Identify needs to be run twice to persist data

[gamatos]

Hello everyone,

i have this problems since i started using jellyfin.

I import a lot of foreign movies to jellyfin. Some of them are not picked up by the automatic identification process, no worries there.

For these movies that cannot be found i experience the following behaviour.

I manually identify them using imdb id.
The metadata is picked up, titles, actors the whole thing.I can see it in my libraries when i browse or when i enter a movie to see its details.

The moment i import another movie that cannot be automatically identified, the metadata from the previous manually identified movie is gone, its back to its original state. I do the whole manual identify process again for the previous movie i imported, and then the metadata is finally persisted for ever.

Any clue how to approach this? Thank you for your time

[Venson]

Hey @gamatos

Could you share a server logfile from the time when this happend?
Please upload to an external website like pastebin or alike and remove all personal information like DomainNames/Ips/Usernames.

[gamatos]

here you go https://pastebin.com/iaZ562Eu

/media/movies/Η μοίρα μιας γυναίκας (196 was added, i had to manually identify it, i then imported a few more movies that were automatically identified, then /media/movies/Η παραστρατημένη (1966)/1814_-_Η_ΠΑΡΑΣΤΡΑΤΗΜΕΝΗ_-_1966_-_TVRip_-_720x536.mp4 and /media/movies/Η νύφη τοσκασε (1962)/0330_-_Η_ΝΥΦΗ_ΤΟΣΚΑΣΕ_-_1962_-_DVDRip_-_640x480.mp4 had to also be manually identified.

At the end the aforementioned manually identified movies went back to being not identified

[Venson]

@gamatos

One thing i immediately see is that the filenames do not match the recommended structure.
Please rename your movie files according to this:
https://jellyfin.org/docs/general/server/media/movies

[gamatos]

I am going to try renaming the files before import, but does anything in my log files explain the behavior im experiencing? If all the filenames have the same format, why does this only happen in a subset of them?

[Venson]

@gamatos

Jellyfins discovery process is a bit complicated and prone to a bit ... moody.

I see a lot of loglines like this:


[2023-07-03 09:05:10.645 +00:00] [INF] [78] Emby.Server.Implementations.IO.LibraryMonitor: "Η παριζιάνα (1969)" ("/media/movies/Η παριζιάνα (1969)") will be refreshed.
[2023-07-03 09:05:11.178 +00:00] [INF] [78] Emby.Server.Implementations.Library.LibraryManager: Removing item, Type: "Movie", Name: "1814_-_Η_ΠΑΡΑΣΤΡΑΤΗΜΕΝΗ", Path: "/media/movies/Η παριζιάνα (1969)/1814_-_Η_ΠΑΡΑΣΤΡΑΤΗΜΕΝΗ_-_1966_-_TVRip_-_720x536.mp4", Id: 619b991e-71e7-1df7-0af0-9e70dfd3a4af


A refresh is requested by the FileSystemWatcher and then the file cannot be found anymore. Why that is the reason the log does not say. But an educated guess would be twofold:
- Issues with the underlying storage system. If that is a mounted device, it might be faulty
- Jellyfin is known to be a bit unreliable with non englisch character sets in several instances. This could also be an issue with your OS or dotnet.

Renaming thoese files in accordance with the documentation will bring some clarity.

[gamatos]

thank you for your explanation, i would like to believe that your second guess is right. I will try transliterating the names from greek to english to see what happens. I have imported 3500 english movies and never experienced a problem with them, only the greek ones

[Venson]

@gamatos yea greek is a first one for me but i had lots and lots of people come to the chat with Japanese's and Chinese charsets and there (can) be tons of issues with that. Still, renaming the files to the jellyfin recommended file structure should be your first to-go in my opinion.

For example:

/media/movies/Η ΠΑΡΑΣΤΡΑΤΗΜΕΝΗ (1966)/Η ΠΑΡΑΣΤΡΑΤΗΜΕΝΗ (1966) - 720x536.mp4


I know Jf does not like (sometimes) years to prefix movie names.