The SQL PASS Performance Virtual Chapter has put together an exceptional event this week called the Summer Performance Palooza, an encore to the highly attended Winter Palooza held last December. The Palooza will take place this Thursday, June 27, with an all-star lineup of SQL experts that promises to deliver a full day of high quality SQL Server Performance sessions. The day begins with three morning sessions if you live in the Americas, starting at 10am Central Time/3pm GMT. These sessions, while not beginner level, cover foundational topics in SQL Server such as indexing, statistics and bad performance habits that are required knowledge for any DBA. Jes, Erin and Aaron will show you practices and techniques that you can put to use today to achieve top database performance.
10am Jes Borland Index Methods You’re Not Using
11am Erin Stellato Updates to Statistics and the Effect on Query Performance
12pm Aaron Bertrand 10 Bad Habits that can Kill Performance
These sessions are followed by Wendy Pastrick showing us how to use our old friend Performance Monitor in a virtualized environment. This is important for so many DBAs dealing with virtual databases today, and knowing how to interpret PerfMon data is vital to pinpoint virtual DB performance problems.
1pm Wendy Pastrick Using Performance Monitor for your Virtual Servers
The afternoon delivers three sessions that show you how to approach, attack, and solve SQL Server performance issues. Janis, Brian and Brent will take you beyond the obvious and get you digging into areas of SQL Server you may not have known existed.
2pm Janis Griffin Looney Tuner? No, there IS a method to my madness!
3pm Brian Flynn Pivoting Performance Data & Pinpointing Problems
4pm Brent Ozar How the SQL Server Engine Thinks
Finally, Todd Kleinhas shows us how to quickly and easily build a SQL Server Lab using virtual machines to help us solve our performance issues. Make sure to catch this session if you are in need of spinning up environments to run through specific scenarios or to facilitate repeatable testing.
5pm Todd Kleinhas How to Build an Affordable SQL Server Lab
The Performance VC wants to thank our sponsor Confio whose support as our sponsor has helped us get this done for you. Click here to go to the Performance Virtual Chapter web site and register. See you Thursday!
A common method of developing applications today involves a lot of web developers and no data modelers. Even now when relational databases are a proven commodity, it’s hard to believe how messed up they get.
One of my clients has a vendor application that follows this pattern, essentially no database design for a high-volume application. They break all the rules and then some; no foreign key constraints, poor relations among tables, and they create tables on the fly, implementing their own strange partitioning scheme. As inevitably happens, it broke yesterday. Large amounts of locking and blocking ground the system to a halt.
I received a call, dug into it, and determined they were doing a full table scan on 46M rows more than one per second. An obvious index was missing. I built it, recompiled the stored procedure where the problem surfaced, and they were off and running. A familar problem for DBAs, but I have to say I’m surprised how often I still see this kind of thing happen, and it seems to be worse in the web-centric world of software development.
Dejected but unbowed, I enjoin the DBAs of the world who blog, speak, write and work in the trenches to continue educating, reviewing, and fixing this all too common error!