Repository navigation
Implement rdiff-backup support #59
Description
Activity
I'ld guess for a simple implementation that would be also the way to go with
borg-importimporting such backups.With a lot of effort one could theoretically maybe avoid that full extractions get written to disk, but I guess that would be too much work for just single-time use (per user). It's not just the file names and content, we also would need to care for all the other metadata.
In case users happen to stumble over this issue, here is a rather quick & dirty script that allows one to import incremental
rdiff-backupbackups into Borg: https://gist.github.com/PhrozenByte/78ae546db81cbc09f71b298159c6d66dIt also uses the approach of first extracting everything and then using that to create a Borg archive. I used this script 2,5 years ago to import some old backups, it worked just fine back then, but again: It's a quick & dirty hack and was written for my specific use case. I highly recommend creating a backup of the
rdiff-backuprepo (yes, a backup of backups…). If it doesn't fit one's needs, it might at least be a starting point for one's own script. Feel free to share your script(s), too.
https://rdiff-backup.net/ is another common backup tool. Implementing the importer might be a bit easier due to the fact that it's also written in python.
I've written a shell script that does it, however it's quite cumbersome as it requires the full extraction of a snapshot for borg to back it up, and rdiff-backup isn't very smart on restore (it restores a file even if it's unchanged).