Surviving Server Overload
Learn how to tune a PostgreSQL server the only way that matters: when the server is running terribly and you have no idea what's wrong. This is a unique chance to see the habits and tools used by someone who faces that problem constantly, watching as they make adjustments and tracking changes during a live demo, logged into a production server.
There are plenty of training sessions that talk about tuning performance on your PostgreSQL server. Speaker Greg Smith even wrote a whole book on that topic.
One day, you're going to find your PostgreSQL database server is unresponsive. The only question is whether you will be escaping afterwards with your job and reputation intact, and you're not going to have a leisurely period to study then. This session gives you a unique opportunity to prepare. Before it happens to you (again?), you can watch how an expert who does this exact type of support and consulting every day responds to them. Greg is going to login to a server that at the beginning is performing terribly. By the end, it will be running smoothly. You get to watch every step of the way, to see what change were made, and what information collected prompted them.
The material covered will introduce you to:
- Proactive monitoring you should be doing before your problems start
- Tools and techniques to locate the source of problems on a slow server
- PostgreSQL specific monitoring tools
- Good performance tuning methodology
Attendees should already have basic knowledge of how to use PostgreSQL. The troubleshooting session will happen on a Linux server, using a mix of command line and web tools. Most of the techniques will also apply to other platforms too. Slides summarizing changes made will be provided, so you can focus more on the demo rather than taking notes.