Docker Apps and UUID issue with NVIDIA GPU after upgrade to 24.10 or 25.04

TrueNAS General
, ,
65

For TrueNAS Scale 25.10 (Goldeneye) I used:

midclt call app.update APP_NAME ‘{“values”: {“resources”: {“gpus”: {“use_all_gpus”: false, “nvidia_gpu_selection”: {“PCI_SLOT”: {“use_gpu”: true, “uuid”: “GPU_UUID”}}}}}}’

1 Like
66

hi and thank you it worked for me too!
do any of you know how do i check that it actuallyworks in my jellyfin?

67

when you play a file that should transcode you can use the webshell or an ssh session and

nvidia-smi

if a service like jellyfin or plex is using the gpu it will tell you at the bottom of the output like this

+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.172.08             Driver Version: 570.172.08     CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA T400 4GB                Off |   00000000:27:00.0 Off |                  N/A |
| 37%   48C    P0            N/A  /   31W |     223MiB /   4096MiB |     99%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A         2259904      C   /usr/lib/jellyfin-ffmpeg/ffmpeg         219MiB |
+-----------------------------------------------------------------------------------------+

or you can use the jellyfin player. At the bottom right near the fullscreen button is a cogwheel icon. click on playback information, a screen should pop up and will tell you if the file is bein transcoded and the reason why it gets transcoded.

1 Like
69

thnaks i played with it and now it works
thx!

70

Thanks a lot for this thread, it stopped working for me in jellyfin after i changed GPU to 1660 TI and updated to 25.10.

Command to find out details of the GPU:

midclt call app.gpu_choices | jq

My output:

{
  "0000:07:00.0": {
    "vendor": "NVIDIA",
    "description": "NVIDIA GeForce GTX 1660 Ti",
    "error": null,
    "vendor_specific_config": {
      "uuid": "GPU-0ea5b7e6-5d2a-c514-a260-c136f9625e55"
    },
    "gpu_details": {
      "addr": {
        "pci_slot": "0000:07:00.0",
        "domain": "0000",
        "bus": "07",
        "slot": "00"
      },
      "description": "NVIDIA Corporation TU116 [GeForce GTX 1660 Ti]",
      "devices": [],
      "vendor": "NVIDIA",
      "uses_system_critical_devices": false,
      "critical_reason": null,
      "available_to_host": true
    },
    "pci_slot": "0000:07:00.0"
  }
}

Safter copying pci slot: 0000:07:00.0 and uuid: GPU-0ea5b7e6-5d2a-c514-a260-c136f9625e55 command that solved my issue was:

midclt call -j app.update jellyfin-truenas '{"values": {"resources": {"gpus": {"use_all_gpus": false, "nvidia_gpu_selection": {"0000:07:00.0": {"use_gpu": true, "uuid": "GPU-0ea5b7e6-5d2a-c514-a260-c136f9625e55"}}}}}}'

Thanks again and have a good server everyone :slight_smile: