Perks and Pitfalls of Postgres with Pgpool

Postgres is a great database, but you can only squeeze so much out of a single server.  Pgpool (really, Pgpool-II) offers a lot of features to help scale Postgres past a single server including connection pooling, load balancing reads, and several models for high availability and replication.  This talk will focus on using it with Postgres's streaming replication --introducing Flail Replication as an optimal model in this context-- but will cover some of the other features for comparison.

As useful as Pgpool can be, it also has its idiosyncracies that give new users opportunities for mistakes.  The audience will hear about some of Pgpool's drawbacks and failure scenarios with suggestions for dealing with them.  This includes configuration notes, experiences with monitoring and addressing imperfect behavior.