I am currently using two backups with Restic, one for local external drive and one for remote B2. I was wondering how the copy feature would compare to this method. One advantage I see with my approach is I have two independent backups so if one fails, the other likely won’t suffer the same fate.
Is using copy more efficient than running a second backup job for an offsite copy?
I couldn’t answer your question but I follow the same strategy as you. Two backups with restic, one local to an external drive and the other to the cloud.
It seems better than doing a single simultaneous backup to two different repositories.
I have two independent backups, at different times, with independent reports as well.
Also you can check part of this thread:
The main benefit of two independent backups is that random bitflips during a backup likely only affect one copy. The copy command verifies data while copying a snapshot, but that cannot detect bitflips to occurred while creating the backup. That is, from a reliability perspective, two independent backups are better.
Regarding efficiency that’s more difficult to answer, instead of walking the filesystem and looking for change files, copy has t o process a snapshot and copy changed data. In many cases that doesn’t make a large difference, but it depends on the backup data set. It’s definitely faster to copy a backup of a large VM image than backing it up twice.
I make a copy to an external hard disk once a week (approx), then “check” the copy with the “–read-data” option to make sure all is well.
This external hard disk is not permanently connected so doing this ensures that it has all the intermediate snapshots that the main backup has.
It also avoids me having to run the forget+prune on the second backup; I only need to do that on the main backup.
I use the same strategy but once a week I use restic browser to download a random folder from the repository.