So the annual UKOUG technology conference has come and gone for yet another year. This time it was in a new location, having moved away from its regular berth in Birmingham. Manchester is not a city I’m that familiar with, the only previous time I had been there was a trip with my wife, and she drove me to a near death experience with an oncoming tram.
Thankfully this time the only near death experiences for me this year was laptop failure during a demo, though I did hear of some folks getting into a scrape or two.
Yes, there were less people about than previous years, but that can be explained by the Apps folks having split off to their own conference. The venue was I thought pretty reasonable, while some rooms were pretty small, as lots of people commented, better presenting in a full small room than a fairly empty large room. So for me, the venue worked fine. Manchester is likely to be harder for more people to get to than Birmingham though, and I will be very interested which city the conference will be in next year, I understand it will be moving again.
So onto my experience at the conference. Certainly, my conference experience has changed a lot over the past few years, and the meeting people aspect has come more to the fore. I had a strange conference in terms of sessions I attended: I barely attended a database session! It was all storage/IO or operating systems. In some ways, you could look at this as a bit alarming, as I’m not sure the value add is necessarily at that level of the stack, however for me it’s really where my career has been for most of the time.
The quality of the presentations I did see was outstanding, and that for me is a critical thing of the UKOUG Annual Conference: the quality of the presentations. I don’t think I saw a bad one. The highlights for me where the following three though:
Luca’s work is just awesome, and he has developed some latency visualisations tools, which were very interesting to see.
Round tables can either be really eye-opening or fairly uninteresting. It all depends on who is contributing what. This roundtable was in the eye-opening camp. Joel Goodman chaired this excellently and it was great to see the only 12c Grid Infrastructure implementation in the room was one done by e-dba and my colleague Svetoslav Gyurov.
This was actually given at Oak Table World running alongside the UKOUG conference. This was an excellent, eye-opening presentation on hadoop.
The other thing that is important to me at conferences, compared to say 6-7 years ago is presenting. I’m not sure I’d like to go to UKOUG and not present. Sadly, this year due to other commitments from colleagues, I ended up picking up 4 presentations. This is too much for me to focus on at once. Thankfully though, 3 of them were on Exadata, and one of those a panel discussion. But 4 is quite a stressful amount.
So, some thoughts on my presentations.
This is a presentation I have delivered quite a few times and am very familiar with the content, thankfully this went well and I was reasonably happy with it in the delivery, and I was able to add some stuff on the (now) newly available Exadata X4′s to keep it fresh! You can grab this presentation if you are interested. Be aware there is lots of text to read in the notes field, even if the slide is fairly minimal.
Next was 2 on Tuesday:
I stepped in at the last minute for this one, and did 20 minutes on Exadata. I did not like this presentation – it was too much of a cut ‘n’ shut job. One attendee complained afterwards that there was not any content on actually using Exalytics with Exadata.
I couldn’t fail to agree.
So, I had not heard of Linux Containers until about 2 weeks prior to this presentation, but it turned out this was the one I had most been looking forward to. I’d done the most work leading up to UKOUG on this presentation as I had to learn it from scratch in the 2 weeks (as well as the other presentations, and the day-to-day job!). This was meant to show what Containers were, why you might use them and then demo them in action. I really thought the slides looked cracking (someone please take me aside and have a word if I’m out to lunch on this), and though I was hesitant about the content at times, I thought I’d almost pulled it off.
Then the demos, which at first were going fine, started to kill my laptop. and I mean kill. Not just the demos not working, but the laptop being totally utterly unresponsive, so I could not even get back to powerpoint from my VM. It was horrible. The presentation, though about 80% done, just came to an abrupt car crash of an ending. It took about 15 minutes after this to even power of my laptop!
As Tom Kyte put it the next day:
Image courtesy of Marc Fielding.
Both Andy and Frits are awesome and it was a privilege and a pleasure to share the stage with them. We had an excellent discussion and lots of interaction with the audience which I think made it a really worthwhile hour.
I’d love to do something like that again!
Again, image courtesy of Marc Fielding.
It was a great conference, I really had a great time, and I feel lucky and indeed proud to be part of a great Oracle community and to know so many outstanding individuals.
I’ve been using OS X for a very long time, and one of the Applications that is invaluable in my day to day work is VirtualBox. It’s a great application allowing you to work with various Oracle versions and products, and I have quite a few linux VMs with differing DB versions inside them.
I’m also a bit rash when it comes to upgrading my OS X version, and with OS X Mavericks released on the 22nd of October and best of all FREE, it was a hard combination to resist. Well, resist I did for all of a morning! As I had recently upgraded to VirtualBox 4.30 I was quietly confident there would be no issues….
Oh boy, I often get into trouble with my sunny optimism.
The first issue you might hit with VirtualBox 4.30 on OS X 10.9, could be a “kernel driver not loaded” has actually been well covered in this forum posting. The script works quite well at loading drivers.
However, when I did this, I was still getting this error pop up, and no VMs starting:
This was not good. Now, I then reinstalled VirtualBox 4.30 and voila I could start a VM again. Happy days. Then, as this was on my desktop, I close it down for the evening, and come back the next day. Same issue as above, no VM’s will start. Grrrr. Once again a reinstall fixes the issue.
I knew something was getting loaded differently at boot time, than when VirtualBox was freshly installed. The forum posting above gave me a clue as to what it might be. So it was kextstat to the rescue!
kexstat will show you what kernel extensions you have running. Just like the linux lsmod command.
So I ran kextstat after a fresh a reboot:
jarneil> kexstat . . 118 3 0xffffff7f8226b000 0x43000 0x43000 org.virtualbox.kext.VBoxDrv (4.2.18) 119 0 0xffffff7f822ae000 0x5000 0x5000 org.virtualbox.kext.VBoxNetFlt (4.2.18) 120 0 0xffffff7f822b3000 0x6000 0x6000 org.virtualbox.kext.VBoxNetAdp (4.2.18) 121 0 0xffffff7f822b9000 0x8000 0x8000 org.virtualbox.kext.VBoxUSB (4.2.18) . .
So among a whole bunch of other kernel extensions there were the 4 VirtualBox modules and all loaded from the previous version I had been running before upgrading to 4.30 which was *days* before I had upgraded to OS X 10.9!
Now it was just a case of making sure this lot did not load again, in fact they clearly were not removed properly upon upgrading from 4.2.18 to 4.30.
I used locate to find out these 4.2.18 kext were being loaded from /Library/Extensions there was a directory for each kext above in there. If you read the Info.plist file for them it was clear these were the 4.2.18 versions.
After doing rm -rf on these old VirtualBox directories in the /Library/Extensions directory I can now happily reboot and *still* have my VM’s working!
I’ve just come back from my 5th trip to Oracle OpenWorld. While there is something very special about your first trip to OpenWorld, this has been my favourite trip of the lot. In previous years I’ve blogged about what’s been going on every day, but for me, twitter and @jarneil is where I tend to be more active these days, but I thought I’d give a general overview of my OpenWorld experiences this year.
I was quite lucky this year to get a presentation selected “Exadata From Installation to Go Live: Experiences From Banking And Retail”. I was delighted by the number of attendees who turned up to it, well over 150 I reckon. while I have presented many times at various UKOUG events over the year, but I feel it’s slightly different presenting in the US, the attendees were very keen to ask questions and many came up to chat at the end. It certainly makes it all feel worthwhile as a presenter, and I’ll be submitting a presentation for next year for sure.
OakTable World has really become a fixed staple of my OpenWorld experience, a conference within a conference, where some of the most knowledgeable folks in the Oracle community present in a bit more technical detail than you get in the general conference. It really was tremendously well organised by @kylehailey and it was a real privelege to be able to attend this.
The big announcement in the Larry Ellison keynote was of course the new Oracle database in memory option. While it sounds like this is a long way from being production ready, it does sound like potentially an awesome feature. It did however make me wonder whether it might start to cannibalize Exadata sales: why worry if you don’t have smart scans when you are querying your hot data in memory.
On the Exadata front, the conference was a bit of damp squib, there was no new hardware announced, apparently due to the timeframe of the recent Intel chip release – it was too late for Oracle. Also there was no new Exadata Storage Server Software announced either. This really did take me by surprise the current version has been around for 9 months or so, and I really expected an update. Now, there were sneak peaks regarding what is coming and there is some good stuff in the offing, in particular compression of data in flashcache, done in hardware on the card, should essentially ensure you get more bang for your flashcache. There is also QOS for networks
To ensure offloading occurs for different versions of databases running on the compute nodes there is going to be a bit of an architectural change with the software running on the cells, with essentially an offloading process for the different database versions. It sounds like this may take a while to appear, and it’s a bit disappointing to say the least, that 12c has been released with no offloading support.
I thought I’d mention a few of my favourite presentations, in addition to the OakTable World stuff above, that I saw. First up is from Maria Colgan and Jonathan Lewis called optimizer bootcamp. It not only was highly educational, but was also superbly entertaining – definitely the outstanding presentation of the week. I also enjoyed “Oracle Exadata: What’s New, What’s Coming” by Juan Loaiza. Juan is an excellent presenter and comes across as highly knowledgeable about the product. Finally I’d like to mention “Solving the Toughest Oracle Exadata Problems by Using Tips from Oracle Development” mostly presented by Scott Heisey. This really did give you a lot of tips on where to look and what to do when something goes wrong with your exadata. If the slides for this become available, I’d highly recommend taking a look if you have to manage an Exadata box.
The Social Side
What truly made this year a memorable experience though, was the social side. After 5 years, I now know quite a few people and really it is superb being able to talk to some of the smartest folks in the Oracle community. It is this that really made it a great year: from the Oracle Ace Dinner and speaking to a whole bunch of Enkitec folks, to the pythian party and one of the best stories I heard at OpenWorld from Karl Arao, to an e-dba dinner and some great stories about Oracle UK from James Anthony and finally the great blogger meetup and speaking with Luca Canali, oh and not forgetting dinner at the Franciscan Crab to the finale with Martin Nash in the W and a couple of mojhito.