[Plugin] Nvidia-Driver

[CorneliousJD]

2 minutes ago, ich777 said:

The default one.

 

If you are planing to use them in Docker containers for HW acceleration or LLMs not much difference at all.

 

Latest is fine.

 

Wonderful, thanks for the quick reply. I'll be using it just for HW transcoding and maybe some LLM tinkering in the future.

 

I'm just not clear if I need the nvidia.conf file in the modprobe.d since "latest" right now seems to be the open source driver -- it mentions workstation cards and geforce cards, but the Tesla is technically neither, and is a "Datacenter card" 

 

I think I'm going to just bit the bullet and order the card now.

Edited March 17 by CorneliousJD

[ich777]

3 minutes ago, CorneliousJD said:

I'm just not clear if I need the nvidia.conf file in the modprobe.d since "latest" right now seems to be the open source driver

Can you send me a screenshot please?

Latest should never be the open source one.

 

The open source option is a dedicated option and not bound to latest, production, new feature branch or a static version listed below. The version numbers might be the same but the open source is a completely separate package and should never be latest.

 

Just stick with latest, that should be fine.

[CorneliousJD]

1 minute ago, ich777 said:

Can you send me a screenshot please?

Latest should never be the open source one.

 

The open source option is a dedicated option and not bound to latest, production, new feature branch or a static version listed below. The version numbers might be the same but the open source is a completely separate package and should never be latest.

 

Just stick with latest, that should be fine.

Ok thanks - that's where my confusion was, the Latest and Open Source were the same version number, and I assumed that it was auto-selecting the open source option. 

 

I understand how that I look at it a bit deeper (Today is the first day I've seen this plugin page option)

 

Sorry for the confusion, but thanks for the explanation!  

[ich777]

2 minutes ago, CorneliousJD said:

Sorry for the confusion, but thanks for the explanation!  

No worries, I'm here to help.

 

As said, latest should be fine or if you are more comfortable you can also select the prodcution branch.

[Florris]

Hi everyone, I used my nvidia gpu for a while now together with my jellyfin docker, but from one day to another it stopped working. I made an issue on the jellyfin docker's gh but apparently its an nvidia issue, here is how I start the docker together with the issue.

 

docker run
  -d
  --name='jellyfin'
  --net='bridge'
  -e TZ="Europe/Berlin"
  -e HOST_OS="Unraid"
  -e HOST_HOSTNAME="Bocchi"
  -e HOST_CONTAINERNAME="jellyfin"
  -e 'JELLYFIN_PublishedServerUrl'='192.168.0.5'
  -e 'PUID'='99'
  -e 'PGID'='100'
  -e 'UMASK'='022'
  -l net.unraid.docker.managed=dockerman
  -l net.unraid.docker.webui='http://[IP]:[PORT:8096]'
  -l net.unraid.docker.icon='https://raw.githubusercontent.com/linuxserver/docker-templates/master/linuxserver.io/img/jellyfin-logo.png'
  -p '8096:8096/tcp'
  -p '8920:8920/tcp'
  -p '7359:7359/udp'
  -p '1900:1900/udp'
  -v '/mnt/user/Jellyfin':'/data/tvshows':'rw'
  -v '/mnt/user/Jellyfin':'/data/movies':'rw'
  -v '/mnt/user/jellyfin_data/':'/config':'rw'
  -e NVIDIA_DRIVER_CAPABILITIES=all
  -e NVIDIA_VISIBLE_DEVICES=all
  --runtime=nvidia
  --gpus all 'lscr.io/linuxserver/jellyfin'

0025478b06b8336843ca753b9d4126ff63a73feb1187fe44d44b95b27a50a348
docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: initialization error: open failed: /proc/sys/kernel/overflowuid: permission denied: unknown.

 

(there are no logs since it won't even start)

 

Also I'm not sure which parameters are needed, but no matter which I try it fails to start.

bocchi-diagnostics-20240318-1055.zip

[ich777]

39 minutes ago, Florris said:

Also I'm not sure which parameters are needed, but no matter which I try it fails to start.

Remove your Jellyfin container entirely, go to the Docker page, select Add Container from the bottom, select your Jellyfin container from the drop down and deploy it again.

 

If that doesn't help try to boot your machine without UEFI, check if you have Above 4G Decoding and Resizable BAR Support in your BIOS enabled.

[Laydin]

Hello All, i'm a little bit of an Unraid newbie and i'm trying to unlock the Nvidia transcoding, however i'm not sure how to get to the container screen that is in the instructions here: 

 

I'm not sure if i'm missing something or i need to add a container, which i'm not sure how to do. Or maybe my card is incompatible?

Any advice is greatly appreciated.

 

Thanks.

[Fastcompjason]

I too am having a recent issue where nVidia HW transcoding used to work perfect, but recently I have been troubleshooting system issues and now that I think I have my system stable I have noticed that my Plex container will NOT use the nVidia GPU at all for transcoding and will only transcode video at all if I disable the option in Plex setting for using HW transcoding if available (then it will fall back to CPU transcoding). I like the previous poster have the same green info (different specific GPU of course) in the Nvidia info section of the nVidia driver settings option page in UNraid but when I have the plex setting enabled for "Use hardware acceleration when available" and "Use hardware-accelerated video encoding" Plex will NOT transcode at all and just sits there with a black player screen and no usage of either CPU or GPU (as shown in GPU statistics on the Dashboard of Unraid). What am I missing? To my knowledge I haven't had any changes to the docker edit settings (advanced) for anything and especially didn't change anything for GPU arguments. Can someone spot an issue? One strange thing is that in the nVidia driver settings page, I see that it shows "

latest: v550.40.07

Production Branch: Not found for this unRAID version!"

(I am on Unraid 6.12.6)        

[ich777]

2 hours ago, Laydin said:

I'm not sure if i'm missing something or i need to add a container, which i'm not sure how to do. Or maybe my card is incompatible?

No your card is not incompatible.

Can you please post a screenshot from your Docker template for Plex?

Are you sure that you have entered the Plex claim and see the card in the settings from Plex?

What is does the transcoder log say?

 

BTW are you sure that you are forcing a transcode, if yes how do you test? How do you know that transcoding is not working?

[ich777]

1 hour ago, Fastcompjason said:

I am on Unraid 6.12.6

Please uninstall the plugin, upgrade to 6.12.8, install the plugin again and reboot or restart the Docker service.

[Laydin]

53 minutes ago, ich777 said:

No your card is not incompatible.

Can you please post a screenshot from your Docker template for Plex?

Are you sure that you have entered the Plex claim and see the card in the settings from Plex?

What is does the transcoder log say?

 

BTW are you sure that you are forcing a transcode, if yes how do you test? How do you know that transcoding is not working?

 

I created a Nvidia docker using the plex template as a base for my template choices:

 

and i did add the 2 variables to the Nvidia template:

 

I have nothing in my transcoder log, nor could i get it to show up on my plex transcoder settings.

I also checked in the plex dashboard. I'm not sure what i'm missing to get it to work:

Edited March 20 by Laydin

[ich777]

44 minutes ago, Laydin said:

I also checked in the plex dashboard. I'm not sure what i'm missing to get it to work:

You have to force a transcode, you are using direct play, on your ROKU device you have to lower the quality below 1080p and transcoding should kick in.

 

You also have to have --runtime=nvidia in the template (in Extra Parameters with advanced view enabled in the template).

[Laydin]

1 hour ago, ich777 said:

You have to force a transcode, you are using direct play, on your ROKU device you have to lower the quality below 1080p and transcoding should kick in.

 

You also have to have --runtime=nvidia in the template (in Extra Parameters with advanced view enabled in the template).

Thank You!! I set up the docker as a separate thing versus in the plex docker, it is now working!

[Coreybau]

I recently got a Quadro P5000 and installed it in my server replacing my Quadro P4000. I have a GTX 1050 I use to run the GUI so didn't think anything was wrong until IU went to see what the latest driver was. When I looked I could only see my GTX 1050. So I took out the P5000 and put the P4000 back in. Still only the 1050 is showing. I read for days on this topic and nothing I tried worked. The P4000 worked fine until I tried to switch it out. I was on 6.12.5 so upgraded to 6.12.8. Also installed the Docker fix. Still only the GTX 1050 is showing. So I rebooted this last time and gave up to just ask the forum what is going on. I could always see the P4000, P5000, and a borrowed P6000 in the devices but none of them show up in the nvidia-drivers plug in.
Help please

crbunraid-diagnostics-20240321-0608.zip

[ich777]

57 minutes ago, Coreybau said:

Help please

No support because of this line in your syslog (as also pointed out in the first post - big red text right on top) :

Mar 21 05:45:17 CRBUnraid emhttpd: /usr/local/emhttp/plugins/user.scripts/backgroundScript.sh "/tmp/user.scripts/tmpScripts/unlock nvidia/script" >/dev/null 2>&1

 

Remove the script, reboot and post new Diagnostics.

[Coreybau]

Thanks for the assist. Removed the User Script. Rebooted. Here is my Diags. The problem is still there. 

crbunraid-diagnostics-20240321-1255.zip

[ich777]

2 minutes ago, Coreybau said:

The problem is still there.

Can you spot the issue:

Mar 21 12:40:32 CRBUnraid kernel: NVRM: GPU 0000:05:00.0: GPU does not have the necessary power cables connected.
Mar 21 12:40:32 CRBUnraid kernel: NVRM: GPU 0000:05:00.0: RmInitAdapter failed! (0x24:0x1c:1556)
Mar 21 12:40:32 CRBUnraid kernel: NVRM: GPU 0000:05:00.0: rm_init_adapter failed, device minor number 0

 

[mattiscg]

Do i need this plugin if i want to use my nvida graphic card for Plex hardware accleration? 

[ich777]

2 minutes ago, mattiscg said:

Do i need this plugin if i want to use my nvida graphic card for Plex hardware accleration? 

I would recommend that you read the first post but yes, you do need that plugin if you want to use your Nvidia card for hardware accelerated transcoding in your Plex Docker container.

 

In the second post in this thread is how to set it up in the Docker container.

[Coreybau]

I saw the problem and it didn't make sense. It still doesn't. Everything was connected correctly. So I corrected the other things you said but when I got home I put the P5000 back in. And it starts working. So things turned out the way I wanted in the first place. Thank you again.

[ich777]

15 minutes ago, Coreybau said:

So things turned out the way I wanted in the first place. Thank you again.

Are you really sure that the power supply is up to the task and all cables are all the way plugged in?

 

Anyways, glad that everything is now working as expected for you.

[Coreybau]

I am sure. Everything was plugged in and seated correctly. The card was in the z820 for years. I just wanted to upgrade. Used the adapter for the same power in the P5000 and it just worked. Go figure. 

[captainkegs]

On 3/12/2024 at 12:08 PM, ich777 said:

Please upgrade to 6.12.8 and then post new Diagnostics, before upgrading make sure to go to the Plugins in Unraid, look if you have a Plugins Error tab, if yes delete the plugin and upgrade (please also check your notifications if you got a message before rebooting to upgrade to the new version that the new version is downloaded.

After the upgrade see if the Nvidia driver got installed, if not Reinstall it from the CA App.

 

Please don't do anything manual, this will make troubleshooting much more complicated.

 

Do the steps from above and see if it installs correctly after the upgrade.

 

Upgrading worked. I had a few other issues but was able to work through them. Thanks!

[Ace319]

Hi I have a NVIDIA Quadro FX 580 GPU that im using solely for the purpose of being able to see the GUI on a monitor attached to the server.  Not using to for transcoding or anything as my mobo (asrock p67 extreme 4 gen 3) does not have integrated video output for my i5-2500k.

 

Monitor is connected via displayport to the vid card and I can see the bios>select unraid boot to GUI> then once its all done and booted all I see is a blinking cursor.  Not sure if this was due to a recent update but things have been fine for a long time (i was able to see the GUI and main unraid page just fine).  

 

I tried this plug-in but looks like i need a legacy NVIDIA-Linux-x86_64-340.108.  So how do I install this?  I am quite green to terminal use and what not so I really dont want to mess anything up.  

[ich777]

14 minutes ago, Ace319 said:

but looks like i need a legacy NVIDIA-Linux-x86_64-340.108

Please read the first recommended post on top.

 

15 minutes ago, Ace319 said:

So how do I install this?  I am quite green to terminal use and what not so I really dont want to mess anything up.  

There is no easy way to do that.

 

16 minutes ago, Ace319 said:

Monitor is connected via displayport to the vid card and I can see the bios>select unraid boot to GUI> then once its all done and booted all I see is a blinking cursor.

Sorry but I can‘t help here since this card is simply not supported.

Maybe try to get a slightly newer card like a Quadro P400 or similar that is supported by the plugin.

I would still recommend that you get something like a Nvidia T400 since it‘s a recent card.