[rdiff-backup-users] An idea

[Dave Kelly]

Hello everyone.  First of all, the new Wiki is great.  I think it will be a
valuable resource.  Thank you!

Second, I have an idea I've been toying with, and I want to run it up the
flagpole and see who salutes.

I currently have a backup box that basically "runs the show" as far as
backups go.  By that I mean that none of the machines it backs up really
have any say about what gets backed up and when.  It is all run out of the
"mothership".  I am backing up 2 Windows servers, 3 Novell volumes, and a
Linux box with this (how's that for variety?!).

My idea consists of writing some sort of config file where I can have
standard config items, like a servername, server IP, username, password,
mount point (for the novell & windows shares), etc.  Maybe one config file
per server.  Then an outside python script that parses all the config files,
and runs the backups.  This would make it much easier to add and remove
servers, because all you'd have to do is add a config for the server, and
the script would pick it up.

I've got pieces of this going on right now...mostly a directory structure
for holding everything.  Here is what I'm using:

/backup
   /conf
      global.conf
      /server1
          include
          exclude
          backup.conf
      /serverX
          include
          exclude
          backup.conf
   /bin
      (backup script goes here)
   /mnt
      /server1
      /serverX
   /repository
      /server1
         (Here's where the backups are made to)
      /serverX

For the windows and novell servers, my script mounts them under the
appropriate folder under /mnt, and proceeds with the backup as normal.

Thoughts on this?

In conjunction with this, I'm hacking together (and hacking is definitely
the best description here) a Webmin module (http://www.webmin.com/) that
would allow me to add and remove servers via editing the conf files.  Right
now, all it does is lists all the servers that I have configured under the
conf directory, but it shows the last time of backup, what type of server it
is (according to my conf file), and I can view the summary files for any
backup I've done.  I'm almost to a point where I can download a file, so I'm
going to work on an interface where you can browse the list of backed up
files, choose an archive date, and it'll restore that file to a temp file or
something and download it to your local drive for restores.

Basically, my ultimate goal would be to have a graphical interface to
rdiff-backup, eventually with permissions handling so foolish users could
restore that Word document they just deleted without having to call me.  :)

Looking for thoughts and comments on all of this.  Thanks!

-D