Is it just my instance, or does Jellyfin have difficulty skipping (forwards or backwards), pausing - especially for extended periods of time, or restarting a video where I last left off?
I have no issues on my end with any of 6 different clients. How are you running jellyfin? Docker or actual install?
rootless podman. There will also be some overhead from the nginx reverse proxy as well. I get 2Gig Fiber from my ISP, so that’s probably not the issue.
No issues on my end. Though I’m running it on “bare metal” and have it behind caddy(reverse proxy) and WireGuard (VPN). Some of the clients can be buggy and local network conditions can cause some oddities. I have had tracking break for me when using the windows client on a previous version. And will occasionally notice similar issues when I’m away from home using cellular or using suspect WiFi.
I experience this sometimes, but I’m sure it’s a problem with my server hardware, not a particularly strong machine.
Yeah. It sounds like my setup is very similar. The server it runs on is a RHEL 9.6 host, and accessing it from my public domain means you hit a Linode VPS, which uses Nginx to reverse-proxy the connection to the physical machine via a Wireguard tunnel.
I use an AMD Ryzen 5 4650G and ECC memory.
Try with different source material. Maybe it’s not an issue of Jellyfin, but of the underlying file.
Sorry it’s taken so long for me to get back to you. I didn’t really do a lot of testing of this. But turns out that it seems that my internet connection was just bad.
Did you get that 4650g to do video transcoding?
Yeah. Although, it can only encode. I have to use software decode because it doesn’t support AV1, which is the codec in which my media is stored on the disk.
What Linux configurations are people running here for Jellyfin? My JF server is one of the last of my machines I still need to kick Fedora off of (along with my primary desktop, which I hate to take down and rebuild while waiting for Cinnamon to be working on OpenMandriva). Currently it’s on a Dell Optiplex 5040 with 16G memory.
My setup is very similar. I run Nginx on a Linode at the “nanode” tier, but it uses the Wireguard tunnel to relay requests between the server and the open internet.
Once I get back home, I’ll SSH into my homelab and share my Quadlet File(s) and container environment files. I am not sure if they’ll be what you’re looking for in your case though.
So I had a similar issue. I had an 8th gen Intel IGPU, with massive transcoding issues. Same exact behavior. seeking, pausing, etc… were all fraught with waits.
Once I went out and grabbed a used A380 and got everything configured for it, everything was functionally instant.
I have a 7th gen Ryzen system with 64GB of ram running TrueNAS. I use the official jellyfin container through the truenas Apps section, and I’ve got it limited to 4GB of memory and 4 cores. Works fine, with the A380 passed to the container.
Yeah. I currently use an A380 to encode the media with my media managment destkop system. I am waiting on Intel to release a damn B770 or something so I can downcycle it into the homelab. I wonder how easy it will be to make this work with two GPUs though.
Though in the case for my original complaint, the client device in which I was using was doing direct-play because it has the hardware to be able to decode AV1.
Just point jellyfin to the GPU device node (in /dev/dri
) that correlates to the GPU you want it to use.
You’re sure it’s using direct play? Well, in that case, I can’t really speak to the issues there. Might be disk access latency or something?
I’m pretty sure it was client internet connection at the time. I was in a very rural place and only had like 2 bars of hotspot service. I have been playing other content on that device using my home network, and it was much better. Also, transcoding is not configured on my Jellyfin instance at the moment, so it wouldn’t have played at all.
â•─jellyfin@aurelion-core-server ~
╰─$ cat -p ~/.config/containers/systemd/jellyfin-server.container 130 ↵
# vim: filetype=systemd
[Unit]
Description=Jellyfin Server
Wants=network-online.target
After=network-online.target
[Container]
AutoUpdate=registry
ContainerName=jellyfin-server
EnvironmentFile=/home/jellyfin/jellyfin.env
AddDevice=/dev/dri/:/dev/dri/
PublishPort=8096:8096/tcp
UserNS=keep-id
Volume=/home/jellyfin/app/cache:/cache:rw,Z
Volume=/home/jellyfin/app/config:/config:rw,Z
Volume=/home/jellyfin/media/Movies:/Movies:ro
Volume=/home/jellyfin/media/TV:/TV:ro
Volume=/home/jellyfin/Music:/Music:ro
Image=docker.io/jellyfin/jellyfin:latest
[Service]
SuccessExitStatus=0 143
[Install]
WantedBy=multi-user.target default.target
â•─jellyfin@aurelion-core-server ~
╰─$ cat -p ~/jellyfin.env
JELLYFIN_PublishedServerUrl=https://media.linuxdragon.dev
Note: The jellyfin
user account on my server doesn’t have sudo privileges, but it does have more privileges than other systems because it’s a part of the media
group. Like all other services, it’s also in the systemd-journal
group.
I have this issue sometimes but its only with certain files and when its going through the reverse proxy. When I go to it directly I dont have the issue. I also have this issue with emby so I think its not jellyfin related.
Honestly, I’ve been thinking about creating my own media server software in C/C++, Go, or Rust. I really want to write a project in a compiled language, but I don’t really have any good ideas. Plus any project I start will have to have an intrinsic value to me, and right now, that means it will have to revolve around my homelab.
Edit: Oh and the learning opportunites for processing media in code will be super valuable to me tbh.