I got a look a new prototype for the next generation Exadata last week while doing some work with a company in Europe. Apparently there is a big push to be environmentally friendly there now and so Oracle is trying to come up with a model that uses less power and is biodegradable. The word on the street is that it won’t be available until after release 2 of the 12c database.
The new model has a few drawbacks though. For one thing, it only lasts a few weeks before you must either replace it or higher some rocket surgeon consultants to come in and tune it. From the early version of the prototype I saw, it does appear to be smaller and more tasty than previous models though.
That’s a picture of the lead designer (JP) showing off the prototype. The code name for the project is “Exanana” by the way. The new model should be available in select supermarkets after lunch (err launch). Here’s another picture of JP and one of the other designers (Paul) hamming it up for the camera.
I probably should have saved this post for April 1st!
I know no one really likes the term “tuning” these days, but it’s a short catchy word that gets the idea across. So I’ll just stick with it for the title of this post.
Note that this is one of those posts that’s not really supposed to be about how to solve a particular problem. It’s really just a story about a distraction that I ran into and I how I thought about getting around the issue and then ultimately resolving the root cause. Maybe you will find it instructive to see the process.
So I have this script that I use occasionally (paramon.sql) to see what parallel query slaves are doing. Unfortunately the script doesn’t have a header in it, but I’m pretty sure I lifted it from Randolf Geist. I can’t find it on his blog anywhere, but it looks like his style of writing SQL, and PX Query is something he’s written a lot about, so I’m pretty sure that’s where I got it. (Update: see Jonathan Lewis’s comment below attributing the script to Andy Brooker) Anyway, the script has worked great for me in the past but I recently noticed that it was really sluggish on a couple of 11gR2 DB’s running on Exadata. Here’s an example:
-bash-3.2$ !sql
sqlp
SQL*Plus: Release 11.2.0.3.0 Production on Mon Jan 14 12:23:15 2013
Copyright (c) 1982, 2011, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,
Data Mining and Real Application Testing options
INSTANCE_NAME STARTUP_TIME CURRENT_TIME DAYS SECONDS
---------------- -------------------------- -------------------------- ------- ----------
dbm1 09-JAN-2013 03:25 14-JAN-2013 12:23 5.37 464246
SYS@dbm1> set SQLPROMPT "11.2.0.3> "
11.2.0.3>
11.2.0.3> @paramon
Enter value for status:
Node Name Status Pid Sid Parent OSUSER Schema CHILD_WAIT PARENT_WAIT
----- ---- ---------- ----- ----- ------ ------------------------------ ---------- ------------------------------ ------------------------------
P000 AVAILABLE 35
P001 AVAILABLE 36
P002 AVAILABLE 37
P003 AVAILABLE 38
P004 AVAILABLE 39
P005 AVAILABLE 40
P006 AVAILABLE 41
P007 AVAILABLE 42
P008 AVAILABLE 43
P009 AVAILABLE 44
P010 AVAILABLE 45
P011 AVAILABLE 46
P012 AVAILABLE 47
P013 AVAILABLE 48
P014 AVAILABLE 49
P015 AVAILABLE 50
P016 AVAILABLE 51
P017 AVAILABLE 52
P018 AVAILABLE 53
P019 AVAILABLE 54
P020 AVAILABLE 55
P021 AVAILABLE 56
P022 AVAILABLE 57
P023 AVAILABLE 58
P024 AVAILABLE 59
P025 AVAILABLE 60
P026 AVAILABLE 61
P027 AVAILABLE 62
P028 AVAILABLE 63
P029 AVAILABLE 64
P030 AVAILABLE 65
P031 AVAILABLE 66
32 rows selected.
Elapsed: 00:00:23.11
So on this 11g DB it took 23 seconds to run the query. On one of my 10g DB’s though the performance was stellar.
Just a quick note to let you know I’ll be speaking at the Hotsos Symposium in March in Dallas. I’ve attended every year for the past 6 or 7 years and spoken at several of them. It has consistently been the best performance oriented Oracle conference I’ve attended. This year will be no different with the likes of Maria Colgan, Karen Morton, Tom Kyte, Tim Gorman, Cary Millsap, Frits Hoogland, Gwen Shapira, Alex Gorbachev, Kyle Haily, Carlos Sierra, Kellyn Pot’Vin, Mark Farnham, Toon Koppelaars, Andy Zitelli, Neil Gunther, Stephan Haisley, Marco Gralike, Steven Feuerstein and a host of others. Looks like several of the speakers (including me) plan to be talking about new performance oriented features of the soon to be released Oracle 12c database, so it should be very interesting. One of the best things about the conference is the chance to talk to people (including the speakers) at the breaks. And by the way, while the speaker list is impressive, there are always a large number of highly talented people in attendance that are not speaking. I routinely learn as much from conversations between the sessions as I do listening to the presentations. I highly recommend the symposium to anyone that is interested in Oracle performance. Here’s a link to the main page where you can find the complete list of speakers and their topics and register for the conference.
Just a quick note to let you know that I updated the presentations section of my blog. Most of the talks I’ve done over the last 3 years, including Hotsos Symposium, RMOUG Training Days, Oracle Open World, UKOUG Conference and Enkitec Extreme Exadata Expo (E4) are now on the site.
Just a quick note to point out that the official white paper version of the TCO study I talked about in a previous post (Exadata vs. IBM P-Series) is now available on the Oracle web site here:
The most interesting part of the study (in my opinion) is the quotes from the participants that are using Exadata. These quotes provide some insight into how people feel about the platform after having systems in production for a while (note that all the interviews were done prior to the release of the X3-2). I should also point out that these customers are not typical Oracle reference customers. They were interviewed by a third party and promised anonymity as part of the study.
Here’s a link to a video of Mogens Norgaard discussing the upcoming OakTable World conference (click on the picture of Mogens below). Note that the name of the conference has been changed from Oracle Closed World to OakTable World.
Fellow EnkiteciesKarl Arao and Tanel Poder will both be presenting at the event along with a bunch of other well known OakTable members. I may even do a 10-15 minute TED style talk at lunch time on Monday. Hope to see you there.
It’s been about two and a half years since Enkitec took delivery of our first Exadata. (I blogged about it here: Weasle Stomping Day) Getting our hands on Exadata was very cool for all of us geeks. A lot has changed since then, but we’re still a bunch of geeks at heart and so this week we indulged our geekdom once again with the delivery of our Big Data Appliance (BDA). In case you haven’t heard about it, Oracle has released an engineered system that is designed to host “Big Data” (which is not my favorite term, but I’ll have to save that for some other time). The Hadoop ecosystem has taken off in the last couple of years and this is Oracle’s initial foray into the arena. The BDA comes loaded with 18 servers, each sporting 36 Terabytes of storage for a whopping total of 648 Terabytes. It also comes with Cloudera’s distribution of Hadoop (and software from various other open source projects that are part of the Hadoop ecosystem). We’re very excited to start working with Cloudera and Oracle in this exciting new approach to managing large data sets. Anyway here’s a few pictures:
The rack’s pretty heavy with all the disk drives. One of the delivery guys said he had a full rack of EMC drives that actually fell through the floor of the office building they delivered it to (no one was hurt). Fortunately we didn’t have any mishaps. And at a couple of thousand pounds, we will not be moving it around to see how it looks next to the coffee table (like we do with slightly less heavy pieces of furniture at home).
This is me and my buddy Pete Cassidy (Oracle Instructor Extraordinaire) messing around.
Another picture of me and Pete. Not as good as the other one, but I love the shoes!
Tim Fox loves the big power cables.
The BDA also has a handy beer shelf (this is the top secret new feature).
The BDA cabinet has a lock and of course the key’s were in a well label plastic bag. I had Andy Colvin hold up the label so I could take his picture. I called the shot “DoorKeyAndy”. – seemed appropriate 😉
The 2012 EnkitecExtreme Exadata Expo is behind us now. Our video guy (Bob) has been working diligently for the last week or so to get the presentations edited. They will be made available to the attendees shortly. We have already posted a video of the opening session. It is me interviewing Cary Millsap about his impressions of Exadata. One of the things I have found most interesting about Exadata is how it makes very experienced Oracle performance guys re-think things. It’s fun watching them being exposed to Exadata in an intimate way (not just Power Point). The reactions are interesting. There is usually a desire to try to break it although it’s generally harder than it appears, at least initially. I got to watch Cary for a few days (along with several of his Method-R guys) when he had his first exposure and we talk about that during the video. Cary made a profound impact on me and my career and I think his insight into the Exadata architecture is worth watching. Hopefully the fact that I have a great face for radio will not deter you (too much). The video has the obligatory Enkitec logo and endless loop, non-descript music intro, but other than that, I certify it as marketing free! 😉
Be sure and check out enkitec.tv for other videos too.
Earlier this year I participated in a Total Cost of Ownership (TCO) study that was run by a company called the FactPoint Group. The goal was to compare the cost of purchasing and running Oracle on Exadata vs. the cost of purchasing and running Oracle on IBM P-Series hardware. The findings are published here:
Fair warning, the study was funded by Oracle and the first 10 minutes are a sales pitch for Exadata, but the data collected by FactPoint for the TCO study fills the rest of the presentation.. That part was very interesting (to me anyway). They interviewed people from 5 companies which had been running in production on Exadata for at least 6 months and 5 companies that had production systems running on IBM P-Series machines. As a bonus, 2 of the Exadata customers had previously run their systems on IBM P-Series hardware or currently had other systems running on P-Series machines.
The TCO calculations were based on equivalent systems from a CPU count basis. That is, they priced out an IBM P7 system that had the same number of cores as an Exadata Half Rack and used that for the calculations. I didn’t think that was really a fair comparison since the performance of the two systems was vastly different. In fact, one of the customers that had moved from IBM to an Exadata had the following to say:
“This data warehouse was originally on 12 a P570 and the nightly load took 5 hours. With Exadata it takes 30 minutes with the Exadata 1⁄4 rack running at only 5% CPU utilization. Weekly stats updating went from 60 hours to 35 minutes. Backups reduced from 14 hours to 45 minutes.”
This is only one of many quotes from the customers included in the presentation. The quotes were quite revealing, and to my way of thinking, were the most interesting part of the presentation. Another oddity of the cost calculations was that they included RAC licenses in the Exadata costs while not including RAC in the IBM costs. Again I felt that this was an unfair comparison as RAC provided HA capabilities to the Exadata platform that the IBM platform simple didn’t have. Nevertheless, the study found that Exadata was about 35% less expensive. I wonder what it would have been if they had tried to cost an IBM system that performed as well as an Exadata and provided the same HA capabilities.
Just a quick update on my speaking schedule for the rest of the year (I hope this is all).
Big Data Technology Briefing – Texas Rangers Stadium (Arlington, TX) – 8/23/12
I’ll be talking about Hadoopie stuff.
Amazon Commit – Seattle, WA – 9/24/12
I’ll be talking about Controlling Execution Plans and also about Creative Problem Solving.
Oracle Open World – San Francisco – 10/2/12
I’ll be talking about Hadoop and Exadata.
UKOUG – Birmingham, England – 12/4/12
I’ll be talking about Controlling Execution Plans (again).
Presumably 12c DB will be released by then so some brand new stuff should be in this one.
Oracle DW Global Leaders Forum – Dallas – 12/13/12
I saw a flyer saying I was talking about “Big Data”.
