[Support] ich777 - Jellyfin AMD/Intel/Nvidia [DEPRECATED]

[JonathanM]

6 minutes ago, RYZEN5000 said:

i dont want to mess around so no more updates for me, mabye once every few months from here on out.

I'm sure that will solve your issues.

[ich777]

10 hours ago, RYZEN5000 said:

ich777, the guy responsible.

I'm not entirely sure that I'm responsible for that that your container updates that often because the last stable was built 2 months (translated from: "2 Monate") and 18 days (translated from: " 18Tage") ago and no update was pushed since then (as you can see from the screenshot):

 

As said above I'm also using the container and I have no issue if I click on (what is basically the same when an update happens) or when I bring up the context menu and click on .

 

10 hours ago, RYZEN5000 said:

Thats frustrating to me, I spend many hours configuring things the custom way that I like them.

Maybe there is some configuration error in your template? Can you post screenshots from your template? Never experienced a problem wiping anything.

 

10 hours ago, RYZEN5000 said:

I cant chose a different container because I just bought a nvidia 1050ti and this is the only container with the guid input to use it.

That's not true, you can also use @binhex or @linuxserver.io container, they both support Nvidia transcoding, I can tell for sure because I've tried both with NVENC and they work fine.

 

10 hours ago, RYZEN5000 said:

The instructions given are clearly to leave it blank and that simply won't work.

I already changed the instructions as said above, again thank you for the suggestion, you have to remove the entire entry with the marked button in the screenshot (the red asterisk was done on purpose):

 

10 hours ago, RYZEN5000 said:

I see software everywhere, not just your container that isn't functioning as it was decscribed to and I am in a battle to keep a brand new server running, everytime i turn around its always an update the program unraid offered that spoiled my work, I am battling it but I am getting frustrated with working full time to keep it running.

I don't have any problems with Nextcloud or any other container that I run on my Server, regardless if it is from @binhex, @linuxserver.io or anybody else.

 

Maybe there is some configuration issue in your templates that causes your configuration is wiped after an update, maybe even the shares in combination with the "Use Cache" setting is set in a different way that it is not intended how it should be.

 

10 hours ago, RYZEN5000 said:

BTW my server is a brand new beautiful glowing rgb monster. Its a work of art.

I turned off all RGB because I don't need RGB in my closet...

[PatientXero]

I found that the stock template needs to be modified if you are using NVIDIA transcoding. The additional parameter of NVIDIA_DRIVER_CAPABILITIES with the key of 'all' fixed by transcoding issues. The maintainer needs to fix their template, changing Nvidia Driver Capabilities to NVIDIA_DRIVER_CAPABILITIES.

 

Now the other thing, is that transcoding does not kick in unless you choose a lower bitrate than what the file is originally encoded at. I've watched my GPU dashboard to verify that.

Edited November 24, 2021 by PatientXero

[ich777]

8 hours ago, PatientXero said:

I found that the stock template needs to be modified if you are using NVIDIA transcoding. The additional parameter of NVIDIA_DRIVER_CAPABILITIES with the key of 'all' fixed by transcoding issues. The maintainer needs to fix their template, changing Nvidia Driver Capabilities to NVIDIA_DRIVER_CAPABILITIES.

Exactly, in templates where these vatiables are missing you have to manually add them, see here (second post):

 

[PatientXero]

On 11/23/2021 at 10:55 PM, ich777 said:

Exactly, in templates where these vatiables are missing you have to manually add them, see here (second post):

 

That's not what I meant. You already have 'Nvidia Driver Capabilities' in your template, which isn't properly formatted. If you formatted it to "NVIDIA_DRIVER_CAPABILITIES", that would fix the issue. Telling someone to add it doesn't help your case, and is is sloppy development. 

Edited November 25, 2021 by PatientXero
Reformat

[ich777]

15 hours ago, PatientXero said:

That's not what I meant. You already have 'Nvidia Driver Capabilities' in your template, which isn't properly formatted. If you formatted it to "NVIDIA_DRIVER_CAPABILITIES", that would fix the issue. Telling someone to add it doesn't help your case, and is is sloppy development. 

I really don't get what you mean, I think you are talking about the name which is "Nvidia Driver Capabilities" and is filled in because it looks way nicer to the user than not filling it in and showing only the bare key entry "NVIDIA_DRIVER_CAPABILITIES", it is formatted properly from what I see:

 

On 11/23/2021 at 10:57 PM, PatientXero said:

The additional parameter of NVIDIA_DRIVER_CAPABILITIES with the key of 'all' fixed by transcoding issues.

Also please keep in mind that the key always has to be the variable, in that case "NVIDIA_DRIVER_CAPABILITIES" and the value always has to be the value from the variable, in this case "all".

 

 

BTW Thank you for the roses...

[Zaphragor]

Hi, 

 

I'm on unRAID 6.9.2 and get a green-screen if I write the amdgpu.conf and reboot. 

 

I'm currently using a Ryzen 2400G without any dedicated PCIe graphics card. I'm pretty new to unRAID and not really sure how to debug this error. 

 

I tried to change my main graphic output in the BIOS, but that won't help. 

 

PCIe ACS override is set to "Both".

 

Installing Radeon Top causes the same error. 

 

Any ideas what could cause this?

[RYZEN5000]

On 11/22/2021 at 3:52 PM, RYZEN5000 said:

ich777, the guy responsible. Hi, and binhex calm yourself. I am not meaning to be abusive but this constant updating is counterproductive. I set the updates to once a month in my settings, no point building up my server to have it broken down every few days with a new update. the containers wont run after. Thats frustrating to me, I spend many hours configuring things the custom way that I like them. This is not your average low end hardware its a special build to me. The cost was extreme all brand new parts so of course I'm going to get upset when its out of service from clicking update. Its not just you, the nextcloud update wont connect to the database after it updates i have to reinstall both nextcloud and postgresql14. All the other apps I use are fine after an update. Im backin up all the config folders so downtime is minimal but I was up all night last night before i got it to reinstalled because it opened and ran but screen was stuck at loading spiral with only some of my media library running after the update. I had to assign new permissions and completely delete the entire Jellyfin Nvidia folder in appdata before the new installation would work. The new installation I did just this morning has been scanning my media all day and still isnt finished. resulting in me doing this every few days and my drives never get to spin down. The only solution is to stop updating. I have a modest media collection of 1200 movies and several dozen complete tv series.

 

I cant chose a different container because I just bought a nvidia 1050ti and this is the only container with the guid input to use it.

the binhex container i used before this one and for the record it always worked even after updates, so calm down your container is all good. 

 

I tried everything to get the container to start during an install and im talking about what goes into that device field* I have been putting /dev/nvidia0 becuase it wont start with it cleared. it also wont start with the card guid number for the card in the field, I have the guid in the nvidia visible devices field already. it does work with /dev/dri and it works with nvidia-uvm but the processor load seems a few digits higher with that one. The instructions given are clearly to leave it blank and that simply won't work. It transcodes with /dev/nvidia0 but its still using 8-11% cpu which is low compared to the %75 cpu it was using before I put in the new 1050TI.

 

I sympathis with you guys for not getting paid, that really sucks but I look around and I see software everywhere, not just your container that isn't functioning as it was decscribed to and I am in a battle to keep a brand new server running, everytime i turn around its always an update the program unraid offered that spoiled my work, I am battling it but I am getting frustrated with working full time to keep it running. I am spending 12 hours a day on unraid, whether its watching movies editing, adding to my media collection, I am building something great that alot of other people don't have I have super high expectations of the software, and only the best highest quality parts money could buy went into my build.

 

Light 4 apps asked to update all at once and I was like ok I'll try it but omw I will never let them all upodate again. I'm not going to update them when they are working fine. The Nextcloud update the last one that came out was aweful, before the update I was getting 70-99mbs upload speeds and after I was only uploading at %17 to %60 sometimes %70 peak. Why did they mess with it when it was soo fast??? Took forever to upload at 60mbs hovering around 40 most of the time all over the place. I want 99mbs back!!

BTW my server is a brand new beautiful glowing rgb monster. Its a work of art.

Update, I ran all the updates today, unnasigned devices Plugin, postgresql14, nextcloud and nothing broke. They did fix the previous issues with the updates crashing the app requiring deletion of the appdata folder and reinstall. I have not gotten an update from Nvidia Jellyfin yet but it is working fine since i last reinstalled it. Server has run fine with no updates for 6 days straight it ran Tdarr and went untill the power went out this afternoon resetting the timer. 

 

It would be nice to see a new development in the tv app to make custom themes for the TV instead of the same old background. If we could manipulate the tv app after login to reflect what Jellyfin skin looks like on the PC. Is there a way to do that?

 

Agkserver

https://photos.app.goo.gl/snTSrb96zEeJ5Qyf8

[ich777]

2 hours ago, RYZEN5000 said:

I have not gotten an update from Nvidia Jellyfin yet but it is working fine since i last reinstalled it.

If you want to test if it is working after an update or at least simulate a update from the container do the following:

1. Go to the Docker page in unRAID
2. Turn on Advanced View on the upper right corner
3. At the container click on

Don't forget to turn of Advanced View again on the Docker page because that produces actually a load on your server even if no one is logged in.

 

2 hours ago, RYZEN5000 said:

t would be nice to see a new development in the tv app to make custom themes for the TV instead of the same old background. If we could manipulate the tv app after login to reflect what Jellyfin skin looks like on the PC. Is there a way to do that?

I would actually recommend to create an issue on the official Jellyin Github page because I'm not a developer from Jellyfin itself and that is out of my control: Click

[ich777]

4 hours ago, Zaphragor said:

I'm currently using a Ryzen 2400G without any dedicated PCIe graphics card. I'm pretty new to unRAID and not really sure how to debug this error. 

I would recommend to create a Bug Report here: Click

 

4 hours ago, Zaphragor said:

I'm on unRAID 6.9.2 and get a green-screen if I write the amdgpu.conf and reboot. 

4 hours ago, Zaphragor said:

Installing Radeon Top causes the same error. 

My plugin does basically the same as when you create the amdgpu.conf file but more direct, I don't think that is related to the plugin itself and more a hardware/configuration issue. Please create a post on the Bug Report sub forums.

[Zaphragor]

Thank you very much for the fast reply! 

 

Yes I will do so. 

[rodan5150]

Anyone have luck with tone mapping using a Quadro P400? Hardware transcoding seems to work fine as long as I disable tone mapping. It would be nice to have tone mapping, but obviously not a deal breaker. Just a slight annoyance since my 4K HDR content looks pretty bad if it is transcoded without tone mapping. 

[ich777]

3 hours ago, rodan5150 said:

Anyone have luck with tone mapping using a Quadro P400?

No issue here with a Nvidia T400.

 

Do you chnaged any of the settings?

[rodan5150]

10 hours ago, ich777 said:

No issue here with a Nvidia T400.

 

Do you chnaged any of the settings?

Not sure of the decode/encode difference between P400 and T400, that may be a factor.

But no, I just have it default. I've attached screen shots.

 

Thanks for the quick reply, you are a great asset to this community!

[ich777]

13 hours ago, rodan5150 said:

But no, I just have it default. I've attached screen shots.

That's really strange, I've also tried it now and it doesn't work either, but I also have to say that it seems now I've messed up my Jellyfin installation because I can't tone map any more, will look into this.

[rodan5150]

On 12/13/2021 at 1:58 AM, ich777 said:

That's really strange, I've also tried it now and it doesn't work either, but I also have to say that it seems now I've messed up my Jellyfin installation because I can't tone map any more, will look into this.

Thanks again for your help. No rush on this, I'm primarily a Plex user, I'm just in testing phase with Jellyfin at this point. Looking to have a backup option for Plex. I've been a Plex user for a decade or so, and I get worried about the direction they are headed in sometimes. Plus, I love to support the open source community, so Jellyfin makes a lot of sense to me vs Emby.

[ich777]

1 hour ago, rodan5150 said:

Emby

I have to say that I'm a Emby user but also have Jellyfin as a backup plan and Plex only for testing if the Nvidia Drivers work or various other things to test.

[Reflexion]

Thanks for the effort; ich777 for the easy creation with HW tranc.
What I noticed when rebuilding with the current unRAID beta is that you still have to use "CA Cleanup Appdata" otherwise you will take old data with you...

I also had the problem with playback with the QSV setting, so make sure that VAAPI device;
/dev/dri/renderD128

...however, there is something I miss and at least "back then" was actively running, RAM transc. I had at that time another container "dee31797's Repository ) container " under it it was as in the picture to recognize possible RAM transcodierung to use 50% was taken, what my NVMe already^^ ...in Jellyfin one had to change then "/config/transcodes" in "/transcode", finished,.

unfortunately... this does not work anymore. I had to reinstall the container because the settings were shot in Jellyfin.

Edited December 18, 2021 by Reflexion

[ich777]

26 minutes ago, Reflexion said:

What I noticed when rebuilding with the current unRAID beta is that you still have to use "CA Cleanup Appdata" otherwise you will take old data with you...

I don't understand what you mean by that...

If you want to set up a new instance from the container you always have to delete the whole container and the folder for Jellyfin that was created in your appdata directory.

 

27 minutes ago, Reflexion said:

I also had the problem with playback with the QSV setting, so make sure that VAAPI device;
/dev/dri/renderD128

That depends on the model from your iGPU if QSV is working or not, I would recommend to use VA-API if QSV won't work.

 

28 minutes ago, Reflexion said:

...however, there is something I miss and at least "back then" was actively running, RAM transc. I had at that time another container "dee31797's Repository ) container " under it it was as in the picture to recognize possible RAM transcodierung to use 50% was taken, what my NVMe already^^

The main issue with Jellyfin is that if the transcoding directory runs out of space the playback will stop too, see this Github Issue: Click

However you can use whatever path you want inside the container you only have to set it first:

1. Open Jellyfin
2. Go to your Admin Settings
3. Go to Playback
4. At the transcoding path enter the path you want to use, in your case I think this will be '/transcode'

Don't forget to map the path in the template as in your screenshot.

[Reflexion]

@ich777correct, deleting manually or with CA Cleanup Appdata works, but unfortunately not the classic deletion of the container.

Correct, I have adjusted '/transcode', saved and rebooted... but after that there is no more playback and the settings for HW encoding are emptied.

With the omission of RAM transc. and the classic VAAPI, it works well so far, to the sorrow of my NVMe's.

[ich777]

Just now, Reflexion said:

@ich777correct, deleting manually or with CA Cleanup Appdata works, but unfortunately not the classic deletion of the container.

Why should this work?

If you want to start new you have to do this even for any other container too.

 

Just now, Reflexion said:

Correct, I have adjusted '/transcode', saved and rebooted... but after that there is no more playback and the settings for HW encoding are emptied.

This is not normal, please check your settings, also keep in mind with my container you have to set your appdata share to Use Cache "Only" or "Prefer" if you don't customized the path because it searches for it on the Cache.

[jafi]

Hi,

 

I have used jellyfin before, but updated my homelab. So now I am testing this version of jellyfin, because I have AMD iGPU. I can connect to jelly via lan (IP) fine, but not via reverseproxy (public domain). 

 

With public domain I get message: "select server" and url is https://jellyfin.mydomain.com/web/index.html#!/selectserver.html

 

I use nginx proxy manager and all other services work fine.

I have activated "Allow remote connections to this server".

Also I added custom variable to the tempalte: JELLYFIN_PublishedServerUrl, value: jellyfin.mydomain.com

[ich777]

19 minutes ago, jafi said:

I have used jellyfin before, but updated my homelab.

20 minutes ago, jafi said:

With public domain I get message: "select server" and url is https://jellyfin.mydomain.com/web/index.html#!/selectserver.html

Please try to delete the browser cache.

 

This can either be a browser issue (caching) or issue with your with your Reverse Proxy if you can reach it through your internal IP.

[Cicatrix]

Hey there,

 

tried to Play a Video via Jellyfin-App in my Smartphone (Xiaomi Mi 9T pro). Working with Pentium G6605. Video works great via Fire Stick 4k on a Samsung 4k TV, but in the Phone it is Just some colors. Attached Screenshot of the colors and the Video information. Intel GPU Top is installed, path /Dev/dri. Amy ideas?

 

Edit:

OK, Changed from Quick Sync to VAAPI and now it Is working "OK". Sorry and Ty.

Edited December 22, 2021 by Cicatrix

[limital]

i have a strange problem with jellyfin. Im using an amd ryzen 4650g and configured the hardware acceleration but i'm getting strange colorful flickering on all my media. i found out that the flickering disappears when i disable hardware encoding in settings. Any idea why that happens?

Edited December 25, 2021 by limital
picture added