disabled systemd service re-enabled after updates

[jimbothigpen]

Is it expected behavior for jellyfin updates on Debian, using the https://repo.jellyfin.org/debian repo, to re-enable the systemd service if it's been disabled by the user? Other .deb packaged software seems to respect the state of systemd services as set by the user, but jellyfin updates always re-enable the service and try to start it up if it's stopped. Might not be a huge issue for many people, but I'm trying to run a sort of makeshift HA setup, with 3 LXCs using shared storage for jellyfin data, relying on keepalived to start the service on the proper host. If I don't first kill keepalived and ensure jellyfin is stopped on all 3 hosts, updates have a habit of corrupting the sqlite databases (when the service is re-enabled and started after an update on a system that is not supposed to be running jellyfin at the time).

If the behavior of updates -- re-enabling and starting the disabled service -- can't be changed to take into account the state of the service as set by the user, does anyone know of some sort of apt acrobatics that would prevent it?

[TheDreadPirate]

I chatted with the team and the deb package is SUPPOSED to respect the current state of the service when upgrading. I am also interested in this since my main instance is in Docker, but I maintain an apt install for testing. As such, the service stays disabled and is only started for testing but both gets re-enabled and started after each upgrade.

I submitted a github ticket for jellyfin-packaging.

https://github.com/jellyfin/jellyfin-pac.../issues/50