Talk Information
A lot of little things all new in 9.1

11-01-2011 07:00 PM
Robert Bernier

One of the cool facets of PostgreSQL is there's always something new and cool to discover. But what may surprise you is that it isn't just the 'big' features that can pop your eyes. There are many jewels that one can overlook in that proverbial data beach of sand.

Starting with the most recent version of postgres 9.1, we're going to look at a number of concepts and features that takes only a couple of minutes to review but can potentially prove quite labour, and maybe even life, saving. Some of the techniques will be brand new, while other features have been there for a number of years. All them are quite cool for both new and seasoned DBAs.

The talk will present each feature in less than a couple of minutes followed by an equally short, but easy, example of using it:

Here's a concept list:

  • performance tuning
  • labout saving techniques
  • disaster recovery
  • system backups and upgrading
  • monitoring

Here's a very small list of what will be covered:

  • creating and using foreign tables
  • multiple collation columns in a single table
  • contrib modules
  • unlogged tables
  • common table expressions
  • leveraging the pg_* suite of utilties
  • pg_upgrade (eliminate dumping and restoring the data cluster between major pg releases)
  • tuning pg (editing postgresql.conf; lightning approach)
  • function declarations
    •   in/out parameters
    •   polymorphism
    •   overloading
  • newer and more powerful triggers
  • loading a server with a damaged cluster
  • pg_stat_activity and its variants
  • serialization and locks
  • using explain