[sf-perl] [ot] pruning rsnapshot backups

David Alban extasia at extasia.org
Thu Mar 25 12:21:29 PDT 2010

hi mark,

thanks for your response.

until now, i've been finding the largest files in the latest snapshot
and, if they don't need to be backed up, i delete them from all trees
(except the original :-).  so it's not a matter of finding the
"largest" trees, it's just a matter of keeping a small subset of those
generated to this point.

anyway, thanks for confirming the independence of each tree!


p.s.  related amusing story.  i once tried to rsync an rsnapshot
"metatree", that is, a tree containing multiple snapshots to different
filesystem partition, but i forgot to use the --hard-links option (to
preserve them) and i filled the destination partition. :-)

On Thu, Mar 25, 2010 at 11:49 AM, Mark Grimes <mgrimes at cpan.org> wrote:
> Each tree is made up of files that may share inodes with other trees.
> Deleting one file will not delete its contents as long as there is
> another file that references that inode. You should be able to delete
> one tree without impacting others***.
> That said, since the trees sharing inodes, they typically don't take
> up all that much room. If you are just deleting blindly you might have
> to delete a whole bunch of trees to make an impact. I have a q&d perl
> script that I use to see which trees are really taking up space. Often
> you can spot something odd in just a few backups that are taking up a
> bunch of room. You can find the script at:
> http://www.peculier.com/dl/du-bank
> Hope it is helpful.
> *** I am not all that familiar with rsnapshot specifically, but I use
> the rsync --link-dest in my own backup scripts.

Live in a world of your own, but always welcome visitors.

More information about the SanFrancisco-pm mailing list