Clustering - handling MySQL server failures

[Dustin Cloues]

I'm working on setting up mysql clustering. We have two dedicated database servers, and we want to set them up so that our application can continue uninterrupted if one fails. Unfortunately, I'm not familiar enough with the MySQL clustering to decide what is the best solution. Our application can not use the NDB API directly; instead, it must connect to a MySQL server which should use the cluster for storage.

Both database servers will be storage nodes in the cluster. We have a router that can provide load balancing / failover, and I'm thinking that the best solution is to set up a MySQL instance on each db server (so each computer will be a storage node and an api client). Either one of the servers could be the management server.

Is there any better way to implement this? I've looked over every piece of documentation that I can find and it seems like it is, but since I don't have any experience with MySQL clustering, I don't know for sure. Any opinions or advice would be appreciated.

Thanks,
Dustin Cloues