Login

Fate · 2024-11-20, 04:03 PM

(2024-11-20, 03:54 PM)iwishiknewmore Wrote: here are ffmpeg logs, most recent ones. https://privatebin.net/?78f167e25991f99a...HoYrXjSLgw
system is as follows:
OS: proxmox with 12600k cpu
virtualized on proxmox in vm's: opnsense, openm,ediavault and proxmox backup server. And then a debian lxc container with docker installed where the whole arr stack runs. this lxc has 4cpus, 8gb ram, 8gb swap and 40gb boot disk
media is stored on 1 16tb drive with all config, transcodes etc on a 2tb nvme.

ive since noticed that with no streams from jellyfin all ui's are responsive, but with even one stream they take much longer to load, i dont know why this is as this wasnt the case before

that log looks fine can you paste a few older ones?

I assume you have the igp passthrough as it is using it via quicksync, right?

Can you post a screenshot of the jellyfin ---> administrator-> dashbaord-> playback--> transcoding page?

Thanks!

TheDreadPirate · 2024-11-20, 04:19 PM

Does it buffer for all videos? Or just some? The particular log you shared is an audio only transcode.

Do you pass the GPU into the LXC and docker?

iwishiknewmore · 2024-11-20, 05:19 PM

yes igpu is passed through to lxc and then to jellyfin, indeed using quicksync.

here is older transcode log: https://privatebin.net/?18b73f37156208ee...9nHQD8Y5JY
and here older directstream log: https://privatebin.net/?3cbbb051d8601bb2...4hd5Aywkkv
here is transcode setting: https://pasteboard.co/MWzhiaUGrlKS.png

and yes, all videos seem to buffer, it gets worse the higher the bitrate, starting around 7Mb/s but thats not even 1MB/s so really shouldnt be an issue at all. happens on desktop app and web for both transcodes and directstreams, only android app isnt affected.

TheDreadPirate · 2024-11-20, 05:41 PM

As far as transcoding, everything is happening more than fast enough. Are you able to do a speed test for the LXC? This sounds like a network issue. Either with speed or with packet loss.

iwishiknewmore · 2024-11-20, 07:46 PM

the lxc container gets max up and download speed of 200Mb/s up and down, qbit can reach this but is limited to 100 up so it doesnt take up all the bandwidth, is there a way to test for packet loss?

TheDreadPirate · 2024-11-20, 08:12 PM

If this problem happens for all clients, including a browser, we can try using the browser console to trouble shoot.

Open Jellyfin in your browser, press F12 to open the dev console, go to the "Console" tab.

Initiate playback in the main windows and monitor the console for activity when it buffers.

iwishiknewmore · 2024-11-20, 11:37 PM

iwishiknewmore · 2024-11-21, 12:36 PM

i think it has to be a network issue then right? i just have no idea what or why as everything worked flawlessly before

TheDreadPirate · 2024-11-21, 03:56 PM

Did you install a VPN? Change routers?

iwishiknewmore · 2024-11-21, 05:33 PM

i have my isp router and then a router in front of homelab, as mentioned opnsense, both havnt changed since the start. jellyfin is exposed through the same vpn qbit is, and has been since the start, ive changed this to see if it fixes stuff but it hasnt. services are accessed through tailscale, which is the only thing i can think of that could be the issue

Login

Username/Email:
Password:


Or login with a social network below

Login

Username/Email:
Password:


Or login with a social network below

Buffering