HA / Load Balancing / 2 Servers.

Here's a hypothetical I'll throw up in there for you.

1. Run a bash/perl script which constantly pings the active/primary box.
2. If the ping fails, issue a remote reboot to the active/primary box and start ipaliases on the standby box.

Thus you now have a fool proof Failover system for when hardware issues occur, preventing prolonged downtime. You may want to implement a service monitoring tool rather than a basic ping monitoring but this is just a thought up in the air for you to go on.

* Filesystems need to be synced constantly (drbd/netfs/smbd).
* Ofcourse you don't want ipaliases starting by itself on boot up.
You want ipaliases initialized by the perl/bash script.