Typically, redundancy is achieved via database replication and some file replication. Then, you point your main DNS record to the primary server and set up some way of failing over to the other server in case the first one doesn't respond. After that, you need to decide if you are failing over...