-
Story
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
None
-
False
-
None
-
False
-
-
in ETCD-476 and OCPBUGS-18032 we found that the scheduler is fairly wild to schedule pods to different CP nodes when we do not supply a node selector.
While we don't want to support hostPath volumes, we should add the ability to ensure the retention runs on the same place as we eventually take the backup.
Additionally, this feature could be useful for some customers when taking one-off backups.
AC:
- API change to both periodic and one-off backup CRDs
- client-go update
- e2e test update