While investigating a problem from a customer case, we noticed that in some environments velero was failing to create podvolumerestore resources from restore process, which was breaking restic restore of volumes.

After some digging, it turned out to be a niche bug in how Velero handles restorations of pods with labels set by plugins, which in our case the kubevirt plugin that we enabled by default was affecting. Because of this, when a user has the kubevirt plugin installed and tries to use restic restores, the restore was not working as expected without any error to the user.

For QE verification:

Set `kubevirt` in the defaultPlugin list, and run a restic restore workload and confirm that the data in the volume is restored properly.