A recent project of mine entailed replicating a RETS database on a local MySQL database server. The client had a new real-estate mobile search app in development and wanted to have a local copy of the RETS database for search queries instead of a remote RETS server. [...] Replicating RETS data rather than using the live version is slightly complex, however. In order to replicate the RETS data into your own local database, a series of processes are needed in order to make sure the data you have is both updated and in sync with the server.

He breaks it down into three main steps with some sample code for each:

• Grabbing the complete database once
• Keeping the data in sync
• Ensuring the cron job is executed on regular intervals

Recently I was asked if PECL/mysqlnd_ms should be used to add MySQL replication support to a yet to be developed PHP application. The mysqlnd plugin, which supports all PHP MySQL extensions (PDO, mysqli, mysql), stood up against a classical, simple, proven and fast approach: one connection for reads, one connection for writes. Let's compare. This is a bit of an unfair challenge, because PECL/mysqlnd_ms was designed as a drop-in for existing applications, not optimized for those starting from scratch, *yell*... The plugin stands up quite well, anyway!

He starts with a look at the "classical pattern" of using a factory or singleton to make a database object instance that gives back different connections for reads versus writes (slave vs master). The mysqlnd_ms plugin allows you to define configuration settings to tell the queries to automatically go to certain places for different actions. For example, you could use "master_on_write" to tell it to use a master node if you're doing an INSERT or UPDATE versus a SELECT. He also shows a more complex example using a SQL hint and one issue that might come from the "human element" - not paying attention to database character sets.

In 2008, Paul M. Jones announced the release of an SQL adapter that allows Solar users to connect to master/slave database installations. My first reaction was: Great! This will inspire other FOSS developers to create similar components. And guess what, it did. I wrote my own ReplicationAdapter. It's not great, but it's flexible enough to support the most commonly used replication scenarios:

He gives examples of how it can be used for single-master replication, multi-master replication and how the adapter can be used to create a distributed caching system for database connections. A brief Zend Framework-based example is also included.

Now, most people will say you should never do this. And lots of time they are right. And once upon a time I agreed with them without question. Then I started living in the real world where sometimes you have to do things you never thought you would. Here are the two places where I stores files in a database.

His places are in Phorum, using the method to keep the system's database and file attachments contained and easier to move around, and in the replication of their databases, making it simple to have images for the deals on dealnews transferable to the various slave databases.