-
Bug
-
Resolution: Obsolete
-
Major
-
OADP 1.2
-
False
-
-
False
-
ToDo
-
-
-
Moderate
-
5
-
0
-
Very Likely
-
0
-
Customer Escalated, Customer Facing
-
None
-
Unset
-
Unknown
-
No
Description of problem:
- When deleting backups with 'velero' tool, volumeSnapshot and volumeSnapshotContents are not getting deleted from neither OCP nor storage backend
Version-Release number of selected component (if applicable):
- OCP 4.12.20
- OADP 1.2
- IBM VPC Block CSI driver 5.0.16_127
How reproducible:
- All the time on customer's environment
Steps to Reproduce:
1. Run a backup against a namespace which contains a PVC provisioned by the CSI driver but do not have a running Pod which uses it to avoid mounting the volume on a node to simulate the failure.
2. Backup will complete with Partially failed status which is expected since the volume was not able to be snapshoted due to CSI limitation [1], volsnap and volsnapcontent objects will be created.
3. Delete backup with 'velero backup delete <backup-name>' command
Actual results:
- volumeSnapshot and volumeSnapshotContent objects will remain in the OCP cluster.
- OADP won't delete these objects from the OCP cluster.
Expected results:
- Velero / OADP should be able to delete volumeSnapshot and volumeSnapshotContent objects when createSnapshot call fails and volsnap stays with ReadyToUse=false
Additional info:
[1] https://cloud.ibm.com/docs/openshift?topic=openshift-vpc-volume-snapshot#vpc-create-snapshot