Nov 15: OOW Day 5
The final day. I didn't wake up until 5 and had a stack of emails relating to the performance problem I'm discussing at the moment. Once I'd got through those, I bumped into APC checking out of the hotel. He'll be on his way to the airport now. I have to offer him a big thank you, though, for reminding me of the final Real World Performance presentation at 8:30 this morning.
What a treat, particularly from a purely selfish point of view. "When to use the appropriate database technology" was Andrew Holdsworth's over-amibitious attempt to discuss a raft of features and where their use might help performance. I say over-ambitious, because he recognised himself how impossible that would be in an hour, so made the wise decision to limit himself to three topics.
It was fascinating stuff and I had a strangely split reaction to it.
a) Bl**dy h*ll, I'm going around telling people they won't get much joy with Degrees of Parallelism much higher than single digits and these guys are using DOPs in excess of 100! That kind of blows away anything I've got to say on the subject.
b) Actually, what I've written about is completely true for lots of sites because, a majority of customers have 1 or 2 HBAs operating at 100MB/sec. He described this as effectively switching off PX as a viable option.
There was so much excellent content here that I'd rather blog or write about it properly later. I am so glad APC reminded me, because it was my presentation of the week!
(Oh, and at least he agreed with me about not using parallel_adaptive_multi_user
)
After that, I skipped a presentation to catch up on some blogging and then I was torn. I'd promised my current client that I would attend "Best Practices for Planning, Testing and Deploying Oracle's CPU" because patching is becoming more important at work. However, I knew Kevin Closson was speaking at the same time and I fancied attending that. In the end, I went for the CPU presentation. I'm not sure I made the right call, but there were a few interesting bits and pieces in there and I'll probably cascade them back to the team when I get back on-site, but too boring for inclusion here. One statement that may not have been announced yet was that 9 months after a new minor release is available (e.g. 10.2.0.4) CPUs will not be available for the previous minor release (e.g. 10.2.0.3). That sounds like yet more upgrade workload. Oh, goody!
Next was a burger (same place as last time!) a re-charge of my laptop and back in time for Tim Hall's Unconference presentation "The Oracle DBA - A Dying Breed?" There's a really good crowd, it's a very interactive survey of the state of the DBA/Developer relationship and, as I expected, Tim's an energetic and funny speaker. (It's one of the reasons I switched my schedule - I need entertainment on the last afternoon.) He even managed to withstand a barrage of loud music from across the hall. Oh, but I think he had a very modern, helpful, friendly bunch of developers and DBAs because I didn't recognise the responses as being representative of my experience. It was like a Developer/DBA love-in for goodness' sake!
Good stuff, Tim!
After this I might catch one more presentation, then the end of conference social in the Howard Street tent, then it's on to an Oak Table dinner kindly organised by Jared Still. I doubt there'll be much more blogging between now and my flight home tomorrow, so that's all for now, folks!
What a treat, particularly from a purely selfish point of view. "When to use the appropriate database technology" was Andrew Holdsworth's over-amibitious attempt to discuss a raft of features and where their use might help performance. I say over-ambitious, because he recognised himself how impossible that would be in an hour, so made the wise decision to limit himself to three topics.
- Parallel Query, Partitions and Hashing
- Optimiser
- Session and Connection Management
It was fascinating stuff and I had a strangely split reaction to it.
a) Bl**dy h*ll, I'm going around telling people they won't get much joy with Degrees of Parallelism much higher than single digits and these guys are using DOPs in excess of 100! That kind of blows away anything I've got to say on the subject.
b) Actually, what I've written about is completely true for lots of sites because, a majority of customers have 1 or 2 HBAs operating at 100MB/sec. He described this as effectively switching off PX as a viable option.
There was so much excellent content here that I'd rather blog or write about it properly later. I am so glad APC reminded me, because it was my presentation of the week!
(Oh, and at least he agreed with me about not using parallel_adaptive_multi_user
)After that, I skipped a presentation to catch up on some blogging and then I was torn. I'd promised my current client that I would attend "Best Practices for Planning, Testing and Deploying Oracle's CPU" because patching is becoming more important at work. However, I knew Kevin Closson was speaking at the same time and I fancied attending that. In the end, I went for the CPU presentation. I'm not sure I made the right call, but there were a few interesting bits and pieces in there and I'll probably cascade them back to the team when I get back on-site, but too boring for inclusion here. One statement that may not have been announced yet was that 9 months after a new minor release is available (e.g. 10.2.0.4) CPUs will not be available for the previous minor release (e.g. 10.2.0.3). That sounds like yet more upgrade workload. Oh, goody!
Next was a burger (same place as last time!) a re-charge of my laptop and back in time for Tim Hall's Unconference presentation "The Oracle DBA - A Dying Breed?" There's a really good crowd, it's a very interactive survey of the state of the DBA/Developer relationship and, as I expected, Tim's an energetic and funny speaker. (It's one of the reasons I switched my schedule - I need entertainment on the last afternoon.) He even managed to withstand a barrage of loud music from across the hall. Oh, but I think he had a very modern, helpful, friendly bunch of developers and DBAs because I didn't recognise the responses as being representative of my experience. It was like a Developer/DBA love-in for goodness' sake!
Good stuff, Tim!
After this I might catch one more presentation, then the end of conference social in the Howard Street tent, then it's on to an Oak Table dinner kindly organised by Jared Still. I doubt there'll be much more blogging between now and my flight home tomorrow, so that's all for now, folks!
Oracle OpenWorld 2007 has come and gone and from the Real-World Performance Group’s perspective we’d consider it a successful one. The content of this year’s presentations seems to have gone over quite well as shown by the number of ...
Tracked: Nov 19, 23:59