Master->Slave/Master->Slave, something not considered in replication design

[Tony Ford]

Say I have 3 mysql servers, mysql-1, mysql-2, and mysql-3 (All running 4.0.23).

mysql-1 is a master, and mysql-2 is it's slave.

mysql-2 is a master, and mysql-3 is it's slave.

Seems by design if a write occurs on mysql-1 it is replicated to mysql-2, but, that query is not replicated to mysql-3.

I'm guessing that replication queries performed on a slave are not written to that slaves binary log, and therefore will not be written to a slave underneath that. I don't see any reason why these queries shouldn't be written to the binary log?

This is useful to me because I have 2 applications, mysql-1 and mysql-2. The two apps share some data, so I replicate certain tables between the two. Thing is, the app running off of mysql-2 is large, so I replicate to 2 more slaves under that, to distribute reads across more machines. So, if an update occurs on mysql-1 I need it to also execute on mysql-2 AND mysql-3.

Be great to hear something from the replication team at mysql ab. Thanks in advance.