This week’s water falling from the skies isn’t turning into snow. And on days when it’s clear, my car is covered with tree pollen. While they means something different to most people, to me these are the signs that the spring conference season is about to start. There’s a conference in North America during each of the next three months containing serious PostgreSQL content, and this year I’m making each of them.
PgEast 2011 takes place in New York City starting just over two weeks from now. If you’re nearby, it’s not too late to make travel plans. The location is so easy to reach via train. The conference venue is right on top of New York’s Penn Station, a major train hub. All the major East Coast cities from Boston to Richmond are easy rides from Penn. The conference is also next to a PATH train stop, with trains that lead to and from north New Jersey. Those easily connect up with the whole New Jersey Transit system.
This year’s conference is huge, with 5 to 6 parallel PostgreSQL sessions during many time slots. There’s even a MongoDB track on top of that. This conference also has a large number of half and full day training sesions going on; these are paid for separately from the main conference. If you work with PostgreSQL, I would expect at least one of these sessions is worth arriving early for.
I decided to do something really different from the usual training format. Rather than going through a normal slide series, my Surviving Server Overload training session is going to spend most of its time staring at the console or monitoring tools of an overloaded PostgreSQL server. The idea is that you’ll get to see what I look for when trying to find performance problems, and learn a bit about the methodology I use to resolve them. Doing this sort of tuning work is essential to many PostgreSQL systems once their usage increases. And it’s normally learned only way one: stressfully, while trying to figure out why the server went down under load. My class lets you see what that will look like before it happens, and leave you with some techniques and ideas to adopt that can postpone a performance outage from happening to you–and leave you better prepared when one does happen.
The way the schedule has worked out, you could make a full day of training by seeing my talk in the morning, then either the Replication training from Magnus or the PostgreSQL 9 session from Robert Treat in the afternoon. There are a few combos that would fit together well possible here. There are also some alternatives that take up the whole day in one session, such as Bruce’s administration primer.
Tough choices. If any of these training sessions interest you, may sure to book them as soon as possible. Class sizes are pretty limited and several sessions are expected to sell out.
In April, there is a whole PostgreSQL track at the increasingly misnamed O’Reilly MySQL Conference this April in Santa Clara, CA. I’m doing a talk there on hardware benchmarking that closes with some notes about MySQL vs. PostgreSQL performance testing. Many other popular PostgreSQL speakers will be there too; it’s a “keep your enemies closer” sort of thing, us mixing in with our mortal enemies at Oracle.
And the last of the spring conferences is PGCon in Ottawa, Canada. The schedule for this conference just went up, and I’m doing another half-day session about performance tuning (this one a more traditional slide based talk) on the first day of the conference. The tutorial days at PGCon are often a bit under-attended relative to the main conference. With Robert Haas doing a unique talk about database introspection as the other half of the first day, this year you definitely need to arrive early to PGCon if you want to see all of the good talks being presented.