Description of problem:

Backup or restore fails with the message:

found a backup with status "InProgress" during the server starting, mark it as "Failed"

Version-1.4.3

How reproducible:

When creating a OADP DPA for the first with the azure plugin specified the Velero deployment created will lack the mounted ConfigMap cloud-credentials-azure.

When a BSL is added to the spec.backupLocations field this will trigger the ConfigMap cloud-credentials-azure to be added as a volumeMount and volume regardless of type, including S3, regardless of whether or not the BSL.

This triggers a Velero deployment modification and Velero pod deletion and recreate with the new value.

If a Backup or Restore is created during the Velero Pod replacement the backup or restore may failure with .status.failureReason in the following format:

found a backup with status "InProgress" during the server starting, mark it as "Failed"

Steps to Reproduce:
1. Create a DPA with azure plugin
2. Add a backupLocation in the DPA.
3. Trigger a new Backup or Restore. Depending on the race condition timing the backup or restore will fail with the above error.

Actual results:

Operation may fail due to Velero pod restart.

Expected results:

Velero Deployment should be created with ConfigMap cloud-credentials-azure already added to the Velero Deployment if the azure plugin is included to minimize Velero Pod restarts.

Additional info:

Workaround is to retry.