[Plugin] Linuxserver.io - Unraid Nvidia


Recommended Posts

1 hour ago, saarg said:

What does Plex say?

Plex "now playing" was showing hw transcoding, but playback was intermittent and stalling. nvidia-smi watch was showing no instances at all. All the gpu id stuff looked correct in the container. I rolled back to nvidia 6.8.3 to keep my users happy.

 

Anybody else have issues? I was hoping to see if linux kernel 5.7 would solve my AMD audio/usb passthrough issues. I don't want to have to compile my own stuff. 

Link to comment
1 hour ago, saarg said:

Post your docker run command and a screenshot of the Nvidia plugin page.

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='plex' --net='host' -e TZ="America/Los_Angeles" -e HOST_OS="Unraid" -e 'VERSION'='docker' -e 'NVIDIA_VISIBLE_DEVICES'='GPU-54955277-9286-fded-8f31-020baf163fa1' -e 'TCP_PORT_32400'='32400' -e 'TCP_PORT_3005'='3005' -e 'TCP_PORT_8324'='8324' -e 'TCP_PORT_32469'='32469' -e 'UDP_PORT_1900'='1900' -e 'UDP_PORT_32410'='32410' -e 'UDP_PORT_32412'='32412' -e 'UDP_PORT_32413'='32413' -e 'UDP_PORT_32414'='32414' -e 'PUID'='99' -e 'PGID'='100' -v '/mnt/user/media/Movies/':'/movies':'rw' -v '/mnt/user/media/TV Shows/':'/tv':'rw' -v '':'/music':'rw' -v '/config/transcode':'/transcode':'rw' -v '/mnt/user/media/Comedy/':'/comedy/':'rw' -v '/mnt/user/appdata/plex':'/config':'rw' --runtime=nvidia 'linuxserver/plex'

cbc630a408b5e6d77fcbb4ab61c3d4f45bb0c3db1dfd4cc95a24da0ea6ff0178

The command finished successfully!

 

MhtHxC0.png

Link to comment
10 hours ago, waymon said:

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='plex' --net='host' -e TZ="America/Los_Angeles" -e HOST_OS="Unraid" -e 'VERSION'='docker' -e 'NVIDIA_VISIBLE_DEVICES'='GPU-54955277-9286-fded-8f31-020baf163fa1' -e 'TCP_PORT_32400'='32400' -e 'TCP_PORT_3005'='3005' -e 'TCP_PORT_8324'='8324' -e 'TCP_PORT_32469'='32469' -e 'UDP_PORT_1900'='1900' -e 'UDP_PORT_32410'='32410' -e 'UDP_PORT_32412'='32412' -e 'UDP_PORT_32413'='32413' -e 'UDP_PORT_32414'='32414' -e 'PUID'='99' -e 'PGID'='100' -v '/mnt/user/media/Movies/':'/movies':'rw' -v '/mnt/user/media/TV Shows/':'/tv':'rw' -v '':'/music':'rw' -v '/config/transcode':'/transcode':'rw' -v '/mnt/user/media/Comedy/':'/comedy/':'rw' -v '/mnt/user/appdata/plex':'/config':'rw' --runtime=nvidia 'linuxserver/plex'

cbc630a408b5e6d77fcbb4ab61c3d4f45bb0c3db1dfd4cc95a24da0ea6ff0178

The command finished successfully!

 

MhtHxC0.png

That looks correct. What doesn't look correct is your transcode bind mount. You have mapped /config/transcode on unraid, which is not a user share in unraid so it gets created in ram. Dopu should fix that.

You can also tr to add a new variable for NVIDIA_DRIVER_CAPABILITIES and set it to all and see if that changes anything.

Link to comment
15 hours ago, saarg said:

That looks correct. What doesn't look correct is your transcode bind mount. You have mapped /config/transcode on unraid, which is not a user share in unraid so it gets created in ram. Dopu should fix that.

You can also tr to add a new variable for NVIDIA_DRIVER_CAPABILITIES and set it to all and see if that changes anything.

 

Wherever I set it I get this:

what(): Codecs: Initialize: 'boost::filesystem::temp_directory_path: Not a directory: "/mnt/cache/plextrans"'

I've tried my appdata folder and my media folder where tv and movies are.  Both say the above error.

What is "Dopu?"

 

Thanks again for helping me.

Link to comment
10 hours ago, waymon said:

 

Wherever I set it I get this:


what(): Codecs: Initialize: 'boost::filesystem::temp_directory_path: Not a directory: "/mnt/cache/plextrans"'

I've tried my appdata folder and my media folder where tv and movies are.  Both say the above error.

What is "Dopu?"

 

Thanks again for helping me.

Isn't dopu self explanatory? 😂 It should have been saying you.

 

We can't read your mind, so you need to post the docker run command and also what you do and where the error message pops up.

Link to comment

FIXED it i have no idea how just the old reset everything and try again

hi 

i have just upgraded to a ryzen for my system and i am unable to get the nvidia driver to see the p400 card 

i get the following error 

"NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running."

 

quick system overview 

b450 mortherboard 

pci slot 1 p400

pci slot 2 lis card 

pci slot 3 quadro nvs 295 

 

slot three is set to primary video device and it give an out put on video 

 

the cards are both showing in the device list 

 

[1022:43c7] 02:07.0 PCI bridge: Advanced Micro Devices, Inc. [AMD] 400 Series Chipset PCIe Port (rev 01)

[8086:1539] 03:00.0 Ethernet controller: Intel Corporation I211 Gigabit Network Connection (rev 03)

[1000:0072] 05:00.0 RAID bus controller: Broadcom / LSI SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

[10de:06fd] 07:00.0 VGA compatible controller: NVIDIA Corporation G98 [Quadro NVS 295] (rev a1)

IOMMU group 1:[1022:1482] 00:02.0 Host bridge: Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge

IOMMU group 2:[1022:1482] 00:03.0 Host bridge: Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge

[1022:1483] 00:03.1 PCI bridge: Advanced Micro Devices, Inc. [AMD] Starship/Matisse GPP Bridge

[10de:1cb3] 08:00.0 VGA compatible controller: NVIDIA Corporation GP107GL [Quadro P400] (rev a1)

[10de:0fb9] 08:00.1 Audio device: NVIDIA Corporation GP107GL High Definition Audio Controller (rev a1)

 

i have removed the driver and reinstalled it same error 

i tried with just the p400 in same error

i changed the slot of the p400 same error

 

anybody got any ideas 

Edited by andy harris
Link to comment

I installed a new GPU recently into my unraid server, and I have been getting hundreds of these errors in the system log:

 

NVRM: GPU - RmInitAdapter failed!
NVRM: GPU - rm_init_adapter failed, device minor number 0
NVRM: GPU - Failed to copy vbios to system memory

 

image.png.675f5e09cfad0ad13c072138768e451c.png 

 

New GPU is a GTX 1660Ti

Running Nvidia UnRaid 6.8.3 with Nvidia 440.59 drivers

 

I was not having any of these errors with my previous GPU, and I do not have any issues with my new GPU in another two machines when I tested

 

From what I have been able to see during troubleshooting, is the GPU keeps going online and offline, and as per the video it seems to be when the cached memory in unraid gets all used up, due to file transfers happening. Perhaps there is a tunable of some sort to prevent all the memory from getting used?

 

I did a screen capture when it was happening, its boring AF but it does show the device going offline and online when the cached memory goes up and down. Keep an eye on the memory usage and the green available sliver at the top.
 

 

There seems to be a number of other posts about this, but can't really see that its been resolved.

 

Any help or suggestions would be appreciated

Link to comment
30 minutes ago, Thompson8 said:

Hi, Is there someway of completely uninstalling this? ever since I've installed it I can't seem to get any of my GPU Passthrough VM's to work. They just crash the VM tab and I have to force stop from the dashboard page.

 

Thanks

Load back up the plugin and select to install the stock OS build.

Link to comment
On 6/26/2020 at 4:40 AM, saarg said:

Isn't dopu self explanatory? 😂 It should have been saying you.

 

We can't read your mind, so you need to post the docker run command and also what you do and where the error message pops up.

 

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='plex' --net='host' -e TZ="America/Los_Angeles" -e HOST_OS="Unraid" -e 'VERSION'='docker' -e 'NVIDIA_VISIBLE_DEVICES'='all' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -e 'TCP_PORT_32400'='32400' -e 'TCP_PORT_3005'='3005' -e 'TCP_PORT_8324'='8324' -e 'TCP_PORT_32469'='32469' -e 'UDP_PORT_1900'='1900' -e 'UDP_PORT_32410'='32410' -e 'UDP_PORT_32412'='32412' -e 'UDP_PORT_32413'='32413' -e 'UDP_PORT_32414'='32414' -e 'PUID'='99' -e 'PGID'='100' -v '/mnt/user/media/Movies/':'/movies':'rw' -v '/mnt/user/media/TV Shows/':'/tv':'rw' -v '':'/music':'rw' -v '/mnt/cache/plextrans':'/transcode':'rw' -v '/mnt/user/media/Comedy/':'/comedy/':'rw' -v '/mnt/user/appdata/plex':'/config':'rw' --runtime=nvidia 'linuxserver/plex'

97e7392d31c195872fe8f672ee5e4c187cb7ac4a63868458b9a5418f96e75cdb

The command finished successfully!

AScQsQU.png

hOtFJY9.pngeJHFera.png

 

😞

Edited by waymon
Link to comment
8 minutes ago, waymon said:

 

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='plex' --net='host' -e TZ="America/Los_Angeles" -e HOST_OS="Unraid" -e 'VERSION'='docker' -e 'NVIDIA_VISIBLE_DEVICES'='all' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -e 'TCP_PORT_32400'='32400' -e 'TCP_PORT_3005'='3005' -e 'TCP_PORT_8324'='8324' -e 'TCP_PORT_32469'='32469' -e 'UDP_PORT_1900'='1900' -e 'UDP_PORT_32410'='32410' -e 'UDP_PORT_32412'='32412' -e 'UDP_PORT_32413'='32413' -e 'UDP_PORT_32414'='32414' -e 'PUID'='99' -e 'PGID'='100' -v '/mnt/user/media/Movies/':'/movies':'rw' -v '/mnt/user/media/TV Shows/':'/tv':'rw' -v '':'/music':'rw' -v '/mnt/cache/plextrans':'/transcode':'rw' -v '/mnt/user/media/Comedy/':'/comedy/':'rw' -v '/mnt/user/appdata/plex':'/config':'rw' --runtime=nvidia 'linuxserver/plex'

97e7392d31c195872fe8f672ee5e4c187cb7ac4a63868458b9a5418f96e75cdb

The command finished successfully!

AScQsQU.png

hOtFJY9.pngeJHFera.png

 

😞

 

 

You have the variable NVIDIA_VISIBLE_DEVICES set incorrectly. It should be the GPU UUID  and NOT set to all

 

read #3 in the How to Utilise the GPU in LinuxServer.io Emby, Jellyfin or Plex Docker Containers section of the following post

 

  • Haha 1
Link to comment
1 hour ago, DavidSpek said:

Not sure if this has been requested already, but could the driver version be updated to 450.36.06? There is a bug with Plex where the GPU stays in power state P0 after transcoded streams are terminated and this is supposed to be fixed in the new Nvidia driver. 

https://forums.plex.tv/t/stuck-in-p-state-p0-after-transcode-finished-on-nvidia/387685/101

440.100 is the latest driver for linux. Current unraid, 6.8.3, is using 440.59 and the driver version will not be bumped until next build.

The 6.9.0 beta release might have 440.100.

We use the slackbuild drivers, so even though nvidia release a driver, it might not be the latest one for slackware.

Link to comment
10 minutes ago, saarg said:

440.100 is the latest driver for linux. Current unraid, 6.8.3, is using 440.59 and the driver version will not be bumped until next build.

The 6.9.0 beta release might have 440.100.

We use the slackbuild drivers, so even though nvidia release a driver, it might not be the latest one for slackware.

Given the power management fixes in the new 450.36.06 driver (and assuming also newer ones once they are released), would it be possible to add this driver once it is available through SlackBuild? I know the usual procedure it to have one release for each unraid build unless there is a critical bug, but given the fact that many use their GPU's for Plex transcoding and power usage in P8 is ~30% of the power usage in P0 wit a P2000 I think it is debatable if this should be considered critical. Reading through the thread I linked earlier the issue with Plex and power management was actually something Nvidia needed to fix in their driver and was not caused by an issue in Plex (even though Emby and Jellyfin don't seem to have this issue), which also supports the notion that this is a bug in the driver. 

Link to comment

GTX 1060 not working for 4K HEVC

 

My GeForce GTX 1060 3GB installed fine with these instructions. 

NVidia driver 440.59, unRAID 6.8.3

Most everything works.

But I cannot transcode an HEVC 4K video. It just hangs.

I confirmed that the video is H.265 (HEVC) 4:2:0, so it should work according to https://developer.nvidia.com/video-encode-decode-gpu-support-matrix

 

My questions.

1. Is there a fix? 

2. If not, how can I prevent it from using hardware transcoding for HEVC 4K videos (and only those)?

 

Thanks.

 

Link to comment
1 hour ago, Pducharme said:

I have a GTX 1060 6GB and it seems to work... strange yours doesn't.

Very strange. I just upgraded to 6.9.0-beta22 with the upgraded NVidia drivers, and it still doesn't work. 

I hope I can troubleshoot it.

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