Chapter 14. Scaling with Replication

Sometimes when trying to improve database performance, the most practical approach is to add more copies of the data and spread out the load over them all. There are a variety of PostgreSQL projects that replicate data across multiple nodes that you might be able to use for that purpose. Also, high performance systems tend to come with high availability requirements, with their respective overhead, too. Learning how to co-exist with your replication software may impact your performance tuning work, even if the copies are not being used actively for load distribution.

Note that none of the mature tools mentioned in this chapter support scaling up for high write volume usefully. The best explored solution in that category is using pl/Proxy, as described in the partitioning chapter. Write volume scaling is also a goal of Postgres-XC. In general, when we're talking about scaling a database upwards by adding nodes, this is allowing more...