Welcome to the Event Horizon Mail Cluster Project

Right now all we have is a simple patch which makes it possible for postfix, the primary mta, to read from a MySQL database server farm.

A list of databases is supplied to postfix in the configuration files. For each query that postfix needs to make, a server is selected randomly from the list. This means that there is a equal number of queries going to each server. High availablity is also achieved. If that server is down then the next one on the list is tried. When the end of the list is reached, postfix will go to the beginning of the list and keep trying until all the servers on the list have been tried once.

By doing things this way, the load on the databases is spread between the servers.

The patch is on the sourceforge site: Here

Some of the advantages are: