In the previous chapters, you saw how to build complex slave replications. However, in all those solutions, there was a Single Point of Failure (SPOF): the master node.

The solution to avoid having a single master is to build a dual master replication. It's common to have a simple dual master replication architecture. However, it's important to do things right when you're managing this kind of thing. Compared to other solutions, you can't afford to make mistakes and should be very careful with these kinds of tasks in production usage. To be as strict as possible, we're going to use cluster tools for some solutions; load balancer and features will only be available beginning with MariaDB 10.