As I run this server on a MySQL powered solution (a Hibernate-based web-app running on JBoss4 on Linux), I’ve wondered occasionally what it would take to introduce fault tolerance at the database level in case of server failure.
OnLamp.com have an article by Giuseppe Maxia called “Advanced MySQL Replication Techniques” that shows what is involved to set up replication between multiple database server instances to introduce high availability of a MySQL database.