i would be nice like rclone to be a kind of bridge for ssds/ram to the hard drive as incoming and recent folders.
in the form of a cache or after a time delay and everything about a moutpoint would be like unionfs or mergerfs.
but with better performance due to assigned cache.
example:
ssd 1tb as incoming /mnt/ssds
hdd 20tb as recent /mnt/hdd
mounted as /mnt/all
if the memory space would be full or a definable frame of for example 5gb (ssd area) which it then begins to move automatically (hdd area)
when i do it via unionfs or mergersfs the directory disappears briefly until it is completely moved. or it takes him a while to catch on
and it is visible again via /mnt/all
maybe rclone can even do it already and I missed something?
For me it is important that the ssd can unfold fully but that everything remains visible via /mnt/all that is simultaneously read and write mount.
especially in times of m.2 ssds with write rates of 2gb/s and higher.
I imagine I'm not the only one interested.
if it should be inappropriate please delete the entry thank you in the preface.
especially when there are many small files on the hdd which is mounted in the union. if i now have zb 4000 files/folders on the hdd the listing over the ftp as an example takes 6-10 seconds. because I always access my mounted drives via rclone (cloud, local etc) from outside. and the hard disk flushed the cache after a few minutes without access.
The cache backend will work for many small files I think. It isn't well maintained at the moment though so we are developing something new which isn't finished yet!