Remove kubevirt/kubevirt e2e workarounds once their upstream PRs merge

openshift/oadp-operator#2404's kdm e2e suite carries three workarounds for still-open kubevirt/kubevirt bugs. Each should be removed once its blocking PR merges and a release picks it up.

(migtools/kubevirt-datamover-controller#207, migtools/kubevirt-datamover-controller#208, and migtools/kubevirt-datamover-controller#212 are expected to merge shortly and are intentionally not tracked here.)

Workarounds to remove

1. Stuck VMB finalizer blocks namespace deletion forever (kubevirt/kubevirt#18724)

   • Workaround: tests/e2e/lib/virt_helpers.go's ClearStuckVMBFinalizers (line 1674) / IsNamespaceDeletedClearingStuckVMBFinalizers (line 1715), called from tests/e2e/backup_restore_suite_test.go's deleteNamespace (line 421)
   • Remove once: kubevirt/kubevirt#18725 merges. kubevirt/kubevirt#18289 alone is not sufficient -- it reduces how often the bug fires but leaves a timing/version-skew gap (confirmed by cross-check), so it complements kubevirt/kubevirt#18725 rather than replacing it.

2. VirtualMachineBackup status freeze during target PVC attach (CNV-85377/CNV-89684)

   • Workaround: tests/e2e/lib/virt_helpers.go's NudgeVmiToTriggerResync (line 1054) and VMBHasNoConditions (line 1646), called from tests/e2e/virt_backup_restore_suite_test.go (lines 476, 482, 510)
   • Remove once: kubevirt/kubevirt#18949 merges.

3. Stale reconcile can mark an already-completed VirtualMachineBackup Failed with SourceLost

   • Workaround: tests/e2e/lib/apps.go's GetNamespaceEventMessages (line 386), called from tests/e2e/virt_backup_restore_suite_test.go (lines 507, 530)
   • Remove once: kubevirt/kubevirt#18957 merges.

Not tracked here (already a permanent improvement, not removable scaffolding)

openshift/oadp-operator#2404's hyperConvergedGVR() (preferring HCO's v1 API over v1beta1) stays useful even after kubevirt/hyperconverged-cluster-operator#4552 merges, so it's not a candidate for removal.

[!Note]
Responses generated with Claude