Hi, is it possible to speed up full maintenance via parallelism?
When I run kopia maintenance run --full it seems to iterate through all the blobs sequentially to look for unreferenced content. While this happens I have tiny blips of network and cpu activity, I’d like to crank both up significantly. I’ve tried looking for parallelism flags for the maintenance command but I haven’t found any. I’d like to process many blobs in parallel which (if possible) should greatly speed up full maintenance.
I’m using an rclone backend on top of storj and I’ve figured out how to greatly speed up kopia snapshot fix invalid-files --commit by updating my repository config to add the following to my rclone config–with this I can max out my cpu and the command runs much faster:
...
"rcloneArgs": ["--checkers=16", "--transfers=128"],
"listParallelism": 256,
...