25 minutes ago, LakersFan said:

Thank you!

Please read the post above yours and you have your answer (especially the last paragraph).

Thanks @ich777 . I initially read that post but thought it was a different issue. So I rolled it back and it works. Basically what you're saying though, is I'll no longer be able to update this particular GPU, right?

Just now, LakersFan said:

Basically what you're saying though, is I'll no longer be able to update this particular GPU, right?

It seems now that Nvidia made 10 series now legacy which basically means that you are stuck on driver version 580, however please keep in mind that Pascal so to speak 10 series cards are now almost 10 years old.

It's also clearly visible in the syslog:

Dec 5 07:26:34 Tower kernel: NVRM: The NVIDIA NVIDIA GeForce GTX 1060 6GB GPU installed in this system is

Dec 5 07:26:34 Tower kernel: NVRM: supported through the NVIDIA 580.xx Legacy drivers. Please

Dec 5 07:26:34 Tower kernel: NVRM: visit http://www.nvidia.com/object/unix.html for more

Dec 5 07:26:34 Tower kernel: NVRM: information. The 590.44.01 NVIDIA driver will ignore

Dec 5 07:26:34 Tower kernel: NVRM: this GPU. Continuing probe...

Dec 5 07:26:34 Tower kernel: NVRM: The NVIDIA NVIDIA GeForce GTX 1050 Ti GPU installed in this system is

Dec 5 07:26:34 Tower kernel: NVRM: supported through the NVIDIA 580.xx Legacy drivers. Please

Dec 5 07:26:34 Tower kernel: NVRM: visit http://www.nvidia.com/object/unix.html for more

Dec 5 07:26:34 Tower kernel: NVRM: information. The 590.44.01 NVIDIA driver will ignore

Dec 5 07:26:34 Tower kernel: NVRM: this GPU. Continuing probe...

Dec 5 07:26:34 Tower kernel: NVRM: No NVIDIA GPU found.

On 12/5/2025 at 9:08 PM, ich777 said:

It seems now that Nvidia made 10 series now legacy which basically means that you are stuck on driver version 580, however please keep in mind that Pascal so to speak 10 series cards are now almost 10 years old.

It's also clearly visible in the syslog:

Dec 5 07:26:34 Tower kernel: NVRM: The NVIDIA NVIDIA GeForce GTX 1060 6GB GPU installed in this system is

Dec 5 07:26:34 Tower kernel: NVRM: supported through the NVIDIA 580.xx Legacy drivers. Please

Dec 5 07:26:34 Tower kernel: NVRM: visit http://www.nvidia.com/object/unix.html for more

Dec 5 07:26:34 Tower kernel: NVRM: information. The 590.44.01 NVIDIA driver will ignore

Dec 5 07:26:34 Tower kernel: NVRM: this GPU. Continuing probe...

Dec 5 07:26:34 Tower kernel: NVRM: The NVIDIA NVIDIA GeForce GTX 1050 Ti GPU installed in this system is

Dec 5 07:26:34 Tower kernel: NVRM: supported through the NVIDIA 580.xx Legacy drivers. Please

Dec 5 07:26:34 Tower kernel: NVRM: visit http://www.nvidia.com/object/unix.html for more

Dec 5 07:26:34 Tower kernel: NVRM: information. The 590.44.01 NVIDIA driver will ignore

Dec 5 07:26:34 Tower kernel: NVRM: this GPU. Continuing probe...

Dec 5 07:26:34 Tower kernel: NVRM: No NVIDIA GPU found.

is it fixed that you can set the driver version to a different , but the plugin will eventually switch back to "latest"?

I guess that would be annoying, untested (i also got the error with my 1050ti)

Edited December 8, 2025Dec 8 by NewDisplayName

4 minutes ago, NewDisplayName said:

I hope its fixed that you can change the driver version

Cool, didn't know that you can't change the driver version. :D

5 minutes ago, NewDisplayName said:

but the plugin will eventually switch back to "latest"?

For now yes because I haven't changed legacy versions, who knows what Nvidia does next and if they stick to that that driver 580 is now legacy but I will fix that eventually.

Hi - I've had a notification email a couple of times over the last few weeks saying...

Event: Nvidia Driver
Subject: Notification
Description: Found new Nvidia Driver v580.105.08 but a download error occurred! Please try to download the driver manually!
Importance: alert

However, v580.105.08 is the version I have installed. Whenever this happens, my log has several of these lines listed...

Dec  9 09:56:39 unRAID1 kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x22:0x51:884)
Dec  9 09:56:39 unRAID1 kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0
Dec  9 09:57:35 unRAID1 kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x22:0x51:884)
Dec  9 09:57:35 unRAID1 kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

And shortly after, if I go to the Nvidia driver page I get no cards found. Appreciated the GPU stats plugin is separate, but on the dashboard no cards listed (Vendor command returned unparseable data).

Sometimes it comes back, maybe bouncing on and off. I've rebooted and it's gone back to being stable for a while.

This is an A2000 (after upgrading from 1050ti), not used in VM, latest version of Unraid.

It's not catastrophic, but up until recently my machine's been pretty stable. Could easily be something I've left misconfigured though.

Thanks!

unraid1-diagnostics-20251209-0958.zip

2 minutes ago, Cessquill said:

Sometimes it comes back, maybe bouncing on and off. I've rebooted and it's gone back to being stable for a while.

This is a common issue with newer cards it seems.

Please try to add:

nvidia-persistenced

to your go file and reboot.

However most of the times it seems some kind of hardware compatibility issue.

Make sure your BIOS is up to date, you have Resizable BAR Support and above 4G Decoding enabled in your BIOS.

Check that you PSU is up to the task (please don't forget that PSUs degrade over time).

8 minutes ago, Cessquill said:

Hi - I've had a notification email a couple of times over the last few weeks saying...

This is a different issue and not related to the other issue.

Can you also please send a screenshot from your plugin page please.

3 minutes ago, ich777 said:

Can you also please send a screenshot from your plugin page please.

I'll add the line to my go file (can I put it in a user script on array start?).

I'll check the BIOS - pretty sure it's on the last update SuperMicro did but I'll check the settings. To be honest it's getting a little old and needs replacing, but that's a job for next year. The PSU should be up to it, but again, age...

Thank you for your time

6 minutes ago, Cessquill said:

I'll add the line to my go file (can I put it in a user script on array start?).

The command needs to be executed exactly once, when the server starts.

I would leave it in the go file.

Oh and please execute it once if you haven't rebooted yet. Few users said that this solved their issues.

22 hours ago, ich777 said:

The command needs to be executed exactly once, when the server starts.

I would leave it in the go file.

Oh and please execute it once if you haven't rebooted yet. Few users said that this solved their issues.

Command added to go file, BIOS confirmed to be up to date (2021), BIOS is up to date, above 4G Decoding enabled, but I do not have Resizable BAR Support available. Fingers crossed that buys a bit of time until a full system upgrade.

I will look at the PSU too since I added an extra drive a few weeks ago - might have tipped it over the edge 🤔

Thank you for your help again.

4 hours ago, Cessquill said:

I will look at the PSU too since I added an extra drive a few weeks ago - might have tipped it over the edge 🤔

Yeah, these new GPUs can produce pretty significant Power Spikes in your system and this could be one reason why the GPU falls from the bus and doesn't recover properly.

10 minutes ago, ich777 said:

Yeah, these new GPUs can produce pretty significant Power Spikes in your system and this could be one reason why the GPU falls from the bus and doesn't recover properly.

Good point, thank you. I've just checked - I've got a 920w power supply and it's powering 24 drives + parity + SSDs + GPU. If it was 920w new it could well be struggling.

It's a SuperMicro chassis, so not a regular PSU. I've just seen I can get a 1U 1280W supply that fits and is still part of their quiet range, so that might be worth a try (used prices are worth the risk).

EDIT: Used good condition 1200w supply - £24, just purchased. There's also 1400w, but they're still £250+

Edited December 10, 2025Dec 10 by Cessquill

I’m trying to get my GPU to work with Jellyfin.

While troubleshooting, I noticed that the GPU was detected at first. But as soon as I started Jellyfin and played a movie, the GPU disappeared. When I ran watch nvidia-smi in the Unraid terminal, it reported “no GPU.” After that, the GPU also disappeared from the Unraid NVIDIA plugin—it went from showing up normally to not being detected at all.

Dec 11 07:43:28 Happy2Me kernel: NVRM: Rate limiting GSP RPC error prints for GPU at PCI:0000:01:00 (printing 1 of every 30). The GPU likely needs to be reset.

Dec 11 07:45:24 Happy2Me kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x25:0xffff:1740)

Dec 11 07:45:24 Happy2Me kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

Dec 11 07:45:24 Happy2Me kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x62:0x40:2790)

Dec 11 07:45:24 Happy2Me kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

Dec 11 07:45:26 Happy2Me kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x62:0x40:2790)

Its an GeForce GTX 1650

Edited December 10, 2025Dec 10 by tombstone

3 minutes ago, tombstone said:

Its an GeForce GTX 1650

I just briefly read a post in some thread that the newest nvidia driver dropped support for all 1x series cards. Or maybe it was just 10xx cards, can't remember exactly. But maybe you need to roll back to a previous driver? Check what cards your driver supports.

13 minutes ago, strike said:

I just briefly read a post in some thread that the newest nvidia driver dropped support for all 1x series cards. Or maybe it was just 10xx cards, can't remember exactly. But maybe you need to roll back to a previous driver? Check what cards your driver supports.

Thanks, but I just checked and it seems like it’s still supported, but the plugin is using the beta version?

7 hours ago, tombstone said:

But as soon as I started Jellyfin and played a movie, the GPU disappeared.

Please post your diagnostics.

7 hours ago, strike said:

Or maybe it was just 10xx cards

This.

All GTX 10 und 9 series are now legacy. All Turing based cards, including GTX16 series should work just fine bisides the point that the driver seems pretty unstable with the newer driver releases.

Don't suppose you could update the Open Source Driver to "590.44.01" (source: https://github.com/NVIDIA/open-gpu-kernel-modules/releases/tag/590.44.01) could ya?

Also, been getting Out of Memory errors recently with the error referencing Nvidia.server-diagnostics-20251211-0351.zip

7 minutes ago, AnnabellaRenee87 said:

Also, been getting Out of Memory errors recently with the error referencing Nvidia.

This refers to Kernel Memory not strictly speaking RAM.

Please make sure that you've enabled Above 4G Decoding and Resizable BAR Support in your BIOS.

Please also make sure that you are on the latest BIOS version available, most of the times this points to firmware related issue.

Not much that I can do about, you'll find plenty of these errors on the Nvidia Developer forums and most of the times this happens to users on AMD platforms, please maybe report that on the Nvidia Developer forums since this is directly related to the driver.

10 minutes ago, AnnabellaRenee87 said:

Don't suppose you could update the Open Source Driver to "590.44.01" (source: https://github.com/NVIDIA/open-gpu-kernel-modules/releases/tag/590.44.01) could ya?

I could, but usually I compile the drivers once when a Unraid version is released, are there any new features that you need or why do you need the newer driver?

12 minutes ago, ich777 said:

This refers to Kernel Memory not strictly speaking RAM.

Please make sure that you've enabled Above 4G Decoding and Resizable BAR Support in your BIOS.

Please also make sure that you are on the latest BIOS version available, most of the times this points to firmware related issue.

Not much that I can do about, you'll find plenty of these errors on the Nvidia Developer forums and most of the times this happens to users on AMD platforms, please maybe report that on the Nvidia Developer forums since this is directly related to the driver.

I could, but usually I compile the drivers once when a Unraid version is released, are there any new features that you need or why do you need the newer driver?

I was hoping it would fix some of those memory errors, I just told the Fix Comon Problems plugin to igneore the errors, I've got 256 GB of RAM so I know I'm not running out of memeory.

I've also got above 4G decoding and Resizable Bar enabled.

I've also got this as an option, it's Disable right now.

Not sure if that would help or not.

this is what I get before running jellyfin

4 hours ago, ich777 said:

Please post your diagnostics.

I have download it, can I pm you 1st time doing this and now don't know what could be on this file?

58 minutes ago, AnnabellaRenee87 said:

I was hoping it would fix some of those memory errors, I just told the Fix Comon Problems plugin to igneore the errors, I've got 256 GB of RAM so I know I'm not running out of memeory.

I've also got above 4G decoding and Resizable Bar enabled.

I've also got this as an option, it's Disable right now.

Not sure if that would help or not.

I turned it on, seems to have helped, will test for a bit, don't know if it will cause any performance issues or not.

EDIT: it defaulted to Disabled, Changed it to Enabled.

Edited December 11, 2025Dec 11 by AnnabellaRenee87

2 hours ago, AnnabellaRenee87 said:

I was hoping it would fix some of those memory errors, I just told the Fix Comon Problems plugin to igneore the errors, I've got 256 GB of RAM so I know I'm not running out of memeory.

As said that has nothing to do with System Memory, that refers to Kernel Memory, so to speak reserved memory for the Kernel in System Memory and the message is actually produced by the Nvidia Driver, so nothing that I could do about.

2 hours ago, AnnabellaRenee87 said:

Not sure if that would help or not.

Try it.

As another test, please post nvidia-persistenced in your go file and execute it once (so that you don't have to reboot).

13 hours ago, tombstone said:

Thanks, but I just checked and it seems like it’s still supported, but the plugin is using the beta version?

If you have selected "latest" on the plugins settings page, then yes, you will be on the beta driver because latest means always the latest available driver version, which not always must mean that you are on a Beta driver, but currently Beta is the latest available driver.

You know that the plugin has a settings page right?

1 hour ago, tombstone said:

I have download it, can I pm you 1st time doing this and now don't know what could be on this file?

Nope, I don't accept any PMs, there is usually nothing you have to hide in that file.

Please drag the Diagnostics file into the textbox here on the forums.

19 hours ago, ich777 said:

If you have selected "latest" on the plugins settings page, then yes, you will be on the beta driver because latest means always the latest available driver version, which not always must mean that you are on a Beta driver, but currently Beta is the latest available driver.

You know that the plugin has a settings page right?

Nope, I don't accept any PMs, there is usually nothing you have to hide in that file.

Please drag the Diagnostics file into the textbox here on the forums.

OK thanks for looking into this for me

happy2me-diagnostics-20251211-2017.zip