How to replacing ECK data node with minimal downtime/data moving?

we running an ES cluster using ECK installed on AWS EKS, with each ES instance per node.
We often need to update the underlying AWS worker nodes for many scenarios:

  • fixing runC CVE-2024-21626
  • upgrade K8S version - which happens every 5-6 months

How to replacing k8s node with minimal data migrate to other nodes (we have many TB) ?

Is there a way to create new node, then replacing the old pod on old node with new pod on new node, and detach VPC from old node, attach the existing VPC? If that is possible then we can afford having down time like < 1 min per node.

PS: we using karpenter to manage node, if that matters.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.