"...there is absolutely nothing to gain from writing to more than one of the servers at a time." Actually I think there is (and I have been using it for more than a year now,) if your application is designed for HA then you probably use some kind of MySQL proxy system or other app fail-over mechanism. Then you will be able to write to any cluster node. For example: If a node of the circular multi-master cluster fails then your app will continue to run -and write data! Human intervention -best in my opinion- or very complex automated stored procedure methods exist to recover -or at least reconnect a smaller circle of the full cluster. Meanwhile you app is still running on a subset of the cluster :). Caveat: You probably should use other multiple backup systems of your MySQL data. And please, your apps should be able to handle failure correctly. This also means that your app should not rely on the MySQL cluster for mission critical roles, or -shudder- life or property critical roles!