[Support] ich777 - Jellyfin AMD/Intel/Nvidia


74 posts in this topic Last Reply

Recommended Posts

1 hour ago, pzen said:

Yes I restarted after making the change to the go file and I just do the modprobe amdgpu command manually in addition

Can you do "ls /dev/dri" and post the output here?

This looks all good and the driver is loaded, are you on Unraid 6.9.0?

Are you sure that you set the iGPU to be the first boot device or better speaking have you installed also a dedicated GPU in your system?

Link to post
  • Replies 73
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Summary: Support Thread for ich777 Jellyfin - AMD/Intel/Nvidia DockerHub: https://hub.docker.com/r/ich777/jellyfin   Jellyfin is a Free Software Media System that puts you in control of

From "well-informed sources" i know it was also successfully tested on Ryzen 5 Pro 4650g  

thanks for your advise mate, you manage to fix my problem. Tested play and hardware transcoding / direct play, restart the cointaner few time, no issue at all.    

Posted Images

Hello

 

Can you do "ls /dev/dri" and post the output here?

 

Directory / dri not created should we do it manually ? or normally this is done when setting up the installation of Jellyfin AMD / Intel / Nvidia ?

 

See the screenshot below

 

This looks all good and the driver is loaded, are you on Unraid 6.9.0?

 

Yes I am using 6.9.0 (updated to the latest version today)

 

Are you sure that you set the iGPU to be the first boot device or better speaking have you installed also a dedicated GPU in your system?

 

Yes I have a dedicated graphics card, a Sapphire Radeon HD 4850, however I do not have enough knowledge of unraid (on the other hand I have so far not had the need to declare it to Unraid, is this important?)

 

Thank you for taking the time to help me

1.png

Edited by pzen
Link to post
8 minutes ago, pzen said:

Directory / dri not created should we do it manually ?

 

No. If the driver and hardware are working well together the dri directory (and its contents) automatically appears in /dev. I don't think amdgpu is the best driver module for the HD 4850. A better match is the older radeon module but it doesn't have the same advanced features as amdgpu.

Link to post

Hello

 

Thanks for the information

 

I am new to Unraid, is it possible to install the drivers manually? (Retrieving a file to start its execution in Unraid via a terminal)

 

If so could you indicate the commands or a link explaining how to do this?

 

Have a good day

Link to post
7 minutes ago, pzen said:

Thanks for the information

 

I am new to Unraid, is it possible to install the drivers manually? (Retrieving a file to start its execution in Unraid via a terminal)

 

If so could you indicate the commands or a link explaining how to do this?

Yes, you simply type in 'modprobe radeon' then it should be activated.

 

Keep in mind that you are running a fairly old CPU (I think they where released somewhere around 2005) and I don't think that you can make use of it in Jellyfin...

Link to post

thank you very much

 

After using the 'modprobe radeon' command I finally have access to the application (I'm happy)

 

On the other hand, in the event of a restart of the server, it is necessary to restart pr 'modprobe radeon’ How can I get the command to be taken directly when starting the server?

 

And I send you the points entered in Jellyfin did not hesitate to tell me if I made a mistake (picture below)

 

And last question how to verify that the garphic card is solicited?

 

 

4.png

3.png

2.png

Link to post
57 minutes ago, pzen said:

On the other hand, in the event of a restart of the server, it is necessary to restart pr 'modprobe radeon’ How can I get the command to be taken directly when starting the server?

Put the command in the go file.

 

57 minutes ago, pzen said:

And last question how to verify that the garphic card is solicited?

If you force a transcode of a stream then the CPU should not be hit that hard (but it could be that you didn't notice it because the CPU is actually really old and only dual core I think), but keep in mind that if you CPU is not capable of a fromat it will fall back to CPU transcoding and you have no benefit.

 

For example if the iGPU is used in a Ryzen 3200G for transcoding of a 4K file to 1080p the CPU hovers around 10-15% (that's for transcoding the audio, subtitle,...), if you try to transcode a file without the iGPU the CPU utilisation is about 80-90%.

 

EDIT: I don't think your CPU is capable of transcoding HEVC and I even don't think it's capable of transcoding any or most of h264 profiles.

Link to post

Hello,

 

Indeed my configuration is old.

 

I put this machine together with piece recovery and secondhand components so that I could experience Unraid and its surroundings.

 

Maybe later I will assemble a configuration with newer components.

 

Thank you very much for the help you have given

 

Have a good day

Link to post

Hello, thanks for your working. But I have some issues using Jellyfin. Once I finished the installation and open the WebUI, it will get stucked. The error log as below:

 

[2021-03-10 20:35:25.612 +08:00] [INF] [83] Jellyfin.Api.Auth.CustomAuthenticationHandler: AuthenticationScheme: "CustomAuthentication" was challenged.
[2021-03-10 20:35:25.612 +08:00] [INF] [81] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.612 +08:00] [INF] [81] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.612 +08:00] [INF] [81] Jellyfin.Api.Auth.CustomAuthenticationHandler: AuthenticationScheme: "CustomAuthentication" was challenged.
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: AuthenticationScheme: "CustomAuthentication" was challenged.
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: "CustomAuthentication" was not authenticated. Failure message: "Invalid token."
[2021-03-10 20:35:25.613 +08:00] [INF] [82] Jellyfin.Api.Auth.CustomAuthenticationHandler: AuthenticationScheme: "CustomAuthentication" was challenged.
 

Could you help to analysis this? Thanks a lot

Link to post
41 minutes ago, UDMark said:

But I have some issues using Jellyfin.

Was Jellyfin running before or is this a brand new installation?

 

EDIT: Tried it now and got no problem upgraded my Jellyfin instance from 1.6.4 to 1.7.0:

grafik.png.4432806b14f06b886c041345fda69a3a.png

 

 

It may be possible that you actually have to downgrade to 1.6.4 and uninstall the Plugins that you have manually installed if you have installed any manually since some are not compatible with the newer version and will stop the server from starting.

Link to post
1 hour ago, sgraaf said:

Quick question: what is the "Cache path" exactly? Is this the transcodes path?

This is where the server Caches for example pictures and other media that doesn't have to be persistent and yes you can set that path also to your transcoding directory.

Got to Server settings -> Playback -> Transcoding path.

Link to post

I have my Cache path set to a directory within the appdata. I think it would not be a good idea to leave it unset because it would then cache the pictures inside the container itself, taking up space within the docker.img file.

 

39965975_ScreenShot2021-03-12at00_38_23.png.41d4a82d99d2bc4255e1805b8d245cbf.png

 

Link to post
1 hour ago, John_M said:

What I'm missing now is the AMD equivalent of Intel GPU top, namely radeontop. Could it be installed via a plugin, in the same way?

Nice, didn't know that it's named radeontop :)

I will look into this in the next few days. :)

Link to post

I will link my original Reddit post here (

In short, my nvidia GTX 1660 is not being used for hardware transcodes citing a [hevc @ 0x55ec83325180] No device available for decoder: device type cuda needed for codec hevc. Here is a snippet of logs I captured..

 

The 4K HEVC SDR movie transcode log has the following:

[matroska,webm @ 0x55ec830e1180] Could not find codec parameters for stream 4 (Subtitle: hdmv_pgs_subtitle (pgssub)): unspecified size
Consider increasing the value for the 'analyzeduration' and 'probesize' options

[AVHWDeviceContext @ 0x55ec837f1800] cu->cuInit(0) failed -> CUDA_ERROR_UNKNOWN: unknown error Device creation failed: -1313558101.
[hevc @ 0x55ec83325180] No device available for decoder: device type cuda needed for codec hevc. Stream mapping:
Stream #0:0 -> #0:0 (hevc (native) -> h264 (h264_nvenc))
Stream #0:1 -> #0:1 (copy)
Device setup failed for decoder on input stream #0:0 : Unknown error occurred

 

The 4K HEVC HDR Movie transcode log as the following:

[AVHWDeviceContext @ 0x5625c1e2e880] cu->cuInit(0) failed -> CUDA_ERROR_UNKNOWN: unknown error Device creation failed: -1313558101.
[hevc @ 0x5625c1e2ff40] No device available for decoder: device type cuda needed for codec hevc. Stream mapping:
Stream #0:0 -> #0:0 (hevc (native) -> h264 (h264_nvenc))
Stream #0:1 -> #0:1 (eac3 (native) -> aac (native))
Device setup failed for decoder on input stream #0:0 : Unknown error occurred

 

Previous transcode logs showed the following:

[AVHWDeviceContext @ 0x55b64d1ac1c0] Failed to get number of OpenCL platforms: -1001.
Device creation failed: -19.
Failed to set value 'opencl=ocl:0.0' for option 'init_hw_device':
No such device Error parsing global options: No such device

 

Any ideas of what could be my problem? I have the nvidia drivers installed, I can see my GPU on the dashboard and its activity, I have restarted and gotten a fresh JellyFin download, nvidia-smi returns my gtx.

Link to post

@Waddoo tried it now on my own on my DevServer with an GTX1050Ti and it should work:

grafik.png.99ebc9082b09b09025e1cd5fb53a0fa9.png

 

Here are a few screenshots from my template (please check also if you got no leading space at Nvidia Visible Devices) :

grafik.thumb.png.ba04671b8fbbd78877131a5b8e756eb6.png

grafik.thumb.png.e2a30b196dca97d4afc018cf462d73b3.png

grafik.thumb.png.296adcd02844ecad361a553c588ee49c.png

 

This is the output from 'nvidia-smi' when I transcode from 4K to 480p:

grafik.png.e3a1298da36151b613a4c6cfca1373bd.png

 

This is the output form 1080p to 480p:

grafik.png.2cf80b69990849d8f539704e53472b53.png

 

Here are my settings from Jellyfin:

grafik.thumb.png.68702681d929c3582eb7bdf77041eb43.png

 

 

Have you eventually activated the last checkbox 'Tone Mapping active'?

Can you send a few screenshots from your settings?

Does this also happen when you transcode from 1080p to something different?

 

I think something isn't working at passing through your Nvidia card to the container.

Link to post

@ich777 I have setup the docker with the images below, let me know what you think is my next best step forward. I did remove the /dev/dri since it read to remove it but apart from that I think my settings are matching yours. 

Quote

Have you eventually activated the last checkbox 'Tone Mapping active'?

I tried it with that checked and without, same result of the JellyFin session recognizing it needs transcode but video is frozen. 

 

I have included as many screenshots as I can showing what I see. I have also included a log file of a 1080p video that has the same fate as the 4K HEVC SDR. Sitting frozen with the timeline moving both on my player and dashboard, but no video/audio and no transcode process running. 

Screen Shot 2021-03-16 at 00.54.03.png

Screen Shot 2021-03-16 at 00.54.50.png

Screen Shot 2021-03-16 at 00.55.33.png

Screen Shot 2021-03-16 at 00.57.30.png

Screen Shot 2021-03-16 at 00.58.59.png

Log.txt

Link to post
36 minutes ago, Waddoo said:

@ich777 I have setup the docker with the images below, let me know what you think is my next best step forward. I did remove the /dev/dri since it read to remove it but apart from that I think my settings are matching yours. 

Have you also added '--runtime=nvidia' at Extra Parameters (to see that you have to enable the 'Advanced View' with the switch in the top right corner of the template page).

Link to post
4 hours ago, ich777 said:

Have you also added '--runtime=nvidia' at Extra Parameters (to see that you have to enable the 'Advanced View' with the switch in the top right corner of the template page).

If only it were easy as a missing parameter. Yes, I did remember to add —runtime=nvidia. Separated by a space next to the restart statement like in your image. I could give the entire server a reboot but not sure why the docker is t getting the gpu passed. 
 

if I typed the nvidia gpu ID incorrectly, wouldn’t the docker not boot?

 

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker create --name='Jellyfin-AMD-Intel-Nvidia' --net='bridge' -e TZ="America/Chicago" -e HOST_OS="Unraid" -e 'NVIDIA_VISIBLE_DEVICES'='GPU-201c0321-0f64-c185-4e78-ded96ed40231' -e 'NVIDIA_DRIVER_CAPABILITIES'='all' -p '8096:8096/tcp' -p '8920:8920/tcp' -v '/mnt/user/appdata/Jellyfin-AMD-Intel-Nvidia':'/config':'rw' -v '/mnt/user/Entertainment/Movies/':'/movies':'rw' -v '/mnt/user/Entertainment/TV Shows/':'/tv':'rw' -v '/mnt/user/Entertainment/Music/':'/music':'rw' -v '/mnt/user/JellyFin_Cache/':'/cache':'rw' --restart=unless-stopped --runtime=nvidia 'ich777/jellyfin' 

16c5f50a532e94b845cdfdb57e499aadc775f10a5e0a23650adc64bb9779173b

The command finished successfully!

Edited by Waddoo
Added docker start
Link to post
6 minutes ago, Waddoo said:

Yes, I did remember to add —runtime=nvidia.

These are two "-" not one...

 

6 minutes ago, Waddoo said:

I could give the entire server a reboot but not sure why the docker is t getting the gpu passed.

Can you try to open up a console window from the container itself (Click on the icon of Jellyfin and then click on Console) and then type in 'nvidia-smi'.

 

6 minutes ago, Waddoo said:

if I typed the nvidia gpu ID incorrectly, wouldn’t the docker not boot?

The container should still be able to start up.

Link to post

See my docker command from my previous post, shows all the parameters on startup. 
 

I have added a photo of the Jellyfin console and nvidia-Smi result. 
 

log.log
above is the ffmpeg log from a restarted docker to grab the docker command. 

Would you consider scheduling a time to do a screen share and that way see the Docker image on the fly?

3D429E90-85AA-4D0F-94E1-E1C9387F4A81.png

Edited by ich777
moved log into file to not bloat the thread
Link to post

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.