-
Bug
-
Resolution: Unresolved
-
Critical
-
CNV v4.17.0, CNV v4.16.1, CNV v4.15.0
-
1
-
False
-
-
False
-
CNV v4.99.0.rhel9-1288, CNV v4.15.7.rhel9-32
-
---
-
---
-
-
Storage Core Sprint 259, CNV Storage 260, Storage Core Sprint 261
-
None
Description of problem:
If there is a problem during VirtualMachineRestore, like the one described in https://issues.redhat.com/browse/CNV-47105, it is very difficult for the user to get the VM back. The system does not reconcile after deleting this object, the VM points to wrong disks and has a leftover .status.restoreInProgress
Version-Release number of selected component (if applicable):
4.16.1
How reproducible:
Always
Steps to Reproduce:
1. Get into a situation that a VirtualMachineRestore hangs and cannot progress. I'm using https://issues.redhat.com/browse/CNV-47105 as example $ oc get vmrestore resotre-snapshot-cyan-cockroach-53-1724733464456 NAME TARGETKIND TARGETNAME COMPLETE RESTORETIME ERROR resotre-snapshot-cyan-cockroach-53-1724733464456 VirtualMachine rhel8-aqua-asp-20 false 2. Delete the VirtualMachineRestore object Now we have 2 problems A) The VM was not changed back to use the volumes prior to the restore, its stuck pointing to new restore-xyz volumes that may or may not have been created. B) There is a leftover status that prevents VM start, requiring a manual patch $ oc get vm rhel8-aqua-asp-20 -o yaml | yq '.status.restoreInProgress' "resotre-snapshot-cyan-cockroach-53-1724733464456" $ virtctl start rhel8-aqua-asp-20 Error starting VirtualMachine Internal error occurred: admission webhook "virtualmachine-validator.kubevirt.io" denied the request: Cannot start VM until restore "resotre-snapshot-cyan-cockroach-53-1724733464456" completes The system should do better to recover and reconcile from such scenarios, without support intervention.
Actual results:
VMs unable to start
Expected results:
VMs able to abort restores and start on previous state
Additional info:
- clones
-
CNV-48787 [4.17] Unable to start VM after stuck/failed VirtualMachineRestore
- MODIFIED
- relates to
-
CNV-47105 Unable to restore a snapshot if the original DataVolume clone source is from a namespace/pvc that was deleted
- ON_QA
-
CNV-48692 [4.17] Unable to restore a snapshot if the original DataVolume clone source is from a namespace/pvc that was deleted
- ON_QA
-
CNV-48693 [4.15] Unable to restore a snapshot if the original DataVolume clone source is from a namespace/pvc that was deleted
- Verified
- links to
-
RHEA-2024:138842 OpenShift Virtualization 4.15.7 Images
- mentioned on