How to freeze/lock repo when check --read-data fails?

Getting Help
1

I setup restic a couple months ago, and have been really happy with it. Recently I added autorestic, and went from cron to systemd to automate several steps.

One thing I hoped to do – I run various check --read-data-subset ops regularly, and email the results, but my instinct says that I can do better by freezing/locking the repo if a real error is found.

First question: is there value in freezing the repo – in other words, can new backups make anything worse? Or should I let backups, checks and forget/prunes carry on?

Second question: is there a known/good way to block/freeze? Notions I had before coming here:

  • wrap my scripts with my own lock/prereq, but that feels silly
  • || chmod 000 on the repo directory and/or index

All feedback & ideas welcome!

Cheers,
Ben

2

New backups don’t modify existing files. So they can’t damage the repository further, but depending on the kind of damage, the new snapshots won’t be restorable. Although, there’s a good chance that at least a large part of the snapshots are fine.

prune verifies the repository integrity before making any modifications, although it’s less thorough than check --read-data, but it nevertheless includes lots of checks to not make things worse.

1 Like