Fix: can’t lock file ‘/var/lock/qemu-server/lock-VMID.conf’ – got timeout

An action on a VM, such as qm stop, a start, a config change or a backup, fails after a while with a message like this (here for VM 104):

can't lock file '/var/lock/qemu-server/lock-104.conf' - got timeout

What it means

Every task that changes a VM first takes a short-lived file lock, /var/lock/qemu-server/lock-<vmid>.conf, on the node. The error says your task waited for that lock and gave up because another process held it the whole time. This is a different mechanism from the lock: line in the VM config (as in “VM is locked (backup)”), so qm unlock usually does not help here.

Likely causes

1. Another task on the same VM is still running: a start, shutdown, backup, snapshot or migration. A guest shutdown can wait a long time if the guest OS ignores the request.
2. A task is hung, for example a start that waits forever for storage, or an external backup tool’s task that never finished.
3. Cluster or storage problems that make operations stall, such as lost quorum or an unreachable network storage.

Fix it step by step

1. Look for running tasks for that VM in the task list at the bottom of the web UI, or on the node: pvenode task list --source active.
2. If a legitimate task is running (for example a backup or a slow shutdown), wait for it to finish, or stop it from the task viewer if you are sure that is safe. Then retry.
3. If no task appears but the error continues, see which process holds the lock file: fuser -v /var/lock/qemu-server/lock-104.conf or lsof /var/lock/qemu-server/lock-104.conf, if those tools are installed, then inspect it with ps -fp <PID>.
4. Check the logs around the time of the stuck task: journalctl -u pvedaemon --since "1 hour ago".
5. Check cluster and storage health: pvecm status and pvesm status. Fix any lost quorum or inactive storage first.
6. As a last resort, users report that restarting the API daemon clears stuck worker tasks: systemctl restart pvedaemon. Running guests keep running, but tasks managed by that daemon are interrupted, so check the task list first.

pvenode task list --source active
fuser -v /var/lock/qemu-server/lock-104.conf
ps -fp <PID>
journalctl -u pvedaemon --since "1 hour ago"
pvecm status
pvesm status

Do not delete lock files by hand. The lock exists to stop two tasks from changing the same VM at once; removing it while a task still runs can corrupt the configuration or disks.

If it keeps coming back

• Look for overlapping schedules, such as backups, snapshots or external tools acting on the same VM at the same time, and spread them out.
• Install and enable the QEMU guest agent so shutdowns and backups complete reliably.
• Investigate slow or unreliable storage, which makes every task take longer and hold the lock longer.

Official documentation

• pvenode(1): task list and task status
• Proxmox forum: can’t lock file … got timeout