Dec 5 2002


                    LINUX NEWS
            http://www.Cramsession.com
          December 5, 2002 -- Issue #110


TABLE OF CONTENTS

1) Sean’s Notes

2) Linux News

Group Urges Limits on Open Source
Linux Jobs *Are* Out There
Linux Networx Builds Kick Ass Clusters
Advanced Server Gets More Reliable

3) Linux Resources

How to Get Hired as an Open Source Developer
Linux Cluster Information
Just Look At The Numbers
IDE RAID Examined
SUN's Capacity On Demand

4) App o’ the Week

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

Serebra Learning Corporation knows that it’s true, you get paid more if you have the skills. Learn at your own pace with our dynamic training programs for the skills needed to succeed in today’s IT market. The Best Way to Learn Anything, Anywhere, Anytime. Check out this month’s specials!

http://ad.brainbuzz.com/?RC153&AIX32

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

For information on how to advertise in this newsletter please contact mailto:adsales@CramSession.com or visit http://cramsession.com/marketing/default.asp


1) Sean’s Notes

Have 500 systems you need installed identically? Nor do I. But the techniques you’d use to do this are helpful in other ways. Say you had a lab box that you reinstall often, but always the same way (or maybe a few ways). Or maybe you just want a way of duplicating your current installation, such as when you’re moving out of test and into production on a new box. Through a technique known as “Kickstart” (Red Hat specific, I know), you can do this very easily.

Kickstarting hinges around one file, ks.cfg, which contains a list of the answers to all the questions that get thrown at you during the installation, everything from your firewall config down to your package selection.

There are two ways to go about making ks.cfg. The first is to use “ksconfig”, a GUI tool that lets you check off all the options, and generate the output file. If you’re looking to clone an existing installation, later versions of Red Hat dump a file called “anaconda-ks.cfg” in /root, which is, in a nutshell, the Kickstart file containing all the options you selected in the last installation. Subject to a couple of caveats, you can pick up that file, and clone your system hundreds of times with one floppy. Of course, any changes you made after the system came up won’t take effect unless you script them in yourself; we’ll see this later.

There is a third option for generating ks.cfg, namely starting from scratch, but I highly recommend grabbing something that works and editing from there. Kickstart is picky, and you’ll save yourself a lot of grief if you start off with something that’s at least close to working.

ks.cfg has three sections. First is the command section, where you have to specify if you’re doing an install or upgrade, if the network, X, or other components are to be configured, and other questions that pop up during the install. Next is the list of packages you want installed. Finally, there is a section containing scripts that get run before and after the install. As usual, blank lines are ignored, and # means a comment. Here is a very basic file: I’ll point you toward the complete list of commands at the end.

install or upgrade?

install

where is the install media?

cdrom

Languages being used? For some reason, you need both lines.

lang en_US langsupport –default en_US en_US

keyboard type

keyboard us

mouse? microsoft is a popular one, and –device specifies the

port, ignoring the leading /dev/

mouse microsoft –device ttyS0

give a root password, I’m using “password”

rootpw password

firewall settings, I’m disabling it for now

firewall –disabled

basic shadowed passwords, MD5 hashes

authconfig –enableshadow –enablemd5

timezone setting

timezone Canada/Central

Those basic options should be self-explanatory. The fun part is the disk related stuff:

where does the bootloader go, and which one to use?

default is GRUB, unless you use –useLilo

bootloader –location=mbr

VERY DANGEROUS! This will clear your partition table!

make sure you don’t need the data on the drive!

clearpart –all

build the partitions. For ease of demonstration,

I’m making a 128MB swap file, and making the rest the root

by saying that the partition is 1MB, but can grow to fit the disk

part swap –size 128 part / –fstype ext3 –size=1 –grow

Make sense? The first command tells the install program where to put the boot loader. After that, the partition table is wiped clean, and the two partitions are created. After that, it’s a matter of telling Red Hat what packages you want. Note you can select both package names, or package groups by prefixing the name with @

%packages # Select the groups of packages @Network Support @Dialup Support @NFS File Server @Anonymous FTP Server @SQL Database Server @Web Server @Router / Firewall @DNS Name Server @Network Managed Workstation @Utilities @Software Development @Kernel Development # individual packages httpd

Finally, we can run a couple of commands after the install finishes:

%post echo “System installed on date” » /etc/motd echo “nameserver 192.168.1.1” > /etc/resolv.conf echo “domain example.com” » /etc/resolv.conf

Before you go thinking that I’m pulling all of these commands out of thin air, browse on over to the official Red Hat instructions for the complete list:

http://www.redhat.com/docs/manuals/linux/RHL-8.0-Manual/custom-guid e/ch-kickstart2.html

Now that we have our glorious ks.cfg, the easiest way to perform a Kickstart is to copy the file onto a boot disk:

dd if=/mnt/cdrom/images/bootnet.img of=/dev/fd0

mcopy ks.cfg a:\ks.cfg

The first command uses the bootnet.img file from the first Red Hat CD to create a boot floppy. Since the boot disk is actually an MS-DOS filesystem, you can use the mcopy command to copy the ks.cfg to the disk without needing to mount it.

Finally, boot your test computer with this boot floppy. Remember, it’s going to automatically clear your partition table and install Linux, so make sure you don’t need that disk! (This is where VMWare is especially helpful). At the first prompt, enter:

linux ks=floppy

To start a Kickstart from the floppy. Stand back, and watch it go!

There are many more options available. For example, I store the Red Hat installation tree on my workstation and export it with NFS or FTP. Rather than the “cdrom” directive, you can pass the NFS or FTP information and install over the network.

If you play with this long enough, you’re bound to run into some bugs. Believe me, there are lots. If you miss a directive, the install will stop and ask you for the missing information. If you mess up a keyword, it will crash, giving you a cryptic error message. I spent half an hour trying to figure what I had done wrong, when it turns out I had forgotten the comment symbol in front of what should have been a comment. Still, the ability to bring up a test box to a known state with barely any intervention is worth the time investment.

That’s a quick look at Kickstart, folks. It’s a powerful tool for mass roll outs, or creating identical installations of lab equipment. Anyone looking to sit the RHCE certification lab exam will want to know this topic inside and out. Since it’s fully scriptable, you can have your final product as polished as you want.

Long live the Penguin,

Sean swalberg@cramsession.com


2) Linux News


Group Urges Limits on Open Source

The “Initiative For Software Choice” suggests that a defence contractor’s report on the benefits of Open Source may be incorrect. Not surprisingly, Microsoft and Intel are major backers. Cisco is the one that surprises me, as they’ve been written up as huge users of Free software. Et tu, CompTIA?

http://news.com.com/2100-1001-975578.html?tag_top


Linux Jobs Are Out There

The surveys are in, and the results show that Linux skills are in demand… but what in particular? What combinations of skills are being sought? Read on to find out.

http://www.linuxplanet.com/linuxplanet/print/4564/


Linux Networx Builds Kick Ass Clusters

“Aerodynamics engineers with Boeing’s Expendable Launch Systems division in Huntington Beach, CA used a 96-node cluster of PCs with Advanced Micro Devices 850-MHz Athlon processors running Red Hat Linux, rather than a $500,000 supercomputer, to keep costs low in pursuit of its goal. Linux cluster-management company Linux Networx helped to develop the environment.”

http://www.informationweek.com/story/IWK20021127S0035


Advanced Server Gets More Reliable

Red Hat is taking a cue from the OSDL Carrier Grade Linux Working Group, who are looking to bring telco-standard reliability to Linux. I’m looking forward to seeing the enhancements coming from Red Hat, and the markets they’ll try to get into with their enterprise offerings.

http://www.redhat.com/about/presscenter/2002/press_elf_rhas.html


3) Linux Resources


How to Get Hired as an Open Source Developer

Though the example here is for a company that uses a lot of Open Source tools, the advice is good for anyone looking to get a job in development. Blasting your resume out doesn’t seem to be the key; rather, a targeted approach is necessary.

http://www.theopenenterprise.com/story/TOE20021202S0001


Linux Cluster Information

Perdue’s computing services group has taken hundreds of recycled computers and turned them into massive Linux clusters for research. I don’t know what’s more disappointing – that their “recycled” computers are better than my desktop, or that I don’t get to play on the cluster. Either way, some interesting information on how a cluster is used in a real environment.

http://www.itap.purdue.edu/rcs/linuxclusters.cfm


Just Look At The Numbers

“This paper provides quantitative data that, in many cases, using open source software / free software is a reasonable or even superior approach to using their proprietary competition according to various measures. This paper examines market share, reliability, performance, scalability, security, and total cost of ownership.” An excellent paper.

http://www.dwheeler.com/oss_fs_why.html


IDE RAID Examined

Though SCSI is still king of storage, many of the computers we encounter will still have IDE. It’s cheap, it works, and it, too, can have the redundancy usually enjoyed in server systems without having to resort to software RAID. This article looks at the performance of various IDE-RAID cards, and talks a bit about RAID itself.

http://tech-report.com/reviews/2002q4/ideraid/index.x?pg=1


SUN’s Capacity On Demand

The utility model for computing has been seen in various forms over the years. The concept is intriguing… these whitepapers give a bit more information on how Sun plans to pull it off.

http://www.sun.com/datacenter/cod/


4) App o’ the Week

I’ve often pointed out scripts that help in managing Red Hat patches, but this is the first I’ve come across for other distros (Debian users, apt-get upgrade doesn’t count!). This time, it’s a script for SuSE.

http://fou4s.gaugusch.at/


(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: mailto:leave-linuxnews-3825955Y@list.cramsession.com

To subscribe to this newsletter and many others visit our site at: http://newsletters.cramsession.com/signup/default.asp