Summary

Hot Standby allows us to run queries against the type of standby databases used in early
versions only for high availability failover. The WAL used for creating standby databases logs block changes, and can only be used to make a complete copy of a database cluster, not a subset. Using log-shipping replication will result in some archiving lag between the standby and the master, but this is resolved in synchronous replication. Replicating database statements instead is more flexible and allows features such as version upgrades, but some statements don't replicate easily and this approach has higher overhead than write-ahead log shipping. Slony, Londiste, and Bucardo are mature solutions for replications. pgpool-II can be used to do load balancing across multiple read-only slaves, as well as some forms of synchronous replication. There are many additional replication...