Created attachment 1715070 [details]
Description of problem:
When a PVC is migrated using Copy -> Snapshot, the migration fails.
Version-Release number of selected component (if applicable):
SOURCE CLUSTER: OCP 4.2
TARGET CLUSTER: OCP 4.5
REPLICATION REPOSITORY: AZURE
Steps to Reproduce:
1. Create a namespace with an application that uses PVCs
oc new-project bztest
oc new-app mysql-persistent
2. Create a migration plan, and chose "Copy" and "Volume snapshot" method when configuring the PVC migration.
The migration fails. We can find no erros in the restore logs. This is the status of the PVCs in the target cluster after the failure
$ oc get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
mysql Lost pvc-c2cc0043-f767-11ea-b3f5-0022488efde9 0 managed-premium 4m31s
We can see in the restore log that the volume was skipped.
time="2020-09-15T15:30:17Z" level=info msg="Adding PV pvc-c2cc0043-f767-11ea-b3f5-0022488efde9 as an additional item to restore" cmd=/velero logSource="pkg/restore/add_pv_from_pvc_action.go:66" pluginName=velero restore=openshift-migration/f8ce4020-f767-11ea-afb5-03fb410600cc-7cz8b
time="2020-09-15T15:30:17Z" level=info msg="Skipping persistentvolumes/pvc-c2cc0043-f767-11ea-b3f5-0022488efde9 because it's already been restored." logSource="pkg/restore/restore.go:844" restore=openshift-migration/f8ce4020-f767-11ea-afb5-03fb410600cc-7cz8b
The migration should have finished without errors and the PVC should be able to find the snapshot PV in the target cluster.
The volume can be found in the source cluster's resource group and in the replication repository's resource group. But not in the target cluster's resource group.
this is not a bug, because we missed AZURE_RESOURCE_GROUP