Hi Phil,
Are you currently using the vanilla "mysql" role or are you using the "mysql_cluster" role. If the former, you should just be able to add a new machine with the "mysql_slave" role. I'll admit it's been quite some time since I've done this, having really only run PostgreSQL in my last few production deployments. But, reading through the source that seems to be it.
If you already have a slave present, you'll need to update the generated rubber config to account for the new slave's server_id. If this is your first slave, the generated Rubber config should be sufficient (master gets server_id 1 and the slave gets server_id 2).
Bootstrapping the machine with the mysql_slave role should take care of configuring replication:
So, the steps would be (from memory -- I don't have a MySQL cluster to test against at the moment):
$ ALIAS=new_machine cap rubber:create
$ ALIAS=new_machine ROLES=mysql_slave cap rubber:roles:add
$ ALIAS=new_machine cap rubber:bootstrap
$ cap deploy
Please let me know if you have any other questions.