[Plugin] Nvidia-Driver


ich777

1063 posts in this topic Last Reply

Recommended Posts

  • Replies 1.1k
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Nvidia-Driver (only Unraid 6.9.0beta35 and up) This Plugin is only necessary if you are planning to make use of your Nvidia graphics card inside Docker Containers. If you only want to use yo

To utilize your Nvidia graphics card in your Docker container(s) the basic steps are:   Add '--runtime=nvidia' in your Docker template in 'Extra Parameters' (you have to enable 'Advanced

Thanks @bytchslappa for the heads up. Everyone using a User Script for Nvidia GPU Power State just edit the script to: #!/bin/bash nvidia-persistenced fuser -v /dev/nvidia* and leave the s

Posted Images

Just now, bradtn said:

"No devices were found "

Can you check if you power cables for the card are seated correctly, please also try to reseat the card in the PCIe slot.

 

Have you also installed a VM to try if the card works in there?

Link to post
24 minutes ago, ich777 said:

Can you check if you power cables for the card are seated correctly, please also try to reseat the card in the PCIe slot.

 

Have you also installed a VM to try if the card works in there?

I'm not home atm but I did a quick remote login and edited the one vm I do have and I am able to select the Nvidia card in that menu. Have not reseated etc yet. Seeing as I can see it in the vm I shouldn't have to? 

Link to post
Just now, bradtn said:

I'm not home atm but I did a quick remote login and edited the one vm I do have and I am able to select the Nvidia card in that menu. Have not reseated etc yet. Seeing as I can see it in the vm I shouldn't have to? 

I'm also able to see the card if the 8+6pin power connector is not seated right but I can't use it properly.

 

Have you anything done to your syslinux config?

Link to post
5 minutes ago, ich777 said:

I'm also able to see the card if the 8+6pin power connector is not seated right but I can't use it properly.

 

Have you anything done to your syslinux config?

Only thing I've done is add and remove drives and do a new config for that. I'll definitely check power when I get home then

Link to post
Just now, bradtn said:

Only thing I've done is add and remove drives and do a new config for that. I'll definitely check power when I get home then

Eventually something got uplugged or something...

Try also to reseat the card.

Link to post
15 hours ago, ich777 said:

Eventually something got uplugged or something...

Try also to reseat the card.

so I did the reseat and power reseat and upon rebooting into the system and eagerly opening the nvidia driver plugin I screamed with joy! my card was detected as it was before! I went to go resetup my plex docker as it once was. Completed that and went to test the transcoding and run the nvidia-smi watch command and BOOM "no devices were found". Check the plugin page again, gpu disappeared. Rebooted again for fun , no devices were found. Just as a side note my motherboard bios is up to date checked that as well and I also had a DP cable plugged into the gpu as i did read that could cause issues. I checked system log and I did get a bunch of errors that ive never seen before , hopefully you understand it? Thanks for everything

 

I did find this post online which seems to reference the exact issue but this is all chinese to me and they seem to reference that a much earlier driver iteration than the plugin provides "fixes" it

 

https://forums.developer.nvidia.com/t/linux-driver-410-73-gtx-980-nvrm-rminitadapter-failed/67231/25

 

Dec 17 02:58:02 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:58:02 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:58:02 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:58:02 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0
Dec 17 02:58:03 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:58:03 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:58:03 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:58:03 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0
Dec 17 02:58:19 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:58:19 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:58:19 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:58:19 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0
Dec 17 02:58:20 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:58:20 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:58:20 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:58:20 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0
Dec 17 02:59:06 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:59:06 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:59:07 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:59:07 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0
Dec 17 02:59:07 Godzilla kernel: resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000c0000-0x000dffff window]
Dec 17 02:59:07 Godzilla kernel: caller _nv000709rm+0x1af/0x200 [nvidia] mapping multiple BARs
Dec 17 02:59:07 Godzilla kernel: NVRM: GPU 0000:08:00.0: RmInitAdapter failed! (0x26:0xffff:1239)
Dec 17 02:59:07 Godzilla kernel: NVRM: GPU 0000:08:00.0: rm_init_adapter failed, device minor number 0

 

nvidia error.PNG

Edited by bradtn
Link to post
42 minutes ago, bradtn said:

so I did the reseat and power reseat and upon rebooting into the system and eagerly opening the nvidia driver plugin I screamed with joy! my card was detected as it was before! I went to go resetup my plex docker as it once was. Completed that and went to test the transcoding and run the nvidia-smi watch command and BOOM "no devices were found". Check the plugin page again, gpu disappeared. Rebooted again for fun , no devices were found. Just as a side note my motherboard bios is up to date checked that as well and I also had a DP cable plugged into the gpu as i did read that could cause issues. I checked system log and I did get a bunch of errors that ive never seen before , hopefully you understand it? Thanks for everything

 

I did find this post online which seems to reference the exact issue but this is all chinese to me and they seem to reference that a much earlier driver iteration than the plugin provides "fixes" it

 

https://forums.developer.nvidia.com/t/linux-driver-410-73-gtx-980-nvrm-rminitadapter-failed/67231/25

Can you give me your syslinux configuration (in the Main tab click on 'Flash' then scroll a little down and you should see the config, please post a screenshot here).

Link to post
1 minute ago, ich777 said:

@bradtn can you try to uninstall the plugin itself and reboot and see if this error pops up again?

 

Then try to eventually boot a VM with the card assigned.

Are you booting in Legacy or UEFI mode?

Booting uefi mode atm. I'll try the plugin removal now 

Link to post
11 minutes ago, ich777 said:

@bradtn can you try to uninstall the plugin itself and reboot and see if this error pops up again?

 

Then try to eventually boot a VM with the card assigned.

Are you booting in Legacy or UEFI mode?

Plugin removed no more errors in log, editing my one vm and simply switching from vnc to my 1070 and starting produced this. 

Screenshot_20201217-082522_Chrome.jpg

Link to post
Just now, bradtn said:

Plugin removed no more errors in log, editing my one vm and simply switching from vnc to my 1070 and starting produced this.

Seems like there is something wrong with the graphics card...

Have you bound the card do VFIO or something else?

 

The last thing I can recommend is to reset the BIOS and see if this helps, but do this only if you are familar with it!

Link to post
2 minutes ago, ich777 said:

Seems like there is something wrong with the graphics card...

Have you bound the card do VFIO or something else?

 

The last thing I can recommend is to reset the BIOS and see if this helps, but do this only if you are familar with it!

I don't even know what vfio means or is so I suspect no. I'm familiar with bios but define what you mean by reset. Also I do find it strange that so many other people in the old thread started having same issue as me? 

Edited by bradtn
Link to post
7 minutes ago, bradtn said:

I'm familiar with bios but define what you mean by reset.

You can do a hard reset by pulling out the main power line and pulling out the CMOS battery from the motherboard and wait for a minute or you can reset it in the BIOS menu itself but I would recommend to do the first way so that everything is cleared out.

 

7 minutes ago, bradtn said:

Also I do find it strange that so many other people in the old thread started having same issue as me?

This has nothing to do with the plugin itself since the plugin downloads and installs the driver, you can also try my 6.8.3 build from the Unraid-Kernel-Helper thread (first post on the bottom are the prebuilt images).

Also one thing to note if the plugin is now uninstalled and the card isn't working in a VM I can only think of another issue that is not related to the plugin itself.

 

No, not many people, one with the P2000 that was dead on arrival and never worked, the other issues are all resolved.

 

EDIT: Are you using a riser cable or anything like that?

Link to post
24 minutes ago, ich777 said:

You can do a hard reset by pulling out the main power line and pulling out the CMOS battery from the motherboard and wait for a minute or you can reset it in the BIOS menu itself but I would recommend to do the first way so that everything is cleared out.

 

This has nothing to do with the plugin itself since the plugin downloads and installs the driver, you can also try my 6.8.3 build from the Unraid-Kernel-Helper thread (first post on the bottom are the prebuilt images).

Also one thing to note if the plugin is now uninstalled and the card isn't working in a VM I can only think of another issue that is not related to the plugin itself.

 

No, not many people, one with the P2000 that was dead on arrival and never worked, the other issues are all resolved.

 

EDIT: Are you using a riser cable or anything like that?

Nope no riser cables etc. I do have a seperate pc if I put this gpu in that pc and gpu works what would that mean for unraid? In respects to the BIOS reset. Do I pose any risk to my unraid setup in any fashion by doing this? If so what precautions should I take/make? 

Link to post
3 minutes ago, bradtn said:

I do have a seperate pc if I put this gpu in that pc and gpu works what would that mean for unraid?

You can of course try that first, this would be best if you have a spare PC.

 

I also have an old GTX960 with the same error,  if I try it on a different machine I can get sometimes a video output but the driver fails to install in the end.

 

7 minutes ago, bradtn said:

In respects to the BIOS reset. Do I pose any risk to my unraid setup in any fashion by doing this? If so what precautions should I take/make?

No, if you configure your SATA same as before and set the server back to USB boot everything should went fine.

Please eventually check if you got the option above 4G Decoding in your BIOS, if so please enable it.

Link to post
2 minutes ago, ich777 said:

You can of course try that first, this would be best if you have a spare PC.

 

I also have an old GTX960 with the same error,  if I try it on a different machine I can get sometimes a video output but the driver fails to install in the end.

 

No, if you configure your SATA same as before and set the server back to USB boot everything should went fine.

Please eventually check if you got the option above 4G Decoding in your BIOS, if so please enable it.

4g decoding what heading is this typically under? 

Link to post

Ok, so long story short;

Today I upgraded my homeserver, which is running unraid, from a z97 board + i7 4790 to a x99 Board with a i7 6900k.

I was using the 4790´s iGPU for encoding, and at least I think everything was running fine. Not 100% sure if my problems are related to teh upgrade but at least thats what I suspect.

When I tried to connect to my plex server it wasnt online because the Docker wouldnt start. Deleted the /dev/dri thingy from the menu and it started porperly.

But I found a problem, on my living room TV the image looks as it should, on my Nvidia Shield it looks okay but with some artefacting or whatever that is, and on my Office TV there is a lot of artefacting, it just looks horrible.

I´m not 100% sure but I think this is happening because it´s not using the intel iGPU hardwareencode anymore because the 6900k doesnt have one.

So I put in my old GTX 960 to use hardware encoding on it instead of the iGPU.

I was looking for Tutorials on youtube on how to get my Nvidia Card working in unraid and plex, but unfortunately some stuff happened here apparently in November and the nvidia unraid plugin everyone was using successfully is not available anymore. I cant find it in the community Apps thingy.

So now I found this here and it says I need to be on a beta version to use my GPU now.

 

Isn´t there a better way of getting my card to work in my plex docker, without upgrading to a beta?

I dont want to be on a beta version to use a feature that was available before without a beta.

Pls help.

 

Edit

I looked a bit further into it:
My living room TV, phone, PC and Nvidia shield use direct play in Plex.

My Samsung TV in my Office just doesnt, even if I set it to "on".

When I set my Livingroom TV to not use direct play it looks just as awful as the samsung TV.

When set to not using direct play the dashboard says 1080p(h264) converting, thats when it looks awful.

I´m not even sure anymore if using the GPU for encoding will make it better, but again, I feel like I didnt have this problem before and it definately won´t get worse.

 

Any tips?

Edited by derWinky
Link to post
1 hour ago, derWinky said:

Ok, so long story short;

Today I upgraded my homeserver, which is running unraid, from a z97 board + i7 4790 to a x99 Board with a i7 6900k.

I was using the 4790´s iGPU for encoding, and at least I think everything was running fine. Not 100% sure if my problems are related to teh upgrade but at least thats what I suspect.

When I tried to connect to my plex server it wasnt online because the Docker wouldnt start. Deleted the /dev/dri thingy from the menu and it started porperly.

But I found a problem, on my living room TV the image looks as it should, on my Nvidia Shield it looks okay but with some artefacting or whatever that is, and on my Office TV there is a lot of artefacting, it just looks horrible.

I´m not 100% sure but I think this is happening because it´s not using the intel iGPU hardwareencode anymore because the 6900k doesnt have one.

So I put in my old GTX 960 to use hardware encoding on it instead of the iGPU.

I was looking for Tutorials on youtube on how to get my Nvidia Card working in unraid and plex, but unfortunately some stuff happened here apparently in November and the nvidia unraid plugin everyone was using successfully is not available anymore. I cant find it in the community Apps thingy.

So now I found this here and it says I need to be on a beta version to use my GPU now.

 

Isn´t there a better way of getting my card to work in my plex docker, without upgrading to a beta?

I dont want to be on a beta version to use a feature that was available before without a beta.

Pls help.

 

Edit

I looked a bit further into it:
My living room TV, phone, PC and Nvidia shield use direct play in Plex.

My Samsung TV in my Office just doesnt, even if I set it to "on".

When I set my Livingroom TV to not use direct play it looks just as awful as the samsung TV.

When set to not using direct play the dashboard says 1080p(h264) converting, thats when it looks awful.

I´m not even sure anymore if using the GPU for encoding will make it better, but again, I feel like I didnt have this problem before and it definately won´t get worse.

 

Any tips?

For the hardware that can use direct play, use that. No point transcoding something that can be natively played on the device.

 

For the devices that you have to enable transcoding to get it to work, try changing the resolution/bitrate the transcode outputs.

What does the now playing page says about the transcoding? From codec to which codec.

Link to post

I cannot get the driver to work on 6.9RC1. I had the linuxserver driver working on 6.83. I removed that driver and then upgraded to 6.9RC1 following the instructions at the beginning of this thread to install the Nvidia plugin. The screenshot below shows that I have the plugin installed. 

image.thumb.png.2ebe8c7a6754d84b2d195cf9876b0fd7.png

 

I get nothing when I click on the Nvidia driver as seen below;

image.thumb.png.aa5b43758b2dc5dedb6c3643cd0621f5.png

 

Here is the output of nvidia-smi. It took it about 10 minutes to run, but eventually gave the information below;

image.png.45c6030278c4482c543055f7ab1a1740.png

 

image.thumb.png.9c1b329a330d816717f2173e8e5bf65a.png

 

I tried some of the suggestions in this thread such as reseating the card and making sure that was not bound with VFIO. I also rebooted multiple times in addition to disabling Docker. Any other suggestions? 

 

best regards,

Rob

 

 

 

image.png

Edited by unixdork
better screenshot
Link to post
10 hours ago, derWinky said:

Isn´t there a better way of getting my card to work in my plex docker, without upgrading to a beta?

I dont want to be on a beta version to use a feature that was available before without a beta.

Pls help.

You can one of the prebuilt images (first post at the bottom) from here:

 

Download the archive and extract and replace the bz* files from your USB Boot device with the ones from the archive.

To see the UUID of your GPU install, after rebooting your server, the Unraid-Kernel-Helper-Plugin (not the Docker).

 

 

1 hour ago, unixdork said:

I cannot get the driver to work on 6.9RC1. I had the linuxserver driver working on 6.83. I removed that driver and then upgraded to 6.9RC1 following the instructions at the beginning of this thread to install the Nvidia plugin. The screenshot below shows that I have the plugin installed.

Can it be that you have PiHole or something installed in your network?

The Plugin page just shows a empty page (when you click on the Nvidia-Driver symbol)?

 

EDIT: Anyways wouldn't it be better to use the Intel iGPU for this (no restriction of HW transconding streams).

Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.