	I still need to update the docs for the wombat 2.0 release - Keep an eye on 
http://www.alleft.com/projects/wombat for updates. In the most simplistic terms, 
installation involves:

1) Configure wombat.conf with the settings you require and the hosts that
    need to be backed up.

2) Make sure the hosts to be backed up have rsync installed and are accessible via
    ssh with key authentication (there are other ways, but this is the easiest)

3) Install wombat.pl into cron and set it up to run every hour or so.

	In addition, the backup server will need to have several perl modules 
installed:


use threads;
use threads::shared;
use Thread::Conveyor;
use Thread::Serialize;
use File::Rsync;
use Getopt::Long;
use Time::Local;
use Fcntl qw(:DEFAULT :flock);
use DBI;
use DBIx::Sequence;
use MLDBM(DB_File);

	If you're missing any of these, you should be able to install them with
something simple like: "perl -MCPAN -e 'install threads'", etc.

	If you're upgrading from the original release, there are some significant changes:

1) I've changed to using the DBI driver and requiring a database, so you'll need Postgres
	or MySQL. The first time you run the new version, wombat will automatically import
	your old job data. You can move /var/lib/wombat if you don't want it to do that.

2) wombat supports some limited command line options (more to come). As a result
        of this, running "wombat.pl" will just output the usage. You'll need to
        run "wombat.pl --run" to actually start a job. This is also how you can
        specify the path to a conf file: "wombat.pl --run /my/path/and/config.conf"

3) "wombat.pl --report [hours]" will show you a report of all jobs run in the last
        [hours]. This is one area where I'm curious to get feedback. I'm planning
        on adding more reporting capabilities, say reporting by host, target, last
        succeeded, last failed, etc. I recommend setting a cron job to run
        "wombat.pl --report 24" once a day.

4) New option: "synctimeout [seconds]" will forcibly stop a rsync after the number
        of seconds specified.

5) New option: rsync_opts { option value } will over-ride the rsync options. These
        are passed thru to File::Rsync, so anything it understands can be specified.
        For example:

  rsync_opts {
        timeout 1800    #Set a rsync I/O timeout of 30 minutes. This is different
                        # than synctimeout!
        exclude "CVS/"  #Exclude all CVS directories
        bwlimit 20      #Cap the bandwidth used to 20 KBytes/sec
  }

6) New option: mirror(list). Mirrors the target space to the (possibly remote) path.
   Mirroring works by breaking the list of writeo and mirror targets pairwise modulo
   the length of writeto. So:

   writeto(a,b)
   mirror(A,B,C) ->	a->A, b->B, a->C

   Remote hosts can be specified by just specfying the target in host:path
   or host::share notation

7) New option: mirror_opts for over-riding rsync defaults for mirror jobs

	That's just an overview - have a look at the Changelog if you want a detailed
breakdown. If you have any questions or want to be kept informed about updates, feel
free to join the mailing list at:

http://www.alleft.com/mailman/listinfo/wombat

	Or you can email me at:	petersp@alleft.com

Enjoy!
