I’ve been using restic + rclone for many years now to back up about 3 TB of files on my NAS to Backblaze B2. I was revisiting my backup scripts recently and was wondering whether it is better to run separate “restic backup” commands for each file path to back up on my NAS or whether to combine all of the paths into a single “restic backup” command.
From my experimentation, it seems there is no repository size difference either way, but restic does treat the snapshots as different. I get “no parent snapshot found” if I change the path list in the command line arguments and the first backup is slow. The size of the files added is actually very small so deduplication seems to be working the same either way.
Are there pros/cons to doing it one way or the other that I might be missing here, apart from the obvious that multiple commands take longer to run than a single command?
IMO it is purely cosmetic thing. Whatever is better for you is ok. First backup after changing path indeed takes longer as all files have to be rescanned. But it is one off thing. And deduplication works across all repo anyway. So nothing new beyond what really changes will be stored.
followed by prune, it treats snapshots with different path lists as separate snapshots for the purpose of maintaining telescopic schedules. That makes things a bit awkward to deal with if/when the path list changes if I run the backup as a single command. I will have to go manually forget older snapshots with the older path list.
@kapitainsky thank you for these pointers! This looks like exactly what I need. Not sure how I missed these options until now. I will change my backup scripts to use tags with the backup and forget commands.
If you don’t have a specific reason to split up the paths you back up into multiple backup commands, don’t. No need to solve a problem that doesn’t exist.