MySQL replication is quite different to replication on other DBMSes I’ve used (Sybase). It’s pretty easy to setup, it’s quick and solves many problems. However one thing it can’t do which frustrates me enormously is provide the ability to replicate the contents of multiple source databases into a single destination database. I’ve used this type of functionality before, mainly to allow independent nodes to replicate to a central reporting node.
However, MySQL has almost all the funtionality to do this already. It’s just built into the mysql daemon, mysqld, and not separate from it.
The functionality is there: connectivity to a master (the existing I/O thread), the storage of the replication changes and connectivity to the local db. It’s this last part which currently is part of the mysqld and is not able to talk to an external mysqld. If this were possible you could have a separate mysql …
[Read more]