HA Resource Agent Triggers Shutdown on Stopped VM during Snapshot Backup, Causes "VM is locked (backup)" Error

[Herzen]

Proxmox VE Version: 9.2.4
Proxmox Backup Server: 4.1
HA Enabled: Yes, with VM set to "stopped"
Affected Component: HA Manager (pve-ha-crm/pve-ha-lrm) + Backup System (vzdump)

Issue description:
When backing up a powered-off VM with HA state set to stopped using snapshot mode to PBS, the HA resource agent intervenes DURING the backup process (not after), attempting to shut down the VM while it is still locked for backup, resulting in the error:

task started by HA resource agent
TASK ERROR: VM is locked (backup)

As you can see from the timestamps, the HA attempt to change the VM status happens during the backup job:

INFO:

Backup Job Log:

VM Name: cloneformigrate
INFO: include disk 'scsi0' 'DatastoreShared:vm-101-disk-0' 17G
INFO: creating Proxmox Backup Server archive 'vm/101/2026-10-06T00:03:49Z'
INFO: starting kvm to execute backup task
INFO: starting backup via QMP command
INFO: started backup task 'ecfe1d5c-71a6-4752-9bf3-fb936b8c64e0'
INFO: scsi0: dirty-bitmap status: created new
INFO:  13% (2.3 GiB of 17.0 GiB) in 3s, read: 794.7 MiB/s, write: 0 B/s
INFO:  24% (4.2 GiB of 17.0 GiB) in 6s, read: 654.7 MiB/s, write: 0 B/s
INFO:  37% (6.4 GiB of 17.0 GiB) in 9s, read: 733.3 MiB/s, write: 0 B/s
INFO:  51% (8.7 GiB of 17.0 GiB) in 12s, read: 798.7 MiB/s, write: 0 B/s
INFO:  60% (10.3 GiB of 17.0 GiB) in 15s, read: 520.0 MiB/s, write: 0 B/s
INFO:  69% (11.9 GiB of 17.0 GiB) in 18s, read: 550.7 MiB/s, write: 0 B/s
INFO:  80% (13.7 GiB of 17.0 GiB) in 21s, read: 613.3 MiB/s, write: 0 B/s
INFO:  92% (15.7 GiB of 17.0 GiB) in 24s, read: 690.7 MiB/s, write: 0 B/s
INFO: 100% (17.0 GiB of 17.0 GiB) in 27s, read: 446.7 MiB/s, write: 0 B/s
INFO: Waiting for server to finish backup validation...
INFO: backup is sparse: 3.46 GiB (20%) total zero data
INFO: backup was done incrementally, reused 17.00 GiB (100%)
INFO: transferred 17.00 GiB in 100 seconds (174.1 MiB/s)
INFO: stopping kvm after backup task
INFO: adding notes to backup
INFO: prune older backups with retention: keep-last=14, keep-monthly=1
INFO: running 'proxmox-backup-client prune' for 'vm/101'
INFO: pruned 1 backup(s) not covered by keep-retention policy
INFO: Finished Backup of VM 101 (00:01:55)
INFO: Backup finished at 2026-10-06 02:05:44

HA interaction log:

Start Time: 2026-10-06 02:04:04
End Time: 2026-10-06 02:04:04
task started by HA resource agent
TASK ERROR: VM is locked (backup)

 

[gfngfn256]

Assuming your backup is OK (which it appears to be), the task started by HA resource agent is to be expected as the backup starts the kvm, as shown in your log:

INFO: starting kvm to execute backup task

so the HA task gets activated as it sees the VM running.

Then the:

TASK ERROR: VM is locked (backup)

will also naturally occur (gracefully) as it is locked by the backup process & can be safely ignored.
I don't see an error here but rather graceful logging.

 

[Herzen]

I don't see any “error” either, but it might be a useful hint for the developer to add an ignore condition for this specific and fairly common case.

This behavior triggers a frightening, blood red alarm that could scare new users or cloud provider customers. I thought it might be useful to share this with the community because I couldn't find any related topics on the forum.