I'm not quite sure how to do the license transfer.

Use the USB creator to install unraid to the new flash drive

Copy the config folder from your existing one onto it, replacing any existing files

Boot with the new drive

Follow the license transfer wizard

Edited February 26Feb 26 by Kilrah

19 hours ago, DCoy said:

Hey, I am very new to this and I am trying to set up a plex server and want to use my old GTX 1070 to transcode. I have the GPU installed in the server and installed the Nvidia driver app but it wont recognize my GPU. Any advice on how I can fix this issue?

Please post your Diagnostics.

On 2/27/2026 at 1:29 AM, ich777 said:

Please post your Diagnostics.

how do I do that? I think I did it here: Uploading Attachment...

Hi

Just upgraded my server to 7.2.4, I've got a GTX 1660 Super and an RTX 5060 TI, but the OpenSource driver (needed due to the 5XXX card) only detects the 5060.

Just for checking, I've tried the production branch, and no GPU was detected.

Attaching my diagnostics...

small-diagnostics-20260228-2045.zip

Thanks in advance

3 hours ago, gshlomi said:

Just for checking, I've tried the production branch, and no GPU was detected.

Seems like something is wrong:

Feb 28 20:42:50 Small kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 239

Feb 28 20:42:50 Small kernel:

Feb 28 20:42:50 Small kernel: nvidia 0000:0c:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=none

Feb 28 20:42:50 Small kernel: NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:

Feb 28 20:42:50 Small kernel: NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0d:00.0)

Feb 28 20:42:50 Small kernel: nvidia 0000:0d:00.0: probe with driver nvidia failed with error -1

Feb 28 20:42:50 Small kernel: NVRM: The NVIDIA probe routine failed for 1 device(s).

Feb 28 20:42:50 Small kernel: NVRM: loading NVIDIA UNIX Open Kernel Module for x86_64 590.48.01 Release Build (dvs-builder@U22-I3-AE18-23-3) Mon Dec 8 13:05:00 UTC 2025

Feb 28 20:42:50 Small kernel: nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for x86_64 590.48.01 Release Build (dvs-builder@U22-I3-AE18-23-3) Mon Dec 8 12:51:03 UTC 2025

Please try to update your BIOS, make sure that Resizable BAR Support and Above 4G Decoding is enabled.

The culprit seems to be however Resizable BAR Support.

5 hours ago, DCoy said:

how do I do that? I think I did it here:

It seems that the upload didn't went all the way through.

59 minutes ago, ich777 said:

Seems like something is wrong:

Feb 28 20:42:50 Small kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 239

Feb 28 20:42:50 Small kernel:

Feb 28 20:42:50 Small kernel: nvidia 0000:0c:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=none

Feb 28 20:42:50 Small kernel: NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:

Feb 28 20:42:50 Small kernel: NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0d:00.0)

Feb 28 20:42:50 Small kernel: nvidia 0000:0d:00.0: probe with driver nvidia failed with error -1

Feb 28 20:42:50 Small kernel: NVRM: The NVIDIA probe routine failed for 1 device(s).

Feb 28 20:42:50 Small kernel: NVRM: loading NVIDIA UNIX Open Kernel Module for x86_64 590.48.01 Release Build (dvs-builder@U22-I3-AE18-23-3) Mon Dec 8 13:05:00 UTC 2025

Feb 28 20:42:50 Small kernel: nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for x86_64 590.48.01 Release Build (dvs-builder@U22-I3-AE18-23-3) Mon Dec 8 12:51:03 UTC 2025

Please try to update your BIOS, make sure that Resizable BAR Support and Above 4G Decoding is enabled.

The culprit seems to be however Resizable BAR Support.

Thanks, enabled Above 4G Decoding & Resizable BAR support and both cards are visible now.

I haven't noticed these settings as being prerequisites in the first post, but I might have missed something.

Recently installed a quadro p4000 in my dell R730xd and i'm unable to get unraid to recognize the card. It is seen by idrac and shows up in unraid but after installing nvidia driver it cannot find the device.

1. Unraid version: 7.2.4

2. GPU: Quadro P4000 (GP104GL) in Dell R730XD slot 4 x16

3. lspci output

   1. 82:00.0 VGA compatible controller: NVIDIA Corporation GP104GL [Quadro P4000] (rev a1)

      82:00.1 Audio device: NVIDIA Corporation GP104 High Definition Audio Controller (rev a1)

4. dmesg | grep nvidia

   1. [ 37.984748] nvidia: loading out-of-tree module taints kernel.

      [ 37.984767] nvidia: module license 'NVIDIA' taints kernel.

      [ 37.984776] nvidia: module license taints kernel.

      [ 38.266216] nvidia-nvlink: Nvlink Core is being initialized, major device number 238

      [ 38.276493] nvidia 0000:82:00.0: enabling device (0000 -> 0003)

      [ 38.281753] nvidia 0000:82:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=none

      [ 38.579148] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 580.105.08 Wed Oct 29 22:15:26 UTC 2025

      [ 38.583121] [drm] [nvidia-drm] [GPU ID 0x00008200] Loading driver

      [ 38.584422] [drm] Initialized nvidia-drm 0.0.0 for 0000:82:00.0 on minor 1

      [ 162.321401] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.

5. Tried all 580.xx versions

6. Symptoms: "No devices were found" in plugin + nvidia-smi "failed to communicate" or empty

7. No VFIO binding, VM Manager enabled but GPU not passed through

homethrall-diagnostics-20260307-2124.zip

5 hours ago, Cslis said:

Recently installed a quadro p4000 in my dell R730xd and i'm unable to get unraid to recognize the card. It is seen by idrac and shows up in unraid but after installing nvidia driver it cannot find the device.

I think you know what to do:

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: GPU does not have the necessary power cables connected.

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: RmInitAdapter failed! (0x24:0x1c:1590)

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: rm_init_adapter failed, device minor number 0

6 hours ago, ich777 said:

I think you know what to do:

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: GPU does not have the necessary power cables connected.

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: RmInitAdapter failed! (0x24:0x1c:1590)

Mar 7 21:14:08 HomeThrall kernel: NVRM: GPU 0000:82:00.0: rm_init_adapter failed, device minor number 0

After more troubleshooting turns out the power cable wasn't jumped causing it to not be recognized. After jumping the sense pin it's now fully detected!

sometimes it's really the simplest things 🤦

thank you i was going crazy trying to figure that out.

On 8/3/2025 at 10:27 PM, zanetheinsane said:

I am also getting this exact same issue this morning with my Tesla P4 (GP104GL). The only change was updating Unraid to 7.1.4.

Nvidia Info:

Nvidia Driver Version: 575.64.05

Open Source Kernel Module: No

Installed GPU(s):
0:
Tesla P4
01:00.0
GPU-1452e99e-5e98-1d02-48fa-434f381c4ea3

The plugin has no issues seeing the card, and Plex is configured to correctly detect it and use it. When Plex starts using it to transcode, the stream will eventually fail when it gets the "off the bus" error and will lock up the Plex docker.

I initially thought it was because I needed to update the Nvidia Driver Version to the latest (previously it was on v575.57.08).

Prior to the 7.1.4 upgrade, no issues transcoding or streaming (usually 4+ streams at once) and had about 30 days uptime. I've tested it with just the Plex docker running and nothing else and it still throws this error.

Aug 3 13:36:30 plex2 kernel: NVRM: GPU at PCI:0000:01:00: GPU-1452e99e-5e98-1d02-48fa-434f381c4ea3

Aug 3 13:36:30 plex2 kernel: NVRM: GPU Board Serial Number: 0421519083008

Aug 3 13:36:30 plex2 kernel: NVRM: Xid (PCI:0000:01:00): 79, GPU has fallen off the bus.

Aug 3 13:36:30 plex2 kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.

Aug 3 13:36:30 plex2 kernel: NVRM: GPU 0000:01:00.0: GPU serial number is 0421519083008.

Aug 3 13:36:30 plex2 kernel: NVRM: A GPU crash dump has been created. If possible, please run

Aug 3 13:36:30 plex2 kernel: NVRM: nvidia-bug-report.sh as root to collect this data before

Aug 3 13:36:30 plex2 kernel: NVRM: the NVIDIA kernel module is unloaded.

Aug 3 13:36:30 plex2 kernel: NVRM: Xid (PCI:0000:01:00): 154, GPU recovery action changed from 0x0 (None) to 0x1 (GPU Reset Required)

Was hoping it was an issue where I could just roll back to a previous driver version and not have to revert the Unraid update.

Did you ever resolve this? i have the same issue

Mar 10 16:04:22 VeNASaur kernel: NVRM: GPU at PCI:0000:01:00: GPU-e892f186-cabe-3ff7-aff8-58db14d53509

Mar 10 16:04:22 VeNASaur kernel: NVRM: GPU Board Serial Number: 0323116030975

Mar 10 16:04:22 VeNASaur kernel: NVRM: Xid (PCI:0000:01:00): 79, GPU has fallen off the bus.

Mar 10 16:04:22 VeNASaur kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.

Mar 10 16:04:22 VeNASaur kernel: NVRM: GPU 0000:01:00.0: GPU serial number is 0323116030975.

Mar 10 16:04:22 VeNASaur kernel: NVRM: A GPU crash dump has been created. If possible, please run

Mar 10 16:04:22 VeNASaur kernel: NVRM: nvidia-bug-report.sh as root to collect this data before

Mar 10 16:04:22 VeNASaur kernel: NVRM: the NVIDIA kernel module is unloaded.

Mar 10 16:04:22 VeNASaur kernel: NVRM: Xid (PCI:0000:01:00): 154, GPU recovery action changed from 0x0 (None) to 0x1 (GPU Reset Required)

GTX 1080

Nvidia Driver Version: 580.126.18

nvidia-persistenced in go file, pcie_aspm=off pcie_port_pm=off in syslinux, set gen 3 on the pcie slot

When this happens it locks up my jellyfin container and forces me to do an unclean shutdown.

lspci | grep -i nvidia still shows my GPU connected but nvidia-smi throws this error:

Unable to determine the device handle for GPU0: 0000:01:00.0: Unknown Error

i also got these errors before adding the syslinux line for ASPM, they just preceded the NVRM ones. after adding the syslinux line these messages disappeared:

Mar 6 16:22:26 VeNASaur kernel: pcieport 0000:00:01.0: AER: Correctable error message received from 0000:00:01.0

Mar 6 16:22:26 VeNASaur kernel: pcieport 0000:00:01.0: PCIe Bus Error: severity=Correctable, type=Physical Layer, (Receiver ID)

Mar 6 16:22:26 VeNASaur kernel: pcieport 0000:00:01.0: device [8086:a70d] error status/mask=00000001/00002000

Mar 6 16:22:26 VeNASaur kernel: pcieport 0000:00:01.0: [ 0] RxErr

I thought i could use my old 1080 for some light transcoding without having to buy a new GPU, i know it is old but it should still be supported on the 580.x branch right?

Diagnostics attached.

venasaur-diagnostics-20260311-0202.zip

Edited March 11Mar 11 by MammothJerk

5 hours ago, MammothJerk said:

When this happens it locks up my jellyfin container and forces me to do an unclean shutdown.

If this happens usually something is wrong with the hardware or at least some incompatibility issue.

Please try to load the default settings in the BIOS if possible, after that make sure that the boot order is correct, Resizable BAR Support and Above 4G Decoding is enabled.

If you have these AER issues in your log you can actually disable them since at times the cause issues. Add this to your syslinxu.config (at the end of your append line) :

pcie=noaer

However I would not recommend enabling ASPM when you get AER errors afterwards since this indicates some kind of Firmware issue.

Since you have Xid error 79 logged you can take a look here what this means, however it can be all:

https://docs.nvidia.com/deploy/xid-errors/analyzing-xid-catalog.html

However please also check if your Power Supply is still up to the task since they degrade over time and a few issues where solved by swapping the power supply for a new one.

On 3/10/2026 at 9:05 PM, MammothJerk said:

Did you ever resolve this? i have the same issue

GTX 1080

Nvidia Driver Version: 580.126.18

nvidia-persistenced in go file, pcie_aspm=off pcie_port_pm=off in syslinux, set gen 3 on the pcie slot

When this happens it locks up my jellyfin container and forces me to do an unclean shutdown.

lspci | grep -i nvidia still shows my GPU connected but nvidia-smi throws this error:

Unable to determine the device handle for GPU0: 0000:01:00.0: Unknown Error

i also got these errors before adding the syslinux line for ASPM, they just preceded the NVRM ones. after adding the syslinux line these messages disappeared:

I thought i could use my old 1080 for some light transcoding without having to buy a new GPU, i know it is old but it should still be supported on the 580.x branch right?

Diagnostics attached.

venasaur-diagnostics-20260311-0202.zip

Hi Guys,

I seem to have gotten a related issue. I have an nVidia 1050 GPU. Yesterday I upgraded to unRAID v.7.2.4 from version 7.1.2. I had transcoding working on 7.1.2 and then also in 7.2.4. Last night, I got an error about downloading a new version of the driver. I manually updated by removing the old plugin, and installing the latest version v595.45.04. Transcoding did not work....! After about 2 hours w/ ChatGPT, I suggested restoring from my flash drive's yesterday backup. It analyzed this and, in summary ,it told me to go down to v580.119.02 of the driver. This got transcoding back in business... I will paste some of its conclusions below.

Issue could be specific to Pascal-generation GTX 10xx GPUs
GPU model (GTX 1050)

1. Unraid version (7.2.4)

2. driver that broke it (595.45.04)

3. driver that fixed it (580.119.02)

4. Plex container (linuxserver/plex)

5. confirmation that NVENC works again

@ich777 are you aware of this...? Would this affect other generations of GPUs? And other future versions of the driver?

I have hard-locked this driver version, so I have no concerns having driver update itself and breaking. Hopefully this will help others if they run into the same issue.

Oh... there is a little mystery. Before upgrading unRAID, I had Extra Parameter --runtime=nvidia. Plex worked perfectly. After upgrading, without me realizing, this crashed the Plex docker every time. I removed the parameter and Plex worked perfectly again. After all this rigamarole with the driver version issue; it is the opposite situation. If I do NOT have this in ExtraParameters, Plex will crash upon startup. If I do have it, the Plex works nicely. Not sure what this means...

Thank you all!

39 minutes ago, hernandito said:

I manually updated by removing the old plugin

What did you do exactly? What do you mean manually updating by removing the old plugin?

39 minutes ago, hernandito said:

After about 2 hours w/ ChatGPT

Your time would be better invested making a small post here because ChatGPT has proven to be almost always wrong about the Nvidia Driver for Unraid.

40 minutes ago, hernandito said:

driver that broke it (595.45.04)

Man... Driver version 590+ won't work with 10 series cards anymore, they will only work with RTX20 series. :D

You can only use drivers including 580 series with your 10 series cards, they are now legacy and you should consider buying a new card somewhere in the near future.

ChatGPT has proven to make false suggestions again and wasted your time, sorry for you about that but you would be better of making a short post here next time. :)

Next time please include your Diagnostics.

14 minutes ago, ich777 said:

What did you do exactly? What do you mean manually updating by removing the old plugin?

Your time would be better invested making a small post here because ChatGPT has proven to be almost always wrong about the Nvidia Driver for Unraid.

Man... Driver version 590+ won't work with 10 series cards anymore, they will only work with RTX20 series. :D

You can only use drivers including 580 series with your 10 series cards, they are now legacy and you should consider buying a new card somewhere in the near future.

ChatGPT has proven to make false suggestions again and wasted your time, sorry for you about that but you would be better of making a short post here next time. :)

Next time please include your Diagnostics.

I learned a big lesson… I followed the old tutorial from SpaceInvader about transcoding with plex. He mentioned adding the —runtime=nvidia. Worked great… when I upgraded to 7.2. I asked AI, do I need the driver plugin since I thought unRAID now includes it… it told me “no, you can remove it”. So I did… and of course it was wrong… and I had to re-install again and it had the new version, which did not work.

I know it’s an old card… Do you remember unRAID during its infancy…? The premise was, “do you have an old pc and old hdds? You can make into a file/media server!” This was before dockers or anything. Fast forward to a week ago, i had an old GPU in a box under my bed…. And i thought, I can use it to transcode for my Plex friends. They always run into glitches and I get the calls; it’s not working! I thought it couldn’t hurt. I can’t bring myself to spend money on a GPU just for transcoding every once in a while. Oh… and I just started to get more 4K, Atmos, hdr10+, hecv, because I upgraded tv and avr. Seems like with latest plex, my friends were having a hard time playing my content…. Myself? I use Kodi htpc’s to play direct from my shares. AI told me that plex has been making changes and it was being stricter about what it would play. I went back to a release from a couple weeks back, and sure enough, the newer media worked again. And I thought hardware transcoding could help that.

Am I the only one addicted to AI? It’s almost like I ask it everything, like “which hand to use when I go to the bathroom….” Anyway, let’s stay on topic here…

Thank you my friend.

Hi,

On Unraid 7.2.3, I upgraded from open source nvidia kernel module 580.95.05 to open source nvidia kernel module 590.48.01, and run into problems with ollama not working with both GPU (3060, and 5060).

See the below in the dmesg:

[ 1816.413718] NVRM: iovaspaceDestruct_IMPL: 1 left-over mappings in IOVAS 0x400

[ 1816.435493] NVRM: GPU1 nvAssertFailedNoLog: Assertion failed: pIOVAS != NULL @ io_vaspace. c:592

[ 1816.435576] NVRM: GPU1 nvAssertFailedNoLog: Assertion failed: pIOVAS != NULL @ io_vaspace. c:601

[ 2374.905012] NVRM: GPU1 kbifInitLtr_GB202: LTR is disabled in the hierarchy

I would like to go back to the previous version of open source nvidia module 580.95.05 but it is no longer available as a selection under " open source driver " in the nvidia driver settings page, and has been replaced with 590.48.01, which I am having problems with. Selecting the 580.95.05 version there seems to load the nvidia proprietary version and not the open source version.

I did some searching, and it found the some notes on editing /boot/config/plugins/nvidia-driver/settings.cfg and setting driver_version=580.95.05 however that does not seem to be working either.

Is there anyway to just undo / go back to the previous driver version for now?

edit: figured it out:

downloaded both of the below

wget https://api.github.com/repos/unraid/unraid-nvidia-driver/releases/tags/6.12.54-Unraid/nvos-580.95.05-6.12.54-Unraid-1.txz

https://github.com/unraid/unraid-nvidia-driver/releases/download/6.12.54-Unraid/nvos-580.95.05-6.12.54-Unraid-1.txz.md5

and copy both into

/boot/config/plugins/nvidia-driver/packages/6.12.54

next edit

/boot/config/plugins/nvidia-driver/settings.cfg

first_installation=no

driver_version=latest_nos

disable_xconfig=false

update_check=false

finally reboot

Edited March 19Mar 19 by jc179
update answer

On 1/20/2026 at 11:47 AM, ich777 said:

Oh sorry I've overlooked your comment... :/

Yes, but this has nothing to do with the plugin, you have to either add an entry to your syslinux.config or you have to add a modprobe.d file to Unraid, please issue this command from a Unraid terminal and reboot afterwards:

mkdir -p /boot/config/modprobe.d
echo "options nvidia-drm modeset=1" > /boot/config/modprobe.d/nvidia-drm.conf

After the reboot you should be able to do this from an Unraid terminal to verify it was applied correctly:

cat /sys/module/nvidia_drm/parameters/modeset

and it should either revert Y IIRC.

So I am hoping someone can help - I am trying to get this modeset=1 configured so I can run the Eden template for nintendo switch emulation from the appstore, however, I cannot seem to get the nvidia GPU to passthrough. I've attached my diagnostics.

Am I doing something incorrectly?

media-server-diagnostics-20260319-1310.zip

Hey Team,

I've been having an issue with my notifications for a while.

ENOENT: no such file or directory, scandir '/tmp/notifications/unread'

I booted into Safe Mode, and the issue went away. I have been slowly removing plugins to identify which one is breaking my notifications, and I think it's the Nvidia driver plugin. It was the last one removed because I wanted my Jellyfin Docker to stay up. Once I removed it, the issue didn't happen. When I reinstalled it, the notifications issue came back.

Thank you

homeserver-diagnostics-20260320-1216.zip

6 hours ago, pete748 said:

and I think it's the Nvidia driver plugin.

The Nvidia driver Plugin doesn't do anything to the /tmp directory other than saving the version numbers.

6 hours ago, pete748 said:

I wanted my Jellyfin Docker to stay up

Was Jellyfin on while testing?

Are you sure it isn't Jellyfin because Inknow some people using the /tmp directory in the host for the transcoding cache and that can also cause that.

On 3/11/2026 at 7:55 AM, ich777 said:

If this happens usually something is wrong with the hardware or at least some incompatibility issue.

Please try to load the default settings in the BIOS if possible, after that make sure that the boot order is correct, Resizable BAR Support and Above 4G Decoding is enabled.

If you have these AER issues in your log you can actually disable them since at times the cause issues. Add this to your syslinxu.config (at the end of your append line) :

pcie=noaer

However I would not recommend enabling ASPM when you get AER errors afterwards since this indicates some kind of Firmware issue.

Since you have Xid error 79 logged you can take a look here what this means, however it can be all:

https://docs.nvidia.com/deploy/xid-errors/analyzing-xid-catalog.html

However please also check if your Power Supply is still up to the task since they degrade over time and a few issues where solved by swapping the power supply for a new one.

i removed the GPU from the jellyfin container and haven't gotten another crash yet... ill keep testing.

Is there a way I can use an older version of the driver? v580.xx is too new for my kepler card. So I would need to get an older driver. I have no Idea how I could do this. Note: I do not want the card to do transcoding as this would be a waste of time. I want to use it to do console gpu rendering for VM machines.

5 minutes ago, ftrueck said:

Is there a way I can use an older version of the driver? v580.xx is too new for my kepler card. So I would need to get an older driver.

No since the older drivers won't compile against newer Kernels and therefore you can't select an older driver, please also see the pinned post:
https://forums.unraid.net/topic/98978-plugin-nvidia-driver/page/193/#findComment-1551263

Oh, well. Then I probably have to learn to live without the GPU rendering for now. Thank you for the quick response though!

8 minutes ago, ftrueck said:

Oh, well. Then I probably have to learn to live without the GPU rendering for now. Thank you for the quick response though!

Maybe look into something newer like a Nvida T400 or similar since these cards are future proof, at least until Nvidia deprecates them again but I don't think that this will happen soon since this card for example is based on Turing.