Doug's Oracle Blog

Hotsos Symposium 2012 Summary

dougburns@yahoo.com (Doug Burns) — Sun, 29 Apr 2012 14:13:00 PDT

For those of you who don't follow Twitter (and, let's face it, I still think there are solid reasons to avoid the beast or at least treat it with caution), the 10th Annual Hotsos Symposium finished almost two months ago. It's just taken me a while to blog about it! (Unfortunately, blogging activity has taken a back-seat compared to, well, a bunch of other activities ...)

I'm going to try to talk less about the various dinners and meet-ups with names who most people don't recognise anyway, but Hotsos probably has the highest percentage at any conference of people I know and don't see often (maybe not UKOUG) so there was plenty of food, chat and some somewhat more gentle drinking than most conferences. Or at least the nights finish much earlier than I'm used to at other conferences!

Monday

As this was the 10th Hotsos Symposium, it kicked off with some presentations to 10 year veteran attendees and, much more impressively, 10 year veteran *speaker* Wolfgang Breitling. It was a shame that Cary Millsap wasn't also around to pick up his jacket. Stephan Haisley would have been in the 10-year crowd but couldn't make it. Well, he sort of made it by playing drums via live feed and then it was onto Gary Goodman's keynote presentation which was all about focussing on the right issues for your organisation, rather than getting distracted by too much detail and not being able to see the forest for the trees.

They'd lined up Maria Colgan for the single-stream technical session immediately after the keynote and, despite having seen a lot of the content of "Inside The 11g Optimizer: Removing The Mystery" before (although Maria would no doubt claim it was entirely new), she was as entertaining and informative as always. I was discussing it with a fellow speaker afterwards and we agreed that it's unusual to hear someone who speaks so quickly and energetically but covers all the points properly too. Well, almost all. Alex Gorbachev tweeted his disenchantment that she didn't cover the One-Pass Distinct Sampling algorithm, although all the geeky detail is available for anyone who is interested enough. Anyone wanting to hear Maria present a lot more should attend next years Symposium Training Day as she's the planned speaker.

Iordan Iotzov - "Advanced Methods For Managing Statistics Of Volatile Tables In Oracle" was next for me. It wasn't quite what I was expecting because of my own focus on frequent partition-exchange operations rather than standard DML and I'm not sure how much the after-DML trigger techniques used would translate very well to my own work, but it was thought-inspiring stuff, nonetheless.

Despite having left myself with far too much work to do on my own presentations and so needing to spend a good bit of Monday and Tuesday on that, I wasn't going to miss Cary Millsap talking about "Instrumentation: Why You Should Bother". Which was a good job because, as well as his usual high standard, he had a lot to say about the importance of using DBMS_APPLICATION_INFO which really comes into it's own when you start using the ASH Analytics I was due to cover in my second presentation. I was also able to perform a couple of useful functions. First I helped carry Alex Gorbachev's ridiculously heavy gift from Cary back to his seat (I'm starting to wonder if Cary just takes heavy stuff with him whenever he knows I'll be there to help carry it) and I was also able to model the 'I Can Help You Trace It' t-shirt which I'd worn for a bit of free publicity/support, but hadn't thought of the likely implications of wearing it into the same room as Cary's presentation. Regardless of the embarrassments, Cary was as thought-provoking as always and it was over too quickly.

All I had time left for after that was a brief visit in to see Mark Rittman make his Symposium debut with "Inside Oracle Exalytics And Oracle TimesTen In-Memory Database". It was interesting to see another experienced speaker go through the unusually nervous experience of stepping up on to the podium in Dallas. I don't quite know what it is, but it certainly feels more scary and challenging than most conferences, but I'm convinced it brings the best out of speakers. Certainly, although I could tell Mark was more nervous than usual, he also seemed very focussed and deserved the beer I bought him afterwards before we headed off to the Enkitec Exadata Panel

As well as beers, there was a good spread of Exadata talent on that panel (and in the room, I suppose) and I think the informal atmosphere worked well. (Picture me telling Jonathan Lewis to shut up before proceeding to heckle about Exadata OLTP which, of course, drew a well-deserved retaliatory rebuke!)

Although there were probably a couple of beers after all that, most of my attention was now focussed on my own presentation so I had to skip dinner with a few Oak Table chums. I knew it would be worth it in the end, though ....

TUESDAY

By waking up early and skipping breakfast, I managed to attend a few bits and pieces of presentations here and there, kicking off with Jonathan Lewis' - "The Beginners Guide to Becoming an Expert". It's an umbrella presentation that I've seen quite a few times now but the detailed content changes every time so is never boring. As well as a live demo of human reaction times using a bunch of barely-organised volunteers ;-), the gist of the presentation was that you need to learn about the basic mechanics of Oracle so that you can understand what you would reasonably expect Oracle to do in different circumstances and then verify those initial assumptions. Or something like that - it *was* a long time ago now!

Although I managed to pop my head through a couple of doors to see how fellow presenters like Alex Gorbachev and Christian Antognini were getting on, I was primarily focussed on my laptop and making sure that my stats presentation was ready to go. Because I was trying a slightly unusual introduction involving apologies for talking to my mates, thanks for all the help I'd got and introducing Paddington Bear to the audience, I needed to think about some non-technical issues too. The main problem with the presentation was that there was probably too much repetition of the previous year's presentation on the same subject, but the evidence I keep seeing suggests that I probably need to keep banging on about partition stats for a while yet!

As I finished and slinked off for my smoke, Maria Colgan suggested we have a quick word afterwards and my initial (and vocal) reaction was a humorous but despairing 'Oh, what *now*?'. i.e. What mistake had I made this time? LOL

For the record, she pointed out that the GRANULARITY parameter value required to maintain Incremental Stats does not *have* to be AUTO, but *must* include Global Stats. e.g. ALL or GLOBAL AND PARTITION. I've not tested this properly yet, but it is a useful clarification because I might want to choose to only gather stats on a single partition whilst maintaining the Incremental Global Stats (by using GLOBAL AND PARTITION), rather than leaving it to DBMS_STATS to determine which Partitions need stats gathered on them. (Oh, and if I've *still* got this wrong, I'm sure she'll be along to correct me imminently!)

At least with the first presentation out of the way, I was able to catch a couple of more sessions in between worrying about my demos. First was Tanel Poder - Exadata Performance Method - during which Tanel talked a lot of sense. Simply installing your existing applications on to Exadata might make them go two or three times quicker, based on a simple hardware refresh but as most people will have heard by now - if you're not benefitting from Smart Scan properly then don't expect to see the blistering performance that you've heard people rave about.

My final session of the day was Wolfgang Breitling talking about his SQL Tuning Kit: A Guide To Diagnosing Performance Issues. As someone who probably uses fewer performance analysis scripts than I used to (which can be both a good or a bad thing), I was interested in comparing the particular data Wolfgang was looking at and whether that information was available in OEM these days and I'd say that in most cases, all of the useful stuff is and that I personally find it quicker *these days* than a bunch of scripts. However, there's definitely still a place for useful scripts here and there, Wolfgang had a good box-full and, as we were discussing later, his don't require the Tuning and Diagnostics Pack licenses or anything more than sqlplus, perl and maybe a little Java. (I've had this conversation several times recently though - take-up of Diagnostics and Tuning seems to be pretty wide at the sites I and my fellow UK consultants/contractors work at these days so this is less of an issue than some people would have you believe.) Good presentation, though, even if I should have worn my glasses to see the tiny writing!

The evening ended with a combination of good drinks and conversation and the regular Symposium Party Night. Terrific fun as always although I'd suggest that this photo illustrates

- I'm duller than I thought I was - playing around with my phone when there's supposed to be a party on. Probably tweeting
- Jonathan Lewis is clearly duller still as I've decided I can't be bothered with this chat. (Or maybe he can't be bothered with mine?)
- Free glasses of wine tend to be fuller than the paid-for variety! Thanks sponsors!
- I'm a geek really - check out my Sinclair ZX81 T-Shirt. Thanks Mads!

Ultimately, though, it's difficult to enjoy a party fully when you have demos to worry about the next day

WEDNESDAY

Wednesday morning was an utter write-off for me because I was so nervous about how my OEM 12c demo would go but determined not to screw it up after the car-crash that was my previous attempt at an OEM presentation at the Symposium. (Note that I've provided a link to my *post* about it, not the *video* that exists! In fairness, Marco Gralike did manage to catch the small amount of that hour when the presentations actually worked.) I walked through all of the demos time and again in my room and captured screen-shots that I could show people instead if it went wrong again but it really ruins what I'm trying to show if I can't use an interactive seat-of-my-pants approach. I was also determined to assuage for my poor first attempt so I tried to cover all of the angles, including the specific one of making sure I was performing the demos with my wireless net connection *disabled*

Imagine my sinking heart when, even allowing myself 15 or 20 minutes to setup (I love lunch-breaks! Just not lunch), my carefully prepared laptop VM wouldn't drive the projector properly. Going through a whole restart procedure solved the problem but it was a nerve-wracking process whilst a growing audience watched on. Checking later, it seemed that quite a few of the big audience had attended the previous minor disaster, so I reckon they were all sadists

In the end it didn't all go perfectly and there are areas I could and will improve but it was such a relief to get through the whole experience in one piece and hopefully show people some new, cool and useful stuff with some sensible suggestions on getting the best use from ASH Analytics. There were certainly lots of people who were very complimentary afterwards and very excited about using OEM 12c but, as I think I said to every one of them, it's the new tools that are the star here. Giving presentations on cool subjects that you really believe in is the most fun. So thanks to JB et. al. at Oracle for coming up with another winner.

As I've said many times, watching any presentation immediately before or after one of my own is something I really struggle with, either through nerves on the front-end or post-presentation adrenalin come-down afterwards, but I simply had to catch Paul Matuszyk talking about Oracle 11g Extended Statistics. There's a bit of a story behind this. Paul and I used to work for a certain large satellite television company in Scotland and when I went off to present at Hotsos, he was really interested. Being independent contractors we moved on and lost touch a little but then he showed up at the Symposium one year because he'd always wanted to go after what I'd told him about it. (Note to certain User Groups - maybe you gain more new attendees from speaker recommendations than you might think! I know I've encouraged a whole bunch of people who have later turned up at conferences ...) But he was particularly interested in presenting and after what I think was one failed submission attempt, he was presenting this year! (Note to people thinking of submitting abstracts - the Hotsos Symposium Agenda is not *just* about existing speakers!)

His presentation on Extended Statistics contained some really detailed and useful information and, sitting next to Maria Colgan throughout, it all seemed accurate too. He was perhaps a little nervous as it was his first Hotsos presentation, but he did an amazing job and I'm glad I got to see it!

Which left time for just one more presentation - Kerry Osborne on Real World Exadata. Initially I was a little disappointed when Kerry pointed out that instead of the customer performance case studies he'd planned to present on, he was going to discuss an informal survey he'd conducted looking at the usage patterns of Enkitec's many Exadata customers. As it turned out, I found the presentation utterly fascinating, enjoyable and a really good break from the usual technical presentations. I can't remember all the detailed numbers but, as usual, I was surprised by just how many customers are buying quarter racks and how many Exadata implementations are for systems that aren't classic Data Warehouse or Reporting systems. There was a debate later about whether optimizer_index_cost_adj has any place in the 21st century (and I have a plan to blog about this soon) but the bulk of the presentation was just sensible real world stuff that I've come to expect from Kerry. Nice way to wrap up the conference.

Well, kind of. As usual, the Symposium is a conference that's still just about small enough to have a final short Farewell session before heading off for some relaxed Mexican food and cocktails with lots of friends. Great end to the main conference.

THURSDAY

Although the weather had been a bit hit-and-miss earlier in the week, Thursday dawned to extremely strong wind and rain. I was quite taken aback (having never experienced anything like it in March at Hotsos), although others were there 'the year that it snowed'! It softened the blow a little to hear that Texas was in the midst of a bad drought and needed the water and to realise that I was likely to spend most of the day indoors listening to Jonathan Lewis' Training Day, catching up with things I'd missed whilst not in the office or catching up with some sleep. Man, I was *tired* and only managed to last until about 2pm!

Unsurprisingly, I really enjoyed what I saw though and the good thing about the Training Day is all of the material is printed to take away so I was able to review anything that I'd missed later on. There were a lot of topics covered in very quick succession, so it was a lot to take in! Good stuff.

The evening was spent catching up on email, finally spending a bit more time with a few friends and slowly wading through tons of mail But it was good to catch my breath and have one more nights sleep before the long flight home.

Conclusions

Although I hadn't originally planned to attend and only stepped in at the last minute to fill in for Randolf Geist, I enjoyed the Symposium as much as usual which is largely to do with the perfectly sized number of attendees and lots of friends in attendance and the brilliant organisation skills of all the Hotsos team, particulaly Rhonda and Becky! The agenda was as excellent as always and so Cary Millsap deserves a note of thanks for helping to keep the standard high. As usual, thanks to the Oracle ACE Director program at OTN for helping me to travel over there and a final note of thanks to all those who spent time listening to my presentations. I hope you got something useful from them.

Usual disclosure: My travel and accommodation expenses were covered by the Oracle ACE Director program.

Vote Peter Robson!

dougburns@yahoo.com (Doug Burns) — Thu, 01 Mar 2012 14:18:13 PST

The UKOUG Council elections are in progress and I think User Groups are a critical part of the community. Peter Robson is both a friend and the type of person I admire and want to represent me within the UKOUG but, sadly, the grumpy old sod is possibly the least Web 2.0 chap I know so, concerned that his campaigning efforts might be hindered, I asked him if he'd like to make his pitch here. Over to you, Peter ....

Thanks to Doug for giving me this impromptu platform for a bit more shouting about my candidacy for the Council of the UK Oracle Users Group. It’s only right, as it was he who first persuaded me to stand for the Board (as it then was) all those years ago.

So if you are not a member of the UKOUG, I suppose you’d better stop reading now, unless I can persuade you to join up! My pitch here is to first of all, persuade you, as a member of UKOUG, to at least vote, and then secondly, hopefully convince you that I should be one of your five votes.

I have been actively involved in the UKOUG for about eight years now, both as a director of the original Board, and as an active member of the Scottish User Group committee. Now that I’m retired, I have the time to spend on these things, and my word, it can eat up one’s time, make no mistake! But it is so worthwhile. Time after time we have found that the one thing which members value are the opportunities to get together and talk about their work. Just ask Doug – he always turns up to the Birmingham Conference. Indeed, he is so obsessed with networking that he will even travel to San Francisco for Oracle OpenWorld! We can’t rival OOW, but our conferences, and SIGs, do provide the best example for the Oracle crowd to network here in the UK.

Personally, I have been able to bring lots of experience to input towards the organisation of these events by virtue of having presented in most of the European User Group conferences, as well as a few in North America, not to mention the Chris Date seminars that I also organise across the UK and Europe. Close to my heart is the Scottish community, which languished for many years, but is now thriving. Our annual conference is June 13 – Tom Kyte is presenting – make a note and come up to Scotland!

You can read the usual huff and puff about me on the election pages, but now I want to say something about the sort of person that you should look to elect to the Council. Most of all we need passionate, committed individuals, and not people who think they might try it for the benefit of their CV. We need people with the time to devote to the business of UKOUG, people who REALLY have the time to give. This is so important, as not only do we have a large user base to offer a service to, but we have a large office staff of salaried people who depend on us for their livelihood. No way are the office staff coasting – they are all stretched, and give of their very best. Indeed, it has been an absolute pleasure to work with them over the past years. We can only do our very best for them, which is why I stress that we need committed members of Council with sufficient time to devote to the job. With such people, and the skills and expertise of the office staff, I see no reason why UKOUG cannot thrive and grow.

Thanks for reading this stuff – did I ask you to vote for me? Yeah, go on, I would really appreciate that!

Peter Robson

Hotsos Symposium 2012

dougburns@yahoo.com (Doug Burns) — Tue, 07 Feb 2012 13:37:00 PST

Oh, well ... having decided that I was going to skip the Symposium this year, everything changed.

My friend Randolf Geist had to cancel his attendance so when he saw my previous post he asked if I'd be prepared to step in with a couple of presentations so that he wouldn't feel quite so bad about letting people down. I asked for a little while to think about it (because all of my original reasons for not attending still existed) but in the end was happy to help out.

To be crystal clear, Randolf offered this option to the Hotsos folk because it can be difficult lining up replacement speakers so late in the day and they decided that they wanted to go with it.

In other words, before people start slagging me off on Twitter or whatever for taking agenda spots that should have gone to other speakers. (Difficult to believe, I know, but stranger things have happened.) ... I was asked by a friend to help. I said yes. We discussed whether we should offer this option and Randolf made it clear that it was just an option that they could go with or not. No skin off my nose, frankly. Anyone who is unhappy with that should perhaps take it up with those who organise the agenda because I'm sick of feeling apologetic for actions I've taken in good faith! (Sorry, I'm probably just a little touchy and grouchy at the moment because I'm ill in bed ...)

Anyway, back to a more even mood .... The page here shows the two presentations I'll be giving.

The final Early Bird Registration expires this Friday so if you plan to go, now would be a good time to register.

Hopefully see you there.

Randolf Geist on 11g Incremental Statistics

dougburns@yahoo.com (Doug Burns) — Mon, 16 Jan 2012 23:12:00 PST

Well it wasn't the post I planned to return to technical matters with.

Lots of readers here have asked me when I'm going to get round to writing about 11g Incremental Statistics as part of the stats series. Although Incrementals are on my To Do list, I wanted to finish off the stats copying posts first. In any case, Randolf Geist got there already so I'll cross it off my list and point you towards his post instead.

Yes, I know there have been a lot of Incrementals posts already by people like Robin Moffat and John Hallas, but Randolf's post maps most closely on to the post I planned, which is an overview of Incrementals that highlights some of the practicalities of using them in "The Real World". I'd particularly draw attention to a couple of aspects which I think people keep misunderstanding.

The first time you implement Incrementals on a table, Oracle will have to trawl through the entire table in order to build the initial synposes. This has always seemed obvious to me - how can you incrementally build on synposes that haven't been created yet? But the long duration initial gather seems to surprise people and they decide that Incrementals are 'slow'.
Incrementals are a replacement for GRANULARITY=>'GLOBAL AND PARTITION' and not 'PARTITION'! Expecting an option which gathers Partition stats and then goes around updating synposes to perform as well as a simple partition gather is unrealistic^*. Any performance improvement needs to be measured against both gathering the Partition stats and maintaining the Global stats. Incrementals will almost definitely be quicker than that! I prefer to think of Incrementals not so much as a performance improvement (because most people probably didn't regather Global statistics every time they gathered individual Partition statistics because they didn't have the time on an active system), but an improvement to the quality of your Global stats because you can now afford to maintain them with the same frequency as your Partition stats, rather than scheduling an out-of-hours Global stats gather or depending on the inaccurate NDVs that result from the previous aggregation process.

Good post, anyway. Thanks Randolf!
^* However, it's fair to say that Oracle have continued trying to improve the performance of synposis maintenance.

Upcoming Conferences

dougburns@yahoo.com (Doug Burns) — Sun, 15 Jan 2012 21:45:00 PST

At this time of year, I'm usually panicking about my Hotsos presentation/paper and up to my eyeballs in test scripts. I enjoy it and the absence feels a little strange, but I decided to do something different this year.

Although it's the 10th Symposium and I'm sure it will be as educational and fun as ever, I decided not to submit an abstract for a variety of reasons, one of which was the cluster of conferences that occur during the spring. Although the Oracle ACE Director program helps with travel expenses, the rest is self-funded including the time off work so I need to pick and choose my conferences carefully. For years Hotsos, Openworld and the UKOUG were fixtures but that means I've missed out on the wonders of Miracle spring events and the various Scandinavian boat trips!

Of course, almost as soon as I decided I wouldn't be going to Dallas the regret started to set in as there are people who I often only see at Hotsos who I won't be able to catch up with thise year and I think some of my best presentations have been there. It got worse when I looked at the list of confirmed speakers and spotted an old friends name. Sigh, having encouraged Paul Matuszyk to attend the Symposium while we were working together in Scotland he enjoyed attending a few times and decided he wanted to present there one day. Now that he has been accepted, I won't be there to see it (By the way, this reinforces a point I've made to several people. Although you will see a lot of familiar faces presenting, there's always been space for new presenters as far as I'm aware. Mark Rittman is another 'new' speaker there this year.)

However, now that the decision's been made, I need to let it go and move on to those conferences I hope to attend instead.

I had a terrific time at the OUGN (Norwegian User Group) spring conference on the boat last year so I was delighted when they invited me back to speak at this year's event.

My presentations there will be an existing one from OOW and UKOUG and a new one ...

Performance and Stability with Oracle 11g SQL Plan Management

This presentation discusses the use of Oracle 11g SQL Plan Management to safely upgrade a complex and critical Data Warehouse system, ensuring stable performance whilst taking full advantage of optimizer improvements in Oracle Database 11gR2.

Subjects covered include -

SQL Plan Management fundamentals
Controlling execution plan evolution
Using SQL Plan Management to resolve complex performance challenges
The interaction between bind variables, histograms, Adaptive Cursor Sharing and SQL Plan Management.

The presentation will focus on real world experiences of both the strengths and limitations of SPM.

Falling in love all over again - OEM 12c Performance Page Enhancements
Oracle RDBMS 10g introduced significant instrumentation enhancements and the OEM 10g pages illustrated them to make performance analysis much easier. OEM 12c is the latest evolutionary stage of the Performance Pages and this
presentation shows the new features that make them even more compelling and useful in more situations.

I'm also hoping to attend Miracle Openworld because it's a long time since I was in Denmark and I realised last year just how much I've missed this conference. I submitted the Statistics on Partitioned Objects abstract because I think it would be good for that particular conference and because it's still what I'm spending most of my time on in the office at the moment. Having said that, I haven't heard anything back yet and it looks like the agenda already has some pretty interesting presentations on there, so I might not make the cut this time.

Still tempted to go, though