Red Hat Bugzilla – Bug 1258659
[RFE] provide better monitoring to allow to know to which disk the VM is writing to after failed live snapshot.
Last modified: 2017-09-28 04:35:54 EDT
== Description of problem ==
A failed snapshot creation or LSM (Live Storage Migration) requires the VM to be restarted.
The logs display:
"Failed to create live snapshot 'Auto-generated for Live Storage Migration' for VM 'test-vm'. VM restart is recommended."
It's nice to have this bug showing this hint, but restarting a VM because of a temporary network problem while doing a LSM is not affordable in production.
== How reproducible ==
Easy to reproduce.
== Steps to Reproduce ==
1. Configure host to fail during snapshot creation (we need to reproduce a "Failed to create live snapshot 'Auto-generated for Live Storage Migration' for VM 'XXXXXX'. VM restart is recommended." warning).
2. Create a snapshot (process must fail)
3. Check snapshots and compare with virsh output
== Actual results ==
Current snapshot displayed inside oVirt is wrong.
Check real active snapshot via virsh.
Workarround is to restart the VM.
== Expected results ==
oVirt must show correct results, without requiring a VM restart.
== Additional info ==
When a live snapshost fails, we should query QEMU and adjust engine/vdsm's definition of the VM to match the "real" definition.