Win 11 VM - GPU Passthrough - Crashes after longer idle periods

[WoRie]

Hi,

 

after wasting a quarter of a year due to a broken GPU and subsequent issues due to this, I replaced the GPU (RTX 4090, water cooled) and was finally able to use the VM as intended. I passed through 32 of the 64 gigs of my RAM, and all but one core/HT core from my xeon CPU.

 

Everything is running fine, even for several hours, even with breaks in between sessions for several hours. I can either connect completly local with an HDMI cable connected to the passedthrough GPU as if I weren't running everything through virtualization in Unraid, or through parsec.

 

However, if I don't shut down the vm and leave it idling for longer periods, say over night for example, I come back to my system running all passed through CPU cores on 100% and the VM unresponsive. I then have to hard shutdown the vm through unraid (which works fine), my CPU usage drops back down to normal values. I can then reboot the VM and use it as before, no issues whatsoever.

 

The event viewer in windows only shows an error, that windows was shut down unexpectedly the last time, but nothing else in the period where the crash would have occured. The most recent one was at 02:03 in the morning, where I was fast asleep.

 

Any idea where I could start the trouble shooting? Diags are attached.

wonas-diagnostics-20230824-0928.zip