One element that is absolutely required for any highly available PostgreSQL installation is replication. It does not matter if we have a SAN that provides disk redundancy, nor is DRBD or other block-level replication sufficient to protect our investment. Duplicating and backing up data is always a good practice, but when it comes to availability, we need online copies of the database.

Similarly, if other departments need data that resides in our OLTP database, how can we provide it safely? In ideal circumstances, we can supply a copy of the necessary tables. This way, we don't strain the primary database with ad hoc report-based queries. A new DBA might try to accomplish this by building a synchronization library or performing scheduled extracts and copies into a remote database. However, there are easier ways!

PostgreSQL gives us methods to build and maintain a fully online copy of our primary database. Furthermore, there are existing utilities to duplicate tables when we don't need a copy of the whole database. In this chapter, we will utilize PostgreSQL replication as well as third-party table-synchronization tools. Building the best stack requires familiarity with the tools available.