I have a question about the backup strategy. I have a folder with 300GB of pictures and a folder with 50GB of documents, that I back up via rclone to onedrive. Do I put both folders in one restic repository or do I make two backup restic repositories for whatever reason? And that’s exactly the reason I can’t find objectively right now. I can’t find a reason why and I can’t find a reason why not. Do you have any ideas or advantages and disadvantages for one or the other method? I know that restic can do both, but I’m more interested in the strategy.
The main thing you get within one repository is the deduplication of data. If you have 300 GB of binary picture data this is very different from your 50 GB of documents, so putting them in the same repository for deduplication purposes is basically a moot point.
On the other hand, why would you not put them in the same repository? Everything backed up on one client generally goes into one repository. One reason to split backups into multiple repositories would be e.g. if you have multiple users/clients backing up data and don’t want them to be able to access each others’ data. If both clients back up to the same repository, they will have access to each others’ data even if they use separate keys for accessing the repository.
In the example you mentioned, I would put it all in one repository (and of course also back up to a second repository for purposes of having two backups of my data, but that’s a different matter).
Another reason to split them up would be if the two data sets had different “importance”.
I backup my financial+medical information separately from pictures etc. Not only do I use a different passphrase for that repo, but I also copy it to a few more places (cloud, external HDD, whatever) for even more redundancy than for the other one.
It also helps that my fin+med data is much (much!) smaller than my pictures so I can keep more copies of it, e.g., even one in my gdrive, with its 15 GB limit. (If you’ve heard of “3-2-1 backup”, I go “5-3-2” or something for my fin+med data!)