Updating an Amazon EKS cluster Kubernetes version
When a new Kubernetes version is available in Amazon EKS, you can update your Amazon EKS cluster to the latest version.
We recommend that, before you update to a new Kubernetes version, you review the
information in Amazon EKS Kubernetes versions
and also review in the update steps in this topic. If you're updating to version 1.22,
you must make the changes listed in Kubernetes version 1.22 prerequisites to your cluster before
updating it.
New Kubernetes versions sometimes introduce significant changes. Therefore, we recommend that you test the behavior of your applications against a new Kubernetes version before you update your production clusters. You can do this by building a continuous integration workflow to test your application behavior before moving to a new Kubernetes version.
The update process consists of Amazon EKS launching new API server nodes with the updated Kubernetes version to replace the existing ones. Amazon EKS performs standard infrastructure and readiness health checks for network traffic on these new nodes to verify that they're working as expected. If any of these checks fail, Amazon EKS reverts the infrastructure deployment, and your cluster remains on the prior Kubernetes version. Running applications aren't affected, and your cluster is never left in a non-deterministic or unrecoverable state. Amazon EKS regularly backs up all managed clusters, and mechanisms exist to recover clusters if necessary. We're constantly evaluating and improving our Kubernetes infrastructure management processes.
To update the cluster, Amazon EKS requires up to five free IP addresses from the subnets that you specified when you created your cluster. Amazon EKS creates new cluster elastic network interfaces (network interfaces) in any of the subnets that you specified. The network interfaces may be created in different subnets than your existing network interfaces are in, so make sure that your security group rules allow required cluster communication for any of the subnets that you specified when you created your cluster. If any of the subnets that you specified when you created the cluster don't exist, don't have enough free IP addresses, or don't have security group rules that allows necessary cluster communication, then the update can fail.
Even though Amazon EKS runs a highly available control plane, you might experience minor service interruptions during an update. For example, assume that you attempt to connect to an API server around when it's terminated and replaced by a new API server that's running the new version of Kubernetes. You might experience API call errors or connectivity issues. If this happens, retry your API operations until they succeed.
Update the Kubernetes version for your Amazon EKS cluster
To update the Kubernetes version for your cluster
-
Compare the Kubernetes version of your cluster control plane to the Kubernetes version of your nodes.
-
Get the Kubernetes version of your cluster control plane with the
kubectl version --shortcommand.kubectl version --short -
Get the Kubernetes version of your nodes with the
kubectl get nodescommand. This command returns all self-managed and managed Amazon EC2 and Fargate nodes. Each Fargate pod is listed as its own node.kubectl get nodes
Before updating your control plane to a new Kubernetes version, make sure that the Kubernetes minor version of both the managed nodes and Fargate nodes in your cluster are the same as your control plane's version. For example, if your control plane is running version
1.21and one of your nodes is running version1.20, then you must update your nodes to version1.21. We also recommend that you update your self-managed nodes to the same version as your control plane before updating the control plane. For more information, see Updating a managed node group and Self-managed node updates. To update the version of a Fargate node, first delete the pod that's represented by the node. Then update your control plane. Any remaining pods will update to the new version after you redeploy them. -
-
By default, the pod security policy admission controller is enabled on Amazon EKS clusters. Before updating your cluster, ensure that the proper pod security policies are in place. This is to avoid potential security issues. You can check for the default policy with the
kubectl get psp eks.privilegedcommand.kubectl get psp eks.privilegedIf you receive the following error, see default pod security policy before proceeding.
Error from server (NotFound): podsecuritypolicies.extensions "eks.privileged" not found If the Kubernetes version that you originally deployed your cluster with was Kubernetes
1.18or later, skip this step.You might need to remove a discontinued term from your CoreDNS manifest.
-
Check to see if your CoreDNS manifest has a line that only has the word
upstream.kubectl get configmap coredns -n kube-system -o jsonpath='{$.data.Corefile}' | grep upstreamIf no output is returned, this means that your manifest doesn't have the line. If this is the case, skip to the next step. If the word
upstreamis returned, remove the line. -
Remove the line near the top of the file that only has the word
upstreamin the configmap file. Don't change anything else in the file. After the line is removed, save the changes.kubectl edit configmap coredns -n kube-system -o yaml
-
-
Update your cluster using
eksctl, the AWS Management Console, or the AWS CLI.Important -
If you're updating to version
1.22, you must make the changes listed in Kubernetes version 1.22 prerequisites to your cluster before updating it. -
Because Amazon EKS runs a highly available control plane, you can update only one minor version at a time. For more information about this requirement, see Kubernetes Version and Version Skew Support Policy
. Assume that your current version is 1.20and you want to update to 1.22. Then, you must first update your cluster to1.21and then later update it from1.21to1.22. -
Make sure that the
kubeleton your managed and Fargate nodes are at the same Kubernetes version as your control plane before you update. We recommend that your self-managed nodes are at the same version as the control plane. They can be only up to one version behind the current version of the control plane. -
If your cluster is configured with a version of the Amazon VPC CNI plugin that is earlier than
1.8.0, then we recommend that you update the plugin to version1.11.2before updating your cluster to version1.21or later. For more information, see Updating the Amazon VPC CNI plugin for Kubernetes add-on or Updating the Amazon VPC CNI plugin for Kubernetes self-managed add-on.
-
-
After your cluster update is complete, update your nodes to the same Kubernetes minor version as your updated cluster. For more information, see Self-managed node updates and Updating a managed node group. Any new pods that are launched on Fargate have a
kubeletversion that matches your cluster version. Existing Fargate pods aren't changed. -
(Optional) If you deployed the Kubernetes Cluster Autoscaler to your cluster before updating the cluster, update the Cluster Autoscaler to the latest version that matches the Kubernetes major and minor version that you updated to.
-
Open the Cluster Autoscaler releases
page in a web browser and find the latest Cluster Autoscaler version that matches your cluster's Kubernetes major and minor version. For example, if your cluster's Kubernetes version is 1.22find the latest Cluster Autoscaler release that begins with1.22. Record the semantic version number (<1.22.n>) for that release to use in the next step. -
Set the Cluster Autoscaler image tag to the version that you recorded in the previous step with the following command. If necessary, replace
with your own value.1.22.nkubectl -n kube-system set image deployment.apps/cluster-autoscaler cluster-autoscaler=k8s.gcr.io/autoscaling/cluster-autoscaler:v1.22.n
-
-
(Clusters with GPU nodes only) If your cluster has node groups with GPU support (for example,
p3.2xlarge), you must update the NVIDIA device plugin for KubernetesDaemonSet on your cluster with the following command. kubectl apply -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/v0.9.0/nvidia-device-plugin.yml -
Update the Amazon VPC CNI plugin for Kubernetes, CoreDNS, and
kube-proxyadd-ons. If you updated your cluster to version1.21or later, than we recommend updating the add-ons to the minimum versions listed in Service account tokens.-
If you updated your cluster to version
1.18, you can add Amazon EKS add-ons. For instructions, see Adding the Amazon VPC CNI Amazon EKS add-on, Adding the CoreDNS Amazon EKS add-on, and Adding the kube-proxy Amazon EKS add-on. To learn more about Amazon EKS add-ons, see Amazon EKS add-ons. -
If you updated to version
1.19or later and are using Amazon EKS add-ons, in the Amazon EKS console, select Clusters, then select the name of the cluster that you updated in the left navigation pane. Notifications appear in the console. They inform you that a new version is available for each addon that has an available update. To update an add-on, select the Add-ons tab. In one of the boxes for an add-on that has an update available, select Update now, select an available version, and then select Update. -
Alternately, you can use the AWS CLI or
eksctlto update the Amazon VPC CNI plugin for Kubernetes, CoreDNS, and kube-proxy Amazon EKS add-ons.
-
Kubernetes version 1.22 prerequisites
A number of deprecated beta APIs (v1beta1) have been removed in
version 1.22 in favor of the GA (v1) version of those same
APIs. As noted in the Kubernetes version 1.22 API and
Feature removal blog1.22.
Before updating your cluster to Kubernetes version 1.22, make sure to
do the following:
-
Change your YAML manifest files and clients to reference the new APIs.
-
Update custom integrations and controllers to call the new APIs.
-
Make sure that you use an updated version of any third-party tools. These tools include ingress controllers, service mesh controllers, continuous delivery systems, and other tools that call the new APIs. To check for discontinued API usage in your cluster, enable audit control plane logging and specify
v1betaas an event filter. Replacement APIs are available in Kubernetes for several versions. -
If you currently have the AWS Load Balancer Controller deployed to your cluster, you must update it to version
2.4.1before updating your cluster to Kubernetes version1.22.
When you update clusters to version 1.22, existing persisted
objects can be accessed using the new APIs. However, you must migrate manifests
and update clients to use these new APIs. Updating the clusters prevents
potential workload failures.
Kubernetes version 1.22 removes support from the following beta APIs.
Migrate your manifests and API clients based on the following information:
| Resource | Beta version | GA version | Notes |
|---|---|---|---|
ValidatingWebhookConfiguration
MutatingWebhookConfiguration |
admissionregistration.k8s.io/v1beta1 |
admissionregistration.k8s.io/v1 |
|
CustomResourceDefinition |
apiextensions.k8s.io/v1beta1 |
apiextensions.k8s.io/v1 |
|
APIService |
apiregistration.k8s.io/v1beta1 |
apiregistration.k8s.io/v1 |
None |
TokenReview |
authentication.k8s.io/v1beta1 |
authentication.k8s.io/v1 |
None |
SubjectAccessReview LocalSubjectAccessReview
SelfSubjectAccessReview |
authorization.k8s.io/v1beta1 |
authorization.k8s.io/v1 |
spec.group is renamed to
spec.groups |
CertificateSigningRequest |
certificates.k8s.io/v1beta1 |
certificates.k8s.io/v1 |
|
|
|
coordination.k8s.io/v1beta1 |
coordination.k8s.io/v1 |
None |
|
|
|
networking.k8s.io/v1 |
|
|
|
networking.k8s.io/v1beta1 |
networking.k8s.io/v1 |
None |
|
|
rbac.authorization.k8s.io/v1beta1 |
rbac.authorization.k8s.io/v1 |
None |
PriorityClass |
scheduling.k8s.io/v1beta1 |
scheduling.k8s.io/v1 |
None |
CSIDriver CSINode StorageClass
VolumeAttachment |
storage.k8s.io/v1beta1 |
storage.k8s.io/v1 |
None |
To learn more about the API removal, see the Deprecated API migration guide

