Relationship between S3 operations and uplink

I see, you may reduce the rclone memory footprint to reduce a parallelism, but it likely will affect the speed.
Hm. You may try to use xargs command to run rclone in parallel, but with a reduced number of threads/parallelism for the each instance, because right now we do not have an option in the uplink CLI to check the destination, it will always upload, independently of does it changed or not…
Hm… It wouldn’t work too I guess, if the source is exact for the each instance…
I passed your question to the team.

The only thing which I thinks about is to use some backup tools instead, like Duplicacy, restic or HashBackup, they can do a hashed snapshots (only a difference) and reduce your storage costs (because of a bigger packs), this also should increase the speed for recovering.

1 Like