[6.11.3] GPU passthrough vm now hangs and then entire unraid server reboots

Nothing in the logs as it is post reboot do you have a syslog server setup?

I'm afraid not. I'm turning on the local syslog server now so will get some more logs ASAP

 

 

Edited November 11, 2022 by harley-c

actually... with the constant parity checks I seem to have had a disk failure. I've replaced it and will let this recover before any further testing so it will probably be tomorrow before I get any more logs

tower-diagnostics-20221116-2043.zipanother diagnostics file atached. as far as i've seen so far it only seems to be my vm with the GPU passed through. Seems stable when this is not runnning, and seems stable when other regular vms are running

 

syslog in ther diag file seemed to be truncated so i included full file from the flash drive also

 

didn't seem to really be any messages before the flood of messages after reboot, but see if anything stands out

 

thanks

 

syslog-from-flash.txt

Just crashed again and I managed to get this in the vm log before the host rebooted 

 

Searching around seems to keep pointing me to some info on AMD GPU reset bugs. I'm using an NVidia 3060Ti though ... has anyone seen that same sort of issue with nVidia?

Seeing the same error on my setup on unraid 6.11.5
- Using a Nvidia 3060
 

I can install and create a VM and pass the GPU to the VM guest. But when i update the GPU to the latest drivers, it crashes the server.

 

Did you find a solution? Did reverting back to 6.11.0 revert the issue?

Edited February 4, 2023 by Jeff Balagot