[Plugin] Nvidia-Driver

[yakboyslim]

Complete. Still not working, but at least Nvidia knows I am on the most current and maybe can troubleshoot further with them as well.

[yakboyslim]

5 hours ago, yakboyslim said:

Complete. Still not working, but at least Nvidia knows I am on the most current and maybe can troubleshoot further with them as well.

Got it working finally. Rolled back my BIOS version from 1.70 to 1.40. Took me forever to get it to boot, but eventually found out CSM is broken in this older BIOS version.

 

Disabled CSM, unraid booted fine and everything works now! So apparently Asrock broke multiple GPU in an update at some point. Either way, I have one GPU working in dockers, and one passed through to the VM and everything else seems to work as before!

 

Thanks for the help! Now to learn all these other PCIE passthroughs...

[ich777]

11 hours ago, yakboyslim said:

Got it working finally. Rolled back my BIOS version from 1.70 to 1.40. Took me forever to get it to boot, but eventually found out CSM is broken in this older BIOS version.

I hope you don't mind me asking this question but are you willing to upgrade to unRAID v6.10.0-rc2 and see if it everything works there too.

 

This will actually help me a lot... and you always can downgrade back to 6.9.2 again if you want to.

I have a few reports about issues with two Nvidia cards like from @Wingede.

[yakboyslim]

On 11/10/2021 at 7:43 AM, ich777 said:

I hope you don't mind me asking this question but are you willing to upgrade to unRAID v6.10.0-rc2 and see if it everything works there too.

I'd be willing to. I have never done an upgrade before so I might need a guide, etc. but I would be willing to try it.

[ich777]

6 minutes ago, yakboyslim said:

I'd be willing to. I have never done an upgrade before so I might need a guide, etc. but I would be willing to try it.

A new Nvidia driver dropped yesterday, IIRC v470.86

 

To upgrade unRAID to 6.10.0-rc2 simply go to Tools -> Update OS -> in the drop down select "Next" wait a little and a upgrade button should appear.

[AceRimmer]

Is there any way to shut down the Nvidia driver plugin without rebooting? I've used the plugin for Tdarr but my GPU is passed back to my main VM now and my sys log is endlessly filling up.

 

[ich777]

4 minutes ago, AceRimmer said:

Is there any way to shut down the Nvidia driver plugin without rebooting? I've used the plugin for Tdarr but my GPU is passed back to my main VM now and my sys log is endlessly filling up.

No, that's for example one thing why I don't recommend using one card with Docker and VMs, but I think you are running into this issue because of something else.

 

• Do you have the GPU Statistics plugin installed?
• Do you still use Tadarr currently to convert files?
• Do you still use any other container that uses the card when the VM is not turned on?

[AceRimmer]

1 minute ago, ich777 said:

No, that's for example one thing why I don't recommend using one card with Docker and VMs, but I think you are running into this issue because of something else.

 

• Do you have the GPU Statistics plugin installed?
• Do you still use Tadarr currently to convert files?
• Do you still use any other container that uses the card when the VM is not turned on?

 

I understand, my other card is an AMD and its not great for gaming, hence using my Nvidia card for pretty much everything. 

• Do you have the GPU Statistics plugin installed? Yes
• Do you still use Tadarr currently to convert files? Tdarr and Tdarr node dockers are installed but both are turned off and not set to auto run
• Do you still use any other container that uses the card when the VM is not turned on? No

[ich777]

8 minutes ago, AceRimmer said:

my other card is an AMD and its not great for gaming

Sell the card and buy a Nvidia T400, you should get them brand new below MSRP, I payed for mine about € 135,-

Then you can use the card for transcoding <- this card is also very power efficient, is based on Turing and has a few RTX cores from what I know.

 

8 minutes ago, AceRimmer said:

Do you have the GPU Statistics plugin installed? Yes

Try to uninstall the plugin and reload the Dashboard page if you have a tab with it open.

It is more than possible that the GPU Statistics plugin is causing that.

[AceRimmer]

On 11/11/2021 at 7:37 PM, ich777 said:

Sell the card and buy a Nvidia T400, you should get them brand new below MSRP, I payed for mine about € 135,-

Then you can use the card for transcoding <- this card is also very power efficient, is based on Turing and has a few RTX cores from what I know.

 

Try to uninstall the plugin and reload the Dashboard page if you have a tab with it open.

It is more than possible that the GPU Statistics plugin is causing that.

 

Ok ill uninstall and monitor the logs. 

Thanks for the GPU advice, i'll check it out now. 

 

-------------------------------------------------------------------------------------------------------------------------------------------------------

 

UPDATE: 

Uninstalling the GPU Statistics plugin cleared up the log. I also found a good price for a Nvidia T400 coming in at €137.10 so im going to go with that card. Benchmarks look good, it seems to perform close to the Nvidia P620 so its great value for money considering its only a 30w card. Thank you @ich777

Edited November 12, 2021 by AceRimmer

[yakboyslim]

1 hour ago, ich777 said:

A new Nvidia driver dropped yesterday, IIRC v470.86

 

To upgrade unRAID to 6.10.0-rc2 simply go to Tools -> Update OS -> in the drop down select "Next" wait a little and a upgrade button should appear.

6.10.0-rc2 and the v470.86 work great for me.

[ich777]

6.10.0-rc2 and the v470.86 work great for me.

The VM also works flawlessly? May I ask if you can upload a Diagnostocs from your 6.10.0-rc2 installation?

Sent from my C64

[mortist]

Hi Team,

 

I am looking for some support with Plex Hardware transcoding. 

In short, I have followed all of the available guides and I think I have everything setup properly, but when Plex begins to transcode I can see the activity start in 'watch nvidia-smi' and then the GPU goes idle again.

 

I do have Plex Pass, I do have the Nvidia driver installed.

Using unRAID 6.9.2 

Dell R530

Quadro P2000

 

I have tried the Plex docker, Binhex and Linuxserver, all seem to behave the same. I see them call the GPU and then it goes idle shortly afterwards. 

 

I have tried all three Nvidia code branches, but I found no change.

 

I appreciate any support or guidance!

 

The Plex logs are the only clue I have found so far:

 

Nov 11, 2021 06:43:22.715 [0x1529ff51eb38] DEBUG - [Transcode] Codecs: testing h264 (decoder) with hwdevice vaapi
Nov 11, 2021 06:43:22.716 [0x1529ff51eb38] DEBUG - [Transcode] Codecs: hardware transcoding: testing API vaapi
Nov 11, 2021 06:43:22.716 [0x1529ff51eb38] DEBUG - [Transcode] Codecs: hardware transcoding: opening hw device failed - probably not supported by this system, error: Generic error in an external library
Nov 11, 2021 06:43:22.717 [0x1529ff51eb38] DEBUG - [Transcode] Codecs: testing h264 (decoder) with hwdevice nvdec
Nov 11, 2021 06:43:22.718 [0x1529ff51eb38] DEBUG - [Transcode] Codecs: hardware transcoding: testing API nvdec

 

Including screenshots and logs:

 

 

Plex_Media_server_short.txt tower-diagnostics-20211111-0633.zip

[ich777]

12 hours ago, mortist said:

I am looking for some support with Plex Hardware transcoding.

From what I see in the logs you configured everything correctly.

You've cut the Plex log a little to early, what does it say after it probed NVENC?

So the GPU is called only for a short time amd then goes to idle if I understamd that right?

Pleasy check if you have Throttle transcoding turned on in Plex, if yes this is a pretty normal behaviour since it transcodes a few minutes ahead and if the buffer runs empty it transcodes again a few minutes ahead and so on...

Sent from my C64

 

[mortist]

8 hours ago, ich777 said:

From what I see in the logs you configured everything correctly.

You've cut the Plex log a little to early, what does it say after it probed NVENC?

So the GPU is called only for a short time amd then goes to idle if I understamd that right?

Pleasy check if you have Throttle transcoding turned on in Plex, if yes this is a pretty normal behaviour since it transcodes a few minutes ahead and if the buffer runs empty it transcodes again a few minutes ahead and so on...

Sent from my C64

 

 

 

Thanks for your support.

 

I've included more logs as attachments.

 

In the Plex logs around 7:45:40 I started a normal playback without transcoding and the stream is fine.

 

At approximately 7:46:00 I tried to force the client to transcode to a lower quality and the video simply goes black and never starts.

 

You can see that Plex calls Nvidia for 1-2 seconds, but then the GPU goes idle again. 

 

I cannot find anything more than "Generic error in an external library". 

 

Is there some other log where more GPU details are stored?

 

 

Transcode throttle buffer is set to 6000

 

 

nvidia_Console_output.txt Plex_Media_server_short_11132021.txt nvidia-smi_-l_1.txt nvidia_status.txt

[ich777]

40 minutes ago, mortist said:

In the Plex logs around 7:45:40 I started a normal playback without transcoding and the stream is fine.

 

At approximately 7:46:00 I tried to force the client to transcode to a lower quality and the video simply goes black and never starts.

 

You can see that Plex calls Nvidia for 1-2 seconds, but then the GPU goes idle again. 

 

I cannot find anything more than "Generic error in an external library". 

 

Is there some other log where more GPU details are stored?

This looks like a Plex related error and I simply can't help.

You are trying this from a PC or better speaking from a Web Client or am I wrong?

Try it instead from a mobile device where the Plex app is installed, had this exact same issues in the past and this is related to the Browser/WebClient.

 

If you are willing to test you can try and install my Jellyfin container and pass through your GPU and enable HW transcoding there (how-to is in the second post from this thread).

[mortist]

2 hours ago, ich777 said:

This looks like a Plex related error and I simply can't help.

You are trying this from a PC or better speaking from a Web Client or am I wrong?

Try it instead from a mobile device where the Plex app is installed, had this exact same issues in the past and this is related to the Browser/WebClient.

 

If you are willing to test you can try and install my Jellyfin container and pass through your GPU and enable HW transcoding there (how-to is in the second post from this thread).

 

 

Ah man, Solid advice.
 

Yes I was using a Chrome Web browser and it would fail every time.

 

I tried from my phone and boom- worked no issue. 

 

I also have another client running right now, a FireTV transcoding with no issues.

 

Well, that's bad news for anyone trying to HW transcode in Chrome!

 

I'm going to leave HW transcoding turned on. 

 

Thanks for your advice! I wasn't even thinking to try a different client.

 

 

[nilsf2000]

I know it is not officially supported but I am running Unraid in a VM on Proxmox and passing a Quadro P400 to Unraid.

The Device is correctly appearing unter system devices but the Nvidia Driver ist not detecting it.

 

'Nvidia-smi' is also outputting No devices were found

 

Anyone got any Tips?

 

(The GPU was working fine when I was running Unraid directly on the Host and the GPU is outputting the Bios etc. correctly so I don't thing it is a Hardware Failure. I switched the GPU to another Pcie slot with no change.)

 

Thanks for any help

[wgstarks]

I’m using my GTX-1660 Super with the TDARR docker for transcoding and it seems to be working ok in my initial testing but this plugin doesn’t display any info while transcoding. Any suggestions for fixing this?

[ich777]

9 hours ago, nilsf2000 said:

I know it is not officially supported but I am running Unraid in a VM on Proxmox and passing a Quadro P400 to Unraid.
The Device is correctly appearing unter system devices but the Nvidia Driver ist not detecting it.
 
'Nvidia-smi' is also outputting No devices were found
 
Anyone got any Tips?
 
(The GPU was working fine when I was running Unraid directly on the Host and the GPU is outputting the Bios etc. correctly so I don't thing it is a Hardware Failure. I switched the GPU to another Pcie slot with no change.)
 
Thanks for any help

Please post you Diagnostics.

Sent from my C64
 

[ich777]

I’m using my GTX-1660 Super with the TDARR docker for transcoding and it seems to be working ok in my initial testing but this plugin doesn’t display any info while transcoding. Any suggestions for fixing this?

What do you mean exactly? The Nvidia Driver plugin doesn't display any information about usage.

Sent from my C64

[wgstarks]

7 hours ago, ich777 said:

What do you mean exactly? The Nvidia Driver plugin doesn't display any information about usage.

Sent from my C64
 

I had thought that GPU Statistics on the dashboard was linked to this plugin. Actually, after several crashes last night while changing driver versions the statistics seem to be working. The production branch (v470.86) seems to be very unstable though. My system froze when trying to do a clean reboot to install the driver and then crashed again sometime during the night. I’ve switched back to latest branch. Hopefully that will resolve this.🤞

[ich777]

4 minutes ago, wgstarks said:

I had thought that GPU Statistics on the dashboard was linked to this plugin.

It is linked to this topic:

 

4 minutes ago, wgstarks said:

The production branch (v470.86) seems to be very unstable though.

Can't confirm this since I'm running it for almost 4 days and have no crashes whatsoever, are you sure that the Nvidia Driver plugin causes this?

I'm on unRAID v6.10.0-rc2 with Nvidia driver v470.86

 

If you experience any crashing please post your Diagnostics.

 

6 minutes ago, wgstarks said:

I’ve switched back to latest branch.

Latest is actually the "Next Feature Branch" at time of writing.

[wgstarks]

10 minutes ago, ich777 said:

Latest is actually the "Next Feature Branch" at time of writing.

Yeah. I noticed this last night in the settings and thought it would probably be better to switch to the production branch for stability. A few seconds after changing the setting on the plugin my shares disappeared, reappeared and then were gone again, so I tried to do a clean reboot. It froze. Had to force a hard shutdown. After rebooting a parity check started and my intention was to investigate the crash after the parity check had completed but the system crashed again at 2:23am according to automated notifications. When I woke this morning I noticed that my server was frozen again. Had to do another hard shutdown to resolve this (no access via web, ipmi, ssh).

 

Since the entire event was immediately preceded by me changing the desired plugin version I decided to switch back to the latest branch which I had been using for several days without any issues. Really not sure what the underlying cause of the event was. I know the driver doesn’t actually get installed until the system reboots so I’m assuming that changing branches in the plugin settings shouldn’t have any effect on the system. The TDAR docker was actually running and using the GPU for transcoding when I changed the setting initially. I wasn’t planning to do the reboot until it had finished but that didn’t work out.🤨

[ich777]

5 minutes ago, wgstarks said:

A few seconds after changing the setting on the plugin my shares disappeared, reappeared and then were gone again, so I tried to do a clean reboot. It froze.

This seems like a issue from Tdarr from what I know, please see here:

Please also note that switching the branches in the Nvidia Driver plugin does nothing else then change the version in the config file from the driver plugin which is located on the USB Boot device.

 

7 minutes ago, wgstarks said:

When I woke this morning I noticed that my server was frozen again. Had to do another hard shutdown to resolve this (no access via web, ipmi, ssh).

Have you access to a display that is connected directly to the server. I'm really curious what causes this...

 

8 minutes ago, wgstarks said:

The TDAR docker was actually running and using the GPU for transcoding when I changed the setting initially. I wasn’t planning to do the reboot until it had finished but that didn’t work out.🤨

I think this is more related to Tdarr (issue linked above):

 

 

Have you tried Unmanic yet if this happens there too?