Why does file operation in cache takes a long time?

So I have rclone setup for gdrive with cache unencrypted. Using cache-tmp-upload-path I’m uploading my files to gdrive-cache.
Say my files are in /home/aaa/media and cache-tmp-upload-path is /home/aaa/rclone-upload.
I would do:

mv /home/aaa/media/* /mnt/Gdrive-cache

This operation takes a long time to finish despite media and rclone-upload folder reside in the same hard drive.
Shouldn’t the logic be like : mv from a to b just like in linux and then do rclone stuff such as making a list of files to upload in rclone?
So shouldn’t command mv above finish almost instantly when rclone upload folder is in the same drive as source file?

No, it’s a copy as you using the fuse file system to make the “move” which in this case is a copy and delete. It doesn’t know it’s on the same file system.

Ok. So how about I ask different question. Why not have a way for user to mv from media to rclone-upload folder and then rclone takes care the rest?

That’s exactly what I had been doing via rclone move script or cron job in the much early version of rclone before it had any of much of the current feature. I guess what I’m trying to say is the current way of copying it first onto fuse filesystem just to end into local cache folder which most likely reside in RAM or in my case the same hdd might not be the most ideal way of doing it

You have a point there…

You could go back to using rclone move then use rclone rc to invalidate the upload directory. That won’t get the chunks into the cache though…