finally I got some time on hand for setting up proper backups. Since I heard about Restic (on a chaos radio talkshow) I was hyped! My Restic setup is running and so far and I understood handling keys, authentication and did some test backup/restores to get a grasp how things work. Now comes the part of planning the backup plan of where to put what.
Mostly I want to backup the important stuff (that can’t be replaced) of my main computer and some Linux servers I have.
Can you help me with setting/selecting the repos. Is it better to use just one repo for everything, or should I create a repo per client (?).
Another question: What is a good timeframe for the forget and --prune command? Once a week or once a Month?
The question of repos in my view depends mostly on the main advantage of using only one repo: deduplication. If you potentially have the same files across all machines, I would use one repo. Otherwise you can use different ones. It doesn’t really matter much in everyday life but having different ones probably makes you a little more flexible storage-wise.
When you do forget and prune doesn’t matter. By now it’s so fast that it can usually happen easily at night. But that obviously depends on the amount of data you have. I do it once a week and also I use a mirror site of the backup (rsynced) to continuously check --read-data-subset and see if all is good.
One thing to keep in mind though is that each host using the same restic repository, can read all data stored in the repository. If you cannot ensure that all hosts can trust each other, then separate repositories are better even though this won’t allow deduplication between hosts.
No. append-only only prevents the deletion of data, not reading it. Multiple hosts using an append-only repository still can read every bit of data in the repository, they just cannot delete anythin.