[scanner] fix: add pre-deploy PV cleanup steps to deploy-vllm-d#20566
Conversation
…ploy-vllm-d The deploy-vllm-d job was missing five steps that deploy-pok-prod uses to clear stuck PVCs/PVs before attempting Helm release recovery. Without them, a stuck or orphaned PersistentVolume causes 'helm rollback' to fail (the release has no clean revision to roll back to), which blocks all subsequent deploy steps. Add the following steps, adapted to the kc namespace used by vllm-d: - Clean up stuck persistent volumes - Validate persistent volumes before deploy - Pre-deploy cluster health check - Clean up stuck pods - Wait for volume detachment Fixes #20548 Signed-off-by: scanner <scanner@kubestellar.io> Signed-off-by: scanner-bot <scanner-bot@kubestellar.io>
kubestellar-prow
Bot
added
the
dco-signoff: yes
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
✅ Deploy Preview for kubestellarconsole canceled.
|
|
👋 Hey @kubestellar-hive[bot] — thanks for opening this PR!
This is an automated message. |
kubestellar-prow
Bot
added
the
size/L
…over stuck Helm release When the kc Helm release is in 'failed' state with no good prior revision, 'helm rollback' itself fails and the Recover stuck Helm release step exits 1, blocking all subsequent deploy steps. Add a helm uninstall fallback so the next 'helm upgrade --install' step can perform a clean fresh install rather than stopping the pipeline. Applies the same fix to both deploy-vllm-d and deploy-pok-prod which share the identical recovery pattern. Fixes #20548 Signed-off-by: kubestellar-hive[bot] <kubestellar-hive[bot]@users.noreply.github.com>
|
Thank you for your contribution! Your PR has been merged. Check out what's new:
Stay connected: Slack #kubestellar-dev | Multi-Cluster Survey |
|
Post-merge build verification passed ✅ Both Go and frontend builds compiled successfully against merge commit |
Summary
The
deploy-vllm-djob was missing five pre-deploy steps thatdeploy-pok-prodalready uses to clear stuck PVCs/PVs before Helm release recovery. Without them, a stuck or orphaned PersistentVolume causeshelm rollbackto fail (no clean revision to roll back to), blocking all subsequent deploy steps on every main merge.Added to
deploy-vllm-d(adapted to thekcnamespace):Test plan
deploy-vllm-dproceeds past PV/PVC cleanup when the cluster is healthy (no-op path)Fixes #20548