Cache backend very slow with a lot of "chunk retry storage" messages

Yeah, I just haven’t upgraded to 1.45 yet there but I will. Plex is same version on both servers.

Yes, and a different key for each server.

I understand usage is likely driven by Plex, I just was wondering if in windows the difference in mounting mechanism might impact usage. I’ve noticed this discrepancy over the versions and thought it might have gone away in Windows but looking at over the last few days usage that’s obviously not the case. But you’re right, as soon as I turn off my Plex server on the Windows box the usage goes away so it’s not the file system but Plex itself or the interaction between the two.

Perhaps its even a difference between Plex behaviour on Windows and Linux. I notice if the mount drops on Linux then it doesn’t go through marking everything that’s unavailable to be deleted (if sometihng is marked unavailable restoring the mount and restarting Plex marks the files as available again) whereas if this happens on Windows this doesn’t happen and it is still marked for deletion until those paths are scanned again.

Do you have this checked on one and not the other?

image

No, only thing checked for scans on both is - Run a partial scan when changes are detected

I’ve upgraded to latest version of WinFSP (prerequisite for rclone mount on Windows), upgraded rclone to 1.45, remounted and restarted Plex to see if this has any impact.

You should be able to check the Plex Logs and see what’s going on as well.