Skip to content

Force-delete wedged pod so Recreate finishes in one pass - #2080

Open
alex-zaitsev wants to merge 1 commit into
0.27.4from
fix/2078-recreate-force-delete-pod
Open

alex-zaitsev wants to merge 1 commit into
0.27.4from
fix/2078-recreate-force-delete-pod

Conversation

@alex-zaitsev

Copy link
Copy Markdown
Member

Summary

  • Honor the scale-to-0 wait in doDeleteStatefulSet. If a pod does not terminate in time, force-delete it and continue the existing Delete + Create path so Recreate completes in the same pass.
  • Fixes hosts left at Replicas=0 when a ClickHouse or Keeper pod hangs on shutdown (Recreate can take a whole shard (or Keeper quorum) down when a pod is slow to delete #2078).
  • Adds unit tests for the escalate path, pod NotFound, a real pod-delete failure, a successful wait, and Recreate still creating in the same pass.

… pass

A scale-to-0 wait timeout used to abort Recreate and leave the host at Replicas=0. Escalate by deleting the pod so Delete+Create can complete in the same reconcile.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant