I want to setup 2 MySql instances with failover.

I am thinking about doing this with both instances of the RDBMS pointing to a shared drive (a SAN).  One instance will be active, and one will be passive, by having one RDBMS instance started and one stopped.

If the active RDBMS failed we would then like to automatically stop that instance, and start the other RDBMS which would now point to the same mount point on the SAN, we should now have failover without the cost/ complication of replication.

The MySql RDBMS’s would run on HP-UX, and with HP ServiceGuard controlling the shutdown/ startup of MySql scripts in the event of failures. There would be a floating IP address that pointed to a subnet with both RDBMS’s on it, HP ServiceGuard would then ensure that whichever RDBMS was active was where the floating IP address was ultimately pointing to.

I would also welcome comments on the sanity of this strategy. Or alternative strategies using a SAN”

Your best bet might be to read the second edition of the book
“High Performance MySQL”, which is due out in June. ISBN is 0596101716.

RAID could give you synchronized copies on two (or more) drives.
Otherwise, if you need synchronization, you maybe need real-time replication..?

Why do you want to failover to an out-of date copy of your database?

Welcome to the ExpressionEngine forums, Michael. This is more of a HowTo question, so I’m going to move it there.