Search

OakieTags

Who's online

There are currently 0 users and 41 guests online.

Recent comments

network

Importance of Network to the Cloud

I’ve been discussing for years about the importance of network to database performance, especially once I started working on VLDBs, (Very Large Databases) but its a topic that often is disregarded.  Now that I’m working more and more in the cloud, it’s become more evident the importance of the network to our survival.

NFS Performance Issues at TCP level

What happens with I/O requests over NFS and more specifically with Oracle? How does NFS affect performance and what things can be done to improve performance?

What happens at the TCP layer when I request with dd an 8K chunk of data off an NFS mounted file system?

Here is one example:

I do a

dd if=/dev/zero of=foo bs=8k count=1

where my output file is on an NFS mount, I see the TCP send and receives from NFS server to client as:

(the code is in dtrace and runs on the server side, see #2970a6;" href="https://sites.google.com/site/oraclemonitor/tcp-d">tcp.d for the code)

#2970a6;" href="http://dboptimizer.com/wp-content/uploads/2011/04/nfs3.png">

[Oracle] SQL*Net researching - Setting Session Data Unit (SDU) size and how it can go wrong

Introduction

I come across this SDU issue from time to time by doing oracle database consulting work and so i think it is worth to write a blog about it. Basically this blog is about how to verify that the SDU setting is considered (or not) at all. It also covers some basics and SAP suggestions.

 

Lets' start with an explanation and some SAP information about the SDU, before we go on with researching and testing several settings.

 

OSP #2c: Build a Standard Platform from the Bottom-Up

This is the fourth of twelve articles in a series called Operationally Scalable Practices. The first article gives an introduction and the second article contains a general overview. In short, this series suggests a comprehensive and cogent blueprint to best position organizations and DBAs for growth.

OSP #2c: Build a Standard Platform from the Bottom-Up

This is the fourth of twelve articles in a series called Operationally Scalable Practices. The first article gives an introduction and the second article contains a general overview. In short, this series suggests a comprehensive and cogent blueprint to best position organizations and DBAs for growth.

OSP #2c: Build a Standard Platform from the Bottom-Up

This is the fourth of twelve articles in a series called Operationally Scalable Practices. The first article gives an introduction and the second article contains a general overview. In short, this series suggests a comprehensive and cogent blueprint to best position organizations and DBAs for growth.

TCP Trace Analysis for NFS

How do we know where latency comes from when  there is a disparity in reported I/O latency on  the I/O subsystem and that of the latency reported on the  client box requesting the I/O.

For example if I have an Oracle database requesting I/O  and Oracle says an 8Kb request takes 50 ms yet the I/O storage subsystem says 8Kb I/Os are taking 1ms (averages) , then where does the 49  extra ms come from?

When the I/O subsystem is connected to Oracle via NFS  then there are a lot of layers that could be causing the extra latency.

Screen Shot 2013-08-23 at 1.35.20 PM

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

OakTable…

I was recently nominated and approved as a member of the OakTable Network .

Do you ever get that feeling that one day people are going to realize you don’t have a clue what you are talking about? I think that day just got a little closer. :)

Cheers

Tim…

unplumb (or unbinding) NICs on Linux

I’ve been quiet for a long time now, but this entry hopefully will shake the cobwebs off and get me back into the habit.

I recently had a need to “unplumb” (from Solaris fame) or make interfaces on Linux “disappear” from the ifconfig list. It could be that I don’t know how to completely deconfigure an interface, but I didn’t find any methods to unassign an IP address from a Linux Ethernet interface after it was assigned. You can take interfaces down (ifconfig eth3 down) and reconfigure them to assign different addresses, but not remove the address completely.

Fetch as Much as You Can

In my “Everything DBAs Need to Know About TCP/IP” presentation, I discussed the issue of fetch size – tuning the number of rows your application will get from Oracle in a single “fetch” call.

The problem is interesting when you want to get a large number of rows from the DB to the application. Cary Millsap would say that no one should want to do that, but Cary obviously never had to work with statistical analysis application. These things demand large amounts of blood raw data to do their thing.

If this is too slow for your users, you should look at statistics like “round trips”. If the number of round-trips looks completely off the wall, like 200,000 round trips for 2M rows, someone might have left JDBC’s default fetch size of 10.