Log Buffer #134: A Carnival of the Vanities for DBAs

Posted in: Technical Track

Welcome to the 134th edition of Log Buffer, the weekly review of database blogs. There’s no time to lose, so let’s begin—with MySQL.

The big news this week, the epoch-shattering event, is of course, Monty Widenius’s departure from Sun, his time to move on. Perhaps not much of a surprise, given his famously underwhelmed response to 5.1’s GA release.

But Monty, your thunder has been stolen! MySQL boss Marten Mickos is leaving Sun too. So says Matthew Aslett on 451 CAOS Theory. No official announcement, but this news is “confirmed” according to Matt’s source.

Maybe a conference is in order. On the MySQL Performance Blog, Baron Schwartz announces the Percona Performance Conference 2009, taking place in Sanata Clara, California on April 22 and 23, coincident with the MySQL Conference & Expo.

On the same blog, Peter Zaitsev shares the funniest bug ever.

Matt Yonkovit, possessor of a Big DBA Head, gives the first of a series of what he calls the five-minute DBA. Matt introduces this character thus: “Their seems to be a lot of folks out their who end up responsible for fixing or maintaining a MySQL database who are not really DBA’s. These guys and gals become DBA’s five minutes at time during the day generally when something goes wrong.  . . .  They tend to be looking for the quick fix, something that can be done in five minutes or less so they can get back to their other important tasks.” In this part, Matt offers answers to the question, “if you were on a deserted island and could only change 1 database parameter what would it be?”

You might use Domas Mituzas’s just-released mydumper, of which Domas says: “I had that idea that there’s no good tool to do logical dump of MySQL data for large sites – mysqldump doesn’t provide too much of I/O pressure, mk-parallel-dump is closer, but it doesn’t do consistent snapshots, uses same mysqldump, as well as is written in Perl (haha!), and I just wanted something new to hack, as proof of concept.” His tests suggest it performs faster than both mysqldump and maatkit’s dump facility.

Colin Charles reports that MySQL has been deemed a hot skill for 2009, by Gartner.

In that light, a SQL Server DBA might want to broaden his or her skillset.
Jason Massie offers a brief MySQL Cheatsheet for SQL DBAs.

Systems Engineering and RDBMS covers new views and functions for checking dependency order in SQL 2008, beginning, “Checking for dependencies and dependency order in versions prior to SQL Server 2008 was not very reliable.  . . .  In SQL Server 2008, the support is much more robust . . . ”

Along the same lines, Mike Weiner of the of the SQLCAT blog outlines what he believes are the top 10 SQL Server 2008 features for the database administrator.

Dan Guzman has a thorough report and analysis of what he calls the “UPSERT” Race Condition With MERGE, as introduced in SQL Server 2008.

SSIS isn’t the newest thing around, but as Andy Leonard writes, “The SSIS Expression Language is one of the steeper slopes on the SSIS learning curve. Many database professionals have never been exposed to the syntax.” If you’re one of those struggling your way up the curve, Andy’s Introduction to the SSIS Expression Language is for you.

Kevin Kline introduces Scalable SQL Server Grid with XKoto. Could this be MSSQL’s answer to Oracle RAC? he asks. The informed comment from the Kevin’s readers is very worthwhile.

Time for some Oracle stuff. Jason Arneil advises, be careful how you open your physical standby, for a little bug might fly out.

On The Quadro Blog, Alex Fatkulin speculates on edition-based redefinition, a feature mentioned in the documentation for 11gR1, but not yet available.

As another Alex—Gorbachev—writes, 11g might be in your future, in the wake of 10gR1 falling out of premier support.

As always, it is a lot to keep track of. H.Tonguç Yılmaz knows this, and he wonders about the purpose of Oracle’s “DBA 2.0” buzz-phrase. He writes, ” . . . Oracle DBAs lives are getting much more complicated day by day and what Oracle marketing stuff is forcing IT managers to believe is completely the opposite . . . ”

You want the Knuth? Can you handle the Knuth? Cary Millsap can. Now reading Donald Knuth might seem like a backward approach to making things simpler, but Cary found that Knuth’s formula, “premature optimization is the root of all evil,” may be helpful even to database programmers.

Chen Shapira responded with a Knuth-inspired item on the psychology of instrumentation, looking at how and why compilers, runtimes, and DBMSs continue to short-shrift their users on instrumentation.

Responses came to that in turn, from Cary (On the Usefulness of Software Instrumentation) and Tom Kyte (Instrumentation).

Back to Chen for a moment. She gives a sneak preview of her Hotsos presentation, in her article on the dining philosophers problem, which she uses as a way to think about concurrency issues.

From PostgreSQL blogs, two excellent articles. On The Scale-Out Blog, Robert Hodges illustrates simple HA with PostgreSQL point-in-time recovery.

And on the zillablog, Robert Treat discusses how best to handle controlled failover scenarios when using PITR warm standby.

Here’s a DB2 item, courtesy Craig Mullins: a DB2 Performance Monitoring Overview.

It sure is a lot to take in, isn’t it? Here’s a little something from Tom Kyte to lighten the load of every DBA, an SQL Joke: “A SQL query walks into a bar . . . ” It’s Tom’s punchline—click through!

That’s all for now. If I missed your favourite DB blogs from this week, please link to them in the comments. Till next time!

Want to talk with an expert? Schedule a call with our team to get the conversation started.

About the Author

Dave Edwards is the Communications Specialist for the Pythian Group.

No comments

Leave a Reply

Your email address will not be published. Required fields are marked *