Upstream Issue: https://github.com/vmware-tanzu/velero/issues/9344

Problem Description:

VolumeGroupSnapshots tested with Velero v1.17.0 does not respect volumePolicies. When using a common label (e.g., app.kubernetes.io/instance) for volume group snapshots, PVCs that should be excluded by volumePolicy (such as NFS PVCs) are still included in the grouping logic, causing backup failures.

Reproduction:

Using a volumePolicy to skip NFS volumes:

version: v1
volumePolicies:
* conditions:
    nfs: {}
  action:
    type: skip

And a backup configured with VolumeGroupSnapshot label:

spec:   resourcePolicy:     kind: configmap
    name: velero-volume-policies
  volumeGroupSnapshotLabelKey: app.kubernetes.io/instance

Error:

Errors:
  Velero:    name: /data-mysql-0 message: /Error backing up item error: /error executing custom action (groupResource=persistentvolumeclaims, namespace=group-test, name=data-mysql-0): rpc error: code = Unknown desc = failed to determine CSI driver for PVCs in VolumeGroupSnapshot group "group-test": found multiple CSI drivers: linstor.csi.linbit.com and nfs.csi.k8s.io

Expected Behavior:

1. 1. Velero should respect the volumePolicy and skip the NFS PVC (which works correctly for non-group snapshots)
      2. If a group snapshot is not possible, it should gracefully fall back to a normal backup instead of skipping the backup of the PVC

Upstream Status:

• State: Open
• Assignee: shubham-pampattiwar
• Labels: VolumeGroupSnapshot, 1.18-candidate
• Velero Version: v1.17.0