rsnapshot: rsync-Based Filesystem Snapshot

Snapshot filesystems, which offer instant-availability backups from multiple
time points, are a great idea. But migrating to a new filesystem is a major
deal. rsnapshot works on an
ext3 filesystem, using rsync and hard links to create
something that looks and works like multiple instant-availability full backups
(as with a snapshot system), but that takes up the space of only one full
backup plus incremental changes.

That “full backup space” is of course still the gotcha — basically,
you’ll need a spare disk. However, external disks are increasingly cheap; and
rsnapshot can be used remotely with ssh. So you could use
one large central disk to keep snapshots of your most important data from
multiple clients.

rsnapshot is relatively easy to configure. Check
/etc/rsnapshot.conf for the minor edits needed (make sure to
uncomment the “interval” lines), then try a test run from the command line
with rsnapshot hourly. The interval names don’t actually matter;
they’re just used to define how many snapshots of each type are kept. The
actual run intervals are defined using the crontab — a reasonable
crontab example is provided.

The first run may take a while (as always with a first run of
rsync). Subsequent runs will be much faster as the files are
updated from the snapshot. Usefully, it’s possible to set it up so any
user can pull his or her own backups, rather than relying on root access — thus
minimizing your work later on!

rsnapshot should run on nearly any Unix-type system (requiring
only perl and rsync), and is available as a package for most
of the major Linuxes.

This article was first published on ServerWatch.com.