2002 10 24

                    LINUX NEWS
          October 24, 2002 -- Issue #104


1) Sean’s Notes

2) Linux News

Would You Like Linux With That?
Mandrake 9.0 Review
Automating Manufacturing Processes
Linux IP Telephony to be Demonstrated

3) Linux Resources

In a Jiffy
Linux Basics
New DNS and BIND Book
How to Really Screw Up a Linux Installation
Creating a Chroot Jail

4) App o’ the Week

~~~~~~~~~~~~~~~~~~~~~~ ADVERTISEMENT ~~~~~~~~~~~~~~~~~~~~~~~

Gain study time and enhance your learning! Hear hundreds of certification exam questions on audio CD or cassette. Learn while you commute to and from work, exercise, or walk the dog. Ideal for those times when you can’t read. 90-day money-back guarantee if you are not happy.


For information on how to advertise in this newsletter
please contact mailto:adsales@CramSession.com or visit

1) Sean's Notes

Every so often I'll run into a story that gets me into the
disaster recovery frame of mind.  This time around, a water main
broke outside the wall of a company's data centre.  The
resulting water jet broke a hole in the wall, and demolished
much of their operation.  However, they had prepared for such a
disaster, and were able to stay in business.


Take it with a grain of salt, some of this company's business
seems to be supply chain, so it's a feather in their hat that
their products were so versatile.

A water main break?  Who would have guessed?  Often, we plan for
a power supply failing, maybe our ISP falling over, but rarely
do we think "what if this office weren't here tomorrow?"

This article also sparked discussion on the NANOG mailing list:


Some of the messages in the thread talk about other water-based
threats.  If you're in a multi-level building, and there is a
fire in the floor above, where is the water from the sprinklers
going to go?  Or, worse, you have a fire of your own and have to
shut down your servers?

Before we rush out and throw money at this problem, it's
probably good to do a risk assessment.  The questions I'd be
asking are:

- What is the business's tolerance to outage?  Minutes?  Hours?

Sometimes it's easy to quantify an outage in terms of $/hr, but
sometimes it's not as easy, such as dealing with human lives.

- What services are essential to offer, and how will they be
offered?  What sort of service level is expected?

If your business is taking orders over the web, is it acceptable
to have customers phone in orders while you're recovering?  If
your business has to run on computers, is a delay in service
acceptable?  If a task once responded instantly when 5000 users
used it, but took 10 seconds when you were in disaster mode, is
this OK?

- How does the application fail?  What does it need to run?

In the networking world, failover is pretty straightforward.
Networks can be moved around the globe in seconds or minutes.
If the application that drives the business caves in because it
doesn't have the data it needs, technology can't help it.

These questions aren't technical questions, they're business
questions, which is good, because disaster recovery is a
business problem.

What are business problems doing in a Linux newsletter?  Well,
the fact of the matter is that our skills as systems and network
admins are leveraged by the business to solve their problems.
All the Linux skills in the world won't help if you can't apply
them to the problem.  Sometimes we just have to step back from
the details of what we're doing and look at the big picture.

I liken this disaster recovery planning to insurance.  The
business is spending money on equipment and people to hedge off
any losses in the event that something awful happened to the
primary data centre.  Part of the planning might end up saying
that you'll just have to accept the risk.  Maybe you realize
that for N dollars you can guard against many risks, but to
really take care of every contingency will require 10\*N dollars.
At that point, it's up to management to authorize the extra
spending, or say "we've done all we can".  However, one of our
jobs as admins is to highlight the risks, and discuss possible
ways of mitigating it.  You'll find that even though you can't
solve a particular problem, discussing it and realizing that you
can't solve it is a whole lot better than being left in the dark.

As a closing note, one excellent example of a fault tolerant
system is the global DNS.  Did you know a massive DOS attack was
launched this week against most of the root servers?  I didn't
until I read it online, because everything worked like normal.
The thirteen root servers are spread around the globe, each able
to take a 300% load.  Even with most of them out of commission,
the Internet can still go, because the data on the root servers
is cached on the top level domain servers.  This will not be
feasible in every situation, but DNS shows that it can be done.
Not bad for something built 20-odd years ago, eh?

Long live the Penguin,


2) Linux News

Would You Like Linux With That?

Burger King has announced that their POS systems in Puerto Rico
will be Linux-based. The application is all web based, making
Linux the natural choice. This article goes over some of the
technology and hardware involved.


Mandrake 9.0 Review

ExtremeTech takes a look at Mandrake 9.0. Some good stuff here,
a screenshot or two, but also a description of the differences
between all the different offerings available from Mandrake.


Automating Manufacturing Processes

Linux is finding its way into more than just your desktops and
servers. Manufacturing processes, once the domain of
programmable logic controllers (PLCs) are starting to benefit
from the Penguin, too.


Linux IP Telephony to be Demonstrated

Bayonne is a project to let Linux play with telephony cards, in
the hopes of creating a PBX or IVR environment. Right now, this
is the domain of big, expensive hardware, so this project will
certainly have a market when it is successful.


3) Linux Resources

In a Jiffy

The fundamental unit of time in Linux is the jiffy, which
currently stands at 10ms. Everything hinges on it, especially
task scheduling. The 2.5 kernels allowed this number to be
changed. The code to do so was ported back to 2.4, and the
author answers some questions about the patch in an easy-to-
understand manner.


Linux Basics

This introductory document takes the reader through many common
tasks that occur on a Linux box. Though the domain name would
suggest it's for Debian users, the instructions aren't very


New DNS and BIND Book

O'Reilly is releasing a new book on DNS and BIND. Their first
book on this subject was excellent, and I'm looking forward to
seeing this one. Part of the promotion behind the release of
this book is the posting of some excerpts. Well worth a read if
you're into BIND.


How to Really Screw Up a Linux Installation

Here are the chronicles of Mary Robinson, a frequent contributer
to Cramsession's content. She installed Linux a while ago, and
wrote up some of the problems she encountered. I remember
working through some of this with her, and believe me, it's
better to learn from her mistakes than go through that again!


Creating a Chroot Jail

Here are some generic instructions to keep in mind whenever
you're looking at creating a chroot jail. The idea behind the
jail is that you can run an application inside of it, and if it
is compromised, the attacker can't get out to the rest of your


4) App o' the Week

Now here is a cool utility! If you have a command that involves
pipes, it'll report on the speed and progress. Very interesting


(C) 2002 BrainBuzz.com, Inc. All Rights Reserved.

          This message is from CramSession

You are currently subscribed to the following list
   Hottest Linux News and Resources
   as: sean@ertw.com

To un-subscribe from this newsletter by e-mail,
   send a blank email message to:

To subscribe to this newsletter and many others visit
our site at: