OH WOW! Right now I am slapping myself on my back so hard that it hurts! In an earlier blog post on I/O Optimization and Solid State Disk (SSD), I referenced a recent IBM DM Magazine article and discussed an improved weighting formula for determining which tablespaces would benefit most from SSD storage. WELL, I just created a masterpiece SQL Snapshot command that implements best practices weighting factor calculations for SSD! This SQL Snapshot command is amazing...

[ more.. ]

DB2 LUW Database Professionals: If you can only get budget to attend ONE class this DECADE, I strongly encourage you to do whatever it takes to get yourself enrolled in my NEW Ed Seminar DB2 for LUW Performance Analysis and Tuning Workshop being given at IDUG in Tampa, Florida, on 10 May 2010.

For this lab, you should bring with you copies of SYSIBMADM Snapshot Catalog Views (detailed instructions and scripts will be provided separately). In this lab, you will receive dozens and dozens and dozens of SQL Snapshot commands (on CDROM) to analyze your database's performance data, you will learn what the numbers and metrics mean, and you will graduate fully equipped to conquer amazing performance tuning miracles for your organization...

[ more.. ]

IBM is generating a lot of buzz with DB2 pureScale, and there is plenty of press to read on Database and Hardware wars. The DB2Night Show™ Episode #4 featured special guests from the IBM Toronto Lab who gave us an outstanding introduction to DB2 pureScale - catch the WMV replay if you missed it. On the heels of DB2 9.7 Oracle enablement, DB2 pureScale is an exciting response to Oracle RAC. But is pureScale right for your organization, or should you be considering Xkoto GRIDSCALE instead? First and foremost, IBM DB2 PureScale is NOT a Disaster Recovery solution, but Xkoto GRIDSCALE is. How else are these technologies different?

[ more.. ]

IBM Data Management Magazine Q1 2010 arrived today and I was excited to find an article on Unleashing the value of Solid State Drives. Of course, I was thrilled to see DBI's new ad as well (grin). The article puts forth several good ideas about I/O tuning and begs for greater details. Whether you have SSD or not, the principle ideas of the article still apply to "faster storage" versus "slower storage". Let's take a deeper look...

[ more.. ]

In the wake of recent claims by Larry E, I'm joining Craig Mullins and re-cycling a former blog post with updates...

29 Jan, 2010 - Redrock Shores, CA. In late breaking news today, Orickle announced preliminary plans for the next version of its world dominating database 13G. Version 13G will include new features such as DB3 LUW Enablement and nano compression. A senior Orickle executive who requested anonymity said "We skipped 12 because 13 is a lucky number, and 'G' now stands for Global domination. We lost some serious market share to DB3 9.7 but we believe we can win global database dominance back with our superior marketing, sales, and negotiation tactics." Continued...

[ more.. ]

Recently a customer wrote me and asked if there was a way to run a query that would report on any pending (deferred) DB CFG or DBM CFG changes. More specifically, he wanted to create an alert in DBI's Brother-Hawk™ that would alert him if any changes would occur upon the next database restart. With V9+ SQL Snapshots, this is easy...

[ more.. ]

This blog post is for developers, testers, and QA professionals. If you are a production/operations DBA, it is okay if you'd like to read it too so that you can be aware of the proposed suggestions and best practices. I think we can all agree that discovering performance problems in production is NOT ideal, so let's see what can be done to arrest issues before they become detrimental, damaging, or costly to your organization. We'll begin by busting some myths and unwise practices...

[ more.. ]

Very few DB2 LUW performance blog posts have delivered so many emails to my Inbox as the prior posting on VOLATILE tables. One of our blog readers brought some very important information to my attention that I feel an obligation to share...

[ more.. ]

Do you have a table that has statistical mood swings? Sometimes the table has a few rows, and sometimes it has many? Or would you like to strongly encourage the DB2 optimizer to use available indexes? VOLATILE isn't a nice description for a person, but you can assign this label to a moody table and often get excellent results. More...

[ more.. ]

Watch The DB2Night Show™ Episode #10 on 8 January 2010 at 10am CST! Our topic is "DEAR IBM..." and DB2 LUW users from around the globe have sent their comments and questions into our show! Our guests include senior IBM executives Sal Vella, Tim Vincent, and Drew Bradstock!

You can obtain your free seats to our virtual studio audience and find replays of previous episodes at http://www.DB2NightShow.com

The mission of The DB2Night Show is to Educate, Inform, and Entertain. In just a few months, we've had hundreds of DB2 professionals in our virtual studio audience and show replays have been watched thousands of times. If you haven't joined us yet, please be our guest! And if you've become one of our many loyal "groupie" followers, thanks for watching!

Cheers,
Scott Hayes

The IBM IOD Conference will be buzzing with news on DB2 for LUW 9.8 pureScale - amongst many other breakthrough announcements. But if you are not able to attend IOD this year, or if you are but want to be amongst "the first to know" the details, don't miss episode #4 of The DB2Night Show™ - Special guests Matt Huras and Aamer Sachedina from the IBM Toronto Lab will take us on a deep dive into this new release of DB2 LUW that offers breakthrough scalability without limits. Join us on 22 October 2009 at 9am CENTRAL/10am EASTERN for this in-depth first look at IBM's latest performance miracle. Claim Your Front Row Seat HERE. More...

[ more.. ]

We heard it hear first - live from IDUG Europe in Rome Italy! In the wake of Orickle's 13G announcement, clearly IBM has its boxing gloves on!

DB2 LUW pureScale is somewhat akin to DB2 z/OS sysplex data sharing - multiple "member" hosts share storage and are interconnected by a high speed switch, although, in DB2 LUW's case, the switch is facilitated by innovative software running on a robust Power series server. The architecture and scalability is ideal for organization's having extremely high OLTP transaction processing requirements. More...

[ more.. ]

IDUG is a great conference for learning from other users, consultants, and top shelf IBM people. It is also a great place to meet with vendors to discover the latest products and services that are available. DBI and our Italian partner Expertise4IT look forward to meeting you in the exhibit hall.

And just in case you would like to invite some of your associates or executives who are unable to attend the full conference, just click THIS LINK and print your FREE pass!

Also, have you signed up yet for my Ed Seminar "300 Fast Fabulous DB2 Performance Facts"? The SQL Snapshots I'll be sharing with attendees are worth the price of the class alone!

Cheers, Scott

Are you Joe the DBA? Or Jane? Do you have questions about DB2 LUW Performance? Would you like to learn about DB2 Performance in real-time? Join the studio audience for The DB2Night Show and get the latest tips, techniques, and answers to your questions --- like this question about PCTFREE that we recently received.

[ more.. ]

Keep your skills sharp and your databases optimized - attend IDUG Europe in Rome to advance your career, learn the latest DB2 tips, tricks, and techniques, and improve your business performance. Visit the IDUG Europe web site for details! And be sure to sign up for my Ed Seminar "300 Fast Fabulous DB2 Facts and Performance Tips"!

[ more.. ]

We just completed our Webinar on DB2 LUW Index Design Best Practices and Case Studies. Clustering Indexes were discussed, and one of the Webinar participants sent an email asking this question:

Great webinar today, thank you. A quick question on clustering indexes. When trying to determine the best columns by looking at the highest aggregate sort time, what should one do if there is already a cluster index on the table? That is, what to do if the existing cluster index is suspected to not be the right one? Here's my answer...

[ more.. ]

Hello DB2 LUW Performance fans,

Just a quick note to let you know you can save $100 off your IBM IOD 2009 Conference registration by using Promotion Code G09BOND. This promo code is valid for a long as IOD registrations are open and available.

[ more.. ]

Despite all the hype about AUTOMATIC tuning, a remarkably large number of DB2 LUW customers are still suffering from inadequate and sub-optimal performance. AUTOMATIC tuning can provide some benefits, BUT it only works well when the database is free of significant physical design defects.

We have heard from DB2 LUW users that their ETLs take too long. Index changes were the solution.

We have heard that response times were too slow. Index changes were the solution.

We have heard that CPU utilization is too high, or that there are periodic spikes in CPU utilization. Index changes were the solution.

We have seen OLTP applications appear to "seize up" or "stall". Index changes were the solution.

We have seen DB2 users struggle to complete year end and quarter end reports on a timely basis. Again, Index changes were the solution...

[ more.. ]

It was about a year ago that a DBA in Sydney Australia asked me if I knew of a way to correlate database tuning efforts to Energy savings. IBM's Smarter Planet campaign gave me renewed inspiration to research this important topic, for the fourth pillar of Smarter Planet involves seeking new ways to improve efficiency and lower costs of existing processes. What you are about to learn might be a real eye opener.

[ more.. ]

I pushed my father in a wheel chair for the first time last month. As I looked at the man who was my Scout Master, who helped me achieve Eagle rank, who taught me to bowl, and inspired me with strong core values of helpfulness and community service, these wheel chair excursions marked a pivotal inspiration point for contemplating my own current health, and have given rise to my own self-improvement goals that you can read about in my personal blog. This year's new IDUG presentations and Ed Seminars will also include new information on DB2 Health monitoring. So, in this blog post, let's look into the health and fitness of your databases.

[ more.. ]

I was recently contacted by a DBA at a large insurance company. After upgrading from DB2 V8.2 to 9.1.0.3, the DBA activated the Self Tuning Memory Manager (STMM) in the hope that it would solve all performance woes. To the contrary, performance got worse, and the DBA noted a marked increase in SORT overflows. He even opened a PMR with IBM. But this isn't a case of STMM gone crazy; something else is going on, but what? Let's take a look at the numbers and discover the truth.

[ more.. ]

Congratulations to the IDUG Europe Conference Planning Committee (CPC)! IDUG Europe 2008 was one of the best IDUG conferences ever! On 17 October 2008, I taught the one day Ed Seminar "DB2 LUW Performance Diagnosis Learning Lab." This post contains a review of one of the participant's data and findings. The data comes from a busy production database that runs a banking application.

[ more.. ]

Congratulations on hunting your elephants. The prior blog post was the most popular ever; it has received well over 3,000 hits and continues to ascend in popularity. I readily concede that elephants can wreak havoc in your databases, but there exists a much more dangerous predator of your database's performance: mosquitoes.

Mosquitoes are, on an individual basis, small, inexpensive statements with an apparently low timeron cost, but, when they are executed with high frequency, they will suck the life out of your system and degrade the performance of your business. Two blog posts ago, the topic of statement workload analysis was introduced. If you have not read it yet, please do so now.

[ more.. ]

Hopefully you read the prior blog post on elephants and mosquitoes which discusses different approaches to statement performance analysis. If you haven't read it yet, please do so now. In this post, we will look at ways to hunt and kill your elephants. And, by no means do I favor cruelty to animals in any way, this is just a metaphor. Let us imagine that your phone just rang and your boss is screaming "What's happening RIGHT NOW?!?!?!!?!?" ...

[ more.. ]

It's said that there is more than one way to skin a cat, meaning there are multiple ways to accomplish the same objective. As for the saying, I don't think this is very kind to cats even though I prefer dogs. The prior blog post discussed making adjustments to CHNGPGS_THRESH to reduce, avoid, or mitigate transient "brown outs" in transaction throughput. There is another way...

[ more.. ]