I am back from my third visit to UKOUG and as always, the trip was wonderful, the conference was excellent and if I could have stayed, I would have. Of course, after dealing with trains all morning and then being stuck in an airplane seat watching movies for hours on end, sleeping in my own bed last night was absolute heaven. But the UK must be a wonderful place to live: the people are lovely, the architecture is gorgeous, and I love walking through the streets or riding the trains to find unexpected views and places.
However, I'm not sure I like the compressed schedule this year as there were too many simultaneous presentations that I wanted to attend. As usual, there were times when I didn't get to any sessions due to work, but this time if I was out for an hour, I missed four great presentations instead of one or two. And during my presentation, there were at least three that I wanted to attend and obviously couldn't, so I am very grateful to the people who did show up in my room. If there had been a few less, I think I would have invited them to follow me over to Eric Grancher's room so we could learn about how CERN is using AWR together.
This year, the most important lesson I learned was in my own session, and the lesson is that variance is too big of a topic to be presented within another topic. My topic was 'Embedded Oracle: the Real Self-Managing Database', aka what needs to be done differently when embedding the Enterprise Edition of Oracle in a customer product to minimize or eliminate the need for administrative tasks. One of the key differences is that we instrument the code with the Instrumentation Library for Oracle (ILO). I've made some changes to the ILO code to collect the elapsed time on all executions of key processes and then I can use the resulting data to analyze the performance history and trends, which is very useful when you don't have real time access to evaluate a database issue. For a 45 minute session on embedded Oracle, this level of info along with the code changes would have been enough. Instead, I attempted to cover using variance metrics as well, and I'm afraid I just confused the audience. Once we got into their questions, it was clear that they would have been better served by hearing about variance in multiple kinds of databases, which is how I present it in the 'Variance as a Tool' presentation. Later in the 'Meet the Speakers' forum, we joked that I needed to institute a prerequisite system for my attendees, but seriously, it is my responsibility to cover the material effectively and I fell short this time around. I'll be more careful about how I combine the subject matter in the future.
For those who would like to learn more about measuring variance and why it's useful, I have contributed a chapter for the new Oak Table book due to release early next year. My chapter focuses on the importance of measurement and considering the predictability of a process in addition to the response time. I've also promised Apress that I will deliver my changed ILO code as a package to be downloaded from their website. If any of those who attended my session have questions that didn't get answered, they are welcome to contact me directly at the email address given on the last page of the slides, or at my gmail account via Oracle-L.
As for the sessions I attended, Active Data Guard, ASH and AWR are my primary focus for my current project, so I aimed to attend any session on those subjects. Larry Carpenter gave two top notch sessions on Active Data Guard, and since I've been using his Oracle Data Guard 11g Handbook as a reference for our installation, configuration and documentation, both sessions were exceptionally useful to me. His book is very, very good, and one of my favorite things is that he encourages people to 'feed their inner geek' and learn what is behind the GUI tool methods. One of my pet annoyances is when people rely on the GUI to the extent that they never learn what the tool is doing behind the pretty picture, so I loved the 'feed the geek' references. Tools are great and 'clicking on the big stuff' can be the fastest way to solve the immediate problem, but that doesn't mean that we should be dependent on the tool. Perhaps everyone can aspire to become DBA version 2.5: use the tools effectively but be able to manage just fine without them too. As you probably guessed, I attended Graham Woods' 'The ASHes of Time' session :) He's always good, but the combination of info he provided in this one really does the topic justice so if you didn't make it, be sure to get the slides and watch for it at the next event you attend.
Which reminds me, I don't think mine are up yet ... need to check that first thing in the morning.
2 years ago