate with the extender
HTTPTimeout specifies the timeout duration for a call to the extender. Filter timeout fails the scheduling of the pod. Prioritize
timeout is ignored, k8s/other extenders priorities are used to select the node.
nodeCacheCapable[Required] bool
NodeCacheCapable specifies that the extender is capable of caching node information,
so the scheduler should only send minimal information about the eligible nodes
assuming that the extender already cached full details of all nodes in the cluster
ManagedResources is a list of extended resources that are managed by
this extender.
A pod will be sent to the extender on the Filter, Prioritize and Bind
(if the extender is the binder) phases iff the pod requests at least
one of the extended resources in this list. If empty or unspecified,
all pods will be sent to this extender.
If IgnoredByScheduler is set to true for a resource, kube-scheduler
will skip checking the resource in predicates.
ignorable[Required] bool
Ignorable specifies if the extender is ignorable, i.e. scheduling should not
fail when the extender returns an error or is not reachable.
ExtenderTLSConfig contains settings to enable TLS with extender
Field
Description
insecure[Required] bool
Server should be accessed without verifying the TLS certificate. For testing only.
serverName[Required] string
ServerName is passed to the server for SNI and is used in the client to check server
certificates against. If ServerName is empty, the hostname used to contact the
server is used.
certFile[Required] string
Server requires TLS client certificate authentication
keyFile[Required] string
Server requires TLS client certificate authentication
caFile[Required] string
Trusted root certificates for server
certData[Required] []byte
CertData holds PEM-encoded bytes (typically read from a client certificate file).
CertData takes precedence over CertFile
keyData[Required] []byte
KeyData holds PEM-encoded bytes (typically read from a client certificate key file).
KeyData takes precedence over KeyFile
caData[Required] []byte
CAData holds PEM-encoded bytes (typically read from a root certificates bundle).
CAData takes precedence over CAFile
SchedulerName is the name of the scheduler associated to this profile.
If SchedulerName matches with the pod's "spec.schedulerName", then the pod
is scheduled with this profile.
Plugins specify the set of plugins that should be enabled or disabled.
Enabled plugins are the ones that should be enabled in addition to the
default plugins. Disabled plugins are any of the default plugins that
should be disabled.
When no enabled or disabled plugin is specified for an extension point,
default plugins for that extension point will be used if there is any.
If a QueueSort plugin is specified, the same QueueSort Plugin and
PluginConfig must be specified for all profiles.
PluginConfig is an optional set of custom plugin arguments for each plugin.
Omitting config args for a plugin is equivalent to using the default config
for that plugin.
PluginConfig specifies arguments that should be passed to a plugin at the time of initialization.
A plugin that is invoked at multiple extension points is initialized once. Args can have arbitrary structure.
It is up to the plugin to process these Args.
PluginSet specifies enabled and disabled plugins for an extension point.
If an array is empty, missing, or nil, default plugins at that extension point will be used.
Enabled specifies plugins that should be enabled in addition to default plugins.
If the default plugin is also configured in the scheduler config file, the weight of plugin will
be overridden accordingly.
These are called after default plugins and in the same order specified here.
Disabled specifies default plugins that should be disabled.
When all default plugins need to be disabled, an array containing only one "*" should be provided.
Plugins include multiple extension points. When specified, the list of plugins for
a particular extension point are the only ones enabled. If an extension point is
omitted from the config, then the default set of plugins is used for that extension point.
Enabled plugins are called in the order specified here, after default plugins. If they need to
be invoked before default plugins, default plugins must be disabled and re-enabled here in desired order.
Bind is a list of plugins that should be invoked at "Bind" extension point of the scheduling framework.
The scheduler call these plugins in order. Scheduler skips the rest of these plugins as soon as one returns success.
Resources to consider when scoring.
The default resource set includes "cpu" and "memory" with an equal weight.
Allowed weights go from 1 to 100.
Weight defaults to 1 if not specified or explicitly set to 0.
ClientConnectionConfiguration contains details for constructing a client.
Field
Description
kubeconfig[Required] string
kubeconfig is the path to a KubeConfig file.
acceptContentTypes[Required] string
acceptContentTypes defines the Accept header sent by clients when connecting to a server, overriding the
default value of 'application/json'. This field will control all connections to the server used by a particular
client.
contentType[Required] string
contentType is the content type used when sending data to the server from this client.
qps[Required] float32
qps controls the number of queries per second allowed for this connection.
burst[Required] int32
burst allows extra queries to accumulate when a client is exceeding its rate.
JSONOptions contains options for logging format "json".
Field
Description
splitStream[Required] bool
[Experimental] SplitStream redirects error messages to stderr while
info messages go to stdout, with buffering. The default is to write
both to stdout, without buffering.
LeaderElectionConfiguration defines the configuration of leader election
clients for components that can run with leader election enabled.
Field
Description
leaderElect[Required] bool
leaderElect enables a leader election client to gain leadership
before executing the main loop. Enable this when running replicated
components for high availability.
leaseDuration is the duration that non-leader candidates will wait
after observing a leadership renewal until attempting to acquire
leadership of a led but unrenewed leader slot. This is effectively the
maximum duration that a leader can be stopped before it is replaced
by another candidate. This is only applicable if leader election is
enabled.
renewDeadline is the interval between attempts by the acting master to
renew a leadership slot before it stops leading. This must be less
than or equal to the lease duration. This is only applicable if leader
election is enabled.
retryPeriod is the duration the clients should wait between attempting
acquisition and renewal of a leadership. This is only applicable if
leader election is enabled.
resourceLock[Required] string
resourceLock indicates the resource object type that will be used to lock
during leader election cycles.
resourceName[Required] string
resourceName indicates the name of resource object that will be used to lock
during leader election cycles.
resourceNamespace[Required] string
resourceName indicates the namespace of resource object that will be used to lock
during leader election cycles.
Maximum number of nanoseconds (i.e. 1s = 1000000000) between log
flushes. Ignored if the selected logging backend writes log
messages without buffering.
verbosity[Required] uint32
Verbosity is the threshold that determines which log messages are
logged. Default is zero which logs only the most important
messages. Higher values enable additional messages. Error messages
are always logged.
[Experimental] Options holds additional parameters that are specific
to the different logging formats. Only the options for the selected
format get used, but all of them get validated.
VModuleConfiguration
(Alias of []k8s.io/component-base/config/v1alpha1.VModuleItem)
VModuleConfiguration is a collection of individual file names or patterns
and the corresponding verbosity threshold.
DefaultPreemptionArgs
DefaultPreemptionArgs holds arguments used to configure the
DefaultPreemption plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta3
kind string
DefaultPreemptionArgs
minCandidateNodesPercentage[Required] int32
MinCandidateNodesPercentage is the minimum number of candidates to
shortlist when dry running preemption as a percentage of number of nodes.
Must be in the range [0, 100]. Defaults to 10% of the cluster size if
unspecified.
minCandidateNodesAbsolute[Required] int32
MinCandidateNodesAbsolute is the absolute minimum number of candidates to
shortlist. The likely number of candidates enumerated for dry running
preemption is given by the formula:
numCandidates = max(numNodes * minCandidateNodesPercentage, minCandidateNodesAbsolute)
We say "likely" because there are other factors such as PDB violations
that play a role in the number of candidates shortlisted. Must be at least
0 nodes. Defaults to 100 nodes if unspecified.
InterPodAffinityArgs
InterPodAffinityArgs holds arguments used to configure the InterPodAffinity plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta3
kind string
InterPodAffinityArgs
hardPodAffinityWeight[Required] int32
HardPodAffinityWeight is the scoring weight for existing pods with a
matching hard affinity to the incoming pod.
KubeSchedulerConfiguration
KubeSchedulerConfiguration configures a scheduler
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta3
kind string
KubeSchedulerConfiguration
parallelism[Required] int32
Parallelism defines the amount of parallelism in algorithms for scheduling a Pods. Must be greater than 0. Defaults to 16
(Members of DebuggingConfiguration are embedded into this type.)
DebuggingConfiguration holds configuration for Debugging related features
TODO: We might wanna make this a substruct like Debugging componentbaseconfigv1alpha1.DebuggingConfiguration
percentageOfNodesToScore[Required] int32
PercentageOfNodesToScore is the percentage of all nodes that once found feasible
for running a pod, the scheduler stops its search for more feasible nodes in
the cluster. This helps improve scheduler's performance. Scheduler always tries to find
at least "minFeasibleNodesToFind" feasible nodes no matter what the value of this flag is.
Example: if the cluster size is 500 nodes and the value of this flag is 30,
then scheduler stops finding further feasible nodes once it finds 150 feasible ones.
When the value is 0, default percentage (5%--50% based on the size of the cluster) of the
nodes will be scored.
podInitialBackoffSeconds[Required] int64
PodInitialBackoffSeconds is the initial backoff for unschedulable pods.
If specified, it must be greater than 0. If this value is null, the default value (1s)
will be used.
podMaxBackoffSeconds[Required] int64
PodMaxBackoffSeconds is the max backoff for unschedulable pods.
If specified, it must be greater than podInitialBackoffSeconds. If this value is null,
the default value (10s) will be used.
Profiles are scheduling profiles that kube-scheduler supports. Pods can
choose to be scheduled under a particular profile by setting its associated
scheduler name. Pods that don't specify any scheduler name are scheduled
with the "default-scheduler" profile, if present here.
Extenders are the list of scheduler extenders, each holding the values of how to communicate
with the extender. These extenders are shared by all scheduler profiles.
NodeAffinityArgs
NodeAffinityArgs holds arguments to configure the NodeAffinity plugin.
AddedAffinity is applied to all Pods additionally to the NodeAffinity
specified in the PodSpec. That is, Nodes need to satisfy AddedAffinity
AND .spec.NodeAffinity. AddedAffinity is empty by default (all Nodes
match).
When AddedAffinity is used, some Pods with affinity requirements that match
a specific Node (such as Daemonset Pods) might remain unschedulable.
NodeResourcesBalancedAllocationArgs
NodeResourcesBalancedAllocationArgs holds arguments used to configure NodeResourcesBalancedAllocation plugin.
Resources to be managed, the default is "cpu" and "memory" if not specified.
NodeResourcesFitArgs
NodeResourcesFitArgs holds arguments used to configure the NodeResourcesFit plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta3
kind string
NodeResourcesFitArgs
ignoredResources[Required] []string
IgnoredResources is the list of resources that NodeResources fit filter
should ignore. This doesn't apply to scoring.
ignoredResourceGroups[Required] []string
IgnoredResourceGroups defines the list of resource groups that NodeResources fit filter should ignore.
e.g. if group is ["example.com"], it will ignore all resource names that begin
with "example.com", such as "example.com/aaa" and "example.com/bbb".
A resource group name can't contain '/'. This doesn't apply to scoring.
DefaultConstraints defines topology spread constraints to be applied to
Pods that don't define any in pod.spec.topologySpreadConstraints.
.defaultConstraints[*].labelSelectors must be empty, as they are
deduced from the Pod's membership to Services, ReplicationControllers,
ReplicaSets or StatefulSets.
When not empty, .defaultingType must be "List".
DefaultingType determines how .defaultConstraints are deduced. Can be one
of "System" or "List".
"System": Use kubernetes defined constraints that spread Pods among
Nodes and Zones.
"List": Use constraints defined in .defaultConstraints.
Defaults to "System".
VolumeBindingArgs
VolumeBindingArgs holds arguments used to configure the VolumeBinding plugin.
Field
Description
apiVersion string
kubescheduler.config.k8s.io/v1beta3
kind string
VolumeBindingArgs
bindTimeoutSeconds[Required] int64
BindTimeoutSeconds is the timeout in seconds in volume binding operation.
Value must be non-negative integer. The value zero indicates no waiting.
If this value is nil, the default value (600) will be used.
Shape specifies the points defining the score function shape, which is
used to score nodes based on the utilization of statically provisioned
PVs. The utilization is calculated by dividing the total requested
storage of the pod by the total capacity of feasible PVs on each node.
Each point contains utilization (ranges from 0 to 100) and its
associated score (ranges from 0 to 10). You can turn the priority by
specifying different scores for different utilization numbers.
The default shape points are:
0 for 0 utilization
10 for 100 utilization
All points must be sorted in increasing order by utilization.
Extender holds the parameters used to communicate with the extender. If a verb is unspecified/empty,
it is assumed that the extender chose not to provide that extension.
Field
Description
urlPrefix[Required] string
URLPrefix at which the extender is available
filterVerb[Required] string
Verb for the filter call, empty if not supported. This verb is appended to the URLPrefix when issuing the filter call to extender.
preemptVerb[Required] string
Verb for the preempt call, empty if not supported. This verb is appended to the URLPrefix when issuing the preempt call to extender.
prioritizeVerb[Required] string
Verb for the prioritize call, empty if not supported. This verb is appended to the URLPrefix when issuing the prioritize call to extender.
weight[Required] int64
The numeric multiplier for the node scores that the prioritize call generates.
The weight should be a positive integer
bindVerb[Required] string
Verb for the bind call, empty if not supported. This verb is appended to the URLPrefix when issuing the bind call to extender.
If this method is implemented by the extender, it is the extender's responsibility to bind the pod to apiserver. Only one extender
can implement this function.
enableHTTPS[Required] bool
EnableHTTPS specifies whether https should be used to communicate with the extender
HTTPTimeout specifies the timeout duration for a call to the extender. Filter timeout fails the scheduling of the pod. Prioritize
timeout is ignored, k8s/other extenders priorities are used to select the node.
nodeCacheCapable[Required] bool
NodeCacheCapable specifies that the extender is capable of caching node information,
so the scheduler should only send minimal information about the eligible nodes
assuming that the extender already cached full details of all nodes in the cluster
ManagedResources is a list of extended resources that are managed by
this extender.
A pod will be sent to the extender on the Filter, Prioritize and Bind
(if the extender is the binder) phases iff the pod requests at least
one of the extended resources in this list. If empty or unspecified,
all pods will be sent to this extender.
If IgnoredByScheduler is set to true for a resource, kube-scheduler
will skip checking the resource in predicates.
ignorable[Required] bool
Ignorable specifies if the extender is ignorable, i.e. scheduling should not
fail when the extender returns an error or is not reachable.
ExtenderTLSConfig contains settings to enable TLS with extender
Field
Description
insecure[Required] bool
Server should be accessed without verifying the TLS certificate. For testing only.
serverName[Required] string
ServerName is passed to the server for SNI and is used in the client to check server
certificates against. If ServerName is empty, the hostname used to contact the
server is used.
certFile[Required] string
Server requires TLS client certificate authentication
keyFile[Required] string
Server requires TLS client certificate authentication
caFile[Required] string
Trusted root certificates for server
certData[Required] []byte
CertData holds PEM-encoded bytes (typically read from a client certificate file).
CertData takes precedence over CertFile
keyData[Required] []byte
KeyData holds PEM-encoded bytes (typically read from a client certificate key file).
KeyData takes precedence over KeyFile
caData[Required] []byte
CAData holds PEM-encoded bytes (typically read from a root certificates bundle).
CAData takes precedence over CAFile
SchedulerName is the name of the scheduler associated to this profile.
If SchedulerName matches with the pod's "spec.schedulerName", then the pod
is scheduled with this profile.
Plugins specify the set of plugins that should be enabled or disabled.
Enabled plugins are the ones that should be enabled in addition to the
default plugins. Disabled plugins are any of the default plugins that
should be disabled.
When no enabled or disabled plugin is specified for an extension point,
default plugins for that extension point will be used if there is any.
If a QueueSort plugin is specified, the same QueueSort Plugin and
PluginConfig must be specified for all profiles.
PluginConfig is an optional set of custom plugin arguments for each plugin.
Omitting config args for a plugin is equivalent to using the default config
for that plugin.
PluginConfig specifies arguments that should be passed to a plugin at the time of initialization.
A plugin that is invoked at multiple extension points is initialized once. Args can have arbitrary structure.
It is up to the plugin to process these Args.
PluginSet specifies enabled and disabled plugins for an extension point.
If an array is empty, missing, or nil, default plugins at that extension point will be used.
Enabled specifies plugins that should be enabled in addition to default plugins.
If the default plugin is also configured in the scheduler config file, the weight of plugin will
be overridden accordingly.
These are called after default plugins and in the same order specified here.
Disabled specifies default plugins that should be disabled.
When all default plugins need to be disabled, an array containing only one "*" should be provided.
Plugins include multiple extension points. When specified, the list of plugins for
a particular extension point are the only ones enabled. If an extension point is
omitted from the config, then the default set of plugins is used for that extension point.
Enabled plugins are called in the order specified here, after default plugins. If they need to
be invoked before default plugins, default plugins must be disabled and re-enabled here in desired order.
Bind is a list of plugins that should be invoked at "Bind" extension point of the scheduling framework.
The scheduler call these plugins in order. Scheduler skips the rest of these plugins as soon as one returns success.
MultiPoint is a simplified config section to enable plugins for all valid extension points.
Plugins enabled through MultiPoint will automatically register for every individual extension
point the plugin has implemented. Disabling a plugin through MultiPoint disables that behavior.
The same is true for disabling "*" through MultiPoint (no default plugins will be automatically registered).
Plugins can still be disabled through their individual extension points.
In terms of precedence, plugin config follows this basic hierarchy
Specific extension points
Explicitly configured MultiPoint plugins
The set of default plugins, as MultiPoint plugins
This implies that a higher precedence plugin will run first and overwrite any settings within MultiPoint.
Explicitly user-configured plugins also take a higher precedence over default plugins.
Within this hierarchy, an Enabled setting takes precedence over Disabled. For example, if a plugin is
set in both multiPoint.Enabled and multiPoint.Disabled, the plugin will be enabled. Similarly,
including multiPoint.Disabled = '*' and multiPoint.Enabled = pluginA will still register that specific
plugin through MultiPoint. This follows the same behavior as all other extension point configurations.
Resources to consider when scoring.
The default resource set includes "cpu" and "memory" with an equal weight.
Allowed weights go from 1 to 100.
Weight defaults to 1 if not specified or explicitly set to 0.
UtilizationShapePoint represents single point of priority function shape.
Field
Description
utilization[Required] int32
Utilization (x axis). Valid values are 0 to 100. Fully utilized node maps to 100.
score[Required] int32
Score assigned to given utilization (y axis). Valid values are 0 to 10.
6.12.12 - kubeadm Configuration (v1beta2)
Overview
Package v1beta2 defines the v1beta2 version of the kubeadm configuration file format.
This version improves on the v1beta1 format by fixing some minor issues and adding a few new fields.
A list of changes since v1beta1:
"certificateKey" field is added to InitConfiguration and JoinConfiguration.
"ignorePreflightErrors" field is added to the NodeRegistrationOptions.
The JSON "omitempty" tag is used in a more places where appropriate.
The JSON "omitempty" tag of the "taints" field (inside NodeRegistrationOptions) is removed.
See the Kubernetes 1.15 changelog for further details.
Migration from old kubeadm config versions
Please convert your v1beta1 configuration files to v1beta2 using the "kubeadm config migrate" command of kubeadm v1.15.x
(conversion from older releases of kubeadm config files requires older release of kubeadm as well e.g.
kubeadm v1.11 should be used to migrate v1alpha1 to v1alpha2; kubeadm v1.12 should be used to translate v1alpha2 to v1alpha3;
kubeadm v1.13 or v1.14 should be used to translate v1alpha3 to v1beta1)
Nevertheless, kubeadm v1.15.x will support reading from v1beta1 version of the kubeadm config file format.
Basics
The preferred way to configure kubeadm is to pass an YAML configuration file with the --config option. Some of the
configuration options defined in the kubeadm config file are also available as command line flags, but only
the most common/simple use case are supported with this approach.
A kubeadm config file could contain multiple configuration types separated using three dashes (---).
kubeadm supports the following configuration types:
The list of configuration types that must be included in a configuration file depends by the action you are
performing (init or join) and by the configuration options you are going to use (defaults or advanced customization).
If some configuration types are not provided, or provided only partially, kubeadm will use default values; defaults
provided by kubeadm includes also enforcing consistency of values across components when required (e.g.
--cluster-cidr flag on controller manager and clusterCIDR on kube-proxy).
Users are always allowed to override default values, with the only exception of a small subset of setting with
relevance for security (e.g. enforce authorization-mode Node and RBAC on API server)
If the user provides a configuration types that is not expected for the action you are performing, kubeadm will
ignore those types and print a warning.
Kubeadm init configuration types
When executing kubeadm init with the --config option, the following configuration types could be used:
InitConfiguration, ClusterConfiguration, KubeProxyConfiguration, KubeletConfiguration, but only one
between InitConfiguration and ClusterConfiguration is mandatory.
The InitConfiguration type should be used to configure runtime settings, that in case of kubeadm init
are the configuration of the bootstrap token and all the setting which are specific to the node where kubeadm
is executed, including:
nodeRegistration, that holds fields that relate to registering the new node to the cluster;
use it to customize the node name, the CRI socket to use or any other settings that should apply to this
node only (e.g. the node ip).
apiServer, that represents the endpoint of the instance of the API server to be deployed on this node;
use it e.g. to customize the API server advertise address.
The ClusterConfiguration type should be used to configure cluster-wide settings,
including settings for:
Networking, that holds configuration for the networking topology of the cluster; use it e.g. to customize
pod subnet or services subnet.
Etcd configurations; use it e.g. to customize the local etcd or to configure the API server
for using an external etcd cluster.
kube-apiserver, kube-scheduler, kube-controller-manager configurations; use it to customize control-plane
components by adding customized setting or overriding kubeadm default settings.
The KubeProxyConfiguration type should be used to change the configuration passed to kube-proxy instances deployed
in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kube-proxy/ or
https://pkg.go.dev/k8s.io/kube-proxy/config/v1alpha1#KubeProxyConfiguration
for kube proxy official documentation.
The KubeletConfiguration type should be used to change the configurations that will be passed to all kubelet instances
deployed in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kubelet/ or
https://pkg.go.dev/k8s.io/kubelet/config/v1beta1#KubeletConfiguration
for kubelet official documentation.
Here is a fully populated example of a single YAML file containing multiple
configuration types to be used during a kubeadm init run.
apiVersion:kubeadm.k8s.io/v1beta2kind:InitConfigurationbootstrapTokens:- token:"9a08jv.c0izixklcxtmnze7"description:"kubeadm bootstrap token"ttl:"24h"- token:"783bde.3f89s0fje9f38fhf"description:"another bootstrap token"usages:- authentication- signinggroups:- system:bootstrappers:kubeadm:default-node-tokennodeRegistration:name:"ec2-10-100-0-1"criSocket:"/var/run/dockershim.sock"taints:- key:"kubeadmNode"value:"someValue"effect:"NoSchedule"kubeletExtraArgs:v:4ignorePreflightErrors:- IsPrivilegedUserlocalAPIEndpoint:advertiseAddress:"10.100.0.1"bindPort:6443certificateKey:"e6a2eb8581237ab72a4f494f30285ec12a9694d750b9785706a83bfcbbbd2204"---apiVersion:kubeadm.k8s.io/v1beta2kind:ClusterConfigurationetcd:# one of local or externallocal:imageRepository:"registry.k8s.io"imageTag:"3.2.24"dataDir:"/var/lib/etcd"extraArgs:listen-client-urls:"http://10.100.0.1:2379"serverCertSANs:- "ec2-10-100-0-1.compute-1.amazonaws.com"peerCertSANs:- "10.100.0.1"# external:# endpoints:# - "10.100.0.1:2379"# - "10.100.0.2:2379"# caFile: "/etcd/kubernetes/pki/etcd/etcd-ca.crt"# certFile: "/etcd/kubernetes/pki/etcd/etcd.crt"# keyFile: "/etcd/kubernetes/pki/etcd/etcd.key"networking:serviceSubnet:"10.96.0.0/16"podSubnet:"10.244.0.0/24"dnsDomain:"cluster.local"kubernetesVersion:"v1.12.0"controlPlaneEndpoint:"10.100.0.1:6443"apiServer:extraArgs:authorization-mode:"Node,RBAC"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:FilecertSANs:- "10.100.1.1"- "ec2-10-100-0-1.compute-1.amazonaws.com"timeoutForControlPlane:4m0scontrollerManager:extraArgs:"node-cidr-mask-size": "20"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:Filescheduler:extraArgs:address:"10.100.0.1"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:FilecertificatesDir:"/etc/kubernetes/pki"imageRepository:"registry.k8s.io"useHyperKubeImage:falseclusterName:"example-cluster"---apiVersion:kubelet.config.k8s.io/v1beta1kind:KubeletConfiguration# kubelet specific options here---apiVersion:kubeproxy.config.k8s.io/v1alpha1kind:KubeProxyConfiguration# kube-proxy specific options here
Kubeadm join configuration types
When executing kubeadm join with the --config option, the JoinConfiguration type should be provided.
The JoinConfiguration type should be used to configure runtime settings, that in case of kubeadm join
are the discovery method used for accessing the cluster info and all the setting which are specific
to the node where kubeadm is executed, including:
NodeRegistration, that holds fields that relate to registering the new node to the cluster;
use it to customize the node name, the CRI socket to use or any other settings that should apply to this
node only (e.g. the node IP).
APIEndpoint, that represents the endpoint of the instance of the API server to be eventually deployed on this node.
networking holds configuration for the networking topology of the cluster.
kubernetesVersion[Required] string
kubernetesVersion is the target version of the control plane.
controlPlaneEndpoint[Required] string
controlPlaneEndpoint sets a stable IP address or DNS name for the control plane; it
can be a valid IP address or a RFC-1123 DNS subdomain, both with optional TCP port.
In case the controlPlaneEndpoint is not specified, the advertiseAddress + bindPort
are used; in case the controlPlaneEndpoint is specified but without a TCP port,
the bindPort is used.
Possible usages are:
In a cluster with more than one control plane instances, this field should be
assigned the address of the external load balancer in front of the
control plane instances.
In environments with enforced node recycling, the controlPlaneEndpoint
could be used for assigning a stable DNS to the control plane.
dns defines the options for the DNS add-on installed in the cluster.
certificatesDir[Required] string
certificatesDir specifies where to store or look for all required certificates.
imageRepository[Required] string
imageRepository sets the container registry to pull images from.
If empty, registry.k8s.io will be used by default; in case of kubernetes version is
a CI build (kubernetes version starts with ci/) gcr.io/k8s-staging-ci-images
is used as a default for control plane components and for kube-proxy, while
registry.k8s.io will be used for all the other images.
useHyperKubeImage[Required] bool
useHyperKubeImage controls if hyperkube should be used for Kubernetes components
instead of their respective separate images.
DEPRECATED: As hyperkube is itself deprecated, this fields is too. It will be
removed in future kubeadm config versions, kubeadm will print multiple warnings
when this set to true, and at some point it may become ignored.
featureGates[Required] map[string]bool
featureGates contains the feature gates enabled by the user.
clusterName[Required] string
The cluster name.
ClusterStatus
ClusterStatus contains the cluster status. The ClusterStatus will be stored in
the kubeadm-config ConfigMap in the cluster, and then updated by kubeadm when
additional control plane instance joins or leaves the cluster.
apiEndpoints currently available in the cluster, one for each control
plane/API server instance.
The key of the map is the IP of the host's default interface.
InitConfiguration
InitConfiguration contains a list of elements that is specific "kubeadm init"-only runtime
information.
bootstrapTokens is respected at kubeadm init time and describes a set of bootstrap tokens to create.
This information IS NOT uploaded to the kubeadm cluster ConfigMap, partly because of its sensitive nature.
localAPIEndpoint represents the endpoint of the API server instance that's deployed on this control plane node.
In HA setups, this differs from ClusterConfiguration.controlPlaneEndpoint in the sense that ControlPlaneEndpoint
is the global endpoint for the cluster, which then load-balances the requests to each individual API server. This
configuration object lets you customize what IP/DNS name and port the local API server advertises it's accessible
on. By default, kubeadm tries to auto-detect the IP of the default interface and use that, but in case that process
fails you may set the desired value here.
certificateKey[Required] string
certificateKey sets the key with which certificates and keys are encrypted prior to being uploaded in
a secret in the cluster during the uploadcerts init phase.
JoinConfiguration
JoinConfiguration contains elements describing a particular node.
nodeRegistration holds fields that relate to registering the new
control-plane node to the cluster
caCertPath[Required] string
caCertPath is the path to the SSL certificate authority used to
secure comunications between a node and the control-plane.
Defaults to "/etc/kubernetes/pki/ca.crt".
controlPlane defines the additional control plane instance to be deployed
on the joining node. If nil, no additional control plane instance will be deployed.
expires specifies the timestamp when this token expires. Defaults to being set
dynamically at runtime based on the ttl. expires and ttl are mutually exclusive.
usages[Required] []string
usages describes the ways in which this token can be used. Can by default be used
for establishing bidirectional trust, but that can be changed here.
groups[Required] []string
groups specifies the extra groups that this token will authenticate as when/if
used for authentication.
BootstrapTokenDiscovery is used to set the options for bootstrap token based discovery
Field
Description
token[Required] string
token is a token used to validate cluster information fetched from
the control-plane.
apiServerEndpoint[Required] string
apiServerEndpoint is an IP or domain name to the API server from which information
will be fetched.
caCertHashes[Required] []string
caCertHashes specifies a set of public key pins to verify when token-based discovery
is used. The root CA found during discovery must match one of these values.
Specifying an empty set disables root CA pinning, which can be unsafe.
Each hash is specified as ":", where the only currently supported type is "sha256".
This is a hex-encoded SHA-256 hash of the Subject Public Key Info (SPKI) object in
DER-encoded ASN.1. These hashes can be calculated using, for example, OpenSSL.
unsafeSkipCAVerification[Required] bool
unsafeSkipCAVerification allows token-based discovery without CA verification via
caCertHashes. This can weaken the security of kubeadm since other nodes can
impersonate the control-plane.
BootstrapTokenString is a token of the format abcdef.abcdef0123456789 that is used
for both validation of the practically of the API server from a joining node's point
of view and as an authentication method for the node in the bootstrap phase of
"kubeadm join". This token is and should be short-lived
ControlPlaneComponent holds settings common to control plane component of the cluster
Field
Description
extraArgs[Required] map[string]string
extraArgs is an extra set of flags to pass to a control plane component.
A key in this map is the flag name as it appears on the command line except
without leading dash(es).
file is used to specify a file or URL to a kubeconfig file from which to load
cluster information.
bootstrapToken and file are mutually exclusive.
tlsBootstrapToken[Required] string
tlsBootstrapToken is a token used for TLS bootstrapping.
If bootstrapToken is set, this field is defaulted to .bootstrapToken.token,
but can be overridden.
If file is set, this field must be set in case the KubeConfigFile does not
contain any other authentication information.
ImageMeta allows to customize the image used for components that are not
originated from the Kubernetes/Kubernetes release process
Field
Description
imageRepository[Required] string
imageRepository sets the container registry to pull images from.
If not set, the imageRepository defined in ClusterConfiguration will be used.
imageTag[Required] string
imageTag allows for specifying a tag for the image.
In case this value is set, kubeadm does not change automatically the
version of the above components during upgrades.
localAPIEndpoint represents the endpoint of the API server instance
to be deployed on this node.
certificateKey[Required] string
certificateKey is the key that is used for decryption of certificates after
they are downloaded from the secret upon joining a new control plane node.
The corresponding encryption key is in the InitConfiguration.
(Members of ImageMeta are embedded into this type.)
ImageMeta allows to customize the container used for etcd.
dataDir[Required] string
dataDir is the directory etcd will place its data.
Defaults to "/var/lib/etcd".
extraArgs[Required] map[string]string
extraArgs are extra arguments provided to the etcd binary when run
inside a static pod.
A key in this map is the flag name as it appears on the
command line except without leading dash(es).
serverCertSANs[Required] []string
serverCertSANs sets extra Subject Alternative Names (SANs) for the
etcd server signing certificate.
peerCertSANs[Required] []string
peerCertSANs sets extra Subject Alternative Names (SANs) for the
etcd peer signing certificate.
NodeRegistrationOptions holds fields that relate to registering a new control-plane
or node to the cluster, either via "kubeadm init" or "kubeadm join".
Field
Description
name[Required] string
name is the .Metadata.Name field of the Node API object that will be created
in this kubeadm init or kubeadm join operation.
This field is also used in the CommonName field of the kubelet's client certificate
to the API server.
Defaults to the hostname of the node if not provided.
criSocket[Required] string
criSocket is used to retrieve container runtime information. This information will
be annotated to the Node API object, for later re-use.
taints specifies the taints the Node API object should be registered with.
If this field is unset, i.e. nil, in the kubeadm init process it will be defaulted with
a control-plane taint for control-plane nodes. If you don't want to taint your control-plane
node, set this field to an empty list, i.e. taints: [], in the YAML file. This field is
solely used for Node registration.
kubeletExtraArgs[Required] map[string]string
kubeletExtraArgs passes through extra arguments to the kubelet. The arguments here are
passed to the kubelet command line via the environment file kubeadm writes at runtime for
the kubelet to source. This overrides the generic base-level configuration in the
'kubelet-config-1.X' ConfigMap.
Flags have higher priority when parsing. These values are local and specific to the node
kubeadm is executing on.
A key in this map is the flag name as it appears on the command line except without leading dash(es).
ignorePreflightErrors[Required] []string
ignorePreflightErrors provides a list of pre-flight errors to be ignored when the
current node is registered.
6.12.13 - kubeadm Configuration (v1beta3)
Overview
Package v1beta3 defines the v1beta3 version of the kubeadm configuration file format.
This version improves on the v1beta2 format by fixing some minor issues and adding a few new fields.
A list of changes since v1beta2:
The deprecated "ClusterConfiguration.useHyperKubeImage" field has been removed.
Kubeadm no longer supports the hyperkube image.
The "ClusterConfiguration.DNS.Type" field has been removed since CoreDNS is the only supported
DNS server type by kubeadm.
Include "datapolicy" tags on the fields that hold secrets.
This would result in the field values to be omitted when API structures are printed with klog.
Add "InitConfiguration.SkipPhases", "JoinConfiguration.SkipPhases" to allow skipping
a list of phases during kubeadm init/join command execution.
Add "InitConfiguration.NodeRegistration.ImagePullPolicy" and "JoinConfiguration.NodeRegistration.ImagePullPolicy"
to allow specifying the images pull policy during kubeadm "init" and "join".
The value must be one of "Always", "Never" or "IfNotPresent".
"IfNotPresent" is the default, which has been the existing behavior prior to this addition.
Add "InitConfiguration.Patches.Directory", "JoinConfiguration.Patches.Directory" to allow
the user to configure a directory from which to take patches for components deployed by kubeadm.
Move the BootstrapToken* API and related utilities out of the "kubeadm" API group to a new group
"bootstraptoken". The kubeadm API version v1beta3 no longer contains the BootstrapToken* structures.
Migration from old kubeadm config versions
kubeadm v1.15.x and newer can be used to migrate from v1beta1 to v1beta2.
kubeadm v1.22.x and newer no longer support v1beta1 and older APIs, but can be used to migrate v1beta2 to v1beta3.
Basics
The preferred way to configure kubeadm is to pass an YAML configuration file with the --config option. Some of the
configuration options defined in the kubeadm config file are also available as command line flags, but only
the most common/simple use case are supported with this approach.
A kubeadm config file could contain multiple configuration types separated using three dashes (---).
kubeadm supports the following configuration types:
The list of configuration types that must be included in a configuration file depends by the action you are
performing (init or join) and by the configuration options you are going to use (defaults or advanced
customization).
If some configuration types are not provided, or provided only partially, kubeadm will use default values; defaults
provided by kubeadm includes also enforcing consistency of values across components when required (e.g.
--cluster-cidr flag on controller manager and clusterCIDR on kube-proxy).
Users are always allowed to override default values, with the only exception of a small subset of setting with
relevance for security (e.g. enforce authorization-mode Node and RBAC on api server)
If the user provides a configuration types that is not expected for the action you are performing, kubeadm will
ignore those types and print a warning.
Kubeadm init configuration types
When executing kubeadm init with the --config option, the following configuration types could be used:
InitConfiguration, ClusterConfiguration, KubeProxyConfiguration, KubeletConfiguration, but only one
between InitConfiguration and ClusterConfiguration is mandatory.
The InitConfiguration type should be used to configure runtime settings, that in case of kubeadm init
are the configuration of the bootstrap token and all the setting which are specific to the node where
kubeadm is executed, including:
NodeRegistration, that holds fields that relate to registering the new node to the cluster;
use it to customize the node name, the CRI socket to use or any other settings that should apply to this
node only (e.g. the node ip).
LocalAPIEndpoint, that represents the endpoint of the instance of the API server to be deployed on this node;
use it e.g. to customize the API server advertise address.
The ClusterConfiguration type should be used to configure cluster-wide settings,
including settings for:
networking that holds configuration for the networking topology of the cluster; use it e.g. to customize
Pod subnet or services subnet.
etcd: use it e.g. to customize the local etcd or to configure the API server
for using an external etcd cluster.
kube-apiserver, kube-scheduler, kube-controller-manager configurations; use it to customize control-plane
components by adding customized setting or overriding kubeadm default settings.
The KubeProxyConfiguration type should be used to change the configuration passed to kube-proxy instances
deployed in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kube-proxy/ or
https://pkg.go.dev/k8s.io/kube-proxy/config/v1alpha1#KubeProxyConfiguration
for kube-proxy official documentation.
The KubeletConfiguration type should be used to change the configurations that will be passed to all kubelet instances
deployed in the cluster. If this object is not provided or provided only partially, kubeadm applies defaults.
See https://kubernetes.io/docs/reference/command-line-tools-reference/kubelet/ or
https://pkg.go.dev/k8s.io/kubelet/config/v1beta1#KubeletConfiguration
for kubelet official documentation.
Here is a fully populated example of a single YAML file containing multiple
configuration types to be used during a kubeadm init run.
apiVersion:kubeadm.k8s.io/v1beta3kind:InitConfigurationbootstrapTokens:- token:"9a08jv.c0izixklcxtmnze7"description:"kubeadm bootstrap token"ttl:"24h"- token:"783bde.3f89s0fje9f38fhf"description:"another bootstrap token"usages:- authentication- signinggroups:- system:bootstrappers:kubeadm:default-node-tokennodeRegistration:name:"ec2-10-100-0-1"criSocket:"/var/run/dockershim.sock"taints:- key:"kubeadmNode"value:"someValue"effect:"NoSchedule"kubeletExtraArgs:v:4ignorePreflightErrors:- IsPrivilegedUserimagePullPolicy:"IfNotPresent"localAPIEndpoint:advertiseAddress:"10.100.0.1"bindPort:6443certificateKey:"e6a2eb8581237ab72a4f494f30285ec12a9694d750b9785706a83bfcbbbd2204"skipPhases:- addon/kube-proxy---apiVersion:kubeadm.k8s.io/v1beta3kind:ClusterConfigurationetcd:# one of local or externallocal:imageRepository:"registry.k8s.io"imageTag:"3.2.24"dataDir:"/var/lib/etcd"extraArgs:listen-client-urls:"http://10.100.0.1:2379"serverCertSANs:- "ec2-10-100-0-1.compute-1.amazonaws.com"peerCertSANs:- "10.100.0.1"# external:# endpoints:# - "10.100.0.1:2379"# - "10.100.0.2:2379"# caFile: "/etcd/kubernetes/pki/etcd/etcd-ca.crt"# certFile: "/etcd/kubernetes/pki/etcd/etcd.crt"# keyFile: "/etcd/kubernetes/pki/etcd/etcd.key"networking:serviceSubnet:"10.96.0.0/16"podSubnet:"10.244.0.0/24"dnsDomain:"cluster.local"kubernetesVersion:"v1.21.0"controlPlaneEndpoint:"10.100.0.1:6443"apiServer:extraArgs:authorization-mode:"Node,RBAC"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:FilecertSANs:- "10.100.1.1"- "ec2-10-100-0-1.compute-1.amazonaws.com"timeoutForControlPlane:4m0scontrollerManager:extraArgs:"node-cidr-mask-size": "20"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:Filescheduler:extraArgs:address:"10.100.0.1"extraVolumes:- name:"some-volume"hostPath:"/etc/some-path"mountPath:"/etc/some-pod-path"readOnly:falsepathType:FilecertificatesDir:"/etc/kubernetes/pki"imageRepository:"registry.k8s.io"clusterName:"example-cluster"---apiVersion:kubelet.config.k8s.io/v1beta1kind:KubeletConfiguration# kubelet specific options here---apiVersion:kubeproxy.config.k8s.io/v1alpha1kind:KubeProxyConfiguration# kube-proxy specific options here
Kubeadm join configuration types
When executing kubeadm join with the --config option, the JoinConfiguration type should be provided.
The JoinConfiguration type should be used to configure runtime settings, that in case of kubeadm join
are the discovery method used for accessing the cluster info and all the setting which are specific
to the node where kubeadm is executed, including:
nodeRegistration, that holds fields that relate to registering the new node to the cluster;
use it to customize the node name, the CRI socket to use or any other settings that should apply to this
node only (e.g. the node ip).
apiEndpoint, that represents the endpoint of the instance of the API server to be eventually deployed on this node.
networking holds configuration for the networking topology of the cluster.
kubernetesVersion string
kubernetesVersion is the target version of the control plane.
controlPlaneEndpoint string
controlPlaneEndpoint sets a stable IP address or DNS name for the control plane.
It can be a valid IP address or a RFC-1123 DNS subdomain, both with optional TCP port.
In case the controlPlaneEndpoint is not specified, the advertiseAddress + bindPort
are used; in case the controlPlaneEndpoint is specified but without a TCP port,
the bindPort is used.
Possible usages are:
In a cluster with more than one control plane instances, this field should be
assigned the address of the external load balancer in front of the
control plane instances.
In environments with enforced node recycling, the controlPlaneEndpoint could
be used for assigning a stable DNS to the control plane.
dns defines the options for the DNS add-on installed in the cluster.
certificatesDir string
certificatesDir specifies where to store or look for all required certificates.
imageRepository string
imageRepository sets the container registry to pull images from.
If empty, registry.k8s.io will be used by default.
In case of kubernetes version is a CI build (kubernetes version starts with ci/)
gcr.io/k8s-staging-ci-images will be used as a default for control plane components
and for kube-proxy, while registry.k8s.io will be used for all the other images.
featureGates map[string]bool
featureGates contains the feature gates enabled by the user.
clusterName string
The cluster name.
InitConfiguration
InitConfiguration contains a list of elements that is specific "kubeadm init"-only runtime
information.
kubeadm init-only information. These fields are solely used the first time kubeadm init runs.
After that, the information in the fields IS NOT uploaded to the kubeadm-config ConfigMap
that is used by kubeadm upgrade for instance. These fields must be omitempty.
bootstrapTokens is respected at kubeadm init time and describes a set of Bootstrap Tokens to create.
This information IS NOT uploaded to the kubeadm cluster configmap, partly because of its sensitive nature
localAPIEndpoint represents the endpoint of the API server instance that's deployed on this
control plane node. In HA setups, this differs from ClusterConfiguration.controlPlaneEndpoint
in the sense that controlPlaneEndpoint is the global endpoint for the cluster, which then
load-balances the requests to each individual API server.
This configuration object lets you customize what IP/DNS name and port the local API server
advertises it's accessible on. By default, kubeadm tries to auto-detect the IP of the default
interface and use that, but in case that process fails you may set the desired value here.
certificateKey string
certificateKey sets the key with which certificates and keys are encrypted prior to being
uploaded in a Secret in the cluster during the uploadcerts init phase.
skipPhases []string
skipPhases is a list of phases to skip during command execution.
The list of phases can be obtained with the kubeadm init --help command.
The flag "--skip-phases" takes precedence over this field.
nodeRegistration holds fields that relate to registering the new
control-plane node to the cluster.
caCertPath string
caCertPath is the path to the SSL certificate authority used to secure
comunications between a node and the control-plane.
Defaults to "/etc/kubernetes/pki/ca.crt".
controlPlane defines the additional control plane instance to be deployed
on the joining node. If nil, no additional control plane instance will be deployed.
skipPhases []string
skipPhases is a list of phases to skip during command execution.
The list of phases can be obtained with the kubeadm join --help command.
The flag --skip-phases takes precedence over this field.
BootstrapTokenDiscovery is used to set the options for bootstrap token based discovery
Field
Description
token[Required] string
token is a token used to validate cluster information fetched from the
control-plane.
apiServerEndpoint string
apiServerEndpoint is an IP or domain name to the API server from which
information will be fetched.
caCertHashes []string
caCertHashes specifies a set of public key pins to verify when token-based discovery
is used. The root CA found during discovery must match one of these values.
Specifying an empty set disables root CA pinning, which can be unsafe.
Each hash is specified as ":", where the only currently supported type is
"sha256". This is a hex-encoded SHA-256 hash of the Subject Public Key Info (SPKI)
object in DER-encoded ASN.1. These hashes can be calculated using, for example, OpenSSL.
unsafeSkipCAVerification bool
unsafeSkipCAVerification allows token-based discovery without CA verification
via caCertHashes. This can weaken the security of kubeadm since other nodes can
impersonate the control-plane.
ControlPlaneComponent holds settings common to control plane component of the cluster
Field
Description
extraArgs map[string]string
extraArgs is an extra set of flags to pass to the control plane component.
A key in this map is the flag name as it appears on the command line except
without leading dash(es).
file is used to specify a file or URL to a kubeconfig file from which to load
cluster information.
bootstrapToken and file are mutually exclusive.
tlsBootstrapToken string
tlsBootstrapToken is a token used for TLS bootstrapping.
If bootstrapToken is set, this field is defaulted to .bootstrapToken.token, but
can be overridden. If file is set, this field must be set in case the KubeConfigFile
does not contain any other authentication information
ImageMeta allows to customize the image used for components that are not
originated from the Kubernetes/Kubernetes release process
Field
Description
imageRepository string
imageRepository sets the container registry to pull images from.
If not set, the imageRepository defined in ClusterConfiguration will be used instead.
imageTag string
imageTag allows to specify a tag for the image.
In case this value is set, kubeadm does not change automatically the version of
the above components during upgrades.
localAPIEndpoint represents the endpoint of the API server instance to be
deployed on this node.
certificateKey string
certificateKey is the key that is used for decryption of certificates after
they are downloaded from the secret upon joining a new control plane node.
The corresponding encryption key is in the InitConfiguration.
(Members of ImageMeta are embedded into this type.)
ImageMeta allows to customize the container used for etcd.
dataDir[Required] string
dataDir is the directory etcd will place its data.
Defaults to "/var/lib/etcd".
extraArgs map[string]string
extraArgs are extra arguments provided to the etcd binary when run
inside a static Pod. A key in this map is the flag name as it appears on the
command line except without leading dash(es).
serverCertSANs []string
serverCertSANs sets extra Subject Alternative Names (SANs) for the etcd
server signing certificate.
peerCertSANs []string
peerCertSANs sets extra Subject Alternative Names (SANs) for the etcd peer
signing certificate.
NodeRegistrationOptions holds fields that relate to registering a new control-plane or
node to the cluster, either via "kubeadm init" or "kubeadm join"
Field
Description
name string
name is the .metadata.name field of the Node API object that will be created in this
kubeadm init or kubeadm join operation.
This field is also used in the CommonName field of the kubelet's client certificate to
the API server.
Defaults to the hostname of the node if not provided.
criSocket string
criSocket is used to retrieve container runtime info.
This information will be annotated to the Node API object, for later re-use
taints specifies the taints the Node API object should be registered with.
If this field is unset, i.e. nil, in the kubeadm init process it will be defaulted
with a control-plane taint for control-plane nodes.
If you don't want to taint your control-plane node, set this field to an empty list,
i.e. taints: [] in the YAML file. This field is solely used for Node registration.
kubeletExtraArgs map[string]string
kubeletExtraArgs passes through extra arguments to the kubelet.
The arguments here are passed to the kubelet command line via the environment file
kubeadm writes at runtime for the kubelet to source.
This overrides the generic base-level configuration in the 'kubelet-config-1.X' ConfigMap.
Flags have higher priority when parsing. These values are local and specific to the node
kubeadm is executing on. A key in this map is the flag name as it appears on the
command line except without leading dash(es).
ignorePreflightErrors []string
ignorePreflightErrors provides a list of pre-flight errors to be ignored when
the current node is registered.
imagePullPolicy specifies the policy for image pulling during kubeadm "init" and
"join" operations.
The value of this field must be one of "Always", "IfNotPresent" or "Never".
If this field is unset kubeadm will default it to "IfNotPresent", or pull the required
images if not present on the host.
Patches contains options related to applying patches to components deployed by kubeadm.
Field
Description
directory string
directory is a path to a directory that contains files named
"target[suffix][+patchtype].extension".
For example, "kube-apiserver0+merge.yaml" or just "etcd.json". "target" can be one of
"kube-apiserver", "kube-controller-manager", "kube-scheduler", "etcd". "patchtype" can
be one of "strategic" "merge" or "json" and they match the patch formats supported by
kubectl.
The default "patchtype" is "strategic". "extension" must be either "json" or "yaml".
"suffix" is an optional string that can be used to determine which patches are applied
first alpha-numerically.
expires specifies the timestamp when this token expires. Defaults to being set
dynamically at runtime based on the ttl. expires and ttl are mutually exclusive.
usages []string
usages describes the ways in which this token can be used. Can by default be used
for establishing bidirectional trust, but that can be changed here.
groups []string
groups specifies the extra groups that this token will authenticate as when/if
used for authentication
BootstrapTokenString is a token of the format abcdef.abcdef0123456789 that is used
for both validation of the practically of the API server from a joining node's point
of view and as an authentication method for the node in the bootstrap phase of
"kubeadm join". This token is and should be short-lived.
JSONOptions contains options for logging format "json".
Field
Description
splitStream[Required] bool
[Experimental] SplitStream redirects error messages to stderr while
info messages go to stdout, with buffering. The default is to write
both to stdout, without buffering.
[Experimental] InfoBufferSize sets the size of the info stream when
using split streams. The default is zero, which disables buffering.
VModuleConfiguration
(Alias of []k8s.io/component-base/config/v1alpha1.VModuleItem)
Appears in:
VModuleConfiguration is a collection of individual file names or patterns
and the corresponding verbosity threshold.
CredentialProviderConfig
CredentialProviderConfig is the configuration containing information about
each exec credential provider. Kubelet reads this configuration from disk and enables
each provider as specified by the CredentialProvider type.
providers is a list of credential provider plugins that will be enabled by the kubelet.
Multiple providers may match against a single image, in which case credentials
from all providers will be returned to the kubelet. If multiple providers are called
for a single image, the results are combined. If providers return overlapping
auth keys, the value from the provider earlier in this list is used.
CredentialProvider represents an exec plugin to be invoked by the kubelet. The plugin is only
invoked when an image being pulled matches the images handled by the plugin (see matchImages).
Field
Description
name[Required] string
name is the required name of the credential provider. It must match the name of the
provider executable as seen by the kubelet. The executable must be in the kubelet's
bin directory (set by the --image-credential-provider-bin-dir flag).
matchImages[Required] []string
matchImages is a required list of strings used to match against images in order to
determine if this provider should be invoked. If one of the strings matches the
requested image from the kubelet, the plugin will be invoked and given a chance
to provide credentials. Images are expected to contain the registry domain
and URL path.
Each entry in matchImages is a pattern which can optionally contain a port and a path.
Globs can be used in the domain, but not in the port or the path. Globs are supported
as subdomains like *.k8s.io or k8s.*.io, and top-level-domains such as k8s.*.
Matching partial subdomains like app*.k8s.io is also supported. Each glob can only match
a single subdomain segment, so *.io does not match *.k8s.io.
A match exists between an image and a matchImage when all of the below are true:
Both contain the same number of domain parts and each part matches.
The URL path of an imageMatch must be a prefix of the target image URL path.
If the imageMatch contains a port, then the port must match in the image as well.
defaultCacheDuration is the default duration the plugin will cache credentials in-memory
if a cache duration is not provided in the plugin response. This field is required.
apiVersion[Required] string
Required input version of the exec CredentialProviderRequest. The returned CredentialProviderResponse
MUST use the same encoding version as the input. Current supported values are:
credentialprovider.kubelet.k8s.io/v1alpha1
args []string
Arguments to pass to the command when executing it.
Env defines additional environment variables to expose to the process. These
are unioned with the host's environment, as well as variables client-go uses
to pass argument to the plugin.
CredentialProviderConfig is the configuration containing information about
each exec credential provider. Kubelet reads this configuration from disk and enables
each provider as specified by the CredentialProvider type.
providers is a list of credential provider plugins that will be enabled by the kubelet.
Multiple providers may match against a single image, in which case credentials
from all providers will be returned to the kubelet. If multiple providers are called
for a single image, the results are combined. If providers return overlapping
auth keys, the value from the provider earlier in this list is used.
KubeletConfiguration
KubeletConfiguration contains the configuration for the Kubelet
Field
Description
apiVersion string
kubelet.config.k8s.io/v1beta1
kind string
KubeletConfiguration
enableServer[Required] bool
enableServer enables Kubelet's secured server.
Note: Kubelet's insecure port is controlled by the readOnlyPort option.
Default: true
staticPodPath string
staticPodPath is the path to the directory containing local (static) pods to
run, or the path to a single static pod file.
Default: ""
httpCheckFrequency is the duration between checking http for new data.
Default: "20s"
staticPodURL string
staticPodURL is the URL for accessing static pods to run.
Default: ""
staticPodURLHeader map[string][]string
staticPodURLHeader is a map of slices with HTTP headers to use when accessing the podURL.
Default: nil
address string
address is the IP address for the Kubelet to serve on (set to 0.0.0.0
for all interfaces).
Default: "0.0.0.0"
port int32
port is the port for the Kubelet to serve on.
The port number must be between 1 and 65535, inclusive.
Default: 10250
readOnlyPort int32
readOnlyPort is the read-only port for the Kubelet to serve on with
no authentication/authorization.
The port number must be between 1 and 65535, inclusive.
Setting this field to 0 disables the read-only service.
Default: 0 (disabled)
tlsCertFile string
tlsCertFile is the file containing x509 Certificate for HTTPS. (CA cert,
if any, concatenated after server cert). If tlsCertFile and
tlsPrivateKeyFile are not provided, a self-signed certificate
and key are generated for the public address and saved to the directory
passed to the Kubelet's --cert-dir flag.
Default: ""
tlsPrivateKeyFile string
tlsPrivateKeyFile is the file containing x509 private key matching tlsCertFile.
Default: ""
tlsCipherSuites []string
tlsCipherSuites is the list of allowed cipher suites for the server.
Values are from tls package constants (https://golang.org/pkg/crypto/tls/#pkg-constants).
Default: nil
tlsMinVersion string
tlsMinVersion is the minimum TLS version supported.
Values are from tls package constants (https://golang.org/pkg/crypto/tls/#pkg-constants).
Default: ""
rotateCertificates bool
rotateCertificates enables client certificate rotation. The Kubelet will request a
new certificate from the certificates.k8s.io API. This requires an approver to approve the
certificate signing requests.
Default: false
serverTLSBootstrap bool
serverTLSBootstrap enables server certificate bootstrap. Instead of self
signing a serving certificate, the Kubelet will request a certificate from
the 'certificates.k8s.io' API. This requires an approver to approve the
certificate signing requests (CSR). The RotateKubeletServerCertificate feature
must be enabled when setting this field.
Default: false
authentication specifies how requests to the Kubelet's server are authenticated.
Defaults:
anonymous:
enabled: false
webhook:
enabled: true
cacheTTL: "2m"
authorization specifies how requests to the Kubelet's server are authorized.
Defaults:
mode: Webhook
webhook:
cacheAuthorizedTTL: "5m"
cacheUnauthorizedTTL: "30s"
registryPullQPS int32
registryPullQPS is the limit of registry pulls per second.
The value must not be a negative number.
Setting it to 0 means no limit.
Default: 5
registryBurst int32
registryBurst is the maximum size of bursty pulls, temporarily allows
pulls to burst to this number, while still not exceeding registryPullQPS.
The value must not be a negative number.
Only used if registryPullQPS is greater than 0.
Default: 10
eventRecordQPS int32
eventRecordQPS is the maximum event creations per second. If 0, there
is no limit enforced. The value cannot be a negative number.
Default: 5
eventBurst int32
eventBurst is the maximum size of a burst of event creations, temporarily
allows event creations to burst to this number, while still not exceeding
eventRecordQPS. This field canot be a negative number and it is only used
when eventRecordQPS > 0.
Default: 10
enableDebuggingHandlers bool
enableDebuggingHandlers enables server endpoints for log access
and local running of containers and commands, including the exec,
attach, logs, and portforward features.
Default: true
enableContentionProfiling bool
enableContentionProfiling enables lock contention profiling, if enableDebuggingHandlers is true.
Default: false
healthzPort int32
healthzPort is the port of the localhost healthz endpoint (set to 0 to disable).
A valid number is between 1 and 65535.
Default: 10248
healthzBindAddress string
healthzBindAddress is the IP address for the healthz server to serve on.
Default: "127.0.0.1"
oomScoreAdj int32
oomScoreAdj is The oom-score-adj value for kubelet process. Values
must be within the range [-1000, 1000].
Default: -999
clusterDomain string
clusterDomain is the DNS domain for this cluster. If set, kubelet will
configure all containers to search this domain in addition to the
host's search domains.
Default: ""
clusterDNS []string
clusterDNS is a list of IP addresses for the cluster DNS server. If set,
kubelet will configure all containers to use this for DNS resolution
instead of the host's DNS servers.
Default: nil
nodeStatusUpdateFrequency is the frequency that kubelet computes node
status. If node lease feature is not enabled, it is also the frequency that
kubelet posts node status to master.
Note: When node lease feature is not enabled, be cautious when changing the
constant, it must work with nodeMonitorGracePeriod in nodecontroller.
Default: "10s"
nodeStatusReportFrequency is the frequency that kubelet posts node
status to master if node status does not change. Kubelet will ignore this
frequency and post node status immediately if any change is detected. It is
only used when node lease feature is enabled. nodeStatusReportFrequency's
default value is 5m. But if nodeStatusUpdateFrequency is set explicitly,
nodeStatusReportFrequency's default value will be set to
nodeStatusUpdateFrequency for backward compatibility.
Default: "5m"
nodeLeaseDurationSeconds int32
nodeLeaseDurationSeconds is the duration the Kubelet will set on its corresponding Lease.
NodeLease provides an indicator of node health by having the Kubelet create and
periodically renew a lease, named after the node, in the kube-node-lease namespace.
If the lease expires, the node can be considered unhealthy.
The lease is currently renewed every 10s, per KEP-0009. In the future, the lease renewal
interval may be set based on the lease duration.
The field value must be greater than 0.
Default: 40
imageMinimumGCAge is the minimum age for an unused image before it is
garbage collected.
Default: "2m"
imageGCHighThresholdPercent int32
imageGCHighThresholdPercent is the percent of disk usage after which
image garbage collection is always run. The percent is calculated by
dividing this field value by 100, so this field must be between 0 and
100, inclusive. When specified, the value must be greater than
imageGCLowThresholdPercent.
Default: 85
imageGCLowThresholdPercent int32
imageGCLowThresholdPercent is the percent of disk usage before which
image garbage collection is never run. Lowest disk usage to garbage
collect to. The percent is calculated by dividing this field value by 100,
so the field value must be between 0 and 100, inclusive. When specified, the
value must be less than imageGCHighThresholdPercent.
Default: 80
volumeStatsAggPeriod is the frequency for calculating and caching volume
disk usage for all pods.
Default: "1m"
kubeletCgroups string
kubeletCgroups is the absolute name of cgroups to isolate the kubelet in
Default: ""
systemCgroups string
systemCgroups is absolute name of cgroups in which to place
all non-kernel processes that are not already in a container. Empty
for no container. Rolling back the flag requires a reboot.
The cgroupRoot must be specified if this field is not empty.
Default: ""
cgroupRoot string
cgroupRoot is the root cgroup to use for pods. This is handled by the
container runtime on a best effort basis.
cgroupsPerQOS bool
cgroupsPerQOS enable QoS based CGroup hierarchy: top level CGroups for QoS classes
and all Burstable and BestEffort Pods are brought up under their specific top level
QoS CGroup.
Default: true
cgroupDriver string
cgroupDriver is the driver kubelet uses to manipulate CGroups on the host (cgroupfs
or systemd).
Default: "cgroupfs"
cpuManagerPolicy string
cpuManagerPolicy is the name of the policy to use.
Requires the CPUManager feature gate to be enabled.
Default: "None"
cpuManagerPolicyOptions map[string]string
cpuManagerPolicyOptions is a set of key=value which allows to set extra options
to fine tune the behaviour of the cpu manager policies.
Requires both the "CPUManager" and "CPUManagerPolicyOptions" feature gates to be enabled.
Default: nil
cpuManagerReconcilePeriod is the reconciliation period for the CPU Manager.
Requires the CPUManager feature gate to be enabled.
Default: "10s"
memoryManagerPolicy string
memoryManagerPolicy is the name of the policy to use by memory manager.
Requires the MemoryManager feature gate to be enabled.
Default: "none"
topologyManagerPolicy string
topologyManagerPolicy is the name of the topology manager policy to use.
Valid values include:
restricted: kubelet only allows pods with optimal NUMA node alignment for
requested resources;
best-effort: kubelet will favor pods with NUMA alignment of CPU and device
resources;
none: kubelet has no knowledge of NUMA alignment of a pod's CPU and device resources.
single-numa-node: kubelet only allows pods with a single NUMA alignment
of CPU and device resources.
Policies other than "none" require the TopologyManager feature gate to be enabled.
Default: "none"
topologyManagerScope string
topologyManagerScope represents the scope of topology hint generation
that topology manager requests and hint providers generate. Valid values include:
container: topology policy is applied on a per-container basis.
pod: topology policy is applied on a per-pod basis.
"pod" scope requires the TopologyManager feature gate to be enabled.
Default: "container"
qosReserved map[string]string
qosReserved is a set of resource name to percentage pairs that specify
the minimum percentage of a resource reserved for exclusive use by the
guaranteed QoS tier.
Currently supported resources: "memory"
Requires the QOSReserved feature gate to be enabled.
Default: nil
runtimeRequestTimeout is the timeout for all runtime requests except long running
requests - pull, logs, exec and attach.
Default: "2m"
hairpinMode string
hairpinMode specifies how the Kubelet should configure the container
bridge for hairpin packets.
Setting this flag allows endpoints in a Service to loadbalance back to
themselves if they should try to access their own Service. Values:
"promiscuous-bridge": make the container bridge promiscuous.
"hairpin-veth": set the hairpin flag on container veth interfaces.
"none": do nothing.
Generally, one must set --hairpin-mode=hairpin-veth to achieve hairpin NAT,
because promiscuous-bridge assumes the existence of a container bridge named cbr0.
Default: "promiscuous-bridge"
maxPods int32
maxPods is the maximum number of Pods that can run on this Kubelet.
The value must be a non-negative integer.
Default: 110
podCIDR string
podCIDR is the CIDR to use for pod IP addresses, only used in standalone mode.
In cluster mode, this is obtained from the control plane.
Default: ""
podPidsLimit int64
podPidsLimit is the maximum number of PIDs in any pod.
Default: -1
resolvConf string
resolvConf is the resolver configuration file used as the basis
for the container DNS resolution configuration.
If set to the empty string, will override the default and effectively disable DNS lookups.
Default: "/etc/resolv.conf"
runOnce bool
runOnce causes the Kubelet to check the API server once for pods,
run those in addition to the pods specified by static pod files, and exit.
Default: false
cpuCFSQuota bool
cpuCFSQuota enables CPU CFS quota enforcement for containers that
specify CPU limits.
Default: true
cpuCFSQuotaPeriod is the CPU CFS quota period value, cpu.cfs_period_us.
The value must be between 1 us and 1 second, inclusive.
Requires the CustomCPUCFSQuotaPeriod feature gate to be enabled.
Default: "100ms"
nodeStatusMaxImages int32
nodeStatusMaxImages caps the number of images reported in Node.status.images.
The value must be greater than -2.
Note: If -1 is specified, no cap will be applied. If 0 is specified, no image is returned.
Default: 50
maxOpenFiles int64
maxOpenFiles is Number of files that can be opened by Kubelet process.
The value must be a non-negative number.
Default: 1000000
contentType string
contentType is contentType of requests sent to apiserver.
Default: "application/vnd.kubernetes.protobuf"
kubeAPIQPS int32
kubeAPIQPS is the QPS to use while talking with kubernetes apiserver.
Default: 5
kubeAPIBurst int32
kubeAPIBurst is the burst to allow while talking with kubernetes API server.
This field cannot be a negative number.
Default: 10
serializeImagePulls bool
serializeImagePulls when enabled, tells the Kubelet to pull images one
at a time. We recommend not changing the default value on nodes that
run docker daemon with version < 1.9 or an Aufs storage backend.
Issue #10959 has more details.
Default: true
evictionHard map[string]string
evictionHard is a map of signal names to quantities that defines hard eviction
thresholds. For example: {"memory.available": "300Mi"}.
To explicitly disable, pass a 0% or 100% threshold on an arbitrary resource.
Default:
memory.available: "100Mi"
nodefs.available: "10%"
nodefs.inodesFree: "5%"
imagefs.available: "15%"
evictionSoft map[string]string
evictionSoft is a map of signal names to quantities that defines soft eviction thresholds.
For example: {"memory.available": "300Mi"}.
Default: nil
evictionSoftGracePeriod map[string]string
evictionSoftGracePeriod is a map of signal names to quantities that defines grace
periods for each soft eviction signal. For example: {"memory.available": "30s"}.
Default: nil
evictionPressureTransitionPeriod is the duration for which the kubelet has to wait
before transitioning out of an eviction pressure condition.
Default: "5m"
evictionMaxPodGracePeriod int32
evictionMaxPodGracePeriod is the maximum allowed grace period (in seconds) to use
when terminating pods in response to a soft eviction threshold being met. This value
effectively caps the Pod's terminationGracePeriodSeconds value during soft evictions.
Note: Due to issue #64530, the behavior has a bug where this value currently just
overrides the grace period during soft eviction, which can increase the grace
period from what is set on the Pod. This bug will be fixed in a future release.
Default: 0
evictionMinimumReclaim map[string]string
evictionMinimumReclaim is a map of signal names to quantities that defines minimum reclaims,
which describe the minimum amount of a given resource the kubelet will reclaim when
performing a pod eviction while that resource is under pressure.
For example: {"imagefs.available": "2Gi"}.
Default: nil
podsPerCore int32
podsPerCore is the maximum number of pods per core. Cannot exceed maxPods.
The value must be a non-negative integer.
If 0, there is no limit on the number of Pods.
Default: 0
enableControllerAttachDetach bool
enableControllerAttachDetach enables the Attach/Detach controller to
manage attachment/detachment of volumes scheduled to this node, and
disables kubelet from executing any attach/detach operations.
Note: attaching/detaching CSI volumes is not supported by the kubelet,
so this option needs to be true for that use case.
Default: true
protectKernelDefaults bool
protectKernelDefaults, if true, causes the Kubelet to error if kernel
flags are not as it expects. Otherwise the Kubelet will attempt to modify
kernel flags to match its expectation.
Default: false
makeIPTablesUtilChains bool
makeIPTablesUtilChains, if true, causes the Kubelet ensures a set of iptables rules
are present on host.
These rules will serve as utility rules for various components, e.g. kube-proxy.
The rules will be created based on iptablesMasqueradeBit and iptablesDropBit.
Default: true
iptablesMasqueradeBit int32
iptablesMasqueradeBit is the bit of the iptables fwmark space to mark for SNAT.
Values must be within the range [0, 31]. Must be different from other mark bits.
Warning: Please match the value of the corresponding parameter in kube-proxy.
TODO: clean up IPTablesMasqueradeBit in kube-proxy.
Default: 14
iptablesDropBit int32
iptablesDropBit is the bit of the iptables fwmark space to mark for dropping packets.
Values must be within the range [0, 31]. Must be different from other mark bits.
Default: 15
featureGates map[string]bool
featureGates is a map of feature names to bools that enable or disable experimental
features. This field modifies piecemeal the built-in default values from
"k8s.io/kubernetes/pkg/features/kube_features.go".
Default: nil
failSwapOn bool
failSwapOn tells the Kubelet to fail to start if swap is enabled on the node.
Default: true
memorySwap configures swap memory available to container workloads.
containerLogMaxSize string
containerLogMaxSize is a quantity defining the maximum size of the container log
file before it is rotated. For example: "5Mi" or "256Ki".
Default: "10Mi"
containerLogMaxFiles int32
containerLogMaxFiles specifies the maximum number of container log files that can
be present for a container.
Default: 5
configMapAndSecretChangeDetectionStrategy is a mode in which ConfigMap and Secret
managers are running. Valid values include:
Get: kubelet fetches necessary objects directly from the API server;
Cache: kubelet uses TTL cache for object fetched from the API server;
Watch: kubelet uses watches to observe changes to objects that are in its interest.
Default: "Watch"
systemReserved map[string]string
systemReserved is a set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G)
pairs that describe resources reserved for non-kubernetes components.
Currently only cpu and memory are supported.
See http://kubernetes.io/docs/user-guide/compute-resources for more detail.
Default: nil
kubeReserved map[string]string
kubeReserved is a set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G) pairs
that describe resources reserved for kubernetes system components.
Currently cpu, memory and local storage for root file system are supported.
See https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/
for more details.
Default: nil
reservedSystemCPUs[Required] string
The reservedSystemCPUs option specifies the CPU list reserved for the host
level system threads and kubernetes related threads. This provide a "static"
CPU list rather than the "dynamic" list by systemReserved and kubeReserved.
This option does not support systemReservedCgroup or kubeReservedCgroup.
showHiddenMetricsForVersion string
showHiddenMetricsForVersion is the previous version for which you want to show
hidden metrics.
Only the previous minor version is meaningful, other values will not be allowed.
The format is <major>.<minor>, e.g.: 1.16.
The purpose of this format is make sure you have the opportunity to notice
if the next release hides additional metrics, rather than being surprised
when they are permanently removed in the release after that.
Default: ""
systemReservedCgroup string
systemReservedCgroup helps the kubelet identify absolute name of top level CGroup used
to enforce systemReserved compute resource reservation for OS system daemons.
Refer to Node Allocatable
doc for more information.
Default: ""
kubeReservedCgroup string
kubeReservedCgroup helps the kubelet identify absolute name of top level CGroup used
to enforce KubeReserved compute resource reservation for Kubernetes node system daemons.
Refer to Node Allocatable
doc for more information.
Default: ""
enforceNodeAllocatable []string
This flag specifies the various Node Allocatable enforcements that Kubelet needs to perform.
This flag accepts a list of options. Acceptable options are none, pods,
system-reserved and kube-reserved.
If none is specified, no other options may be specified.
When system-reserved is in the list, systemReservedCgroup must be specified.
When kube-reserved is in the list, kubeReservedCgroup must be specified.
This field is supported only when cgroupsPerQOS is set to true.
Refer to Node Allocatable
for more information.
Default: ["pods"]
allowedUnsafeSysctls []string
A comma separated whitelist of unsafe sysctls or sysctl patterns (ending in *).
Unsafe sysctl groups are kernel.shm*, kernel.msg*, kernel.sem, fs.mqueue.*,
and net.*. For example: "kernel.msg*,net.ipv4.route.min_pmtu"
Default: []
volumePluginDir string
volumePluginDir is the full path of the directory in which to search
for additional third party volume plugins.
Default: "/usr/libexec/kubernetes/kubelet-plugins/volume/exec/"
providerID string
providerID, if set, sets the unique ID of the instance that an external
provider (i.e. cloudprovider) can use to identify a specific node.
Default: ""
kernelMemcgNotification bool
kernelMemcgNotification, if set, instructs the the kubelet to integrate with the
kernel memcg notification for determining if memory eviction thresholds are
exceeded rather than polling.
Default: false
shutdownGracePeriod specifies the total duration that the node should delay the
shutdown and total grace period for pod termination during a node shutdown.
Default: "0s"
shutdownGracePeriodCriticalPods specifies the duration used to terminate critical
pods during a node shutdown. This should be less than shutdownGracePeriod.
For example, if shutdownGracePeriod=30s, and shutdownGracePeriodCriticalPods=10s,
during a node shutdown the first 20 seconds would be reserved for gracefully
terminating normal pods, and the last 10 seconds would be reserved for terminating
critical pods.
Default: "0s"
shutdownGracePeriodByPodPriority specifies the shutdown grace period for Pods based
on their associated priority class value.
When a shutdown request is received, the Kubelet will initiate shutdown on all pods
running on the node with a grace period that depends on the priority of the pod,
and then wait for all pods to exit.
Each entry in the array represents the graceful shutdown time a pod with a priority
class value that lies in the range of that value and the next higher entry in the
list when the node is shutting down.
For example, to allow critical pods 10s to shutdown, priority>=10000 pods 20s to
shutdown, and all remaining pods 30s to shutdown.
The time the Kubelet will wait before exiting will at most be the maximum of all
shutdownGracePeriodSeconds for each priority class range represented on the node.
When all pods have exited or reached their grace periods, the Kubelet will release
the shutdown inhibit lock.
Requires the GracefulNodeShutdown feature gate to be enabled.
This configuration must be empty if either ShutdownGracePeriod or ShutdownGracePeriodCriticalPods is set.
Default: nil
reservedMemory specifies a comma-separated list of memory reservations for NUMA nodes.
The parameter makes sense only in the context of the memory manager feature.
The memory manager will not allocate reserved memory for container workloads.
For example, if you have a NUMA0 with 10Gi of memory and the reservedMemory was
specified to reserve 1Gi of memory at NUMA0, the memory manager will assume that
only 9Gi is available for allocation.
You can specify a different amount of NUMA node and memory types.
You can omit this parameter at all, but you should be aware that the amount of
reserved memory from all NUMA nodes should be equal to the amount of memory specified
by the node allocatable.
If at least one node allocatable parameter has a non-zero value, you will need
to specify at least one NUMA node.
Also, avoid specifying:
Duplicates, the same NUMA node, and memory type, but with a different value.
zero limits for any memory type.
NUMAs nodes IDs that do not exist under the machine.
memory types except for memory and hugepages-
Default: nil
enableProfilingHandler bool
enableProfilingHandler enables profiling via web interface host:port/debug/pprof/
Default: true
enableDebugFlagsHandler bool
enableDebugFlagsHandler enables flags endpoint via web interface host:port/debug/flags/v
Default: true
seccompDefault bool
SeccompDefault enables the use of RuntimeDefault as the default seccomp profile for all workloads.
This requires the corresponding SeccompDefault feature gate to be enabled as well.
Default: false
memoryThrottlingFactor float64
MemoryThrottlingFactor specifies the factor multiplied by the memory limit or node allocatable memory
when setting the cgroupv2 memory.high value to enforce MemoryQoS.
Decreasing this factor will set lower high limit for container cgroups and put heavier reclaim pressure
while increasing will put less reclaim pressure.
See http://kep.k8s.io/2570 for more details.
Default: 0.8
registerWithTaints are an array of taints to add to a node object when
the kubelet registers itself. This only takes effect when registerNode
is true and upon the initial registration of the node.
Default: nil
registerNode bool
registerNode enables automatic registration with the apiserver.
Default: true
SerializedNodeConfigSource
SerializedNodeConfigSource allows us to serialize v1.NodeConfigSource.
This type is used internally by the Kubelet for tracking checkpointed dynamic configs.
It exists in the kubeletconfig API group because it is classified as a versioned input to the Kubelet.
CredentialProvider represents an exec plugin to be invoked by the kubelet. The plugin is only
invoked when an image being pulled matches the images handled by the plugin (see matchImages).
Field
Description
name[Required] string
name is the required name of the credential provider. It must match the name of the
provider executable as seen by the kubelet. The executable must be in the kubelet's
bin directory (set by the --image-credential-provider-bin-dir flag).
matchImages[Required] []string
matchImages is a required list of strings used to match against images in order to
determine if this provider should be invoked. If one of the strings matches the
requested image from the kubelet, the plugin will be invoked and given a chance
to provide credentials. Images are expected to contain the registry domain
and URL path.
Each entry in matchImages is a pattern which can optionally contain a port and a path.
Globs can be used in the domain, but not in the port or the path. Globs are supported
as subdomains like '.k8s.io' or 'k8s..io', and top-level-domains such as 'k8s.'.
Matching partial subdomains like 'app.k8s.io' is also supported. Each glob can only match
a single subdomain segment, so *.io does not match *.k8s.io.
A match exists between an image and a matchImage when all of the below are true:
Both contain the same number of domain parts and each part matches.
The URL path of an imageMatch must be a prefix of the target image URL path.
If the imageMatch contains a port, then the port must match in the image as well.
defaultCacheDuration is the default duration the plugin will cache credentials in-memory
if a cache duration is not provided in the plugin response. This field is required.
apiVersion[Required] string
Required input version of the exec CredentialProviderRequest. The returned CredentialProviderResponse
MUST use the same encoding version as the input. Current supported values are:
credentialprovider.kubelet.k8s.io/v1beta1
args []string
Arguments to pass to the command when executing it.
Env defines additional environment variables to expose to the process. These
are unioned with the host's environment, as well as variables client-go uses
to pass argument to the plugin.
enabled allows anonymous requests to the kubelet server.
Requests that are not rejected by another authentication method are treated as
anonymous requests.
Anonymous requests have a username of system:anonymous, and a group name of
system:unauthenticated.
mode is the authorization mode to apply to requests to the kubelet server.
Valid values are AlwaysAllow and Webhook.
Webhook mode uses the SubjectAccessReview API to determine authorization.
clientCAFile is the path to a PEM-encoded certificate bundle. If set, any request
presenting a client certificate signed by one of the authorities in the bundle
is authenticated with a username corresponding to the CommonName,
and groups corresponding to the Organization in the client certificate.
swapBehavior configures swap memory available to container workloads. May be one of
"", "LimitedSwap": workload combined memory and swap usage cannot exceed pod memory limit
"UnlimitedSwap": workloads can use unlimited swap, up to the allocatable limit.
JSONOptions contains options for logging format "json".
Field
Description
splitStream[Required] bool
[Experimental] SplitStream redirects error messages to stderr while
info messages go to stdout, with buffering. The default is to write
both to stdout, without buffering.
Maximum number of nanoseconds (i.e. 1s = 1000000000) between log
flushes. Ignored if the selected logging backend writes log
messages without buffering.
verbosity[Required] uint32
Verbosity is the threshold that determines which log messages are
logged. Default is zero which logs only the most important
messages. Higher values enable additional messages. Error messages
are always logged.
[Experimental] Options holds additional parameters that are specific
to the different logging formats. Only the options for the selected
format get used, but all of them get validated.
VModuleConfiguration
(Alias of []k8s.io/component-base/config/v1alpha1.VModuleItem)
CredentialProviderRequest includes the image that the kubelet requires authentication for.
Kubelet will pass this request object to the plugin via stdin. In general, plugins should
prefer responding with the same apiVersion they were sent.
Field
Description
apiVersion string
credentialprovider.kubelet.k8s.io/v1alpha1
kind string
CredentialProviderRequest
image[Required] string
image is the container image that is being pulled as part of the
credential provider plugin request. Plugins may optionally parse the image
to extract any information required to fetch credentials.
CredentialProviderResponse
CredentialProviderResponse holds credentials that the kubelet should use for the specified
image provided in the original request. Kubelet will read the response from the plugin via stdout.
This response should be set to the same apiVersion as CredentialProviderRequest.
cacheKeyType indiciates the type of caching key to use based on the image provided
in the request. There are three valid values for the cache key type: Image, Registry, and
Global. If an invalid value is specified, the response will NOT be used by the kubelet.
cacheDuration indicates the duration the provided credentials should be cached for.
The kubelet will use this field to set the in-memory cache duration for credentials
in the AuthConfig. If null, the kubelet will use defaultCacheDuration provided in
CredentialProviderConfig. If set to 0, the kubelet will not cache the provided AuthConfig.
auth is a map containing authentication information passed into the kubelet.
Each key is a match image string (more on this below). The corresponding authConfig value
should be valid for all images that match against this key. A plugin should set
this field to null if no valid credentials can be returned for the requested image.
Each key in the map is a pattern which can optionally contain a port and a path.
Globs can be used in the domain, but not in the port or the path. Globs are supported
as subdomains like '.k8s.io' or 'k8s..io', and top-level-domains such as 'k8s.'.
Matching partial subdomains like 'app.k8s.io' is also supported. Each glob can only match
a single subdomain segment, so *.io does not match *.k8s.io.
The kubelet will match images against the key when all of the below are true:
Both contain the same number of domain parts and each part matches.
The URL path of an imageMatch must be a prefix of the target image URL path.
If the imageMatch contains a port, then the port must match in the image as well.
When multiple keys are returned, the kubelet will traverse all keys in reverse order so that:
longer keys come before shorter keys with the same prefix
non-wildcard keys come before wildcard keys with the same prefix.
For any given match, the kubelet will attempt an image pull with the provided credentials,
stopping after the first successfully authenticated pull.
AuthConfig contains authentication information for a container registry.
Only username/password based authentication is supported today, but more authentication
mechanisms may be added in the future.
Field
Description
username[Required] string
username is the username used for authenticating to the container registry
An empty username is valid.
password[Required] string
password is the password used for authenticating to the container registry
An empty password is valid.
CredentialProviderRequest includes the image that the kubelet requires authentication for.
Kubelet will pass this request object to the plugin via stdin. In general, plugins should
prefer responding with the same apiVersion they were sent.
Field
Description
apiVersion string
credentialprovider.kubelet.k8s.io/v1beta1
kind string
CredentialProviderRequest
image[Required] string
image is the container image that is being pulled as part of the
credential provider plugin request. Plugins may optionally parse the image
to extract any information required to fetch credentials.
CredentialProviderResponse
CredentialProviderResponse holds credentials that the kubelet should use for the specified
image provided in the original request. Kubelet will read the response from the plugin via stdout.
This response should be set to the same apiVersion as CredentialProviderRequest.
cacheKeyType indiciates the type of caching key to use based on the image provided
in the request. There are three valid values for the cache key type: Image, Registry, and
Global. If an invalid value is specified, the response will NOT be used by the kubelet.
cacheDuration indicates the duration the provided credentials should be cached for.
The kubelet will use this field to set the in-memory cache duration for credentials
in the AuthConfig. If null, the kubelet will use defaultCacheDuration provided in
CredentialProviderConfig. If set to 0, the kubelet will not cache the provided AuthConfig.
auth is a map containing authentication information passed into the kubelet.
Each key is a match image string (more on this below). The corresponding authConfig value
should be valid for all images that match against this key. A plugin should set
this field to null if no valid credentials can be returned for the requested image.
Each key in the map is a pattern which can optionally contain a port and a path.
Globs can be used in the domain, but not in the port or the path. Globs are supported
as subdomains like '.k8s.io' or 'k8s..io', and top-level-domains such as 'k8s.'.
Matching partial subdomains like 'app.k8s.io' is also supported. Each glob can only match
a single subdomain segment, so *.io does not match *.k8s.io.
The kubelet will match images against the key when all of the below are true:
Both contain the same number of domain parts and each part matches.
The URL path of an imageMatch must be a prefix of the target image URL path.
If the imageMatch contains a port, then the port must match in the image as well.
When multiple keys are returned, the kubelet will traverse all keys in reverse order so that:
longer keys come before shorter keys with the same prefix
non-wildcard keys come before wildcard keys with the same prefix.
For any given match, the kubelet will attempt an image pull with the provided credentials,
stopping after the first successfully authenticated pull.
AuthConfig contains authentication information for a container registry.
Only username/password based authentication is supported today, but more authentication
mechanisms may be added in the future.
Field
Description
username[Required] string
username is the username used for authenticating to the container registry
An empty username is valid.
password[Required] string
password is the password used for authenticating to the container registry
An empty password is valid.
WebhookAdmission provides configuration for the webhook admission controller.
Field
Description
apiVersion string
apiserver.config.k8s.io/v1
kind string
WebhookAdmission
kubeConfigFile[Required] string
KubeConfigFile is the path to the kubeconfig file.
6.13 - Scheduling
6.13.1 - Scheduler Configuration
FEATURE STATE:Kubernetes v1.25 [stable]
You can customize the behavior of the kube-scheduler by writing a configuration
file and passing its path as a command line argument.
A scheduling Profile allows you to configure the different stages of scheduling
in the kube-scheduler.
Each stage is exposed in an extension point. Plugins provide scheduling behaviors
by implementing one or more of these extension points.
You can specify scheduling profiles by running kube-scheduler --config <filename>,
using the
KubeSchedulerConfiguration (v1beta3
or v1)
struct.
Note: KubeSchedulerConfiguration v1beta2
is deprecated in v1.25 and will be removed in v1.26. Please migrate KubeSchedulerConfiguration to
v1beta3 or v1
before upgrading Kubernetes to v1.25.
Profiles
A scheduling Profile allows you to configure the different stages of scheduling
in the kube-scheduler.
Each stage is exposed in an extension point.
Plugins provide scheduling behaviors by implementing one
or more of these extension points.
You can configure a single instance of kube-scheduler to run
multiple profiles.
Extension points
Scheduling happens in a series of stages that are exposed through the following
extension points:
queueSort: These plugins provide an ordering function that is used to
sort pending Pods in the scheduling queue. Exactly one queue sort plugin
may be enabled at a time.
preFilter: These plugins are used to pre-process or check information
about a Pod or the cluster before filtering. They can mark a pod as
unschedulable.
filter: These plugins are the equivalent of Predicates in a scheduling
Policy and are used to filter out nodes that can not run the Pod. Filters
are called in the configured order. A pod is marked as unschedulable if no
nodes pass all the filters.
postFilter: These plugins are called in their configured order when no
feasible nodes were found for the pod. If any postFilter plugin marks the
Pod schedulable, the remaining plugins are not called.
preScore: This is an informational extension point that can be used
for doing pre-scoring work.
score: These plugins provide a score to each node that has passed the
filtering phase. The scheduler will then select the node with the highest
weighted scores sum.
reserve: This is an informational extension point that notifies plugins
when resources have been reserved for a given Pod. Plugins also implement an
Unreserve call that gets called in the case of failure during or after
Reserve.
permit: These plugins can prevent or delay the binding of a Pod.
preBind: These plugins perform any work required before a Pod is bound.
bind: The plugins bind a Pod to a Node. bind plugins are called in order
and once one has done the binding, the remaining plugins are skipped. At
least one bind plugin is required.
postBind: This is an informational extension point that is called after
a Pod has been bound.
multiPoint: This is a config-only field that allows plugins to be enabled
or disabled for all of their applicable extension points simultaneously.
For each extension point, you could disable specific default plugins
or enable your own. For example:
You can use * as name in the disabled array to disable all default plugins
for that extension point. This can also be used to rearrange plugins order, if
desired.
Scheduling plugins
The following plugins, enabled by default, implement one or more of these
extension points:
ImageLocality: Favors nodes that already have the container images that the
Pod runs.
Extension points: score.
NodeUnschedulable: Filters out nodes that have .spec.unschedulable set to
true.
Extension points: filter.
NodeResourcesFit: Checks if the node has all the resources that the Pod is
requesting. The score can use one of three strategies: LeastAllocated
(default), MostAllocated and RequestedToCapacityRatio.
Extension points: preFilter, filter, score.
NodeResourcesBalancedAllocation: Favors nodes that would obtain a more
balanced resource usage if the Pod is scheduled there.
Extension points: score.
VolumeBinding: Checks if the node has or if it can bind the requested
volumes.
Extension points: preFilter, filter, reserve, preBind, score.
Note:score extension point is enabled when VolumeCapacityPriority feature is
enabled. It prioritizes the smallest PVs that can fit the requested volume
size.
VolumeRestrictions: Checks that volumes mounted in the node satisfy
restrictions that are specific to the volume provider.
Extension points: filter.
VolumeZone: Checks that volumes requested satisfy any zone requirements they
might have.
Extension points: filter.
NodeVolumeLimits: Checks that CSI volume limits can be satisfied for the
node.
Extension points: filter.
EBSLimits: Checks that AWS EBS volume limits can be satisfied for the node.
Extension points: filter.
GCEPDLimits: Checks that GCP-PD volume limits can be satisfied for the node.
Extension points: filter.
AzureDiskLimits: Checks that Azure disk volume limits can be satisfied for
the node.
Extension points: filter.
PrioritySort: Provides the default priority based sorting.
Extension points: queueSort.
DefaultBinder: Provides the default binding mechanism.
Extension points: bind.
DefaultPreemption: Provides the default preemption mechanism.
Extension points: postFilter.
You can also enable the following plugins, through the component config APIs,
that are not enabled by default:
CinderLimits: Checks that OpenStack Cinder
volume limits can be satisfied for the node.
Extension points: filter.
Multiple profiles
You can configure kube-scheduler to run more than one profile.
Each profile has an associated scheduler name and can have a different set of
plugins configured in its extension points.
With the following sample configuration, the scheduler will run with two
profiles: one with the default plugins and one with all scoring plugins
disabled.
Pods that want to be scheduled according to a specific profile can include
the corresponding scheduler name in its .spec.schedulerName.
By default, one profile with the scheduler name default-scheduler is created.
This profile includes the default plugins described above. When declaring more
than one profile, a unique scheduler name for each of them is required.
If a Pod doesn't specify a scheduler name, kube-apiserver will set it to
default-scheduler. Therefore, a profile with this scheduler name should exist
to get those pods scheduled.
Note: Pod's scheduling events have .spec.schedulerName as the ReportingController.
Events for leader election use the scheduler name of the first profile in the
list.
Note: All profiles must use the same plugin in the queueSort extension point and have
the same configuration parameters (if applicable). This is because the scheduler
only has one pending pods queue.
Plugins that apply to multiple extension points
Starting from kubescheduler.config.k8s.io/v1beta3, there is an additional field in the
profile config, multiPoint, which allows for easily enabling or disabling a plugin
across several extension points. The intent of multiPoint config is to simplify the
configuration needed for users and administrators when using custom profiles.
Consider a plugin, MyPlugin, which implements the preScore, score, preFilter,
and filter extension points. To enable MyPlugin for all its available extension
points, the profile config looks like:
One benefit of using multiPoint here is that if MyPlugin implements another
extension point in the future, the multiPoint config will automatically enable it
for the new extension.
Specific extension points can be excluded from MultiPoint expansion using
the disabled field for that extension point. This works with disabling default
plugins, non-default plugins, or with the wildcard ('*') to disable all plugins.
An example of this, disabling Score and PreScore, would be:
Starting from kubescheduler.config.k8s.io/v1beta3, all default plugins
are enabled internally through MultiPoint.
However, individual extension points are still available to allow flexible
reconfiguration of the default values (such as ordering and Score weights). For
example, consider two Score plugins DefaultScore1 and DefaultScore2, each with
a weight of 1. They can be reordered with different weights like so:
In this example, it's unnecessary to specify the plugins in MultiPoint explicitly
because they are default plugins. And the only plugin specified in Score is DefaultScore2.
This is because plugins set through specific extension points will always take precedence
over MultiPoint plugins. So, this snippet essentially re-orders the two plugins
without needing to specify both of them.
The general hierarchy for precedence when configuring MultiPoint plugins is as follows:
Specific extension points run first, and their settings override whatever is set elsewhere
Plugins manually configured through MultiPoint and their settings
Default plugins and their default settings
To demonstrate the above hierarchy, the following example is based on these plugins:
Plugin
Extension Points
DefaultQueueSort
QueueSort
CustomQueueSort
QueueSort
DefaultPlugin1
Score, Filter
DefaultPlugin2
Score
CustomPlugin1
Score, Filter
CustomPlugin2
Score, Filter
A valid sample configuration for these plugins would be:
Note that there is no error for re-declaring a MultiPoint plugin in a specific
extension point. The re-declaration is ignored (and logged), as specific extension points
take precedence.
Besides keeping most of the config in one spot, this sample does a few things:
Enables the custom queueSort plugin and disables the default one
Enables CustomPlugin1 and CustomPlugin2, which will run first for all of their extension points
Disables DefaultPlugin1, but only for filter
Reorders DefaultPlugin2 to run first in score (even before the custom plugins)
In versions of the config before v1beta3, without multiPoint, the above snippet would equate to this:
While this is a complicated example, it demonstrates the flexibility of MultiPoint config
as well as its seamless integration with the existing methods for configuring extension points.
With the v1beta2 configuration version, you can use a new score extension for the
NodeResourcesFit plugin.
The new extension combines the functionalities of the NodeResourcesLeastAllocated,
NodeResourcesMostAllocated and RequestedToCapacityRatio plugins.
For example, if you previously used the NodeResourcesMostAllocated plugin, you
would instead use NodeResourcesFit (enabled by default) and add a pluginConfig
with a scoreStrategy that is similar to:
In Kubernetes versions before v1.23, a scheduling policy can be used to specify the predicates and priorities process. For example, you can set a scheduling policy by
running kube-scheduler --policy-config-file <filename> or kube-scheduler --policy-configmap <ConfigMap>.
This scheduling policy is not supported since Kubernetes v1.23. Associated flags policy-config-file, policy-configmap, policy-configmap-namespace and use-legacy-policy-config are also not supported. Instead, use the Scheduler Configuration to achieve similar behavior.
Kubernetes contains several tools to help you work with the Kubernetes system.
crictl
crictl is a command-line
interface for inspecting and debugging CRI-compatible
container runtimes.
Dashboard
Dashboard, the web-based user interface of Kubernetes, allows you to deploy containerized applications
to a Kubernetes cluster, troubleshoot them, and manage the cluster and its
resources itself.
Helm
🛇 This item links to a third party project or product that is not part of Kubernetes itself. More information
Helm is a tool for managing packages of pre-configured
Kubernetes resources. These packages are known as Helm charts.
Use Helm to:
Find and use popular software packaged as Kubernetes charts
Share your own applications as Kubernetes charts
Create reproducible builds of your Kubernetes applications
Intelligently manage your Kubernetes manifest files
Manage releases of Helm packages
Kompose
Kompose is a tool to help Docker Compose users move to Kubernetes.
Use Kompose to:
Translate a Docker Compose file into Kubernetes objects
Go from local Docker development to managing your application via Kubernetes
Kui is a GUI tool that takes your normal
kubectl command line requests and responds with graphics.
Kui takes the normal kubectl command line requests and responds with graphics. Instead
of ASCII tables, Kui provides a GUI rendering with tables that you can sort.
Kui lets you:
Directly click on long, auto-generated resource names instead of copying and pasting
Type in kubectl commands and see them execute, even sometimes faster than kubectl itself
Query a Job and see its execution rendered
as a waterfall diagram
Click through resources in your cluster using a tabbed UI
Minikube
minikube is a tool that
runs a single-node Kubernetes cluster locally on your workstation for
development and testing purposes.
6.14.1 - Mapping from dockercli to crictl
Note:
This section links to third party projects that provide functionality required by Kubernetes. The Kubernetes project authors aren't responsible for these projects, which are listed alphabetically. To add a project to this list, read the content guide before submitting a change. More information.
Note: This page is deprecated and will be removed in Kubernetes 1.27.
crictl is a command-line interface for CRI-compatible container runtimes.
You can use it to inspect and debug container runtimes and applications on a
Kubernetes node. crictl and its source are hosted in the
cri-tools repository.
This page provides a reference for mapping common commands for the docker
command-line tool into the equivalent commands for crictl.
Mapping from docker CLI to crictl
The exact versions for the mapping table are for docker CLI v1.40 and crictl
v1.19.0. This list is not exhaustive. For example, it doesn't include
experimental docker CLI commands.
Note: The output format of crictl is similar to docker CLI, despite some missing
columns for some CLI. Make sure to check output for the specific command if your
command output is being parsed programmatically.
Retrieve debugging information
mapping from docker cli to crictl - retrieve debugging information
docker cli
crictl
Description
Unsupported Features
attach
attach
Attach to a running container
--detach-keys, --sig-proxy
exec
exec
Run a command in a running container
--privileged, --user, --detach-keys
images
images
List images
info
info
Display system-wide information
inspect
inspect, inspecti
Return low-level information on a container, image or task
logs
logs
Fetch the logs of a container
--details
ps
ps
List containers
stats
stats
Display a live stream of container(s) resource usage statistics
Column: NET/BLOCK I/O, PIDs
version
version
Show the runtime (Docker, ContainerD, or others) version information
Perform Changes
mapping from docker cli to crictl - perform changes
docker cli
crictl
Description
Unsupported Features
create
create
Create a new container
kill
stop (timeout = 0)
Kill one or more running container
--signal
pull
pull
Pull an image or a repository from a registry
--all-tags, --disable-content-trust
rm
rm
Remove one or more containers
rmi
rmi
Remove one or more images
run
run
Run a command in a new container
start
start
Start one or more stopped containers
--detach-keys
stop
stop
Stop one or more running containers
update
update
Update configuration of one or more containers
--restart, --blkio-weight and some other resource limit not supported by CRI.
Supported only in crictl
mapping from docker cli to crictl - supported only in crictl
crictl
Description
imagefsinfo
Return image filesystem info
inspectp
Display the status of one or more pods
port-forward
Forward local port to a pod
pods
List pods
runp
Run a new pod
rmp
Remove one or more pods
stopp
Stop one or more running pods
7 - Contribute to K8s docs
Kubernetes welcomes improvements from all contributors, new and experienced!
Manage and publish the documentation parts of the Kubernetes release cycle
Getting started
Anyone can open an issue about documentation, or contribute a change with a
pull request (PR) to the
kubernetes/website GitHub repository.
You need to be comfortable with
git and
GitHub
to work effectively in the Kubernetes community.
flowchart TB
subgraph third[Open PR]
direction TB
U[ ] -.-
Q[Improve content] --- N[Create content]
N --- O[Translate docs]
O --- P[Manage/publish docs parts of K8s release cycle]
end
subgraph second[Review]
direction TB
T[ ] -.-
D[Look over the K8s/website repository] --- E[Check out the Hugo static site generator]
E --- F[Understand basic GitHub commands]
F --- G[Review open PR and change review processes]
end
subgraph first[Sign up]
direction TB
S[ ] -.-
B[Sign the CNCF Contributor License Agreement] --- C[Join sig-docs Slack channel]
C --- V[Join kubernetes-sig-docs mailing list]
V --- M[Attend weekly sig-docs calls or slack meetings]
end
A([fa:fa-user New Contributor]) --> first
A --> second
A --> third
A --> H[Ask Questions!!!]
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,B,C,D,E,F,G,H,M,Q,N,O,P,V grey
class S,T,U spacewhite
class first,second,third white
Figure 1. Getting started for a new contributor.
Figure 1 outlines a roadmap for new contributors. You can follow some or all of the steps for Sign up and Review. Now you are ready to open PRs that achieve your contribution objectives with some listed under Open PR. Again, questions are always welcome!
Some tasks require more trust and more access in the Kubernetes organization.
See Participating in SIG Docs for more details about
roles and permissions.
Your first contribution
You can prepare for your first contribution by reviewing several steps beforehand. Figure 2 outlines the steps and the details follow.
flowchart LR
subgraph second[First Contribution]
direction TB
S[ ] -.-
G[Review PRs from other K8s members] -->
A[Check K8s/website issues list for good first PRs] --> B[Open a PR!!]
end
subgraph first[Suggested Prep]
direction TB
T[ ] -.-
D[Read contribution overview] -->E[Read K8s content and style guides]
E --> F[Learn about Hugo page content types and shortcodes]
end
first ----> second
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,B,D,E,F,G grey
class S,T spacewhite
class first,second white
Figure 2. Preparation for your first contribution.
Read the Contribution overview to
learn about the different ways you can contribute.
SIG Docs is the group of contributors who
publish and maintain Kubernetes documentation and the website. Getting
involved with SIG Docs is a great way for Kubernetes contributors (feature
development or otherwise) to have a large impact on the Kubernetes project.
Join the SIG Docs async Slack standup meeting on those weeks when the in-person Zoom video meeting does not take place. Meetings are always announced on #sig-docs. You can contribute to any one of the threads up to 24 hours after meeting announcement.
Other ways to contribute
Visit the Kubernetes community site. Participate on Twitter or Stack Overflow, learn about local Kubernetes meetups and events, and more.
If you notice an issue with Kubernetes documentation or have an idea for new content, then open an issue. All you need is a GitHub account and a web browser.
In most cases, new work on Kubernetes documentation begins with an issue in GitHub. Kubernetes contributors
then review, categorize and tag issues as needed. Next, you or another member
of the Kubernetes community open a pull request with changes to resolve the issue.
Opening an issue
If you want to suggest improvements to existing content or notice an error, then open an issue.
Click the Create an issue link on the right sidebar. This redirects you
to a GitHub issue page pre-populated with some headers.
Describe the issue or suggestion for improvement. Provide as many details as you can.
Click Submit new issue.
After submitting, check in on your issue occasionally or turn on GitHub notifications.
Reviewers and other community members might ask questions before
they can take action on your issue.
Suggesting new content
If you have an idea for new content, but you aren't sure where it should go, you can
still file an issue. Either:
Choose an existing page in the section you think the content belongs in and click Create an issue.
Provide a clear issue description. Describe what specifically is missing, out of date,
wrong, or needs improvement.
Explain the specific impact the issue has on users.
Limit the scope of a given issue to a reasonable unit of work. For problems
with a large scope, break them down into smaller issues. For example, "Fix the security docs"
is too broad, but "Add details to the 'Restricting network access' topic" is specific enough
to be actionable.
Search the existing issues to see if there's anything related or similar to the
new issue.
If the new issue relates to another issue or pull request, refer to it
either by its full URL or by the issue or pull request number prefixed
with a # character. For example, Introduced by #987654.
Follow the Code of Conduct. Respect your
fellow contributors. For example, "The docs are terrible" is not
helpful or polite feedback.
7.2 - Contributing new content
This section contains information you should know before contributing new
content.
flowchart LR
subgraph second[Before you begin]
direction TB
S[ ] -.-
A[Sign the CNCF CLA] --> B[Choose Git branch]
B --> C[One language per PR]
C --> F[Check out contributor tools]
end
subgraph first[Contributing Basics]
direction TB
T[ ] -.-
D[Write docs in markdown and build site with Hugo] --- E[source in GitHub]
E --- G['/content/../docs' folder contains docs for multiple languages]
G --- H[Review Hugo page content types and shortcodes]
end
first ----> second
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,B,C,D,E,F,G,H grey
class S,T spacewhite
class first,second white
Figure - Contributing new content preparation
The figure above depicts the information you should know
prior to submitting new content. The information details follow.
Contributing basics
Write Kubernetes documentation in Markdown and build the Kubernetes site
using Hugo.
Kubernetes documentation uses CommonMark as its flavor of Markdown.
The source is in GitHub. You can find
Kubernetes documentation at /content/en/docs/. Some of the reference
documentation is automatically generated from scripts in
the update-imported-docs/ directory.
Page content types describe the
presentation of documentation content in Hugo.
In addition to the standard Hugo shortcodes, we use a number of
custom Hugo shortcodes in our
documentation to control the presentation of content.
Documentation source is available in multiple languages in /content/. Each
language has its own folder with a two-letter code determined by the
ISO 639-1 standard
. For example, English documentation source is stored in /content/en/docs/.
For more information about contributing to documentation in multiple languages
or starting a new translation,
see localization.
Pull requests from contributors who haven't signed the CLA fail the automated
tests. The name and email you provide must match those found in
your git config, and your git name and email must match those used for the
CNCF CLA.
Choose which Git branch to use
When opening a pull request, you need to know in advance which branch to base
your work on.
Scenario
Branch
Existing or new English language content for the current release
main
Content for a feature change release
The branch which corresponds to the major and minor version the feature change is in, using the pattern dev-<version>. For example, if a feature changes in the v1.26 release, then add documentation changes to the dev-1.26 branch.
If you're still not sure which branch to choose, ask in #sig-docs on Slack.
Note: If you already submitted your pull request and you know that the
base branch was wrong, you (and only you, the submitter) can change it.
Languages per PR
Limit pull requests to one language per PR. If you need to make an identical
change to the same code sample in multiple languages, open a separate PR for
each language.
Tools for contributors
The doc contributors tools
directory in the kubernetes/website repository contains tools to help your
contribution journey go more smoothly.
7.2.1 - Opening a pull request
Note:Code developers: If you are documenting a new feature for an
upcoming Kubernetes release, see
Document a new feature.
To contribute new content pages or improve existing content pages, open a pull request (PR).
Make sure you follow all the requirements in the
Before you begin section.
If your change is small, or you're unfamiliar with git, read
Changes using GitHub to learn how to edit a page.
If your changes are large, read Work from a local fork to learn how to make
changes locally on your computer.
Changes using GitHub
If you're less experienced with git workflows, here's an easier method of
opening a pull request. Figure 1 outlines the steps and the details follow.
flowchart LR
A([fa:fa-user New Contributor]) --- id1[(K8s/Website GitHub)]
subgraph tasks[Changes using GitHub]
direction TB
0[ ] -.-
1[1. Edit this page] --> 2[2. Use GitHub markdown editor to make changes]
2 --> 3[3. fill in Propose file change]
end
subgraph tasks2[ ]
direction TB
4[4. select Propose file change] --> 5[5. select Create pull request] --> 6[6. fill in Open a pull request]
6 --> 7[7. select Create pull request]
end
id1 --> tasks --> tasks2
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:#fff;
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,1,2,3,4,5,6,7 grey
class 0 spacewhite
class tasks,tasks2 white
class id1 k8s
Figure 1. Steps for opening a PR using GitHub.
On the page where you see the issue, select the pencil icon at the top right.
You can also scroll to the bottom of the page and select Edit this page.
Make your changes in the GitHub markdown editor.
Below the editor, fill in the Propose file change form.
In the first field, give your commit message a title.
In the second field, provide a description.
Note: Do not use any GitHub Keywords
in your commit message. You can add those to the pull request description later.
Select Propose file change.
Select Create pull request.
The Open a pull request screen appears. Fill in the form:
The Subject field of the pull request defaults to the commit summary.
You can change it if needed.
The Body contains your extended commit message, if you have one,
and some template text. Add the
details the template text asks for, then delete the extra template text.
Leave the Allow edits from maintainers checkbox selected.
Note: PR descriptions are a great way to help reviewers understand your change.
For more information, see Opening a PR.
Select Create pull request.
Addressing feedback in GitHub
Before merging a pull request, Kubernetes community members review and
approve it. The k8s-ci-robot suggests reviewers based on the nearest
owner mentioned in the pages. If you have someone specific in mind,
leave a comment with their GitHub username in it.
If a reviewer asks you to make changes:
Go to the Files changed tab.
Select the pencil (edit) icon on any files changed by the pull request.
Make the changes requested.
Commit the changes.
If you are waiting on a reviewer, reach out once every 7 days. You can also post a message in the
#sig-docs Slack channel.
When your review is complete, a reviewer merges your PR and your changes go live a few minutes later.
Work from a local fork
If you're more experienced with git, or if your changes are larger than a few lines,
work from a local fork.
Make sure you have git installed
on your computer. You can also use a git UI application.
Figure 2 shows the steps to follow when you work from a local fork. The details for each step follow.
flowchart LR
1[Fork the K8s/website repository] --> 2[Create local clone and set upstream]
subgraph changes[Your changes]
direction TB
S[ ] -.-
3[Create a branch example: my_new_branch] --> 3a[Make changes using text editor] --> 4["Preview your changes locally using Hugo (localhost:1313) or build container image"]
end
subgraph changes2[Commit / Push]
direction TB
T[ ] -.-
5[Commit your changes] --> 6[Push commit to origin/my_new_branch]
end
2 --> changes --> changes2
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:#fff;
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class 1,2,3,3a,4,5,6 grey
class S,T spacewhite
class changes,changes2 white
Figure 2. Working from a local fork to make your changes.
Fetch commits from your fork's origin/main and kubernetes/website's upstream/main:
git fetch origin
git fetch upstream
This makes sure your local repository is up to date before you start making changes.
Note: This workflow is different than the
Kubernetes Community GitHub Workflow.
You do not need to merge your local copy of main with upstream/main before pushing updates
to your fork.
Create a branch
Decide which branch base to your work on:
For improvements to existing content, use upstream/main.
For new content about existing features, use upstream/main.
For new features in an upcoming Kubernetes release, use the feature branch. For more
information, see documenting for a release.
For long-running efforts that multiple SIG Docs contributors collaborate on,
like content reorganization, use a specific feature branch created for that effort.
If you need help choosing a branch, ask in the #sig-docs Slack channel.
Create a new branch based on the branch identified in step 1. This example assumes the base
branch is upstream/main:
git checkout -b <my_new_branch> upstream/main
Make your changes using a text editor.
At any time, use the git status command to see what files you've changed.
Commit your changes
When you are ready to submit a pull request, commit your changes.
In your local repository, check which files you need to commit:
git status
Output is similar to:
On branch <my_new_branch>
Your branch is up to date with 'origin/<my_new_branch>'.
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)
modified: content/en/docs/contribute/new-content/contributing-content.md
no changes added to commit (use "git add" and/or "git commit -a")
Add the files listed under Changes not staged for commit to the commit:
git add <your_file_name>
Repeat this for each file.
After adding all the files, create a commit:
git commit -m "Your commit message"
Note: Do not use any GitHub Keywords
in your commit message. You can add those to the pull request
description later.
Push your local branch and its new commit to your remote fork:
git push origin <my_new_branch>
Preview your changes locally
It's a good idea to preview your changes locally before pushing them or opening a pull request.
A preview lets you catch build errors or markdown formatting problems.
You can either build the website's container image or run Hugo locally. Building the container
image is slower but displays Hugo shortcodes, which can
be useful for debugging.
If you have not updated your website repository, the website/themes/docsy directory is empty.
The site cannot build without a local copy of the theme. To update the website theme, run:
git submodule update --init --recursive --depth 1
In a terminal, go to your Kubernetes website repository and start the Hugo server:
cd <path_to_your_repo>/website
hugo server --buildFuture
In a web browser, navigate to https://localhost:1313. Hugo watches the
changes and rebuilds the site as needed.
To stop the local Hugo instance, go back to the terminal and type Ctrl+C,
or close the terminal window.
Open a pull request from your fork to kubernetes/website
Figure 3 shows the steps to open a PR from your fork to the K8s/website. The details follow.
flowchart LR
subgraph first[ ]
direction TB
1[1. Go to K8s/website repository] --> 2[2. Select New Pull Request]
2 --> 3[3. Select compare across forks]
3 --> 4[4. Select your fork from head repository drop-down menu]
end
subgraph second [ ]
direction TB
5[5. Select your branch from the compare drop-down menu] --> 6[6. Select Create Pull Request]
6 --> 7[7. Add a description to your PR]
7 --> 8[8. Select Create pull request]
end
first --> second
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
class 1,2,3,4,5,6,7,8 grey
class first,second white
Figure 3. Steps to open a PR from your fork to the K8s/website.
From the head repository drop-down menu, select your fork.
From the compare drop-down menu, select your branch.
Select Create Pull Request.
Add a description for your pull request:
Title (50 characters or less): Summarize the intent of the change.
Description: Describe the change in more detail.
If there is a related GitHub issue, include Fixes #12345 or Closes #12345 in the
description. GitHub's automation closes the mentioned issue after merging the PR if used.
If there are other related PRs, link those as well.
If you want advice on something specific, include any questions you'd like reviewers to
think about in your description.
Select the Create pull request button.
Congratulations! Your pull request is available in Pull requests.
After opening a PR, GitHub runs automated tests and tries to deploy a preview using
Netlify.
If the Netlify build fails, select Details for more information.
If the Netlify build succeeds, select Details opens a staged version of the Kubernetes
website with your changes applied. This is how reviewers check your changes.
GitHub also automatically assigns labels to a PR, to help reviewers. You can add them too, if
needed. For more information, see Adding and removing issue labels.
Addressing feedback locally
After making your changes, amend your previous commit:
git commit -a --amend
-a: commits all changes
--amend: amends the previous commit, rather than creating a new one
Update your commit message if needed.
Use git push origin <my_new_branch> to push your changes and re-run the Netlify tests.
Note: If you use git commit -m instead of amending, you must squash your commits
before merging.
Changes from reviewers
Sometimes reviewers commit to your pull request. Before making any other changes, fetch those commits.
Fetch commits from your remote fork and rebase your working branch:
git fetch origin
git rebase origin/<your-branch-name>
After rebasing, force-push new changes to your fork:
If another contributor commits changes to the same file in another PR, it can create a merge
conflict. You must resolve all merge conflicts in your PR.
Update your fork and rebase your local branch:
git fetch origin
git rebase origin/<your-branch-name>
If your PR has multiple commits, you must squash them into a single commit before merging your PR.
You can check the number of commits on your PR's Commits tab or by running the git log
command locally.
Note: This topic assumes vim as the command line text editor.
Start an interactive rebase:
git rebase -i HEAD~<number_of_commits_in_branch>
Squashing commits is a form of rebasing. The -i switch tells git you want to rebase interactively.
HEAD~<number_of_commits_in_branch indicates how many commits to look at for the rebase.
Output is similar to:
pick d875112ca Original commit
pick 4fa167b80 Address feedback 1
pick 7d54e15ee Address feedback 2
# Rebase 3d18sf680..7d54e15ee onto 3d183f680 (3 commands)
...
# These lines can be re-ordered; they are executed from top to bottom.
The first section of the output lists the commits in the rebase. The second section lists the
options for each commit. Changing the word pick changes the status of the commit once the rebase
is complete.
For the purposes of rebasing, focus on squash and pick.
This squashes commits 4fa167b80 Address feedback 1 and 7d54e15ee Address feedback 2 into
d875112ca Original commit, leaving only d875112ca Original commit as a part of the timeline.
Save and exit your file.
Push your squashed commit:
git push --force-with-lease origin <branch_name>
Contribute to other repos
The Kubernetes project contains 50+ repositories. Many of these
repositories contain documentation: user-facing help text, error messages, API references or code
comments.
If you see text you'd like to improve, use GitHub to search all repositories in the Kubernetes
organization. This can help you figure out where to submit your issue or PR.
Each repository has its own processes and procedures. Before you file an issue or submit a PR,
read that repository's README.md, CONTRIBUTING.md, and code-of-conduct.md, if they exist.
Most repositories use issue and PR templates. Have a look through some open issues and PRs to get
a feel for that team's processes. Make sure to fill out the templates with as much detail as
possible when you file issues or PRs.
What's next
Read Reviewing to learn more about the review process.
7.2.2 - Documenting a feature for a release
Each major Kubernetes release introduces new features that require documentation. New releases also bring updates to existing features and documentation (such as upgrading a feature from alpha to beta).
Generally, the SIG responsible for a feature submits draft documentation of the
feature as a pull request to the appropriate development branch of the
kubernetes/website repository, and someone on the SIG Docs team provides
editorial feedback or edits the draft directly. This section covers the branching
conventions and process used during a release by both groups.
For documentation contributors
In general, documentation contributors don't write content from scratch for a release.
Instead, they work with the SIG creating a new feature to refine the draft documentation and make it release ready.
After you've chosen a feature to document or assist, ask about it in the #sig-docs
Slack channel, in a weekly SIG Docs meeting, or directly on the PR filed by the
feature SIG. If you're given the go-ahead, you can edit into the PR using one of
the techniques described in
Commit into another person's PR.
Find out about upcoming features
To find out about upcoming features, attend the weekly SIG Release meeting (see
the community page for upcoming meetings)
and monitor the release-specific documentation
in the kubernetes/sig-release
repository. Each release has a sub-directory in the /sig-release/tree/master/releases/
directory. The sub-directory contains a release schedule, a draft of the release
notes, and a document listing each person on the release team.
The release schedule contains links to all other documents, meetings,
meeting minutes, and milestones relating to the release. It also contains
information about the goals and timeline of the release, and any special
processes in place for this release. Near the bottom of the document, several
release-related terms are defined.
This document also contains a link to the Feature tracking sheet, which is
the official way to find out about all new features scheduled to go into the
release.
The release team document lists who is responsible for each release role. If
it's not clear who to talk to about a specific feature or question you have,
either attend the release meeting to ask your question, or contact the release
lead so that they can redirect you.
The release notes draft is a good place to find out about
specific features, changes, deprecations, and more about the release. The
content is not finalized until late in the release cycle, so use caution.
Feature tracking sheet
The feature tracking sheet for a given Kubernetes release
lists each feature that is planned for a release.
Each line item includes the name of the feature, a link to the feature's main
GitHub issue, its stability level (Alpha, Beta, or Stable), the SIG and
individual responsible for implementing it, whether it
needs docs, a draft release note for the feature, and whether it has been
merged. Keep the following in mind:
Beta and Stable features are generally a higher documentation priority than
Alpha features.
It's hard to test (and therefore to document) a feature that hasn't been merged,
or is at least considered feature-complete in its PR.
Determining whether a feature needs documentation is a manual process. Even if
a feature is not marked as needing docs, you may need to document the feature.
For developers or other SIG members
This section is information for members of other Kubernetes SIGs documenting new features
for a release.
If you are a member of a SIG developing a new feature for Kubernetes, you need
to work with SIG Docs to be sure your feature is documented in time for the
release. Check the
feature tracking spreadsheet
or check in the #sig-release Kubernetes Slack channel to verify scheduling details and
deadlines.
Open a placeholder PR
Open a draft pull request against the
dev-1.26 branch in the kubernetes/website repository, with a small
commit that you will amend later. To create a draft pull request, use the
Create Pull Request drop-down and select Create Draft Pull Request,
then click Draft Pull Request.
Leave a comment on the related kubernetes/enhancements
issue with a link to the PR to notify the docs person managing this release that
the feature docs are coming and should be tracked for the release.
If your feature does not need
any documentation changes, make sure the sig-release team knows this, by
mentioning it in the #sig-release Slack channel. If the feature does need
documentation but the PR is not created, the feature may be removed from the
milestone.
PR ready for review
When ready, populate your placeholder PR with feature documentation and change
the state of the PR from draft to ready for review. To mark a pull request
as ready for review, navigate to the merge box and click Ready for review.
Do your best to describe your feature and how to use it. If you need help structuring your documentation, ask in the #sig-docs slack channel.
When you complete your content, the documentation person assigned to your feature reviews it.
To ensure technical accuracy, the content may also require a technical review from corresponding SIG(s).
Use their suggestions to get the content to a release ready state.
If your feature is an Alpha or Beta feature and is behind a feature gate,
make sure you add it to Alpha/Beta Feature gates
table as part of your pull request. With new feature gates, a description of
the feature gate is also required. If your feature is GA'ed or deprecated,
make sure to move it from that table to Feature gates for graduated or deprecated features
table with Alpha and Beta history intact.
If your feature needs documentation and the first draft
content is not received, the feature may be removed from the milestone.
All PRs reviewed and ready to merge
If your PR has not yet been merged into the dev-1.26 branch by the release deadline, work with the
docs person managing the release to get it in by the deadline. If your feature needs
documentation and the docs are not ready, the feature may be removed from the
milestone.
7.2.3 - Submitting blog posts and case studies
Anyone can write a blog post and submit it for review.
Case studies require extensive review before they're approved.
The Kubernetes Blog
The Kubernetes blog is used by the project to communicate new features, community reports, and any
news that might be relevant to the Kubernetes community. This includes end users and developers.
Most of the blog's content is about things happening in the core project, but we encourage you to
submit about things happening elsewhere in the ecosystem too!
Anyone can write a blog post and submit it for review.
Submit a Post
Blog posts should not be commercial in nature and should consist of original content that applies
broadly to the Kubernetes community. Appropriate blog content includes:
New Kubernetes capabilities
Kubernetes projects updates
Updates from Special Interest Groups
Tutorials and walkthroughs
Thought leadership around Kubernetes
Kubernetes Partner OSS integration
Original content only
Unsuitable content includes:
Vendor product pitches
Partner updates without an integration and customer story
Have a look at the Markdown format for existing blog posts in the
website repository.
Write out your blog post in a text editor of your choice.
On the same link from step 2, click the Create new file button. Paste your content into the editor.
Name the file to match the proposed title of the blog post, but don’t put the date in the file name.
The blog reviewers will work with you on the final file name and the date the blog will be published.
When you save the file, GitHub will walk you through the pull request process.
A blog post reviewer will review your submission and work with you on feedback and final details.
When the blog post is approved, the blog will be scheduled for publication.
Guidelines and expectations
Blog posts should not be vendor pitches.
Articles must contain content that applies broadly to the Kubernetes community. For example, a
submission should focus on upstream Kubernetes as opposed to vendor-specific configurations.
Check the Documentation style guide for
what is typically allowed on Kubernetes properties.
Links should primarily be to the official Kubernetes documentation. When using external
references, links should be diverse - For example a submission shouldn't contain only links
back to a single company's blog.
Sometimes this is a delicate balance. The blog team
is there to give guidance on whether a post is appropriate for the Kubernetes blog, so don't
hesitate to reach out.
Blog posts are not published on specific dates.
Articles are reviewed by community volunteers. We'll try our best to accommodate specific
timing, but we make no guarantees.
Many core parts of the Kubernetes projects submit blog posts during release windows, delaying
publication times. Consider submitting during a quieter period of the release cycle.
If you are looking for greater coordination on post release dates, coordinating with
CNCF marketing is a more appropriate choice than submitting a blog post.
Sometimes reviews can get backed up. If you feel your review isn't getting the attention it needs,
you can reach out to the blog team via this slack channel
to ask in real time.
Blog posts should be relevant to Kubernetes users.
Topics related to participation in or results of Kubernetes SIGs activities are always on
topic (see the work in the Upstream Marketing Team
for support on these posts).
The components of Kubernetes are purposely modular, so tools that use existing integration
points like CNI and CSI are on topic.
Posts about other CNCF projects may or may not be on topic. We recommend asking the blog team
before submitting a draft.
Many CNCF projects have their own blog. These are often a better choice for posts. There are
times of major feature or milestone for a CNCF project that users would be interested in
reading on the Kubernetes blog.
The official blog is not for repurposing existing content from a third party as new content.
The license for the blog allows
commercial use of the content for commercial purposes, but not the other way around.
Blog posts should aim to be future proof
Given the development velocity of the project, we want evergreen content that won't require
updates to stay accurate for the reader.
It can be a better choice to add a tutorial or update official documentation than to write a
high level overview as a blog post.
Consider concentrating the long technical content as a call to action of the blog post, and
focus on the problem space or why readers should care.
Technical Considerations for submitting a blog post
Submissions need to be in Markdown format to be used by the Hugo generator
for the blog. There are many resources available on how to use
this technology stack.
We recognize that this requirement makes the process more difficult for less-familiar folks to
submit, and we're constantly looking at solutions to lower this bar. If you have ideas on how to
lower the barrier, please volunteer to help out.
The SIG Docs blog subproject
manages the review process for blog posts. For more information, see
Submit a post.
Ensure that your blog post follows the correct naming conventions and the following frontmatter
(metadata) information:
The Markdown file name must follow the format YYYY-MM-DD-Your-Title-Here.md. For example,
2020-02-07-Deploying-External-OpenStack-Cloud-Provider-With-Kubeadm.md.
Do not include dots in the filename. A name like 2020-01-01-whats-new-in-1.19.md causes
failures during a build.
The front matter must include the following:
---layout:blogtitle:"Your Title Here"date:YYYY-MM-DDslug:text-for-URL-link-here-no-spaces---
The first or initial commit message should be a short summary of the work being done and
should stand alone as a description of the blog post. Please note that subsequent edits to
your blog will be squashed into this main commit, so it should be as useful as possible.
Examples of a good commit message:
Add blog post on the foo kubernetes feature
blog: foobar announcement
Examples of bad commit message:
Add blog post
.
initial commit
draft post
The blog team will then review your PR and give you comments on things you might need to fix.
After that the bot will merge your PR and your blog post will be published.
If the content of the blog post contains only content that is not expected to require updates
to stay accurate for the reader, it can be marked as evergreen and exempted from the automatic
warning about outdated content added to blog posts older than one year.
To mark a blog post as evergreen, add this to the front matter:
evergreen:true
Examples of content that should not be marked evergreen:
Tutorials that only apply to specific releases or versions and not all future versions
References to pre-GA APIs or features
Submit a case study
Case studies highlight how organizations are using Kubernetes to solve real-world problems. The
Kubernetes marketing team and members of the CNCF
collaborate with you on all case studies.
Refer to the case study guidelines
and submit your request as outlined in the guidelines.
7.3 - Reviewing changes
This section describes how to review content.
7.3.1 - Reviewing pull requests
Anyone can review a documentation pull request. Visit the pull requests
section in the Kubernetes website repository to see open pull requests.
Reviewing documentation pull requests is a great way to introduce yourself to the Kubernetes
community. It helps you learn the code base and build trust with other contributors.
Comment on positive aspects of PRs as well as changes.
Be empathetic and mindful of how your review may be received.
Assume good intent and ask clarifying questions.
Experienced contributors, consider pairing with new contributors whose work requires extensive changes.
Review process
In general, review pull requests for content and style in English. Figure 1 outlines the steps for
the review process. The details for each step follow.
flowchart LR
subgraph fourth[Start review]
direction TB
S[ ] -.-
M[add comments] --> N[review changes]
N --> O[new contributors should choose Comment]
end
subgraph third[Select PR]
direction TB
T[ ] -.-
J[read description and comments]--> K[preview changes in Netlify preview build]
end
A[Review open PR list]--> B[Filter open PRs by label]
B --> third --> fourth
classDef grey fill:#dddddd,stroke:#ffffff,stroke-width:px,color:#000000, font-size:15px;
classDef white fill:#ffffff,stroke:#000,stroke-width:px,color:#000,font-weight:bold
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,B,J,K,M,N,O grey
class S,T spacewhite
class third,fourth white
Filter the open PRs using one or all of the following labels:
cncf-cla: yes (Recommended): PRs submitted by contributors who have not signed the CLA
cannot be merged. See Sign the CLA
for more information.
language/en (Recommended): Filters for english language PRs only.
size/<size>: filters for PRs of a certain size. If you're new, start with smaller PRs.
Additionally, ensure the PR isn't marked as a work in progress. PRs using the work in progress label are not ready for review yet.
Once you've selected a PR to review, understand the change by:
Reading the PR description to understand the changes made, and read any linked issues
Reading any comments by other reviewers
Clicking the Files changed tab to see the files and lines changed
Previewing the changes in the Netlify preview build by scrolling to the PR's build check
section at the bottom of the Conversation tab.
Here's a screenshot (this shows GitHub's desktop site; if you're reviewing
on a tablet or smartphone device, the GitHub web UI is slightly different):
To open the preview, click on the Details link of the deploy/netlify line in the
list of checks.
Go to the Files changed tab to start your review.
Click on the + symbol beside the line you want to comment on.
Fill in any comments you have about the line and click either Add single comment
(if you have only one comment to make) or Start a review (if you have multiple comments to make).
When finished, click Review changes at the top of the page. Here, you can add
a summary of your review (and leave some positive comments for the contributor!).
Please always use the "Comment"
Avoid clicking the "Request changes" button when finishing your review.
If you want to block a PR from being merged before some further changes are made,
you can leave a "/hold" comment.
Mention why you are setting a hold, and optionally specify the conditions under
which the hold can be removed by you or other reviewers.
Avoid clicking the "Approve" button when finishing your review.
Leaving a "/approve" comment is recommended most of the time.
Reviewing checklist
When reviewing, use the following as a starting point.
Language and grammar
Are there any obvious errors in language or grammar? Is there a better way to phrase something?
Are there any complicated or archaic words which could be replaced with a simpler word?
Are there any words, terms or phrases in use which could be replaced with a non-discriminatory alternative?
Does the word choice and its capitalization follow the style guide?
Are there long sentences which could be shorter or less complex?
Are there any long paragraphs which might work better as a list or table?
Content
Does similar content exist elsewhere on the Kubernetes site?
Does the content excessively link to off-site, individual vendor or non-open source documentation?
Website
Did this PR change or remove a page title, slug/alias or anchor link? If so, are there broken
links as a result of this PR? Is there another option, like changing the page title without
changing the slug?
Does the PR introduce a new page? If so:
Is the page using the right page content type
and associated Hugo shortcodes?
Does the page appear correctly in the section's side navigation (or at all)?
Do the changes show up in the Netlify preview? Be particularly vigilant about lists, code
blocks, tables, notes and images.
Other
For small issues with a PR, like typos or whitespace, prefix your comments with nit:.
This lets the author know the issue is non-critical.
7.3.2 - Reviewing for approvers and reviewers
SIG Docs Reviewers and
Approvers do a few extra things
when reviewing a change.
Every week a specific docs approver volunteers to triage
and review pull requests. This
person is the "PR Wrangler" for the week. See the
PR Wrangler scheduler for more information. To become a PR Wrangler, attend the weekly SIG Docs meeting and volunteer. Even if you are not on the schedule for the current week, you can still review pull
requests (PRs) that are not already under active review.
In addition to the rotation, a bot assigns reviewers and approvers
for the PR based on the owners for the affected files.
Everything described in Reviewing a pull request applies, but Reviewers and Approvers should also do the following:
Using the /assign Prow command to assign a specific reviewer to a PR as needed. This is extra important
when it comes to requesting technical review from code contributors.
Note: Look at the reviewers field in the front-matter at the top of a Markdown file to see who can
provide technical review.
Making sure the PR follows the Content and Style guides; link the author to the relevant part of the guide(s) if it doesn't.
Using the GitHub Request Changes option when applicable to suggest changes to the PR author.
Changing your review status in GitHub using the /approve or /lgtm Prow commands, if your suggestions are implemented.
Commit into another person's PR
Leaving PR comments is helpful, but there might be times when you need to commit
into another person's PR instead.
Do not "take over" for another person unless they explicitly ask
you to, or you want to resurrect a long-abandoned PR. While it may be faster
in the short term, it deprives the person of the chance to contribute.
The process you use depends on whether you need to edit a file that is already
in the scope of the PR, or a file that the PR has not yet touched.
You can't commit into someone else's PR if either of the following things is
true:
If the PR author pushed their branch directly to the
https://github.com/kubernetes/website/
repository. Only a reviewer with push access can commit to another user's PR.
Note: Encourage the author to push their branch to their fork before
opening the PR next time.
The PR author explicitly disallows edits from approvers.
Prow commands for reviewing
Prow is
the Kubernetes-based CI/CD system that runs jobs against pull requests (PRs). Prow
enables chatbot-style commands to handle GitHub actions across the Kubernetes
organization, like adding and removing labels, closing issues, and assigning an approver. Enter Prow commands as GitHub comments using the /<command-name> format.
The most common prow commands reviewers and approvers use are:
Prow commands for reviewing
Prow Command
Role Restrictions
Description
/lgtm
Organization members
Signals that you've finished reviewing a PR and are satisfied with the changes.
/approve
Approvers
Approves a PR for merging.
/assign
Reviewers or Approvers
Assigns a person to review or approve a PR
/close
Reviewers or Approvers
Closes an issue or PR.
/hold
Anyone
Adds the do-not-merge/hold label, indicating the PR cannot be automatically merged.
This GitHub Issue filter
finds issues that might need triage.
Triaging an issue
Validate the issue
Make sure the issue is about website documentation. Some issues can be closed quickly by
answering a question or pointing the reporter to a resource. See the
Support requests or code bug reports section for details.
Assess whether the issue has merit.
Add the triage/needs-information label if the issue doesn't have enough
detail to be actionable or the template is not filled out adequately.
Close the issue if it has both the lifecycle/stale and triage/needs-information labels.
Issues are generally opened and closed quickly.
However, sometimes an issue is inactive after its opened.
Other times, an issue may need to remain open for longer than 90 days.
Issue lifecycle labels
Label
Description
lifecycle/stale
After 90 days with no activity, an issue is automatically labeled as stale. The issue will be automatically closed if the lifecycle is not manually reverted using the /remove-lifecycle stale command.
lifecycle/frozen
An issue with this label will not become stale after 90 days of inactivity. A user manually adds this label to issues that need to remain open for much longer than 90 days, such as those with a priority/important-longterm label.
Handling special issue types
SIG Docs encounters the following types of issues often enough to document how
to handle them.
Duplicate issues
If a single problem has one or more issues open for it, combine them into a single issue.
You should decide which issue to keep open (or
open a new issue), then move over all relevant information and link related issues.
Finally, label all other issues that describe the same problem with
triage/duplicate and close them. Only having a single issue to work on reduces confusion
and avoids duplicate work on the same problem.
Dead link issues
If the dead link issue is in the API or kubectl documentation, assign them /priority critical-urgent until the problem is fully understood. Assign all other dead link issues /priority important-longterm, as they must be manually fixed.
Blog issues
We expect Kubernetes Blog entries to become
outdated over time. Therefore, we only maintain blog entries less than a year old.
If an issue is related to a blog entry that is more than one year old,
close the issue without fixing.
Support requests or code bug reports
Some docs issues are actually issues with the underlying code, or requests for
assistance when something, for example a tutorial, doesn't work.
For issues unrelated to docs, close the issue with the kind/support label and a comment
directing the requester to support venues (Slack, Stack Overflow) and, if
relevant, the repository to file an issue for bugs with features (kubernetes/kubernetes
is a great place to start).
Sample response to a request for support:
This issue sounds more like a request for support and less
like an issue specifically for docs. I encourage you to bring
your question to the `#kubernetes-users` channel in
[Kubernetes slack](https://slack.k8s.io/). You can also search
resources like
[Stack Overflow](https://stackoverflow.com/questions/tagged/kubernetes)
for answers to similar questions.
You can also open issues for Kubernetes functionality in
https://github.com/kubernetes/kubernetes.
If this is a documentation issue, please re-open this issue.
Sample code bug report response:
This sounds more like an issue with the code than an issue with
the documentation. Please open an issue at
https://github.com/kubernetes/kubernetes/issues.
If this is a documentation issue, please re-open this issue.
7.4 - Localizing Kubernetes documentation
This page shows you how to
localize the docs for a
different language.
Contribute to an existing localization
You can help add or improve content to an existing localization. In Kubernetes
Slack you'll find a channel for each localization. There is also a general
SIG Docs Localizations Slack channel
where you can say hello.
Note: If you want to work on a localization that already exists, check
this page in that localization (if it exists), rather than the
English original. You might see extra details there.
Find your two-letter language code
First, consult the ISO 639-1 standard
to find your localization's two-letter language code. For example, the two-letter code for Korean
is ko.
git clone https://github.com/<username>/website
cd website
The website content directory includes sub-directories for each language. The localization you
want to help out with is inside content/<two-letter-code>.
Suggest changes
Create or update your chosen localized page based on the English original. See
translating content for more details.
If you notice a technical inaccuracy or other problem with the upstream (English)
documentation, you should fix the upstream documentation first and then repeat the
equivalent fix by updating the localization you're working on.
Please limit pull requests to a single localization, since pull requests that change
content in multiple localizations could be difficult to review.
Follow Suggesting Content Improvements
to propose changes to that localization. The process is very similar to proposing changes to the
upstream (English) content.
Start a new localization
If you want the Kubernetes documentation localized into a new language, here's what
you need to do.
Because contributors can't approve their own pull requests, you need at least two contributors
to begin a localization.
All localization teams must be self-sustaining. The Kubernetes website is happy to host your work, but
it's up to you to translate it and keep existing localized content current.
You'll need to know the two-letter language code for your language. Consult the
ISO 639-1 standard to find your
localization's two-letter language code. For example, the two-letter code for Korean is
ko.
When you start a new localization, you must localize all the
minimum required content before
the Kubernetes project can publish your changes to the live
website.
SIG Docs can help you work on a separate branch so that you
can incrementally work towards that goal.
Find community
Let Kubernetes SIG Docs know you're interested in creating a localization! Join the
SIG Docs Slack channel and the
SIG Docs Localizations Slack channel.
Other localization teams are happy to help you get started and answer any questions you have.
Please also consider participating in the
SIG Docs Localization Subgroup meeting.
The mission of the SIG Docs localization subgroup is to work across the SIG Docs localization
teams to collaborate on defining and documenting the processes for creating localized contribution
guides. In addition, the SIG Docs localization subgroup will look for opportunities for the
creation and sharing of common tools across localization teams and also serve to identify new
requirements to the SIG Docs Leadership team. If you have questions about this meeting, please
inquire on the SIG Docs Localizations Slack channel.
You can also create a Slack channel for your localization in the kubernetes/community
repository. For an example of adding a Slack channel, see the PR for
adding a channel for Persian.
Join the Kubernetes GitHub organization
Once you've opened a localization PR, you can become members of the Kubernetes GitHub
organization. Each person on the team needs to create their own
Organization Membership Request
in the kubernetes/org repository.
Members of @kubernetes/sig-docs-**-owners can approve PRs that change content within (and only
within) your localization directory: /content/**/.
For each localization, The @kubernetes/sig-docs-**-reviews team automates review assignment for
new PRs.
Members of @kubernetes/website-maintainers can create new localization branches to coordinate translation efforts.
Members of @kubernetes/website-milestone-maintainers can use the /milestoneProw command to assign a milestone to issues or PRs.
Configure the workflow
Next, add a GitHub label for your localization in the kubernetes/test-infra repository. A label
lets you filter issues and pull requests for your specific language.
The Kubernetes website uses Hugo as its web framework. The website's Hugo configuration resides in
the config.toml file.
To support a new localization, you'll need to modify config.toml.
Add a configuration block for the new language to config.toml, under the existing [languages] block.
The German block, for example, looks like:
The value for languageName will be listed in language selection bar. Assign "language name in
native script and language (English language name in latin script)" to languageName.
For example, languageName = "한국어 (Korean)" or languageName = "Deutsch (German)".
languageNameLatinScript can be used to access the language name in latin script and use it in
the theme. Assign "language name in latin script" to languageNameLatinScript. For example,
languageNameLatinScript ="Korean" or languageNameLatinScript = "Deutsch".
When assigning a weight parameter for your block, find the language block with the highest
weight and add 1 to that value.
For more information about Hugo's multilingual support, see
"Multilingual Mode".
Add a new localization directory
Add a language-specific subdirectory to the
content folder in the repository.
For example, the two-letter code for German is de:
mkdir content/de
You also need to create a directory inside data/i18n for
localized strings; look at existing localizations
for an example. To use these new strings, you must also create a symbolic link
from i18n/<localization>.toml to the actual string configuration in
data/i18n/<localization>/<localization>.toml (remember to commit the symbolic link).
For example, for German the strings live in data/i18n/de/de.toml, and
i18n/de.toml is a symbolic link to data/i18n/de/de.toml.
Localize the community code of conduct
Open a PR against the cncf/foundation
repository to add the code of conduct in your language.
Setting up the OWNERS files
To set the roles of each user contributing to the localization, create an OWNERS file inside the
language-specific subdirectory with:
reviewers: A list of kubernetes teams with reviewer roles, in this case, the
sig-docs-**-reviews team created in
Add your localization team in GitHub.
approvers: A list of kubernetes teams with approvers roles, in this case, the
sig-docs-**-owners team created in
Add your localization team in GitHub.
labels: A list of GitHub labels to automatically apply to a PR, in this case, the language
label created in Configure the workflow.
More information about the OWNERS file can be found at go.k8s.io/owners.
# See the OWNERS docs at https://go.k8s.io/owners# This is the localization project for Spanish.# Teams and members are visible at https://github.com/orgs/kubernetes/teams.reviewers:- sig-docs-es-reviewsapprovers:- sig-docs-es-ownerslabels:- language/es
After adding the language-specific OWNERS file, update the root
OWNERS_ALIASES file with the new Kubernetes teams
for the localization, sig-docs-**-owners and sig-docs-**-reviews.
Next, open a pull request (PR) to add a
localization to the kubernetes/website repository.
The PR must include all of the minimum required content before it can
be approved.
For an example of adding a new localization, see the PR to enable
docs in French.
Add a localized README file
To guide other localization contributors, add a new
README-**.md to the top level of
kubernetes/website, where ** is the two-letter language code.
For example, a German README file would be README-de.md.
Provide guidance to localization contributors in the localized README-**.md file.
Include the same information contained in README.md as well as:
A point of contact for the localization project
Any information specific to the localization
After you create the localized README, add a link to the file from the main English README.md,
and include contact information in English. You can provide a GitHub ID, email address,
Slack channel, or other method of contact. You must also provide a link to your
localized Community Code of Conduct.
Launching your new localization
Once a localization meets requirements for workflow and minimum output, SIG Docs will:
Translated documents must reside in their own content/**/ subdirectory, but otherwise follow the
same URL path as the English source. For example, to prepare the
Kubernetes Basics tutorial for translation into German,
create a subfolder under the content/de/ folder and copy the English source:
Translation tools can speed up the translation process. For example, some editors offers plugins
to quickly translate text.
Caution: Machine-generated translation is insufficient on its own. Localization requires extensive human
review to meet minimum standards of quality.
To ensure accuracy in grammar and meaning, members of your localization team should carefully
review all machine-generated translations before publishing.
Source files
Localizations must be based on the English files from a specific release targeted by the localization team.
Each localization team can decide which release to target which is referred to as the target version below.
Revise the comments at the top of the file to suit your localization,
then translate the value of each string. For example, this is the German-language
placeholder text for the search form:
[ui_search_placeholder]
other = "Suchen"
Localizing site strings lets you customize site-wide text and features: for example, the legal
copyright text in the footer on each page.
Language specific style guide and glossary
Some language teams have their own language-specific style guide and glossary.
For example, see the Korean Localization Guide.
Language specific Zoom meetings
If the localization project needs a separate meeting time, contact a SIG Docs Co-Chair or Tech
Lead to create a new reoccurring Zoom meeting and calendar invite. This is only needed when
the team is large enough to sustain and require a separate meeting.
Per CNCF policy, the localization teams must upload their meetings to the SIG Docs YouTube
playlist. A SIG Docs Co-Chair or Tech Lead can help with the process until SIG Docs automates it.
Branching strategy
Because localization projects are highly collaborative efforts, we
encourage teams to work in shared localization branches - especially
when starting out and the localization is not yet live.
For example, an approver on a German localization team opens the localization branch
dev-1.12-de.1 directly against the kubernetes/website repository, based on the source branch for
Kubernetes v1.12.
Individual contributors open feature branches based on the localization branch.
For example, a German contributor opens a pull request with changes to
kubernetes:dev-1.12-de.1 from username:local-branch-name.
Approvers review and merge feature branches into the localization branch.
Periodically, an approver merges the localization branch to its source branch by opening and
approving a new pull request. Be sure to squash the commits before approving the pull request.
Repeat steps 1-4 as needed until the localization is complete. For example, subsequent German
localization branches would be: dev-1.12-de.2, dev-1.12-de.3, etc.
Teams must merge localized content into the same branch from which the content was sourced.
For example:
a localization branch sourced from main must be merged into main.
a localization branch sourced from release-1.24 must be merged into
release-1.24.
Note: If your localization branch was created from main branch but it is not merged into main before
new release branch release-1.25 created, merge it into both main and new release
branch release-1.25. To merge your localization branch into new release branch
release-1.25, you need to switch upstream branch of your localization branch to
release-1.25.
At the beginning of every team milestone, it's helpful to open an issue comparing upstream changes
between the previous localization branch and the current localization branch.
There are two scripts for comparing upstream changes.
upstream_changes.py
is useful for checking the changes made to a specific file. And
diff_l10n_branches.py
is useful for creating a list of outdated files for a specific localization branch.
While only approvers can open a new localization branch and merge pull requests, anyone can open a
pull request for a new localization branch. No special permissions are required.
For more information about working from forks or directly from the repository, see
"fork and clone the repo".
Upstream contributions
SIG Docs welcomes upstream contributions and corrections to the English source.
7.5 - Participating in SIG Docs
SIG Docs is one of the
special interest groups
within the Kubernetes project, focused on writing, updating, and maintaining
the documentation for Kubernetes as a whole. See
SIG Docs from the community github repo
for more information about the SIG.
SIG Docs welcomes content and reviews from all contributors. Anyone can open a
pull request (PR), and anyone is welcome to file issues about content or comment
on pull requests in progress.
You can also become a member,
reviewer, or
approver.
These roles require greater access and entail certain responsibilities for
approving and committing changes. See
community-membership
for more information on how membership works within the Kubernetes community.
The rest of this document outlines some unique ways these roles function within
SIG Docs, which is responsible for maintaining one of the most public-facing
aspects of Kubernetes -- the Kubernetes website and documentation.
SIG Docs chairperson
Each SIG, including SIG Docs, selects one or more SIG members to act as
chairpersons. These are points of contact between SIG Docs and other parts of
the Kubernetes organization. They require extensive knowledge of the structure
of the Kubernetes project as a whole and how SIG Docs works within it. See
Leadership
for the current list of chairpersons.
SIG Docs teams and automation
Automation in SIG Docs relies on two different mechanisms:
GitHub teams and OWNERS files.
GitHub teams
There are two categories of SIG Docs teams on GitHub:
@sig-docs-{language}-owners are approvers and leads
@sig-docs-{language}-reviews are reviewers
Each can be referenced with their @name in GitHub comments to communicate with
everyone in that group.
Sometimes Prow and GitHub teams overlap without matching exactly. For
assignment of issues, pull requests, and to support PR approvals, the
automation uses information from OWNERS files.
OWNERS files and front-matter
The Kubernetes project uses an automation tool called prow for automation
related to GitHub issues and pull requests. The
Kubernetes website repository uses
two prow plugins:
blunderbuss
approve
These two plugins use the
OWNERS and
OWNERS_ALIASES
files in the top level of the kubernetes/website GitHub repository to control
how prow works within the repository.
An OWNERS file contains a list of people who are SIG Docs reviewers and
approvers. OWNERS files can also exist in subdirectories, and can override who
can act as a reviewer or approver of files in that subdirectory and its
descendants. For more information about OWNERS files in general, see
OWNERS.
In addition, an individual Markdown file can list reviewers and approvers in its
front-matter, either by listing individual GitHub usernames or GitHub groups.
The combination of OWNERS files and front-matter in Markdown files determines
the advice PR owners get from automated systems about who to ask for technical
and editorial review of their PR.
How merging works
When a pull request is merged to the branch used to publish content, that content is published to http://kubernetes.io. To ensure that
the quality of our published content is high, we limit merging pull requests to
SIG Docs approvers. Here's how it works.
When a pull request has both the lgtm and approve labels, has no hold
labels, and all tests are passing, the pull request merges automatically.
Kubernetes organization members and SIG Docs approvers can add comments to
prevent automatic merging of a given pull request (by adding a /hold comment
or withholding a /lgtm comment).
Any Kubernetes member can add the lgtm label by adding a /lgtm comment.
Only SIG Docs approvers can merge a pull request
by adding an /approve comment. Some approvers also perform additional
specific roles, such as PR Wrangler or
SIG Docs chairperson.
What's next
For more information about contributing to the Kubernetes documentation, see:
Anyone can contribute to Kubernetes. As your contributions to SIG Docs grow,
you can apply for different levels of membership in the community.
These roles allow you to take on more responsibility within the community.
Each role requires more time and commitment. The roles are:
Anyone: regular contributors to the Kubernetes documentation
Members: can assign and triage issues and provide non-binding review on pull requests
Reviewers: can lead reviews on documentation pull requests and can vouch for a change's quality
Approvers: can lead reviews on documentation and merge changes
Anyone
Anyone with a GitHub account can contribute to Kubernetes. SIG Docs welcomes all new contributors!
Note: Don't send a direct email or Slack direct message to an individual
SIG Docs member. You must request sponsorship before submitting your application.
Open a GitHub issue in the
kubernetes/org repository. Use the
Organization Membership Request issue template.
Let your sponsors know about the GitHub issue. You can either:
Mention their GitHub username in an issue (@<GitHub-username>)
Send them the issue link using Slack or email.
Sponsors will approve your request with a +1 vote. Once your sponsors
approve the request, a Kubernetes GitHub admin adds you as a member.
Congratulations!
If your membership request is not accepted you will receive feedback.
After addressing the feedback, apply again.
Accept the invitation to the Kubernetes GitHub organization in your email account.
Note: GitHub sends the invitation to the default email address in your account.
Reviewers
Reviewers are responsible for reviewing open pull requests. Unlike member
feedback, you must address reviewer feedback. Reviewers are members of the
@kubernetes/sig-docs-{language}-reviews
GitHub team.
Note: To provide non-binding feedback, prefix your comments with a phrase like "Optionally: ".
Edit user-facing strings in code
Improve code comments
You can be a SIG Docs reviewer, or a reviewer for docs in a specific subject area.
Assigning reviewers to pull requests
Automation assigns reviewers to all pull requests. You can request a
review from a specific person by commenting: /assign [@_github_handle].
If the assigned reviewer has not commented on the PR, another reviewer can
step in. You can also assign technical reviewers as needed.
Using /lgtm
LGTM stands for "Looks good to me" and indicates that a pull request is
technically accurate and ready to merge. All PRs need a /lgtm comment from a
reviewer and a /approve comment from an approver to merge.
A /lgtm comment from reviewer is binding and triggers automation that adds the lgtm label.
Becoming a reviewer
When you meet the
requirements,
you can become a SIG Docs reviewer. Reviewers in other SIGs must apply
separately for reviewer status in SIG Docs.
To apply:
Open a pull request that adds your GitHub user name to a section of the
OWNERS_ALIASES file
in the kubernetes/website repository.
Note: If you aren't sure where to add yourself, add yourself to sig-docs-en-reviews.
Assign the PR to one or more SIG-Docs approvers (user names listed under
sig-docs-{language}-owners).
If approved, a SIG Docs lead adds you to the appropriate GitHub team. Once added,
K8s-ci-robot
assigns and suggests you as a reviewer on new pull requests.
Publish contributor content by approving and merging pull requests using the /approve comment
Propose improvements to the style guide
Propose improvements to docs tests
Propose improvements to the Kubernetes website or other tooling
If the PR already has a /lgtm, or if the approver also comments with
/lgtm, the PR merges automatically. A SIG Docs approver should only leave a
/lgtm on a change that doesn't need additional technical review.
Approving pull requests
Approvers and SIG Docs leads are the only ones who can merge pull requests
into the website repository. This comes with certain responsibilities.
Approvers can use the /approve command, which merges PRs into the repo.
Warning: A careless merge can break the site, so be sure that when you merge something, you mean it.
If you ever have a question, or you're not sure about something, feel free
to call for additional review.
Verify that Netlify tests pass before you /approve a PR.
Visit the Netlify page preview for a PR to make sure things look good before approving.
Participate in the
PR Wrangler rotation schedule
for weekly rotations. SIG Docs expects all approvers to participate in this
rotation. See PR wranglers.
for more details.
Becoming an approver
When you meet the
requirements,
you can become a SIG Docs approver. Approvers in other SIGs must apply
separately for approver status in SIG Docs.
To apply:
Open a pull request adding yourself to a section of the
OWNERS_ALIASES
file in the kubernetes/website repository.
Note: If you aren't sure where to add yourself, add yourself to sig-docs-en-owners.
Assign the PR to one or more current SIG Docs approvers.
If approved, a SIG Docs lead adds you to the appropriate GitHub team. Once added,
@k8s-ci-robot
assigns and suggests you as a reviewer on new pull requests.
What's next
Read about PR wrangling, a role all approvers take on rotation.
Use this script to remind contributors that haven't signed the CLA to do so.
Provide feedback on changes and ask for technical reviews from members of other SIGs.
Provide inline suggestions on the PR for the proposed content changes.
If you need to verify content, comment on the PR and request more details.
Assign relevant sig/ label(s).
If needed, assign reviewers from the reviewers: block in the file's front matter.
You can also tag a SIG for a review by commenting @kubernetes/<sig>-pr-reviews on the PR.
Use the /approve comment to approve a PR for merging. Merge the PR when ready.
PRs should have a /lgtm comment from another member before merging.
Consider accepting technically accurate content that doesn't meet the
style guidelines. As you approve the change,
open a new issue to address the style concern. You can usually write these style fix
issues as good first issues.
Using style fixups as good first issues is a good way to ensure a supply of easier tasks
to help onboard new contributors.
Helpful GitHub queries for wranglers
The following queries are helpful when wrangling.
After working through these queries, the remaining list of PRs to review is usually small.
These queries exclude localization PRs. All queries are against the main branch except the last one.
No CLA, not eligible to merge:
Remind the contributor to sign the CLA. If both the bot and a human have reminded them, close
the PR and remind them that they can open it after signing the CLA.
Do not review PRs whose authors have not signed the CLA!
Needs LGTM:
Lists PRs that need an LGTM from a member. If the PR needs technical review, loop in one of the reviewers suggested by the bot. If the content needs work, add suggestions and feedback in-line.
Quick Wins: Lists PRs against the main branch with no clear blockers. (change "XS" in the size label as you work through the PRs [XS, S, M, L, XL, XXL]).
Not against the primary branch: If the PR is against a dev- branch, it's for an upcoming release. Assign the docs release manager using: /assign @<manager's_github-username>. If the PR is against an old branch, help the author figure out whether it's targeted against the best branch.
Helpful Prow commands for wranglers
# add English label
/language en
# add squash label to PR if more than one commit
/label tide/merge-method-squash
# retitle a PR via Prow (such as a work-in-progress [WIP] or better detail of PR)
/retitle [WIP] <TITLE>
When to close Pull Requests
Reviews and approvals are one tool to keep our PR queue short and current. Another tool is closure.
Close PRs where:
The author hasn't signed the CLA for two weeks.
Authors can reopen the PR after signing the CLA. This is a low-risk way to make sure nothing gets merged without a signed CLA.
The author has not responded to comments or feedback in 2 or more weeks.
Don't be afraid to close pull requests. Contributors can easily reopen and resume works in progress. Often a closure notice is what spurs an author to resume and finish their contribution.
To close a pull request, leave a /close comment on the PR.
Note: The k8s-triage-robot bot marks issues as stale after 90 days of inactivity. After 30 more days it marks issues as rotten and closes them. PR wranglers should close issues after 14-30 days of inactivity.
PR Wrangler shadow program
In late 2021, SIG Docs introduced the PR Wrangler Shadow Program. The program was introduced to help new contributors understand the PR wrangling process.
Become a shadow
If you are interested in shadowing as a PR wrangler, please visit the PR Wranglers Wiki page to see the PR wrangling schedule for this year and sign up.
Kubernetes org members can edit the PR Wranglers Wiki page and sign up to shadow an existing PR Wrangler for a week.
Others can reach out on the #sig-docs Slack channel for requesting to shadow an assigned PR Wrangler for a specific week. Feel free to reach out to Brad Topol (@bradtopol) or one of the SIG Docs co-chairs/leads.
Once you've signed up to shadow a PR Wrangler, introduce yourself to the PR Wrangler on the Kubernetes Slack.
7.6 - Documentation style overview
The topics in this section provide guidance on writing style, content formatting
and organization, and using Hugo customizations specific to Kubernetes
documentation.
7.6.1 - Documentation Content Guide
This page contains guidelines for Kubernetes documentation.
If you have questions about what's allowed, join the #sig-docs channel in
Kubernetes Slack and ask!
For information on creating new content for the Kubernetes
docs, follow the style guide.
Overview
Source for the Kubernetes website, including the docs, resides in the
kubernetes/website repository.
Located in the kubernetes/website/content/<language_code>/docs folder, the
majority of Kubernetes documentation is specific to the Kubernetes
project.
What's allowed
Kubernetes docs allow content for third-party projects only when:
Content documents software in the Kubernetes project
Content documents software that's out of project but necessary for Kubernetes to function
Content is canonical on kubernetes.io, or links to canonical content elsewhere
Third party content
Kubernetes documentation includes applied examples of projects in the Kubernetes project—projects that live in the kubernetes and
kubernetes-sigs GitHub organizations.
Links to active content in the Kubernetes project are always allowed.
Kubernetes requires some third party content to function. Examples include container runtimes (containerd, CRI-O, Docker),
networking policy (CNI plugins), Ingress controllers, and logging.
Docs can link to third-party open source software (OSS) outside the Kubernetes project only if it's necessary for Kubernetes to function.
Dual sourced content
Wherever possible, Kubernetes docs link to canonical sources instead of hosting
dual-sourced content.
Dual-sourced content requires double the effort (or more!) to maintain
and grows stale more quickly.
Note: If you're a maintainer for a Kubernetes project and need help hosting your own docs,
ask for help in #sig-docs on Kubernetes Slack.
More information
If you have questions about allowed content, join the Kubernetes Slack #sig-docs channel and ask!
This page gives writing style guidelines for the Kubernetes documentation.
These are guidelines, not rules. Use your best judgment, and feel free to
propose changes to this document in a pull request.
For additional information on creating new content for the Kubernetes
documentation, read the Documentation Content Guide.
Changes to the style guide are made by SIG Docs as a group. To propose a change
or addition, add it to the agenda for an upcoming SIG Docs meeting, and attend the meeting to participate in the
discussion.
Note: Kubernetes documentation uses
Goldmark Markdown Renderer
with some adjustments along with a few
Hugo Shortcodes to support
glossary entries, tabs, and representing feature state.
Language
Kubernetes documentation has been translated into multiple languages
(see Localization READMEs).
The English-language documentation uses U.S. English spelling and grammar.
Documentation formatting standards
Use upper camel case for API objects
When you refer specifically to interacting with an API object, use UpperCamelCase, also known as Pascal case. You may see different capitalization, such as "configMap", in the API Reference. When writing general documentation, it's better to use upper camel case, calling it "ConfigMap" instead.
The following examples focus on capitalization. For more information about formatting API object names, review the related guidance on Code Style.
Do and Don't - Use Pascal case for API objects
Do
Don't
The HorizontalPodAutoscaler resource is responsible for ...
The Horizontal pod autoscaler is responsible for ...
A PodList object is a list of pods.
A Pod List object is a list of pods.
The Volume object contains a hostPath field.
The volume object contains a hostPath field.
Every ConfigMap object is part of a namespace.
Every configMap object is part of a namespace.
For managing confidential data, consider using the Secret API.
For managing confidential data, consider using the secret API.
Use angle brackets for placeholders
Use angle brackets for placeholders. Tell the reader what a placeholder
represents, for example:
Display information about a pod:
kubectl describe pod <pod-name> -n <namespace>
If the namespace of the pod is default, you can omit the '-n' parameter.
Use bold for user interface elements
Do and Don't - Bold interface elements
Do
Don't
Click Fork.
Click "Fork".
Select Other.
Select "Other".
Use italics to define or introduce new terms
Do and Don't - Use italics for new terms
Do
Don't
A cluster is a set of nodes ...
A "cluster" is a set of nodes ...
These components form the control plane.
These components form the control plane.
Use code style for filenames, directories, and paths
Do and Don't - Use code style for filenames, directories, and paths
Do
Don't
Open the envars.yaml file.
Open the envars.yaml file.
Go to the /docs/tutorials directory.
Go to the /docs/tutorials directory.
Open the /_data/concepts.yaml file.
Open the /_data/concepts.yaml file.
Use the international standard for punctuation inside quotes
Do and Don't - Use the international standard for punctuation inside quotes
Do
Don't
events are recorded with an associated "stage".
events are recorded with an associated "stage."
The copy is called a "fork".
The copy is called a "fork."
Inline code formatting
Use code style for inline code, commands, and API objects
For inline code in an HTML document, use the <code> tag. In a Markdown
document, use the backtick (`).
Do and Don't - Use code style for inline code, commands, and API objects
Do
Don't
The kubectl run command creates a Pod.
The "kubectl run" command creates a pod.
The kubelet on each node acquires a Lease…
The kubelet on each node acquires a lease…
A PersistentVolume represents durable storage…
A Persistent Volume represents durable storage…
For declarative management, use kubectl apply.
For declarative management, use "kubectl apply".
Enclose code samples with triple backticks. (```)
Enclose code samples with any other syntax.
Use single backticks to enclose inline code. For example, var example = true.
Use two asterisks (**) or an underscore (_) to enclose inline code. For example, var example = true.
Use triple backticks before and after a multi-line block of code for fenced code blocks.
Use multi-line blocks of code to create diagrams, flowcharts, or other illustrations.
Use meaningful variable names that have a context.
Use variable names such as 'foo','bar', and 'baz' that are not meaningful and lack context.
Remove trailing spaces in the code.
Add trailing spaces in the code, where these are important, because the screen reader will read out the spaces as well.
Note: The website supports syntax highlighting for code samples, but specifying a language is optional. Syntax highlighting in the code block should conform to the contrast guidelines.
Use code style for object field names and namespaces
Do and Don't - Use code style for object field names
Do
Don't
Set the value of the replicas field in the configuration file.
Set the value of the "replicas" field in the configuration file.
The value of the exec field is an ExecAction object.
The value of the "exec" field is an ExecAction object.
Run the process as a DaemonSet in the kube-system namespace.
Run the process as a DaemonSet in the kube-system namespace.
Use code style for Kubernetes command tool and component names
Do and Don't - Use code style for Kubernetes command tool and component names
Do
Don't
The kubelet preserves node stability.
The kubelet preserves node stability.
The kubectl handles locating and authenticating to the API server.
The kubectl handles locating and authenticating to the apiserver.
Run the process with the certificate, kube-apiserver --client-ca-file=FILENAME.
Run the process with the certificate, kube-apiserver --client-ca-file=FILENAME.
Starting a sentence with a component tool or component name
Do and Don't - Starting a sentence with a component tool or component name
Do
Don't
The kubeadm tool bootstraps and provisions machines in a cluster.
kubeadm tool bootstraps and provisions machines in a cluster.
The kube-scheduler is the default scheduler for Kubernetes.
kube-scheduler is the default scheduler for Kubernetes.
Use a general descriptor over a component name
Do and Don't - Use a general descriptor over a component name
Do
Don't
The Kubernetes API server offers an OpenAPI spec.
The apiserver offers an OpenAPI spec.
Aggregated APIs are subordinate API servers.
Aggregated APIs are subordinate APIServers.
Use normal style for string and integer field values
For field values of type string or integer, use normal style without quotation marks.
Do and Don't - Use normal style for string and integer field values
Do
Don't
Set the value of imagePullPolicy to Always.
Set the value of imagePullPolicy to "Always".
Set the value of image to nginx:1.16.
Set the value of image to nginx:1.16.
Set the value of the replicas field to 2.
Set the value of the replicas field to 2.
Referring to Kubernetes API resources
This section talks about how we reference API resources in the documentation.
Clarification about "resource"
Kubernetes uses the word "resource" to refer to API resources, such as pod, deployment, and so on. We also use "resource" to talk about CPU and memory requests and limits. Always refer to API resources as "API resources" to avoid confusion with CPU and memory resources.
When to use Kubernetes API terminologies
The different Kubernetes API terminologies are:
Resource type: the name used in the API URL (such as pods, namespaces)
Resource: a single instance of a resource type (such as pod, secret)
Object: a resource that serves as a "record of intent". An object is a desired state for a specific part of your cluster, which the Kubernetes control plane tries to maintain.
Always use "resource" or "object" when referring to an API resource in docs. For example, use "a Secret object" over just "a Secret".
API resource names
Always format API resource names using UpperCamelCase, also known as PascalCase, and code formatting.
For inline code in an HTML document, use the <code> tag. In a Markdown document, use the backtick (`).
Don't split an API object name into separate words. For example, use PodTemplateList, not Pod Template List.
For more information about Kubernetes API terminologies, please review the related guidance on Kubernetes API terminology.
Code snippet formatting
Don't include the command prompt
Do and Don't - Don't include the command prompt
Do
Don't
kubectl get pods
$ kubectl get pods
Separate commands from output
Verify that the pod is running on your chosen node:
kubectl get pods --output=wide
The output is similar to this:
NAME READY STATUS RESTARTS AGE IP NODE
nginx 1/1 Running 0 13s 10.200.0.4 worker0
Versioning Kubernetes examples
Code examples and configuration examples that include version information should be consistent with the accompanying text.
If the information is version specific, the Kubernetes version needs to be defined in the prerequisites section of the Task template or the Tutorial template. Once the page is saved, the prerequisites section is shown as Before you begin.
To specify the Kubernetes version for a task or tutorial page, include min-kubernetes-server-version in the front matter of the page.
If the example YAML is in a standalone file, find and review the topics that include it as a reference.
Verify that any topics using the standalone YAML have the appropriate version information defined.
If a stand-alone YAML file is not referenced from any topics, consider deleting it instead of updating it.
For example, if you are writing a tutorial that is relevant to Kubernetes version 1.8, the front-matter of your markdown file should look something like:
---title:<your tutorial title here>min-kubernetes-server-version:v1.8---
In code and configuration examples, do not include comments about alternative versions.
Be careful to not include incorrect statements in your examples as comments, such as:
apiVersion:v1# earlier versions use...kind:Pod...
Kubernetes.io word list
A list of Kubernetes-specific terms and words to be used consistently across the site.
Kubernetes.io word list
Term
Usage
Kubernetes
Kubernetes should always be capitalized.
Docker
Docker should always be capitalized.
SIG Docs
SIG Docs rather than SIG-DOCS or other variations.
On-premises
On-premises or On-prem rather than On-premise or other variations.
Shortcodes
Hugo Shortcodes help create different rhetorical appeal levels. Our documentation supports three different shortcodes in this category: Note{{< note >}}, Caution{{< caution >}}, and Warning{{< warning >}}.
Surround the text with an opening and closing shortcode.
Use the following syntax to apply a style:
{{< note >}}
No need to include a prefix; the shortcode automatically provides one. (Note:, Caution:, etc.)
{{< /note >}}
The output is:
Note: The prefix you choose is the same text for the tag.
Note
Use {{< note >}} to highlight a tip or a piece of information that may be helpful to know.
For example:
{{< note >}}
You can _still_ use Markdown inside these callouts.
{{< /note >}}
The output is:
Note: You can still use Markdown inside these callouts.
You can use a {{< note >}} in a list:
1. Use the note shortcode in a list
1. A second item with an embedded note
{{< note >}}
Warning, Caution, and Note shortcodes, embedded in lists, need to be indented four spaces. See [Common Shortcode Issues](#common-shortcode-issues).
{{< /note >}}
1. A third item in a list
1. A fourth item in a list
The output is:
Use the note shortcode in a list
A second item with an embedded note
Note: Warning, Caution, and Note shortcodes, embedded in lists, need to be indented four spaces. See Common Shortcode Issues.
A third item in a list
A fourth item in a list
Caution
Use {{< caution >}} to call attention to an important piece of information to avoid pitfalls.
For example:
{{< caution >}}
The callout style only applies to the line directly above the tag.
{{< /caution >}}
The output is:
Caution: The callout style only applies to the line directly above the tag.
Warning
Use {{< warning >}} to indicate danger or a piece of information that is crucial to follow.
For example:
{{< warning >}}
Beware.
{{< /warning >}}
The output is:
Warning: Beware.
Katacoda Embedded Live Environment
This button lets users run Minikube in their browser using the Katacoda Terminal.
It lowers the barrier of entry by allowing users to use Minikube with one click instead of going through the complete
Minikube and Kubectl installation process locally.
The Embedded Live Environment is configured to run minikube start and lets users complete tutorials in the same window
as the documentation.
Caution: The session is limited to 15 minutes.
For example:
{{< kat-button >}}
The output is:
Common Shortcode Issues
Ordered Lists
Shortcodes will interrupt numbered lists unless you indent four spaces before the notice and the tag.
For example:
1. Preheat oven to 350˚F
1. Prepare the batter, and pour into springform pan.
`{{< note >}}Grease the pan for best results.{{< /note >}}`
1. Bake for 20-25 minutes or until set.
The output is:
Preheat oven to 350˚F
Prepare the batter, and pour into springform pan.
Note: Grease the pan for best results.
Bake for 20-25 minutes or until set.
Include Statements
Shortcodes inside include statements will break the build. You must insert them in the parent document, before and after you call the include. For example:
{{< note >}}
{{< include "task-tutorial-prereqs.md" >}}
{{< /note >}}
Markdown elements
Line breaks
Use a single newline to separate block-level content like headings, lists, images, code blocks, and others. The exception is second-level headings, where it should be two newlines. Second-level headings follow the first-level (or the title) without any preceding paragraphs or texts. A two line spacing helps visualize the overall structure of content in a code editor better.
Headings
People accessing this documentation may use a screen reader or other assistive technology (AT). Screen readers are linear output devices, they output items on a page one at a time. If there is a lot of content on a page, you can use headings to give the page an internal structure. A good page structure helps all readers to easily navigate the page or filter topics of interest.
Do and Don't - Headings
Do
Don't
Update the title in the front matter of the page or blog post.
Use first level heading, as Hugo automatically converts the title in the front matter of the page into a first-level heading.
Use ordered headings to provide a meaningful high-level outline of your content.
Use headings level 4 through 6, unless it is absolutely necessary. If your content is that detailed, it may need to be broken into separate articles.
Use pound or hash signs (#) for non-blog post content.
Use underlines (--- or ===) to designate first-level headings.
Use sentence case for headings. For example, Extend kubectl with plugins
Use title case for headings. For example, Extend Kubectl With Plugins
Paragraphs
Do and Don't - Paragraphs
Do
Don't
Try to keep paragraphs under 6 sentences.
Indent the first paragraph with space characters. For example, ⋅⋅⋅Three spaces before a paragraph will indent it.
Use three hyphens (---) to create a horizontal rule. Use horizontal rules for breaks in paragraph content. For example, a change of scene in a story, or a shift of topic within a section.
Use horizontal rules for decoration.
Links
Do and Don't - Links
Do
Don't
Write hyperlinks that give you context for the content they link to. For example: Certain ports are open on your machines. See Check required ports for more details.
Use ambiguous terms such as "click here". For example: Certain ports are open on your machines. See here for more details.
Write Markdown-style links: [link text](URL). For example: [Hugo shortcodes](/docs/contribute/style/hugo-shortcodes/#table-captions) and the output is Hugo shortcodes.
Write HTML-style links: <a href="/media/examples/link-element-example.css" target="_blank">Visit our tutorial!</a>, or create links that open in new tabs or windows. For example: [example website](https://example.com){target="_blank"}
Lists
Group items in a list that are related to each other and need to appear in a specific order or to indicate a correlation between multiple items. When a screen reader comes across a list—whether it is an ordered or unordered list—it will be announced to the user that there is a group of list items. The user can then use the arrow keys to move up and down between the various items in the list.
Website navigation links can also be marked up as list items; after all they are nothing but a group of related links.
End each item in a list with a period if one or more items in the list are complete sentences. For the sake of consistency, normally either all items or none should be complete sentences.
Note: Ordered lists that are part of an incomplete introductory sentence can be in lowercase and punctuated as if each item was a part of the introductory sentence.
Use the number one (1.) for ordered lists.
Use (+), (*), or (-) for unordered lists.
Leave a blank line after each list.
Indent nested lists with four spaces (for example, ⋅⋅⋅⋅).
List items may consist of multiple paragraphs. Each subsequent paragraph in a list item must be indented by either four spaces or one tab.
Tables
The semantic purpose of a data table is to present tabular data. Sighted users can quickly scan the table but a screen reader goes through line by line. A table caption is used to create a descriptive title for a data table. Assistive technologies (AT) use the HTML table caption element to identify the table contents to the user within the page structure.
This section contains suggested best practices for clear, concise, and consistent content.
Use present tense
Do and Don't - Use present tense
Do
Don't
This command starts a proxy.
This command will start a proxy.
Exception: Use future or past tense if it is required to convey the correct
meaning.
Use active voice
Do and Don't - Use active voice
Do
Don't
You can explore the API using a browser.
The API can be explored using a browser.
The YAML file specifies the replica count.
The replica count is specified in the YAML file.
Exception: Use passive voice if active voice leads to an awkward construction.
Use simple and direct language
Use simple and direct language. Avoid using unnecessary phrases, such as saying "please."
Do and Don't - Use simple and direct language
Do
Don't
To create a ReplicaSet, ...
In order to create a ReplicaSet, ...
See the configuration file.
Please see the configuration file.
View the pods.
With this next command, we'll view the pods.
Address the reader as "you"
Do and Don't - Addressing the reader
Do
Don't
You can create a Deployment by ...
We'll create a Deployment by ...
In the preceding output, you can see...
In the preceding output, we can see ...
Avoid Latin phrases
Prefer English terms over Latin abbreviations.
Do and Don't - Avoid Latin phrases
Do
Don't
For example, ...
e.g., ...
That is, ...
i.e., ...
Exception: Use "etc." for et cetera.
Patterns to avoid
Avoid using "we"
Using "we" in a sentence can be confusing, because the reader might not know
whether they're part of the "we" you're describing.
Do and Don't - Patterns to avoid
Do
Don't
Version 1.4 includes ...
In version 1.4, we have added ...
Kubernetes provides a new feature for ...
We provide a new feature ...
This page teaches you how to use pods.
In this page, we are going to learn about pods.
Avoid jargon and idioms
Some readers speak English as a second language. Avoid jargon and idioms to help them understand better.
Do and Don't - Avoid jargon and idioms
Do
Don't
Internally, ...
Under the hood, ...
Create a new cluster.
Turn up a new cluster.
Avoid statements about the future
Avoid making promises or giving hints about the future. If you need to talk about
an alpha feature, put the text under a heading that identifies it as alpha
information.
An exception to this rule is documentation about announced deprecations
targeting removal in future versions. One example of documentation like this
is the Deprecated API migration guide.
Avoid statements that will soon be out of date
Avoid words like "currently" and "new." A feature that is new today might not be
considered new in a few months.
Do and Don't - Avoid statements that will soon be out of date
Do
Don't
In version 1.4, ...
In the current version, ...
The Federation feature provides ...
The new Federation feature provides ...
Avoid words that assume a specific level of understanding
Avoid words such as "just", "simply", "easy", "easily", or "simple". These words do not add value.
This guide shows you how to create, edit and share diagrams using the Mermaid
JavaScript library. Mermaid.js allows you to generate diagrams using a simple
markdown-like syntax inside Markdown files. You can also use Mermaid to
generate .svg or .png image files that you can add to your documentation.
The target audience for this guide is anybody wishing to learn about Mermaid
and/or how to create and add diagrams to Kubernetes documentation.
Figure 1 outlines the topics covered in this section.
flowchart LR
subgraph m[Mermaid.js]
direction TB
S[ ]-.-
C[build diagrams with markdown] -->
D[on-line live editor]
end
A[Why are diagrams useful?] --> m
m --> N[3 x methods for creating diagrams]
N --> T[Examples]
T --> X[Styling and captions]
X --> V[Tips]
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
classDef spacewhite fill:#ffffff,stroke:#fff,stroke-width:0px,color:#000
class A,C,D,N,X,m,T,V box
class S spacewhite
%% you can hyperlink Mermaid diagram nodes to a URL using click statements
click A "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click C "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click D "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click N "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click T "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click X "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
click V "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgc3ViZ3JhcGggbVtNZXJtYWlkLmpzXVxuICAgIGRpcmVjdGlvbiBUQlxuICAgICAgICBTWyBdLS4tXG4gICAgICAgIENbYnVpbGQ8YnI-ZGlhZ3JhbXM8YnI-d2l0aCBtYXJrZG93bl0gLS0-XG4gICAgICAgIERbb24tbGluZTxicj5saXZlIGVkaXRvcl1cbiAgICBlbmRcbiAgICBBW1doeSBhcmUgZGlhZ3JhbXM8YnI-dXNlZnVsP10gLS0-IG1cbiAgICBtIC0tPiBOWzMgeCBtZXRob2RzPGJyPmZvciBjcmVhdGluZzxicj5kaWFncmFtc11cbiAgICBOIC0tPiBUW0V4YW1wbGVzXVxuICAgIFQgLS0-IFhbU3R5bGluZzxicj5hbmQ8YnI-Y2FwdGlvbnNdXG4gICAgWCAtLT4gVltUaXBzXVxuICAgIFxuIFxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzRGVmIHNwYWNld2hpdGUgZmlsbDojZmZmZmZmLHN0cm9rZTojZmZmLHN0cm9rZS13aWR0aDowcHgsY29sb3I6IzAwMFxuICAgIGNsYXNzIEEsQyxELE4sWCxtLFQsViBib3hcbiAgICBjbGFzcyBTIHNwYWNld2hpdGUiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOnRydWV9" _blank
Figure 1. Topics covered in this section.
All you need to begin working with Mermaid is the following:
Note: You can click on each diagram in this section to view the code and rendered
diagram in the Mermaid live editor.
Why you should use diagrams in documentation
Diagrams improve documentation clarity and comprehension. There are advantages for both the user and the contributor.
The user benefits include:
Friendly landing spot. A detailed text-only greeting page could
intimidate users, in particular, first-time Kubernetes users.
Faster grasp of concepts. A diagram can help users understand the key
points of a complex topic. Your diagram can serve as a visual learning guide
to dive into the topic details.
Better retention. For some, it is easier to recall pictures rather than text.
The contributor benefits include:
Assist in developing the structure and content of your contribution. For
example, you can start with a simple diagram covering the high-level points
and then dive into details.
Expand and grow the user community. Easily consumed documentation
augmented with diagrams attracts new users who might previously have been
reluctant to engage due to perceived complexities.
You should consider your target audience. In addition to experienced K8s
users, you will have many who are new to Kubernetes. Even a simple diagram can
assist new users in absorbing Kubernetes concepts. They become emboldened and
more confident to further explore Kubernetes and the documentation.
Mermaid
Mermaid is an open source
JavaScript library that allows you to create, edit and easily share diagrams
using a simple, markdown-like syntax configured inline in Markdown files.
The following lists features of Mermaid:
Simple code syntax.
Includes a web-based tool allowing you to code and preview your diagrams.
Supports multiple formats including flowchart, state and sequence.
Easy collaboration with colleagues by sharing a per-diagram URL.
Broad selection of shapes, lines, themes and styling.
The following lists advantages of using Mermaid:
No need for separate, non-Mermaid diagram tools.
Adheres to existing PR workflow. You can think of Mermaid code as just
Markdown text included in your PR.
Simple tool builds simple diagrams. You don't want to get bogged down
(re)crafting an overly complex and detailed picture. Keep it simple!
Mermaid provides a simple, open and transparent method for the SIG communities
to add, edit and collaborate on diagrams for new or existing documentation.
Note: You can still use Mermaid to create/edit diagrams even if it's not supported
in your environment. This method is called Mermaid+SVG and is explained
below.
Live editor
The Mermaid live editor is
a web-based tool that enables you to create, edit and review diagrams.
The following lists live editor functions:
Displays Mermaid code and rendered diagram.
Generates a URL for each saved diagram. The URL is displayed in the URL
field of your browser. You can share the URL with colleagues who can access
and modify the diagram.
Option to download .svg or .png files.
Note: The live editor is the easiest and fastest way to create and edit Mermaid diagrams.
Methods for creating diagrams
Figure 2 outlines the three methods to generate and add diagrams.
graph TB
A[Contributor]
B[Inline
Mermaid code added to .md file]
C[Mermaid+SVG
Add mermaid-generated svg file to .md file]
D[External tool
Add external-tool- generated svg file to .md file]
A --> B
A --> C
A --> D
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class A,B,C,D box
%% you can hyperlink Mermaid diagram nodes to a URL using click statements
click A "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggVEJcbiAgICBBW0NvbnRyaWJ1dG9yXVxuICAgIEJbSW5saW5lPGJyPjxicj5NZXJtYWlkIGNvZGU8YnI-YWRkZWQgdG8gLm1kIGZpbGVdXG4gICAgQ1tNZXJtYWlkK1NWRzxicj48YnI-QWRkIG1lcm1haWQtZ2VuZXJhdGVkPGJyPnN2ZyBmaWxlIHRvIC5tZCBmaWxlXVxuICAgIERbRXh0ZXJuYWwgdG9vbDxicj48YnI-QWRkIGV4dGVybmFsLXRvb2wtPGJyPmdlbmVyYXRlZCBzdmcgZmlsZTxicj50byAubWQgZmlsZV1cblxuICAgIEEgLS0-IEJcbiAgICBBIC0tPiBDXG4gICAgQSAtLT4gRFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3giLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click B "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggVEJcbiAgICBBW0NvbnRyaWJ1dG9yXVxuICAgIEJbSW5saW5lPGJyPjxicj5NZXJtYWlkIGNvZGU8YnI-YWRkZWQgdG8gLm1kIGZpbGVdXG4gICAgQ1tNZXJtYWlkK1NWRzxicj48YnI-QWRkIG1lcm1haWQtZ2VuZXJhdGVkPGJyPnN2ZyBmaWxlIHRvIC5tZCBmaWxlXVxuICAgIERbRXh0ZXJuYWwgdG9vbDxicj48YnI-QWRkIGV4dGVybmFsLXRvb2wtPGJyPmdlbmVyYXRlZCBzdmcgZmlsZTxicj50byAubWQgZmlsZV1cblxuICAgIEEgLS0-IEJcbiAgICBBIC0tPiBDXG4gICAgQSAtLT4gRFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3giLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click C "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggVEJcbiAgICBBW0NvbnRyaWJ1dG9yXVxuICAgIEJbSW5saW5lPGJyPjxicj5NZXJtYWlkIGNvZGU8YnI-YWRkZWQgdG8gLm1kIGZpbGVdXG4gICAgQ1tNZXJtYWlkK1NWRzxicj48YnI-QWRkIG1lcm1haWQtZ2VuZXJhdGVkPGJyPnN2ZyBmaWxlIHRvIC5tZCBmaWxlXVxuICAgIERbRXh0ZXJuYWwgdG9vbDxicj48YnI-QWRkIGV4dGVybmFsLXRvb2wtPGJyPmdlbmVyYXRlZCBzdmcgZmlsZTxicj50byAubWQgZmlsZV1cblxuICAgIEEgLS0-IEJcbiAgICBBIC0tPiBDXG4gICAgQSAtLT4gRFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3giLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click D "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggVEJcbiAgICBBW0NvbnRyaWJ1dG9yXVxuICAgIEJbSW5saW5lPGJyPjxicj5NZXJtYWlkIGNvZGU8YnI-YWRkZWQgdG8gLm1kIGZpbGVdXG4gICAgQ1tNZXJtYWlkK1NWRzxicj48YnI-QWRkIG1lcm1haWQtZ2VuZXJhdGVkPGJyPnN2ZyBmaWxlIHRvIC5tZCBmaWxlXVxuICAgIERbRXh0ZXJuYWwgdG9vbDxicj48YnI-QWRkIGV4dGVybmFsLXRvb2wtPGJyPmdlbmVyYXRlZCBzdmcgZmlsZTxicj50byAubWQgZmlsZV1cblxuICAgIEEgLS0-IEJcbiAgICBBIC0tPiBDXG4gICAgQSAtLT4gRFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3giLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
Figure 2. Methods to create diagrams.
Inline
Figure 3 outlines the steps to follow for adding a diagram using the Inline
method.
graph LR
A[1. Use live editor to create/edit diagram] -->
B[2. Store diagram URL somewhere] -->
C[3. Copy Mermaid code to page markdown file] -->
D[4. Add caption]
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class A,B,C,D box
%% you can hyperlink Mermaid diagram nodes to a URL using click statements
click A "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggTFJcbiAgICBBWzEuIFVzZSBsaXZlIGVkaXRvcjxicj4gdG8gY3JlYXRlL2VkaXQ8YnI-ZGlhZ3JhbV0gLS0-XG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdIC0tPlxuICAgIENbMy4gQ29weSBNZXJtYWlkIGNvZGU8YnI-dG8gcGFnZSBtYXJrZG93biBmaWxlXSAtLT5cbiAgICBEWzQuIEFkZCBjYXB0aW9uXVxuIFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click B "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggTFJcbiAgICBBWzEuIFVzZSBsaXZlIGVkaXRvcjxicj4gdG8gY3JlYXRlL2VkaXQ8YnI-ZGlhZ3JhbV0gLS0-XG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdIC0tPlxuICAgIENbMy4gQ29weSBNZXJtYWlkIGNvZGU8YnI-dG8gcGFnZSBtYXJrZG93biBmaWxlXSAtLT5cbiAgICBEWzQuIEFkZCBjYXB0aW9uXVxuIFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click C "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggTFJcbiAgICBBWzEuIFVzZSBsaXZlIGVkaXRvcjxicj4gdG8gY3JlYXRlL2VkaXQ8YnI-ZGlhZ3JhbV0gLS0-XG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdIC0tPlxuICAgIENbMy4gQ29weSBNZXJtYWlkIGNvZGU8YnI-dG8gcGFnZSBtYXJrZG93biBmaWxlXSAtLT5cbiAgICBEWzQuIEFkZCBjYXB0aW9uXVxuIFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
click D "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZ3JhcGggTFJcbiAgICBBWzEuIFVzZSBsaXZlIGVkaXRvcjxicj4gdG8gY3JlYXRlL2VkaXQ8YnI-ZGlhZ3JhbV0gLS0-XG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdIC0tPlxuICAgIENbMy4gQ29weSBNZXJtYWlkIGNvZGU8YnI-dG8gcGFnZSBtYXJrZG93biBmaWxlXSAtLT5cbiAgICBEWzQuIEFkZCBjYXB0aW9uXVxuIFxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMsRCBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ" _blank
Figure 3. Inline Method steps.
The following lists the steps you should follow for adding a diagram using the Inline method:
Create your diagram using the live editor.
Store the diagram URL somewhere for later access.
Copy the mermaid code to the location in your .md file where you want the diagram to appear.
Add a caption below the diagram using Markdown text.
A Hugo build runs the Mermaid code and turns it into a diagram.
Note: You may find keeping track of diagram URLs is cumbersome. If so, make a note
in the .md file that the Mermaid code is self-documenting. Contributors can
copy the Mermaid code to and from the live editor for diagram edits.
Here is a sample code snippet contained in an .md file:
---
title: My PR
---
Figure 17 shows a simple A to B process.
some markdown text
...
{{< mermaid >}}
graph TB
A --> B
{{< /mermaid >}}
Figure 17. A to B
more text
Note: You must include the Hugo Mermaid shortcode
tags at the start and end of the Mermaid code block. You should add a diagram
caption below the diagram.
The following lists advantages of the Inline method:
Live editor tool.
Easy to copy Mermaid code to and from the live editor and your .md file.
No need for separate .svg image file handling.
Content text, diagram code and diagram caption contained in the same .md file.
You should use the local
and Netlify previews to verify the diagram is properly rendered.
Caution: The Mermaid live editor feature set may not support the K8s/website Mermaid feature set.
You might see a syntax error or a blank screen after the Hugo build.
If that is the case, consider using the Mermaid+SVG method.
Mermaid+SVG
Figure 4 outlines the steps to follow for adding a diagram using the Mermaid+SVG method.
flowchart LR
A[1. Use live editor to create/edit diagram]
B[2. Store diagram URL somewhere]
C[3. Generate .svg file and download to images/ folder]
subgraph w[ ]
direction TB
D[4. Use figure shortcode to reference .svg file in page .md file] -->
E[5. Add caption]
end
A --> B
B --> C
C --> w
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class A,B,C,D,E,w box
click A "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgbGl2ZSBlZGl0b3I8YnI-IHRvIGNyZWF0ZS9lZGl0PGJyPmRpYWdyYW1dXG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdXG4gICAgQ1szLiBHZW5lcmF0ZSAuc3ZnIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSAuc3ZnPGJyPmZpbGUgaW4gcGFnZTxicj4ubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbkEgLS0-IEJcbkIgLS0-IENcbkMgLS0-IHdcblxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzIEEsQixDLEQsRSx3IGJveFxuICAgICIsIm1lcm1haWQiOiJ7XG4gIFwidGhlbWVcIjogXCJkZWZhdWx0XCJcbn0iLCJ1cGRhdGVFZGl0b3IiOmZhbHNlLCJhdXRvU3luYyI6dHJ1ZSwidXBkYXRlRGlhZ3JhbSI6dHJ1ZX0" _blank
click B "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgbGl2ZSBlZGl0b3I8YnI-IHRvIGNyZWF0ZS9lZGl0PGJyPmRpYWdyYW1dXG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdXG4gICAgQ1szLiBHZW5lcmF0ZSAuc3ZnIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSAuc3ZnPGJyPmZpbGUgaW4gcGFnZTxicj4ubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbkEgLS0-IEJcbkIgLS0-IENcbkMgLS0-IHdcblxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzIEEsQixDLEQsRSx3IGJveFxuICAgICIsIm1lcm1haWQiOiJ7XG4gIFwidGhlbWVcIjogXCJkZWZhdWx0XCJcbn0iLCJ1cGRhdGVFZGl0b3IiOmZhbHNlLCJhdXRvU3luYyI6dHJ1ZSwidXBkYXRlRGlhZ3JhbSI6dHJ1ZX0" _blank
click C "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgbGl2ZSBlZGl0b3I8YnI-IHRvIGNyZWF0ZS9lZGl0PGJyPmRpYWdyYW1dXG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdXG4gICAgQ1szLiBHZW5lcmF0ZSAuc3ZnIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSAuc3ZnPGJyPmZpbGUgaW4gcGFnZTxicj4ubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbkEgLS0-IEJcbkIgLS0-IENcbkMgLS0-IHdcblxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzIEEsQixDLEQsRSx3IGJveFxuICAgICIsIm1lcm1haWQiOiJ7XG4gIFwidGhlbWVcIjogXCJkZWZhdWx0XCJcbn0iLCJ1cGRhdGVFZGl0b3IiOmZhbHNlLCJhdXRvU3luYyI6dHJ1ZSwidXBkYXRlRGlhZ3JhbSI6dHJ1ZX0" _blank
click D "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgbGl2ZSBlZGl0b3I8YnI-IHRvIGNyZWF0ZS9lZGl0PGJyPmRpYWdyYW1dXG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdXG4gICAgQ1szLiBHZW5lcmF0ZSAuc3ZnIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSAuc3ZnPGJyPmZpbGUgaW4gcGFnZTxicj4ubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbkEgLS0-IEJcbkIgLS0-IENcbkMgLS0-IHdcblxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzIEEsQixDLEQsRSx3IGJveFxuICAgICIsIm1lcm1haWQiOiJ7XG4gIFwidGhlbWVcIjogXCJkZWZhdWx0XCJcbn0iLCJ1cGRhdGVFZGl0b3IiOmZhbHNlLCJhdXRvU3luYyI6dHJ1ZSwidXBkYXRlRGlhZ3JhbSI6dHJ1ZX0" _blank
click E "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgbGl2ZSBlZGl0b3I8YnI-IHRvIGNyZWF0ZS9lZGl0PGJyPmRpYWdyYW1dXG4gICAgQlsyLiBTdG9yZSBkaWFncmFtPGJyPlVSTCBzb21ld2hlcmVdXG4gICAgQ1szLiBHZW5lcmF0ZSAuc3ZnIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSAuc3ZnPGJyPmZpbGUgaW4gcGFnZTxicj4ubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbkEgLS0-IEJcbkIgLS0-IENcbkMgLS0-IHdcblxuICAgIGNsYXNzRGVmIGJveCBmaWxsOiNmZmYsc3Ryb2tlOiMwMDAsc3Ryb2tlLXdpZHRoOjFweCxjb2xvcjojMDAwO1xuICAgIGNsYXNzIEEsQixDLEQsRSx3IGJveFxuICAgICIsIm1lcm1haWQiOiJ7XG4gIFwidGhlbWVcIjogXCJkZWZhdWx0XCJcbn0iLCJ1cGRhdGVFZGl0b3IiOmZhbHNlLCJhdXRvU3luYyI6dHJ1ZSwidXBkYXRlRGlhZ3JhbSI6dHJ1ZX0" _blank
Figure 4. Mermaid+SVG method steps.
The following lists the steps you should follow for adding a diagram using the Mermaid+SVG method:
Create your diagram using the live editor.
Store the diagram URL somewhere for later access.
Generate an .svg image file for the diagram and download it to the appropriate images/ folder.
Use the {{< figure >}} shortcode to reference the diagram in the .md file.
Add a caption using the {{< figure >}} shortcode's caption parameter.
For example, use the live editor to create a diagram called boxnet.
Store the diagram URL somewhere for later access. Generate and download a
boxnet.svg file to the appropriate ../images/ folder.
Use the {{< figure >}} shortcode in your PR's .md file to reference
the .svg image file and add a caption.
Note: The shortcode is the preferred method for adding .svg image files
to your documentation. You can also use the standard markdown image syntax like so:
.
And you will need to add a caption below the diagram.
You should add the live editor URL as a comment block in the .svg image file using a text editor.
For example, you would include the following at the beginning of the .svg image file:
<!-- To view or edit the mermaid code, use the following URL: -->
<!-- https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb ... <remainder of the URL> -->
The following lists advantages of the Mermaid+SVG method:
Live editor tool.
Live editor tool supports the most current Mermaid feature set.
Employ existing K8s/website methods for handling .svg image files.
Environment doesn't require Mermaid support.
Be sure to check that your diagram renders properly using the
local
and Netlify previews.
External tool
Figure 5 outlines the steps to follow for adding a diagram using the External Tool method.
First, use your external tool to create the diagram and save it as an .svg
or .png image file. After that, use the same steps as the Mermaid+SVG
method for adding .svg image files.
flowchart LR
A[1. Use external tool to create/edit diagram]
B[2. If possible, save diagram coordinates for contributor access]
C[3. Generate .svg or.png file and download to appropriate images/ folder]
subgraph w[ ]
direction TB
D[4. Use figure shortcode to reference svg or png file in page .md file] -->
E[5. Add caption]
end
A --> B
B --> C
C --> w
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class A,B,C,D,E,w box
click A "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgZXh0ZXJuYWw8YnI-dG9vbCB0byBjcmVhdGUvZWRpdDxicj5kaWFncmFtXVxuICAgIEJbMi4gSWYgcG9zc2libGUsIHNhdmU8YnI-ZGlhZ3JhbSBjb29yZGluYXRlczxicj5mb3IgY29udHJpYnV0b3I8YnI-YWNjZXNzXVxuICAgIENbMy4gR2VuZXJhdGUgLnN2ZyA8YnI-b3IucG5nIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmFwcHJvcHJpYXRlPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSBzdmcgb3I8YnI-cG5nIGZpbGUgaW48YnI-cGFnZSAubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbiAgICBBIC0tPiBCXG4gICAgQiAtLT4gQ1xuICAgIEMgLS0-IHdcbiAgICBjbGFzc0RlZiBib3ggZmlsbDojZmZmLHN0cm9rZTojMDAwLHN0cm9rZS13aWR0aDoxcHgsY29sb3I6IzAwMDtcbiAgICBjbGFzcyBBLEIsQyxELEUsdyBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ"
click B "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgZXh0ZXJuYWw8YnI-dG9vbCB0byBjcmVhdGUvZWRpdDxicj5kaWFncmFtXVxuICAgIEJbMi4gSWYgcG9zc2libGUsIHNhdmU8YnI-ZGlhZ3JhbSBjb29yZGluYXRlczxicj5mb3IgY29udHJpYnV0b3I8YnI-YWNjZXNzXVxuICAgIENbMy4gR2VuZXJhdGUgLnN2ZyA8YnI-b3IucG5nIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmFwcHJvcHJpYXRlPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSBzdmcgb3I8YnI-cG5nIGZpbGUgaW48YnI-cGFnZSAubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbiAgICBBIC0tPiBCXG4gICAgQiAtLT4gQ1xuICAgIEMgLS0-IHdcbiAgICBjbGFzc0RlZiBib3ggZmlsbDojZmZmLHN0cm9rZTojMDAwLHN0cm9rZS13aWR0aDoxcHgsY29sb3I6IzAwMDtcbiAgICBjbGFzcyBBLEIsQyxELEUsdyBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ"
click C "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgZXh0ZXJuYWw8YnI-dG9vbCB0byBjcmVhdGUvZWRpdDxicj5kaWFncmFtXVxuICAgIEJbMi4gSWYgcG9zc2libGUsIHNhdmU8YnI-ZGlhZ3JhbSBjb29yZGluYXRlczxicj5mb3IgY29udHJpYnV0b3I8YnI-YWNjZXNzXVxuICAgIENbMy4gR2VuZXJhdGUgLnN2ZyA8YnI-b3IucG5nIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmFwcHJvcHJpYXRlPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSBzdmcgb3I8YnI-cG5nIGZpbGUgaW48YnI-cGFnZSAubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbiAgICBBIC0tPiBCXG4gICAgQiAtLT4gQ1xuICAgIEMgLS0-IHdcbiAgICBjbGFzc0RlZiBib3ggZmlsbDojZmZmLHN0cm9rZTojMDAwLHN0cm9rZS13aWR0aDoxcHgsY29sb3I6IzAwMDtcbiAgICBjbGFzcyBBLEIsQyxELEUsdyBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ"
click D "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgZXh0ZXJuYWw8YnI-dG9vbCB0byBjcmVhdGUvZWRpdDxicj5kaWFncmFtXVxuICAgIEJbMi4gSWYgcG9zc2libGUsIHNhdmU8YnI-ZGlhZ3JhbSBjb29yZGluYXRlczxicj5mb3IgY29udHJpYnV0b3I8YnI-YWNjZXNzXVxuICAgIENbMy4gR2VuZXJhdGUgLnN2ZyA8YnI-b3IucG5nIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmFwcHJvcHJpYXRlPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSBzdmcgb3I8YnI-cG5nIGZpbGUgaW48YnI-cGFnZSAubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbiAgICBBIC0tPiBCXG4gICAgQiAtLT4gQ1xuICAgIEMgLS0-IHdcbiAgICBjbGFzc0RlZiBib3ggZmlsbDojZmZmLHN0cm9rZTojMDAwLHN0cm9rZS13aWR0aDoxcHgsY29sb3I6IzAwMDtcbiAgICBjbGFzcyBBLEIsQyxELEUsdyBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ"
click E "https://mermaid-js.github.io/mermaid-live-editor/edit/#eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG4gICAgQVsxLiBVc2UgZXh0ZXJuYWw8YnI-dG9vbCB0byBjcmVhdGUvZWRpdDxicj5kaWFncmFtXVxuICAgIEJbMi4gSWYgcG9zc2libGUsIHNhdmU8YnI-ZGlhZ3JhbSBjb29yZGluYXRlczxicj5mb3IgY29udHJpYnV0b3I8YnI-YWNjZXNzXVxuICAgIENbMy4gR2VuZXJhdGUgLnN2ZyA8YnI-b3IucG5nIGZpbGU8YnI-YW5kIGRvd25sb2FkIHRvPGJyPmFwcHJvcHJpYXRlPGJyPmltYWdlcy8gZm9sZGVyXVxuICAgIHN1YmdyYXBoIHdbIF1cbiAgICBkaXJlY3Rpb24gVEJcbiAgICBEWzQuIFVzZSBmaWd1cmUgc2hvcnRjb2RlPGJyPnRvIHJlZmVyZW5jZSBzdmcgb3I8YnI-cG5nIGZpbGUgaW48YnI-cGFnZSAubWQgZmlsZV0gLS0-XG4gICAgRVs1LiBBZGQgY2FwdGlvbl1cbiAgICBlbmRcbiAgICBBIC0tPiBCXG4gICAgQiAtLT4gQ1xuICAgIEMgLS0-IHdcbiAgICBjbGFzc0RlZiBib3ggZmlsbDojZmZmLHN0cm9rZTojMDAwLHN0cm9rZS13aWR0aDoxcHgsY29sb3I6IzAwMDtcbiAgICBjbGFzcyBBLEIsQyxELEUsdyBib3hcbiAgICAiLCJtZXJtYWlkIjoie1xuICBcInRoZW1lXCI6IFwiZGVmYXVsdFwiXG59IiwidXBkYXRlRWRpdG9yIjpmYWxzZSwiYXV0b1N5bmMiOnRydWUsInVwZGF0ZURpYWdyYW0iOmZhbHNlfQ"
Figure 5. External Tool method steps
The following lists the steps you should follow for adding a diagram using the External Tool method:
Use your external tool to create a diagram.
Save the diagram coordinates for contributor access. For example, your tool
may offer a link to the diagram image, or you could place the source code
file, such as an .xml file, in a public repository for later contributor access.
Generate and save the diagram as an .svg or .png image file.
Download this file to the appropriate ../images/ folder.
Use the {{< figure >}} shortcode to reference the diagram in the .md file.
Add a caption using the {{< figure >}} shortcode's caption parameter.
Here is the {{< figure >}} shortcode for the images/apple.svg diagram:
{{< figure src="/static/images/apple.svg" alt="red-apple-figure" class="diagram-large" caption="Figure 9. A Big Red Apple" >}}
If your external drawing tool permits:
You can incorporate multiple .svg or .png logos, icons and images into your diagram.
However, make sure you observe copyright and follow the Kubernetes documentation
guidelines on the use of third party content.
You should save the diagram source coordinates for later contributor access.
For example, your tool may offer a link to the diagram image, or you could
place the source code file, such as an .xml file, somewhere for contributor access.
For more information on K8s and CNCF logos and images, check out
CNCF Artwork.
The following lists advantages of the External Tool method:
Contributor familiarity with external tool.
Diagrams require more detail than what Mermaid can offer.
Don't forget to check that your diagram renders correctly using the
local and Netlify previews.
Examples
This section shows several examples of Mermaid diagrams.
Note: The code block examples omit the Hugo Mermaid
shortcode tags. This allows you to copy the code block into the live editor
to experiment on your own.
Note that the live editor doesn't recognize Hugo shortcodes.
graph LR;
client([client])-. Ingress-managed <br> load balancer .->ingress[Ingress];
ingress-->|routing rule|service[Service];
subgraph cluster
ingress;
service-->pod1[Pod];
service-->pod2[Pod];
end
classDef plain fill:#ddd,stroke:#fff,stroke-width:4px,color:#000;
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff;
classDef cluster fill:#fff,stroke:#bbb,stroke-width:2px,color:#326ce5;
class ingress,service,pod1,pod2 k8s;
class client plain;
class cluster cluster;
Example 3 - K8s system flow
Figure 8 depicts a Mermaid sequence diagram showing the system flow between
K8s components to start a container.
Figure 8. K8s system flow diagram
Code block:
%%{init:{"theme":"neutral"}}%%
sequenceDiagram
actor me
participant apiSrv as control plane<br><br>api-server
participant etcd as control plane<br><br>etcd datastore
participant cntrlMgr as control plane<br><br>controller<br>manager
participant sched as control plane<br><br>scheduler
participant kubelet as node<br><br>kubelet
participant container as node<br><br>container<br>runtime
me->>apiSrv: 1. kubectl create -f pod.yaml
apiSrv-->>etcd: 2. save new state
cntrlMgr->>apiSrv: 3. check for changes
sched->>apiSrv: 4. watch for unassigned pods(s)
apiSrv->>sched: 5. notify about pod w nodename=" "
sched->>apiSrv: 6. assign pod to node
apiSrv-->>etcd: 7. save new state
kubelet->>apiSrv: 8. look for newly assigned pod(s)
apiSrv->>kubelet: 9. bind pod to node
kubelet->>container: 10. start container
kubelet->>apiSrv: 11. update pod status
apiSrv-->>etcd: 12. save new state
How to style diagrams
You can style one or more diagram elements using well-known CSS nomenclature.
You accomplish this using two types of statements in the Mermaid code.
classDef defines a class of style attributes.
class defines one or more elements to apply the class to.
In the code for
figure 7,
you can see examples of both.
classDef k8s fill:#326ce5,stroke:#fff,stroke-width:4px,color:#fff; // defines style for the k8s class
class ingress,service,pod1,pod2 k8s; // k8s class is applied to elements ingress, service, pod1 and pod2.
You can include one or multiple classDef and class statements in your diagram.
You can also use the official K8s #326ce5 hex color code for K8s components in your diagram.
A caption is a brief description of a diagram. A title or a short description
of the diagram are examples of captions. Captions aren't meant to replace
explanatory text you have in your documentation. Rather, they serve as a
"context link" between that text and your diagram.
The combination of some text and a diagram tied together with a caption help
provide a concise representation of the information you wish to convey to the
user.
Without captions, you are asking the user to scan the text above or below the
diagram to figure out a meaning. This can be frustrating for the user.
Figure 9 lays out the three components for proper captioning: diagram, diagram
caption and the diagram referral.
flowchart
A[Diagram
Inline Mermaid or SVG image files]
B[Diagram Caption
Add Figure Number. and Caption Text]
C[Diagram Referral
Referenence Figure Number in text]
classDef box fill:#fff,stroke:#000,stroke-width:1px,color:#000;
class A,B,C box
click A "https://mermaid-js.github.io/mermaid-live-editor/edit#eyJjb2RlIjoiZmxvd2NoYXJ0XG4gICAgQVtEaWFncmFtPGJyPjxicj5JbmxpbmUgTWVybWFpZCBvcjxicj5TVkcgaW1hZ2UgZmlsZXNdXG4gICAgQltEaWFncmFtIENhcHRpb248YnI-PGJyPkFkZCBGaWd1cmUgTnVtYmVyLiBhbmQ8YnI-Q2FwdGlvbiBUZXh0XVxuICAgIENbRGlhZ3JhbSBSZWZlcnJhbDxicj48YnI-UmVmZXJlbmVuY2UgRmlndXJlIE51bWJlcjxicj5pbiB0ZXh0XVxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMgYm94IiwibWVybWFpZCI6IntcbiAgXCJ0aGVtZVwiOiBcImRlZmF1bHRcIlxufSIsInVwZGF0ZUVkaXRvciI6ZmFsc2UsImF1dG9TeW5jIjp0cnVlLCJ1cGRhdGVEaWFncmFtIjpmYWxzZX0" _blank
click B "https://mermaid-js.github.io/mermaid-live-editor/edit#eyJjb2RlIjoiZmxvd2NoYXJ0XG4gICAgQVtEaWFncmFtPGJyPjxicj5JbmxpbmUgTWVybWFpZCBvcjxicj5TVkcgaW1hZ2UgZmlsZXNdXG4gICAgQltEaWFncmFtIENhcHRpb248YnI-PGJyPkFkZCBGaWd1cmUgTnVtYmVyLiBhbmQ8YnI-Q2FwdGlvbiBUZXh0XVxuICAgIENbRGlhZ3JhbSBSZWZlcnJhbDxicj48YnI-UmVmZXJlbmVuY2UgRmlndXJlIE51bWJlcjxicj5pbiB0ZXh0XVxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMgYm94IiwibWVybWFpZCI6IntcbiAgXCJ0aGVtZVwiOiBcImRlZmF1bHRcIlxufSIsInVwZGF0ZUVkaXRvciI6ZmFsc2UsImF1dG9TeW5jIjp0cnVlLCJ1cGRhdGVEaWFncmFtIjpmYWxzZX0" _blank
click C "https://mermaid-js.github.io/mermaid-live-editor/edit#eyJjb2RlIjoiZmxvd2NoYXJ0XG4gICAgQVtEaWFncmFtPGJyPjxicj5JbmxpbmUgTWVybWFpZCBvcjxicj5TVkcgaW1hZ2UgZmlsZXNdXG4gICAgQltEaWFncmFtIENhcHRpb248YnI-PGJyPkFkZCBGaWd1cmUgTnVtYmVyLiBhbmQ8YnI-Q2FwdGlvbiBUZXh0XVxuICAgIENbRGlhZ3JhbSBSZWZlcnJhbDxicj48YnI-UmVmZXJlbmVuY2UgRmlndXJlIE51bWJlcjxicj5pbiB0ZXh0XVxuXG4gICAgY2xhc3NEZWYgYm94IGZpbGw6I2ZmZixzdHJva2U6IzAwMCxzdHJva2Utd2lkdGg6MXB4LGNvbG9yOiMwMDA7XG4gICAgY2xhc3MgQSxCLEMgYm94IiwibWVybWFpZCI6IntcbiAgXCJ0aGVtZVwiOiBcImRlZmF1bHRcIlxufSIsInVwZGF0ZUVkaXRvciI6ZmFsc2UsImF1dG9TeW5jIjp0cnVlLCJ1cGRhdGVEaWFncmFtIjpmYWxzZX0" _blank
Figure 9. Caption Components.
Note: You should always add a caption to each diagram in your documentation.
Diagram
The Mermaid+SVG and External Tool methods generate .svg image files.
Here is the {{< figure >}} shortcode for the diagram defined in an
.svg image file saved to /images/docs/components-of-kubernetes.svg:
{{< figure src="/images/docs/components-of-kubernetes.svg" alt="Kubernetes pod running inside a cluster" class="diagram-large" caption="Figure 4. Kubernetes Architecture Components >}}
You should pass the src, alt, class and caption values into the
{{< figure >}} shortcode. You can adjust the size of the diagram using
diagram-large, diagram-medium and diagram-small classes.
Note: Diagrams created using the Inline method don't use the
shortcode. The Mermaid code defines how the diagram will render on your page.
If you define your diagram in an .svg image file, then you should use the
{{< figure >}} shortcode's caption parameter.
{{< figure src="/images/docs/components-of-kubernetes.svg" alt="Kubernetes pod running inside a cluster" class="diagram-large" caption="Figure 4. Kubernetes Architecture Components" >}}
If you define your diagram using inline Mermaid code, then you should use Markdown text.
Figure 4. Kubernetes Architecture Components
The following lists several items to consider when adding diagram captions:
Use the {{< figure >}} shortcode to add a diagram caption for Mermaid+SVG
and External Tool diagrams.
Use simple Markdown text to add a diagram caption for the Inline method.
Prepend your diagram caption with Figure NUMBER.. You must use Figure
and the number must be unique for each diagram in your documentation page.
Add a period after the number.
Add your diagram caption text after the Figure NUMBER. on the same line.
You must puncuate the caption with a period. Keep the caption text short.
Position your diagram caption BELOW your diagram.
Diagram Referral
Finally, you can add a diagram referral. This is used inside your text and
should precede the diagram itself. It allows a user to connect your text with
the associated diagram. The Figure NUMBER in your referral and caption must
match.
You should avoid using spatial references such as ..the image below.. or
..the following figure ..
Here is an example of a diagram referral:
Figure 10 depicts the components of the Kubernetes architecture.
The control plane ...
Diagram referrals are optional and there are cases where they might not be
suitable. If you are not sure, add a diagram referral to your text to see if
it looks and sounds okay. When in doubt, use a diagram referral.
Complete picture
Figure 10 shows the Kubernetes Architecture diagram that includes the diagram,
diagram caption and diagram referral. The {{< figure >}} shortcode
renders the diagram, adds the caption and includes the optional link
parameter so you can hyperlink the diagram. The diagram referral is contained
in this paragraph.
Here is the {{< figure >}} shortcode for this diagram:
{{< figure src="/images/docs/components-of-kubernetes.svg" alt="Kubernetes pod running inside a cluster" class="diagram-large" caption="Figure 10. Kubernetes Architecture." link="https://kubernetes.io/docs/concepts/overview/components/" >}}
Figure 10. Kubernetes Architecture.
Tips
Always use the live editor to create/edit your diagram.
Always use Hugo local and Netlify previews to check out how the diagram
appears in the documentation.
Include diagram source pointers such as a URL, source code location, or
indicate the code is self-documenting.
Always use diagram captions.
Very helpful to include the diagram .svg or .png image and/or Mermaid
source code in issues and PRs.
With the Mermaid+SVG and External Tool methods, use .svg image files
because they stay sharp when you zoom in on the diagram.
Best practice for .svg files is to load it into an SVG editing tool and use the
"Convert text to paths" function.
This ensures that the diagram renders the same on all systems, regardless of font
availability and font rendering support.
No Mermaid support for additional icons or artwork.
Hugo Mermaid shortcodes don't work in the live editor.
Any time you modify a diagram in the live editor, you must save it
to generate a new URL for the diagram.
Click on the diagrams in this section to view the code and diagram rendering
in the live editor.
Look over the source code of this page, diagram-guide.md, for more examples.
Check out the Mermaid docs
for explanations and examples.
Most important, Keep Diagrams Simple.
This will save time for you and fellow contributors, and allow for easier reading
by new and experienced users.
7.6.4 - Writing a new topic
This page shows how to create a new topic for the Kubernetes docs.
Before you begin
Create a fork of the Kubernetes documentation repository as described in
Open a PR.
Choosing a page type
As you prepare to write a new topic, think about the page type that would fit your content the best:
Guidelines for choosing a page type
Type
Description
Concept
A concept page explains some aspect of Kubernetes. For example, a concept page might describe the Kubernetes Deployment object and explain the role it plays as an application while it is deployed, scaled, and updated. Typically, concept pages don't include sequences of steps, but instead provide links to tasks or tutorials. For an example of a concept topic, see Nodes.
Task
A task page shows how to do a single thing. The idea is to give readers a sequence of steps that they can actually do as they read the page. A task page can be short or long, provided it stays focused on one area. In a task page, it is OK to blend brief explanations with the steps to be performed, but if you need to provide a lengthy explanation, you should do that in a concept topic. Related task and concept topics should link to each other. For an example of a short task page, see Configure a Pod to Use a Volume for Storage. For an example of a longer task page, see Configure Liveness and Readiness Probes
Tutorial
A tutorial page shows how to accomplish a goal that ties together several Kubernetes features. A tutorial might provide several sequences of steps that readers can actually do as they read the page. Or it might provide explanations of related pieces of code. For example, a tutorial could provide a walkthrough of a code sample. A tutorial can include brief explanations of the Kubernetes features that are being tied together, but should link to related concept topics for deep explanations of individual features.
Creating a new page
Use a content type for each new page
that you write. The docs site provides templates or
Hugo archetypes to create
new content pages. To create a new type of page, run hugo new with the path to the file
you want to create. For example:
hugo new docs/concepts/my-first-concept.md
Choosing a title and filename
Choose a title that has the keywords you want search engines to find.
Create a filename that uses the words in your title separated by hyphens.
For example, the topic with title
Using an HTTP Proxy to Access the Kubernetes API
has filename http-proxy-access-api.md. You don't need to put
"kubernetes" in the filename, because "kubernetes" is already in the
URL for the topic, for example:
In your topic, put a title field in the
front matter.
The front matter is the YAML block that is between the
triple-dashed lines at the top of the page. Here's an example:
---
title: Using an HTTP Proxy to Access the Kubernetes API
---
Choosing a directory
Depending on your page type, put your new file in a subdirectory of one of these:
/content/en/docs/tasks/
/content/en/docs/tutorials/
/content/en/docs/concepts/
You can put your file in an existing subdirectory, or you can create a new
subdirectory.
Placing your topic in the table of contents
The table of contents is built dynamically using the directory structure of the
documentation source. The top-level directories under /content/en/docs/ create
top-level navigation, and subdirectories each have entries in the table of
contents.
Each subdirectory has a file _index.md, which represents the "home" page for
a given subdirectory's content. The _index.md does not need a template. It
can contain overview content about the topics in the subdirectory.
Other files in a directory are sorted alphabetically by default. This is almost
never the best order. To control the relative sorting of topics in a
subdirectory, set the weight: front-matter key to an integer. Typically, we
use multiples of 10, to account for adding topics later. For instance, a topic
with weight 10 will come before one with weight 20.
Embedding code in your topic
If you want to include some code in your topic, you can embed the code in your
file directly using the markdown code block syntax. This is recommended for the
following cases (not an exhaustive list):
The code shows the output from a command such as
kubectl get deploy mydeployment -o json | jq '.status'.
The code is not generic enough for users to try out. As an example, you can
embed the YAML
file for creating a Pod which depends on a specific
FlexVolume implementation.
The code is an incomplete example because its purpose is to highlight a
portion of a larger file. For example, when describing ways to
customize a RoleBinding,
you can provide a short snippet directly in your topic file.
The code is not meant for users to try out due to other reasons. For example,
when describing how a new attribute should be added to a resource using the
kubectl edit command, you can provide a short example that includes only
the attribute to add.
Including code from another file
Another way to include code in your topic is to create a new, complete sample
file (or group of sample files) and then reference the sample from your topic.
Use this method to include sample YAML files when the sample is generic and
reusable, and you want the reader to try it out themselves.
When adding a new standalone sample file, such as a YAML file, place the code in
one of the <LANG>/examples/ subdirectories where <LANG> is the language for
the topic. In your topic file, use the codenew shortcode:
{{< codenew file="<RELPATH>/my-example-yaml>" >}}
where <RELPATH> is the path to the file to include, relative to the
examples directory. The following Hugo shortcode references a YAML
file located at /content/en/examples/pods/storage/gce-volume.yaml.
Note: To show raw Hugo shortcodes as in the above example and prevent Hugo
from interpreting them, use C-style comments directly after the < and before
the > characters. View the code for this page for an example.
Showing how to create an API object from a configuration file
If you need to demonstrate how to create an API object based on a
configuration file, place the configuration file in one of the subdirectories
under <LANG>/examples.
Note: When adding new YAML files to the <LANG>/examples directory, make
sure the file is also included into the <LANG>/examples_test.go file. The
Travis CI for the Website automatically runs this test case when PRs are
submitted to ensure all examples pass the tests.
The Kubernetes documentation follows several types of page content:
Concept
Task
Tutorial
Reference
Content sections
Each page content type contains a number of sections defined by
Markdown comments and HTML headings. You can add content headings to
your page with the heading shortcode. The comments and headings help
maintain the structure of the page content types.
Examples of Markdown comments defining page content sections:
<!-- overview -->
<!-- body -->
To create common headings in your content pages, use the heading shortcode with
a heading string.
Examples of heading strings:
whatsnext
prerequisites
objectives
cleanup
synopsis
seealso
options
For example, to create a whatsnext heading, add the heading shortcode with the "whatsnext" string:
## {{% heading "whatsnext" %}}
You can declare a prerequisites heading as follows:
## {{% heading "prerequisites" %}}
The heading shortcode expects one string parameter.
The heading string parameter matches the prefix of a variable in the i18n/<lang>.toml files.
For example:
i18n/en.toml:
[whatsnext_heading]
other = "What's next"
i18n/ko.toml:
[whatsnext_heading]
other = "다음 내용"
Content types
Each content type informally defines its expected page structure.
Create page content with the suggested page sections.
Concept
A concept page explains some aspect of Kubernetes. For example, a concept
page might describe the Kubernetes Deployment object and explain the role it
plays as an application once it is deployed, scaled, and updated. Typically, concept
pages don't include sequences of steps, but instead provide links to tasks or
tutorials.
To write a new concept page, create a Markdown file in a subdirectory of the
/content/en/docs/concepts directory, with the following characteristics:
Concept pages are divided into three sections:
Page section
overview
body
whatsnext
The overview and body sections appear as comments in the concept page.
You can add the whatsnext section to your page with the heading shortcode.
Fill each section with content. Follow these guidelines:
Organize content with H2 and H3 headings.
For overview, set the topic's context with a single paragraph.
For body, explain the concept.
For whatsnext, provide a bulleted list of topics (5 maximum) to learn more about the concept.
Annotations is a published example of a concept page.
Task
A task page shows how to do a single thing, typically by giving a short
sequence of steps. Task pages have minimal explanation, but often provide links
to conceptual topics that provide related background and knowledge.
To write a new task page, create a Markdown file in a subdirectory of the
/content/en/docs/tasks directory, with the following characteristics:
Page section
overview
prerequisites
steps
discussion
whatsnext
The overview, steps, and discussion sections appear as comments in the task page.
You can add the prerequisites and whatsnext sections to your page
with the heading shortcode.
Within each section, write your content. Use the following guidelines:
Use a minimum of H2 headings (with two leading # characters). The sections
themselves are titled automatically by the template.
For overview, use a paragraph to set context for the entire topic.
For prerequisites, use bullet lists when possible. Start adding additional
prerequisites below the include. The default prerequisites include a running Kubernetes cluster.
For steps, use numbered lists.
For discussion, use normal content to expand upon the information covered
in steps.
For whatsnext, give a bullet list of up to 5 topics the reader might be
interested in reading next.
A tutorial page shows how to accomplish a goal that is larger than a single
task. Typically a tutorial page has several sections, each of which has a
sequence of steps. For example, a tutorial might provide a walkthrough of a
code sample that illustrates a certain feature of Kubernetes. Tutorials can
include surface-level explanations, but should link to related concept topics
for deep explanations.
To write a new tutorial page, create a Markdown file in a subdirectory of the
/content/en/docs/tutorials directory, with the following characteristics:
Page section
overview
prerequisites
objectives
lessoncontent
cleanup
whatsnext
The overview, objectives, and lessoncontent sections appear as comments in the tutorial page.
You can add the prerequisites, cleanup, and whatsnext sections to your page
with the heading shortcode.
Within each section, write your content. Use the following guidelines:
Use a minimum of H2 headings (with two leading # characters). The sections
themselves are titled automatically by the template.
For overview, use a paragraph to set context for the entire topic.
For prerequisites, use bullet lists when possible. Add additional
prerequisites below the ones included by default.
For objectives, use bullet lists.
For lessoncontent, use a mix of numbered lists and narrative content as
appropriate.
For cleanup, use numbered lists to describe the steps to clean up the
state of the cluster after finishing the task.
For whatsnext, give a bullet list of up to 5 topics the reader might be
interested in reading next.
A component tool reference page shows the description and flag options output for
a Kubernetes component tool. Each page generates from scripts using the component tool commands.
A tool reference page has several possible sections:
Note:Hugo Tip: Start Hugo with hugo server --navigateToChanged for content edit-sessions.
Page Lists
Page Order
The documentation side menu, the documentation page browser etc. are listed using Hugo's default sort order, which sorts by weight (from 1), date (newest first), and finally by the link title.
Given that, if you want to move a page or a section up, set a weight in the page's front matter:
title:My Pageweight:10
Note: For page weights, it can be smart not to use 1, 2, 3 ..., but some other interval, say 10, 20, 30... This allows you to insert pages where you want later.
Documentation Main Menu
The Documentation main menu is built from the sections below docs/ with the main_menu flag set in front matter of the _index.md section content file:
main_menu:true
Note that the link title is fetched from the page's linkTitle, so if you want it to be something different than the title, change it in the content file:
main_menu:truetitle:Page TitlelinkTitle:Title used in links
Note: The above needs to be done per language. If you don't see your section in the menu, it is probably because it is not identified as a section by Hugo. Create a _index.md content file in the section folder.
Documentation Side Menu
The documentation side-bar menu is built from the current section tree starting below docs/.
It will show all sections and their pages.
If you don't want to list a section or page, set the toc_hide flag to true in front matter:
toc_hide:true
When you navigate to a section that has content, the specific section or page (e.g. _index.md) is shown. Else, the first page inside that section is shown.
Documentation Browser
The page browser on the documentation home page is built using all the sections and pages that are directly below the docs section.
If you don't want to list a section or page, set the toc_hide flag to true in front matter:
toc_hide:true
The Main Menu
The site links in the top-right menu -- and also in the footer -- are built by page-lookups. This is to make sure that the page actually exists. So, if the case-studies section does not exist in a site (language), it will not be linked to.
Page Bundles
In addition to standalone content pages (Markdown files), Hugo supports Page Bundles.
One example is Custom Hugo Shortcodes. It is considered a leaf bundle. Everything below the directory, including the index.md, will be part of the bundle. This also includes page-relative links, images that can be processed etc.:
Another widely used example is the includes bundle. It sets headless: true in front matter, which means that it does not get its own URL. It is only used in other pages.
For translated bundles, any missing non-content files will be inherited from languages above. This avoids duplication.
All the files in a bundle are what Hugo calls Resources and you can provide metadata per language, such as parameters and title, even if it does not supports front matter (YAML files etc.). See Page Resources Metadata.
The value you get from .RelPermalink of a Resource is page-relative. See Permalinks.
Styles
The SASS source of the stylesheets for this site is stored in assets/sass and is automatically built by Hugo.
In a Markdown page (.md file) on this site, you can add a shortcode to
display version and state of the documented feature.
Feature state demo
Below is a demo of the feature state snippet, which displays the feature as
stable in the latest Kubernetes version.
{{< feature-state state="stable" >}}
Renders to:
FEATURE STATE:Kubernetes v1.25 [stable]
The valid values for state are:
alpha
beta
deprecated
stable
Feature state code
The displayed Kubernetes version defaults to that of the page or the site. You can change the
feature state version by passing the for_k8s_version shortcode parameter. For example:
There are two glossary shortcodes: glossary_tooltip and glossary_definition.
You can reference glossary terms with an inclusion that automatically updates
and replaces content with the relevant links from our glossary.
When the glossary term is moused-over, the glossary entry displays a tooltip.
The glossary term also displays as a link.
As well as inclusions with tooltips, you can reuse the definitions from the glossary in
page content.
The raw data for glossary terms is stored at
the glossary directory,
with a content file for each glossary term.
Glossary demo
For example, the following include within the Markdown renders to
cluster with a tooltip:
A set of worker machines, called nodes,
that run containerized applications. Every cluster has at least one worker node.
The worker node(s) host the Pods that are
the components of the application workload. The
control plane manages the worker
nodes and the Pods in the cluster. In production environments, the control plane usually
runs across multiple computers and a cluster usually runs multiple nodes, providing
fault-tolerance and high availability.
Links to API Reference
You can link to a page of the Kubernetes API reference using the
api-reference shortcode, for example to the
Pod reference:
The content of the page parameter is the suffix of the URL of the API reference page.
You can link to a specific place into a page by specifying an anchor
parameter, for example to the
PodSpec
reference or the
environment-variables
section of the page:
You can make tables more accessible to screen readers by adding a table caption. To add a
caption to a table,
enclose the table with a table shortcode and specify the caption with the caption parameter.
Note: Table captions are visible to screen readers but invisible when viewed in standard HTML.
Here's an example:
{{<tablecaption="Configuration parameters">}}Parameter | Description | Default
:---------|:------------|:-------
`timeout` | The timeout for requests | `30s`
`logLevel` | The log level for log output | `INFO`
{{</table>}}
The rendered table looks like this:
Configuration parameters
Parameter
Description
Default
timeout
The timeout for requests
30s
logLevel
The log level for log output
INFO
If you inspect the HTML for the table, you should see this element immediately
after the opening <table> element:
In a markdown page (.md file) on this site, you can add a tab set to display
multiple flavors of a given solution.
The tabs shortcode takes these parameters:
name: The name as shown on the tab.
codelang: If you provide inner content to the tab shortcode, you can tell Hugo
what code language to use for highlighting.
include: The file to include in the tab. If the tab lives in a Hugo
leaf bundle,
the file -- which can be any MIME type supported by Hugo -- is looked up in the bundle itself.
If not, the content page that needs to be included is looked up relative to the current page.
Note that with the include, you do not have any shortcode inner content and must use the
self-closing syntax. For example,
{{< tab name="Content File #1" include="example1" />}}. The language needs to be specified
under codelang or the language is taken based on the file name.
Non-content files are code-highlighted by default.
If your inner content is markdown, you must use the %-delimiter to surround the tab.
For example, {{% tab name="Tab 1" %}}This is **markdown**{{% /tab %}}
You can combine the variations mentioned above inside a tab set.
Below is a demo of the tabs shortcode.
Note: The tab name in a tabs definition must be unique within a content page.
Tabs demo: Code highlighting
{{<tabsname="tab_with_code">}}{{<tabname="Tab 1"codelang="bash">}}echo "This is tab 1."
{{</tab>}}{{<tabname="Tab 2"codelang="go">}}println "This is tab 2."
{{</tab>}}{{</tabs>}}
{{<tabsname="tab_with_md">}}{{%tabname="Markdown"%}}This is **some markdown.**
{{<note>}}It can even contain shortcodes.
{{</note>}}{{%/tab%}}{{<tabname="HTML">}}<div>
<h3>Plain HTML</h3>
<p>This is some <i>plain</i> HTML.</p>
</div>
{{</tab>}}{{</tabs>}}
Running Kubernetes requires third-party software. For example: you
usually need to add a
DNS server
to your cluster so that name resolution works.
When we link to third-party software, or otherwise mention it,
we follow the content guide
and we also mark those third party items.
Using these shortcodes adds a disclaimer to any documentation page
that uses them.
Lists
For a list of several third-party items, add:
{{% thirdparty-content %}}
just below the heading for the section that includes all items.
Items
If you have a list where most of the items refer to in-project
software (for example: Kubernetes itself, and the separate
Descheduler
component), then there is a different form to use.
Add the shortcode:
{{% thirdparty-content single="true" %}}
before the item, or just below the heading for the specific item.
Version strings
To generate a version string for inclusion in the documentation, you can choose from
several version shortcodes. Each version shortcode displays a version string derived from
the value of a version parameter found in the site configuration file, config.toml.
The two most commonly used version parameters are latest and version.
{{< param "version" >}}
The {{< param "version" >}} shortcode generates the value of the current
version of the Kubernetes documentation from the version site parameter. The
param shortcode accepts the name of one site parameter, in this case:
version.
Note: In previously released documentation, latest and version parameter values
are not equivalent. After a new version is released, latest is incremented
and the value of version for the documentation set remains unchanged. For
example, a previously released version of the documentation displays version
as v1.19 and latest as v1.20.
Renders to:
v1.25
{{< latest-version >}}
The {{< latest-version >}} shortcode returns the value of the latest site parameter.
The latest site parameter is updated when a new version of the documentation is released.
This parameter does not always match the value of version in a documentation set.
Renders to:
v1.25
{{< latest-semver >}}
The {{< latest-semver >}} shortcode generates the value of latest
without the "v" prefix.
Renders to:
1.25
{{< version-check >}}
The {{< version-check >}} shortcode checks if the min-kubernetes-server-version
page parameter is present and then uses this value to compare to version.
Renders to:
To check the version, enter kubectl version.
{{< latest-release-notes >}}
The {{< latest-release-notes >}} shortcode generates a version string
from latest and removes the "v" prefix. The shortcode prints a new URL for
the release note CHANGELOG page with the modified version string.
7.7.1 - Contributing to the Upstream Kubernetes Code
This page shows how to contribute to the upstream kubernetes/kubernetes project.
You can fix bugs found in the Kubernetes API documentation or the content of
the Kubernetes components such as kubeadm, kube-apiserver, and kube-controller-manager.
If you instead want to regenerate the reference documentation for the Kubernetes
API or the kube-* components from the upstream code, see the following instructions:
The reference documentation for the Kubernetes API and the kube-* components
such as kube-apiserver, kube-controller-manager are automatically generated
from the source code in the upstream Kubernetes.
When you see bugs in the generated documentation, you may want to consider
creating a patch to fix it in the upstream project.
Cloning the Kubernetes repository
If you don't already have the kubernetes/kubernetes repository, get it now:
mkdir $GOPATH/src
cd$GOPATH/src
go get github.com/kubernetes/kubernetes
Determine the base directory of your clone of the
kubernetes/kubernetes repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/kubernetes/kubernetes.
The remaining steps refer to your base directory as <k8s-base>.
Determine the base directory of your clone of the
kubernetes-sigs/reference-docs repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/kubernetes-sigs/reference-docs.
The remaining steps refer to your base directory as <rdocs-base>.
Editing the Kubernetes source code
The Kubernetes API reference documentation is automatically generated from
an OpenAPI spec, which is generated from the Kubernetes source code. If you
want to change the API reference documentation, the first step is to change one
or more comments in the Kubernetes source code.
The documentation for the kube-* components is also generated from the upstream
source code. You must change the code related to the component
you want to fix in order to fix the generated documentation.
Making changes to the upstream source code
Note: The following steps are an example, not a general procedure. Details
will be different in your situation.
Here's an example of editing a comment in the Kubernetes source code.
In your local kubernetes/kubernetes repository, check out the default branch,
and make sure it is up to date:
cd <k8s-base>
git checkout master
git pull https://github.com/kubernetes/kubernetes master
Suppose this source file in that default branch has the typo "atmost":
In your local environment, open types.go, and change "atmost" to "at most".
Verify that you have changed the file:
git status
The output shows that you are on the master branch, and that the types.go
source file has been modified:
On branch master
...
modified: staging/src/k8s.io/api/apps/v1/types.go
Committing your edited file
Run git add and git commit to commit the changes you have made so far. In the next step,
you will do a second commit. It is important to keep your changes separated into two commits.
View the contents of api/openapi-spec/swagger.json to make sure the typo is fixed.
For example, you could run git diff -a api/openapi-spec/swagger.json.
This is important, because swagger.json is the input to the second stage of
the doc generation process.
Run git add and git commit to commit your changes. Now you have two commits:
one that contains the edited types.go file, and one that contains the generated OpenAPI spec
and related files. Keep these two commits separate. That is, do not squash your commits.
Submit your changes as a
pull request to the
master branch of the
kubernetes/kubernetes repository.
Monitor your pull request, and respond to reviewer comments as needed. Continue
to monitor your pull request until it is merged.
PR 57758
is an example of a pull request that fixes a typo in the Kubernetes source code.
Note: It can be tricky to determine the correct source file to be changed. In the
preceding example, the authoritative source file is in the staging directory
in the kubernetes/kubernetes repository. But in your situation,the staging directory
might not be the place to find the authoritative source. For guidance, check the
README files in
kubernetes/kubernetes
repository and in related repositories, such as
kubernetes/apiserver.
Cherry picking your commit into a release branch
In the preceding section, you edited a file in the master branch and then ran scripts
to generate an OpenAPI spec and related files. Then you submitted your changes in a pull request
to the master branch of the kubernetes/kubernetes repository. Now suppose you want to backport
your change into a release branch. For example, suppose the master branch is being used to develop
Kubernetes version 1.25, and you want to backport your change into the
release-1.24 branch.
Recall that your pull request has two commits: one for editing types.go
and one for the files generated by scripts. The next step is to propose a cherry pick of your first
commit into the release-1.24 branch. The idea is to cherry pick the commit
that edited types.go, but not the commit that has the results of running the scripts. For instructions, see
Propose a Cherry Pick.
Note: Proposing a cherry pick requires that you have permission to set a label and a milestone in your
pull request. If you don't have those permissions, you will need to work with someone who can set the label
and milestone for you.
When you have a pull request in place for cherry picking your one commit into the
release-1.24 branch, the next step is to run these scripts in the
release-1.24 branch of your local environment.
Now add a commit to your cherry-pick pull request that has the recently generated OpenAPI spec
and related files. Monitor your pull request until it gets merged into the
release-1.24 branch.
At this point, both the master branch and the release-1.24 branch have your updated types.go
file and a set of generated files that reflect the change you made to types.go. Note that the
generated OpenAPI spec and other generated files in the release-1.24 branch are not necessarily
the same as the generated files in the master branch. The generated files in the release-1.24 branch
contain API elements only from Kubernetes 1.24. The generated files in the master branch might contain
API elements that are not in 1.24, but are under development for 1.25.
Generating the published reference docs
The preceding section showed how to edit a source file and then generate
several files, including api/openapi-spec/swagger.json in the
kubernetes/kubernetes repository.
The swagger.json file is the OpenAPI definition file to use for generating
the API reference documentation.
This page shows how to use the update-imported-docs.py script to generate
the Kubernetes reference documentation. The script automates
the build setup and generates the reference documentation for a release.
Before you begin
Requirements:
You need a machine that is running Linux or macOS.
Docker (Required only for kubectl command reference)
Your PATH environment variable must include the required build tools, such as the Go binary and python.
You need to know how to create a pull request to a GitHub repository.
This involves creating your own fork of the repository. For more
information, see Work from a local clone.
Getting the docs repository
Make sure your website fork is up-to-date with the kubernetes/website remote on
GitHub (main branch), and clone your website fork.
mkdir github.com
cd github.com
git clone git@github.com:<your_github_username>/website.git
Determine the base directory of your clone. For example, if you followed the
preceding step to get the repository, your base directory is
github.com/website. The remaining steps refer to your base directory as
<web-base>.
Note: If you want to change the content of the component tools and API reference,
see the contributing upstream guide.
Overview of update-imported-docs
The update-imported-docs.py script is located in the <web-base>/update-imported-docs/
directory.
The script builds the following references:
Component and tool reference pages
The kubectl command reference
The Kubernetes API reference
The update-imported-docs.py script generates the Kubernetes reference documentation
from the Kubernetes source code. The script creates a temporary directory
under /tmp on your machine and clones the required repositories: kubernetes/kubernetes and
kubernetes-sigs/reference-docs into this directory.
The script sets your GOPATH to this temporary directory.
Three additional environment variables are set:
K8S_RELEASE
K8S_ROOT
K8S_WEBROOT
The script requires two arguments to run successfully:
A YAML configuration file (reference.yml)
A release version, for example:1.17
The configuration file contains a generate-command field.
The generate-command field defines a series of build instructions
from kubernetes-sigs/reference-docs/Makefile. The K8S_RELEASE variable
determines the version of the release.
The update-imported-docs.py script performs the following steps:
Clones the related repositories specified in a configuration file. For the
purpose of generating reference docs, the repository that is cloned by
default is kubernetes-sigs/reference-docs.
Runs commands under the cloned repositories to prepare the docs generator and
then generates the HTML and Markdown files.
Copies the generated HTML and Markdown files to a local clone of the <web-base>
repository under locations specified in the configuration file.
Updates kubectl command links from kubectl.md to the refer to
the sections in the kubectl command reference.
When the generated files are in your local clone of the <web-base>
repository, you can submit them in a pull request
to <web-base>.
Configuration file format
Each configuration file may contain multiple repos that will be imported together. When
necessary, you can customize the configuration file by manually editing it. You
may create new config files for importing other groups of documents.
The following is an example of the YAML configuration file:
Open <web-base>/update-imported-docs/reference.yml for editing.
Do not change the content for the generate-command field unless you understand
how the command is used to build the references.
You should not need to update reference.yml. At times, changes in the
upstream source code, may require changes to the configuration file
(for example: golang version dependencies and third-party library changes).
If you encounter build issues, contact the SIG-Docs team on the
#sig-docs Kubernetes Slack channel.
Note: The generate-command is an optional entry, which can be used to run a
given command or a short script to generate the docs from within a repository.
In reference.yml, files contains a list of src and dst fields.
The src field contains the location of a generated Markdown file in the cloned
kubernetes-sigs/reference-docs build directory, and the dst field specifies
where to copy this file in the cloned kubernetes/website repository.
For example:
Note that when there are many files to be copied from the same source directory
to the same destination directory, you can use wildcards in the value given to
src. You must provide the directory name as the value for dst.
For example:
You can run the update-imported-docs.py tool as follows:
cd <web-base>/update-imported-docs
./update-imported-docs.py <configuration-file.yml> <release-version>
For example:
./update-imported-docs.py reference.yml 1.17
Fixing Links
The release.yml configuration file contains instructions to fix relative links.
To fix relative links within your imported files, set thegen-absolute-links
property to true. You can find an example of this in
release.yml.
Adding and committing changes in kubernetes/website
List the files that were generated and copied to <web-base>:
cd <web-base>
git status
The output shows the new and modified files. The generated output varies
depending upon changes made to the upstream source code.
Create a pull request to the kubernetes/website repository. Monitor your
pull request, and respond to review comments as needed. Continue to monitor
your pull request until it is merged.
A few minutes after your pull request is merged, your updated reference
topics will be visible in the
published documentation.
What's next
To generate the individual reference documentation by manually setting up the required build repositories and
running the build targets, see the following guides:
Docker (Required only for kubectl command reference)
Your PATH environment variable must include the required build tools, such as the Go binary and python.
You need to know how to create a pull request to a GitHub repository.
This involves creating your own fork of the repository. For more
information, see Work from a local clone.
The base directory of your clone of the
kubernetes/kubernetes repository is
$GOPATH/src/k8s.io/kubernetes.
The remaining steps refer to your base directory as <k8s-base>.
The base directory of your clone of the
kubernetes/website repository is
$GOPATH/src/github.com/<your username>/website.
The remaining steps refer to your base directory as <web-base>.
The base directory of your clone of the
kubernetes-sigs/reference-docs
repository is $GOPATH/src/github.com/kubernetes-sigs/reference-docs.
The remaining steps refer to your base directory as <rdocs-base>.
Creating versioned directory and fetching Open API spec
The updateapispec build target creates the versioned build directory.
After the directory is created, the Open API spec is fetched from the
<k8s-base> repository. These steps ensure that the version
of the configuration files and Kubernetes Open API spec match the release version.
The versioned directory name follows the pattern of v<major>_<minor>.
In the <rdocs-base> directory, run the following build target:
cd <rdocs-base>
make updateapispec
Building the API reference docs
The copyapi target builds the API reference and
copies the generated files to directories in <web-base>.
Run the following command in <rdocs-base>:
When generating reference documentation for a new release, update the file,
<web-base>/content/en/docs/reference/kubernetes-api/api-index.md with the new
version number.
Open <web-base>/content/en/docs/reference/kubernetes-api/api-index.md for editing,
and update the API reference version number. For example:
---
title: v1.17
---
[Kubernetes API v1.17](/docs/reference/generated/kubernetes-api/v1.17/)
Open <web-base>/content/en/docs/reference/_index.md for editing, and add a
new link for the latest API reference. Remove the oldest API reference version.
There should be five links to the most recent API references.
Locally test the API reference
Publish a local version of the API reference.
Verify the local preview.
cd <web-base>
git submodule update --init --recursive --depth 1# if not already donemake container-serve
Commit the changes
In <web-base> run git add and git commit to commit the change.
Submit your changes as a
pull request to the
kubernetes/website repository.
Monitor your pull request, and respond to reviewer comments as needed. Continue
to monitor your pull request until it has been merged.
Docker (Required only for kubectl command reference)
Your PATH environment variable must include the required build tools, such as the Go binary and python.
You need to know how to create a pull request to a GitHub repository.
This involves creating your own fork of the repository. For more
information, see Work from a local clone.
The kubernetes/kubernetes repository provides the kubectl and kustomize source code.
Determine the base directory of your clone of the
kubernetes/kubernetes repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/k8s.io/kubernetes.
The remaining steps refer to your base directory as <k8s-base>.
Determine the base directory of your clone of the
kubernetes/website repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/<your-username>/website.
The remaining steps refer to your base directory as <web-base>.
Determine the base directory of your clone of the
kubernetes-sigs/reference-docs repository.
For example, if you followed the preceding step to get the repository, your
base directory is $GOPATH/src/github.com/kubernetes-sigs/reference-docs.
The remaining steps refer to your base directory as <rdocs-base>.
In your local k8s.io/kubernetes repository, check out the branch of interest,
and make sure it is up to date. For example, if you want to generate docs for
Kubernetes 1.24.0, you could use these commands:
cd <k8s-base>
git checkout v1.24.0
git pull https://github.com/kubernetes/kubernetes 1.24.0
If you do not need to edit the kubectl source code, follow the instructions for
Setting build variables.
Editing the kubectl source code
The kubectl command reference documentation is automatically generated from
the kubectl source code. If you want to change the reference documentation, the first step
is to change one or more comments in the kubectl source code. Make the change in your
local kubernetes/kubernetes repository, and then submit a pull request to the master branch of
github.com/kubernetes/kubernetes.
PR 56673
is an example of a pull request that fixes a typo in the kubectl source code.
Monitor your pull request, and respond to reviewer comments. Continue to monitor your
pull request until it is merged into the target branch of the kubernetes/kubernetes repository.
Cherry picking your change into a release branch
Your change is now in the master branch, which is used for development of the next
Kubernetes release. If you want your change to appear in the docs for a Kubernetes
version that has already been released, you need to propose that your change be
cherry picked into the release branch.
For example, suppose the master branch is being used to develop Kubernetes
1.25
and you want to backport your change to the release-1.24 branch. For
instructions on how to do this, see
Propose a Cherry Pick.
Monitor your cherry-pick pull request until it is merged into the release branch.
Note: Proposing a cherry pick requires that you have permission to set a label and a
milestone in your pull request. If you don't have those permissions, you will
need to work with someone who can set the label and milestone for you.
Setting build variables
Go to <rdocs-base>. On you command line, set the following environment variables.
Set K8S_ROOT to <k8s-base>.
Set K8S_WEBROOT to <web-base>.
Set K8S_RELEASE to the version of the docs you want to build.
For example, if you want to build docs for Kubernetes 1.24, set K8S_RELEASE to 1.24.
The createversiondirs build target creates a versioned directory
and copies the kubectl reference configuration files to the versioned directory.
The versioned directory name follows the pattern of v<major>_<minor>.
In the <rdocs-base> directory, run the following build target:
cd <rdocs-base>
make createversiondirs
Checking out a release tag in k8s.io/kubernetes
In your local <k8s-base> repository, checkout the branch that has
the version of Kubernetes that you want to document. For example, if you want
to generate docs for Kubernetes 1.24.0, check out the
v1.24 tag. Make sure
you local branch is up to date.
cd <k8s-base>
git checkout v1.24.0
git pull https://github.com/kubernetes/kubernetes v1.24.0
Running the doc generation code
In your local <rdocs-base>, run the copycli build target. The command runs as root:
cd <rdocs-base>
make copycli
The copycli command cleans the temporary build directory, generates the kubectl command files,
and copies the collated kubectl command reference HTML page and assets to <web-base>.
Adding and committing changes in kubernetes/website
Run git add and git commit to commit the files.
Creating a pull request
Create a pull request to the kubernetes/website repository. Monitor your
pull request, and respond to review comments as needed. Continue to monitor
your pull request until it is merged.
A few minutes after your pull request is merged, your updated reference
topics will be visible in the
published documentation.
Docker (Required only for kubectl command reference)
Your PATH environment variable must include the required build tools, such as the Go binary and python.
You need to know how to create a pull request to a GitHub repository.
This involves creating your own fork of the repository. For more
information, see Work from a local clone.
7.8 - Advanced contributing
This page assumes that you understand how to
contribute to new content and
review others' work, and are ready
to learn about more ways to contribute. You need to use the Git command line
client and other tools for some of these tasks.
After you've been contributing to the Kubernetes documentation for a while, you
may have ideas for improving the Style Guide
, the Content Guide, the toolchain used to build
the documentation, the website style, the processes for reviewing and merging
pull requests, or other aspects of the documentation. For maximum transparency,
these types of proposals need to be discussed in a SIG Docs meeting or on the
kubernetes-sig-docs mailing list.
In addition, it can help to have some context about the way things
currently work and why past decisions have been made before proposing sweeping
changes. The quickest way to get answers to questions about how the documentation
currently works is to ask in the #sig-docs Slack channel on
kubernetes.slack.com
After the discussion has taken place and the SIG is in agreement about the desired
outcome, you can work on the proposed changes in the way that is the most
appropriate. For instance, an update to the style guide or the website's
functionality might involve opening a pull request, while a change related to
documentation testing might involve working with sig-testing.
Coordinate docs for a Kubernetes release
SIG Docs approvers
can coordinate docs for a Kubernetes release.
Each Kubernetes release is coordinated by a team of people participating in the
sig-release Special Interest Group (SIG). Others on the release team for a given
release include an overall release lead, as well as representatives from
sig-testing and others. To find out more about Kubernetes release processes,
refer to
https://github.com/kubernetes/sig-release.
The SIG Docs representative for a given release coordinates the following tasks:
Monitor the feature-tracking spreadsheet for new or changed features with an
impact on documentation. If the documentation for a given feature won't be ready
for the release, the feature may not be allowed to go into the release.
Attend sig-release meetings regularly and give updates on the status of the
docs for the release.
Review and copyedit feature documentation drafted by the SIG responsible for
implementing the feature.
Merge release-related pull requests and maintain the Git feature branch for
the release.
Mentor other SIG Docs contributors who want to learn how to do this role in
the future. This is known as "shadowing".
Publish the documentation changes related to the release when the release
artifacts are published.
Coordinating a release is typically a 3-4 month commitment, and the duty is
rotated among SIG Docs approvers.
Serve as a New Contributor Ambassador
SIG Docs approvers
can serve as New Contributor Ambassadors.
New Contributor Ambassadors welcome new contributors to SIG-Docs,
suggest PRs to new contributors, and mentor new contributors through their first
few PR submissions.
Responsibilities for New Contributor Ambassadors include:
After a new contributor has successfully submitted 5 substantive pull requests
to one or more Kubernetes repositories, they are eligible to apply for
membership
in the Kubernetes organization. The contributor's membership needs to be
backed by two sponsors who are already reviewers.
New docs contributors can request sponsors by asking in the #sig-docs channel
on the Kubernetes Slack instance or on the
SIG Docs mailing list.
If you feel confident about the applicant's work, you volunteer to sponsor them.
When they submit their membership application, reply to the application with a
"+1" and include details about why you think the applicant is a good fit for
membership in the Kubernetes organization.
Serve as a SIG Co-chair
SIG Docs members
can serve a term as a co-chair of SIG Docs.
Prerequisites
A Kubernetes member must meet the following requirements to be a co-chair:
Understand SIG Docs workflows and tooling: git, Hugo, localization, blog subproject
Approved by the SIG Docs community either directly or via lazy consensus.
Commit at least 5 hours per week (and often more) to the role for a minimum of 6 months
Responsibilities
The role of co-chair is one of service: co-chairs build contributor capacity, handle process and policy, schedule and run meetings, schedule PR wranglers, advocate for docs in the Kubernetes community, make sure that docs succeed in Kubernetes release cycles, and keep SIG Docs focused on effective priorities.
Responsibilities include:
Keep SIG Docs focused on maximizing developer happiness through excellent documentation
This dashboard is built using Google Data Studio and shows information collected on kubernetes.io using Google Analytics.
Using the dashboard
By default, the dashboard shows all collected analytics for the past 30 days. Use the date selector to see data from a different date range. Other filtering options allow you to view data based on user location, the device used to access the site, the translation of the docs used, and more.
If you notice an issue with this dashboard, or would like to request any improvements, please open an issue.
8 - Docs smoke test page
This page serves two purposes:
Demonstrate how the Kubernetes documentation uses Markdown
Provide a "smoke test" document we can use to test HTML, CSS, and template
changes that affect the overall documentation.
Heading levels
The above heading is an H2. The page title renders as an H1. The following
sections show H3-H6.
H3
This is in an H3 section.
H4
This is in an H4 section.
H5
This is in an H5 section.
H6
This is in an H6 section.
Inline elements
Inline elements show up within the text of paragraph, list item, admonition, or
other block-level element.
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor
incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis
nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu
fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in
culpa qui officia deserunt mollit anim id est laborum.
Inline text styles
bold
italic
bold italic
strikethrough
underline
underline italic
underline bold
underline bold italic
monospace text
monospace bold
Lists
Markdown doesn't have strict rules about how to process lists. When we moved
from Jekyll to Hugo, we broke some lists. To fix them, keep the following in
mind:
Make sure you indent sub-list items 2 spaces.
To end a list and start another, you need a HTML comment block on a new line
between the lists, flush with the left-hand border. The first list won't end
otherwise, no matter how many blank lines you put between it and the second.
Bullet lists
This is a list item
This is another list item in the same list
You can mix - and *
To make a sub-item, indent two spaces.
This is a sub-sub-item. Indent two more spaces.
Another sub-item.
This is a new list. With Hugo, you need to use a HTML comment to separate two
consecutive lists. The HTML comment needs to be at the left margin.
Bullet lists can have paragraphs or block elements within them.
Indent the content to be the same as the first line of the bullet point.
This paragraph and the code block line up with the first B in Bullet
above.
ls -l
And a sub-list after some block-level content
A bullet list item can contain a numbered list.
Numbered sub-list item 1
Numbered sub-list item 2
Numbered lists
This is a list item
This is another list item in the same list. The number you use in Markdown
does not necessarily correlate to the number in the final output. By
convention, we keep them in sync.
Note: For single-digit numbered lists, using two spaces after the period makes
interior block-level content line up better along tab-stops.
This is a new list. With Hugo, you need to use a HTML comment to separate
two consecutive lists. The HTML comment needs to be at the left margin.
Numbered lists can have paragraphs or block elements within them.
Indent the content to be the same as the first line of the bullet
point. This paragraph and the code block line up with the N in
Numbered above.
ls -l
And a sub-list after some block-level content. This is at the same
"level" as the paragraph and code block above, despite being indented
more.
Tab lists
Tab lists can be used to conditionally display content, e.g., when multiple
options must be documented that require distinct instructions or context.
Warning: Headers within tab lists will not appear in the Table of Contents.
Checklists
Checklists are technically bullet lists, but the bullets are suppressed by CSS.
This is a checklist item
This is a selected checklist item
Code blocks
You can create code blocks two different ways by surrounding the code block with
three back-tick characters on lines before and after the code block. Only use
back-ticks (code fences) for code blocks. This allows you to specify the
language of the enclosed code, which enables syntax highlighting. It is also more
predictable than using indentation.
this is a code block created by back-ticks
The back-tick method has some advantages.
It works nearly every time
It is more compact when viewing the source code.
It allows you to specify what language the code block is in, for syntax
highlighting.
It has a definite ending. Sometimes, the indentation method breaks with
languages where spacing is significant, like Python or YAML.
To specify the language for the code block, put it directly after the first
grouping of back-ticks:
ls -l
Common languages used in Kubernetes documentation code blocks include:
bash / shell (both work the same)
go
json
yaml
xml
none (disables syntax highlighting for the block)
Code blocks containing Hugo shortcodes
To show raw Hugo shortcodes as in the above example and prevent Hugo
from interpreting them, use C-style comments directly after the < and before
the > characters. The following example illustrates this (view the Markdown
source for this page).
To format an image, use similar syntax to links, but add a leading !
character. The square brackets contain the image's alt text. Try to always use
alt text so that people using screen readers can get some benefit from the
image.
To specify extended attributes, such as width, title, caption, etc, use the
figure shortcode,
which is preferred to using a HTML <img> tag. Also, if you need the image to
also be a hyperlink, use the link attribute, rather than wrapping the whole
figure in Markdown link syntax as shown below.
Pencil icon
Image used to illustrate the figure shortcode
Even if you choose not to use the figure shortcode, an image can also be a link. This
time the pencil icon links to the Kubernetes website. Outer square brackets enclose
the entire image tag, and the link target is in the parentheses at the end.
You can also use HTML for images, but it is not preferred.
Tables
Simple tables have one row per line, and columns are separated by |
characters. The header is separated from the body by cells containing nothing
but at least three - characters. For ease of maintenance, try to keep all the
cell separators even, even if you heed to use extra space.
Heading cell 1
Heading cell 2
Body cell 1
Body cell 2
The header is optional. Any text separated by | will render as a table.
Markdown tables have a hard time with block-level elements within cells, such as
list items, code blocks, or multiple paragraphs. For complex or very wide
tables, use HTML instead.
{{< mermaid >}}
sequenceDiagram
Alice ->> Bob: Hello Bob, how are you?
Bob-->>John: How about you John?
Bob--x Alice: I am good thanks!
Bob-x John: I am good thanks!
Note right of John: Bob thinks a long<br/>long time, so long<br/>that the text does<br/>not fit on a row.
Bob-->Alice: Checking with John...
Alice->John: Yes... John, how are you?
{{</ mermaid >}}
Produces:
sequenceDiagram
Alice ->> Bob: Hello Bob, how are you?
Bob-->>John: How about you John?
Bob--x Alice: I am good thanks!
Bob-x John: I am good thanks!
Note right of John: Bob thinks a long long time, so long that the text does not fit on a row.
Bob-->Alice: Checking with John...
Alice->John: Yes... John, how are you?
Sidebars and admonitions provide ways to add visual importance to text. Use
them sparingly.
Sidebars
A sidebar offsets text visually, but without the visual prominence of
admonitions.
This is a sidebar.
You can have paragraphs and block-level elements within a sidebar.
You can even have code blocks.
sudo dmesg
Admonitions
Admonitions (notes, warnings, etc) use Hugo shortcodes.
Note:
Notes catch the reader's attention without a sense of urgency.
You can have multiple paragraphs and block-level elements inside an admonition.
| Or | a | table |
Caution: The reader should proceed with caution.
Warning: Warnings point out something that could cause harm if ignored.
Includes
To add shortcodes to includes.
Note:
You need to have a Kubernetes cluster, and the kubectl command-line tool must
be configured to communicate with your cluster. It is recommended to run this tutorial on a cluster with at least two nodes that are not acting as control plane hosts. If you do not already have a
cluster, you can create one by using
minikube
or you can use one of these Kubernetes playgrounds:
This page shows how to install the 
.
