Welcome to the 107th edition of the Log Buffer. My name is Keith Murphy and I am a MySQL database administrator for the Pythian Group. In addition, I am the editor of MySQL Magazine. This is my second go for the Log Buffer, so I must be doing something right!
This week for the open source world brings OSCON in Portland Oregon. There are plenty of MySQL people present and there have been more posting this week from these realms than normal. Also, Lewis Cunningham, among others, posted news that EnterpriseDB released the results of their open source survey at OSCON. The 451 CAOS Theory published their thoughts on the survey.
Of interest to all DBAs is John Duncan’s post about what is called the “five minute rule”. This was introduced in 1987 by Jim Grey. And finally, before we dive into the specific server news, here is a post on Facebook’s project to build a distributed database similar to Google’s BigTable.
Friday saw the release of the summer issue of MySQL Magazine. The highlights of the issue include the first annual MySQL Usage Survey. The magazine is available here. Peter, over at Percona, shows some initial benchmarks for the latest version (0.9.8) of Sphinx. If you aren’t familier with Sphinx, it is a full text search engine that easily integrates with MySQL.
Probably the biggest news this week was the announcement by Brian Aker of “Drizzle”. It is what amounts to a slimmed down version of MySQL server. These comments are from his initial post “Stored Procedures, Views, Triggers, Query Cache, and Prepared Statements are gone for now.” Interesting. My co-worker, Sheeri K. Cabral, posted a video of Brian Aker talking about the Drizzle project at this week’s OSCON. Monty Widenius writes a good summary how Drizzle can/might integrate with the MySQL “ecosphere” at large. It was very good to hear Monty say that Sun management is encouraging this project. There has been some other interesting posts about this as well including Arjen’s thoughts, and Paul McCullagh’s. Brian Moon gives a very thoughtful view on how Drizzle could potentially fit in at dealnews. While I could probably fill up the entire Log Buffer with links to post about Drizzle, I better leave it at that.
Ronald Bradford has an interesting post about “When (n) counts?” People often don’t understand that with integers, floating-point and fixed-precision numbers the (n) means different things. And the Data Charmer brings us an interesting test case of statement and row-based replication.
And finally, for those in the neighborhood, Amit writes about a MySQL Camp being organized for next week in Bangalor, India.
In the big news department for PostgreSQL, Josh Berkus is leaving Sun for other opportunities. It seems the PostgreSQL community now has several choices for certification as pointed out by Robert Treat.
Virag Sharma blogs about a new Oracle Users Group starting up in Hyderabad, India. Meanwhile, the Database Programmer provides insight into what history tables are and how to implement them efficiently. Ready to get deep into the internals? Jonathan Lewis has a good writeup on block sizes and how they can affect your application. For the dbas who work with Oracle RAC, Advait Deo has an intriguing article on Cache Fusion and how it has evolved over the years. For some good information on configuring the TNS listner for Oracle RAC take a look here for some help from Dan Norris. Freek D’Hooge wrote a program to gather and mangage the cbo statistics for your database. He has decided to release it out into the world. Take a look here. An interesting Oracle optimizer puzzle is posed by Karen Morton on her post Optimizer won’t choose Index Fast Full Scan?
SQL Server News:
Greg Low asks if parallel programming is coming to TSQL. Thought provoking and worth a read by any DBA on any platform. Want to know more about high availability SQL Server? Paul Randal points to a video from TechEd on that very topic!! And for a bit of humor, Jason Massie tells us “You know you are a SQL Wonk when…”. I will leave it to you to click through and find out when. For more SQL knowledge take a look here at the Patron Saint of Lost Yaks post on “Why LEN differs from DATALENGTH when using BINARY data”. While I wouldn’t ordinairly have two posts from the same person, I can’t pass up on this post by Jason on how to test for upgrading to SQL Server 2008. Good job Jason! And, finally, Decipher Information Systems has a fabulously detailed post on how to do a regular export of data using a SSIS package.
And that’s all until the next Log Buffer. Please email Dave Edwards if you’d like to edit and publish an edition of your own.