Maybe adding some more description about what's going on here.
That is from what i understood.

Try and restore some contents by asking nicely to backup storages (in the same order).
If the first one does not know the contents, the second one will be asked about it, and so on and so forth.
If some content is not repaired, we simply log the error.

Do we want to ask the backup always in the same order?

Maybe, randomize this to avoid asking always the same backup storage.

It's relevant only if for example, this is a cron which starts this regularly (or something similar).
Otherwise, the randomization can be done more deeply in the code (cf. previous mark).

Due to the very large amount of content to verify, the checker will very likely run continuously. Given that, I suppose it's a better idea to have it as a daemon iterating again and again over the content.