Jump to content

[Plugin] Nvidia-Driver


ich777

Recommended Posts

27 minutes ago, ich777 said:

Sorry but this plugin is not designed for Datacenter GPUs.

You can however fork it and create your own plugin for these cards and maybe publish it to the CA App as long as it doesn't violate the Nvidia EULA.

This would be also beneficial to other users.

 

I would be also glad to help if you need anything.

Providing the enterprise drivers publicly would definitely violate the EULA which is why I suggested the custom URL support or local storage support. That way people that have legitimate access to the driver can compile their own version locally. That way the plugin will never violate any EULA.

 

I didn't want to implement a separate plugin since the fork would be exactly the same but with custom URL/share support for driver packages which is why I asked if you would accept such a PR.

 

image.thumb.png.35ab6cc3cafc2e2d0719f2dff13453cd.png

 

I imagined something like this, where if one would select GRID it would pull the latest compiled version from the local share.

 

The settings could also be enabled through an "Advanced settings" section so regular users with consumer cards won't need to bother. Of course "easy to use" is kind of a hard requirement to fulfill with this, though I could imagine providing a script where users just put the enterprise drivers in a "build" folder on the share, press a button in the UI to compile the driver package and a script would compile it. Then the compiled package would be copied over to the "driver share" for the plugin to use or something like that.

  

27 minutes ago, ich777 said:

Tesla and Quadro cards are already supported by the plugin.

The cards themselves are supported by the consumer drivers yes, but not the vGPU functionality they technically support through the enterprise driver.

 

27 minutes ago, ich777 said:

From my perspective everything that is needed for the cards to work properly should be integrated into the plugin package itself

Yeah it would definitely make sense to include it in the self compiled packages for the users instead of having to install it separately.

Edited by shawly
Link to comment
6 minutes ago, shawly said:

Providing the enterprise drivers publicly would definitely violate the EULA which is why I suggested the custom URL support or local storage support. That way people that have legitimate access to the driver can compile their own version locally. That way the plugin will never violate any EULA.

And this is where the fun begins, this is one reason why my plugin isn't designed for that because it is designed for a completely different use case.

 

For the use in Docker container and not in VMs (vGPU).

 

7 minutes ago, shawly said:

I didn't want to implement a fork since the fork would be exactly the same but with custom URL/share support for driver packages which is why I asked if you would accept such a PR.

But wouldn't it be better to create your own plugin for this specific use case and also customize it for that?

 

You have to understand if I implement this many support question will come in and I simply can't answer them because I don't have the hardware/time/money to even test this all and understand how this is all working.

 

14 minutes ago, shawly said:

I imagined something like this, where if one would select GRID it would pull the latest compiled version from the local share.

This is where the next issue begins, what is if you up- or downgrade your Unraid installation, this share or better speaking this location is not available on boot and the plugin installation will ultimately fail and you have to reinstall the plugin again.

 

15 minutes ago, shawly said:

The cards themselves are supported by the consumer drivers yes, but not the vGPU functionality they technically support through the enterprise driver.

Correct me if I'm wrong but not all Quadro cards are supported or am I wrong about that.

 

16 minutes ago, shawly said:

Yeah it would definitely make sense to include it in the self compiled packages for the users instead of having to install it separately.

...and it would in this case make also sense to customize the plugin in that way so that users can create their vGPUs from the plugin page itself.

 

 

I have to say I find it interesting but I won't change the existing plugin to add vGPU support or the support for GRID cards (I looked up the GRID cards on the Nvidia driver download page and drivers for the K1, K2, K340 & K520 are available for public download but I don't know if that's the driver which is needed).

Link to comment
40 minutes ago, ich777 said:

You have to understand if I implement this many support question will come in and I simply can't answer them because I don't have the hardware/time/money to even test this all and understand how this is all working.

That is true. Though this could be an unsupported or experimental feature. But yeah you're absolutely right.

 

40 minutes ago, ich777 said:

This is where the next issue begins, what is if you up- or downgrade your Unraid installation, this share or better speaking this location is not available on boot and the plugin installation will ultimately fail and you have to reinstall the plugin again.

Aren't shares available on boot or rather when plugins are installed? Of course the users would need to rebuild the driver package for that version. But wouldn't the same scenario happen right now if someone upgraded from an UnRAID version that has the OpenSource driver support to a newer version that doesn't have an OpenSource driver package like the latest version?

40 minutes ago, ich777 said:

Correct me if I'm wrong but not all Quadro cards are supported or am I wrong about that.

These are all the supported cards: https://docs.nvidia.com/grid/gpus-supported-by-vgpu.html

 

40 minutes ago, ich777 said:

I have to say I find it interesting but I won't change the existing plugin to add vGPU support or the support for GRID cards (I looked up the GRID cards on the Nvidia driver download page and drivers for the K1, K2, K340 & K520 are available for public download but I don't know if that's the driver which is needed).

No sadly NVIDIA still locks the vGPU drivers behind a subscription model so the drivers can only legitimately be downloaded if one has access to the enterprise support portal and a license for guests is required as well. Which is why the only way for now would be that the users themselves would need to compile their driver packages for UnRAID.

 

But yeah, you have convinced me that it might make more sense to provide all that functionality in a separate plugin. When I got more time on my hands I'll definitely give it a try. Though I have to say the developer documentation for implementing UnRAID plugins is quite terrible, being scattered over multiple forum threads with lots of outdated info is quite tiresome.

Edited by shawly
Link to comment

Getting this when trying to update by selecting latest radio button

+==============================================================================
| WARNING - WARNING - WARNING - WARNING - WARNING - WARNING - WARNING - WARNING
|
| Don't close this window with the red 'X' in the top right corner until the 'DONE' button is displayed!
|
| WARNING - WARNING - WARNING - WARNING - WARNING - WARNING - WARNING - WARNING
+==============================================================================

----------------Downloading Nvidia Driver Package v470.141.03-----------------
---------This could take some time, please don't close this window!------------

---------------Can't download Nvidia Driver Package v470.141.03----------------

Any ideas?


 

Screenshot_1.png

Screenshot_2.png

Edited by KJThaDon
Link to comment
27 minutes ago, KJThaDon said:

I do not

Can you open up a terminal from Unraid, issue the following command and send me the output:

cat /tmp/nvidia_*

 

From your syslog I see nothing obvious which is preventing the plugin from how it should work, something seems to be wrong anyways on your system, I'm really curious what is wrong there.

 

Have you yet tried to reboot and see if it's the same after a reboot?

Link to comment
6 hours ago, shawly said:

Aren't shares available on boot or rather when plugins are installed?

No, the shares are only available after you start the Array and plugins are installed much earlier in that process.

If you need some help send me a PM and as said, I'm glad to help.

 

6 hours ago, shawly said:

But wouldn't the same scenario happen right now if someone upgraded from an UnRAID version that has the OpenSource driver support to a newer version that doesn't have an OpenSource driver package like the latest version?

The Open Source GPU module is not available because it has an issue where NVENC is not working with that module -> see this GitHub issue: Click

I will start compiling packages for it when this bug is fixed because otherwise it would make no sense to publish such a package yet.

 

6 hours ago, shawly said:

Which is why the only way for now would be that the users themselves would need to compile their driver packages for UnRAID.

Anyways, I would not recommend the package directly on Unraid because Unraid isn't a common purpose server, it should be built in a VM or even better in a Docker container like I do it.

 

6 hours ago, shawly said:

When I got more time on my hands I'll definitely give it a try. Though I have to say the developer documentation for implementing UnRAID plugins is quite terrible, being scattered over multiple forum threads with lots of outdated info is quite tiresome.

If you need something just send over a PM and I'm glad to help.

Link to comment
32 minutes ago, ich777 said:

You already have installed driver version 520.56.06 from what I see from your Diagnostics.

 

What happens when you try to set it to a static version instiead of latest?

 

I did try that a few times also. No difference. Odd that it's showing I have the latest installed, but not on the UI.
 

8 minutes ago, ich777 said:

Can you open up a terminal from Unraid, issue the following command and send me the output:

cat /tmp/nvidia_*

 

From your syslog I see nothing obvious which is preventing the plugin from how it should work, something seems to be wrong anyways on your system, I'm really curious what is wrong there.

 

Have you yet tried to reboot and see if it's the same after a reboot?


 

CUR_PRB=525.60.11
LAST_PRB=515.86.01

CUR_NFB=520.56.06
LAST_NFB=495.46

UPDATED=2023-01-05T07:05Z
470.141.03
515.76
515.86.01
520.56.06
525.53
525.60.11
525.60.13


I will try rebooting again and see what it's showing.

Thank you

Link to comment
Just now, KJThaDon said:

I did try that a few times also. No difference. Odd that it's showing I have the latest installed, but not on the UI.

Can you also give me the output from:

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

 

I think the settings file is messed up but I'm not too sure, I really can't tell what's going on there the output from the other files are also fine and displaying the right version numbers.

Link to comment
28 minutes ago, ich777 said:

Can you also give me the output from:

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

 

I think the settings file is messed up but I'm not too sure, I really can't tell what's going on there the output from the other files are also fine and displaying the right version numbers.

 

update_check=false

Link to comment
54 minutes ago, KJThaDon said:

 

update_check=false

Yeah, there are pretty much all entries missing.

Do you maybe have edited it manually at some point in the past?

If so, all values must be in there like mentioned below.

 

Do you know how to edit this file in the terminal? In your case it should look like this:

first_installation=false
driver_version=520.56.06
disable_xconfig=true
update_check=false

(I assume that you are booting into GUI mode with the Display attached to the Nvidia card, if not change the line disable_xconfig to: disable_xconfig=false)

  • Upvote 1
Link to comment
1 hour ago, ich777 said:

Yeah, there are pretty much all entries missing.

Do you maybe have edited it manually at some point in the past?

If so, all values must be in there like mentioned below.

 

Do you know how to edit this file in the terminal? In your case it should look like this:

first_installation=false
driver_version=520.56.06
disable_xconfig=true
update_check=false

(I assume that you are booting into GUI mode with the Display attached to the Nvidia card, if not change the line disable_xconfig to: disable_xconfig=false)


I have never touched that file. Maybe it was just a failed/interrupted update in the past? I have edited it with those 4 lines and it seems to have updated now! Rebooting server and will report back if all is well.

Thank you!

  • Like 1
Link to comment
10 hours ago, KJThaDon said:


I have never touched that file. Maybe it was just a failed/interrupted update in the past? I have edited it with those 4 lines and it seems to have updated now! Rebooting server and will report back if all is well.

Thank you!


Just wanted to say it's all good now after the file edit. Still not sure what happened exactly, but I was able to update to v525.78.01 today.

Thank you!

  • Like 1
Link to comment
6 hours ago, Thraka said:

Thanks. I wonder why it was working previously. Updated and it's all good now.

I would have worked previously too but only with a manual edit from the settings file.

The main reason why I asked you to upgrade was because 6.9.2 is simply outdated and shouldn't be used anymore.

 

Anyways, glad that it is now working for you!

Link to comment
On 1/1/2023 at 1:08 PM, ich777 said:

I have many people having issues with using GTX 1600 series cards which fall from the bus for no reason or simply wont work at all.

 

Then pick a T600 it has 4GB of VRAM

 

A T400 is more than capable of 3 simultaneous streams if we are not talkig about 4K HFR…

I can transcode 3x 4K streams without any issue on my T400

 

BTW the T400, T600 & T1000 uses the same NVENC chip as the GTX 1600 series.

 

Also keep in mind for transcoding you don‘t need a powerhouse of a graphics card anyways. Even my Intel i5-10600 is capable of 4+ 4K transcodes at once with QuickSync and only uses a fraction of the power what you would need with a Nvidia card.

@ich777 @alturismo

Picked up a used GTX 1660 Super off ebay and swapped the 1070 card.  It runs great!  Chalking this ordeal up to a dead GTX 1070...it was a 2016 build date after all.

That aside, before I installed the new card, I ran a docker version of GPU-BURN and sure enough, the 1070 would not complete the test before throwing a XID: 31 READ_ERROR in the syslog.  [Remember that I was getting XID: 31 WRITE_ERRORs when Plex attempted to transcode before].  I had to manually close the docker to free up resources because the 1070 failed and locked up.

I installed the new GTX 1660 and ran the same GPU-BURN test for 5 mins.  Test completed without error.

I would love to see someone build an UNRAID docker/plugin for the GPU-BURN docker or a version of the CUDA-MEMTEST on github.

 

image.png.177d91e0b7b3c90529ccff96d4a605f2.png

Edited by Jacon
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.

×
×
  • Create New...