Hello, I recently migrated my local media storage to a cloud provider using S3 as the access method. I’m using CloudMounter to mount the S3 bucket as a local directory and created a Plex library based on the media in that directory.
I’m running into an issue which I believe relates to how Plex scans the library. As it scans the files, CloudMounter downloads each one into cache so that it eventually fills my disk. This is a CloudMounter issue, but I believe it’s been caused by Plex reading the file during the scan. Is there any way to modify library scanning behavior so it only looks at file-level attributes like name?
I could be off-base with the exact root cause here, but I can make it happen on demand and, based on my correspondence with CloudMounter support, it would seem to relate the file content being scanned.
I don’t think this is possible. Plex does need to analyze the files to know some basic stuff like what codecs are used and the size of the picture. There are options to tick off to get a little less intensive on the scan but that’s it. My experience with Google File Stream might enlighten you though on how to use Plex to scan mounted cloud drives.
Basically, you need to add the media slowly and skip every part that would make the scan too heavy like the video preview thumbnails.
I have 175tb of media on GD, all added to Plex Media Server on Windows so it does work, just the initial procedure is long and painful.
Thanks for the response! I’ve tried two tools now - CloudMounter and ExpanDrive 7 - with very different results. It appears CloudMounter does a GET of the entire file whenever it’s accessed, which was filling up my local drive with TBs of media files. ExpanDrive 7 appears to handle S3 file accesses much better. I’ve re-added my libraries as ExpanDrive-backed Finder volumes and it took only a couple hours to complete and used only ~150M of local cache.
