AWS launched support for Kubernetes 1.22 for Amazon EKS clusters in April 2022. Rafay customers can now “provision” EKS clusters based on k8s 1.22 and perform “in-place upgrades” of their existing EKS clusters to k8s 1.22.
In this demo, we will see the user experience for Ops/SRE personnel when they perform in-place upgrades of their EKS clusters to k8s 1.22
The “demo-eks-upgrade” EKS cluster is currently based on k8s 1.21 and has two worker nodes in a managed node group.
The administrator is notified that there is a new k8s version available and that they can perform an in-place upgrade of the EKS cluster and the node groups to k8s 1.22.
When the upgrade is initiated, Rafay performs a number of steps in a logical sequence. The first step is a “preflight check” where a number of tests are performed to verify that the cluster is in a state where the upgrade can be safely performed.
Once this step is completed, the second step is to upgrade the AWS managed k8s control plane to k8s 1.22.
Once this step is completed, the third step is to perform a rolling update of critical addons such as the AWS CNI etc in the EKS cluster to a version supported by k8s 1.22
Once this step is completed, the fourth step is to upgrade the worker nodes in the associated EKS node groups to k8s 1.22
Once this step is completed, Rafay performs the final step of verifying that all resources operational on the cluster are still operational and that there was not an inadvertent regression.
Administrators can also view the Kubernetes upgrade history associated with the cluster as it is kept current and up to date.
You can see from the console and zero trust kubectl commands that our EKS cluster has been successfully upgraded “in-place” from k8s v1.21 to v1.22.
Thank you for watching.
