Preserve the sequence around startup
Record when the media storage became available, when Jellyfin started, and when the relevant cleanup task ran. A network share that is working now may still have been absent earlier. Use the operator’s actual logs or service records rather than inferring the earlier state from a successful film played later.
For an illustrative household server, Jellyfin starts before a network disk finishes mounting. The disk returns, and films become accessible again, but a playlist has already disappeared. Those two observations can coexist. They do not mean the video files themselves were erased or prove that a viewer deleted the playlist.
Identify the particular task
Jellyfin’s task documentation names Clean up collections and playlists and says its default includes running at startup. It warns that unavailable media during that operation can result in lost playlists. The guide separately lists library scanning and other maintenance tasks; keep their names distinct in the incident record.
Do not launch every maintenance action to see whether something changes. First find the task history and storage context. Clearing a cache, rescanning and cleaning collections are different operations with different purposes, so a broad sequence can make it harder to understand which one affected the library state.
Restore availability before repeating maintenance
Have the authorised operator verify that the intended storage paths are accessible to the server process, then review the startup and task configuration against that requirement. A path visible in a personal desktop session is useful evidence only if it corresponds to the server’s own access context.
Agree a supported way to ensure required storage is ready before the relevant cleanup occurs. Keep a record of the previous configuration and the expected startup sequence. This guide does not prescribe one service-manager dependency for every deployment or suggest permanently disabling unrelated maintenance.
Treat recovery and prevention separately
If the playlist is already gone, compare available server-state backups and the documented recovery procedure with the operator. Returning the media mount alone does not prove an earlier playlist definition has been restored. Avoid promising automatic recovery from a scan or overwriting current state without checking the saved version.
After the storage and startup arrangement is corrected, inspect the affected playlist and one permitted item after a planned restart. Keep that result separate from merely reaching the dashboard. A concise incident report includes storage readiness, task timing, state recovery and the verified outcome, without exposing private media paths.
Questions about this guide
Does media being accessible now prove it was accessible at startup?
No. Compare storage readiness with the actual task time.
Is collection cleanup the same task as library scanning?
No. Jellyfin lists them separately; identify the operation that actually ran.