[Dirvish] RSYNC backup brain-fart

foner-dirvish at media.mit.edu foner-dirvish at media.mit.edu
Thu Jun 8 00:30:49 UTC 2006


    Date: Mon, 5 Jun 2006 10:13:48 -0700 
    From: "Petcher, Daniel" <danielp at audioprecision.com>

    Somewhere in the Dirvish realm of discussion, someone had mentioned another
    rsync-based backup project that reduces redundancy by storing single copies
    of redundant blocks, using SHA1 checksums. I can't remember the name of this
    project.

    I like Dirvish's ease-of-restore, but I'm currently implementing the
    http://unattended.sourceforge.net <http://unattended.sourceforge.net>
    project for building an unattended OS installer. I've got four different
    languages, two different Windows operating systems, and several different
    standard hardware driver platforms, so MOST of the data within my server's
    10 GB file-share is duplicated (or even quadruplicated, if that's a word.)
    I'd like to run a daily backup with this other product to reduce redundancy
    and then catch its boiled-down file with Dirvish for my weekly offsite
    backup.

    Yes, I HAVE tried googling for this, but I got really bad signal-to-noise
    ratio from my search parameters.

Try faster-dupemerge.  Short but really effective perl script.  You
can find mail I sent about it in the archives to this list, though
searching should turn up the original site.  Works great for me.



More information about the Dirvish mailing list