There is also a package called rdiffbackup which is
based on rsync, which can do a smarter job of incremental
backups.
http://www.nongnu.org/rdiff-backup/
Just a quick "me too" for rdiff-backup. I use it to "automagically" backup a number of servers, all over the country. It works _very well_, and it's ability to recover past versions of files is invaluable. Two "gotchas" I discovered that you may find helpful if you use it:
First, if you plan on doing daily backups and keeping older increments around, there seems to be a limit to how many increments it can correctly handle. The number seems to be somewhere in the low hundreds, but I've never bothered to figure out exactly what it is though. I have little need to save historical data, so my solution was to have it remove increments older than 90 days whenever it does a backup, and it works well for me. Theoretically, one could set it to do a weekly cleanup and then also a more long-term cleanup, maybe removing items that are a year or two old. In this paradigm, you could reach further into the past, though with less granularity.
And second, rdiff-backup responds poorly to overlapping backups. That is, if you have one backup running and then another backup for that same data starts (say, if you are doing daily backups and the first one takes more than 24 hours to complete) both processes dies and leave the increments in a pretty unusable state. The only way I've discovered to rememdy this situation is to delete the increments and start again. Therefore, until you have a good feel for how long your backups take, schedule them unusally far apart, or have your cron script make sure there are no other instances running before it executes.
--
-Regards-
-Quentin Hartman-
_______________________________________________ EUGLUG mailing list [email protected] http://www.euglug.org/mailman/listinfo/euglug
