Hey guys,

This problem has been going for MONTHS and it's driving me insane because just when I think I've fixed it, it happens again.
So I mainly use this card for Plex transcoding, and when the Nvidia driver is detecting the card normally, it works flawlessly. When trandcoding, the GPU statistics plugin correctly displays the Plex icon and nvidia-smi returns the appropriate data. Needless to say, I have correctly set up the Plex docker for the nvidia transcoding. I DON'T have it passed through to a VM.
However every few weeks, the nvidia driver/plugin decides it no longer wants to detect the card, and will not detect it again until I reboot the entire server.

I've attached diags and screenshots. See below.

The card is visible in system devices under IOMMU group 13: [10de:21c4] 01:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1660 SUPER] (rev a1).

However I can find no trace of it anywhere else. Even running 'nvidia-smi' returns 'No devices were found'.
I've double and triple checked that the card is seated correctly. I've deleted the Nvidia plugin/driver, stopped docker, then reinstalled several times. I've rolled back the driver as far back as the plugin will allow (v575.64.05), but still no success.
I'm sick to death of logging into the web UI and discovering that the GPU Statistics plugin has not detected the card (god knows for how long), and I need this fixed PERMANENTLY.

Any ideas???

nas-diagnostics-20251224-0010.zip

Edited December 23, 2025Dec 23 by Corvus

Dec 24 00:10:47 NAS kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x22:0x51:884)

Dec 24 00:10:47 NAS kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

This indicates that the NVIDIA kernel driver is detecting the GPU on the PCIe bus (0000:01:00.0) but it's failing to initialize it. Once this happens, the driver retries continuously and never succeeds.

Typically this can happen

- if the GPU is not being fully reset by a container, and it requires a reboot

- if there are hardware issues with the GPU, like a power problem, overheating, or just failing

6 hours ago, JorgeB said:

Dec 24 00:10:47 NAS kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x22:0x51:884)

Dec 24 00:10:47 NAS kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

This indicates that the NVIDIA kernel driver is detecting the GPU on the PCIe bus (0000:01:00.0) but it's failing to initialize it. Once this happens, the driver retries continuously and never succeeds.

Typically this can happen

- if the GPU is not being fully reset by a container, and it requires a reboot

- if there are hardware issues with the GPU, like a power problem, overheating, or just failing

This also happened with my old 1080ti on another motherboard before I upgraded my Unraid build, so I doubt it's a hardware fault.

What should I do next?

Edited December 23, 2025Dec 23 by Corvus

Most likely option 1 then, or some issue with the Nvidia driver. Recommend asking in the existing plugin support thread; also post the diags there.