PVCs are hanging in termination due to the PV's ONTAP volume not being deleted
Applies to
Trident for Kubernetes/Openshift
Issue
- After having created a Kubernetes/Trident Snapshot and from that, a Kubernetes PVC clone to be used for another pod, terminating the original pod, and deleting its PVC does not fully clean up the PV information in Trident.
- The deletion fails due to the clone-busy Snapshot showing at least two errors:
type: 'Warning' reason: 'VolumeFailedDelete' persistentvolume pvc-UUID is still attached to node WORKER_NODE
API status: failed, Reason: Failed to delete volume \"trident_PREFIX_pvc_UUID\" in Vserver \"VSERVER\" because it has one or more clones.\n\nUse the \"volume clone show -parent-vserver VSERVER -parent-volume trident_PREFIX_pvc_UUID\" command to list clones, and either delete or split the clones and retry the operation.\n\nUse \"volume clone split start -vserver \u003cvserver name\u003e -flexclone \u003cclone name\u003e\" to split clones.\nUse \"volume delete -vserver \u003cvserver name\u003e -volume \u003cclone name\u003e\" to delete clones., Code 15894","level":"error","message":"Unable to delete volume from backend."
- Additionally, the PVCs show as released in the output "
oc|kubectl get pvc -o wide
".