TRID-20819: Snapshot deletion blocked by hidden offline FlexClone on NetApp ASAr2 storage with Trident
Issue
- After deleting a PVC or clone PVC, deletion of the associated Kubernetes VolumeSnapshot/VolumeSnapshotContent fails indefinitely.
- The following entries are seen:
- ONTAP error code
1638555:
level=debug msg="Could not delete the snapshot, going to check if it's busy; error was: API State: failure, Message: Snapshot \"snapshot-xxxx\" of volume \"trident_pvc_xxxx\" in Vserver \"<SVM>\" has not expired or is locked. Use the \"snapshot show -fields owners, expiry-time, snaplock-expiry-time\" command to view the expiry and lock status of the snapshot., Code: 1638555" logLayer=core requestID=xxxx requestSource=CSI workflow="snapshot=delete"snapshot show -fields ownersreports the snapshot owner asvolumeclone.- The deleted clone appears in
volume clone showasoffline, but is absent fromvolume showandstorage-unit show. - Manual attempts to delete or purge the clone fail:
volume clone deletereturnsError: 'delete' is not a recognized command- The volume recovery queue and volume clone split are not available on ASA r2 systems.
- Trident controller logs show repeated split timer resets:
level=warning msg="Initial locked snapshot delete, starting clone split timer." logLayer=core requestID=xxxx requestSource=CSI secondsBeforeSplit=120.00 snapshot=pvc-xxxx/snapshot-xxxx workflow="snapshot=delete"- Periodic backend reconcile triggers every 300–550 seconds, resetting the in-memory split timer and preventing the split from ever completing.
