Better Disaster Recovery

Kevin Fries kfries at
Fri Sep 14 17:01:15 UTC 2007

I just sent a remote install process inquiry, but I have a second
semi-related idea I would like to begin working on and contribute.  If I
do this right (a rarity for me), these two ideas could be combined into

One of the holly grails of enterprise network management is data and
system recovery.  Backup processes backup data, and sometimes settings
(easier in Linux than Windows for sure).  However, I have yet to see a
good enterprise tool in any OS that allows quick, easy, automated system
and/or network recovery.  Here was what I would like to start on, and am
fishing for ideas on how this might work for Ubuntu:

Each machine backs up each night.  The backup consists of /home, /etc,
and a list of all packages installed (I believe apt stores this info
somewhere), and disk layouts.  All of this information should be able to
be fed into a restore process to automatically rebuild a machine in case
of disaster.

The server that handles all the disaster recovery, would then replicate
that data for all machines, including itself, to an offsite location.

Each machine should perform its backup based upon a local /etc setting
to allow it to optimize (desktops could update immediately, but a mail
server would get overloaded if it replicated every file as it was
written, so would take more of a snapshot methodology) for its own
needs.  But the offsite would be replicated based upon FAM.

The eventual goal is to say: "Our enterprise tools in Ubuntu are so
good, your entire building could burn down at 11am, and we can rebuild
your office infrastructure elsewhere inside 3 hours from the time the
replacement computers showed up".  Let Microsoft match that one, lol.

I think Ubuntu has the tools, I want to figure out how to tie them
together in a way to make this happen.

Any ideas, or starting points?

Kevin Fries

More information about the Ubuntu-devel-discuss mailing list