cloning running linux machine through network, strategy and tactics

Joseph Hall perlhoser at gmail.com
Tue Aug 7 14:55:14 MDT 2012


On Tue, Aug 7, 2012 at 2:49 PM, Ryan Byrd <ryanbyrd at gmail.com> wrote:
> Linux Friends:
>
> Picture this: a "friend of mine" just set up a redhat linux machine running
> a website with apache/php/etc. It took a long time to get everything just
> right.
>
> Goal: He'd like so-called clone this live-running machine's files to a new
> freshly installed clean redhat linux machine so he doesn't have to repeat
> the effort. He wants two webservers all set up whereas he has only one
> presently.
>
> He's imagining something like this:
>
> [root at CurrentlyRunningServer ~]$ rsync -ave ssh --delete /* root at NewServer:/
> --exclude={/etc/sysconfig/*,/dev/*,/proc/*,/sys/*,/tmp/*,/run/*,/mnt/*,/media/*,/lost+found,/home/*/.gvfs}
>
> Then he'd just set up the new ip addresses on the new server, and WHAMMO!
> web server #2.
>
> Good idea? bad idea? something better?

Whoa. That's intense. But I have seen scripts that do roughly the same
thing, and so long as you get the right directory structure copied
over, it appears to work. Incidentally, you may want to exclude /boot/
as well (just in case). And even though I'm sure you're sticking with
the same kernel version, you should probably exercise caution with
/lib/modules/.

If you're going to be doing this a lot, you may want to think about
setting up a kickstart file to do the work for you.

-- 
"In order to create, you have to have the willingness, the desire to
be challenged, to be learning." -- Ferran Adria (speaking at Harvard,
2011)


More information about the PLUG mailing list