Having been involved with the UKOUG for more than 10 years now, one of the things that is so nice about returning to the annual UKOUG conference is the sheer number of familiar faces.
It also is a great place to hear about other people’s experiences and an opportunity to broaden your horizons.
Sadly due to my youngest daughter’s terrible timing of being born at this time of the year, I had to miss the fabulous Oaktable Sunday and be at her 4th birthday party instead. Thanks to everyone who told me how good it was. While this did not exactly help, I’m glad others had fun.
One other regular feature of UKOUG conference time is that weather becomes extremely cold. It has been a ridiculously mild autumn/winter in the UK so far this year, but as I soon as I got to Birmingham I’m sure the temperature was 10C colder than in the previous week.
Some highlights that I particularly enjoyed:
Monday at UKOUG had a lot of talks in an Exadata stream, so I spent most of my time swimming there. The highlight presentation for me on Monday, was by Frits Hoogland. It started off looking like there would be a small crowd, but people kept flooding in after the start and by 5 minutes in, there was standing room only. A small explosion of a lightbulb blowing failed to put Frits off his stride in his “Explaining Exadata” presentation, and Frits gave an excellent overview of the components of Exadata.
I really liked how Frits was trying to strip away the magic and mystique surrounding Exadata – as he says it’s not magic just built on standard components. He then delved into how ASM works with the storage cells, and gave a good discussion on flash performance numbers. He Also gave some good information on real life infiniband performance.
Frits had a nice filmed demo highlighting various features of exadata using select of a very large table. Frits’ example of the EHCC showed he could get huge space savings. He had lovely visualisations of the io and database wait profile while doing the scan of the table.
Frits shows that going from 184.108.40.206 to 220.127.116.11 improved serial direct path speed quite substantially. He then showed the improvement that parallel query could bring.
I also attended an Exadata round table, and there were people from a few companies who had implemented Exadata. I had gone hoping the discussions would be very techie, but a lot of the discussions focused around more of the business issues involved in an Exadata implementation – not because there were not loads of techies at the roundtable, but I think this is indicative of how disruptive Exadata is. It is hard to confine it to the normal silos that companies have. Seems like having DBAs sitting in the Unix/SA team to manage Exadata really is the way to do it.
Tuesday really was a superb day at UKOUG, presentation after presentation was just outstanding. The quality of the presenters is outstandingly high. First up I thought I’d try something different and went to Alex Nuijten talking about analytic functions. Now, I’m not much of an SQL coder, but it is good to have a change from the usual presentations I go to: RAC, Dataguard – more towards the “infrastructure” end than the SQL end of being a DBA.
Alex is a superb presentator, was deeply impressed by the quality of the slides. Being on first thing in the morning meant I think he had a tough crowd, and he kept asking for audience feedback but most people were fairly quiet. He explained most of his material really well. As he said himself, this was not an introductory presentation, so while I got lost in a few of the steps/techniques he showed, I’m sure for the vast majority of the audience, starting in a better place than me, this would not have been the case.
Next stand out presentation was by Connor Mcdonald on a year in purgatory, where he had some issues regarding upgrading, and eventually had to perform upgrades over 8 Oracle versions to find a stable release.
All I can say is wow. This was a stunning presentation.
I was on the edge of my seat wanting to know what happened next. That sickening feeling when one of his production upgrades went horribly wrong and he was looking at potential millions of pounds of loss for the business, was just palpable.
Lots of useful tips, including what not to do, but his final message about essentially being on the mainstream in terms of release (OS) I think is so true.
I then had the pleasure of a Tanel Poder troubleshooting masterclass where he discussed a performance problem and as always emphasised that you need to take a systematic approach to problem solving.
As if that was not enough top presenters and presentations I had even more to come in the form of a fairly impromptu parallel query masterclass with Greg Rahn. Greg managed to go on for nearly 2 hours, and there were numerous useful asides. Greg also just has a great way of presenting complex information, I really was astonished by the depth of his technical knowledge, this was in the unconference and there were only a few us lucky enough to attend this one, but I’m certainly glad I did!
After a couple of really late nights, though I admit not as late as some, I was feeling really tired come Wednesday morning, but as I could only stay for a few presentations I thought I would attend the first presentation of the day. I started of the day with Doug Burns and his presentation on partitioned statistics, and I am really glad I did. Doug was sounded a bit underweather earlier in the week, but he was bang on form with this presentation. Doug has a really great rapport with an audience and, maybe this is the Scot in me, but I think he has a lovely delivery – a really good presenter. Also the material was very interesting and I picked up some statistics gather tips – it’s hard to get decent statistics on very large tables even if they are partitioned.
One other new feature of UKOUG 2011 was the lunchtime oaktalks, which were 10 minute lightning talks. These were a lot of fun.
Wednesday was somewhat curtailed for me as I had to return to Oxford to see the tinselled vision of my eldest daughter singing in her school Christmas concert.
All in all, it’s been a great few days.