My source S3 repository is set up with object retention enabled.
I found out that if I run full maintenance with --extend-object-locks on my source repository, and then execute sync-to command to synchronize it to an external S3 replica, object locks at the destination storage are not getting extended. I tried to use --times=true flag, but it didn’t help.
I guess, Kopia doesn’t support it now.
But is it OK to run full maintenance on a regular basis on the replica repository to achieve locks extension? Wouldn’t there be any problems in future with recovery/synchronization/verification from such repo?
Is replica repository a carbon copy of the source one, or there are some differences between them?
It is not something supported out of the box, neither documented. So you are on your own here:) For sure with some level of manual fiddling you can manage it but not sure why to complicate life so much.
Why to run sync at all? It would be much easier and safer IMO simply to run backup to two different repositories.
Well, I have at least a dozen of machines (and more to come) I’d like to backup, and some of them are connected to an internal network, completely isolated from internet for security reasons. Also, setting up a second backup job adds complexity (especially when when VSS/BTRFS/ZFS snapshotting is involved) and increases load on protected machines.
So, it’s doable, but to me it would definitely be easier and more efficient to just add an another cron job with kopia maintenance run --full for the replica repository.