Log buffer 52 – a carnival of the vanities for dbas

“Your honor, your honor” – this week the undoubted privilege of publishing Log Buffer falls to me. Big shout out to Dave Edwards of Pythian for the invite.

In the Oracle community, there’s no doubt in my mind what the biggest issue of the week has been.

An open letter to Larry Ellison from Mark Brinsmead of Pythian has provoked a deluge of signatories and supporting blog articles. Never afraid to put their honest spin on the situation, Howard J Rogers notes, among other things, that Standard Edition users are currently unable to license the feature at all, whilst Doug Burns reckons that Oracle just end up making themselves look small minded.

Maybe the community activity has expedited an ongoing action, maybe it has been just fortuitous timing, but, out of the blue, Oracle are able to reveal to Pythian that AWR can now be disabled after all … but they (Oracle) wouldn’t advise that. Perhaps this is evidence that the community is becoming more influential – credit to everyone, give yourselves a pat on the back but save the (double) whiskies for when it’s a free and integral part of the Enterprise Edition. And hurrah to Oracle for listening and reacting … if that’s what has happened. Or maybe it’s just a case of well done for spotting an anomaly all by yourselves and doing something-that-you-wouldn’t-advise about it.

Elsewhere, when he’s not marvelling at Jonathan Lewis’s commitment Andrew Clarke, is making some worthy points in support of Alex Gorbachev’s battle against any guess (BAAG) movement. Providing evidence further supporting Andrew’s earlier post, Jonathan Lewis illustrates the difference between parse counts,parse calls and “optimisations”. I know I’ve amassed a collection of downloaded scripts over the years, but, in another top quality post, Jonathan reiterates the need to be wary of downloaded SQL and highlights the importance of understanding what you’re about to run. Continuing on the vein of caution and wariness, Niall Litchfield says just be careful out there.

Ever upgraded a critical database without really testing it? Well, Vivek Sharma has and he notes that the default stats behaviour for collecting histograms in 10g can cause issues. Whilst, over at Decipherinfosys, there’s some handy info on how to calculate disk space cost for tables and indexes.

In some vendor agnostic postings, both Slavik Markovich and Lewis Cunningham react to a distasteful incident of a DBA fallen into temptation with his database’s credit card information.

Meanwhile, the mySQL community seems to be gearing up for the latest Open Source Convention – OSCON 2007 in a couple of weeks. Peter Zeitsev lists his top 5 wishes for mySQL. It’s always good to hear about approaches to familiar problems across different databases and Peter describes an approach to implement a counter efficiently in mySQL, while Brian Aker offers a different opinion in response. Kevin Burton has some observations to share from a storage migration from InnoDB to MyISAM.

From the SQL Server frontlines, Robert Davis rejoices that breaking up is hard to do. Something that most of us can relate to, some of us have had to sort out and, of course, something that none of us would violate, Jeff Smith underlines the importance of using the right datatype.

And, finally back in the Oracle realm, I’m sure I speak for many when I say Yes please, Pete.

That was the week that was.

Follow

Get every new post delivered to your Inbox.

Join 72 other followers