I updated my proxmox host a week or so ago and ever since when a backup job runs on the VM with my UHD 770 passed through to it I’ll get an I/O output error and the job fails.
If I manually kick off a backup it’s fine.m after it fails The ZFS pool is healthy. There’s nothing wrong with it, but what I did notice is after the VM is shutdown for the backup on restart it’ll have a kernel panic error about i915 reset requested timeout. So, I think the backup job is only failing because the VM fails to properly boot back up afterwards.
To get the VM to boot again I have to shutdown the rest of my VMs and do a graceful reboot of the proxmox host and then the VM boots happily again.
I do not have the drivers blacklisted because I never had to before, and was trying not to blacklist them because it’s my only GPU for the server so I’d like to retain console access if possible for a break glass situation. Even then I don’t think it’s a blacklist issue because it’s just not resetting it properly now. Does anyone know what it would take to make it start resetting the iGPU again?
If I manually kick off a backup it’s fine.m after it fails The ZFS pool is healthy. There’s nothing wrong with it, but what I did notice is after the VM is shutdown for the backup on restart it’ll have a kernel panic error about i915 reset requested timeout. So, I think the backup job is only failing because the VM fails to properly boot back up afterwards.
To get the VM to boot again I have to shutdown the rest of my VMs and do a graceful reboot of the proxmox host and then the VM boots happily again.
I do not have the drivers blacklisted because I never had to before, and was trying not to blacklist them because it’s my only GPU for the server so I’d like to retain console access if possible for a break glass situation. Even then I don’t think it’s a blacklist issue because it’s just not resetting it properly now. Does anyone know what it would take to make it start resetting the iGPU again?