SOLVED: Movie playback pauses every X minutes

[calthejuggler]

Hello there!

I am having an issue where all the movies on my Jellyfin server seem to pause after a consistent period of time. I can then get it to restart by clicking back to the menu, re-selecting the movie and fast forwarding to the part I was at. For example, if the movie pauses after 1 hour, once I've got it back again, it will then pause at hour 2. If the movies pauses at 45 minutes, it will then pause at 1hr 30m. I assumed this was a memory problem, so I went into Grafana to check and it seems like my cache is filling up, dropping, filling up etc.
[Some of my grafana graphs where the memory cache increases, drops, increases, drops as I restart the movie. The swapfile usage is quite high.]https://pasteboard.co/AzNlSBuTuVrM.png[/img]

From this, I realised my swapfile was only 1GB, which I thought could be the problem, but I have now increased it to 24GB (1.5x my RAM capacity) and I am receiving the same problem:
[A more recent screenshot of my Grafana board, where the memory ran out quickly, but the swapfile is not being used nearly as heavily]https://pasteboard.co/fDNAuga23Avl.png[/img]

Now, I'm not sure what my next move should be an was wondering if there are any common causes of such a problem? I've had a search on the forums, but I couldn't find my exact issue. It's not a major problem, just a minor annoyance It seems to be server related, as it happens no matter which client I use.

My server specs:

[*]- Jellyfin version: 10.10.3
- Installation: Linux Docker (jellyfin/jellyfin:latest) with docker compose
- OS: Debian GNU/Linux 12 (bookworm) x86_64
- CPU: Intel i5-7500 (4) @ 3.800GHz
- GPU: Intel HD Graphics 630
- Memory: 16GB

My compose.yml is as follows:

services:
  jellyfin:
    image: jellyfin/jellyfin:latest
    container_name: jellyfin
    user: 1000:1000
    group_add:
      - 44
      - 105
    network_mode: 'host'
    environment:
      TZ: Europe/London
    volumes:
      - ./configconfig:rw
      - ./cachecache:rw
      - /dev/shmtranscode:rw
      - type: bind
        source: /mnt/media/library
        target: /media
    devices:
      - /dev/dri/card0dev/dri/card0
      - /dev/dri/renderD128dev/dri/renderD128
    restart: 'unless-stopped'

And finally, my Jellyfin log:
https://pastebin.com/EykKLVBD
and my ffmpeg log:
https://pastebin.com/Ahd3XEwp

Both of these logs are from the first Grafana image above. I believe I was watching "Shaun of the Dead"

Happy to provide any more information. I'm very interested to know where the first ports of call to check out are.

Thank you in advance!

[*]EDIT:
[*]This seems to have been solved by turning on "Delete segments" in Admin>Playback>Transcoding. The graphs I have provided above, are still peaking in the same way, but the "Memory committed" graph has an interesting difference:
https://pasteboard.co/COmQIPrLPMKj.png
[*]The red line shows where I upped my swapfile size, I then try to watch something to test which paused at 16:03-ish. Note the absolute cap that causes the straight line from the area chart. After turning on "Delete segments", there is still a peak, but it is "wavy". This makes me feel like the memory just caps out and can't take any more.

[*]Some questions I have now are about why my swapfile doesn't seem to be doing very much at all, but I'm just happy my movies are playing through

[TheDreadPirate]

Caches and buffers are not the same as "used", which remained low. Linux, and Jellyfin in particular, caches aggressively. Memory used for caching is not reserved in the same way and will be released immediately when another app, or Jellyfin itself, needs it. So that is likely not the issue.

Nothing in your logs indicates a server side issue. It is transcoding+tone mapping fast enough. How is the server and client connected to the network? WIFI? Wired? If WiFi, what is the signal strength like on each?