Это видео недоступно.
Сожалеем об этом.
Transform your streaming experience with GPU transcoding on Jellyfin Server
HTML-код
- Опубликовано: 15 авг 2024
- Today we're installing and configuring Jellyfin server on:
- Windows with GUI (and VM)
- Linux CLI
- Proxmox Virtual Machine with GPU Passthrough to the VM
- With a Reverse Proxy for SSL and CrowdSec for Securing Web Access
I discuss what GPU transcoding is, and show you how to use it on your server to increase the number of simulataneous streams. This video can be used for Intel, Nvidia, and AMD cards.
Jellyfin: jellyfin.org/
Jellyfin Ubuntu Single Script: jellyfin.org/d...
GitHub File: github.com/Jam...
Discord: / discord
Twitter: / jimsgarage_
Reddit: / jims-garage
00:00 - Introduction to Jellyfin, Transcoding and Using your NAS
01:09 - To Transcode or Not to Transcode... (GPU vs CPU)
01:58 - Transcoding Recommendations
04:00 - Windows Install
04:40 - Ubuntu Install
04:50 - Proxmox Ubuntu Install
05:11 - VM GPU Passthrough Prerequisites
06:18 - How to Enable GPU Passthrough
07:42 - Docker Compose Configuration Review
10:58 - Jellyfin Web GUI Configuration
12:48 - Validating Installation
Thank you for the video. Informative and well presented as always.
Starting my first server on a HP EliteDesk 800 G3 SFF with i7-7700 in a few days and this video drops right in time.
Will be moving my current Jellyfind from Docker instance on Synology NAS to an instance on Proxmox.
Thanks, Nashnir, I appreciate the feedback. That's exciting, do join my Discord if you'd like any assistance, we have a great group of helpful people now.
Another amazing video, well worth a watch!!
Thanks, Michael. 😁
Thanks for the demo and info, have a great day
Thanks, Chris. You too!
Question for you @Jims-Garage. The command to mount the share you put on the screen at 8:44 is different than the one you used in the TrueNAS video. That video, you edited fstab and then simply used `sudo mount -a'. In THIS video, you didn't edit fstab, but used this other, longer mount command. Does the command used in this Jellyfin video "work better" than the other way? Also, what is the best way to have the Jellyfin or Plex container not start until the mount is complete upon reboot? I currently use rClone in my Plex stack and use "depends on"... how do you handle that? I don't remember seeing it in any of the series, which I've watched several times. Thank you for these tutorials. IMHO, they are the best out there!
@@FloTexaTenn from memory: the commands are the same. One edits the fstab file, the other pipes the data into the fstab file. Same result.
Not sure about the mount order. My understanding is that it's mounted at boot before a docker container (I've never had an issue). There could be an elegant wait solution though.
@@Jims-Garage Thank you for the quick reply. I can certainly Google it, but the command in the Jellyfin video is sudo mount -t cifs -o......xxxxxxx. I was wondering if the -t or -o affected wait time. In other tutorials I've used, plex would spin up before the share mounted, which would result in the library not being available within plex or jellyfin... I'd have to restart the plex/jellyfin containers so they could see the mount. I haven't gone off of rclone because of this. If your way here negates that problem, I'd love to switch to be 100% Jim's Garage! Thanks again. EDIT: The share is on a Synology NAS, FWIW.
Thank you for the informative video:
1- Is the GPU transcoding dependent on the OS/container (thinking about nvidia drivers on linux and the difficulty to pass the gpu to a virtualized environment, but I'm not knowledgeable at all). also, if i do passthrough the gpu to a container running jellyfin, will it be unusable by other services on other containers?
2- is gpu transcoding limited to a number of concurrent transcoding streams? On all hardware (quicksync, NVENC etc)?
You're welcome.
1) You can pass an Nvidia GPU to a VM, the host doesn't need drivers in that scenario (only for an LXC). Only that VM will be able to access it, but all containers on that VM can use the GPU. Another way to do it is to use LXCs, you can have many LXCs to a GPU but it's more complicated (I have a video on it).
2) There are no transcoding limits for Intel, I believe there might be on Nvidia but it changes with series.
@@Jims-Garage thanks you again
Very nicely explained, thank you. One question, is it possible to setup lxc container with docker running jellyfin and support for transcoding? I understand that setup in the video is on VM. Cannot find much info about lxc+docker variant.
Sure, at the end of the day an LXC is just a containerised OS. The only thing that is different is the configuration of the LXC for hardware passthrough.
Good video.
Thanks 👍
Nice video
Cheers, Adam ☺️
Hi Jim starting to learn a lot about these topics with you, great explanations. I’m having an issue with jellyfin I followed everything in detail and pass through proxmox worked to the Ubuntu but jellyfin is not transcoding via hw do I have to run the container as root ? Or with the 1000 user should work ?
When you do lspci do you see the card? Did you enable transcoding in jellyfin?
@@Jims-Garage yes I can see the card and that’s thanks to you I have been struggling with this for a long time . Yes in jellyfin I tried both render cards just like you show on this video exact same names , I have a ser5 belink AMD 5640, says vga Radeon vega series
@@jafandarcia what option did you choose in Jellyfin? Vaapi?
@@Jims-Garage yes I tested both of them vaapi and and I did start selecting everything I saw so I don’t remember the defaults , testing with an iPad some videos just stall hvec format and on tv only direct streaming works
@@jafandarcia worth asking in Discord, I don't have an AMD card to test with.
What is the benefit of hosting the JellyFin server in VM/Docker instance over just on windows?
Nothing pure performance wise, more convenience and the nature of being a server. Typically you'll have a server running Linux with Docker/Kubernetes/LXC and with that comes the benefit of running multiple services on a single host that is permanently on. Containers are a great way to maximise your hardware.
In my case I have a Synology where my media is, but when using docker, the jellyfin doesn't have access to the network folder Z:/ that mounts the NAS network folder. What can I do here?
How did you setup the share within DSM? How did you create the shared folder within Docker?
cant jellyfin take advantage of AMD 780M hw transcoding? or is strictly feature of intels?
AMD is reported as working in Jellyfin, but my understanding is it's less mature and performant than intel (quicksync)
I think someone managed to get a Raspberry Pi to do transcoding, I'd need to find the video though of them showing it working.
Yes, I think it has some support, but not full (there are dependencies on ffmpeg). I had forgotten that.
I can assume an i7-4770K would be sufficient to handle this?
Should be fine for h264 but its version of quicksync doesn't support h265.
Hola me puede hace un video configurando este complemento de jellyfin se llama LDAP porfa
I'm not aware of the LDAP plugin for Jellyfin but I'll take a look (sorry if that's not accurate, I'm working off a translation).
hi id like to use my igpu amd radeon on my gaming pc for tramscpding. cpu is ryzen 5 7600 6 cores
. need help lol
seems to tedious for just for some transcoding...
Thanks for the comment. Which part do you mean? If you don't need transcoding just don't use it, or just use a dedicated machine (albeit it's a waste of valuable system resources).