Log Buffer #142: A Carnival of the Vanities for DBAs

Posted in: Technical Track

Welcome to the 142nd edition of Log Buffer, the weekly review of database blogs.

The SQL Server ‘sphere was a busy place this week. On In Recovery… Paul S. Randall posted his latest straw poll, this time looking into your practices around transaction log size management.

Linchi Shea observed, “In a multi-process/multi-thread system, locking is central to maintain data consistency and keep things in order.  . . . [We] need to begin with understanding the locking behavior of the basic building blocks offered by SQL Server . . .  [One] would think that the basic locking behavior of these building blocks  . . .  would be well documented. Unfortunately, that is not the case at all.” Is it too much to ask for, he wonders, to document the locking behavior?

Kevin Kline is looking for good DMV/database admin queries. You can post or link to yours in the comments.

On thinking outside the box, Peter Larsson offered a little introduction to composable DML, a new tool in 2008. “What is Composable DML? Well, with this tool you can have a statement of UPDATE, DELETE and even MERGE as a data source for your query!”

Michelle Ufford, the SQL Fool, wrote, “[While] I was at the grocery store last night, my mind wandered to SQL Server.” (Yes, that happens to me too.) “This time, I was pondering what happens to a page split if the transaction is rolled back.”

Aaron Alton of The Hobt bravely interposed himself in the ongoing DBA-vs.-developer scuffle to say, “Developer, meet tempdb. tempdb, meet Developer. What’s that, DBA? You’ve never met either? Oh, how rude I’ve been. Allow me to introduce you.” Most cordial of him, I think. He kind of loses his cool though—“For the Love of All That is Good and Pure, tempdb is Your Friend!”

On the Oracle side, our old friend Chen Shapira discovered an unexpected side effect of table shrinking, ” . . . especially when using ‘enable row movement’.  . . .  As the name implies, after enabling row movement, rows get moved around. This can impact the clustering factor of your primary key.” In the comments, Yas links to a post illustrating the culprit here—buffer pinning, by Don Chio.

Chen also brings news of the First International NoCoug SQL Challenge. Although the contest is still open, Rob van Wijk thinks he may have found the winner in Alberto Dell’Era, by whose solution Chen says she was blown away.

Pawel Barut issued a caution on the FOR loop and CONTINUE in Oracle 11g.

Coskan Gundogar similarly found a good reason to move from to or 11G: a resultset bug.

While we’re on bugs (or while they’re on us), let’s look at some MySQL ones. Baron Schwartz found, in MySQL 5.1.33, 4 secret bugs. This follows an item on MySQL and IBM by Baron’s colleague Vadim, in which the latter discovers an unheralded new storage engine in 5.1.33.

On High Availability MySQL, Mark Callaghan unearthed not a hidden, but an obscure bug. “We . . . found an InnoDB table with rows in the secondary index that were missing from the PRIMARY index.”

Anders Karlsson pipes up with his take on what’s wrong with SQL_MODE. Schlomi Noach and Roland Bouman provide links to their own contributions to this discussion.

Ronald Bradford appeared with the latest in his series a beginners look at Drizzle – Datatypes and Tables.

Andrew “gilfster” Gilfrin confessed, I am a Dummy. Or rather, --i-am-a-dummy.

Do we need a MySQL for dummies? Morgan Tocker wonders, who is going to make MySQL easier to use?, with a laundry-list of things that interfere with ease-of-use. Unfortunately, the comments degenerate from a styling of MySQL as “CSV on Steroids” by a PostgreSQL fan.

Well, some people know how to get along. Baron Schwartz is after all, a big MySQL guy. Yet, here he is reporting the highlights PostgreSQL Conference East 2009, Day Two.

In the Database Soup, Josh Berkus has some reminders about pgCon & OSCON Registrations.

And now, a little levity to close this busy edition. Maybe a very little. What do you call a Santa that can turn into a wolf? Click through for the big payoff, courtesy syscomments. Me, I’m just glad there ain’t no sanity clause.

And that is all. As always, I invite you to leave your favourite DB blogs from this week in the comments. And please send me an email if you’d like to publish an edition of Log Buffer on your own blog.

See you in a week’s 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.

4 Comments. Leave new

1. I object to being called old!
2. No one from Pythian is participating in the NoCoug challenge?

David Edwards
April 14, 2009 3:57 pm

1. Chen, I intended no offense. I only meant “old” in the sense of your having been a guest editor of Log Buffer before. And that you are, let’s face it, no spring chicken anymore.

2. Yeah, what about that??

viagens baratas america do sul
September 29, 2016 11:29 am

A humanidade esses critérios fazem parte da rotina de construção dos
dois telejornais e também atuam na seleção
da notícia veiculada através de eles.

David Edwards
April 15, 2009 12:16 pm

Okay—Singer Wang stepped up with an idea based on the central limit theorem. He’s more of a theorist, however, so we’re still waiting on the code.


Leave a Reply

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