The following sequence of events results in failure:

1. Have some template/golden image as PVC in namespace X
2. Allow cloning cross namespace (see additional info)
3. Create a new VM on namespace Y, using Clone PVC option from image from namespace X from step 1
4. Snapshot this VM
5. Delete namespace X
6. Restore the snapshot of the VM

The VirtualMachineRestore gets stuck, as it cannot create the DV anymore.

4.16.1

Always

1. Create an empty PVC named my-disk on a namespace called my-images$ cat disk.yaml 
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: my-disk 
  namespace: my-images
spec:
  accessModes:
    - ReadWriteOnce 
  resources:
    requests:
      storage: 10Gi 
  storageClassName: lvms-ssd

$ oc get pvc -n my-images
NAME      STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   VOLUMEATTRIBUTESCLASS   AGE
my-disk   Bound    pvc-839d1030-c5a7-4ee5-9dd3-4b2018cdcd1a   10Gi       RWO            lvms-ssd       <unset>                 57s

2. Ensure you can clone from my-images to another namespace where VMs are created

3. In the namespace where VMs are created (not my-images), create a new VM from template (I've used RHEL8, but not relevant)
 
 Disk Source: PVC (clone PVC)
 PVC Project: my-images
 PVC Name: my-disk The DV will look like this (I've created mine on homelab namespace)spec:
  preallocation: false
  source:
    pvc:
      name: my-disk
      namespace: my-images
  storage:
    resources:
      requests:
        storage: 10Gi
    storageClassName: lvms-ssd4. 

Ensure the VM created fine

5. In the Web Console, create a VM snapshot of the new VM

$ oc get vmsnapshot
NAME                         SOURCEKIND       SOURCENAME          PHASE       READYTOUSE   CREATIONTIME   ERROR
snapshot-cyan-cockroach-53   VirtualMachine   rhel8-aqua-asp-20   Succeeded   true         4s        

6. Now try to restore that snapshot, also in the Web Console

7. All works

8. Now delete the original my-images/my-disk (its not needed really, the VM is a clone of that)

$ oc delete pvc -n my-images my-disk 
persistentvolumeclaim "my-disk" deleted
$ oc delete project my-images
project.project.openshift.io "my-images" deleted

9. Try to restore the snapshot again, it got stuck here:

$ oc get virtualmachinerestore resotre-snapshot-cyan-cockroach-53-1724733464456 -o yaml
apiVersion: snapshot.kubevirt.io/v1alpha1
kind: VirtualMachineRestore
metadata:
  creationTimestamp: "2024-08-27T04:37:45Z"
  generation: 5
  name: resotre-snapshot-cyan-cockroach-53-1724733464456
  namespace: homelab
  ownerReferences:
  - apiVersion: kubevirt.io/v1
    blockOwnerDeletion: false
    kind: VirtualMachine
    name: rhel8-aqua-asp-20
    uid: 7b75bc2b-e13a-455e-8d9a-5abceb3c957d
  resourceVersion: "37572385"
  uid: d0468b34-488e-4404-8623-906f28d0f7d0
spec:
  target:
    apiGroup: kubevirt.io
    kind: VirtualMachine
    name: rhel8-aqua-asp-20
  virtualMachineSnapshotName: snapshot-cyan-cockroach-53
status:
  complete: false
  conditions:
  - lastProbeTime: null
    lastTransitionTime: "2024-08-27T04:37:45Z"
    reason: 'admission webhook "virtualmachine-validator.kubevirt.io" denied the request:
      namespace my-images does not exist'
    status: "False"
    type: Progressing
  - lastProbeTime: null
    lastTransitionTime: "2024-08-27T04:37:45Z"
    reason: 'admission webhook "virtualmachine-validator.kubevirt.io" denied the request:
      namespace my-images does not exist'
    status: "False"
    type: Ready
  deletedDataVolumes:
  - restore-f5b3cb99-ed07-4597-b875-25fdbfbcd79b-disk-chocolate-pelican-74
  restores:
  - dataVolumeName: restore-d0468b34-488e-4404-8623-906f28d0f7d0-disk-chocolate-pelican-74
    persistentVolumeClaim: restore-d0468b34-488e-4404-8623-906f28d0f7d0-disk-chocolate-pelican-74
    volumeName: disk-chocolate-pelican-74
    volumeSnapshotName: vmsnapshot-4e5c7c55-ca8e-4a31-ae74-fc25dec54073-volume-disk-chocolate-pelican-74

Now the user is unable to restore the VM from a backup as the original (now unrelated) namespace/pvc don't exist anymore 

Unable to restore snapshot of the VM using VirtualMachineRestore

Able to restore snapshot of VM

https://docs.openshift.com/container-platform/4.16/virt/storage/virt-enabling-user-permissions-to-clone-datavolumes.html

The customer (in 4.15) got different error but in the exact same place, as if the source is the problem:

Failed to create restore DataVolume: admission webhook "datavolume-validate.cdi.kubevirt.io" denied the request:  Data volume should have either Source or SourceRef, or be externally populated'

But also failed at the same step of creating the restore DV. But they are on 4.15, this is in latest 4.16.1.

This should work without a manual restore, as customers may need to urgently roll back their VMs.