Using rclone mount or serve with duplicacy?

I’m using a cloud storage that supports only Swift OpenStack which I cannot get to work in duplicacy.

So I wonder if it is a good or a bad idea to use rclone serve or rclone mount to mount the cloud storage and have duplicacy run against that? If this would work, which would be the preferred way, mount as a drive or serve as sftp or webdav for example?

And again, if this would work, is there a way to have a pre-backup script that starts rclone serve before the backup and gracefully stops it once complete (on Windows)? It seems rclone serve for example always runs as a foreground process, so I’m guessing just starting this from the pre-backup script will just freeze the backup.

Grateful for an insight into this.

(see Invalid memory address or nil pointer during init of OpenStack Swift for my problems with using Swift directly with duplicacy)

Intuitively it feels that serve (over sftp) would be preferable. Adding virtual filesystem into the mix seems dangerous – another place where stuff can get cached and become of sync with destination.

That makes sense. Unfortunately as I tried to serve sftp via rclone, the backup operation failed every time at some chunk with an sftp-error of some kind, so this does not seem to be a stable solution (and it was also very slow). :frowning:

Perhaps I’m just going to have to wait for the potential bug regarding swift to be fixed (or to be informed of what I’m doing wrong with the connection string if that’s what’s causing it).