[Plugin] Nvidia-Driver


ich777

Recommended Posts

On 1/22/2021 at 12:34 AM, ich777 said:

I think the P400 is the perfect choice it only draws 30W (but please be sure to get a P400 and not a 400 those two cards look almost the same, a 400 is not capable of hw transcoding).

 

Here are the specs:

https://www.pny.eu/consumer/explore-all-products/nvidia-quadro/895-nvidia-quadro-p400

 

This seems perfectly fine to me.

Is this the first time you try hw transcoding?

Oh man

I got a 1030 instead but now I run into a different issue. It has a massive heatsink that won't be able to fit inside my R420 :). Looking to replace it with a much more slimmer version

IMG_3243.jpeg

Edited by rainfactor
Link to comment

Hi, I am currently running 6.9-rc2 with a GTX 1050ti transcoding works but somehow after a few it crashes and it says under the nvidia plugin panel:

 

Unable to determine the device handle for GPU 0000:01:00.0: GPU is lost. Reboot the system to recover this GPU

 

log:

Jan 25 23:18:36 iNAS kernel: NVRM: GPU at PCI:0000:01:00: GPU-86d42cfe-6f24-1e3c-9422-454ce39a29b6
Jan 25 23:18:36 iNAS kernel: NVRM: Xid (PCI:0000:01:00): 79, pid=22174, GPU has fallen off the bus.
Jan 25 23:18:36 iNAS kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
Jan 25 23:18:36 iNAS kernel: NVRM: A GPU crash dump has been created. If possible, please run
Jan 25 23:18:36 iNAS kernel: NVRM: nvidia-bug-report.sh as root to collect this data before
Jan 25 23:18:36 iNAS kernel: NVRM: the NVIDIA kernel module is unloaded.

 

lspci -k

 

01:00.0 VGA compatible controller: NVIDIA Corporation GP107 [GeForce GTX 1050 Ti] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] GP107 [GeForce GTX 1050 Ti]
        Kernel driver in use: nvidia
        Kernel modules: nvidia_drm, nvidia
01:00.1 Audio device: NVIDIA Corporation GP107GL High Definition Audio Controller (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] GP107GL High Definition Audio Controller

 

Does anyone encounter this problem?  TY.

inas-diagnostics-20210125-2321.zip

Link to comment
15 minutes ago, emersonicus said:

Hi, I am currently running 6.9-rc2 with a GTX 1050ti transcoding works but somehow after a few it crashes and it says under the nvidia plugin panel:

 

 

Does anyone encounter this problem?  TY.

Seems there is something wrong with physical connection itself, have you connected it through a riser cable or something?

Also please try to reseat GPU in the PCIe slot.

 

Have you already tried to reboot?

Link to comment
19 minutes ago, emersonicus said:

Hi, I am currently running 6.9-rc2 with a GTX 1050ti transcoding works but somehow after a few it crashes and it says under the nvidia plugin panel:

 


Unable to determine the device handle for GPU 0000:01:00.0: GPU is lost. Reboot the system to recover this GPU

 

log:


Jan 25 23:18:36 iNAS kernel: NVRM: GPU at PCI:0000:01:00: GPU-86d42cfe-6f24-1e3c-9422-454ce39a29b6
Jan 25 23:18:36 iNAS kernel: NVRM: Xid (PCI:0000:01:00): 79, pid=22174, GPU has fallen off the bus.
Jan 25 23:18:36 iNAS kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
Jan 25 23:18:36 iNAS kernel: NVRM: A GPU crash dump has been created. If possible, please run
Jan 25 23:18:36 iNAS kernel: NVRM: nvidia-bug-report.sh as root to collect this data before
Jan 25 23:18:36 iNAS kernel: NVRM: the NVIDIA kernel module is unloaded.

 

lspci -k

 


01:00.0 VGA compatible controller: NVIDIA Corporation GP107 [GeForce GTX 1050 Ti] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] GP107 [GeForce GTX 1050 Ti]
        Kernel driver in use: nvidia
        Kernel modules: nvidia_drm, nvidia
01:00.1 Audio device: NVIDIA Corporation GP107GL High Definition Audio Controller (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] GP107GL High Definition Audio Controller

 

Does anyone encounter this problem?  TY.

inas-diagnostics-20210125-2321.zip 109.63 kB · 0 downloads

I had the same issue with my gtx 1650 turbo. 
I was unable to fix it. 

Link to comment
4 minutes ago, ich777 said:

Seems there is something wrong with physical connection itself, have you connected it through a riser cable or something?

Also please try to reseat GPU in the PCIe slot.

 

Have you already tried to reboot?

 

Yes I tried many times to reboot and reseat the gpu once. I also tried to use the GPU in VM and everything worked fine not crashes but when I switched it as a transcoding gpu it works then crashes randomly while watching in plex.

Link to comment
1 hour ago, emersonicus said:

 

Yes I tried many times to reboot and reseat the gpu once. I also tried to use the GPU in VM and everything worked fine not crashes but when I switched it as a transcoding gpu it works then crashes randomly while watching in plex.

But you are not running it in a VM and also in Docker at the same time?

 

 

@bellyup & @emersonicus if I build both of you a custom verison of Unraid with the Nvidia drivers integrated can you try it?

Link to comment
1 hour ago, ich777 said:

But you are not running it in a VM and also in Docker at the same time?

 

 

@bellyup & @emersonicus if I build both of you a custom verison of Unraid with the Nvidia drivers integrated can you try it?

I’m sorry you are correct. I’m not running unraid in a vm; however, the issue of it starting to transcode then crashing is what happens to me also. 
 

I was just going to wait for the full 6.9 release before doing anything else but I’m happy to help if you want any further testing. 

Link to comment
12 hours ago, ich777 said:

But you are not running it in a VM and also in Docker at the same time?

 

 

@bellyup & @emersonicus if I build both of you a custom verison of Unraid with the Nvidia drivers integrated can you try it?

 

Hi, yes i am not running both at the same time only run VM once with out the nvidia plugin just to check if the GPU is faulty after I checked that is not a GPU problem, I deleted the VM and installed nvidia plugin and or docker transcoding in plex and it works. I observed when plex is transcoding and tdarr also then suddenly the random crashes appears. Maybe that is correlated to the problem?

Link to comment
2 hours ago, emersonicus said:

I observed when plex is transcoding and tdarr also then suddenly the random crashes appears.

I'm not too familiar with tadarr but is the problem also appearing when you only let Plex transcode?

 

2 hours ago, emersonicus said:

if the GPU is faulty

Have you also installed the proper driver and did you stress test it, since the basic display output works mostly fine but after the driver installation the problems appear like in this post:

 

Are you able or better speaking willing to test a custom image with the drivers builtin? :)

Link to comment
2 hours ago, ich777 said:

I'm not too familiar with tadarr but is the problem also appearing when you only let Plex transcode?

 

Have you also installed the proper driver and did you stress test it, since the basic display output works mostly fine but after the driver installation the problems appear like in this post:

 

Are you able or better speaking willing to test a custom image with the drivers builtin? :)

 

Yes, i tried it with heaven benchmark for 20-30 minutes and didnt crash once.

I am willing to test a custom image with the drivers builtin if there is a guide or hint how to do so. TY! :)
 

Link to comment
On 1/26/2021 at 11:27 AM, emersonicus said:

 

Yes, i tried it with heaven benchmark for 20-30 minutes and didnt crash once.

I am willing to test a custom image with the drivers builtin if there is a guide or hint how to do so. TY! :)
 

Is the build working that I've sent you?

Link to comment
  • 2 weeks later...

Hi,

 

Apologies if this is the wrong place to post,

 

I have been looking into upgrading my current Unraid server from the current 6.8.3 unraid nvidia in which i am using a Nvidia k4000 graphics card for pass through to plex docker.  Setup works perfectly on existing version so given the existing unraid nvidia does not appear to be available anymore i setup a test unraid server using 6.9 rc on a spare system i had lying around however no matter what i try the K4000 graphics card will not be seen on the unraid nvidia plugin.

 

I have tried 2 different k4000 gpu with no success however when i tried with 2 different K620 gpu they are both seen under settings and work fine when either is passthrough to docker. I Checked driver support and the K4000 is listed as supported under that version installed which i believe was V4.55.45.01 but no matter what i try i cannot see the k4000 only the NVIDIA SMI failed as could not communicate with the NVIDIA driver.

 

As a last ditch effort i even tried on both Intel and Ryzen to make sure this was not a platform or motherboard issue.  

 

Is there anyone else who has tried to use a k4000 graphics card with the latest unraid version for nvidia passthrough ?

Link to comment
15 minutes ago, Neil C said:

Is there anyone else who has tried to use a k4000 graphics card with the latest unraid version for nvidia passthrough ?

Passthrough is the wrong expression, use it in Containers fits it more... :)

 

Can you test two things for me?

Hook me up with a short PM and I will send you the links if possible. :)

 

You are now on Unraid 6.9.0rc2?

Link to comment

Hi,

 

As per our PM's this issue is now resolved.  Turned out that the card i was trying to use in the test server and also one i borrowed from friend were quadro 4000's not quadro k4000's like the one used in my main server.  I physically took the card from my working server running 6.8.3 and it was seen without issues so this is entirely my error.

 

Thanks again for your time and support a donation has been sent to support your future work.

 

Best Wishes

 

neil

 

  • Like 1
Link to comment

  

58 minutes ago, ich777 said:

Have you installed this in a Docker container or in a VM?

If so please send me the resources or what you have installed exactly so that I can look into it.

 

Hi @ich777

 

Thank you for getting back to me.

 

I am doing this via a Docker Container. Which I got the instructions from here https://www.spxlabs.com/blog/2020/5/22/unraid-obs-capture-encoding-and-streaming-server

 

I do have NDI setup, and I followed the instructions on the first page to add NVIDIA Visible Container etc 

 

but I am not sure if I am supposed to do this to both containers that get installed via the link above or not? 

 

The reason why I say this is because when I add the relevant nvidia driver info to the docker-obs-ndi I can see select NVENC in OBS but then the screen is blank even though NDI is on. And I can see my CPU usage drop drastically (i'm assuming GPU is doing it's work) but as mentioned, screen is blank. 

 

BUT, if I remove it from docker-obs-ndi, and then use the NVIDIA Variables in the 'edit' section of docker-obs I have image and even though NVENC is selectable it isn't using the GPU but CPU still. I feel I am close but at a dead end. 

 

If you need more info let me know. Just too add when I type nvidia-smi into Terminal it doesn't show anything is used which is odd but could explain the issue. But not sure.

Edited by z0ki
Link to comment
20 minutes ago, z0ki said:

BUT, if I remove it from docker-obs-ndi, and then use the NVIDIA Variables in the 'edit' section of docker-obs I have image and even though NVENC is selectable it isn't using the GPU but CPU still. I feel I am close but at a dead end. 

I think this is because the Container hasen't built in the necessary dependencies for Nvidia hardware acceleration and I would ask on the Github of the creator of the container, not every container has everything built into it for NVENC.

 

You also need the libraries to run NVENC, to pass through the runtime is not always enough.

 

I hope that makes sense to you...

Link to comment
1 hour ago, ich777 said:

I think this is because the Container hasen't built in the necessary dependencies for Nvidia hardware acceleration and I would ask on the Github of the creator of the container, not every container has everything built into it for NVENC.

 

You also need the libraries to run NVENC, to pass through the runtime is not always enough.

 

I hope that makes sense to you...

 

Yeah it does a little thanks mate. So I guess it isn't as cut throat as I thought it would be. I am surprised there isn't a somewhat "official" container for this sort of thing considering how great OBS is via NDI. I was hoping to avoid building a dedicated streaming PC and use my unraid server but I guess it is what it is I suppose. 

Edited by z0ki
Link to comment

Hello I have recently updated to 6.9.0-rc2 and have a GTX970 and a GTX750ti installed. Both were previously visible in 6.8.3 and linuxserver.io's plugin. Since upgrading to rc2 I am not able to see this plugin anymore so I switched.

 

Unfortunately after installation it does not recognize my cards. Under system deviced they are still listed and also driver  v455.45.01 is installed.

210213 unraid nvidia driver plugin 1 install log.png

210213 unraid nvidia driver plugin 2 settings page.png

210213 unraid nvidia driver plugin 3 system devices.png

210213 unraid nvidia driver plugin 4 nvidia-smi.png

Link to comment
4 hours ago, Rayce185 said:

Unfortunately after installation it does not recognize my cards. Under system deviced they are still listed and also driver  v455.45.01 is installed.

Have you already tried already to reboot?

If the command nvidia-smi fails it seems there is something wrong with the download or the installation of the package.

Link to comment
19 minutes ago, ich777 said:

Have you already tried already to reboot?

If the command nvidia-smi fails it seems there is something wrong with the download or the installation of the package.

I have rebooted multiple times already, yes. There has been no change.

 

Maybe it has to do with the 5.10 kernel and the driver being for 5.9?

Is there any way to manually update the driver to the current 460.39, since it is supposed to support kernel 5.10?

https://www.nvidia.com/Download/driverResults.aspx/170134/en-us

Edited by Rayce185
Link to comment

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.