I am running ubuntu with casa os. I was previously running an intel apu (the name has slipped me I will update the post when I can with this info). Recently I got a 1650 that I installed for nvenc transcoding. It seems all the proper drivers are installed but my jellyfin container still fails playback anytime with it turned on.

I have reinstalled the container with the nvidia device variable and no dice. I have also tried installing the nvidia cintainer toolkit but that didn't work either. I am at a loss for trying to get nvenc to work.

Any help is appreciated!

EDIT: here is the ffmpeg log file

https://gofile.io/d/9nsBFq

EDIT 2: It was a problem with my docker compose file! I didn't list the needed devices from the jellyfin documentation. I thought the Container was detecting the gpu but it wasn't. Docker exec nvidia-smi is your friend!

EDIT 3: so now it doesnt kick me out saying the playback failed but its just a black screen with 4k media

EDIT 4: nvm i just didnt have the proper jellytin transcode settings set lol

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 1 points 1 month ago* (last edited 1 month ago) (1 children)

What's the output of nvidia-smi on host? What is "casa os"? Do you have the Nvidia container cuda runtime installed? What's your docker-compose.yml look like?

For me on Debian it worked pretty much out of the box, I just installed the cuda toolkit, blacklisted nouveau, and added this to docker-compose.yml:

environment:
  NVIDIA_DRIVER_CAPABILITIES: all
  NVIDIA_VISIBLE_DEVICES: all
deploy:                                       
  resources:
    reservations:                       
      devices:
      - capabilities: [gpu]

It works both in Immich and Jellyfin.

[–] [email protected] 2 points 1 month ago* (last edited 1 month ago) (1 children)

Casaos is just a server ui for ubuntu to manage docker and other settings:

https://casaos.io/

Here is the output of nvidia-smi (sorry for the picture if my monitor. It's late and I blanking on my lemmy pass and cannot sign in on the pc itself to send a screen cap lol):

I used several different jellyfin images posted in casaos appstore along with the official nvidia documentation. It still does not want to work:

services:
 jellyfin:
   image: jellyfin/jellyfin
   user: 1000:1000
   network_mode: 'host'
   volumes:
     - /DATA/AppData/jellyfin/config:/config
     - /DATA/AppData/jellyfin/cache:/cache
     - /DATA/AppData/jellyfin/media:/media
     - /mnt/drive1/media:/mnt/drive1/media
     - /mnt/drive2/Jellyfin:/mnt/drive2/Jellyfin
     - /mnt/drive3:/mnt/drive3
     - /mnt/drive4/media:/mnt/drive4/media
     - /mnt/drive5/jellyfin:/mnt/drive5/jellyfin
     - /mnt/drive6/jellyfin:/mnt/drive6/jellyfin
   runtime: nvidia
   deploy:
     resources:
       reservations:
         devices:
           - driver: nvidia
           - /dev/nvidia-caps:/dev/nvidia-caps
           - /dev/nvidia0:/dev/nvidia0
           - /dev/nvidiactl:/dev/nvidiactl
           - /dev/nvidia-modeset:/dev/nvidia-modeset
           - /dev/nvidia-uvm:/dev/nvidia-uvm
           - /dev/nvidia-uvm-tools:/dev/nvidia-uvm-tools
             count: all
             capabilities: [gpu] ```

[–] [email protected] 1 points 4 weeks ago* (last edited 4 weeks ago)

Have you tried removing all that stuff?

I'm talking about:

         - driver: nvidia
            - /dev/nvidia-caps:/dev/nvidia-caps
            - /dev/nvidia0:/dev/nvidia0
            - /dev/nvidiactl:/dev/nvidiactl
            - /dev/nvidia-modeset:/dev/nvidia-modeset
            - /dev/nvidia-uvm:/dev/nvidia-uvm
            - /dev/nvidia-uvm-tools:/dev/nvidia-uvm-tools
              count: all

And just leave it with:

  deploy:
      resources:
        reservations:
          devices:
              count: all
              capabilities: [gpu]

Finally, make sure you have the cuda docker toolkit installed.

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html

Make sure to then test with:

sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi

As per:

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/sample-workload.html

EDIT: NVM see you fixed it! Could you post your final docker-compose.yml for reference? I would like to dive deep and figure out why some people need to add more to their docker-compose.yml and some don't.