Borg repo on local server > migrate to external drive

[cmonty14]

Hello,
as of now I want to setup Borg using a repo on a locally mounted remote file system (WebDAV).
This is very clear and there are no issues.

Can I "migrate" this local server repo to a dedicated remote server running Borg in server mode at a later point in time?
If yes, how do I have to proceed then?

THX

[infectormp]

Reading the FAQ may help you
https://borgbackup.readthedocs.io/en/stable/faq.html#can-i-copy-or-synchronize-my-repo-to-another-location

[ThomasWaldmann]

The docs given by @infectormp rather talk about having an additional regularly rsynced "copy" of the repo and the potential issues with that.

The question from the top post is rather a one time migration.

I'ld suggest doing a testing phase for the webdav-mounted stuff. I personally have very limited trust in the quality, reliability and performance of such filesystems, so better try that first.

For later: the repo format used on a "borg server" (by the internal "borg serve" command) is the same as a local repo, so you just need to move the borg repo directory to the new place (e.g. using rsync).

You might need to adapt permissions (user/group/mode) on the repo files on the borg server, do the borg server setup and change the repo url you use on the clients.

After switching from the local repo to the ssh:// repo url, the borg client will notice the location change and ask you whether this is ok -> y.

[v12mike]

IMHO https://borgbackup.readthedocs.io/en/stable/faq.html#can-i-copy-or-synchronize-my-repo-to-another-location is incomplete, as it does not really cover the case where we want an off-line doomsday backup that can be restored if all of our machines currently running borg get nuked.

In our case we have a production machine interacting with users on the internet, borg runs on that machine creating regular backups to an on-machine repo. It also backs-up to second repo an off-site server and the backup logs are manually checked for sanity every day.

What we want is to archive the off-site repo to a tape drive, every few days and have the ability to build a new machine and restore our original server in the case that everything else goes wrong. There do not seem to be any clear, trustworthy instructions for archiving and restoring a disk image of a repo.

[infectormp]

@v12mike this one? https://borgbackup.readthedocs.io/en/stable/deployment/image-backup.html

[ThomasWaldmann]

@v12mike There is nothing unusual with backing up a borg repo:

• backup the repo directory with all the files
• make sure you create the backup from a CONSISTENT state (e.g. create a snapshot of the filesystem while no borg backup to that repo is running, e.g. using borg with-lock ..., and then back up the snapshotted files).