Jellyfin Forum
Cannot start server after upgrade: FFmpeg validation: The process returned no result - Printable Version

+- Jellyfin Forum (https://forum.jellyfin.org)
+-- Forum: Support (https://forum.jellyfin.org/f-support)
+--- Forum: Troubleshooting (https://forum.jellyfin.org/f-troubleshooting)
+--- Thread: Cannot start server after upgrade: FFmpeg validation: The process returned no result (/t-cannot-start-server-after-upgrade-ffmpeg-validation-the-process-returned-no-result)

Pages: 1 2


RE: Cannot start server after upgrade: FFmpeg validation: The process returned no result - void - 2025-03-06

+1, I have jellyfin running on a debian box with unattended upgrades and recently rebooted the box to find jellyfin wasn't launching. After a bit of debugging it turns out the team split jellyfin-ffmpeg into jellyfin-ffmpeg7, jellyfin-ffmpeg6, and jellyfin-ffmpeg5. The new binary location is at /usr/lib/jellyfin-ffmpeg/ffmpeg and appending the dir to path or symlinking works fine.

If the team could fix this up it would be much appreciated <3


RE: Cannot start server after upgrade: FFmpeg validation: The process returned no result - void - 2025-03-06

As it turns out jellyfin failing to start seems to be the database migration from the EFCore Refactor according to journalctl, please don't symlink ffmpeg. Apologies


RE: Cannot start server after upgrade: FFmpeg validation: The process returned no result - mwl - 2025-03-06

In this specific instance, it turned out that the problem is that Windows Server Core is missing a couple of dependencies that FFmpeg relies on (which causes it to simply fail silently).  JF already ships with a custom build, so it may be possible to build without those dependencies, but it also might not, depending on whether they're needed for required features.  Either way, in my case I reimaged the VM again but with the Desktop Experience, and now everything works again.

In an unrelated gripe, this is now the third time I've had to move a VM from Core to Desktop Experience.  Both of the previous two were down to some configuration management not being available remotely and Server Core not shipping with MMC as a fallback for these.