spare eSata Sheevaplug and a spare 1TB drive
Don't leave them lying around in a drawer, power them up at least once a month, otherwise when the crisis strikes and you need to rely on them....
a hot-standby approach which two eSata Sheevaplugs
I suggest you take a look at 'High Availability Clustering' for a few ideas, you would need to scale it down to a 'miniature' version for the Sheevas, but the concepts (and the software you would use) are still sound. The Sheeva supports watchdog, make use of it.
Test, test, test....and when you are sure you have got everything right, pull the plug on the master and observe! don't wait for the (real thing) drama to turn into a crisis
how to keep track of changes of programs and parameters
'rsync' will work this out and transfer for you (no need to copy whole images, although initially cloning the master system is a very good idea, think boot sectors, 'blkid' and everything that entails), and could easily run from cron.
You will need to take care when backing up the master when it is running, or the clone will be missing the cached program data from the master, but will get other live data, lock files, etc, of a running system for processes which do not exist that can cause problems (when you do boot it). I am unfamiliar with Freeswitch so cannot be specific, sorry.
My personal preference at home is to frequently rsync all the nodes onto my workstation via the network (big disk on workstation, small file-systems on the nodes, shell scripts save all the typing). Once a month'ish I connect an external big disk to my workstation and rsync the whole internal disk across, and then stash the drive in a safe place.
If you are not comfortable using the command line (I expect you are), 'unison' is a friendly front end to 'rsync'.
