I have a restic repository located on Computer A that is not accessible from the internet and the backup is too big for an external drive. I want to run restic backup on Computer B and back up some files directly into the repository stored on Computer A. I don’t want to just copy the files from B to A.
Is there a recommended way to do this securely and safely? Can I copy the repository metadata directories (everything except the data/ folder) to Computer B, run restic backup there, then merge the new data back into Computer A by merging the data/ folder and copying other dirs back. Will that work? Is it safe or likely to corrupt the repo?
Is there a better or safer way?
Can Computer B reach Computer A over a network?
Unfortunately I don’t have a way to connect them via network.
No.
If you have no network connectivity between A and B at all, you need to find a way to mirror the files on A to B using some offline sync/transfer solution. Then backup those files locally on B.
Consider this a “refinement” of the “No” in my previous reply.
Well… Maybe it can be done, but if so there are a lot of caveats. I would never attempt such a thing.
Why no?
Restic is supposed to create only the new packs, and I update the repository metadata, why it won’t works?
Is there another solution to backup the files from computer B while preserving the metadata?
I doubt there is any “recommended“ way taking into account your setup limitations.
But your idea might work. Maybe:) Simply try. Then check all repo and try restore to make sure that things work. Worst case you lose your repository. Plan for such scenario.
If it was my data and something I care about I would rather buy bigger external disk to make sure that all repo can be copied there. Extra bonux would be that I would have two copies of my repository. On computer A and external disk.
How would you copy the data between the computers?
Actually I already tried it and it kind of work.
I had a missing pack, but managed to restore the new snapshot.
But I am afraid that it was more luck then a represent case.
The data that contians only new blobs can fit in external drive. The full data directory can’t
Other option would be to initiate a new repo on an external drive with the same chunker polynomial as your repo A.
Then you can copy snapshots from this repo to repo A without losing deduplication benefits.
Still IMO buying bigger external drive is the best and most practical option.
1 Like
It sounds like you have an interesting use case and there may be ways to accomplish it with your current setup, but I’m afraid that none of them will be reliable.
Actually I didn’t think about the copy snapshots, but it can work, thanks!
Just for technical curiosity, why it’s won’t work? The data directory should be imutable so jt will only add files
Any reason you can’t network them?
I also have a PC that I want to backup. It is on the network. I had been assuming that each PC needed its’ own repository on the server. Then I was looking at a listing of snapshots and noticed the HOST column and a light came on. I’ve seen nothing in the manual about it.
My question: Can the repository be used by more than one PC?
@Clarence , please read the docs. Then, if something is still not clear, create your own topic/thread. Kind Regards
Experiment first!
I did a search for “HOST” and found nothing that told me what I wanted to know, but I should have tried it to see what would happen. I sometimes learn more that way.