Uncategorized

I had to think of ‏@OyvindIsene, a great ambassador of the Norwegian Oracle User Group when I typed the heading for this post. Unlike him I have not actively been looking for new challenges but sometimes things just develop, and in my case that was a great turn of events. I am very happy to have signed on the dotted line and in a couple of weeks will join Enkitec in Europe.

How did that happen? During an Oracle conference I met Andy Colvin together with some of his colleagues during a break in the busy schedule. I already knew and respected Enkitec as a great company with lots of seriously experienced DBAs. I feel fortunate to actually know some of them already from email and other social media exchanges.

I’ve had my week in Palo Alto with the Delphix people. I really don’t know where the time went to – but I had a lot of interesting conversations with a number of very able people; plenty of time to experiment; and I’ve even been able to install Delphix on my laptop (as a 64-bit Open Solaris 10 VM under VMWare). I liked the product, and I was impressed with the team they’ve got working on it.

We’ve pencilled in an intial online webinar for Friday 5th April which will probably feature an informal chat where Kyle Hailey and I talk about my impressions of the product and what I’ve done with it so far. Then, in about a month’s time, we’ll have a more technical discussion.

In the meantime, it occurred to me that my next blog post ought to be about LOBs as this would be a nice way to introduce you to one of the key ideas behind DxFS (the Delphix-extended file system that’s based on ZFS).

March 20, 2013, Wednesday, 10:00 am-11:30 am PST

Speaker:   Rick Caccia, Vice President Strategy at Delphix

To Register:  Click Here

Delphix accelerates SAP project delivery by eliminating redundant infrastructure and slow processes.

Delivering application projects quickly can immediately impacts revenues and earnings.  Most IT organizations can only fund a small number of projects in a given year which limits the ability of businesses to capture these market opportunities and operational efficiencies.

I’m heading off to Heathrow airport later on today to fly out to San Francisco for my week of  experimenting with Delphix. I’ve done a little preparation work, of course, including browsing around the Internet to read about related technologies. Some of the material I found was very interesting, so I thought I’d go publish a few of the links that might be useful to other people.

It’s quite surprising to see how long the necessary core technology has been around; and yet there seems to have been minimal follow-up on the possibilities the technology  makes available – perhaps because of the specific  hardware, or special skills needed to put put together a working solution.

So we recently reviewed a table with no fewer than 23 indexes on it. A combination of alter index… monitoring usage and reviewing DBMS_HIST_SQL_PLAN for plans that referenced the indexes found 8 indexes that were never used. As this table is also heavily updated we removed the indexes identified as unused. This was, fortunately, a [...]

Thanks to everyone that came along.

The slides can be found here

by Matt Hayward

Last week I attended the Rocky Mountain Oracle User’s Group Training Days Conference in Denver, Colorado. RMOUG TD is one of the largest independent Oracle User’s Groups for Oracle database technology and draws a large number of Oracle Aces and Ace Directors, Oracle OakTable members and senior Oracle corporate technology leaders.

On Monday I attended a session by Cary Millsap from Method-R Corporation. In this full day session, Cary talked about how to use Oracle’s trace files to accurately diagnose performance problems related to specific business transactions by profiling the execution of database activity at the detailed call level. Using a variety of tools, Cary sliced and diced Oracle trace files to pinpoint problems related to disk, network and cpu performance for a variety of business transactions. Cary’s company, Method R, develops tools that make it easy to analyze trace files both from the command line as well as from within Oracle SQL Developer.

Introduction

When there is a disparity in reported I/O latency between the NFS server and Oracle database on the NFS client the question arises:

Where does the difference in latency come from between NFS Server and Oracle’s timing of pread?

One strategy is to take tcp traces on NFS server and NFS client at the same time, during a period of load that shows the disparity.

The two trace files can then be analyzed to show latency at each side and the delta of latency between packets found in both traces

Files Required

Parsing script is ( thanks to Matt Amdur from Delphix for the  core code)

https://github.com/khailey/tcpdump/blob/master/parsetcp.pl

by Matt Hayward

Given a description of gas dynamics and the atmosphere, you would be hard to pressed forecast tornadoes. The term emergence denotes the phenomena of surprising behaviors arising in complex systems.  Modern storage systems are complex, and full of emergent behavior that make forecasting application I/O performance fiendishly difficult.