DBA Cartoons Don Burleson

Here are some Oracle DBA cartoons that you might enjoy:

IT managers are motivated by minimizing costs, reducing risk, and avoiding downtime.  Redesigning an Oracle database is risky, time-consuming and expensive, and sometimes managers decide to move a poorly-designed Oracle database to faster servers (more CPU and RAM) or faster disks (solid state disk (SSD).

Because the Oracle STATSPACK and AWR reports where designed for use by Oracle Corporate technical support, and not by Oracle customers, some of the AWR and STATSPACK report metrics remain undocumented and very mysterious.  For example, the AWR report has "JOX SGA heap latch", a database internals measurement that is somewhat undocumented.

As Oracle Predictive modeling allows for future forecasting of behavior, many Oracle DBA's take-on a role like the nightly TV weatherman.  And, like the weatherman, they base their predictions on sound statistical models but they dwell in the world of probabilities, not absolute facts.

Mom!  The DBA say that I don't have to clean my room! 
It'll just get messy again, and I'll have to do it over-and-over.

There are two schools of thought about periodic Oracle maintenance.  One claims that Oracle tables and indexes rarely become sub-optimal and seldom need re-building.  The other-side maintains that regularly-scheduled table and index reorganizations are unavoidable, and that most high-DML objects show increasing rates of I/O as the internal structures become sub-optimal from high insert/update/delete activity.

Ever wonder about that anonymous expert on the Oracle forum?

In this age of instant publishing, the old adage "Don't believe everything you read", is more true than ever before.  Verifiable identifies are impossible, and you are left wondering if that "tip" you just found is from a credible source or from a malicious DBA who wants to see you crash your database.  For details on verifying the credibility of web information, see Stephen Andert's and Don Burleson's book "Web Stalkers - Protection from Internet Psychopaths", a comprehensive and entertaining survey of crime and deception on the web.

Can't we just blame it on a poor application design?

There is a great debate between Oracle tuning professionals.  The "old-school" maintains that 80% of performance latency is within the application layer (and not the database), while the "new-school" recognizes that the Oracle DBA often has no authority to change the application, except at the database-level.

He wants to spend $50k on hardware and fix the performance problem tomorrow?
That's absurd, the core problem is sub-optimal application code. 
Tell him we need to spend $200k and six months to redesign the application.

"I didn't need no stinkin' High School to be an Oracle DBA"

Here we see the age-old argument about the importance of academics in Oracle technology. One side of the argument says that the dedication, hard work and perseverance required to graduate from college graduate (not to mention computer science skills), makes the "college-educated" DBA more productive. Engineers must have a degree (and a license), and lawyers must have a JD degree.  So, what the value of a technology degree to an Oracle professional? The other side of the coin says that experience-alone can be the best predictor of Oracle BA job performance. "I can't help but wonder if Bill Gates had taken a Data Structures, Algorithms and Operating System Design course in Harvard, that Windows might be a more reliable platform." - DB