Advert – Oracle 10g Performance Analysis

As promised earlier, here’s a blog post about some of the thinking behind the course I’ll be teaching in Prague on 14th/15th April (assuming enough people are interested in attending).

Why I decided to do it
Several factors appealed to me when Oracle first asked whether I’d be interested :-

1) It’s about two and a half years since the last full course I taught and I miss it sometimes, so I’d been thinking of returning to teaching anyway. I taught an average of 10 full weeks per year between about 2001 and 2005 and it’s a completely different challenge from the usual technical work.

2) I can choose my own subject and write my own material. Working with Learning Tree, the support available from the rest of the business (e.g. Product Development and Marketing) was very useful but could be quite restrictive because they weren’t interested in developing courses that might not have a mass audience. I’m glad I was involved in course development while I was there, though, because it’ll come in useful now. Working with Oracle UK (a long time ago) was worse because I had no involvement in course development whatsoever.

3) The technical challenges that day-to-day work at customer sites throw up can be interesting, but business pressures sometimes mean that there isn’t time to delve into them deeply enough for my own enjoyment. Anyone who has written or taught courses will tell you that it always adds to your own knowledge because you have to make sure that all the questions have been addressed. To teach a subject, you need to know it really well.

Why I chose this subject
So, when I realised I could choose, I had a feeling I might end up picking something performance-related, not least because it’s what I’m really interested in. The other possible direction was professional DBA work (i.e. How the best DBAs handle the job at big sites) but it just felt too amorphous and (not necessarily my choice) the course is part of the ‘Masterclass’ series so I feel the technical content needs to be high. At the same time, I wanted the course to be both bang up to date (so even experienced attendees might learn something new) and timeless (because fundamentals always count and I think I have useful experience). So, 10g Performance Analysis it is and, by 10g, what I really mean is ASH/AWR/ADDM/SQL Profiles and the other nifty tools that are available these days. Yes, only to those who are licensed to use them, but I think a DBA should understand even those tools they might not be using at the moment. Fortunately, I’ve been able to use them extensively at my current site.

What to expect
The course outline is fairly detailed and useful as a personal planning aid and sales tool. (Hint: don’t even begin to write something as complex as a course without having a pretty detailed outline at the start. That’s something I learnt from the Learning Tree Product Development people.) but it might be useful for me to highlight specific areas. Here are my three top objectives, in addition to the basic course content.

1. To put the 10g ‘Automatic’ features in context with the short but meandering history of Oracle performance analysis. The overall approach needn’t be dictated by the tools and, even with the best tools, if you don’t take the right approach you’ll never get anywhere (or, worse, you’ll get to the wrong place … fast!).
2. To include as many examples as possible – some from my day-to-day work and some constructed artificially, because I think both have their place. Coincidentally, a conversation about Swingbench cropped up on the Oak Table mailing list just as I was mulling over the examples and it inspired me to take another look. I’d tried messing about with Swingbench in the past but never had enough time to dedicate to it. I might blog about it soon; we’ll see.
3. A late addition – I’ve been very impressed by some of the 11g features in this area and I see no reason not to include them. I know it says 10g in the title, but there will be tons of 10g content so I think it can only add to the course. Think of it as a 10g-and-above course 😉

Why two days?
I spent some time thinking about whether this should be a single or two day course. I hate artificial padding as much as the next techie but, having taught a few courses, I also understand the importance of time to be able to explain things properly, including how these features can actually be used rather than just regurgitating the manuals. I also think it’s important for attendees to have the opportunity to discuss the issues, rather than just absorb swathes of detail.

I also remembered that a very basic introductory presentation at work on this subject ran to a full three hours, with no significant examples. I want to be able to include some detailed examples and case studies because I feel they’ve been a very notable exception from every performance tuning course I’ve been involved within the past. Without them, it’s just a run through of features and wouldn’t deserve the Master Class title. I’m pretty certain that I’ll have more than enough material to fill two days.

There’s a lot of work to get through, but I’m looking forward to it immensely. I’ll be blogging a lot less over the next three months but, rest assured, I expect a few related technical posts to show up here eventually.

Register here