Josh Berkus will describe the most important GUCS, or global settings, which most PostgreSQL administrators need to know to tune their database, log what's going on, and decide on special options. He will also go over which GUCS were added, changed, or introduced in 8.4 and what they are for.
Postgres relies heavily on statistics collected from data tables to execute queries in the most efficient way possible. Despite the relative simplicity of these statistics, the actual process by which they are collected and used to generate plans is a bit enigmatic. In the next 50 minutes, I will go through the assumptions that the planner makes about data access patterns, the heuristics that generate and apply the summary statistics, and a theorem from probability theory that is central to the planner's estimates. Finally, if time permits, I will show how a custom selectivity function can dramatically improve the planner accuracy for odd data distributions.
My hope is that a clear understanding of the process by which Postgres collects and applies statistics will remove some of the black magic from tuning Postgres and spur the development of better analysis techniques.