[Plugin] Nvidia-Driver

[corfoto4]

18 minutes ago, ich777 said:

it fails to initialize fully:

Thank you. Hopefully Nvidia fixes the drivers. I'll have to get the shutdown thing working.

[ich777]

11 minutes ago, corfoto4 said:

Thank you. Hopefully Nvidia fixes the drivers. I'll have to get the shutdown thing working.

There are many different solutions to this. What came in my mind too is that you can turn on Resume After Power Loss in the BIOS and buy such a Timer Adapter thingi for the Wall (I think such a Shelly also supports a Timer functions if I'm not mistaken) so that when it turns on again the Server also turns on, you just have to make sure to turn the Server off early enough so that the Timer doesn't cuts the power before it fully shut down.

[TheUnraidNoob]

Hi everyone. Sorry if this is a noob error. I have installed a Nvidia RTX2060 into my server and I wish to use this for transcoding in Jellyfin. The GPU shows up in my System Devices, so I assume this means that unraid is recognising the GPU is installed:

 

 

I have installed the Nvidia-Driver plugin following the instructions pinned to page 1 of this topic by ich777. When I go to Plugins and click the Nvidia-Driver icon (Step 4 of the instructions), all I am getting is a blank screen. There is a troubleshooting guide in the instructions, but this says that there should still be something that shows up on this screen. Does anyone know what I might have done incorrectly here?

 

 

[alturismo]

2 hours ago, TheUnraidNoob said:

Does anyone know what I might have done incorrectly here?

start from scratch

 

uninstall the plugin, remove the existing files from your USB drive

 

 

reboot,  install again and WAIT until the done message comes up, reboot and you should be good to go.

[ich777]

3 hours ago, TheUnraidNoob said:

When I go to Plugins and click the Nvidia-Driver icon (Step 4 of the instructions), all I am getting is a blank screen.

Please post your Diagnostics too.

[TheUnraidNoob]

2 hours ago, alturismo said:

start from scratch

 

uninstall the plugin, remove the existing files from your USB drive

 

 

reboot,  install again and WAIT until the done message comes up, reboot and you should be good to go.

 

 

Thank you alturismo. I have followed these steps (including waiting for the done message to appear when installing the plugin) but I still have the same issue. Appreciate the feedback nonetheless!

[TheUnraidNoob]

2 hours ago, ich777 said:

Please post your Diagnostics too.

Hello ich777. My diagnostics zip is attached. Thank diagnostics-20221221-1726.zipyou for getting back to me!

[ich777]

39 minutes ago, TheUnraidNoob said:

Hello ich777. My diagnostics zip is attached. Thank diagnostics-20221221-1726.zipyou for getting back to me!

Something went really wrong with your installation.

 

Please do the following:

1. Remove the plugin
2. Execute this command from a Unraid terminal:

   rm -rf /boot/config/plugins/nvidia-driver

3. Reboot your server
4. Install the plugin from the CA App again and wait for the DONE button to appear, this can take a really long time depending on your internet connection (!!!don't close the pop up with the red X!!!)

 

I would strongly recommend before doing that all that you first upgrade to Unraid 6.11.5, you are still on Unraid 6.9.2 and that version is really out of date and missing some important information for me to diagnose what went wrong from the Diagnostics.

 

If it doesn't work after that please post the Diagnostics again.

[brainreplaced]

Hi everyone. I apologize for the question, I know it has been raised many times, but I could not find an answer 😞

I build a server and trying to set up a gtx 1660 super to work with a plex. I installed the drivers, the card was not recognized, removed and reinstalled - the same result. Please, help. Screenshots and logs are attached.

The card is recognized in unraid

 

 

brainserver-diagnostics-20221221-1916.zip

 

[gzibell]

What would cause GPU Link/PCIe Generation to change?  Been pulling my hair out trying to figure out why transcoding won't work sometimes with my 1660 super.  In testing, I pulled the card, put it in my backup server, installed nvidia, gpu stats, plex set it all up the same, same video file and it works flawlessly.  

 

In my primary server if I force the quality via plex web to 720 2MB before starting the video it'll hardware transcode and play just fine.  If I set the video quality to max and user reccomended for home streaming, then start the video it'll direct play, but then if I pick 720 2MB, I briefly see GPU activity, then the process drops, and nothing ever plays.  When I run nvidia-smi -q during this, I see the following in PCIe Generation.  Prior to attempting to play the video it was 33334. And if I force quality in plex web back to 720 2MB and play the video again, it'll play just fine and then I will see the 33334 again.  

 

I am assuming at this point it has something to do with a BIOS setting but I am not ever sure what to look for at this point.  I have 2 GPU's in the server so in Dual mode both are 4.0 x8 slots.  I do have another slot that is off the  X570 Chipset which is 4.0 x4 and moving the GPU to that slot is about the only idea I have left at this point. 

 

        GPU Link Info
            PCIe Generation
                Max                       : 3
                Current                   : 1
                Device Current            : 1
                Device Max                : 3
                Host Max                  : 4

ungeo-diagnostics-20221221-1043.zip

Edited December 21, 2022 by gzibell
Added diags

[ich777]

1 hour ago, brainreplaced said:

I build a server and trying to set up a gtx 1660 super to work with a plex. I installed the drivers, the card was not recognized, removed and reinstalled - the same result. Please, help. Screenshots and logs are attached.

This is the error:

Dec 21 19:07:25 brainserver kernel: nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
Dec 21 19:07:25 brainserver kernel: nvidia 0000:02:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
Dec 21 19:07:25 brainserver kernel: NVRM: The NVIDIA GPU 0000:02:00.0
Dec 21 19:07:25 brainserver kernel: NVRM: (PCI ID: 10de:21c4) installed in this system has
Dec 21 19:07:25 brainserver kernel: NVRM: fallen off the bus and is not responding to commands.
Dec 21 19:07:25 brainserver kernel: nvidia: probe of 0000:02:00.0 failed with error -1
Dec 21 19:07:25 brainserver kernel: NVRM: The NVIDIA probe routine failed for 1 device(s).
Dec 21 19:07:25 brainserver kernel: NVRM: None of the NVIDIA devices were initialized.

 

This seems more like an hardware incompatibility issue to me.

If possible try this:

1. Search for BIOS updates, if a newer BIOS is available please update your BIOS
2. Switch to Legacy Boot mode instead of UEFI
3. Force the PCIe slot to Gen3 where the card is installed

 

May I ask why do you want to use a Nvidia GPU when you have the iGPU, the new Intel iGPUs are pretty much capable of the same amount of transcodes like your 1650 (depending on the source quality) and as a benefit you save a lot of power.

[ich777]

55 minutes ago, gzibell said:

What would cause GPU Link/PCIe Generation to change?  Been pulling my hair out trying to figure out why transcoding won't work sometimes with my 1660 super.  In testing, I pulled the card, put it in my backup server, installed nvidia, gpu stats, plex set it all up the same, same video file and it works flawlessly.  

I can also only recommend the following things like for the user from above:

1. Search for BIOS updates, if a newer BIOS is available please update your BIOS
2. Switch to Legacy Boot mode instead of UEFI
3. Force the PCIe slot to Gen3 where the card is installed

 

58 minutes ago, gzibell said:

f I set the video quality to max and user reccomended for home streaming, then start the video it'll direct play, but then if I pick 720 2MB, I briefly see GPU activity, then the process drops, and nothing ever plays.

Do you use the Plex Web client for playing, I had back in the days multiple users which reported that the transcoding stops when switching too often between qualities, native clients in the meantime where working just fine.

 

Anyways have you yet took a look at the Plex logs, sounds like that the transcoding process crashed.

[gzibell]

2 hours ago, ich777 said:

I can also only recommend the following things like for the user from above:

1. Search for BIOS updates, if a newer BIOS is available please update your BIOS
2. Switch to Legacy Boot mode instead of UEFI
3. Force the PCIe slot to Gen3 where the card is installed

 

Do you use the Plex Web client for playing, I had back in the days multiple users which reported that the transcoding stops when switching too often between qualities, native clients in the meantime where working just fine.

 

Anyways have you yet took a look at the Plex logs, sounds like that the transcoding process crashed.

 

BIOS is up-to-date

Forced to Gen3 - No Change

Also removed one video card - No Change

Disabled CSM - Enabled Above 4G Decoding and Resizable BAR - No Change (Had to disable to change those 2 settings)

Re-Enabled CMS so I could switch back to Legacy - Had switched to UEIF to see if that resolved - No Change

Also tested with current plex version, and we back a few releases to see if there was an issue with plex but no change there either. 

 

Have reviewed plex logs and don't see anything to interesting other than the transcoding process just stops. 

 

Generaly no, I don't use plex web but used it for testing purposes.  This does not seem to be an issue in native plex clients. Do have some users that use plex web and was creating issues for them so had to disable HW transcoding in the meantime.  

 

Is this just a compatibility issue where my MB isn't playing nice?  Is it time to just bit the bullet and go for a 12th gen with quicksync?

 

 

 

Edited December 21, 2022 by gzibell

[TheUnraidNoob]

13 hours ago, ich777 said:

Something went really wrong with your installation.

 

Please do the following:

1. Remove the plugin
2. Execute this command from a Unraid terminal:

   rm -rf /boot/config/plugins/nvidia-driver

3. Reboot your server
4. Install the plugin from the CA App again and wait for the DONE button to appear, this can take a really long time depending on your internet connection (!!!don't close the pop up with the red X!!!)

 

I would strongly recommend before doing that all that you first upgrade to Unraid 6.11.5, you are still on Unraid 6.9.2 and that version is really out of date and missing some important information for me to diagnose what went wrong from the Diagnostics.

 

If it doesn't work after that please post the Diagnostics again.

 

 

Hello, and thank you for getting back to me. I followed the steps above, including upgrading my Unraid OS to 6.11.5. Same problem as before. This time I am unable to generate diagnostics. When I run diagnostics from the GUI, it is getting stuck at:

 

/usr/bin/nvidia-smi --query 2>/dev/null|todos >>'/mimir-diagnostics-20221221-2210/system/nvidia-smi.txt'

 

I also tried running the "diagnostics" command from a terminal window, but same problem. I left it running for several hours, but it still hasn't generated a diagnostics zip file. Any further help you are willing to offer me will be greatly appreciated.

[ich777]

1 hour ago, gzibell said:

Is this just a compatibility issue where my MB isn't playing nice?

I have a lot of users with not working GTX1600 series cards... Why, I really don't know...

 

For some they work and for some not, I have also multiple reports that they fall from the bus but mostly from users with AMD systems.

 

I could however be a compatibility issue with the motherboard but there is too little information out there that I can tell that for sure, I only can tell that a lot of users with GTX1600 series cards have issues.

[ich777]

5 minutes ago, TheUnraidNoob said:

I also tried running the "diagnostics" command from a terminal window, but same problem. I left it running for several hours, but it still hasn't generated a diagnostics zip file. Any further help you are willing to offer me will be greatly appreciated.

If this fails that bad something is wrong with the card, please check if it is connected to power and also maybe try to re-seat it in the PCIe slot.

 

Please uninstall the plugin, reboot, pull the Diagnostics (without the plugin installed and after a reboot it will just work fine) and then install the plugin again and post your syslog in a text file if possible.

 

Are you sure that the card is working?

[Skitals]

On 12/18/2022 at 12:42 AM, ich777 said:

You can try to disable AER error correction, won‘t hurt anything.

I also have to disable it on my server because otherwise my PCIe cards won‘t work properly (this is most of the times caused by some weired or better speaking wrong integration in the BIOS).

Simply add this to your syslinux.conf:

pci=noaer

 

 

The Quadro P400 came today. Plugged it in and it just worked, I didn't need to tweak anything in my bios. UEFI booting unraid, factory settings for Above 4G Decoding, Resizable BAR, ASPM, C-States, PCIe gen, etc.

 

I can finally stream the UHD discs that I ripped w/ makemkv to my phone without crippling my cpu  Also working with comskip, so no cpu spikes after each dvr recording.

 

Thanks for your assistance, and the plugin.

[ich777]

8 minutes ago, Skitals said:

The Quadro P400 came today. Plugged it in and it just worked, I didn't need to tweak anything in my bios. UEFI booting unraid, factory settings for Above 4G Decoding, Resizable BAR, ASPM, C-States, PCIe gen, etc.

 

I can finally stream the UHD discs that I ripped w/ makemkv to my phone without crippling my cpu  Also working with comskip, so no cpu spikes after each dvr recording.

Thank you for the update really appreciate it.

 

Glad to hear that everything is now working for you!

[gzibell]

28 minutes ago, ich777 said:

I have a lot of users with not working GTX1600 series cards... Why, I really don't know...

 

For some they work and for some not, I have also multiple reports that they fall from the bus but mostly from users with AMD systems.

 

I could however be a compatibility issue with the motherboard but there is too little information out there that I can tell that for sure, I only can tell that a lot of users with GTX1600 series cards have issues.

That would make sense as my primary is an AMD rig and my backup is an Intel rig.  Works on the Intel, not on the AMD.  Going to give it a shot in the slot controlled by the chipset and if that fails, will swap the 1660 with my 1060 and see if that does the trick.  Thanks for everything you do!

[TheUnraidNoob]

3 hours ago, ich777 said:

If this fails that bad something is wrong with the card, please check if it is connected to power and also maybe try to re-seat it in the PCIe slot.

 

Please uninstall the plugin, reboot, pull the Diagnostics (without the plugin installed and after a reboot it will just work fine) and then install the plugin again and post your syslog in a text file if possible.

 

Are you sure that the card is working?

Hello, and thank you for the continued support. I have reseated the card, and reconnected the 8-pin power connector to the card and to the PSU. Same issue upon reinstalling the plugin. I have attached the diagnostics file (generated after removing the plugin and rebooting) and also the syslog file (generated after reinstalling the plugin and attempting to access the plugin settings page).

 

The card is definitely working:
   1. It was the card being used in my gaming rig until a week ago when I upgraded. It was stored in an anti-static bag until being installed into the server.

   2. I am able to view the server bios settings on a monitor connected to a HDMI port on the card.

syslog.txt mimir-diagnostics-20221222-0925.zip

[ich777]

5 hours ago, TheUnraidNoob said:

I have reseated the card, and reconnected the 8-pin power connector to the card and to the PSU. Same issue upon reinstalling the plugin.

The syslog after the install tells me that something went very wrong with the card in this system and the module crashes basically.

But I'm not entirely sure what is causing this.

 

The fact that you can see a display output tells not much if the card is properly working since the card is not fully initialized at this stage (BIOS & Terminal), you have to actually install the drivers for the card and put a 3D load on it to see if it's properly working.

Anyways, even if the card is working in another system it is possible that some form of hardware compatibility issue is causing this (I don't know much about those Intel motherboards that you are using, I've only heard that there can happen some weird things with certain hardware).

 

I would recommend that you try to install the card in another system (maybe Windows would be best), install the drivers, put some 3D load on it (Furmark, 3D Mark,...) and see if it is stable for about 10 minutes.

If that is working in the other system please create another Unraid USB Boot device and boot this exact system with this USB Boot device (don't start the Array, just register for Trail), install the CA App and try to install the Nvidia driver there <- this will not change anything to the system which you test the card in, as long as you don't start the Array.

I know this procedure is really tedious but with this you ensure that the hardware is working and if it's a hardware compatibility issue.

[ich777]

8 hours ago, gzibell said:

Going to give it a shot in the slot controlled by the chipset and if that fails, will swap the 1660 with my 1060 and see if that does the trick.  Thanks for everything you do!

Please let me know if that solves your issue.

[TheUnraidNoob]

1 hour ago, ich777 said:

The syslog after the install tells me that something went very wrong with the card in this system and the module crashes basically.

But I'm not entirely sure what is causing this.

 

The fact that you can see a display output tells not much if the card is properly working since the card is not fully initialized at this stage (BIOS & Terminal), you have to actually install the drivers for the card and put a 3D load on it to see if it's properly working.

Anyways, even if the card is working in another system it is possible that some form of hardware compatibility issue is causing this (I don't know much about those Intel motherboards that you are using, I've only heard that there can happen some weird things with certain hardware).

 

I would recommend that you try to install the card in another system (maybe Windows would be best), install the drivers, put some 3D load on it (Furmark, 3D Mark,...) and see if it is stable for about 10 minutes.

If that is working in the other system please create another Unraid USB Boot device and boot this exact system with this USB Boot device (don't start the Array, just register for Trail), install the CA App and try to install the Nvidia driver there <- this will not change anything to the system which you test the card in, as long as you don't start the Array.

I know this procedure is really tedious but with this you ensure that the hardware is working and if it's a hardware compatibility issue.

 

I will give all of this a go when I get access to a different computer. Thank you very much for the support. I will update this forum with the result!

[gzibell]

9 hours ago, ich777 said:

Please let me know if that solves your issue.

No luck in the chipset slot.  Same behavior.  Swapped in my 1060 got it to work once randomly, then back to the same. Removed driver, checked plugins folder on flash, rebooted, re-installed no change, tried a bunch of plex versions no luck.  Same PCIe drop seen in SMI. 

 

I think I did solve it though....12700 with rush processing and next day delivery😁...QuickSync here I come. 

[jamescon]

I am unable to install the Nvidia-Driver plugin.  I keep receiving the error "Can't download Nvidia Driver Package".  Below is the full message that I receive.  Any suggestions?

 

Thanks!

 

-----

 

 

plugin: installing: nvidia-driver.plg Executing hook script: pre_plugin_checks plugin: downloading: nvidia-driver.plg ... done plugin: downloading: nvidia-driver-2022.10.05.txz ... done +==============================================================================

| Installing new package /boot/config/plugins/nvidia-driver/nvidia-driver-2022.10.05.txz +==============================================================================

Verifying package nvidia-driver-2022.10.05.txz.

Installing package nvidia-driver-2022.10.05.txz:

PACKAGE DESCRIPTION: Package nvidia-driver-2022.10.05.txz installed. +==============================================================================

| 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 v525.60.13------------------

----------This could take some time, please don't close this window!------------

--------------Can't download Nvidia Driver Package v525.60.13-----------------

 

plugin: run failed: /bin/bash Executing hook script: post_plugin_checks