Multiple VMs randomly give a warning during backup

[crawrj]

Hello All. Noob back with another odd issue. I get warnings like this, WARN: unable to check guest fs freeze status - VM 203 qga command 'guest-fsfreeze-status' failed - got timeout sporadically on a daily basis but not all the time or consistently. It can be different VMs at different times on different hosts and it has happened on two different installs of PBS, One original test VM and now on the dedicated host. The backups say they successfully complete. Is this anything I should be concerned with? If so, any ideas on what I should do? Sorry I am new to this so if you need something else, please let me know.

PVE is clustered on 9.2.10
Kernel is 7.0.14-12
virtio 0.1.271
VMs Server 2025 Standard *Current

PBS is 4.2.5
Kernel is 7.0.14-12

 

[longer]

Just my 2 cents, I might be wrong:
My understanding this means pve (i think specifically pvestatd/qmeventd) tries to access the QMP socket of the VM but it couldn't and timeout, this can occurs for many reasons, for example such as another service blocks the socket of qmp. I think as long as the backup finishes successfully it's not an issue.

 

[Chris]

Hello All. Noob back with another odd issue. I get warnings like this, WARN: unable to check guest fs freeze status - VM 203 qga command 'guest-fsfreeze-status' failed - got timeout sporadically on a daily basis but not all the time or consistently.

The guest-fsfreeze-status command is used to determine if a fsfreeze is still ongoing inside the guest VM. The timeout here show that the guest does not respond to the status requests send by the PVE host, which waits 10 minutes at most for this. Please check the systemd jorunal of the host from around the backup time and look especially for logs related to the qemu-guest-agent. That might tell why there was no response from the VM.

Is this anything I should be concerned with? If so, any ideas on what I should do? Sorry I am new to this so if you need something else, please let me know.

The fsfreeze assures that all data still buffered by the guest OS inside the VM is being persistet to the filesystem, which might take some time. This improves backup consistency for the VM by assuring all in-flight data is being written. If this does not happen, the backup snapshot might not be fully consistent, which however modern journaled filesystems can handle. So best is to find out why the freeze command fails, but backups without it will still be fine and restorable assuming a journaled filesystem.

 

[crawrj]

Thanks for the responses. The backups do complete and verify. I will check the logs to see if anything shows up, but since it is happening on multiple hosts, PBS servers, VMs, not all the time, and not consistently in any way, I am not sure what I could possibly do about it. It couldn't be one single issue with all of those things being true right?