[Plugin] Linuxserver.io - Unraid Nvidia


Recommended Posts

1 hour ago, Fiservedpi said:

Is there any way to roll back the driver version 'm getting the error my device is not compatible with this driver version. Do I have to rollback the whole Nvidia-unraid build?

 

** I should note the GPU appears to be working with-in my VM but am unable to Passthrough to any Dockers "No Cuda enabled devices available"

 

 

Using built-in stream user interface
-> Detected 24 CPUs online; setting concurrency level to 24.
WARNING: You do not appear to have an NVIDIA GPU supported by the 430.14 NVIDIA Linux graphics driver installed in this system.  For further details, please see the appendix SUPPORTED NVIDIA GRAPHICS CHIPS in the README available on the Linux driver download page at www.nvidia.com.
-> Not installing a kernel module; skipping the "is an NVIDIA kernel module loaded?" test.
-> Installing NVIDIA driver version 430.14.
-> Skipping check for conflicting rpms.
WARNING: The nvidia-uvm module will not be installed. As a result, CUDA will not function with this installation of the NVIDIA driver.
WARNING: The nvidia-drm module will not be installed. As a result, DRM-KMS will not function with this installation of the NVIDIA driver.
WARNING: You specified the '--no-kernel-module' command line option, nvidia-installer will not install a kernel module as part of this driver installation, and it will not remove existing NVIDIA kernel modules not part of an earlier NVIDIA driver installation.  Please ensure that an NVIDIA kernel module matching this driver version is installed separately.

 

We only build releases with the current version of nvidia driver available. If you want to roll back drivers, you'll need to roll back the unraid build, though this isn't something we recommend or support for obvious reasons.

Link to comment
On 7/12/2019 at 2:56 PM, Squid said:

diagnostics might be a clue if you catch them while the error happens, but I suspect it may be a transient thing

Gtx760, right I like to be in the most current build guess my card just isn't up to snuff

Edited by Fiservedpi
Link to comment

First of all. Thank you to those of you that have made this project a reality.

This is my 2nd trial to get this to work....

SuperMicro X9DA6
Asus Nvidia GTX1060 3gb
Nvidia Build 6.7.2

I've tried in the past to get Binhex's EMBY transcoding to work, but in the EMBY Preferred Hardware Decoders, it doesn't show any options for hardware decoding. Yes I am a premiere member.


Okay, I'll try Linuxserver's Jellyfin. It's what I'd eventually like to switch to when they get their clients on Android, Firestick, and IOS.

I previously had Binhex's Jellyfin, so I unstalled that, and went to install Linuxserver's.
So I entered the info needed(I think) to get my Nvidia card to work. But, got an error:
 

# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='jellyfin' --net='bridge' -e TZ="America/Chicago" -e HOST_OS="Unraid" -e 'NVIDIA_VISIBLE_DEVICES'='GPU-fbc079d3-5b53-7d3c-d6aa-90e0ef8532c9' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -e 'PUID'='99' -e 'PGID'='100' -p '8095:8096/tcp' -p '8920:8920/tcp' -p '1900:1900/udp' -v '/mnt/user/video/movies/':'/movies':'rw' -v '/mnt/user/video/TV/':'/tv':'rw' -v '/mnt/user/Music/':'/music':'rw' -v '/mnt/user/appdata/jellyfin':'/config':'rw' --runtime=nvidia 'linuxserver/jellyfin' 

66b1831b22f62bb8733972879fccf5b6ac270665f1b088665a7d2871220b8bc5
/usr/bin/docker: Error response from daemon: driver failed programming external connectivity on endpoint jellyfin (79b24634eba4eda02806cb82604376eec86365fb7d1356949d5285084c4dd893): Error starting userland proxy: listen tcp 0.0.0.0:8920: bind: address already in use.

I tried re-downloading, and re-installing Nvidia 6.7.2 to no avail.... Same error.... What am I doing wrong?

Capture - driver.PNG

Capture - error.PNG

Link to comment
On 7/9/2019 at 12:19 AM, ramblinreck47 said:

From what ChuckPA said on the Plex Forums, it sounds like a major update to the transcoder is being finished up and hopefully deployed in the near future. I’m thinking they’ll address this, as well as finally provide official NVDEC support and fix the issues they’ve been having with 9th Gen Intel QuickSync. He made it seem like a significant change for the transcoder.

looks like you were right since yesterday after i updated my plex docker my gpu goes back to p8 state as soon as its done transcoding. Now finally my core and mem clocks goes back to idle.

Edit: sorry i was wrong it's decode only thing, it only goes back to p8 state if it was being used for some decode only process which i guess is only when plex is generating video preview thumbnails but once u use it for transcoding or encoding then it never goes back p8 states.

once again im sorry for being stupid😅

On 7/14/2019 at 3:13 AM, Fiservedpi said:

Gtx760, right I like to be in the most current build guess my card just isn't up to snuff

and you know what the irony is with nvidia, my gtx 750 runs just fine probably because 750 runs on maxwell and 760 runs on kepler but isn't its supposed to be other way around.😅

5 hours ago, buellmule said:

First of all. Thank you to those of you that have made this project a reality.

This is my 2nd trial to get this to work....

SuperMicro X9DA6
Asus Nvidia GTX1060 3gb
Nvidia Build 6.7.2

I've tried in the past to get Binhex's EMBY transcoding to work, but in the EMBY Preferred Hardware Decoders, it doesn't show any options for hardware decoding. Yes I am a premiere member.


Okay, I'll try Linuxserver's Jellyfin. It's what I'd eventually like to switch to when they get their clients on Android, Firestick, and IOS.

I previously had Binhex's Jellyfin, so I unstalled that, and went to install Linuxserver's.
So I entered the info needed(I think) to get my Nvidia card to work. But, got an error:
 


# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='jellyfin' --net='bridge' -e TZ="America/Chicago" -e HOST_OS="Unraid" -e 'NVIDIA_VISIBLE_DEVICES'='GPU-fbc079d3-5b53-7d3c-d6aa-90e0ef8532c9' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -e 'PUID'='99' -e 'PGID'='100' -p '8095:8096/tcp' -p '8920:8920/tcp' -p '1900:1900/udp' -v '/mnt/user/video/movies/':'/movies':'rw' -v '/mnt/user/video/TV/':'/tv':'rw' -v '/mnt/user/Music/':'/music':'rw' -v '/mnt/user/appdata/jellyfin':'/config':'rw' --runtime=nvidia 'linuxserver/jellyfin' 

66b1831b22f62bb8733972879fccf5b6ac270665f1b088665a7d2871220b8bc5
/usr/bin/docker: Error response from daemon: driver failed programming external connectivity on endpoint jellyfin (79b24634eba4eda02806cb82604376eec86365fb7d1356949d5285084c4dd893): Error starting userland proxy: listen tcp 0.0.0.0:8920: bind: address already in use.

I tried re-downloading, and re-installing Nvidia 6.7.2 to no avail.... Same error.... What am I doing wrong?

Capture - driver.PNG

Capture - error.PNG

i wouldn't suggest going with jellyfin just a week i also gave it a try as with plex and emby you have to pay for hwacell but it was so glitchy for me that i uninstalled it in a matter of minutes. 

Edited by Max
Link to comment
3 hours ago, buellmule said:

First of all. Thank you to those of you that have made this project a reality.

This is my 2nd trial to get this to work....

SuperMicro X9DA6
Asus Nvidia GTX1060 3gb
Nvidia Build 6.7.2

I've tried in the past to get Binhex's EMBY transcoding to work, but in the EMBY Preferred Hardware Decoders, it doesn't show any options for hardware decoding. Yes I am a premiere member.


Okay, I'll try Linuxserver's Jellyfin. It's what I'd eventually like to switch to when they get their clients on Android, Firestick, and IOS.

I previously had Binhex's Jellyfin, so I unstalled that, and went to install Linuxserver's.
So I entered the info needed(I think) to get my Nvidia card to work. But, got an error:
 


# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='jellyfin' --net='bridge' -e TZ="America/Chicago" -e HOST_OS="Unraid" -e 'NVIDIA_VISIBLE_DEVICES'='GPU-fbc079d3-5b53-7d3c-d6aa-90e0ef8532c9' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -e 'PUID'='99' -e 'PGID'='100' -p '8095:8096/tcp' -p '8920:8920/tcp' -p '1900:1900/udp' -v '/mnt/user/video/movies/':'/movies':'rw' -v '/mnt/user/video/TV/':'/tv':'rw' -v '/mnt/user/Music/':'/music':'rw' -v '/mnt/user/appdata/jellyfin':'/config':'rw' --runtime=nvidia 'linuxserver/jellyfin' 

66b1831b22f62bb8733972879fccf5b6ac270665f1b088665a7d2871220b8bc5
/usr/bin/docker: Error response from daemon: driver failed programming external connectivity on endpoint jellyfin (79b24634eba4eda02806cb82604376eec86365fb7d1356949d5285084c4dd893): Error starting userland proxy: listen tcp 0.0.0.0:8920: bind: address already in use.

I tried re-downloading, and re-installing Nvidia 6.7.2 to no avail.... Same error.... What am I doing wrong?

Capture - driver.PNG

Capture - error.PNG

The port you chose is already in use, so choose a port that is not in use and it should work.

Link to comment
3 hours ago, Max said:

looks like you were right since yesterday after i updated my plex docker my gpu goes back to p8 state as soon as its done transcoding. Now finally my core and mem clocks goes back to idle.

and you know what the irony is with nvidia, my gtx 750 runs just fine probably because 750 runs on maxwell and 760 runs on kepler but isn't its supposed to be other way around.😅

i wouldn't suggest going with jellyfin just a week i also gave it a try as with plex and emby you have to pay for hwacell but it was so glitchy for me that i uninstalled it in a matter of minutes. 

Which docker do you use? Sounds very interesting with the P-states!

Link to comment
1 hour ago, pappaq said:

Which docker do you use? Sounds very interesting with the P-states!

sorry i was wrong it's decode only thing, it only goes back to p8 state if it was being used for some decode only process which i guess is only when plex is generating video preview thumbnails but once u use it for transcoding or encoding then it never goes back p8 states.

once again im sorry for being stupid😅

Link to comment
15 minutes ago, Max said:

sorry i was wrong it's decode only thing, it only goes back to p8 state if it was being used for some decode only process which i guess is only when plex is generating video preview thumbnails but once u use it for transcoding or encoding then it never goes back p8 states.

once again im sorry for being stupid😅

And I tried all plex dockers this evening to prove your point! Emby it is then! Thank you anyways!

Link to comment

hi having a few issues getting transcending to work with plex 

setup 

p400 

unraid 6.7.2

unraid nvidia

 1013876965_unriadnvidia.PNG.c48ad4c6160be4b7ebfeaa3841e687d3.PNG

 

plex parameters 

paramiters.thumb.PNG.093a95651f9fa1fa28c85ee656970168.PNG

 

Capture1.thumb.PNG.cac14289cce68faac94bff73c79dde6c.PNG

 

 

everything looks like it OK adding the parameters passes  

Capture.PNG.4b30d8e86202ad82ab50a9304b73ce46.PNG

 

but no matter what i do i get get streams to encode via gpu they are all using the cpu (i know this barbecue cpu usage goes through the roof 

 

any help would be appreciated 

 

Link to comment
1 hour ago, andy harris said:

hi having a few issues getting transcending to work with plex 

setup 

p400 

unraid 6.7.2

unraid nvidia

 1013876965_unriadnvidia.PNG.c48ad4c6160be4b7ebfeaa3841e687d3.PNG

 

plex parameters 

paramiters.thumb.PNG.093a95651f9fa1fa28c85ee656970168.PNG

 

Capture1.thumb.PNG.cac14289cce68faac94bff73c79dde6c.PNG

 

 

everything looks like it OK adding the parameters passes  

Capture.PNG.4b30d8e86202ad82ab50a9304b73ce46.PNG

 

but no matter what i do i get get streams to encode via gpu they are all using the cpu (i know this barbecue cpu usage goes through the roof 

 

any help would be appreciated 

 

Do you have plex pass?

And did you add the --runtime=Nvidia to the extra parameters box?

Link to comment
2 minutes ago, j0nnymoe said:

is this variable correct? seems like it isn't.

image.png.6d173e4fcb5fc8dcc19dff9dd679c497.png

sorry must have been one of the many time i have tried from scratch to get it working it does read the correct value 

Capture3.thumb.PNG.cea38ae06b8cadfdec9bd070e680c87f.PNG

 

still does not work it just seams to send cpu usage through the roof and nothing in nvidia-smi

Link to comment
18 minutes ago, andy harris said:

sorry must have been one of the many time i have tried from scratch to get it working it does read the correct value 

Capture3.thumb.PNG.cea38ae06b8cadfdec9bd070e680c87f.PNG

 

still does not work it just seams to send cpu usage through the roof and nothing in nvidia-smi

Are you copy/pasting stuff from the forum to setup these variables? that can cause issues.

  • Upvote 1
Link to comment
4 hours ago, j0nnymoe said:

Are you copy/pasting stuff from the forum to setup these variables? that can cause issues.

right like i have sort of fixed it plex will now trans code via the p400 but i am still getting cpu usage through the roof (is that normal) i was expecting it to lower my cpu usage 

 

i have no idea what i did different i just deleted the packages and started again 

Edited by andy harris
Link to comment
5 hours ago, andy harris said:

right like i have sort of fixed it plex will now trans code via the p400 but i am still getting cpu usage through the roof (is that normal) i was expecting it to lower my cpu usage 

 

i have no idea what i did different i just deleted the packages and started again 

That's because plex doesn't take full advantage of the GPU yet, it only uses NVENC, not NVDEC. So the decoding get's done via the CPU. Plus audio is done via the CPU too.

Link to comment

Has anyone been able to get this ti work running unsaid as a VM in esxi?  I have a 1600ti card that i installed put in pass through mode added it to unraid.  I boot fine but I am getting "Unable to determine the device handle for GPU 0000:0B:00.0: Unknown Error" .  researching this out says to edit the VM  and add hypervisor.cpuid.v0 = "FALSE", and then unraid boots extremely slow and then fails fails to fully boot.  Anyone have any pointers?

Link to comment
6 minutes ago, dertbv said:

Has anyone been able to get this ti work running unsaid as a VM in esxi?  I have a 1600ti card that i installed put in pass through mode added it to unraid.  I boot fine but I am getting "Unable to determine the device handle for GPU 0000:0B:00.0: Unknown Error" .  researching this out says to edit the VM  and add hypervisor.cpuid.v0 = "FALSE", and then unraid boots extremely slow and then fails fails to fully boot.  Anyone have any pointers?

Pretty sure someone else had this same issue before. Search the thread.

Link to comment

Just a quick stop in to say THANKS to LSIO folks and @CHBMB for this. I installed it last night and used an old 1GB gt 710 with plex. Seems to be working without issue even though Nvidia-smi:

 

Never shows a process (but it is working because the reported temp changes on the gpu and cpu load is less

Power used doesn't ever change from N/A

seems stuck at P0 but it's a 19 watt card, so, I guess thats ok

 

 

Using this old cheap card GPU took my ml30 get 9 from pegging 3 threads at 100% for about 2 minutes to transcode while creating a buffer to about 50 seconds of pegging 2 threads to transcode while creating a buffer. For fun I added the nvdec script and now a single thread never pushes above 25%, there is essentially no buffer lag meaning near instantaneous playback of transcodes, and it only uses about 200MB of the cards onboard ram. 

 

If you're using this in a home environment with only 1-2 transcodes anticipated, I don't see why you would need anything more powerful. It's a very economical and effective implementation of hw transcoding in plex.

 

 

 

Link to comment

like 99.99% of others reading this thread that haven't yet installed this, I have a working Unraid system already.  

 

I'm thinking of using this "Nvidia Unraid" option so I can add a Nvidia Quadro P2000 so I can get hardware decoding for plex which runs in a docker.

 

Is there a pointer in how to migrate my system while keeping my array, etc?  My understanding is you have to do a clean install of this and can't "upgrade" from a normal unraid setup to this.  I just re-read the top of the thread and realized that this looks incredibly simple. I just install the "Nvidia Unraid" plugin, then choose the 6.7.2 version to install from the plugin settings and it downloads and installs?!?  That seems way too easy i'm skeptical. Is this correct?

 

Also, is there any hard in installing this ahead of getting my P2000, or shoudl i wait to get that, install it, then install Nvidia Unraid?

Edited by tmchow
Link to comment

@tmchow it upgrades your current system from the pluping page, everything will be as it is when it rebooted but with the nvidia drivers installed, also plex only encodes videos using nvenc it still uses the cpu to decode videos, i believe emby uses full dec/enc. there is a script for nvdec but it has been know to be buggy, i had tried it and crashed my system twice with it before i stopped using it

Link to comment
5 minutes ago, fr05ty said:

@tmchow it upgrades your current system from the pluping page, everything will be as it is when it rebooted but with the nvidia drivers installed, also plex only encodes videos using nvenc it still uses the cpu to decode videos, i believe emby uses full dec/enc. there is a script for nvdec but it has been know to be buggy, i had tried it and crashed my system twice with it before i stopped using it

since this is a plex server, the decoding isn't relevant is it? Or am i wrong here.

Link to comment
3 hours ago, fr05ty said:

@tmchow it upgrades your current system from the pluping page, everything will be as it is when it rebooted but with the nvidia drivers installed, also plex only encodes videos using nvenc it still uses the cpu to decode videos, i believe emby uses full dec/enc. there is a script for nvdec but it has been know to be buggy, i had tried it and crashed my system twice with it before i stopped using it

Conversely, I've not had any issues running the Plex script at all (touch wood).

3 hours ago, fr05ty said:

if you are transcoding the video yes, direct play no

Presumably if one is always using direct play, the Nvidia card would be redundant, no?

 

 

Your assumptions are correct - it's an easy install - install plugin, select version, install customised Unraid, reboot.  Unraid will look and feel exactly the same. To my knowledge there's no harm installing it without a card, but there's also not much point.

Link to comment
  • trurl locked this topic
Guest
This topic is now closed to further replies.