Despite my lackluster performance at a talk last month (about Emacs, what else?) the Boulder LUG let me attend another meeting. Thanks!
I was fascinated by this talk — I know little about databases and enjoy hearing about them. Although it ranged strangely between the very generic (high-level comparison of MySQL and PostgreSQL) and the incredibly specific (rundown of command line options to psql), Kevin’s depth of knowledge and cool stories made this well worth attending.
Postgres is a good product. My impression, overall, is that while it is missing some gloss that Oracle has, it is basically “enterprise-ready”. Kevin has a project now that gets one terabyte of new data per day, and Postgres handles that fine. Refreshingly, Kevin did not spend much time defending the choice of an open source product over a proprietary one. I guess at a LUG that would seem unnecessary — I thought so, too, until somebody asked why bother.
A database is a big programming environment in its own right. One take away for me was: learn something before starting. Kevin said he’s run across projects that did not vacuum regularly and were running with a lot of bloat — basic stuff. Based on his talk and the questions, some of which I did not understand at all, I would say that if I were running a company I would just hire a DBA and move on.
Kevin is a fan of Slony-I for replication. This came up over and over through the talk, as did table partitioning (which I had never heard of before).