UKOUG Days 4 and 5

I woke up on the last full day of the conference with a couple of small things left to do on my second presentation. I was confident, though, because it was just a bit of tidying up. Mmmm, that turned out be a mistake

First, though was Graham Wood's "Cool Stuff with Hot ASHes". I thought it was a good overview of Active Session History with some more advanced snippets thrown in. I particularly liked the reminder of where people tend to go wrong when using ASH data, primarily trying to extract too much meaning from the timings in the individual session samples rather than COUNTing them as a proxy for time. (See here for a brief discussion about that, but it probably merits a future post here too.) He showed more graphic examples from OEM than I remember from past presentations, which I'm convinced are easier for jobbing DBAs to get their heads round, and some interesting visualizations of DB Time that are part of ongoing database performance UI research that I hope we'll see in future releases. Good stuff, but with a full hour rather than 45 minutes, maybe there would have been even more examples.

I skipped the next session for some final preparations of my own because I knew I really needed to tidy up one section in particular on initial configuration. There's a funny story behind this actually. Although I had my original Audit Vault configuration document to refer to, I'd left it at home and I hadn't been able to have one last run through because the AV environment at work has been broken for a week or two. Lesson 1? Never, ever present things from memory and old documentation - always have one last manual run-through in a real environment to check! Lesson 2? Never be too focussed on one presentation, to the detriment of the other.

But I had good reason for the latter, that I touched on during the presentation. When it was UKOUG abstract submission time, I was working on implementing Audit Vault in Pre-Production in preparation for the real deal. Between then and the conference I had moved on to other things, so I suggested to the UKOUG that I do something else but they preferred I stick to the original abstract. Which is another way of saying that my heart wasn't really in it any more. Still, I'd hit enough interesting things in the Pre-Production implementation to hang a presentation around and I decided to use a very different style to the performance presentation - straight bullet points about lessons learned.

As I started to speak, I noticed Tammy Bednar who (amongst other things) is the Audit Vault Product Manager sit down in the back row. Talk about pressure! On the whole, I suppose it was a reasonable presentation and I hope people got the odd useful detail from it (not least the difference between Database Vault and Audit Vault!) but it was very far from my best. That's not just natural self-criticism, but backed up by people like Marco, who have seen me present a few times.

Which made it all the more pleasing when Tammy introduced herself afterwards and complimented me on the presentation. Any 'deployment lessons' presentation is bound to focus on the problems and quirks, rather than the more boring stuff that just worked ;-), so I was impressed that she took it all on the chin with a determination to address problems rather than spin them. A class act and no wonder she's a Product Manager.

How classy she is became even more apparent an hour or two later when she emailed me a list of corrections to that configuration section that I'd rushed at the last minute. The easy approach would have been to be defensive about the product and attack my mistakes in public (because, believe me, there were some). Instead, she notified me quietly later and when I've sifted through them and made sure everything's just right, I'll upload the slides and post the procedure here. What I find funny is that section was called "Initial Configuration is Weird", and focussed on how everyone who starts working with Audit Vault seems to find it tricky at first so I feel less bad about getting the details mixed up On the one hand, it was one of the more important sections of the presentation to get right and I shouldn't have rushed it near the end, on the other, it's telling that I got it wrong, even though I've worked on this!

Finishing my work for the week had the usual effect of tiredness hitting me like a sledge-hammer, but there were three or four presentations I still wanted to see. In the end, I managed two, which wasn't too bad

First was Robyn Sands' "Stopping the Madness with Root Cause Analysis". I'm a big admirer of Robyn's work (from a distance, electronically) and we share interests in the field, but I've never heard her speak. She's an extremely listenable, natural speaker, but I wish she'd been in a better room, with a slide-clicker that worked, a microphone and I'd been able to hear her a little more. I wouldn't want to hear a grating, insistent, loud voice ... but when you have the calm, confident voice that Robyn has, I want it to be louder. Regardless, I found myself nodding in agreement with what she had to say and it occurred to me that she was describing in formal terms how I think I practise being a DBA. It was better in formal terms, though, and I think it might prove really useful to me in explaining to managers what I think are the obvious benefits of my approach to DBA work in a way that makes more sense to them. Excellent stuff, but that clicker drove me crazy

A quick lunch and I arranged to meet Kurt after his presentation. (I skipped it because the subject matter wasn't of that much interest to me, but I hear he knocked them dead.) Instead, I attended "CSI Oracle – Investigate Past Performance Problems Using ASH, AWR, ADDM, and Optimizer Statistics" by Patty Charlebois. As well as knowing Patty a little from a couple of trips to Hotsos, the subject area's pretty dear to my heart so I try to attend any related presentations to see how other people are using the Diagnostics pack. Patty's a great speaker and her introduction was so entertaining that I really enjoyed listening to her. In the end, I suppose I was a little disappointed because it could have done with a bit more depth on the actual problem resolution process and finished pretty early too (but it's always better to finish early than late!). It was a fair presentation, but I was left wanting more somehow.

I finally got a chance to catch up with the Belgians (becoming something of a conference tradition and as more and more people appeared in All Bar One, the likelihood of me returning to attend any more presentations became slimmer and slimmer. I was disappointed when quite a few people walked in and said how great Robyn's second presentation was so I'm sorry I missed that.

All in all, it was a late, late night followed by a very hungover Friday. I planned to attend Christian Antognini's Execution Plans Masterclass and then Tanel's in the afternoon, but it was standing room only in the first (for a 2 hour Masterclass!) so I spent the rest of the day saying cheerio to a fee people and catching up with emails before heading home. I could probably have gone home on Friday, but it was probably worth staying if only to catch up with more people on Thursday night.