- Newest
- Most votes
- Most comments
Hello there,
I have replicated this:
- Created the Fargate cluster version 1.26,
- Rolled out a deployment
- Upgraded cluster to 1.27
- Restarted the deployment using command : kubectl rollout restart deployment/<deployment_name> -n <namespace>
- All the nodes in the namespace updated to 1.27 (v1.27.1-eks-2f008fe)
Please try restarting your deployment in all namespaces.
Also from the Kubernetes documentation: https://kubernetes.io/releases/version-skew-policy/#kubelet
The kubelet version 1.26 should not cause any problems since kubelet is supported at 1.27, 1.26, and 1.25
If restarting the deployment does not solve the problem, we would require details that are non-public information. Please open a support case with AWS using the following link: https://support.console.aws.amazon.com/support/home#/case/create
Hi Thubelihle,
I should've followed up on this post, but yesterday, while investigating another issue, I did restart some deployments and noticed the kubelet version updated for those nodes. I restarted the rest of the deployments and all the nodes are now in sync. My guess is that there is a lag in the matching kubelet version being available for the fargate nodes.
I've yet to find any documentation regarding EKS + Fargate + Kubelet version rollouts, but in the future, we're going to wait to upgrade the cluster until we can confirm a matching kubelet version is available. We do understand that kubelet versions have a +/- 1 version support window, but there's no guarantee that future updates won't contain a breaking change that require them to be in sync.
Relevant content
- Accepted Answerasked 7 months ago
- asked 2 years ago
- Accepted Answerasked a year ago
- AWS OFFICIALUpdated 8 months ago
- AWS OFFICIALUpdated 8 months ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated a year ago