Recovering from partial data loss

[Artimus 1]

Yesterday I accidentally lost my main movie collection.  Stupid mistake on my part (and rsync with --delete on the wrong folders).

Before I realized what had happened,  I thought maybe an index went bad.  I took the opportunity to upgrade my server from 4.8.11 to 5.9.1.   It wasn't until after the upgrade that I realized what I had done.

Fortunately,  I do have a backup of the entirely volume!   I am restoring the content to its correct location, and emby is starting to discover it again.  However, It seems that it sees everything as "new".  There are several movies that emby incorrectly identified and I had corrected. These are are now showing up as the wrong movie again.   How can I correct this?

I should iterate that "only" the movie collection was lost.   The entirety of the emby server install, including db were unaffected.   I did also back that up separately while upgrading, but it was after the content went missing.   I have a another backup of the emby install from a previous upgarde but that was likely over a year ago.

 

 

[Luke 42077]

HI, what exactly did you do?

[Artimus 1]

Correction,  updated to 4.9.1.90.

I was moving content into the movie folder,  using an ssh terminal on my phone in a rush.  the environment variables were different in this setting.  Instead of mirroring a sub directory,  I mirrored the parent (deleting all content that was not in the source).

The backup I had of the Movies directory was fairly recent,  so I just copied it all back in, and fixed permissions.  The folder and file names are the same, but timestamps were not preserved.  It looks like emby removed each movie's entry on startup and recreated them as new.

All in all, not the end of the world.  I'll have to do a few hours of curating to fix incorrect titles.   That said, emby still seems to recognize the watched state of movies.  Custom collections were all empty, but I see they are starting to re-populate now.  Perhaps the scraping process is pulling in missing data that is needed to connect the dots.

 

I'll wait until everything is fully scraped before doing anything else.  it is still in progress.   I'm a bit overdue to rework this old setup.  NFS v3 vs v4  permissions has always been an issue, but thats a "me" problem, not emby.   Was hoping to hold off and see if emby would add support for and s3 library like mino or similar.

 

[Lessaj 467]

If you store your NFO files next to the media it would have retained the original import date since that would be stored in the NFO, even if it removed and readded it. Watch status is based on external IDs from the metadata providers for movie/TV libraries.

[Artimus 1]

Just a quick update.

After the library completed scraping,  everything is back to in order.    The only issues I had remaining where 1 movie that was misidentified (due to wrong year in the folder name).   And some custom collections.    I dont have enough instances to say for sure, but it seemed like movies that belonged to more than one collection now only belong to one.   no big deal.