VM with GPU passthrough runs for a few minutes and then hard reboots my server

[dexxy]

I recently switched out my GTX 1070 for an RTX 2060 Super pulled from an OEM machine. Ever since, I've been having instability issues. My Windows VM would refuse to boot, or with eventual tweaking, would boot but then freeze a few minutes after launch. Also, unraid occasionally become unresponsive when I tried to reboot the VM if it was stable long enough to do so.

 

I stopped trying to launch the VM for a few months, and eventually gave up on fixing it and created a new windows VM using the 2060. The windows setup process will run fine for a few minutes before hard rebooting my server. The entire machine goes down in an abrupt and unclean restart. The same problem occurs when I create a PopOS VM, so I don't think it's a windows-specific issue.

 

I've been pulling my hair out trying to solve this, but haven't had any luck. I'm not seeing anything relevant in syslog, either.

 

Any ideas? I'm attaching diags, vm settings, and my IOMMU groupings below

 

 

diagnostics-20211010-2348.zip

Edited October 11, 2021 by dexxy
made settings easier to visualize

[dexxy]

For anyone that finds this in the future, I ended up switching back to the old graphics card and starting a new windows VM from scratch. Seems to be working fine.