mirror of
https://github.com/jellyfin/jellyfin.git
synced 2026-05-04 18:09:12 +03:00
Unable to resume videos #6522
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @soultaco83 on GitHub (Nov 28, 2024).
This issue respects the following points:
Description of the bug
After playing a video and returning I am unable to resume where I left off. Myself and users on my server have to restart the video from start and skip to where they left off. This has been tested with a remux and transcoding file as of now. Updated to latest master commit as of this issue creation.
Reproduction steps
Skip forward in a video enough to trigger a resume point
Back out of the playing video
refresh the page
Attempt to play the video where you left off.
What is the current bug behavior?
Currently the video does not resume playback after starting video from a previous point
What is the expected correct behavior?
Video picks up where it was left off at
Jellyfin Server version
Master
Specify commit id
1272bb9a84Specify unstable release number
No response
Specify version number
No response
Specify the build version
2024112811
Environment
Jellyfin logs
FFmpeg logs
Client / Browser logs
The HLS error appears to be irrelevant to this happening as the video plays if restarted from beginning and error appears as well

Relevant screenshots or videos
https://github.com/user-attachments/assets/24cb4e60-60be-4d08-a057-7424c00b1b1a
Additional information
FFmpeg.Transcode-2024-11-28_08-05-00_8e73c3203f25690a207053b221aa4b75_c41dc374.log
log_20241128.log
@fubar-coder commented on GitHub (Nov 28, 2024):
I have the same problem with Jellyfin 10.10.3.
@gnattu commented on GitHub (Nov 28, 2024):
My theory: your disk is slow and the video size is high, which takes a few seconds to do the seeking for ffmpeg and to send the first frame, but it is already too long and the client timed out. You can verify that by either test with a small (<1GB low bitrate but long enough) file, or make a test library on SSD.
@nyanmisaka commented on GitHub (Nov 28, 2024):
It seems the fMP4-HLS container is disabled. Can u enable it and try again?
@soultaco83 commented on GitHub (Nov 29, 2024):
@nyanmisaka I did have the Prefer fMP4-HLS turned off. Turned it back on and videos are resuming once again. Do you happen to know what this setting does exactly?
@gnattu This was working fine for awhile. Some shows that are smaller are still taking a minute but do load after a few seconds. Is there anyway to possibly increase the timeout period? As nyan stated turning on the fMP4-HLS seems to have resolved it for now.
@gnattu commented on GitHub (Dec 8, 2024):
fmp4 uses fast seek (only seek to key frames) while mpegts is using accurate seek (seek to the exact frame) and the latter is more IO and compute intensive. This timeout is not easy to customize unless you want to rebuild the web by yourself. If you don't have special requirements then using fmp4 is a good enough workaround for you.
@soultaco83 commented on GitHub (Mar 14, 2025):
Closing issue. Enabling fmp4 is working for me and i have just set the default user settings in my docker image to just have that one for everyone.