What if the original file is modified or deleted? #167
|
Thanks a lot for this. I haven't set it up yet but reading up on it. Apologies if I missed this already being covered. I have Tdarr running on my shows/movies so it's common for a file to be .mp4 and then become .mkv. Plex handles this with a refresh but curious what that would look like with PlexCache With FUSE, I assume Tdarr would see the file that's in the cache (via PlexCache) and would convert that one, Plex would see it modified and update itself, so all good there. But what about the original 'archived' file on the array? I assume it would still be the .mp4? I was planning to exclude .plexcached files fromm Tdarr but wouldn't necessarily have to. Though I suspect that would make it transcode both files which would be unnecessary. |
Replies: 3 comments 1 reply
|
I see no issue with this, plexcache will detect the file change after plex updates and when its time to move the file back from the cache drive it will move the file to array and delete the now outdated file. |
|
Thanks Brandon. So is this the expected workflow?
So for 9 based on my understanding of your comment, PlexCache doesn't care if the cached file and the original files are the same (or even same file extension)? It will delete whatever it has mapped as the "source" for the file it's moving out of the cache? |
|
You've got the FUSE side exactly right. Yes, the And yes, excluding |
You've got the FUSE side exactly right.
Yes, the
.plexcachedwill still be the old.mp4right after Tdarr runs. It stays stale until PlexCache touches it again. On the next eviction (when the file moves back to the array), PlexCache's restore logic matches.plexcachedbackups by media identity rather than exact filename. So it seesMovie.mp4.plexcachedas the backup-of-record for the newMovie.mkv, deletes the stale.mp4.plexcached, and writes the new.mkvto the array. It's the same code path that handles Sonarr/Radarr quality upgrades.And yes, excluding
*.plexcachedfrom Tdarr is the right call. Otherwise you'd transcode the slow-array backup for no benefit, and end up with a stale-tra…