borgbackup
Retry option for remote repositories
I got cloud storage for offsite backups and while doing the initial upload (which takes days) the provided experienced some network problems. I restarted the backup run about ten times. Would it be possible for borg to retry automatically if it sees that the ssh connection goes down?
Somewhat related to #1424 in that both require distinguishing between connection errors and other problems.
That would be great, I often experience transient ssh connection breaks due to unstable uplinks and need to manually restart multiple times as well.
This would be soooo useful for residential internet connections with daily force reconnects. I've seen no (literally not a single) backup solution that implements SFTP/SSH retries. The internet is such a volatile thing, I think this should be possible for such long running tasks as backups.
I'm willing to help with this.
borg 1.4 will have more detailed exit codes, so guess it might be possible to solve that already on the script level - see the latest beta/rc if you want to test that.
if the connection breaks down, this is not just some "retry": the ssh client / server process pair is gone, the remote IP might have changed.
so, just starting the same borg command again in the script if it detects a connection loss might be the easiest solution.
