GPU loss on backup

kizzmansky

New Member
May 25, 2024
4
0
1
Hi,

I Have successfully setup GPU (Nvidia RTX A2000) to work with LXC but having wildest issue that I cannot even begin to understand how to track down. Proxmox host (even without any running LXCs or VMs) is loosing sight of it a soon as backup job is ran.

On a freshly booted host "nvidia-smi" output is normal, as soon as I run backup "nvidia-smi" gives me this error:

"Failed to initialize NVML: No supported GPUs were found
Unable to determine the number of GPUs"

Output of "lspci | grep -i nvidia" doesn't change.

Once that happens LXC cannot use the GPU. Everything works after reboot till next backup.

Can someone please point me in the right direction to try to track down the problem? Pulling my hair out.

Proxmox 8.4.1
Nvidia RTX2000 6gb with latest driver
 
Have you maybe configured PCI pass trough for that GPU to one of the VMs? Note that stopped VMs have to be started (although in frozen state) in order to backup the attached block devices.