SOLVED: Media scan stuck at 96%

[TheDreadPirate]

Let's backup a bit. Can you describe your setup?

TrueNAS Scale? On what hardware? Is there any other virtualization? Like TrueNAS is running as a VM in Proxmox?

Jellyfin is running in a container on TrueNAS scale?

Is the media storage local to TrueNAS?

[Efficient_Good_5784]

Thanks for the reply. Container is run using user "apps" (ID 568 for both user and group).
Owner of the files on the file system is 568. To verify that permissions on the media files shouldn't be a problem, I set the permissions to 777 recursively.
Still the same result with that.

Looking at the screenshot, I'm assuming those are the media directories for Jellyfin.
Can you explain what the "u308946" group is and why is it paired with "apps" for the owner of each directory?

Also, how are you setting permissions to 777? With Truenas, you need to switch to root in the shell to apply settings like that. The default admin account is missing some privileges in the name of security.

Finally, as I stated before, make sure that all paths leading to the media libraries have at the very least read permissions. It doesn't matter if the directory you want Jellyfin to use is owned by Jellyfin if the Jellyfin user can't traverse through each directory leading up to it.

[manicraft1001]

TrueNAS is running directly bare metal. No virtualization. CPU is I3 12100, 64GB memory, xfs with 3 data disks, 1 cache and one log.

Media storage is local. The user group that you see on the files is from the user who downloaded the files from my external backup server when I switched to TrueNas.

admin has no permission to set file permissions, but I use sudo to set the permissions.

To ensure that Jellyfin has permissions, I logged in to the container shell, navigated to my library directory and created a file with touch. Using rm, I can also delete the file.

I'm glad for your help so far. Let's hope that we can figure it out together, as I really don't see what could be wrong. I want to reiterate that this setup has worked previously fine with the same permission setup. It's weird that it broke. I assume that something happened implicitly and broke it - although I don't remember any specific changes that could have done that.

[TheDreadPirate]

Is the jellyfin database located on a ZFS volume?

[manicraft1001]

Yes, both media and configuration is on the same volume.
Since there were Jellyscrub errors in the log, I deactivated the plugin and restarted just to make sure. Still running into the same freeze at 96%.

[manicraft1001]

Another update: I and my friend looked at the logs. Although I don't know the specifics of the scheduler in Jellyfin, I suspect that I have a broken database.
Jellyfin seems to be trying to access non existing files and directories. Remember: I moved them on the disk. If Jellyfin didn't pick this up correctly or only partially, this would explain the behaviour that we see?

There's also a message about the Scheduler - seems like it cannot finish something. Instead of allocating your time to fix these conflicts manually, I'd suggest to delete the deployment, delete all configuration and start from scratch. Sure, I'd love to fix the problem, but I hate wasting time if it's just easier with a different way.

[TheDreadPirate]

Do you have any collections that contained movies that you moved?

[manicraft1001]

What do you mean with collections? Yes, I did move movies. Not all of them. I grouped them into an Anime Movie group and "normal" Movie group.

Are you referring to a list of files (eg. to move them back?)

[TheDreadPirate]

Collections are like playlists used to group movies together. Like a "Harry Potter" collection or "Star Wars" collection. Collections use the absolute file path so if you had a collection that looked at the old path it could cause issues with scan completion. But I might be mis-remembering something about how that behaves.

Is the ZFS volume your database is on an SSD?

[manicraft1001]

I did have music playlist. No collections. The Vdev does not have SSDs in the data pool, but two SSD as cache and log.
Doubt that this should matter though for Jellyfin. Speeds should be sufficient.

I'll try out two things evening (19:00 CEST+1) and see what happens:
- Move back files to their previous location and recheck if media scan works again
- Delete deployment and redeploy in TrueNAS