Hi all,
I have spent more time than I’d like to admit at trying to get my Debian 13 VM to initialise the GPU that has been passed through via the host. It shows both in the “lspci” command and via “nvidia-detect” within the VM. My feeling is the error is memory based like others I have seen with similar error codes.
I am running Proxmox 9.0.5 on my very old PC, it is the only GPU in the system and I am trying to pass the GTX970 from the host. By all accounts this should be working. I have included a singular screencap below of a dmesg output (new user means I cannot add more…):
dmesg | grep -E “nvidia|error|nvrm”
At the moment I fundamentally just don’t know what the actual problem is and I guess just been going in circles. I have been searching for hours and trying anything that is remotely similar to my current issue. Before using proxmox, the GPU would work on Debian 12 (bare metal) and I have followed all the steps I have used before. I have also:
- confirmed the GPU precence with:
- Nvidia-detect
- lspci -nnk
- systemctl status nvidia-pesistenced.service (though it “fails to open” pcie)
- Followed Debian wiki for driver install.
- Tried the memory hook script as per (mentioned on another similar nvidia forum post):
- I have reinstalled the 535 drivers and ensured the kernel was given the mok key - though image above suggests that hasnt happened.
- and followed many other things accross youtube and forums
I guess I am mainly asking, is this a driver issue, memory issue, or proxmox issue? I just need some guidance on the root cause and/or even solutions if there are some. Thanks heaps.
nvidia-bug-report.log.gz (106.8 KB)