Docker Apps GPU passthrough "UUID: Not Found"

TrueNAS General
, ,
1

Hello there,

This is my first post here, I hope I provide all information required.
I did search for the issue below, but I couldn’t find any that matched this exactly.
If there is, please let me know and i’ll continue there.

Last week I’ve upgraded TrueNas Scale from 24.x to 25.10.4 and also upgraded my gpu from a MSI GTX 1070 Aero to a newer MSI RTX 5060 Ventus 2x.

I’ve ran into the issue that Truenas didn’t respond to the new gpu and threw an error on boot
“This pci i/o region assigned to your nvidia device is invalid“

After some digging around more posts suggested doing a fresh install of TrueNas with disabling CSM in the bios. It did remove the above mentioned error and I don’t see any gpu error’s on boot.

So after I reloaed my system configuration file and all apps were deployed again. Now my gpu shows up in the apps/plex/edit “GPU Configuration“ screen to be forwarded to the docker container, but on save there is a new exception which can’t seem to fix.

Below error message is from /var/log/app_lifecycle.log

“[2026/06/16 20:03:34] (ERROR) app_lifecycle.compose_action():58 - Failed ‘up’ action for ‘plex’ app: Network ix-plex_default Creating\n Network ix-plex_default Created\n Container ix-plex-plex-1 Creating\n Container ix-plex-plex-1 Created\n Container ix-plex-plex-1 Starting\nError response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: could not apply required modification to OCI specification: error modifying OCI spec: failed to create the automatic CDI modifier: failed to generate CDI spec for mode “auto”: failed to construct device spec generators: failed to get device handle from UUID: Not Found: unknown\n“

  1. I made sure the GPU is not isolated under advanced settings
  2. went to the app screen unchecked “install nvidia drivers“ and saved, waited 5 minutes and rechecked the same option and the rebooted the machine.
  3. nvidia-smi returns “No devices were found“
  4. As mentioned earlier the gpu now shows up in the app edit window “NVIDIA GeForce RTX 5060“

Searching for this error message gave me no direct results, but I found this post (prepend https):
”://forums.truenas.com/t/docker-apps-and-uuid-issue-with-nvidia-gpu-after-upgrade-to-24-10-or-25-04/22547”

ran the midclt call app.gpu_choices | jq command which also shows the gpu.
Then ran midclt call app.update APP_NAME ‘{“values”: {“resources”: {“gpus”: {“use_all_gpus”: false, “nvidia_gpu_selection”: {“PCI_SLOT”: {“use_gpu”: true, “uuid”: “GPU_UUID”}}}}}}’

But it has not changed anything regarding the error message, I’m unsure what my next step should be. I hope you can point me in the right direction.

2

I did find a matching post with no real resolution.
://forums.truenas.com/t/gpu-shows-up-in-the-device-list-but-cant-be-accessed/66151
It looks like i have to wait for TrueNas to add support?

Addition information form midclt call app.gpu_choices | jq

{
  "0000:09:00.0": {
    "vendor": "NVIDIA",
    "description": "NVIDIA GeForce RTX 5060",
    "error": null,
    "vendor_specific_config": {
      "uuid": "GPU-689b992f-fe70-439e-6e56-a8da096663e0"
    },
    "gpu_details": {
      "addr": {
        "pci_slot": "0000:09:00.0",
        "domain": "0000",
        "bus": "09",
        "slot": "00"
      },
      "description": "NVIDIA Corporation VGA compatible controller",
      "devices": [
        {
          "pci_id": "10DE:2D05",
          "pci_slot": "0000:09:00.0",
          "vm_pci_slot": "pci_0000_09_00_0"
        },
        {
          "pci_id": "10DE:22EB",
          "pci_slot": "0000:09:00.1",
          "vm_pci_slot": "pci_0000_09_00_1"
        }
      ],
      "vendor": "NVIDIA",
      "uses_system_critical_devices": false,
      "critical_reason": null,
      "available_to_host": true
    },
    "pci_slot": "0000:09:00.0"
  }
}