GPU available in isolation but unavailable to App on Scale 25.04.2.6 running on bare metal DELL R730XD

Apps and Virtualization
1

Hi all,

I’ve tried to read up, but no similar issue has popped up.

I’m running TrueNAS Scale 25.04.2.6 on bare metal DELL R730XD running (among others) Emby and I tried to add an NVIDIA Quadro P2000 today to allow Emby to use it.

After adding the GPU to the server and booting back up, it appeared in the GPU Isolation menu just fine. I selected it, isolated it and went to the Emby App Settings, scrolled down to the Resource Allocation section to add it to Emby. There I can’t select it. The only option I have is to pass through all non-NVIDIA GPUs.

What should I do to make it appear in the Emby App Settings?

Also, if I’m typing nvidia-smi in the shell, it is not installed… is this something I should add manually? I usually don’t add anything to the TrueNAS installation manually, how do people get that working?

Am I missing something?

Any help is greatly appreciated!

2

To use a GPU in an App - do not isolate it. It should be just available.

Isolation is for use in a VM.

As an additional point a P2000 is no longer supported by TrueNAS after V25.10 as the nvidia drivers no longer support it. There is a way of making it work, but its no longer officially supported (this was an Nvidia decision, not a IX one)

3 Likes
3

This is one of those things that doesn’t make sense but it does. You’d think isolating it would save it for the vm but nope, it’s the opposite.

True about the discontinued support, all Nvidia. They decided to split driver support in half and leave old stuff on an old driver. Ix had to follow suit.

4

It does… but for the VM, not docker :stuck_out_tongue:

5

Thank you! It does make sense, I understand now.

However. To use with transcoding in Emby, the P2000 is a more than capable choice. In fact, I have another server with 2 of these bad boys and I’m running a small LLM on them (10GB VRAM) to use with automation, etc. This is a more than capable card still.

How can I switch the installed NVIDIA driver to the older/other one that still would support the P2000?

On Ubuntu, I was perfectly able to install a driver that supports it, it shouldn’t be impossible here I guess.

Any help is greatly appreciated!

Cheers,

Gabriel

6

By the way… you write 25.10… I’m running 25.04.2.6 so it should still be supported?

I can’t make it appear even if I removed the isolation.

7

I just discovered that I have “Install NVIDIA Drivers” unchecked in Apps / Coniguration / Settings.

Another issue now is that I can’t seem to be able to turn it on as it has an issue with the Address Pools set there (I haven’t set them I guess, that’s all default)…

Hmmm.

G.