Issue with the set up of my Intel A380

[MegaThorx]

If I run the vainfo within the container it fails with the following error message.




My setup is like the following:

Proxmox <=> Ubuntu VM <=> Jellyfin container

After plugging GPU into the server I enabled rebar and passed the device through to the VM. Within the VM I installed the drivers following the guide by Intel (https://dgpu-docs.intel.com/driver/insta...tall-steps) (https://ass.merx.dev/bronze-cheap-coral/direct). After that I verified that it is working within the VM by running vainfo (https://ass.merx.dev/phony-secondhand-cuttlefish/direct). Then I adjusted the docker compose file with the device (https://ass.merx.dev/crooked-bisque-hartebeest/direct).

I'm not sure if the step of installing the drivers on the VM is correct, but without them it won't show the /dev/dri/renderD128 device.

[TheDreadPirate]

Double check your proxmox config with our docs here.

https://jellyfin.org/docs/general/admini...on-proxmox

You also need to install all the Intel drivers and media drivers on Proxmox.

[MegaThorx]

Uninstalling the driver on the VM and passing /dev/dri to the container resolved the issue. But I'm now considering a different setup in which I'm running jellyfin and tdarr directly on the VM, which should allow me to transcode my library and transcode the streams as needed.

@TheDreadPirate I'm not running the container on the proxmox host. The container is running within the VM which runs on Proxmox.