What’s the best way to clear a single item from the cache?
I can see an issue with something like:
Mar 27 16:21:49 gemini rclone[20638]: smu5ej34ujbdoip1cm3mlk92q4/61aitk3dc8289libd64fml2j1e7ff28r1kaoea517akip6h0t8qg/lal8gdvdl267lvqia6vt27ot0k8phu0ukp1ff2i37gvi4mgae0rhrfli66p1efs7hqpq8m3m2hdoa: error refreshing object in : in cache fs Google drive root 'media': object not found
That was when I was uploading some items via a rclone upload to a different spot and I’ve since moved to the tmp_upload method and just caching the writes.
I had a few items that I need to expire but didn’t want to nuke my whole cache. I started with the --rc but couldn’t find any docs on how to actually use it.
as I found the docs in the github for some more details.
@remus.bunduc - Is there a reason you can’t go GD->DECRYPT->CACHE? It would make the decrypted file names appear which would make life much easier in trying to track stuff back and figure out where the issues are. I saw there have been a lot of posts in terms of making sure it was GD->CACHE->DECRYPT
I did finally figure out the proper way to do it. I kept trying expire an item that wasn’t in the cache. I tried using vfs/forget to remove the item and it kept coming back.
Finally, I just expire the directory that the shows were in and that worked.
When I ran my plex analyze, I saw what files were stuck being analyzed and that helped me fix it. Much better than rebuilding the whole cache so the rc commands are very useful!
For anyone reading, cache/expire takes encrypted or decrypted path names happily.
That being said, as with you I too have log entries like that, often looping repeatedly for ages as an application tries to find something missing. It would be nice if that log message triggered a cache purge of that file/parent directory if needed until the error stopped?