Until 2 days ago, I had things running (mostly) the way I wanted with 2 nvidia GPUs passed through to 2 separate windows VMs.
I was unable to connect to those machines sometime yesterday, but was able to connect to other VMs without any issue.
I was able to access those machines after rebooting the server.
When the machines went down again randomly today, I rebooted the server, but only started up 1 of the VMs with passed through GPU, and it worked fine - but lost connection again a few hours later.
I tried to start up the other VM (with passed through GPU), and it's now throwing up the below error
internal error: Unknown PCI header type '127' for device '0000:01:00.0'
Interestingly, the CPU for the VM that I cannot access is stuck at 100%
Has anyone else come across this?