File: issue65.html

package info (click to toggle)
lg-issue65 2-1
links: PTS
area: main
in suites: sarge
size: 3,780 kB
ctags: 230
sloc: sh: 201; perl: 133; makefile: 34
file content (13368 lines) | stat: -rw-r--r-- 543,145 bytes
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<TITLE> Linux Gazette Table of Contents LG #65</TITLE>
<META NAME="robots" CONTENT="noindex, nofollow">
</HEAD>

<BODY BGCOLOR="#FFFFFF" TEXT="#000000" LINK="#0000FF" VLINK="#0000AF"
ALINK="#FF0000" >
<center>
<!-- A HREF="http://www.linuxgazette.com/">
<H1><IMG SRC="../gx/newlogo.jpg" ALT="LINUX GAZETTE" border="0"></H1></A> --> 
<A HREF="http://www.linuxgazette.com/">
<H1><IMG ALT="LINUX GAZETTE" SRC="../gx/lglogo.png" 
	WIDTH="600" HEIGHT="124" border="0"></H1></A> 
<H2>April 2001, Issue 65 &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
Published by <I>Linux Journal</I></H2> 

<A HREF=../index.html>Front Page</A> &nbsp;|&nbsp;
<A HREF=../index.html>Back Issues</A> &nbsp;|&nbsp;
<A HREF=../faq/index.html>FAQ</A> &nbsp;|&nbsp;
<A HREF=../mirrors.html>Mirrors</A> &nbsp;|&nbsp;
<A HREF=http://www.linuxgazette.com/search.html>Search</A>
</center>

<!--=================================================================-->

<table>
<TR>
<TD COLSPAN="3" WIDTH="100%" BGCOLOR="#000000"><IMG
SRC="../gx/hammel/cleardot.gif" ALT="" HSPACE=30 VSPACE=2 HEIGHT=1
WIDTH=1 ALIGN=LEFT></TD>
<TD></TD>
</TR>
<tr><td>
<table>
<tr>
<td VALIGN="top"><H1><font color="#BB0000">Visit Our Sponsors:</font></H1></td></tr>



<!-- *** BEGIN sponsors *** -->
<TR><TD>
	<A HREF="http://www.linuxnetworx.com"><IMG ALT="Linux NetworX" SRC="../gx/sponsors/linuxnetworx.gif" BORDER="0" VSPACE="10"></A>
</TD></TR>

<TR><TD>
	<A HREF="http://www.tuxtops.com"><IMG ALT="Tuxtops" SRC="../gx/sponsors/tuxtops.gif" BORDER="0" VSPACE="10"></A>
</TD></TR>

<TR><TD>
	<A HREF="http://www.elinux.com/xwb00120"><IMG ALT="eLinux.com" SRC="../gx/sponsors/elinux.gif" BORDER="0" VSPACE="10"></A>
</TD></TR>


<!-- *** END sponsors *** -->




</table>
</td>
<!--==================================================================-->
<TD ROWSPAN="3" WIDTH="2" BGCOLOR="#000000" NOSAVE><IMG
SRC="../gx/hammel/cleardot.gif" ALT="" HSPACE=1 HEIGHT=1 WIDTH=1%></TD>
<td VALIGN=top>
<H1><font color="#BB0000">Table of Contents:</font></H1>




<!-- *** BEGIN toc *** -->
<UL>
	<LI>  <a HREF="lg_mail65.html">The MailBag</A> 
	<LI>  <a HREF="lg_bytes65.html">News Bytes</A> 
	<UL>
		<LI>  <a HREF="lg_bytes65.html#distro">Distro News</A> 
		<LI>  <a HREF="lg_bytes65.html#general">News in General</A> 
		<LI>  <a HREF="lg_bytes65.html#software">Software Announcements</A> 
	</UL>
	<LI>  <a HREF="lg_answer65.html">The Answer Gang</A> , <EM>by The <EM>Linux Gazette</EM> Answer Gang</EM>
	<LI>  <a HREF="lg_tips65.html">More 2-Cent Tips</A> 
	<LI>  <a HREF="chan.html">Your Own Home Domain With ADSL</A> , <EM>by Ray Chan</EM>
	<LI>  <a HREF="collinge.html">HelpDex</A> , <EM>by Shane Collinge</EM>
	<LI>  <a HREF="correa.html">Interview with Linux Today's Marty Pitts</A> , <EM>by Fernando Ribeiro Correa</EM>
	<LI>  <a HREF="jenkins.html">Internet Printing - Another Way</A> , <EM>by Graham Jenkins</EM>
	<LI>  <a HREF="joshi.html">Parallel Processing on Linux with PVM and MPI</A> , <EM>by Rahul U. Joshi</EM>
	<LI>  <a HREF="lilly.html">Web Portals Bank on Open-Source Infrastructure</A> , <EM>by Ned Lilly</EM>
	<LI>  <a HREF="nielsen.html">Finding my computer at home from the outside</A> , <EM>by Mark Nielsen</EM>
	<LI>  <a HREF="okopnik.html">Learning Perl, part 3</A> , <EM>by Ben Okopnik</EM>
	<LI>  <a HREF="padala.html">So You Like Color !!!  (The mysterious ^[[ characters)</A> , <EM>by Pradeep Padala</EM>
	<LI>  <a HREF="puryear.html">Book Review: <EM>Networking Printing</EM></A> , <EM>by Dustin Puryear</EM>
	<LI>  <a HREF="steffler.html">Making Smalltalk: OO Thinking</A> , <EM>by Jason Steffler</EM>
	<LI>  <a HREF="stumpel.html">A Private Home Network</A> , <EM>by Jan Stumpel</EM>
	<LI>  <a HREF="sunil.html">Speeding Up Your Net Browsing with PDNSD Domain Name Caching</A> , <EM>by Sunil Thomas Thonikuzhiyil</EM>
	<LI>  <a HREF="lg_backpage65.html">The Back Page</A> 
	<UL>
		<LI>  <a HREF="lg_backpage65.html#authors">About This Month's Authors</A> 
		<LI>  <a HREF="lg_backpage65.html#notlinux">Not Linux</A> 
	</UL>
</UL>

<!-- *** END toc *** -->




</td>
</tr>
<TR>
<TD COLSPAN="3" WIDTH="100%" BGCOLOR="#000000"><IMG
	SRC="../gx/hammel/cleardot.gif" 
	ALT="-------------------------------------------------------------" 
	HSPACE=30 VSPACE=2 HEIGHT=1 WIDTH=1 ALIGN=LEFT></TD>
<TD></TD>
</TR>
</table> <!-- ******************************************* -->

<H3 ALIGN="center"><EM>Linux Gazette</EM> Staff and The Answer Gang</H3>
<BLOCKQUOTE>
<STRONG>Editor:</STRONG> Michael Orr<BR>
<STRONG>Technical Editor:</STRONG> Heather Stern<BR>
<STRONG>Senior Contributing Editor:</STRONG> Jim Dennis<BR>
<STRONG>Contributing Editors:</STRONG>
Ben Okopnik, Dan Wilder, Don Marti
</BLOCKQUOTE>

<TABLE WIDTH="100%"><TR>
<TD WIDTH="100%" BGCOLOR="#000000"><IMG
SRC="../gx/hammel/cleardot.gif" ALT="" HSPACE=30 VSPACE=2 HEIGHT=1
WIDTH=1 ALIGN=LEFT></TD>
</TR></TABLE>

<P> <!-- ************************************************** -->
<table>
<tr><td>
<A HREF="issue65.txt.gz">TWDT 1 (gzipped text file)</A><BR>
<A HREF="issue65.html">TWDT 2 (HTML file)</A><BR>
are files containing the entire issue: one in text format, one in HTML. 
They are provided 
strictly as a way to save the contents as one file for later printing in
the format of your choice; 
there is no guarantee of working links in the HTML version.
 </td></tr>
<!--==================================================================-->
<TR>
<TD COLSPAN="3" WIDTH="100%" BGCOLOR="#000000"><IMG
SRC="../gx/hammel/cleardot.gif" ALT="" HSPACE=30 VSPACE=2 HEIGHT=1
WIDTH=1 ALIGN=LEFT></TD>
<TD></TD>
</TR>
<tr><td>
<center>
<I>Linux Gazette</I><img alt="[tm]" src="../gx/tm.gif">, 
<A HREF="http://www.linuxgazette.com/">http://www.linuxgazette.com/</A><BR> 
This page maintained by the Editor of <I>Linux Gazette</I>,
<A HREF="mailto: gazette@ssc.com"> gazette@ssc.com</A>
<P> 
<H5>Copyright &copy; 1996-2001 Specialized Systems Consultants, Inc.</H5>
</center>
</td></tr>
<TR>
<TD COLSPAN="3" WIDTH="100%" BGCOLOR="#000000"><IMG
SRC="../gx/hammel/cleardot.gif" ALT="" HSPACE=30 VSPACE=2 HEIGHT=1
WIDTH=1 ALIGN=LEFT></TD>
<TD></TD>
</TR>
</table>

<center>
<H1><A NAME="wanted"><IMG ALIGN=MIDDLE ALT="" SRC="../gx/mailbox.gif">
The Mailbag</A></H1> <BR>
<!-- BEGIN wanted -->

</center>
<P> <hr> <P>
<!-- =================================================================== -->
<center><H3><font color="maroon">HELP WANTED -- Article Ideas</font></H3></center>
<P>
<P> Send tech-support questions, answers and article ideas to The Answer Gang
&lt;<A HREF="mailto:linux-questions-only@ssc.com">linux-questions-only@ssc.com</A>&gt;.  Other mail (including
questions or comments about the <EM>Gazette</EM> itself) should go to
&lt;<A HREF="mailto:gazette@ssc.com">gazette@ssc.com</A>&gt;.  All material
sent to either of these addresses will be considered for publication in the
next issue.  Please send answers to the original querent too, so that s/he
can get the answer without waiting for the next issue.

<P> Unanswered questions might appear here.  Questions with
answers--or answers only--appear in The Answer Gang, 2-Cent Tips, or here,
depending on their content.  <em><font color="#990000">There is no guarantee 
that questions will <strong>ever</strong> be answered, especially if not 
related to Linux.</font></em>

<P> <STRONG>Before asking a question, please check the
<A HREF=../faq/index.html><I>Linux Gazette</I> FAQ</A> to see if it has been
answered there.</STRONG>

<P> <HR> <P>
<!--====================================================================-->

<!-- BEGIN HELP WANTED -- Article Ideas -->

<UL>
<!-- index_text begins -->
<li><A HREF="#wanted/1"
	><strong>I've downloaded the ISO file.  Now what do I do with it?  I've burned a CD and it won't boot with it.</strong></a>
<li><A HREF="#wanted/2"
	><strong>Fetchmail question</strong></a>
<li><A HREF="#wanted/3"
	><strong>nuovo progetto</strong></a>
<li><A HREF="#wanted/4"
	><strong>syslog-ng</strong></a>
<li><A HREF="#wanted/5"
	><strong>rebooting using rsh command</strong></a>
<li><A HREF="#wanted/6"
	><strong>query</strong></a>
<li><A HREF="#wanted/7"
	><strong>Defining Keyboard Shortcuts</strong></a>
<li><A HREF="#wanted/8"
	><strong>upgrade2.2.14 to 2.4 kernel documention</strong></a>
<li><A HREF="#wanted/9"
	><strong>Linux Sockets Stuck in FIN_WAIT</strong></a>
<li><A HREF="#wanted/10"
	><strong>CD-Writing with an ATAPI CDR Mini-HOWTO</strong></a>
<!-- index_text ends -->
</UL>
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/1"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">I've downloaded the ISO file.  Now what do I do with it?  I've burned a CD and it won't boot with it.</FONT></H3>
Thu, 8 Mar 2001 03:29:25 -0000
<BR>EK<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20I've%2520downloaded%2520the%2520ISO%2520file.%2520%2520Now%2520what%2520do%2520I%2520do%2520with%2520it%3F%2520%2520I've%2520burned%2520a%2520CD%2520and%2520it%2520won't%2520boot%2520with%2520it.&cc=endak@yahoo.com">(endak from hotmail.com)</a>


<!-- sig -->

<strong>
<P>
Dear Answer Guy,
</P>
<P>
I really hope you can help me in my quest to change from Windows to Linux.
Here is what I have done so far:
</P>
<P><BLOCKQuote>
I've got a 2 year old standard PC:  PII 333MHz, 128MB RAM, 4.3GB, CD-ROM,
Modem, ESS soundcard.
</BLOCKQuote></P>
<P>
I've downloaded the ISO file from <A HREF="http://www.suse.com/">SuSE</A> and it is called
live-evaluation-i386-70.iso
</P>
<P>
I've put this onto a CD.  I can see the file on the CD under Windows.  When
I try to boot up using the CD ROM it ignores it.  My BIOS is seto to Boot
From CD and it will boot from a Windows OS CD so I know the capability is
there.
</P>
<P>
I'm starting to feel that there is something I need to do to the ISO file
when writing it to the CD.  I've seen mention of doing this by "Burning the
Image" - whatever that is.  I've read the HOW-TO for CD and ISO and it
doesn't explain how to do this under Windows only.  It assumes you have a
working Linux environment.
</P>
<P>
I came across your page, and dozens of others on this subject, and I thought
I had found the holy grail as the question is exactly what I was going to
ask.  The thing is your answer is for a working Linux system - the guy had
this as well as Windows - there is no mention of how to do this under
Windows.  Then the person who asked the original question writes back
saying:
</P>
<BLOCKQuote>
<p>
"Jim,
</P><p>
Thanks for your information. And ironically, shortly (next day) after I
wrote you the email, I did find out what was going wrong and how to fix it.
</P>
<P>
WinOnCD did have the capability, but it was somewhat a "hidden" feature of
sorts.
</P>
<P>
I do appreciate your response, though.
</P>
<P>
-Lewis"
</P>
</BLOCKQuote>
<P>
He didn't even mention how he did it in WinOnCD!!!!  ARRggghhh!  I can find
this software on the Web but I still need the knowledge to find the "hidden
feature".  Oh this is so frustrating., so close yet so far.
</P>
<P>
Please please please can you help me to make my ISO file, which took hours
to download, into a bootable CD so that I can install my first ever Linux
OS?
</P>
<P>
Thanks,
EK.
</P>
</strong>
<blockquote>To trim it down considerably, he has tried Nero Burning software
but couldn't make sense of it, and the image that seems to result doesn't
work.  YaST2 booted from floppy gets as far as reaching for the CD, then 
continues to complain that the CD isn't valid.</blockquote>
<blockquote>
Usually the Gang would suggest a rescue disk - our perennial favorites
wouldn't help, they don't have cdrecord aboard, but you could try muLinux
(<a href="http://mulinux.nevalabs.org/"
>http://mulinux.nevalabs.org/</a>) boot that from floppy, and use the Linux
software on it to mount up your DOS filesystem, then cdrecord according 
to the normal HOWTO to burn the SuSE CD.  I have no idea if it would work 
though it seems worth a shot.
</blockquote>
<blockquote>Would some kind soul out there who lives in both worlds point
us to a reliable CD burning app for Windows, along with some fairly 
simple instructions?  We'd be glad to let you put the article in the
Gazette if it will help enough potential Linux'ers out there. -- Heather
</blockquote>

<!-- end 1 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/2"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Fetchmail question</FONT></H3>
Mon, 19 Mar 2001 08:47:25 -0800
<BR>Rodrigo P Gomez <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Fetchmail%2520question&cc=rpgomez@yahoo.com">(rpgomez from yahoo.com)</a>

<P>
I read Ben Okopnik's <a href="../issue62/okopnik.html">article in the 
February issue</a> of Linux Gazette titled
</P>
<P>
No More Spam! (a "procmail"-based solution with tips on "fetchmail" and
"mutt")
</P>
<P>
and I tried implementing some of his suggestions. My problem seems to be this:
While fetchmail will get my e-mail, and pass it off to procmail, which
delivers it to a designated file, I can't seem to read the e-mail with kmail.
I've set my designated file as a local mailbox for kmail to read, but when I
try to read it using kmail I get 1 of 2 behaviors:
</P>
<ol>
<li> kmail ignores the e-mail in the local mailbox
<li> kmail nukes the contents of the local mailbox, and does not
transfer the contents to my kmail inbox.
</ol>
<P>
I tried setting up kmail to read from the local mailbox with the following
lock file options:
<br>mutt dotlocked
<br>mutt dotlocked priveledged
<br>Procmail lockfile
<br>FCNTL
<br>None
</P>
<P>
Any of the above will cause either behavior 1) or 2).
</P>
<P>
Anyway, I'm hoping you might be able to help me figure out how to
read my local mailbox with kmail.
</P>
<P>
--Rod
<br>P.s. Included is a <a href="misc/wanted/rodrigo.kmail.png">snapshot of 
my kmail configuration</a> for reading the local mailbox into my inbox. 
Hopefully, it will be of some use to you for diagnosing my problem.
<br>P.p.s. Thanks in advance for any help you can give me on this.
<br>P.p.p.s. Here are the configuration files for the various utilities:
(I blanked my password in the .fetchmailrc file in this e-mail, for security
reasons 
<IMG SRC="../gx/dennis/smily.gif" ALT=";-)" 
		height="24" width="20" align="middle">)
</p>
<ul>
<li><a href="misc/wanted/rodrigo.fetchmailrc.txt">.fetchmailrc</a>
<li><a href="misc/wanted/rodrigo.procmailrc.txt">.procmailrc</a>
<li>The top part of <a href="misc/wanted/rodrigo.kmailrc.txt">kmailrc</a>, 
     with the reference to my password for yahoo, blanked out
</ul>
<P>

<!-- end 2 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/3"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">nuovo progetto</FONT></H3>
Thu, 15 Mar 2001 16:38:46 +0100
<BR>frangregorio <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20nuovo%2520progetto&cc=fbaril@tiscalinet.it">(fbaril from tiscalinet.it)</a>


<!-- sig -->

<strong>
<P>
Cari amici della gazette
mi chiamo Francesco Baril&agrave;
Insieme ad alcuni amici vorremmo creare uno shell-wordprocessor emacs
based: qualcuno di voi &egrave; interessato?
<br>Se lo siete mandate un email a <em>fbaril@tiscalinet.it</em>
oppure visitate <em>linux.interpuntonet.it/angolinux</em> al tread 
"nuovo progetto"
<br>grazie
<br>Francesco
</P>
<P>
Dear gazette- friends
I am Francesco Baril&agrave;
I and my friends want create a shell- woprdprocessor emacs-based: is anyone
interested?
<br>mail to <A HREF="mailto:fbaril@tiscalinet.it"
	>fbaril@tiscalinet.it</A> or visit 
<a href="http://linux.interpuntonet.it/angolinux"
	>linux.interpuntonet.it/angolinux</a>
 and see the tread "nuovo progetto"
<br>thanks
<br>Francesco Baril&agrave;
</P>
</strong>

<!-- sig -->

<p>Let him know what you think, folks.  For my own part, we could still
use a wordprocessor that actually works... -- Heather</p>

<!-- end 3 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/4"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">syslog-ng</FONT></H3>
Tue, 27 Mar 2001 18:16:54 +0200
<BR>Greg <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20syslog-ng&cc=greg@hellea.be">(greg from hellea.be)</a>


<!-- sig -->

<P>
Hi, I'm fighting with syslog-ng, trying to centralize the logs of all a
network to a log server. Till now, I'm justing testing it between two
machines ... I've already been on it a long time and still no result,
simply nothing appears in the final log on the logserver ...
I've attached the syslog-ng.conf of the client and the server's one too.
</P>
<P>
You must know that for the server's source, I've already tried to enter
parameters (I mean an ip and a port) but it doesn't work either so at
the moment, I'm trying to use the defaults. The same thing for the
client ...
<br>Can you help me on this?
<br>Thanks
</P>
<ul>
<li><a href="misc/wanted/greg.client.conf.txt">client.conf</a>
<li><a href="misc/wanted/greg.logserver.conf.txt">logserver.conf</a>
</ul>

<!-- end 4 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/5"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">rebooting using rsh command</FONT></H3>
Mon, 26 Feb 2001 11:40:46 -0500
<BR>rob <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20rebooting%2520using%2520rsh%2520command&cc=rob@esgi.com"> (rob from esgi.com)</a>

<!-- sig -->

<strong>
<P>
AnswerGuy,
</P>
<P>
I hope you can help me with this problem.
</P>
<P>
We've upgraded our kernel from 2.0.0 to 2.2.14-5.0 (<A HREF="http://www.redhat.com/">Red Hat</A> 6.2). We have a master node that is connected to a total of 14 nodes. In the past, when we rsh'd into a node and wanted to reboot we would just type at the prompt 'reboot'. We would immediatly be kicked back to the master node prompt while the other node was rebooting. This worked out fine.
</P>
<P>
However, when we rsh into a node and type 'reboot' the terminal hangs, for several minutes, until the node has completed rebooted and then it drops us back to the master node.
We want it to work the way it has in the past. Is there some file or configuration I'm missing, some flag to turn off or on?
</P>
</strong>

<p>Any of you gentle readers with experience in this kind of clustering,
if you have any idea what it might be, then feel free to lend a hand.  We'll
publish your answer here. -- Heather</p>

<!-- end 5 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/6"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">query</FONT></H3>
Sat, 3 Mar 2001 15:44:33 +0530
<BR>Mehul Vora <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20query&cc=mehul@now-india.com">(mehul from now-india.com)</a>


<!-- sig -->

<P>
Hii james, this is mehul from india...
</P>
<P>
need to know if i can control bandwidth on a particular interface using 
ipchains or any other utilities in linux... basically, ill have a 
multi-homed linux workstation and i want to limit bandwidth on one 
interface to 512kbps... is that possible?
</P>
<P>
T&amp;R
<br>mehul
</P>

<!-- sig -->


<!-- end 6 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/7"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Defining Keyboard Shortcuts</FONT></H3>
Sun, 04 Mar 2001 23:04:54 -0500
<BR>Daniel S. Washko <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Defining%2520Keyboard%2520Shortcuts&cc=dann@thelinuxlink.net">(dann from thelinuxlink.net)</a>


<!-- sig -->

<P>
I would like to define keyboard shortcuts not specific to a Window
Manager.  I have read documentation on xmodmap, but the information I
have found only talks about and show examples of changing key actions to
another pre-defined action.  That is, being able to change the caps lock
key to function as an escape key.  How would I go about assigning a
command to a lctrl + F1 key code?
</P>
<P>
The ultimate use for this is to help a fellow LUG member.  He wants to
be able to assign text strings to "hot keys" whereby the text string
would be copied to the window he is working in.  He wants to be able to
save the text strings as he is working so that they can be used in other
session; and have the ability to delete or change these text strings.  I
can proably figure out a script to do this, but I am stuck on defining
the keyboard shortcut.
</P>
<P>
Thanks
</P>
<P>
Daniel S. Washko
</P>

<!-- sig -->


<!-- end 7 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/8"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">upgrade2.2.14 to 2.4 kernel documention</FONT></H3>
Mon, 12 Mar 2001 13:19:29 GMT
<BR>Mark Taylor <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20upgrade2.2.14%2520to%25202.4%2520kernel%2520documention&cc=mky@talk21.com">(mky from talk21.com)</a>


<!-- sig -->

<P>
hello,
</P>
<P>
is there any kernel upgrade from 2.2.1# to 2.4 how-to 's.  I have upgraded the kernel using a general how-to and I have problems mounting the vfat partitions by getting the error "invalid major and minor numbers". I know the way the system deals with special device files has changed at 2.4.  Also the "eth0" ethernet adapter is not being recognised.  Any help in finding good documentation for this process would be greatly appreciated.
</P>
<P>
mark taylor
</P>

<!-- sig -->


<!-- end 8 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/9"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Linux Sockets Stuck in FIN_WAIT</FONT></H3>
Mon, 19 Mar 2001 10:09:45 -0500
<BR>Ken Ramseyer <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Linux%2520Sockets%2520Stuck%2520in%2520FIN_WAIT&cc=ken.ramseyer@lmco.com">(ken.ramseyer from lmco.com)</a>


<!-- sig -->

<P>
On our project if an established client socket connection on a remote
chassis is suddenly terminated (e.g., the chassis is powered off), the
socket connection on the local chassis changes from ESTABLISHED to
FIN_WAIT1.  If we then try to restart our application on the local
chassis, it does not work because the socket connection is stuck in a
FIN_WAIT1.  After 5-10 minutes the socket connection stuck in FIN_WAIT1
clears itself and we can successfully restart our application on the
local chassis, but this wait is too long.
</P>
<P>
Do any of you know how to expedite the process of clearing FIN_WAITs on
a Linux/UNIX chassis under these conditions?  The only we can get them
to clear is to either wait 10 minutes or perform a chassis reboot (sync;
shutdown -r now).  Is there a system call that can be used to tell the
operating system to close/delete/clean-up/remove all socket connections
immediately?
</P>
<P>
The only thing I have run across so far is to possibly make a kernel
change (e.g., change #defines in <TT>.../include/net/tcp.h</TT>), or set a socket
option which causes the sockets to close/clean-up faster.  Note: these
changes may be risky if we shorten the timeouts too much.
</P>
<P>
Any help would be appreciated.
</P>
<P>
Thanks,
<br>Ken Ramseyer 
</P>



<!-- end 9 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="wanted/10"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">CD-Writing with an ATAPI CDR Mini-HOWTO</FONT></H3>
Wed, 21 Mar 2001 22:12:22 +0000
<BR>Louise Plumpton <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20CD-Writing%2520with%2520an%2520ATAPI%2520CDR%2520Mini-HOWTO&cc=louiseplumpton@homewitheva.fsnet.co.uk">(louiseplumpton from homewitheva.fsnet.co.uk)</a>


<!-- sig -->

<P>
Hi
<br>I've read the articles from issue 57 of the Linux gazette, but am unable
to get my cd-rw to work.  I don't think I am managing to emulate scsi
correctly, although I have followed the things sugested. I have a sony
cd-rw (CRX 145E ATAPI) and run mandrake 7.0, I also have a iomega 100Mb
zip drive on hdb,  a dvd-rom on hdc and the cd-rw is on hdd.  This is
what I've done and what the comp says:
</P>
<P>In <TT>/etc/rc.d/rc.local</TT> added</p>
<BLOCKQuote>
<code>/sbin/insmod ide-scsi</code>
</BLOCKQuote>
<P>
In <tt>/etc/conf.modules</tt> added
</p>
<blockquote>
<pre>alias scd0 srmod
alias scsi_hostadapter ide-scsi
options ide-cd ignore=hdd
</pre>
<br> (also tried replacing srmod with sr_mod)
</blockquote>
<P>
In <TT>/etc/lilo.conf</TT> added</p>
<blockquote><tt>append="hdd=ide-scsi"</tt></blockquote>
<P>
then in console typed</p>
<blockquote>
<tt>lilo</tt> (and tried <TT>/sbin/lilo</TT>)
</blockquote>
<P>
then rebooted, then
dmesg gives at the end:
</P>
<Pre>
hdb:&lt;3&gt;ide-scsi: hdb: unsupported command in request queue (0)
end_request: I/O error, dev 03:40 (hdb), sector 0
unable to read partition table
scsi0 : SCSI host adapter emulation for IDE ATAPI devices
scsi : 1 host.
Vendor: IOMEGA    Model: ZIP 100           Rev: 14.A
Type:   Direct-Access                      ANSI SCSI revision: 00
Detected scsi removable disk sda at scsi0, channel 0, id 0, lun 0
sda : READ CAPACITY failed.
sda : status = 0, message = 00, host = 0, driver = 28
sda : extended sense code = 2
sda : block size assumed to be 512 bytes, disk size 1GB.
sda:scsidisk I/O error: dev 08:00, sector 0
unable to read partition table
</Pre>
<P>
For some reason I can't get the machine to emulate scsi on anything
other than hdb.  <tt>cdrecord -scanbus</tt>
only lists the zip drive too.  The <TT>/sbin/insmod ide-scsi</TT> comand also
stops the zip drive from working.  Have  you any ideas as to what might
be going wrong?
</P>
<P>
Many thanks for any advice you can offer
<br>Louise
</P>


<!-- end 10 -->
<a name="mailbag"></a>
<P> <hr> <P>
<!-- =================================================================== -->
<center><H3><font color="maroon">GENERAL MAIL</font></H3></center>
<P> <HR> <P>
<!--====================================================================-->

<!-- BEGIN GENERAL MAIL -->

<UL>
<!-- index_text begins -->
<li><A HREF="#mailbag/1"
	><strong>Sci-Linux Project has a listed e-mail address ..</strong></a>
<li><A HREF="#mailbag/2"
	><strong>Navigation...</strong></a>
<li><A HREF="#mailbag/5"
	><strong>Mailto URL error in LG Mailbag (followup)</strong></a>
<li><A HREF="#mailbag/6"
	><strong>Heather Stern</strong></a>
<li><A HREF="#mailbag/8"
	><strong>The Answer Gang</strong></a>
<!-- index_text ends -->
</UL>
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="mailbag/1"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Sci-Linux Project has a listed e-mail address ..</FONT></H3>
Mon, 26 Feb 2001 10:20:53 -0500 (GMT)
<BR>Manoj Warrier <a href="mailto:gazette@ssc.com&cc=manoj@ipr.res.in&subject:%20Re%3A%20%5BLG%2065%5D%20Sci-Linux%2520Project%2520has%2520a%2520listed%2520e-mail%2520address%2520.."
        >(manoj from ipr.res.in)</A>


<!-- sig -->

<STRONG>
<P>
Hi.
</P>
<P>
This is in reply to <b>ghaverla</b>@<b>freenet.edmonton.ab.ca</b>'s mail about
scilinux.freeservers.com not having a e-mail address on the page.
In fact at the bottom of the page before the disclaimer we have
links to contact us. It appears when i look at it from India
On Netscape. I am one of the maintainers and please do contact
me for your suggestions, ideas, flames, etc ..
</P>
<P>
We are just crystallising our thoughts after the fruitful 
<a href="../issue61/lg_answer61.html#tag/6">discussion</a>
with TAG (Jan 2000 issue) and have decided to first install all the
relevant packages on a RedHat partition of our PC, find the library
dependencies, make a "library farm" and use ldconfig after adding
the libraries in the library farm to the <TT>/etc/ld.so.conf</TT>. To check
what problems this gives we will then make a "zeroth" version of
our CDROM and installation script and try installing all on the
SLACKWARE (my dream distro) partition of our PC. Test the software,
check if anything else goes wrong, etc .. It seemed to work for
scilab and octave when I did a manual check. From what I have read
till now about libraries (not much at all, maybe I should read up
and write an article for the <i>LINUX GAZETTE</i> -&gt; best way to
learn something) this does not seem to be a wrong thing to do.
</P>
</STRONG>

<p>Thanks, Manoj, we'd love to see your article!  -- Heather</p>

<STRONG>
<P>
Another thing is we are not trying to make a platform
independent package or installer, but just want to install some
selected scientific software packages on any old/new Linux PC.
We hope to make a zeroth version, test it at our institute PCs
and only after thorough testing keep it for download at some
voulenteering site or mail it to anyone who pays the posting
and CDROM cost.
</P>
<P>
PS: We are still looking for some server to host our site. We do
thank freeservers, but the it would be nice to host the site
at some server that displays Linux related advts only.
</P>
<P>
Manoj
<br>My Enviornment for scientific computing on Linux page :
<A HREF="http://Scilinux.freeservers.com/"
	>http://Scilinux.freeservers.com</A>
</P>
</STRONG>


<!-- end 1 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="mailbag/2"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Navigation...</FONT></H3>
Thu, 01 Mar 2001 17:51:56 -0500
<BR>James Coleman <a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Navigation...">(jecoleman from upsala.org)</a>


<!-- sig -->

<STRONG>
<P>
...I love what you do but how about putting links to the rest of
LinuxGazette (or at least the front page) at the tops and bottoms of
each Answer Gang article? Thanks!
</P>
<P>
--
Jim Coleman
</p>
</STRONG>

<p>Once upon a time this was the case, but folks wrote in, saying that 
having that main navbar as well as the links within the TAG area was a 
bit confusing... they often hit "next article" when they meant "next TAG 
message".  But, we can try adding back in only the main Index, and see 
how that works. -- Heather</p>

<!-- end 2 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="mailbag/5"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Mailto URL error in LG Mailbag (followup)</FONT></H3>
Fri, 2 Mar 2001 00:25:45 -0800 (PST)
<BR>Anthony E. Greene <a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Mailto%2520URL%2520error%2520in%2520LG%2520Mailbag">(agreene from pobox.com)</a>

<!-- sig -->

<P><STRONG>
I'm glad you're using mailto's that include the TAG address. The problem
is the URL construction was incorrect.
</STRONG></P>

<P><STRONG>
It should be an equals sign (=) after the cc parameter. A quick run through sed
should fix this before you get too many messages pointing out the error.
</STRONG></P>
<P><STRONG>
If there's any way to add the subject, you could make the URLs look like
this:
</STRONG></P>
<P><STRONG><BLOCKQuote>
mailto:user@domain?cc=linux-questions-only@ssc.com&amp;subject=Re:%20SUBJECT
</BLOCKQuote></STRONG></P>
<P><STRONG>
Note that the subject has to go through a filter that replaces spaces with
the hex code "%20" to keep the URL legal. In perl, I'd do this:
</STRONG></P>
<blockquote><STRONG><code>
$mailto_subject =~ s/ /%20/g;
</code></STRONG></blockquote>
<P><STRONG>
You can quickly test the functionality of a mailto URL by typing
variations directly into the address/URL box of the target browser and see
if it calls the mail client as you intended.
</STRONG></P>
<P><STRONG>
Tony
</STRONG></P>
<P>
Thanks, Tony, you win the AnswerBubble for the month.  I like it when
folks not only nail us for a bug (that one was me, I'll go fix
the script so it'll be right next time 
<IMG SRC="../gx/dennis/smily.gif" ALT=";&gt;" 
		height="24" width="20" align="middle"> ... 
but also send us Tip grade material about fixing it.  We're trying it this
month.  I'm sure that our readers will let us know if there's any problems -- Heather
</P>

<!-- end 5 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="mailbag/6"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Heather Stern</FONT></H3>
Tue, 6 Mar 2001 12:15:33 -0500
<BR>Michael Gargiullo <a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Heather%2520Stern">(gargiullo from home.com)</a>

<P>
I just want to say thank you.
</P>
<P>
Her greeting in The Answer Gang is beautiful.  I'm a relative newbie to
Linux (It took me a week of reading and playing to compile a kernel that
would find my NICs).  I understand the open source idea, and love it.  I
deal in 2 other fields where this idea (in a general sense) applies.
Emergency Medical Services where we share ideas, issues, and solutions, not
for money, but for the knowledge of helping others.  Also in restoring old
cars,  Things are learned, forgotten, learned again, and most importantly
past on.  By sharing, we (Linux users) end up with a better app or method.
</P>
<P>
Again, Thank you
</P>
<P>
Remember, there are people out there who really appreciate the work that is
done.
</P>
<P>
- Mike Gargiullo
</P>


<!-- end 6 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="mailbag/8"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">The Answer Gang</FONT></H3>
Fri, 9 Mar 2001 11:08:38 -0800
<BR>Andrew Higgs 
<a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20The%2520Answer%2520Gang">(ahiggs from ps.co.za)</a>

<!-- sig -->

<P><STRONG>
In Issue 64 you invite Ray Taylor to join The Answer Gang. How would one
do that? Is it like a mailing list open to everyone? Can anyone help?
</STRONG></P>
<p>
TAG is run like a mailing list in reverse.  The public sends in
questions, and the subscribers are the answerers.  To join, send e-mail
to <A HREF="mailto:tag-request@ssc.com"
	>tag-request@ssc.com</A> with "subscribe tag me@mysite.com" 
in the message body.
Then just jump in whenever you have something to say.  At the end of the
month, Heather selects some of the messages for publishing. -- Mike
</p>

<!-- end 8 -->
<a name="gaz"></a>
<P> <hr> <P>
<!-- =================================================================== -->
<center><H3><font color="maroon">GAZETTE MATTERS</font></H3></center>
<P> <HR> <P>
<!--====================================================================-->

<!-- BEGIN GAZETTE MATTERS -->

<UL>
<!-- index_text begins -->
<li><A HREF="#gaz/1"
	><strong>SSH article</strong></a>
<li><A HREF="#gaz/2"
	><strong>Mean Thoughts on the Linux Router Project</strong></a>
<!-- index_text ends -->
</UL>
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="gaz/1"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">SSH article</FONT></H3>
Tue, 6 Mar 2001 14:55:51 -0800
<BR>Bryan Henderson <a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20SSH%2520article">(bryanh from giraffe-data.com)</a>

<P>
In the
<A HREF="../issue64/dellomodarme.html">article on ssh, scp, and sftp</A> in the
March issue, there is an important area that isn't covered: client/server
compatibility.
</P>
<P>
If you're just doing a basic ssh (to get a remote shell), you're
using a standard SSH protocol and any program named "ssh" is likely
to work with any remote system that offers a service it calls "ssh."
</P>
<P>
But scp and sftp are not standard protocols.  If you run the scp
program from openssh against a remote system that's running an
original ssh server, it will not work.  (And when I learned this the
hard way, it was very hard indeed:  the error message isn't "this
server doesn't implement this scp protocol."  It is, for reasons
that took a day of debugging to figure out, "invalid file descriptor"!
</P>

<!-- end 1 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="gaz/2"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/envelope.gif">
<FONT COLOR="navy">Mean Thoughts on the Linux Router Project</FONT></H3>
Mon, 26 Mar 2001 09:22:56 -0500 (EST)
<BR>--Mark-- <a href="mailto:gazette@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Mean%2520Thoughts%2520on%2520the%2520Linux%2520Router%2520Project">(mf from agate.net)</a>

<STRONG>
<P>
First off let me apologize to all the developers or others who I have
offended with my <A HREF="../issue64/fevola.html">views</A> on the Linux Router
Project (LRP).  By no means did I want to start a flame war.  The truth is that
I wrote about something outdated.  Second, that article was entirely my own;
not the work or opinions of Linux Gazette.
</P>
<P>
Since I wrote "Mean Thoughts" I have received a great many meaningful and
insightful messages from LRP users and developers.  If I wrote any 
<EM>untrue</EM>
information I want to know about it.  One point of contention for example
was whether the ip command is 'nonstandard.'  This is purely subjective.  If
ip really is standard, it should replace ifconfig or route like ipchains
replaced ipfwadm.
</P>
<P>
Nevertheless my views on the LRP have changed.  I received such an education
that I feel obligated to state for the record I have learned uses for each
of the three main LRP distributions, EigerStein
(<A HREF="http://lrp.steinkuehler.net/DiskImages/Eiger/EigerStein.htm"
	>http://lrp.steinkuehler.net/DiskImages/Eiger/EigerStein.htm</A>),
Oxygen
(<A HREF="http://leaf.sourceforge.net/pub/oxygen/"
	>http://leaf.sourceforge.net/pub/oxygen</A>),
LRP 2.9.8
(<A HREF="http://www.linuxrouter.org/"
	>http://www.linuxrouter.org</A>) --even in embedded systems.
I am not brand-loyal.  Advocacy is fine, but fanaticism has got to go.  I'll
use the best tool for the job, and <EM>how</EM> I determine what is the 'best tool'
is purely subjective.  Five years ago I preferred 3Com to any other NIC.
Why?  Two reasons:  The founder of 3Com invented Ethernet, and the cards
were recognized by all the OSs that the company used.  I knew I would not
have to worry about cross platform compatibility.  Now I prefer SMC.  Why?
Mainly because all the OSs recognize them but also because I can
jumper-select IRQ &amp; I/O on the models I use.
</P>
<P>
Would I write another 'anti-Linux' article?  Sure.  But not one that could
potentially insult anyone like when I said, 'developers wasting time'.
Linux is merely a product.  Windows NT is also a product.  Never mind the
fact that I despise products from Redmond, Washington: I don't think it's a
sin to admit that NT is better than Linux at being a "Domain Controller."
It does not change how much I like Linux.
</P>
<P>
Look at the article and notice its verbosity.  It's an opinion, not a
review.  I did not write it solely for explaining <EM>my</EM> (i.e., not Linux
Gazette == Don't shoot the messenger.) thoughts on the LRP, I also wanted to
express and present other information that may be useful to the Linux
community, for example the bit on standardization.  I did not write it to
maliciously annoy anyone.  Also to my knowledge there is no technically
inaccurate information.  I wrote specifically, "I have not done a lot of
work/research with LRP incarnation at linuxrouter.org as such but I am
familiar with the Materhorn Project."  My mistake was that I equated &lt;Linux
Router Project&gt; with one flavor, Materhorn.
</P>
<P>
I may or may not follow with a "Nice Thoughts on the Linux Router Project"
article. 
<IMG SRC="../gx/dennis/smily.gif" ALT=";)" 
		height="24" width="20" align="bottom">  In any case, I'd like
to put all hard feelings aside and hope that anyone who I have offended
would do the same.
</P>
<P>
Sincerely,
</P>
<P>
Mark Fevola
</P>
</STRONG>
<P>
[Mike]
We received several complaints about the article, feeling that it attacked the
LRP unfairly.  Dave Cinege, the creator of the LRP, was going to write a
response addressing the inaccuracies he felt were in the article, but he did
not have time to finish his letter by press time.  I encourage readers with an
interest in routing to follow the links above to the projects' home pages and
decide for themselves if the LRP and its offspring are right for them.
</P>
<P>
Regarding Oxygen, EigerStein and 2.9.8, 
<A HREF="mailto:dcinege@psychosis.com"
	>Dave</A> writes:
</P>
<BLOCKQUOTE>
They are derivatives of stable releases of LRP, which is currently 2.9.8.
I have been creating a new OS similar to LRP for quite some time now.
Many things have come out of this are a new multi-packaging system (standard?)
that is more powerful then rpm or deb, yet not tied to any specific OS.
</BLOCKQUOTE>
<P>
Regarding the 'ip' command, he writes:
</P>
<BLOCKQuote>
ip allows you to control the extended routing features of 2.2 and 2.4,
IE multiple routing tables. Ifconfig still works for the primary routing table
and interface configuration. ip can replace ifconfig, but ifconfig is still the
known standard.
</BLOCKQuote>
<P>
A few letters questioned <EM>LG</EM>'s editorial policy in allowing this article
to be published.  <EM>LG</EM>'s policy is pretty open.  If an article is about
Linux, contains hard facts or cultural value (e.g., humorous articles, cartoons
and articles about Linux VIPs), covers a topic relevant to a significant
portion of the readership, is not an advertisement in disguise, and would still
be relevant several months from now, we'll probably publish it.  There are
borderline cases, and this was one of them.
</P>
<P>
<EM>LG</EM> does not have a technical review board to screen every article,
although I do send a few questionable articles to The Answer Gang
for comment.  You, our readers, are <EM>LG</EM>'s technical review board,
and usually this system works very well.  99% of <EM>LG</EM>'s articles are
published without complaint.
</P>
<P>
In any case, please remember this article describes <EM>one person's</EM>
experience with certain routing programs.  It's not meant to be gospel,
in spite of the letter I received that said, "But newbies will read it and
think it's gospel!"  That's not how it works.  If you want gospel, read several
people's articles and compare them with your own experience.
</P>
<P>
Another thing this article does is raise the question, just because we <EM>can</EM>
use Linux in a wide variety of routing situations, should we?  Are you choosing
a Linux router because it's the most appropriate solution for the task, or
simply because "we're a Linux-only shop"?  Even if the article failed to
present LRP in a fair light, these are still questions worth asking.
</P>
<P>
As always, if you have any comments about an article, whether good or bad,
send them to <EM>LG</EM> and we will forward them to the author.
</P>

<!-- end 2 -->

<H4 ALIGN="center">"Linux Gazette...<I>making Linux just a little more fun!</I>"</H4>
<HR>

<center>
<table cellpadding=7><tr><td>
<IMG SRC="../gx/bytes.gif" border=1  ALT="News Bytes">
</td><td>
<H3>Contents:</H3>
<ul>
<li><a HREF="#distro">Distro News</A>
<li><a HREF="#general">News in General</a>
<li><a HREF="#software">Software Announcements</a>
</ul>
</td></tr></table>

<STRONG>Selected and formatted by  <A HREF="mailto:michael.conry@softhome.net">Michael Conry</A> and <A HREF="mailto:gazette@ssc.com">Mike Orr</A></STRONG>
</center>


<P> Submitters, send your News Bytes items in 
<FONT SIZE="+2"><STRONG>PLAIN TEXT</STRONG></FONT>
format.  Other formats may be rejected without reading.  You have been
warned!  A one- or two-paragraph summary plus URL gets you a better
announcement than an entire press release.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Linux 2.4.3
</FONT>
</H3> 

<P> Linux 2.4.3 is out.  See the 
<A HREF="http://www.kernel.org/pub/linux/kernel/v2.4/ChangeLog-2.4.3">
changelog</A> or find a 
<A HREF="http://www.kernel.org/mirrors/">kernel mirror</A>.



<P> <hr> <P> 
<!-- =================================================================== -->
<!-- MMMMMMMMMMMMMMMMM -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<font color="green">
<I>Linux Journal</I> and <I>Embedded Linux Journal</I>
</font>

</H3>
<center>
<IMG ALT=" "   SRC="misc/bytes/lj-cover84.png" WIDTH=200 HEIGHT=268 HSPACE="10">
<IMG ALT=" "   SRC="misc/bytes/elj-cover2.jpg" WIDTH=200 HEIGHT=268 HSPACE="10">
</center>


<P> 
The April issue of <A HREF="http://www.linuxjournal.com/"><I>Linux
Journal</I></A> is on newsstands now.
This issue focuses on Internet/Intranet.  Click 
<A HREF="http://www.linuxjournal.com/lj-issues/issue84/index.html">here</A>
to view the table of contents, or 
<A HREF="http://www.linuxjournal.com/subscribe/index.html">here</A> 
to subscribe.
<FONT COLOR="green">All articles through December 1999 are available for
public reading at 
<A HREF="http://www.linuxjournal.com/lj-issues/mags.html">http://www.linuxjournal.com/lj-issues/mags.html</A></FONT>.  
Recent articles are available on-line for subscribers only at 
<A HREF="http://interactive.linuxjournal.com">
http://interactive.linuxjournal.com/</A>.  

<P> The March/April issue of 
<A HREF="http://embedded.linuxjournal.com"><I>Embedded Linux Journal</I></A> was
mailed to subscribers in February.  
Click 
<A HREF="http://embedded.linuxjournal.com/magazine/issue02/">here</A>
to view the table of contents.  Professionals working in the embedded field
in the US, Canada or Mexico can get a <EM>free</EM> subscription by clicking
<A HREF="http://embedded.linuxjournal.com/subscribe/">here</A>.  Paid
subscriptions to other countries are also available.

<BR CLEAR="all">


<a name="distro"></a>
<p><hr><p>
<!-- =================================================================== -->
<center><H3><font color="green">Distro News</font></H3></center>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Caldera
</FONT>
</H3> 

<P> <a href="http://www.calderasystems.com">
Caldera Systems</a> have announced the open beta availability of its new
OpenLinux server product, code-named "Project 42," and an agreement with
Lutris to ship the Enhydra Open Source Java/XML application server with the
new version. The product is based on the new Linux 2.4 kernel and targets
OEMs and VARs.
Project 42 incorporates a secure Web server, a file and print server, and a
set of network infrastructure servers, including DHCP, DNS, and firewall.

<hr noshade width="20%">
<p>
Caldera and <a href="http://www.sco.com">
SCO</a> have unveiled Open UNIX 8, incorporating support for Linux
applications.
Open UNIX 8 will maintain compatibility and continuity with the UnixWare 7
operating system while providing a complete Linux environment. In addition,
the product will incorporate support for the execution of unmodified Linux
Intel Architecture binaries.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Debian
</FONT>
</H3> 

<P> <A HREF="http://www.debian.org/">Debian</A> has chosen Ben Collins as the
new Debian Project Leader (DPL).


<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Progeny Debian
</FONT>
</H3> 

<P> Progeny Linux Systems have announced that
Release Candidate 1 of Progeny Debian is now available for
download.  Progeny Debian is based on woody, the current testing
version of Debian,  and made by a team of leading Debian developers.
Company CEO Ian Murdock, has said that he expects any changes after RC1 to
be bug-fixes and cosmetic improvements

<P> Features of RC1 include:

graphical installation and configuration tools,
a GNOME interface for debconf,
improved hardware detection and USB support,
optional migration to GRUB,
 and automated multiple installations.

<P>  Software included in RC1 includes:
<ul>
    <li>Kernel 2.2.18 (2.4 optional)
    <li>XFree 86 4.0.2.
    <li>glibc 2.2.
    <li>Netscape 4.76 and Mozilla 0.7.
    <li>GNU parted
    <li>Open SSH.
    <li>GRUB 0.5.96.1
</ul>

<P> For more information about Progeny Debian, visit 
<a href="http://www.progeny.com/">www.progeny.com</a>.

<a name="general"></a>



<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">SuSE
</FONT>
</H3> 
<p>
<p>
<a href="http://www.suse.de/en/">SuSE Linux</a> have released apparently
very positive news about the future of Linux in Germany. A poll commissioned
by SuSE showed that 56 percent of the interviewed PC users have heard of
Linux and ten percent already use the alternative operating system at home
or at work. This statistic indicates that, in terms of distribution, Linux
is second to Windows. Furthermore, 23 percent of the computer users consider
switching to Linux when upgrading their equipment.  This information was
obtained from a survey recently conducted by the market research institute
TNS EMNID, Bielefeld, Germany.
<hr noshade width="20%">
<a href="http://www.suse.de/en/">SuSE Linux</a> 
are now offering a new Server Version for professional users.
SuSE presented the SuSE Linux Enterprise Server at CeBIT in
Hanover, Germany. 
SuSE Linux Enterprise Server  is
an operating system streamlined for the utilization in servers. 
It has been optimized for security and stability and
comprises all relevant server services.

<hr noshade width="20%">
SuSE's PowerPC Edition 7.1 will be released in early April.  It has k
ernel 2.4.2 and
ALSA (Advanced Linux Sound Architecture) for PowerMacs.  SuSE's administration
and configuration tool YaST2 is complemented by YOU (YaST Online Update) for
updating individual packages after the install.  

<P> It also has KDE2, XFree86
4.0.2 and SaX2, the expanded graphical configuration tool which ensures a
simple and secure setup of supported graphics cards, is also a new feature. An
improved version of MOL (Mac on Linux), the virtual machine used to start MacOS
in Linux, complements the distribution.

<P> The range of supported IBM computers with PowerPC processors has been
considerably enlarged. SuSE Linux 7.1 PowerPC Edition now runs on IBM
Power3 machines. The possibility to use up to 3 GB RAM and an expanded
multi-processor support provided by Kernel 2.4 make SuSE Linux 7.1
PowerPC Edition especially attractive for IBM pSeries 640. Thus, SuSE
Linux 7.1 PowerPC Edition is the first Linux solution that supports these
computers "out of the box".

<P> The package includes 6 CDs, a 500-page manual and 60 days installation
support for EURO 49.00.


<p><hr><p>
<!-- =================================================================== -->
<center><H3><font color="green">News in General</font></H3></center>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Lion worm (DNS/BIND security alert!)
</FONT>
</H3> 

<P> 
Anyone using BIND should be aware that there is a new worm on the loose. 
The Lion worm attacks certain versions of BIND (the domain name
server program).
<a href="http://www.sans.org/">The Sans Institute</a> have plenty of 
<a href="http://www.sans.org/y2k/lion.htm">information</a> on the worm, and
indicate that Bind versions 8.2, 8.2-P1, 8.2.1, 8.2.2-Px are vulnerable.
BIND 8.2.3-REL has been reported as not being
vulnerable (this information is preliminary and potentially incomplete). 
The BIND vulnerability is the TSIG vulnerability
that was reported back on January 29, 2001.

If you believe your system has been compromised,
the SANS Institute has a program
<A HREF="http://www.sans.org/y2k/lion.htm">Lionfind</A> that detects
it.
Now is a good time to get the latest version of BIND from your
distribution vendor, run named as non-root, or switch to a
BIND alternative.
<p>
It is also worth looking at general security issues. To get an idea of how
security should be done, check out 
<a href="http://project.honeynet.org/challenge/results/">
the results</a> of the 
<a href="http://project.honeynet.org/">Honeynet
</a> forensic challenge. Candidates downloaded the partition images
of a compromised Linux system and had to find out "who, what, when,
where, how". The results show how professionals go about doing these things,
but also how difficult and time consuming recovering from a compromise can
be. The lesson is "BE PREPARED!"


<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Upcoming conferences and events
</FONT>
</H3> 

<P> Listings courtesy <EM>Linux Journal</EM>.  See <EM>LJ</EM>'s
<A HREF="http://noframes.linuxjournal.com/events/">Events</A> page for the
latest goings-on.

<!-- *** BEGIN events table [this line needed by Linux Gazette events.py *** -->

<table cellpadding=5 border=0 width=100%>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top> 
	<b>LINUX Business Expo</b><BR>
	<td valign=top>April 2-5, 2001<BR>Chicago, IL<BR>
	<A HREF="http://www.linuxbusinessexpo.com/" target=_blank>
	http://www.linuxbusinessexpo.com</A><BR>
	</td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Free Web ROI Seminar by Akami Technologies</b><BR>
        <td valign=top>April 3, 2001<BR>Seattle, WA<BR>
        <A HREF="http://www.akamai.com/roitime/" target=_blank>
        http://www.akamai.com/roitime/</A><BR>
        </td></tr>                                                      

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Expo, Madrid</b><BR>
        <td valign=top>April 4-5, 2001<BR>Madrid, Spain<BR>
        <A HREF="http://www.linuxexpomadrid.com/EN/home/" target=_blank>
        http://www.linuxexpomadrid.com/EN/home</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Lugfest IV</b><BR>
        <td valign=top>April 21-22, 2001<BR>Simi Valley, CA<BR>
        <A HREF="http://www.lugfest.org/" target=_blank>
        http://www.lugfest.org</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Expo Road Show</b><BR>
        <td valign=top>April 23-27, 2001<BR>Various Locations<BR>
        <A HREF="http://www.linux-expo.com/" target=_blank>
        http://www.linux-expo.com</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Africa 2001</b><BR>
        <td valign=top>April 24-26, 2001<BR>Johannesburg, South Africa<BR>
        <A HREF="http://www.aitecafrica.com" target=_blank>
        http://www.aitecafrica.com</A><BR>
        </td></tr>                      

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Open Source Development Network Summit</b><BR>
        <td valign=top>April 30 - May 1, 2001<BR>Austin, TX<BR>
        <A HREF="http://osdn.com/conferences/handhelds/" target=_blank>
        http://osdn.com/conferences/handhelds/</A><BR>
        </td></tr>                              

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux for Industrial Applications</b><br>3rd Braunschweiger
	Linux-Tage<BR>
        <td valign=top>May 4-6, 2001<BR>Braunschweig, Germany<BR>
        <A HREF="http://braunschweiger.linuxtage.de/industrie/" target=_blank>
        http://braunschweiger.linuxtage.de/industrie/</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux@Work Europe 2001</b><BR>
        <td valign=top>May 8 - June 15, 2001<BR>Various Locations<BR>
        <A HREF="http://www.ltt.de/linux_at_work.2001/" target=_blank>
        http://www.ltt.de/linux_at_work.2001</A><BR>
        </td></tr>  

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Expo, S&atilde;o Paulo</b><BR>
        <td valign=top>May 9-10, 2001<BR>S&atilde;o Paulo, Brazil<BR>
        <A HREF="http://www.linux-expo.com" target=_blank>
        http://www.linux-expo.com</A><BR>
        </td></tr>   

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>SANS 2001</b><BR><td valign=top>May 13-20, 2001<BR>
	Baltimore, MD<BR>
        <A HREF="http://www.sans.org/SANS2001.htm" target=_blank>
       	http://www.sans.org/SANS2001.htm</A><BR>
        </td></tr>   

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>7th Annual Applied Computing Conference</b><BR>
        <td valign=top>May 14-17, 2001<BR>Santa Clara, CA<BR>
        <A HREF="http://www.annatechnology.com/annatech/HomeConf2.asp" 
	target=_blank>
        http://www.annatechnology.com/annatech/HomeConf2.asp</A><BR>
        </td></tr>  

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Expo, China</b><BR>
        <td valign=top>May 15-18, 2001<BR>Shanghai, China<BR>
        <A HREF="http://www.linux-expo.com/" target=_blank>
        http://www.linux-expo.com</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr> 

<tr><td valign=top>
        <b>SITI International Information Technologies Week</b><br>
	OpenWorld Expo 2001<BR>
        <td valign=top>May 22-25, 2001<BR>Montr&eacute;al, Canada<BR>
        <A HREF="http://www.mediapublik.com/en/" target=_blank>
        http://www.mediapublik.com/en/</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr> 

<tr><td valign=top>
	<b>Strictly e-Business Solutions Expo</b><BR>
	<td valign=top>May 23-24, 2001<BR>Minneapolis, MN<BR>
	<A HREF="http://www.strictlyebusinessexpo.com/" target=_blank>
	http://www.strictlyebusinessexpo.com</A><BR>
	</td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Linux Expo, Milan</b><BR>
        <td valign=top>June 6-7, 2001<BR>Milan, Italy<BR>
        <A HREF="http://www.linux-expo.com" target=_blank>
        http://www.linux-expo.com</A><BR>
        </td></tr>   

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
	<b>USENIX Annual Technical Conference</b><BR>
	<td valign=top>June 25-30, 2001<BR>Boston, MA<BR>
	<A HREF="http://www.usenix.org/events/usenix01/" target=_blank>
	http://www.usenix.org/events/usenix01</A><BR>
	</td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
	<b>PC Expo</b><BR>
	<td valign=top>June 26-29, 2001<BR>New York, NY<BR>
	<A HREF="http://www.pcexpo.com/" target=_blank>www.pcexpo.com</A><BR>
	</td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>     
	<b>Internet World Summer</b><BR><td valign=top>July 10-12, 2001<BR>
	Chicago, IL<BR>
	<A HREF="http://www.internetworld.com/" target=_blank>
	http://www.internetworld.com</A><BR>
	</td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
	<b>O'Reilly Open Source Convention</b><BR>
	<td valign=top>July 23-27, 2001<BR>San Diego, CA<BR>
	<A HREF="http://conferences.oreilly.com/" target=_blank>
	http://conferences.oreilly.com</A><BR>
	</td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>10th USENIX Security Symposium</b><BR>
        <td valign=top>August 13-17, 2001<BR>Washington, D.C.<BR>
        <A HREF="http://www.usenix.org/events/sec01/" target=_blank>
        http://www.usenix.org/events/sec01/</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>HunTEC Technology Expo & Conference</b><br>Hosted by Hunstville IEEE<BR>
        <td valign=top>August 17-18, 2001<BR>Huntsville, AL<BR>
	URL unkown at present<br>
        </td></tr>    

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Computerfest</b><BR>
        <td valign=top>August 25-26, 2001<BR>Dayton, OH<BR>
        <A HREF="http://www.computerfest.com/" target=_blank>
        http://www.computerfest.com</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top> 
	<b>LinuxWorld Conference & Expo</b><BR>
	<td valign=top>August 27-30, 2001<BR>San Francisco, CA<BR>
	<A HREF="http://www.linuxworldexpo.com" target="_blank">
	http://www.linuxworldexpo.com</A><BR>
	</td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>The O'Reilly Peer-to-Peer Conference</b><BR>
        <td valign=top>September 17-20, 2001<BR>Washington, DC<BR>
        <A HREF="http://conferences.oreilly.com/p2p/call_fall.html" 
	target=_blank>
        http://conferences.oreilly.com/p2p/call_fall.html</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
	<b>Linux Lunacy<br>Co-Produced by <i>Linux
	Journal</i> and Geek Cruises</b><BR>
        <A HREF="http://www.linuxjournal.com/lunacy/cruise.html" target=_blank>
        Send a Friend <i>LJ</i> and Enter to Win a Cruise!</A><BR> 
	<td valign=top>October 21-28, 2001<BR>Eastern Caribbean<BR>
	<A HREF="http://www.geekcruises.com/" target=_blank>
	http://www.geekcruises.com</A><BR>
	</td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>LinuxWorld Conference & Expo</b><BR>
        <td valign=top>October 30 - November 1, 2001<BR>Frankfurt, Germany<BR>
        <A HREF="http://www.linuxworldexpo.de/linuxworldexpo/index.html" 
	target=_blank>
        http://www.linuxworldexpo.de/linuxworldexpo/index.html</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>5th Annual Linux Showcase & Conference</b><BR>
        <td valign=top>November 6-10, 2001<BR>Oakland, CA<BR>
        <A HREF="http://www.linuxshowcase.org/" target=_blank>
        http://www.linuxshowcase.org/</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>Strictly e-Business Solutions Expo</b><BR>
        <td valign=top>November 7-8, 2001<BR>Houston, TX<BR>
        <A HREF="http://www.strictlyebusinessexpo.com/" target=_blank>
        http://www.strictlyebusinessexpo.com</A><BR>
        </td></tr> 

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>  

<tr><td valign=top>
        <b>LINUX Business Expo</b><BR>Co-located with COMDEX<br>
        <td valign=top>November 12-16, 2001<BR>Las Vegas, NV<BR>
        <A HREF="http://www.linuxbusinessexpo.com" target=_blank>
        http://www.linuxbusinessexpo.com</A><BR>
        </td></tr>

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr>

<tr><td valign=top>
        <b>15th Systems Administration Conference/LISA 2001</b><BR>
        <td valign=top>December 2-7, 2001<BR>San Diego, CA<BR>
        <A HREF="http://www.usenix.org/events/lisa2001/" target=_blank>
        http://www.usenix.org/events/lisa2001</A><BR>
        </td></tr>  

<tr><td colspan=2><HR size=5 width=100% noshade align=center></td></tr> 


</table>
<!-- *** END events table [this line needed by Linux Gazette events.py *** -->



<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">LinuxFocus
</FONT>
</H3> 

<P> <A HREF="http://www.linuxfocus.org">LinuxFocus</A> is a Linux webzine
that's been around for years, but may not be familiar to some <I>LG</I> readers.
Unlike <I>LG</I>, which is essentially in English with some foreign-language
translations, <I>LF</I> was founded with the goal of providing non-English
speakers with "enough [Linux] information in their native
language that they can join in the Linux community."  Currently, seven 
languages are fully supported and four more are partially supported.  
Translations happen both ways: there are currently six French articles waiting
to be adopted by English translators.  <I>LG</I> fully supports <I>LF</I> and
wishes it success.



<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">NetworX and AMD Supply Cluster to Boeing
</FONT>
</H3> 

<P>The Boeing Company is using a 
<a href="http://www.linuxnetworx.com/">
Linux NetworX</a>
cluster powered by 96 
<a href="http://www.amd.com/">
AMD</a>
Athlon
processors. The system, designed as a high performance cluster, is being
used by Boeing Space & Communications in Huntington Beach, Calif. to run
computational fluid dynamics applications in support of the Delta IV
program. Boeing Delta IV engineers tested multiple processor platforms at
Linux NetworX facilities prior to buying the cluster, and selected the AMD
Athlon for its price and performance advantages.  

<hr noshade width="20%">

<P>Linux NetworX has also announced the  development of LinuxBIOS for the 
Alpha platform. 
In conjunction with the LinuxBIOS Open Source project, Linux NetworX has
replaced SRM firmware on the Alpha platform with a Linux-based BIOS. Users
will now have the ability to boot to Linux directly out of the ROM on the
motherboard.
<P>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Python Software Foundation & Python Cookbook
</FONT>
</H3> 

<P> 
<a href="http://activestate.com/">
ActiveState</a> have announced involvement in 
the launch of a collaborative programming book, the
Python Cookbook, with O'Reilly &amp; Associates.  The Cookbook will be a
repository of reviewed Python recipes that have been contributed
by the Python community for the community. It will be freely available
for download.  For details please go to the 
<a href="http://activestate.com/Corporate/Communications/Releases/Press983583358.html">
website</a>.

Activestate will also be a founding co-sponsor of the new Python Software
Foundation (PSF).  The PSF's is to
provide educational, legal and financial resources to the Python community.
More information is available in the full
<a href="http://activestate.com/Corporate/Communications/Releases/Press983863928.html">
press release</a>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Penguin Computing Selects Arkeia Backup for Linux Servers
</FONT>
</H3> 

<P>
<a href="http://www.arkeia.com">
Knox Software Corp.</a> have announced that
the company has entered into a reseller agreement with Penguin Computing
Inc. Under the agreement, Penguin Computing will now
offer Knox's flagship network backup application, Arkeia, for bundling
with Penguin's pre-configured custom Linux servers. 

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">IBM, Biotech and Linux
</FONT>
</H3> 

<a href="http://www.ibm.com/">IBM</a> has been very active on
the <a href="http://ibm.com/linux">Linux</a> front in recent months.
IBM efforts in the emerging 
<a href="http://www-4.ibm.com/software/webservers/lifesciences/">
biotechnology</a>
marketplace received a boost
with the  announcement that 
<a href="http://www.strubix.com/">Structural Bioinformatics</a>
 has chosen DB2 for Linux as its strategic development platform
for future applications. DB2 will be used to manage more
than two terabytes of high-resolution 3-D protein structures, which are used
in the development of new medicines.
<p> For more information on IBM's Linux developments, refer to their 
<a href="http://ibm.com/linux">website</a>.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">OEone Teams Up with EarthLink
</FONT>
</H3> 

<P> 
<a href="http://www.oeone.com/">OEone</a>
and 
<a href="http://www.earthlink.net">EarthLink</a>
are
<a href="http://www.oeone.com/index.php?SCREEN=news/announcements">
working together</a> to integrate EarthLink's Linux based
Internet access software with
OEone's Internet-computer operating environment platform. 

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">TeamLinux and Muze to Expand Relationship
</FONT>
</H3> 
<P> 
<a href="http://www.team-linux.com">
TeamLinux</a> 
is to expand its relationship with
<a href="http://www.muze.com">Muze Inc.</a>
						
to provide hardware support and
service for its existing and future kiosk customers. 
This new multi-year contract provides for TeamLinux to be
the premiere
provider for all hardware and service. Muze will continue to provide its
proprietary software and be the first level of contact for any Muze system
issues. 

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Sair Linux & GNU Certification
</FONT>
</H3> 
Sair Linux have announced their new 
<a href="http://www.linuxcertification.com">website</a>
The Web  site  has  information about
Sair Linux and GNU Certification.
The company provides training, certification, and educational aids (books
etc.,).
Sair also invites LUG's to sign up and receive a "Welcome Kit" including
T-shirts, brochures, information on Sair Linux, Distro CD's.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Agenda Computing Sell Linux PDA
</FONT> </H3> 

<P>California-based
<a href="http://www.agendacomputing.com/">
Agenda Computing</a> are launching a pure
Linux PDA (personal digital assistant)
to challenge Palm in the war for market share. 
Each Agenda VR3 and VR3r is loaded with unique software and hardware
features like 16MB of Flash memory, which eliminates the problem of data
loss associated with RAM-based units. It also supports 7 languages, is
e-mail compatible, and will send a
memo or message to a printer by wireless infrared transfer. 

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Keyspan Ships 4-port USB Serial Adapter for Linux</FONT>
</H3> 
The 
<a href="http://www.keyspan.com">
Keyspan</a> USB 4-Port Serial Adapter is intended to allow
4 serial devices to be connected to a single USB port. Each of its male DB9
ports allows connection to RS232 serial devices at data rates up to 960
Kbps. In addition to supporting Linux 2.4, the Keyspan USB 4-Port Serial
Adapter works with various Windows flavours. (Note, this is not a review,
consult the company for full details.)

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Linux Links
</FONT>
</H3> 

<P> <A HREF="http://www.galleo.com/">Galleo</A> is a mobile multimedia
communicator.  It's a nifty-looking PDA with e-mail, web and music
capabilities.  Unfortunately, their web site is not so nifty: I can't get the
menu buttons to show.  So click on the Galleo image or try this link to their
<A HREF="http://www.galleo.com/products.html">follow this link</A> to get to
the products page and use the text links from there, and click on "Virtual
Tour".  (Requires Javascript and who knows what [Shockwave?] for the movie.)

<P> The 
<a href="http://www.thedukeofurl.org">
Duke of URL</a>
has the following to offer:

<ul>
<li> <a href="http://www.thedukeofurl.org/reviews/misc/trustix12">
Review of Trustix Secure Linux 1.2</a>
<li><a href="http://www.thedukeofurl.org/reviews/misc/linuxguide9">
Linux Buyer's Guide #9</a>
<li><a href="http://www.thedukeofurl.org/reviews/misc/tribes2preview">
Tribes 2 Linux Preview</a>
<li> <a href="http://www.thedukeofurl.org/reviews/misc/turbolinux61">
TurboLinux Workstation Pro 6.1 Review</a>
<li> <a href="http://www.thedukeofurl.org/reviews/misc/linuxapps">
A Beginner's Guide to Linux Apps</a>
</ul>

<P>


Some links from the folks at ZDNet's
<a href="http://www.anchordesk.co.uk/">Anchordesk UK</a>
<ul>

<li> <a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7107682,00.html">
Linux is too much</a>: a look at how big Linux distributions packed with
thousands of apps are likely to leave the average desktop user dazed and
confused.

<li> 
Two looks at the new kernel:
<a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7107784,00.html">
Linux 2.4: It's here! Now what?</a>  and 
<a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7107835,00.html">
Much ado about kernels</a> 

<li>
<a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7108709,00.html">
How to succeed at selling free software</a>

<li>
<a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7108822,00.html">
Why Linux is like pizza</a>

<li>
<a href="http://www.anchordesk.co.uk/anchordesk/commentary/columns/0,2415,7108937,00.html">
Nautilus: here today. Eazel: gone tomorrow?</a>
</ul>

<P> <A HREF="http://debianhelp.org">debianhelp.org</A> offers, um, help on
Debian.

<p>
<a href="http://www.linuxvalley.com/">
Linux Valley</a>,
an Italian portal for the Linux operating system,
has been updated.
It offers a range of interactive and community services.

<p>
Microsoft says Linux is a 
<a href="http://news.cnet.com/investor/news/newsitem/0-9900-1028-4825719-RHAT.html?tag=ltnc">
threat</a>
to intellectual-property rights. <I>Linux Journal</I> 
<a href="http://noframes.linuxjournal.com/articles/conversations/020.html">
disagrees</a>.
<p>
<a href="http://www.lwn.net/2001/0322/a/paranoid-backup.php3">
Paranoid Backup</a>
is designed to "work with cheap tape drives and cheap tapes
without shoe-shining or losing data; to never overwrite old backups; and to use
as few tapes as possible."
<P> The Pentagon's research agency is preparing to demonstrate a 
<a href="http://www.eet.com/story/OEG20010321S0049">soldier's radio</a>
 designed to provide mobile communications among individual
troops on the battlefield. The network
will be based on the Linux operating system.
Courtesy Slashdot.
<P>
The Linux Expo Birmingham 2001 web site
is now 
<a href="http://www.linuxexpobirmingham.com">
online</a>.  For information on other Linux-Expo events, consult their
<a href="http://www.linux-expo.com/">website</a>.
<p>
<a href="http://www.olinux.com/">OLinux</a> have an  
<a href=http://www.olinux.com.br/artigos/279/1.html>interview</a>
with  Rick Lehrbaum from  LinuxDevices.com. OLinux are also currently
looking for an investor or a company willing to translate and promote OLinux
around the world.
<p>Doug Eubanks has put together a new Linux/RoadRunner 
<a href="http://dougemt.no-ip.com">help site</a>. He aims to consolidate
the various threads in the field.
<a name="software"></a>

<P> An article on 
<A HREF="http://www.internetweek.com/newslead01/lead032901.htm">
Microsoft's complicated licensing terms</A> for enterprise users.
The title for the Slashdot link is, "Microsoft Turning Screws on Customers".

<P> <hr> <P> 
<!-- =================================================================== -->
<center><H3><font color="green">Software Announcements</font></H3></center>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Tom's Root/Boot Updated
</FONT>
</H3> 

<P> Tom Oehser has released a minor, but recommended, update to tomsrtbt.
Current version is now 1.7.218.
Get it from: <a href="http://www.toms.net/rb/">http://www.toms.net/rb/</a>.
This is something everybody should have on hand in case you someday have to
boot from an emergency floppy.



<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">AbsoluteX Now Available for Download
</FONT>
</H3> 

AbsoluteX, "Linux With a Twist," is available for download at
<a href="http://www.absolutex.org/">
www.absolutex.org</a>. 
It was unveiled at the Annual Linux Showcase in Atlanta
and is now available to developers worldwide under LGPL.

AbsoluteX is an X-Window developer toolkit created by 
<a href="http://www.epitera.com">
Epitera</a>
to streamline and facilitate the process of developing
customized GUIs for the Linux operating system. Based on the C++
programming language, AbsoluteX is a standard template library (STL) with
multiple inheritance methodology, efficient messaging, and programming
methods that separate logical and visual aspect class libraries. 

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Loki Games
</FONT>
</H3> 

<P> <a href="http://www.lokigames.com/">Loki Software</a> has  
announced an agreement with developer-driven computer and 
videogame publisher Gathering of Developers to bring the hit PC games 
<a href="http://www.runegame.com">
Rune</a> and 
<a href="http://fakk2.godgames.com">
Heavy Metal: F.A.K.K.2</a> to Linux early this year. Testers are required,
<a href="http://odin.lokigames.com">
register here</a>.

<P> Furthermore,
in a race to GPL freedom, Loki Software, Inc. are releasing
the latest in their line of open source projects: a complete set of
end-user and developer tools for managing software releases.
<ul>
<li><a href="http://www.lokigames.com/development/setup.php3">Setup 1.5.8</a>
includes support for Loki's new
XML product database.  This database allows you to 
determine if your product is complete or corrupt.
<li><a href="http://www.lokigames.com/development/loki_update.php3">Loki Update Tool 1.0.6</a>
allows quick and easy
updates of any Loki software. 
<li><a href="http://www.lokigames.com/development/loki_uninstall.php3">Loki Uninstall Tool 1.0.3</a>
allows anyone to quickly and easily
remove any Loki software they install.
<li><a href="http://www.lokigames.com/development/loki_patch.php3">Loki Patch Tools 1.0.1</a>
will allow you to build patches for
any project, complete with pre/post install scripts.
</ul>

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">"Emerald Isle" Ispell
</FONT>
</H3> 

A new version of the package 
<a href="http://borel.slu.edu/~kps/ispell">"ispell-gaeilge"</a> 
which lets users of
International Ispell check their Irish Gaelic spelling has been released
recently. Developed by Kevin P. Scannell, a mathematician at Saint Louis
University, the package boasts a dictionary of over 200,000 entries
covering the many grammatical variations of Irish language words.

Mentions of it in Irish national newspapers have introduced many
people to the world of Linux for the first time. For people frustrated with
the lack of support for minority languages in Windows, projects such as the
<a href="http://www.linux.ie/projects/mandrake-ga/">
Irish localisation
</a> of
<a href="http://www.linux-mandrake.com">
Mandrake Linux</a> offer real encouragement.
Other projects, such as the spelling checker 
<a href="http://www.gaelspell.com/">GaelSpell</a>,
are improving the tools available to Windows users, but also help all computer
users by providing quality word-lists.


<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">GARLIC Version 1.1 Released
</FONT>
</H3> 

Version 1.1 of garlic, free GPL molecular visualisation program for Linux and
Unix, is available at 
<a href="http://pref.etfos.hr/garlic">
http://pref.etfos.hr/garlic</a>. 
It aims to be ANSI and POSIX compliant and may
be easily ported to different Unix-like platforms.

Garlic may be used to analyse proteins, DNA and other large molecules.
The latest version includes a number of plot-options, 
not available in version 1.0:
helical wheel, averaged hydrophobicity, hydrophobic moment, Venn diagram and
Ramachandran plot. A screenshot
<a href="http://pref.etfos.hr/garlic/gallery">
gallery</a> is available.


<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">The FIEN Group to Sell 
Teamware Office for Linux in the U.S.
</FONT>
</H3> 

<P> <a href="http://www.teamware.com">Teamware Group
</a>, a 
<a href="http://www.fujitsu.com">Fujitsu
</a>
subsidiary, and 
<a href="http://www.thefiengroup.com">
The FIEN Group</a>, a Southern
California-based technology consulting organisation have signed a partner
agreement according to which The FIEN group will sell Teamware Office 5.3 for
Linux groupware to customers across the USA.
Teamware Office 5.3 for Linux includes facilities for electronic mail,
time and resource scheduling, discussion groups as well as document storage
and retrieval. The famous Teamware Office groupware suite has
been in the market since 1989 and was ported to Linux platform in spring
2000.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Opera to be released as ad-ware
</FONT>
</H3> 

<P> 
<a href="http://www.opera.com">
Opera Software
</a> has announced that the final
release  of its Linux browser will be available for free to
all users.  The free version has full functionality but contains banner
ads.  If you don't want banner ads, you can register your free copy for $39, or
buy the adless version for the same price.  If you are interested in this
product, Opera 5 for Linux beta 7 is 
<a href="http://www.opera.com/linux/">
now out</a>.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Open Motif Now Supports Latest Linux 2.4 Kernel Distributions
</FONT>
</H3> 

<P> <a href="http://www.ics.com">
Integrated Computer Solutions</a> has released an updated
version of Open Motif Everywhere. This new release officially incorporates
Open Group Patch 3 and Patch 4 into the Open Motif release. These patches
include bug fixes and updates to the Motif libraries, clients and
the demo source code. RPM (version 4) are also provided for both Red Hat
Linux 7, SuSE Linux 7.1 and other distributions using glibc 2.2. 
The latest ICS
Open Motif binary and source packages are available for free download at
ICS's Motif Community site, 
<a href="http://www.motifzone.net">the MotifZone</a>. They
are also available for $29.95 on ICS's Open Motif Everywhere distribution CD
that can be purchased through 
<a href="http://store.ics.com">the ICS Store</a>.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Kaspersky Lab Introduces the New Version of 
Kaspersky Anti-Virus for Linux
</FONT>
</H3> 

<P> Kaspersky Lab  have announced the release of the new version of Kaspersky
Anti-Virus for Linux (3.0 Build 135.3).
This new version adds several features, including 
installer support for different Linux
distributions,
and a ready-made
solution to integrate centralised virus filtering for Postfix e-mail
gateways.
The new version of Kaspersky Anti-Virus is available for 
<a href="http://www.kaspersky.com/download.asp">download</a> from
the Kaspersky Lab Web site. 
All registered users of previous versions of Kaspersky Anti-Virus for Linux
may upgrade to the new version free of charge.

<P> <hr> <P> 
<!-- =================================================================== -->
<H3><IMG ALT=" "   SRC="../gx/bolt.gif">
<FONT COLOR="green">Other software
</FONT>
</H3> 

<p>Rob Pitman has released a LGPL licensed software package  that provides
a "graphical user interface" between a Java application and an ASCII
terminal.  The package emulates the API of the Java AWT 
and the Swing toolkit. It provides "graphical" widgets such as
Frames, Dialogs, Labels, TextFields and Buttons. One can design the GUI of
an application using any Java IDE and then port it to use a text
interface with little work.
You can get more information about the package at:
<a href="http://www.pitman.co.za/projects/charva/index.html">
http://www.pitman.co.za/projects/charva/index.html</a>.

<hr noshade width="20%">

<p>
<a href="http://mahogany.sourceforge.net/">Mahogany
</a> Version 0.62 is out.
Mahogany is an OpenSource(TM) cross-platform mail and news client. It
supports a range of protocols and standards (POP3, IMAP4, MIME, etc.,),
secure communications via SSL, and can be extended  using its built-in
Python interpreter and loadable modules.

<P> <A HREF="http://www.tuxia.com/">TUXIA</A> specialises in embedded Linux
software suites for Internet and Information appliances.  TASTE (TUXIA
Appliance Synthesis Technology Enabled) is a solution based on Linux Kernel 2.4
with an embedded Mozilla browser and other functionalities, that can be
integrated into any hardware platform.



<!-- *** BEGIN copyright *** -->
<P> <hr> <P> 
<H5 ALIGN=center>
Copyright &copy; 2001, Michael Conry and 
the Editors of <A HREF="mailto:gazette@ssc.com"><I>Linux Gazette</I></A>.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR>
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->


<center>
<H1><A NAME="answer">
	<img src="../gx/dennis/qbubble.gif" alt="(?)" 
		border="0" align="middle">
	<font color="#B03060">The Answer Gang</font>
	<img src="../gx/dennis/bbubble.gif" alt="(!)" 
		border="0" align="middle">
</A></H1> 
<BR>
<H4>By Jim Dennis, Ben Okopnik, Dan Wilder, Breen, Chris, and the Gang,
	the Editors of Linux Gazette... 
	and You!
<br>Send questions (or interesting answers) to
	<a href="mailto:linux-questions-only@ssc.com">linux-questions-only@ssc.com</a>
</H4>
<p><em><font color="#990000">There is no guarantee that your questions
	here will <b>ever</b> be answered.  You can be published anonymously 
	- just let us know!
</font></em></p>
</center>

<p><hr><p>
<!--  endcut ======================================================= -->
<H3>Contents:</H3>
<dl>
<dt><a href="#tag/greeting"
	><strong>&para;: Greetings From Heather Stern</strong></A></dl>

<DL>
<!-- index_text begins -->
<dt><A HREF="#tag/1"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	></a>? --or--
<dd><A HREF="#tag/1"
	><strong>What's this word?</strong></a>

<dt><A HREF="#tag/2"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Unable to Install Linux</strong></a>
<dt><A HREF="#tag/3"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>DNS and telnet</strong></a>
<dt><A HREF="#tag/4"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Help on LILO stopping at LI</strong></a>
<dt><A HREF="#tag/5"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>How can you do a recursive search to find broken symbolic links?</strong></a>
<dt><A HREF="#tag/6"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>BIOS passwords - Bane of my existance</strong></a>
<dt><A HREF="#tag/7"
	><img src="../gx/dennis/bbub.gif" height="28" width="50"
	  alt="(!)" border="0"
	></a>telecommunication in a hospital --or--
<dd><A HREF="#tag/7"
	><strong>Making the Connection</strong></a>
<dt><A HREF="#tag/8"
	><img src="../gx/dennis/bbub.gif" height="28" width="50"
	  alt="(!)" border="0"
	><strong>Setup of Microsoft Outlook Express 5 for Sending of Clear Text</strong></a>
<dt><A HREF="#tag/9"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>icons</strong></a>
<dt><A HREF="#tag/10"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Corrupt Tar Archive</strong></a>
<dt><A HREF="#tag/11"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>masquerade in sendmail is broken.</strong></a>
<dt><A HREF="#tag/12"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>neighbour table overflow</strong></a>
<dt><A HREF="#tag/13"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>VIDEO CARD</strong></a>
<dt><A HREF="#tag/14"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>fat versus inodes</strong></a>
<dt><A HREF="#tag/15"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Installing Linux without cdrom</strong></a>
<dt><A HREF="#tag/16"
	><img src="../gx/dennis/bbub.gif" height="28" width="50"
	  alt="(!)" border="0"
	><strong>Installing RedHat 7.0 and a driver for the Chipset Cirrus CL-GD5436</strong></a>
<dt><A HREF="#tag/17"
	><img src="../gx/dennis/bbub.gif" height="28" width="50"
	  alt="(!)" border="0"
	><strong>cd-writing mini-howto</strong></a>
<dt><A HREF="#tag/18"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	></a>These has been bugging me for a while now --or--
<dd><A HREF="#tag/18"
	><strong>Reading the logs</strong></a>

<dt><A HREF="#tag/19"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	></a>Linux Newbie Frustration --or--
<dd><A HREF="#tag/19"
	><strong>So many users, So few POP accounts</strong></a>

<dt><A HREF="#tag/20"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>script</strong></a>
<dt><A HREF="#tag/21"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Linux Box on windows</strong></a>
<dt><A HREF="#tag/22"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	></a>HD bad clusters --or--
<dd><A HREF="#tag/22"
	><strong>Take a Breath!</strong></a>

<dt><A HREF="#tag/23"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>about the adaptation.</strong></a>
<dt><A HREF="#tag/24"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Changing the "login-sequence" in Linux?</strong></a>
<dt><A HREF="#tag/25"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>Linux, X, Dell Video Card</strong></a>
<dt><A HREF="#tag/26"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>sendmail</strong></a>
<dt><A HREF="#tag/27"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>about a stubborn mount error</strong></a>
<dt><A HREF="#tag/28"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	></a>Here is a very stupid question ... --or--
<dd><A HREF="#tag/28"
	><strong>How do I choose?</strong></a>

<dt><A HREF="#tag/29"
	><img src="../gx/dennis/qbub.gif" height="28" width="50"
	  alt="(?)" border="0"
	><strong>I was wondering</strong></a>

<!-- index_text ends -->
</DL>
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/greeting"><HR WIDTH="75%" ALIGN="center"></A>
<H3 align="left"><img src="../gx/dennis/hbubble.gif" 
	height="50" width="60" alt="(&para;) " border="0"
	>Greetings from Heather Stern</H3>
<!-- begin hgreeting -->
<BLOCKQUOTE>
It's that stormy month of the year again, when people expect us to be silly
in print.
</BLOCKQUOTE>
<BLOCKQUOTE>
I feel silly for saying this but it seems like we have to every month:
</BLOCKQUOTE>
<BLOCKQUOTE><ol>
<li> There is no guarantee that questions will <EM>ever</EM> be answered,
especially if not related to Linux.
<li> HTML attachments drive us nuts...
</ol></BLOCKQUOTE>
<BLOCKQUOTE>
ICANN expressed a desire to make a foolish mess of the entire internet.
I wrote "An Open Letter to ICANN" which has been published in <EM>Linux Journal</EM>
recently: <A HREF="http://www.linuxjournal.com/articles/conversations/0022.html"
	>http://www.linuxjournal.com/articles/conversations/0022.html</A>
</BLOCKQUOTE>
<BLOCKQUOTE>
While we're thinking of messes, and Easter coming up, how about cute
fluffy bunnies?  I cleaned up the ol' home office a bit.  I think
Dust Puppy (<A HREF="http://www.userfriendly.org"
	>http://www.userfriendly.org</A>) can find a girlfriend named
Dust Bunny if he tries hard enough.
</BLOCKQUOTE>
<BLOCKQUOTE>
As you hopefully know by now this is a Linux magazine and we normally only
answer Linux questions.  But, it's the silly month, so once again with
that <a href="#tag/cardboard">cardboard box</a> thread that snuck in ... 
</BLOCKQUOTE>
<BLOCKQUOTE>
And finally, something I've been messing with that makes us all continue
to look foolish for using Linux.  How can we call ourselves a desktop
system when all the word processors suck?  Oh yeah.  We don't.  We just
call it an operating system, apps are for distros.  Well, they still need
to work on it.
</BLOCKQUOTE>
<BLOCKQUOTE>
The first thing you may wonder is why would Ms. My Box Is More Productive
Without Producticity Software even care, anyway?  Well, it so happens
that a friend of mine, who isn't computer oriented in the slightest,
wanted a resum&eacute; and of course since we're close, she asks me.  No
problem, I think.  It's just an <tt>rpm -i</tt> or an <tt>apt-get install</tt>
away.  Right.
</BLOCKQUOTE>
<p>For a more positive view, see <A HREF="#tag/7"
        >Tony's telecommunications article</a> this issue.</p>
<BLOCKQUOTE>
Don't believe me, eh?  Well let's start at the top.  WordPerfect is time
bomb ware.  Their idea of "for personal use" includes dying at 90 days
so you have to go get a registration key, allegedly free.  In my old
shareware days I always avoided timebombs.  You never know if they might
also try to take your documents with them or something.  It's a shame
because I always liked their DOS software.  I may buy it someday, when
I need it for myself, after all, with my consulting biz I guess I don't
count as personal use anyway.  But I resist - my principles don't call
for supporting time bombs.  Grr.
</BLOCKQUOTE>
<BLOCKQUOTE>
I tried StarOffice a few months ago.  It shows many of the worst features
of having originally been a port from the windows version via some translation
library.  Its "everything lives inside the Staroffice Window" mold was
one of the GUI features I was glad to get away from when I left Windows
behind, and its printer configuration is evil and broken.  Okay, when it
finally works it's rather cool to have numerous Avery papers selectable
in the dialog so you can do labels and index cards.  But, it's actually
<EM>easier</EM> to set up a printer with plain old lpr and magicfilter.  Yuck.
</BLOCKQUOTE>
<BLOCKQUOTE>
Applix might be okay.  I dunno, I was in a hurry, and wanted something
a bit smaller.  I guess I just hate the idea that I have to download a
whole suite just to get one part.
</BLOCKQUOTE>
<BLOCKQUOTE>
I think I have LyX installed, I try that.  I do.  It doesn't do a number
of things that need doing.  I tried to do spring margins and it has its
own ideas how wide to make the table.  This will never work.
</BLOCKQUOTE>
<BLOCKQUOTE>
The SIAG people have loose parts.  Their word processor is called Pathetic
Writer.  I tried it... and they're right.  If I recall correctly Wordpad
has more features.  Sigh.
</BLOCKQUOTE>
<BLOCKQUOTE>
How about Abiword?  Those abisuite guys have their head on straight, let's
try it.   So happens Terry already has it on his box since we put Progeny
on his desk.  He tries to use it for one page reports and growls at it
because it can't deal with tabs very well.  Hmmm... anyway, just an ssh
session over there and access it via X, right?  Wrong!  It whines that
a font is missing.  That's insane.  Betel has the most complete font
collection in the house, since it's setup to be our TTF font server...
</BLOCKQUOTE>
<BLOCKQUOTE>
Fine. Install it locally.  (I have to get the whole suite.  Oh well.  Get
a soda, come back.)  One <A HREF="http://www.suse.com/">SuSE</A> style rpm i
coming up!  (wave magic
wand)  uh, this doesn't load at all, even to pop up with the complaint.
No error message in the xterm window, nothing.  Fume.
</BLOCKQUOTE>
<BLOCKQUOTE>
Well, let's try the K office then.  Kword coming up.  Installs sweet enough.
Even runs. (Yay!)  Can't do tables even though it has buttons for it.  Now, we
are talking about everyone's favorite use for spring margins, putting
the dates of your last employ all the way to the right, and since almost
nothing has proper spring margins, can't do it without tables.  At least
it does those long beautiful bars, which I had figured would need tables.
Even when I use just plain white space to push things to the end, the
thing is iffy about whether they show up over there.  If I change the
font anywhere on the line its metrics are a scramble and things fall off
entirely.
</BLOCKQUOTE>
<BLOCKQUOTE>
On the bright side, its preview feature generates very clean Postscript,
not yet encapsulated.  So, being the programmer type that I am, I let Kword
do what it could, and improved the rest in text mode, previewing directly
in ghostview.
</BLOCKQUOTE>
<BLOCKQUOTE>
One shouldn't have to be a programmer to whip together a friend's job
hunting paperwork.  It takes us back to the old days, when a CP/M box
could be a decent terminal for a brighter Postscript printer, if you
slipped it a sneaky enough program.
</BLOCKQUOTE>
<BLOCKQUOTE>
Oddly enough if I had just thrown it together in HTML it would have been
pretty quick.  But that would have been in a plain old text editor too --
since the state of the art in WYSIWYG editors for HTML is about the same.
Bluefish and August seem to have them beat all over the place.  I think
I like Bluefish better, it has a feel very similar to HTMLedPro which I
used when I used to live more closely with that other operating system.
</BLOCKQUOTE>
<BLOCKQUOTE>
If the Dot Com Fallout has made your company foolishly let you go, at
least the Linux world has room for you.  You can check out <EM>Linux Journal</EM>'s
Career Center (<A HREF="http://www.linuxjournal.com/employ"
	>http://www.linuxjournal.com/employ</A>), Geekfinder
(<A HREF="http://www.geekfinder.com"
	>http://www.geekfinder.com</A>), the Sysadmin's Guild (SAGE) Job Center
(<A HREF="http://www.usenix.org/sage/jobs/sage-jobs.html"
	>http://www.usenix.org/sage/jobs/sage-jobs.html</A>), or pay attention to
your local area papers for when major high tech Job Fairs are in your area,
so you can go to them.  There are also some really generic job sites like
Dice.Com (<A HREF="http://www.dice.com"
	>http://www.dice.com</A>) or MonsterBoard (<A HREF="http://www.monsterboard.com"
	>http://www.monsterboard.com</A>).
If you hate the corporate mold, check out some of the project offers at
SourceXchange (<A HREF="http://www.sourcexchange.com"
	>http://www.sourcexchange.com</A>)
or Collab.Net (<A HREF="http://www.collab.net"
	>http://www.collab.net</A>).  Or put up your consulting shingle
by listing yourself at Linuxports (<A HREF="http://www.linuxports.com"
	>http://www.linuxports.com</A>) and getting
listed into a few search engines.
</BLOCKQUOTE>
<BLOCKQUOTE>
Me, I don't have to worry about getting into search engines, do I? 
<IMG SRC="../gx/dennis/smily.gif" ALT=";D" 
		height="24" width="20" align="middle">
Have a happy April!
</BLOCKQUOTE>

<!-- end hgreeting -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/1"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 1 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>What's this word?</H3>


<p><strong>From dana gillen 
</strong></p> 
<p align="right"><strong>Answered By Jonathan Markevich,
 Chris  Gianakopoulos, Breen Mullins, Huibert Alblas, Heather Stern 
<br></strong></p>

<!-- sig -->

<!-- ::
What's this word?
~~~~~~~~~~~~~~~~~
:: -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Can you tell me what PCMICIA stands for? Thanks!
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Jonathan]
Sure.  "People Can't Memorize Computer Industry Acronyms".  (You have an
extra "I" in there)
</BLOCKQUOTE>
<BLOCKQUOTE>
Seriously, I believe it's "Personal Computer Memory Card Inter... &nbsp;  ..." Uh, I
forget the "A", thus proving the previous statement.
</BLOCKQUOTE>
<BLOCKQUOTE>
However, they've since ditched the obscure acronym and now call it "PC
Card", since "Memory" was very rarely what it was about.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Chris G.]
The "A", I believe, stands for association.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Breen]
PCMCIA = Personal Computer Memory Card International Association
</BLOCKQUOTE>

<BLOCKQUOTE>
<dl><dt>
A good resource for ATDA (All Those Dratted Acronyms) is the Babel
File:
<dd><code><A HREF="http://www.geocities.com/ikind_babel/babel/babel.html"
	>http://www.geocities.com/ikind_babel/babel/babel.html</A>
</code></dl>
</BLOCKQUOTE>
<BLOCKQUOTE>
HTH.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Huibert]
Hi, my first post to Tag (here it comes
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE>
In Gnome is a litle utility called Gdict (Foot-&gt;utilities-&gt;Gdict)
If you go to Settings-&gt;preferences-&gt;Server-&gt;Database you can select
V.E.R.A. (Virual Entity of Relevant Acronyms)
It's great for looking up Acronyms (Accept the 'wrong' or funny ones are
not listed, is there a place those are listed)
</BLOCKQUOTE>
<BLOCKQUOTE>
Hope I could help
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather] Yes, funny ones that have been gathered into hacker
lore can be found in The Jargon File:
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQUOTE><code>
<a href="http://www.tuxedo.org/jargon/">http://www.tuxedo.org/jargon</a>
</code></BLOCKQUOTE></BLOCKQUOTE>
<Blockquote>
...although, oddly enough, this one isn't in there.  We'll have to fix 
that!
<IMG SRC="../gx/dennis/smily.gif" ALT=":D" 
		height="24" width="20" align="middle">
</Blockquote>

<!-- end 1 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/2"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 2 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Unable to Install Linux</H3>


<p><strong>From N P
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<!-- sig -->

<strong><p align="center">
see attached <a href="misc/tag/NP.equip-list.txt">equipment list</a>
</p></strong>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I am trying to install RH 6.2 on the WD 12.3GB drive. However, it hangs
during the installation (after the partitions are formatted and progress
dialogs starts).
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
If you're using the graphical install, I suggest that you do not. All
"freezing" problems that I've had with RedHat installations happened with
GUI-based installations; all of them were resolved by going to the
text-based one.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Well not quite true. If i don't select any packages (no X,
compilers, multimedia, etc) to install, it installs fine.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
The program that installs RedHat is a huge, complicated thing that goes
into weird contortions once in a while. If you simply cannot manage to
install a full system by using it, install the basic system and whatever
packages are necessary to dial up and do FTP (those should actually be a
part of the basic system, but I'm not certain), and download a copy of
'rpmfind' &lt;<A HREF="http://www.rpmfind.net/linux/rpmfind/rpmfind.html&gt"
	>http://www.rpmfind.net/linux/rpmfind/rpmfind.html&gt</A>;. This program
will connect to an "RPM server" and download whatever packages you specify,
automatically resolving dependencies in the process. It's a not-quite-as-
powerful knockoff of <A HREF="http://www.debian.org/">Debian</A>'s 'apt' tool, but is actually reasonably mature
and useful.
</BLOCKQUOTE>
<BLOCKQUOTE>
Another option is to try installing another distro; I'm a real Debian
zealot, myself. One of the many reasons that I really like it is that
something like the above procedure is <EM>already</EM> one of the standard
installation options: the base system install takes 5-10 minutes, you tell
'apt' which of the many available servers you want to use, and walk away.
'apt' can use FTP, HTTP, local CDs, or packages right off the HD - and you
can mix-and-match sources however you like. Dependency problems? What are
those? &lt;grin&gt;
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
When the
installation hangs there is no response to any keypress and it doesn't hang
at the same part of the installation i.e. at the beginning (just starting),
middle, or end (seconds to go).
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Hm. Have you run a good memory test on the machine? There are plenty of
tools available, but my favorites are the old DOS "burn-in" tool and
Linux's "memtest86" (interestingly enough, "memtest86" doesn't require
Linux: it is a bootable image that can be run from a floppy!) "memtest86"
is a part of the "hwtools" package, at least under Debian. Run either one
of them for a  minimum of 24 hours.
</BLOCKQUOTE>

<!-- end 2 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/3"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 3 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>DNS and telnet</H3>


<p><strong>From crabe 
</strong></p> 
<p align="right"><strong>Answered By Mike Orr
<br></strong></p>

<!-- sig -->

<P><STRONG>
Hi,
How do you get telnet working on your own machine as referred to in the
DNS HOWTO, i.e telnetting at 127.0.0.1 ?  I got telnet working to reach
my ISP but never got around telnetting 127.0.0.1. So I gave up DNS.
I have looked around all the HOWTOs available, and perhaps it's too
simple for mentionning.
I am running LinuxPPC2000.
Thanks for any answer.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
Are you trying to do a standard telnet ("telnet 127.0.0.1") or telnet
to another port ("telnet 127.0.0.1 53" would be a DNS query)?  Here are
a few possibilities:
</BLOCKQUOTE>
<BLOCKQUOTE><ol>
<li> Nobody is listening on the telnet port.  If so, you'll get an immediate
"connection refused" error.  Telnetd is normally started from inetd.
Uncomment the telnet line in <TT>/etc/inetd.conf</TT> and "killall -HUP inetd".
<li> Your loopback devide is not configured.  What happens when you run
"ping 127.0.0.1"?  If you get no response, do "ifconfig".  There should
be a stanza for device "lo".  If not, run "ifconfig lo 127.0.0.1" and/or
"ifconfig 127.0.0.1 up".  (If you're still running kernel 2.0.x, follow that
with "route add -net 127.0.0.0").  Then look at your network startup scripts to
see why it isn't being activated by default.
<li> Inetd runs telnet through a tcpd wrapper for security, and you're failing
the tcpd check.  This would cause the connection to do nothing (at least
nothing visible) and then disconnect after a couple seconds.  See
"man tcpd" and "man 5 hosts_access".
<li> You are telnetting to port 53 and your nameserver is not running.
If so, you'd get a "connection refused" error.  If you installed named
(bind), find out why it isn't running.
</ol></BLOCKQUOTE>
<BLOCKQUOTE>
The TAG security hawks will send a follow-up if I don't also mention
that telnet is a security risk bla bla bla because it doesn't encrypt
your password or your data.  Think twice before running telnetd, and
think a third time before allowing tcpd to allow telnet connections from
outside your local network.
</BLOCKQUOTE>

<!-- end 3 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/4"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 4 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Help on LILO stopping at LI</H3>


<p><strong>From Alessio Frenquelli 
</strong></p> 
<p align="right"><strong>Answered By  Heather Stern
<br></strong></p>

<!-- sig -->

<p><strong>
Hello,
<br>I start thanking you for any help ... I am stuck at this stage, I am not a
GURU on LINUX and I cannot overcome the problem.
</strong><p>
<P><STRONG>
Therefore I cannot really point to what has been changed or went wrong.
</STRONG></P>
<P><STRONG>
Under the Internet I found many, many errors entries pointing to LILO not
being able to load in a disk that is above the 1024 cylinders
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Yes, it used to be LILO's biggest bug, though not its loudest (that one is
people using it wrong and then wailing what are they going to do now that
their MBR is mangled).
</BLOCKQUOTE>
<BLOCKQUOTE>
But ever since the new version those are old messages.  The normal solution
until it came out, was to create a tiny <TT>/boot</TT> near the beginning of the
free space - even most dual booters could manage to slip a 20 Mb partition
below the boundary.  This works because on the <EM>kernel</EM> and <EM>bootmap</EM> needs
to be below the line;  once the kernel is loaded you are no longer working
with real mode BIOS issues at all, you are fully in protected mode and can
access everything the kernel is built for.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
In my case LILO always worked so far, and I did not surely changed the disk
size, done any repartitioning under Windows NT nor under Linux.
</STRONG></P>
<P><STRONG>
Under Windows, I have run Program =&gt; PartitionMagic =&gt;PartitionInfo and I am
attaching the output of the command to this email in case you need to see in
details my machine's partitioning.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> []
So you normally use partition magic for your dual boot menu, I'm guessing.
If so, that is what is presently in your MBR, and would be overwritten
(***warning warning danger will robinson*** or at least look real carefully
that you've set up stanzas for NT also first!) if you change boot = <TT>/dev/hda</TT>
</BLOCKQUOTE>

<h4>ENVIRONMENT</h4>
<p><strong>
Dual bootable Laptop, Toshiba Tecra 8100; one partition is Windows NT
workstation, the other is Linux RedHat 6.1.
</strong><p>
<h4>PROBLEM</h4>
<p><strong>
Linux 6.1 LILO does not longer boot properly. Just stop at the word "LI".
</strong><p>

<h4>SOME TROUBLESHOOTING INFO</h4>
<strong>
Fortunately I have the boot diskette, and booting from it, I can
successfully get to Linux.
</strong><p>
<p><strong>
When running : &lt;&lt; <TT>/sbin/lilo</TT> &gt;&gt;  I got messages:
</strong><p>

<strong><pre>[root@afrenquelli /etc]# lilo
Warning: device 0x0305 exceeds 1024 cylinder limit
Warning: device 0x0305 exceeds 1024 cylinder limit
Warning: device 0x0305 exceeds 1024 cylinder limit
Warning: device 0x0305 exceeds 1024 cylinder limit
Added linux *
[root@afrenquelli /etc]#
</pre></strong>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
You used to get this before, or you now have a bigger disk than you
used to?
</BLOCKQUOTE>
<BLOCKQUOTE>
By the way, Redhat 6.1 is a bit old, and lilo itself was updated last
year so that 1024 cylinder issues are not a problem for it. (You'll
also want to keep up to date on RH security updates, not quite as
drastic as upgrading the system entirely.)
</BLOCKQUOTE>
<BLOCKQUOTE>
With the newer version, you can add the keyword
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>	LBA32
</font></code></blockquote>
<BLOCKQUOTE>
into the top of your <TT>/etc/lilo.conf</TT> and it would use a different method
to know where things are on the disk.
</BLOCKQUOTE>
<BLOCKQUOTE>
Being a boot loader, it's critical for lilo to know <EM>precisely</EM> where
the kernel resides on your drive.  Moving your kernel file (even if you
then moved it back) or your system maps is a good reason to run <TT>/sbin/lilo.</TT>
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Some machine's characteristics:
</STRONG></P>

<pre><strong>[root@afrenquelli /tmp]# df -k
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/hda7              1510032   1201268    232056  84% 	/
/dev/hda5                23302      2648     19451  12% 	/boot
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
No relation.  Lilo puts its bits into the Master Boot Record, which is
not shown here.  If ordered to, it could also use the superblock but,
that is also not shown here, as it's reserved space for the filesystem
driver.
</BLOCKQUOTE>

<Pre><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>[root@afrenquelli /etc]# lilo
Warning: device 0x0305 exceeds 1024 cylinder limit.
   Use of the 'lba32' option may help on newer (EDD BIOS) systems.
Fatal: sector 19926490 too large for linear mode (try 'lba32' instead)
</pre></strong>
<P><STRONG>
------------------------------------- file <TT>/etc/lilo.conf</TT> contains
</STRONG></P>

<pre><strong>boot=/dev/hda7
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
This says put it in the superblock of the 7 partition ... your /
</BLOCKQUOTE>
<BLOCKQUOTE>
Most people would have it in the MBR ... <TT>/dev/hda</TT>  with no number.
Do you have an NT boot menu pointing you into Linux?  Because I also
notice that you don't have a chain loader stanza, to ask the item
below to offer you your NT boot setup.
</BLOCKQUOTE>

<Pre><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
vga=791
</pre></strong>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
a nice framebuffer text mode 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>

<Pre><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
linear
</pre></strong>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
"this is a big disk"
</BLOCKQUOTE>

<Pre><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
default=linux
image=/boot/vmlinuz-2.2.12-20
        label=linux
        initrd=/boot/initrd-2.2.12-20.img
        read-only
        root=/dev/hda7
</pre></strong>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
The stock redhat kernel, I see.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I did not "fiddle" with Linux at all before this error appeared !
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
WHat do you normally use Linux for?
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
So , I changed in <TT>/etc/lilo.conf</TT> the value "linear" with "lba32", and then
<TT>/sbin/lilo</TT> runs fine with :
</strong></p>

<strong><pre>c[root@afrenquelli /etc]# lilo -v
LILO version 21.6-1, Copyright (C) 1992-1998 Werner Almesberger
Linux Real Mode Interface library Copyright (C) 1998 Josh Vanderhoof
Development beyond version 21 Copyright (C) 1999-2000 John Coffman
Released 16-Dec-2000 and compiled at 17:04:30 on Jan  9 2001.

Reading boot sector from /dev/hda7
Merging with /boot/boot.b
Boot image: /boot/vmlinuz-2.2.12-20
Mapping RAM disk /boot/initrd-2.2.12-20.img
Added linux *
/boot/boot.0307 exists - no backup copy made.
Writing boot sector.
</pre></strong>
<p><strong>
At this stage I "REALLY" hoped that the problem went away, but I still get
only "LI" at boot time,
I can only use the boot diskette to get into Linux.
</strong></p>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
ATTEMPTS TO SOLVE IT
===================
</STRONG></P>
<P><STRONG>
Thinking that the latest version of LILO "could" have fixed this problem, I
have downloaded LILO 21.6.1-1 from
<A HREF="http://rpmfind.net/linux/RPM/contrib/libc6/i386/lilo-21.6.1-1.i386.html"
	>http://rpmfind.net/linux/RPM/contrib/libc6/i386/lilo-21.6.1-1.i386.html</A>
</STRONG></P>
<P><STRONG><CODE>
I have then upgraded my LILO with : "rpm -Uhv &lt;nomefile&gt;.rpm"
</CODE></STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Good deal!  Yay!
</BLOCKQUOTE>
<BLOCKQUOTE>
You should also keep up to date on RedHat security updates for RH 6.1.
(Not directly related to this, just a good idea)
</BLOCKQUOTE>
<BLOCKQUOTE>
Any other recent installs or upgrades?
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
The upgrade completed fine, and then when I try to run <TT>/sbin/lilo</TT> I got:
</STRONG></P>

<pre><strong>[root@afrenquelli /etc]# lilo
Warning: device 0x0305 exceeds 1024 cylinder limit.
   Use of the 'lba32' option may help on newer (EDD BIOS) systems.
Fatal: sector 19926490 too large for linear mode (try 'lba32' instead)
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Even better, I'm glad they provide useful error messages ... that tell you
what to do about an error.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
So , I changed in <TT>/etc/lilo.conf</TT> the value "linear" with "lba32", and then
<TT>/sbin/lilo</TT> runs fine with :
</STRONG></P>

<pre><strong>c[root@afrenquelli /etc]# lilo -v
LILO version 21.6-1, Copyright (C) 1992-1998 Werner Almesberger
Linux Real Mode Interface library Copyright (C) 1998 Josh Vanderhoof
Development beyond version 21 Copyright (C) 1999-2000 John Coffman
Released 16-Dec-2000 and compiled at 17:04:30 on Jan  9 2001.

Reading boot sector from /dev/hda7
Merging with /boot/boot.b
Boot image: /boot/vmlinuz-2.2.12-20
Mapping RAM disk /boot/initrd-2.2.12-20.img
Added linux *
/boot/boot.0307 exists - no backup copy made.
Writing boot sector.
</strong></pre>
<P><STRONG>
At this stage I "REALLY" hoped that the problem went away, but I still get
only "LI" at boot time,
I can only use the boot diskette to get into Linux.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
That's weird 
<IMG SRC="../gx/dennis/unsmily.gif" ALT=":(" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE>
Maybe you need to take the linear or lba32 mark <EM>out</EM> ??
</BLOCKQUOTE>
<h4><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
WHAT'S NEXT ?
</h4>
<P><STRONG>
If I could avoid to rebuild the LINUX partition would be GREATLY
appreciated, since I am not a Linux expert and I would need some guidance
also lots of other software is installed and I would like to avoid to
reinstall the all lot !
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Although I'd like to note at this time that making sure your backups of your
personal data on both OS setups are in current state, and good working
condition... and not stored on the same disk.  Tapes or a CD-R or a stack
of ZIP cartridges maybe.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Could I just de-install LILO and re-install LILO ?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
should be able to.  lilo -u to put the backup bits from <TT>/boot/boot.0307</TT>
back into the superblock <TT>/dev/hda7</TT> (where it hopefully came from).  If
it whines abotu timestamps, lilo -U insists.
</BLOCKQUOTE>
<BLOCKQUOTE>
Then, you should be able to run lilo again to install it as a fresher
instance.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Or should I add something into the <TT>/etc/lilo.conf</TT> and try to run "lilo"
again ?
</STRONG></P>
<P><STRONG>
Could I somehow just rebuild the "booting" portion of the Linux, and if so,
could you please provide detailed instructions on how to do it ?
</STRONG></P>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
lilo has actually very excellent documentation that comes with it... much
finer than many application packages in fact.  If you run 'locate lilo'
a bunch of things will scroll by, and several be from the doc directory
tree somewhere on your disk.  I'm guessing <TT>/usr/doc/lilo-</TT> (some version
number) but I'm not near a RedHat system right now to look.  Anyways
on my <A HREF="http://www.suse.com/">SuSE</A> system it's in <TT>/usr/doc/packages/lilo</TT> ... I have some dvi
files (readable or printable by LaTeX tools) and some compressed postscript
files (but ghostview ... the command gv ... is glad to show these to me).
</BLOCKQUOTE>
<BLOCKQUOTE>
The README, however, is the right place to start, because unlike most
readme files around here (which can be summarized "so this is the foo
program, I created it because bla bla.  It's under the GPL/artistic/whatever
license, see COPYING.  If there are any bugs (hope not) get in contact
with me at ...") it has some serious data in it.  Consider it your quickstart
guide to a working LILO.
</BLOCKQUOTE>
<BLOCKQUOTE>
You've already done a number of the obvious things, so, let us know if the
uninstall/reinstall trick works, and if that readme isn't helpful to you
we may be able to translate it to plainer english.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Kind regards,
<br>alessio
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Our best hopes for your system, Alessio.  We'd really like to hear back
what fixed it, if you manage to solve it.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Ah - an opportunity to shill for one of my scripts! 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE>
Here's where my "doc" script would come in really handy: all you'd have to
do is type something like "doc lil", and it would give you a numbered list
of all the subdirectories in your "doc" directory that start with 'lil'.
Typing that number enters the directory and shows you a numbered file list;
typing one of <EM>those</EM> numbers displays the file, no matter what its format
is. When you exit the viewer, it shows you the list again, and gives you a
chance to "fish around" in subdirectories and other files.
</BLOCKQUOTE>
<BLOCKQUOTE>
I know I sent it into LG as a 2-cent tip quite a while ago, but I believe
I've made a few improvements since then, so here it is:
</BLOCKQUOTE>
<p align="center">See attached <a href="misc/tag/doc.bash.txt">doc.bash.txt</a></p>

<!--              . . . . . . . . . . . . . . . . . . .              -->
<HR WIDTH="40%" ALIGN="center">
<P><strong><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>
Heather,
<br>thank you for your complete and prompt reply.
</strong></p>

<p><strong>
I will provide here some answers from the queries that you had.
At this stage it seems that the problem is caused by one of our product,
that I have recently installed under Linux. After reporting the LI problem
to our support team, I received a reply where they state that this product
ast times seems to affect LILO, and to cause the problem that I have
described to you.
</strong></p>

<p><strong>
I am still awaiting from our support team, if they know how to fix the
problem of LILO.
</strong></p>

<p><strong>
And now to your questions:
</strong></p>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Heather]
 Any recent changes?  The next L should represent finding the second stage
 loader so the kernel can get going.  These could be disk-ish things like
 you caught a virus and successfully cleaned it out, had a hard crash and
 needed to reboot, etc
</blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Yes, I have installed a new Micromuse Package under LINUX. After the
installation I rebooted few times without problem, but then one day, LI
started to happen.
</strong></p>


<pre>
 Warning: device 0x0305 exceeds 1024 cylinder limit
</pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Heather]
You used to get this before, or you now have a bigger disk than you
used to?
</blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
No changes on disk size, or disk partitioning under Linux or under NT. I
never run /sbin/lilo before , so I cannot tell you if this was a problem.
But LILO always worked before.
</strong></p>

<p>
I did not "fiddle" with Linux at all before this error appeared !
</p>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Heather]
WHat do you normally use Linux for?
</blockquote>


<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I use Linux to run Micromuse products, and as I said, the problem started to
happen since I installed one of our package. Other people installed it on
their laptop, but they did not report the error. Support told me that at
times this product is know to cause some problem with LILO, but they are not
too sure.
</strong></p>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Heather]
 That's weird 
<IMG SRC="../gx/dennis/unsmily.gif" ALT=":(" 
		height="24" width="20" align="middle">

<br> Maybe you need to take the linear or lba32 mark *out* ??
</blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
You mean, remove linear or lba32 from lilo.conf and try to run lilo again ?
I will try this.
</strong></p>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Heather]
 Then, you should be able to run lilo again to install it as a fresher
 instance.
</blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
So I should just run :
</strong></p>
<p><strong><ol>
<li>lilo -u or lilo -U
<li>Then how do I re-install LILO ? Just running /sbin/lilo or do I need to
download the LILO package from somewhere and the install it with   "rpm -Uhv
<nomefile>.rpm" or something similar ?
</ol></strong></p>

<p><strong>
Thank you for your help once again, I will keep you posted !

<br>ciao, alessio
</strong></p>

<!--              . . . . . . . . . . . . . . . . . . .              -->
<HR WIDTH="40%" ALIGN="center">

<P><strong><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>
Heather,
</p></strong>

<p><strong>
here is how I got this problem fixed with the help of the support personnel
in our company
</p></strong>

<p><strong>
Thank you once again for your support !!!
</p></strong>

<p><strong>
Support adviced me that the following recovery procedure for LILO problems
is to be used for Laptop that have dual boot partition, NT & Linux, when the
menu received at boot time shows OS LOADER, and then it presents 2 choice ,
Windows NT or Linux.
</p></strong>

<p><strong>
Steps taken:
</p></strong>

<p><strong><ol start="0">
<li> Boot Linux from bootdisk
<li> Run /sbin/lilo -v  (make sure no errors are displayed)
<li> Insert a new diskette and mount it:
<blockquote><code><font color="#000033"><br>	mount -t msdos /dev/fd0 /mnt/floppy
</font></code></blockquote>
<li> df -k gives me 2 filesystems:
<blockquote><code><font color="#000033"><br>	/dev/hda7 Mounted /
<br>	/dev/hda5 Mounted /boot
</font></code></blockquote>
<li> run command :

<blockquote><code><font color="#000033"><br>	dd if=/dev/hda7 of=/tmp/bootsect.lnx bs=512 count=1
</font></code></blockquote>
<li><tt> cp /tmp/bootsect.lnx /mnt/floopy</TT>
<li> shutdown and reboot under NT
<li><tt> copy c:\bootsect.lnx c:\bootsect.lnx_old</tt>
<li><tt> copy a:\bootsect.lnx c:\bootsect.lnx</tt>
<li> reboot under Linux
</ol></strong><p>
<p><strong>
PROBLEM solved, LILO loads correctly.
</strong></p>
<p><strong>
Ciao, alessio
</strong></p>

<!-- end 4 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/5"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 5 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>How can you do a recursive search to find broken symbolic links?</H3>


<p><strong>From bandido 
</strong></p> 
<p align="right"><strong>Answered By  Ben Okopnik, Faber Fedor, Mike Orr
<br></strong></p>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I found the odd broken link after a few upgrades, and was wondering how
can I hunt down any other such beasties, 'ls' doesn't have any suitable
way to delimit, and poking about in man pages for find etc made me quite
nautious.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Faber]
Perhaps you should take some Dramamine. 
<IMG SRC="../gx/dennis/smily.gif" ALT=":-)" 
		height="24" width="20" align="middle">  The man pages are you friend.  If you do a "man find" and then type "<TT>/link</TT>" (that will do a search on the word link) you'll find all kinds of references to the work "link" (the owrk "link" will be highlighted).  Scroll down a couple of pages and you'll find the "type" option.
</BLOCKQUOTE>
<BLOCKQUOTE>
So, to find all the links on your system, you would type
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>      find / -type l
</font></code></blockquote>
<BLOCKQUOTE>
Simple, no? 
<IMG SRC="../gx/dennis/smily.gif" ALT=":-)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
No. The querent was asking how to find <EM>broken</EM> links, not all links.
</BLOCKQUOTE>
<BLOCKQUOTE>
What's needed here is the "symlinks" program, written by Mark Lord. It will
find and classify all the links, hard and soft, in the filesystem. If you
want to see all the dangling (i.e., broken) links on your system, just
type
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>symlinks -r / | grep ^dangling        # Recursive search starting from /
</font></code></blockquote>
<BLOCKQUOTE>
If you want to delete all the broken ones, just enter
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>symlinks -dr /         # Recurse and delete broken links starting from /
</font></code></blockquote>
<BLOCKQUOTE>
For me personally, this wouldn't work too well. I use dangling links as
placeholders; as an example, I've disabled NFS during the boot procedure by
"breaking" the symlink in "<TT>/etc/rc2.d</TT>":
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>S19nfs-common -&gt; ../init.d/nfs-common        # Original link
<br>
<br>S19nfs-common -&gt; ../init.d/nfs-commonXXX     # Dangling!
</font></code></blockquote>
<BLOCKQUOTE>
If I should need to restore NFS, a 5-second fix will do it, without having
to figure out what directory the link should go into, where in the process
it should load (as determined by the number after the 'S'), or where it
should point.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thank you Ben, and others, a google search found "symlinks", although it
only appears to be available for <A HREF="http://www.debian.org/">Debian</A>.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
You could use the "alien" utility to convert it, or simply go to Debian's
page for "symlinks" -
</BLOCKQUOTE>
<BLOCKQUOTE>
&lt;<A HREF="http://packages.debian.org/stable/utils/symlinks.html&gt"
	>http://packages.debian.org/stable/utils/symlinks.html&gt</A>;
</BLOCKQUOTE>
<BLOCKQUOTE>
They always provide a link to the tarball from which the package was made,
and you can compile it yourself. &lt;grin&gt; I <EM>like</EM> Debian. A lot.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
Or, if you don't have the symlinks program available:
</BLOCKQUOTE>

<blockquote><pre>( find / -type l | xargs file ) | grep 'broken symbolic link'
</pre></blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
The symlinnks prog worked a charm, indeed ot cleaned up everything nicely,
changing absolute to relative links too, lovely.
</STRONG></P>
<P><STRONG>
The real issue, is my rampant stupidity, since after downloading symlinks,
lo and behold, I discovered it is part of Mandrake 7.2 which I use.
</STRONG></P>
<P><STRONG>
I had pissed about pouring over man pages trying to find out how to
delimit a search to find the buggers, only to discover my salvation
was close at hand.
</STRONG></P>
<P><STRONG>
I have received several TAG replies, and I must say the 1st was within
45 minutes. Astounding 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
I'll say it's astounding.  Some people don't get answers for
weeks... if at all...
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Keep it up guys.
</STRONG></P>
<P><STRONG>
--
Merp!
</STRONG></P>

<!-- end 5 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/6"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 6 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>BIOS passwords - Bane of my existance</H3>


<p><strong>From Unidentified Querent 
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik, Heather Stern
<br></strong></p>

<P><STRONG>
Can I send the Answer Gang a question and ask that I not be identified?
PLEASE??? Reason: I feel stoopid enough already. Hey, you may decide
that it isn't even a good idea to print this one. I doubt I would...
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
You really have no reason to feel stupid. Not knowing something does not
equate to being stupid; as I tell the students in my classes, "educated is
what you're supposed to be when you come out, not when you come in."
However, I don't think that there will be any problem with honoring your
request: Heather, the TAG's Answer Gal, scrubs off the e-mail addresses
anyway, and I've already removed your name.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Yes. I can strip anybody who wants down to <em>anonymous</em> and I already
make a sincere effort to scrub company references, etc from most things.
Someotimes it matters (like when someone is a spokesperson for the
company of a product we're talking about) but usually it is cruft and
gets cut.
</BLOCKQUOTE>
<BLOCKQUOTE>
As Editor Gal I can make sure this thread is scrubbed thorughly of your
identity, and will.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
It's really TWO questions but the second question is not necessary if
you have an answer to the first (which I doubt).
</STRONG></P>
<P><STRONG>
(1)
I read your "LILO:Password Protected Entries" article in the new
March LinuxGazette. Though I do not have a LILO question, I'd like to
ask you to follow up on something else you touched on in that article.
</STRONG></P>
<P><STRONG>
One of my toys is a CTX EzBook 800 laptop which is currently running
<A HREF="http://www.suse.com/">SuSE</A> 7. A while back, I thought it would be a good idea to block access
to "Lorraine's" BIOS settings. I set the BIOS password so that access to
the BIOS is blocked but booting is not. Good thing. I soon forgot the
password.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Ut oh.  This is below the scope Linux can probably help with, but read on.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
This isn't a HUGE problem since I don't have to access the BIOS very
often but booting from a CDROM is now impossible (without using a boot
floppy) and setting or correcting local time is a real pain in the rump
(see question 2).
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
have you tried setting Linux' date and then:
<br><tt>hwclock --systohc</tt>
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I know BIOS backdoors exist but I've been unable to find one for mine.
</strong></p>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
<TT>/dev/nvram</TT> maaaaybe.  Unfortunately it's laced with some righteous warnings
and most people use it by figuring out what to do with it when they <EM>have</EM>
normal BIOS access.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Lorraine's got a PhoenixBIOS 4.0 Release 6.0.67A dated 1985 - 1997. In
the year or so since I got stoopid, I've scoured the Internet for info
on what the Phoenix backdoor might be - I found nothing. I even
contacted the manufacturer, CTX, to see if they would help. All they
would suggest was popping open the laptop and removing the BIOS battery,
something I'm not sure I'd do even if I knew how to (yeah, I know I'm a
wimp).
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Opening the laptop may be tricky.  The usual rule is There Are Lots Of
Tiny Screw To Get Lost.  Taking notes and not making sudden moves while
it's half open (so you can see where plastic traces are plugged in before
yanking them loose carefully) are both good.
</BLOCKQUOTE>
<BLOCKQUOTE>
But the BIOS is usually a watch battery and about as easy to deal with as
a watch once you have it.
</BLOCKQUOTE>
<BLOCKQUOTE>
You may want to get printouts and take notes during bootup of things that
are BIOS options as far as you know them.  dmesg may help some.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
So... The questions remain: Do you know how to foil this sucker or,
failing that, can you
</STRONG></P>
<P><STRONG>
(2)
Tell me how to reset the BIOS time from within SuSE 7? That'd be a
piece of cake with RedHat's linuxconf but I've yet to find anything in
SuSE that would do the trick. Don't even ask about yast and yast2...
Change time zone, yeah. Change time, no way.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Ah yes, this would be the hwclock command I gave above.  You have to be
root to use it.
</BLOCKQUOTE>
<BLOCKQUOTE>
As for linuxconf, err, I haven't had good luck with it myself.  YaST
(yess, that's really how the command is spelled) is the admin tool
under SuSE, but as you can see, it's really more about installing
stuff, not so much for sysadmin work.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Take a look at the 'cmostool' utility
&lt;<A HREF="http://www.ibiblio.org/pub/Linux/hardware/?M=A&gt"
	>http://www.ibiblio.org/pub/Linux/hardware/?M=A&gt</A>;. It allows you to back up,
modify,  hex dump, etc. the CMOS - as well as deleting the whole thing
(which wipes out the password.)
</BLOCKQUOTE>
<BLOCKQUOTE>
!!! WARNING WARNING WARNING !!!
</BLOCKQUOTE>
<BLOCKQUOTE>
Do not do this if you don't know what you're doing! Wiping your CMOS
<EM>will</EM> make your system unbootable. You must know at least the CHS
(cylinder-head-sector) values for your hard drive, and either know or be
able to figure out the other necessary settings. If you dump your CMOS and
get stuck, you are on your own!
</BLOCKQUOTE>
<BLOCKQUOTE>
Now that I've scared you into twitching fits and heebie-jeebies...
</BLOCKQUOTE>
<BLOCKQUOTE>
Most BIOSs today are auto-configuring, and will either auto-detect or give
you the option of auto-detecting your HD; Phoenix BIOS certainly does that
(it's been my favorite for many years now.) For myself, if I'm going to do
that sort of thing - and I've worked on many, many machines where the
owner had set a BIOS password and forgot it - I'll boot DOS, save a copy
of the settings to a bootable floppy via 'savecmos', and only then blow
away the password via 'cmosedit'. That way, if things go truly awry, I can
at least get back to where I was and try something else. The 'savecmos'
utility (including 'cmosedit') is available all over the Net, e.g.
&lt;<A HREF="http://members.tripod.co.uk/paulc/cmosutil.zip&gt"
	>http://members.tripod.co.uk/paulc/cmosutil.zip&gt</A>;.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
P.S.: I bought this laptop new from Sears (don't laugh) and have the
receipt and everything. Honest!
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
I will be <EM>certain</EM> to stop by and check up on you. Have them ready, and
be afraid. Be <EM>very</EM> afraid. 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Cool.  Makes it lots easier to insure and all.  It's entirely a side
note, but <A HREF="http://www.mobilix.org"
	>http://www.mobilix.org</A> has a nice list of laptop resources.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thanks!!!
<br>Signed: Stoopid
</STRONG></P>

<!-- end 6 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/7"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 7 -->
<H3 align="left"><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>Making the Connection</H3>

<p align="right"><strong>By Anthony E. Greene 
<br></strong></p>

<p><em>Somewhere in the shuffle the original querent's message has been
lost, but basically, they asked about connecting their hospital together,
so that the doctors could communicate with ER and ICU, staff could access
suitable records or charts, etc.  The doctors are not dumb people, but 
they already have a specialty and a job to do, so it has to be a pretty 
clean setup.</em></p>
<!-- sig -->

<!-- ::
Making the Connection
~~~~~~~~~~~~~~~~~~~~~
:: -->
<BLOCKQUOTE>
You could setup a PPP server and use the modems to make dialup PPP
connections. This would allow you to use graphical network applications
such as browsers, FTP clients, and network file managers such as <A HREF="http://www.gnome.org/">GNOME</A>'s
GMC.
</BLOCKQUOTE>

<H4>Distributions and Packages</h4>
<BLOCKQUOTE>
Without knowing more about what resources you have available, I cannot
make specific recommendations. <A HREF="http://www.redhat.com/">Red Hat</A>, Mandrake, <A HREF="http://www.slackware.org/">Slackware</A>, <A HREF="http://www.debian.org/">Debian</A>, SuSe,
and <A HREF="http://www.caldera.com/">Caldera</A> all come with the tools you'll need to setup a network. I have
not used <A HREF="http://linux.corel.com/">Corel</A> but I've read that they left some server and development
packages out. That may be fine for home desktops, but in a business
environment I'd want a distribution that includes everything I might need
and lets me choose what to leave out or disable. You <EM>will</EM> need some
server packages to implement a solution and you will want development
packages available in case you need some tools that are not available in a
package.
</BLOCKQUOTE>
<h4>Data Entry</h4>
<BLOCKQUOTE>
First you need to figure out what applications will be used for data
entry. Eventually, you may find you need a database application, but it
sounds like what you need right now is something that generates documents
that can be shared. If the results are to be typed out as free text, a
text editor is probably the best way to go. The text editors that ship
with GNOME (gedit) and <A HREF="http://www.kde.org/">KDE</A> (kedit) are both adequate, but something like
Nedit has fewer bugs and more power. If you need to use templates for data
entry, you could either create some read-only files as templates or create
templates in a StarOffice for use with its word processor.
</BLOCKQUOTE>
<BLOCKQUOTE>
For something with a little more familiarity to GUI users, AbiWord can
edit plain text, RTF, and simple DOC files. It has a toolbar that any Word
user could use with no problem and is fairly lightweight. AbiWord is part
of GNOME Office and ships with the Ximian (Helixcode) desktop.
</BLOCKQUOTE>
<BLOCKQUOTE>
There are some Open Source medical applications available. Try searching
for them at Freshmeat &lt;<A HREF="http://www.freshmeat.net/&gt"
	>http://www.freshmeat.net/&gt</A>;, Sourceforge
&lt;<A HREF="http://www.sourceforge.net/&gt"
	>http://www.sourceforge.net/&gt</A>;, and Google &lt;<A HREF="http://www.google.com/&gt"
	>http://www.google.com/&gt</A>;.
</BLOCKQUOTE>
<h4>Desktop Applications</h4>
<BLOCKQUOTE>
If you really need an integrated solution for Linux desktops at a minimal
cost, StarOffice is a good choice. The latest version (5.2) is still a
serious memory and resource hog and takes time to startup. But once it's
running, its speed is reasonable, considering its large feature set.
</BLOCKQUOTE>
<BLOCKQUOTE>
I haven't used Applixware, but it is supposed to be very usable and
programmable. The latter may prove useful to you if you plan to use it for
data entry. Applixware is not free, but is a lot less expensive than MS
Office.
</BLOCKQUOTE>
<BLOCKQUOTE>
For intranet browsing and email, I still recommend Netscape Communicator
4.7x. It works fairly well, is stable on non-Java pages, and supports LDAP
and HTML mail. These last two features are very useful in an
organizational mail client. Netscape 6.x does not support LDAP and
StarMail's LDAP interface is too difficult to be useful. An LDAP server is
not too hard to setup for small organizations and is great for maintaining
an organizational address book.
</BLOCKQUOTE>
<h4>Sharing/Publishing</h4>
<BLOCKQUOTE>
After the data is entered, you will need to make it sharable. I suggest
each department have a directory that only they can write to and any
authorized use can read. Setting up these groups and permissions is not
too complicated, but is more than I want to cover here.
</BLOCKQUOTE>
<BLOCKQUOTE>
The key thing about sharing is deciding what protocols you will use to
share. Client applications for FTP and HTTP are easy to use. Both servers
are easy to install. But the permissions scheme for HTTP is separate from
the system user and groups settings. That makes it complicated to setup if
you have multiple groups of users that need different permissions. So I
don't recommend using <A HREF="http://www.apache.org/">Apache</A> and HTTP to share the documents.
</BLOCKQUOTE>
<BLOCKQUOTE>
You can use FTP, but the WuFTPd server that has shipped with many
distributions is an almost constant source of security problems. Just make
sure you disable anonymous logins if you choose to use FTP. Web browsers
are great FTP clients because they can launch external applications to
view documents. The only real problem with FTP is that passwords are sent
over the network unencrypted. On a small, closed network this should not
be a problem.
</BLOCKQUOTE>
<BLOCKQUOTE>
This is probably more than you expected, but it's just enough to get you
started. Running a network will mean learning a lot at first, but it
should run well after it's setup.
</BLOCKQUOTE>


<!-- end 7 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/8"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 8 -->
<H3 align="left"><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>Setup of Microsoft Outlook Express 5 for Sending of Clear Text</H3>


<p align="right"><strong>Answered By Chris Gianakopoulos 
<br></strong></p>
<p><strong>
[Heather] We get so many people who send us perfectly good questions, in
HTML, which drives some of our mailers crazy.  It's not surprising that
someone with a crippled Linux box would reach for a nearby Windows system
to send the mail.  So, here's some help for you.  Utterly self serving,
to help <EM>us</EM> get plaintext 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</strong></p>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > [Chris]
Hey Heather,
</BLOCKQUOTE>
<BLOCKQUOTE>
I glanced at the March 2001 Linux Gazette and noticed your (subtle)
request for the steps needed to set up my Outlook Express mailer to send
clear text.  I read in a textbook (circuits) that a promise made is a debt
unpaid.  I will recoin the phrase to "A request made is a response
unpaid".  Therefore, I will post you a response which attempts to provide
a coherent step of steps to achieve our goal.
</BLOCKQUOTE>
<BLOCKQUOTE>
I actually executed these
steps while typing the steps into a text file using vi for DOS.  I use my
Microsoft machine when I email late nights.  I hope that this is coherent!
Anybody can sanity check me, of course (that's what teams do -- review
each other's work).
Here are the steps.
</BLOCKQUOTE>
<h4>
Steps for Setting Up Microsoft Outlook Express 5 for Sending of Clear Text
</h4>
<ol>
<li> Start up Microsoft Outlook Express.
<li> From the "Tools" menu, select "Options".
<li> An Options dialog box will pop up.
<li> On the Options dialog box, select the "Send" tab.
<li> Under the "Mail Sending Format" section of the dialog box, select
the "Plain Text" radio button.
<li> Press the "Plain Text Settings" button.
<li> A Plain Text Settings dialog box will appear.
<li> For message format, I select the "MIME" radio button.
<li> Check the "Indent the original text with" check box.  This will cause
any included original message to be indented and preceeded with a "&gt;"
sign.
<li>. Select "Automatically wrap text at" with 74 characters. (Ben Okopnik's
suggestion to me).
<li>. Press the OK button of the Plain Text Settings dialog box.
<li>. I make sure that the "Reply to messages using the format in which they
were sent" check box is unchecked.
<li> Press the OK button of the Options dialog box.
</ol>
<BLOCKQUOTE>
You're all set!
<br>-------------------- End of Instructions ----------------------
</BLOCKQUOTE>
<BLOCKQUOTE>
The line lengths of the steps look short because I typed those steps into
a text file, using vi, and I always keep
lines less than 80 characters.  I'm from the old days of using terminals
(not ASR-33 teletypes although once
I had a General Electric Terminet 300 TTY for a printer), so I avoid line
wrap.
</BLOCKQUOTE>
<BLOCKQUOTE>
Thank you and Ben for the encourgement that you give.  I'm
still a cross between a soon to be Linux hacker and an embedded software
hacker (they call me an engineer, but, I think that is questionable).
</BLOCKQUOTE>
<BLOCKQUOTE>
You are a kick-a## teem!  Keep up the good work!
<br>Chris G.
</BLOCKQUOTE>

<!-- end 8 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/9"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 9 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>icons</H3>


<p><strong>From Joseph Ibbitson 
</strong></p> 
<p align="right"><strong>Answered By Thomas Adam 
<br></strong></p>

<P><STRONG>
Hello Gang
</STRONG></P>
<P><STRONG>
I'm an old guy (81) trying to learn Linux with very little computer
experience.  Strangly, with so much help available online I find it easier to
learn Linux than Windows.  However, one problem I have is just what does the
various icons indicate when left clicking on the file tree.  I see gears,
folders, screens with and without locks, apparently sheets of paper, some
with corners folded over,  cubes of assorted colours, etc. etc.  I am running
Mandrake 7.2.  My main problem is that I cannot find any  instruction on how
to navigate the file system when I don't know anything!  I have yet to find a
book that explains the very basics.  Example-how do I find the proper way to
install software.  I have installed ,I beleive ,Sane,  see it listed but how
do I arrange it so I can Use it?
</STRONG></P>
<P><STRONG>
I hope you will exscuse this rambling requests.  I really want to master
Linux but until I can get over the basics I am having trouble.  Any help you
can give me will be sincerely appreciated.  Thank You.
</STRONG></P>
<P><STRONG>
--
Joseph
</STRONG></P>


<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Thomas]
Hi....
</BLOCKQUOTE>
<BLOCKQUOTE>
Judging from your description, I assume that you are using <A HREF="http://www.kde.org/">KDE</A>. The icons that you see, are supposed (although I admit, I have trouble with this) to help you understand what KFM (the KFile Manager) is doing.
</BLOCKQUOTE>
<BLOCKQUOTE>
Cog wheels indicate that the program is executable
folders indicate just that, that they are folders
screens usually indicate that the program is a script of sorts. Try clicking on it once and opening it in a text editor such as "kwrite" or "kedit"
</BLOCKQUOTE>
<BLOCKQUOTE>
In really basic term, the Linux File system, has various components to it....
</BLOCKQUOTE>
<BLOCKQUOTE>
the root of the file system "<TT>/</TT>" holds folders such as:
</BLOCKQUOTE>

<blockquote><pre>etc
home
usr
rootmnt
</pre></blockquote>
<BLOCKQUOTE>
"etc" holds most of the initialisation scripts that loads as linux is booting (i.e. the output from the kernel)
</BLOCKQUOTE>
<BLOCKQUOTE>
"home" is the folder which stores the users work that is on the system
</BLOCKQUOTE>
<BLOCKQUOTE>
"root" is the folder where all of "roots" work is stored. Root is the system admin of a linux computer and has read\write permissions on every file. In other words root controls everything.
</BLOCKQUOTE>
<BLOCKQUOTE>
"mnt" holds the symbolic links to other partitions on your local machine
</BLOCKQUOTE>
<BLOCKQUOTE>
"usr" is the folder which stores main executables, man pages, etc.
</BLOCKQUOTE>
<BLOCKQUOTE>
Using your file manager as before replace what is already entered at the top, and put <TT>/usr/bin</TT>
</BLOCKQUOTE>
<BLOCKQUOTE>
here you'll find a lot of cog wheel icons. This is the main folder which will store all your programs.
</BLOCKQUOTE>
<BLOCKQUOTE>
Since you are using Linux-Mandrake (as do I) installing software is often done by using RPM's (RedHat Package Manager). To install these, insert your CD and at the console, type:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>cd /mnt/cdrom/Mandrake/RPMS
</font></code></blockquote>
<BLOCKQUOTE>
then type:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>ls
</font></code></blockquote>
<BLOCKQUOTE>
and you'll see a huge long list. To install any RPM (regardless of the path..folder that it is stored in) type:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>rpm -i nameofrpm-1.0-0mdk.i586.rpm
</font></code></blockquote>
<BLOCKQUOTE>
and that should install it (assuming there are no module dependencies!)
</BLOCKQUOTE>
<BLOCKQUOTE>
I know this must seem very vague and confusing, but I believe I have started you off....
</BLOCKQUOTE>



<!-- end 9 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/10"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 10 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Corrupt Tar Archive</H3>


<p><strong>From Mohamed Ezz 
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik
<br></strong></p>

<!-- sig -->

<strong><p>
I have 'ftp'ed an 8MB tar archive file and cannot untar it. I did not do a
checksum after the ftp because I know ftp should do this on its own. When I
run: $ tar xvf myfile.tar I get:
</strong></p>

<strong><pre>tar: This does not look like a tar archive
tar: Skipping to next header
tar: 447 garbage bytes ignored at end of archive
tar: Error exit delayed from previous errors
</pre></strong>
<strong><p>
My problem is I lost the files of which the archive is composed of, so I
can't regenerate it. To make things worse, the archive file on the source
machine (from where I did the ftp) was deleted. So the local archive is my
only hope of retrieving my files.
</p></strong>
<strong><p>
Any help is greatly appreciated.
</p></strong>
<strong><p>
Ezz
</p></strong>

<blockquote><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
There might actually be some hope here! If an archive is tarred <em>and</em>
gzipped, the above is exactly the error that will be returned when you try
to untar it without un-gzipping. Try this:
</blockquote>
<blockquote> <blockquote> <code>
tar xvzf myfile.tar
</code> </blockquote> </blockquote>
<blockquote>
Note that the file really should have been called "myfile.tgz", if that's
what it turns out to be.
</blockquote>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello Ben,
That was it! Thank you so much. Excuse my ignorance about the
extension.
Ezz
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
No worries at all, Mohamed; glad I could help.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
Actually, it makes us happy to learn a problem has been fixed.  Thanks
for letting us know.
</BLOCKQUOTE>

<!-- end 10 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/11"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 11 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>masquerade in sendmail is broken.</H3>


<p><strong>From  Clark Ashton Smith 
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<!-- sig -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
In issue 21 there was an article by the "The Answer Guy"
which explained how to use the masquerade feature in a
local sendmail configuration.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Issue 21 is from 1997. Considering how much Linux has changed since those
days, and how closely "sendmail" is tied into all those changes, relying
on information that old is not going to get you good results.
</BLOCKQUOTE>
<BLOCKQUOTE>
On the other hand, the information that is available for setting up
sendmail is generally pretty poor, and not at all intended for the casual
user; it's pretty nightmarish out there.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
It worked fine with Redhat 5.2 Linux, but I just tried
it on Redhat 6.0 and the FEATURE(nodns) reports that it
is a no-op and I should use the service.switch file to
disable dns lookup.  Well after 5 hours of reading
sendmail faqs, newsgroups and tips I am no closer to
making this work.
</STRONG></P>
<P><STRONG>
I have a simple network with a ppp connection to the
internet.  Many folks out there must have similar setups.
</STRONG></P>
<P><STRONG>
Could someone please show us how to get the masquerade
feature working again?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Well, you <EM>could</EM> look at my article in issue 58, called 
"<a href="../issue58/okopnik2.html">Configuring
Sendmail in RedHat 6.2</a>"; this might be a bit more up to date, and tells
you how to do masquerading. Given the situation that you're in, though, you
might want to try installing "masqmail" - all the features you need in
your situation, made to work well with a masquerading setup, works with
multiple ISPs, and it is <EM>much</EM> less complex than "sendmail".
</BLOCKQUOTE>
<BLOCKQUOTE>
If you set it up and it does what you need, you owe the Oracle an article
on your experience. &lt;grin&gt; Just kidding. Hope this helps.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather] Actually, we really could use the article 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="bottom"> I've tried masqmail
and it looked okay, but, its minimal documentation seems to assume that
you'll be using masqdialer to drive your dialup connection.  I never quite
got around to spending the time to make it deal with changing, but non-dialed
connections (such as laptops often encounter). A little AnswerGang message
or even article about your experience setting it up properly without that
assumption, would be really handy.  Alternatively, now you also know about
masqdial too, and it may make dialing into your ISPs easier.
</BLOCKQUOTE>

<!-- end 11 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/12"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 12 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>neighbour table overflow</H3>


<p><strong>From Berg Alexander 
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<P><STRONG>
hi,
</STRONG></P>
<P><STRONG>
i have the problem that i have a running terminal-server (booting
over net via Root-NFS) system. now we want to add a second subnet to the
server, and all should be okay in the config files. BOOTP is working,
TFTP is working but the client is not able to mount the root-fs, with the
error message "neighbour table overflow"... we also have changed the
nfs-server, no luck...
</STRONG></P>
<P><STRONG>
bye
</STRONG></P>
<P><STRONG>
Alexander Berg
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
The message 'neighbor table overflow' is not about your NFS, it's at a
lower layer than that.
</BLOCKQUOTE>
<BLOCKQUOTE>
It means that your arp cache is overflowing because your machine can't
tell who is on its own subnet... its neighbors.  Which usually means your
localhost setup is broken (because lots of applications use networking
internal to your machine - which is <EM>always</EM> on its own subnet, so those
packets should never even escape the computer) or, far less commonly,
that your netmask for your own external address is wrong.
</BLOCKQUOTE>
<BLOCKQUOTE>
Sadly, tftp and network booting are things I'm not so familiar with, so
perhaps one amongst the rest of The Answer Gang can help tell you where
to correct your terminals' localhost setup.
</BLOCKQUOTE>
<BLOCKQUOTE>
Because this happened when you're adding a new subnet, you may find a need
to set up machines on both subnets with ethernet aliases.  When properly
set up then running ifconfig should result in something like this:
</BLOCKQUOTE>

<blockquote><pre>eth0      Link encap:Ethernet  HWaddr so:me:he:xv:al:ue
          inet addr:192.168.129.15  Bcast:192.168.129.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:5939693 errors:0 dropped:0 overruns:0 frame:0
          TX packets:5971444 errors:0 dropped:0 overruns:0 carrier:0
          collisions:8308 txqueuelen:100
          Interrupt:10 Base address:0xff00

eth0:1    Link encap:Ethernet  HWaddr di:ft:he:xv:al:ue
          inet addr:192.168.64.2  Bcast:192.168.64.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          Interrupt:10 Base address:0xff00

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:3924  Metric:1
          RX packets:320906 errors:0 dropped:0 overruns:0 frame:0
          TX packets:320906 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
</pre></blockquote>
<BLOCKQUOTE>
Best of luck!
</BLOCKQUOTE>
<p><em>
 note, Alexander put antispam hooks in his address when mailing us,
so he never saw his emailed response.  We still hope this helps him and
others with the dreaded Neighbour Overflow. 
</em></p>

<!-- end 12 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/13"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 13 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>VIDEO CARD</H3>


<p><strong>From Tracy
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<!-- sig -->

<P><STRONG>
DEAR ANSWER GUY:
</STRONG></P>
<P><STRONG>
Greetings.
</STRONG></P>
<P><STRONG>
Do you know who is making two-monitor cards for 98SE?
</STRONG></P>
<P><STRONG>
Thanks.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Jonathan]
Oooh, this one has "Heather" written all over it.  Let's stand back.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Nope.  Our focus is on Linux, so we don't think much about windows (except
sharing files with it)... it is possible to use multiple monitors with
linux, as long as you have such a card, and it's something supported by
X version 4.  Before we got that folks had to buy a well tuned commercial
X server (think video driver) called "Metro-X" (<A HREF="http://www.metrolink.com"
	>http://www.metrolink.com</A>).
</BLOCKQUOTE>
<BLOCKQUOTE>
So, as a matter of hardware, you might not need one card that does two
monitors at a time; some video cards are okay with having two of the same
kind in the computer, but then you need a video driver that will survive
the experience.  Some past linux'er described his troubles doing this at
(<A HREF="http://www.tarball.net/docs/multihead_fb_howto.html"
	>http://www.tarball.net/docs/multihead_fb_howto.html</A>) -
</BLOCKQUOTE>
<BLOCKQUOTE>
The keyword you probably want to try in the search engines is "dual monitors"
or "multi-headed" video cards.
</BLOCKQUOTE>
<BLOCKQUOTE>
Perhaps a visit to one of the windows magazine sites would be more useful.
With luck they might have had an article on the topic in the last few months
so you could get some reviewer comments about which are any good in your
platform.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
ENTHUSIASTICALLY YOURS,
tracy
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Best of luck and feel free to give us a buzz if you have any linux questions.
</BLOCKQUOTE>

<Strong><p><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>
DEAR HEATHER,
</strong></p>
<p><strong>
Thanks.  Matrox answered too and it appears that Matrox has a card
(G450) with two outlets which allows two monitors to operate as one big
one.
</strong></p>
<p><strong>
ENTHUSIASTICALLY YOURS,
</strong></p>
<p><strong>
tracy
</strong></p>

<!-- end 13 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/14"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 14 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>fat versus inodes</H3>


<p><strong>From narender
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<P><STRONG>
dear sir ,
</STRONG></P>
<P><STRONG>
i want to know why the viruses are so common in dos and
windows while unix is ammune to these ?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
In order to spread effectively, viruses have to gain system-level privileges
and abuse them.  In DOS and Windows, system level priveleges have no "natural"
defenses - all requests for system services are on behalf of the same user,
you.
</BLOCKQUOTE>
<BLOCKQUOTE>
NT has slightly better natural defenses, but also gets some interesting ones.
</BLOCKQUOTE>
<BLOCKQUOTE>
The ability of viruses to spread seems to be enhanced by some other features
which you would otherwise find handy, like the ability of several apps to
share a single macro language.
</BLOCKQUOTE>
<BLOCKQUOTE>
This is why there are so many antivirus companies - even after they've gone
and bought each other up a bunch.  They're in the business of selling immune
systems and the ability to spot that the machine is "ill" before the symptoms
get obvious.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
is it all due to inodes
concept in the unix ?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
No.  UNIX family OS' all expect different applications to run in seperated
memory spaces, called a process.  If a process (even owned by the same user)
tries to wander out of its allowed space it is killed (that's called a
sementation violation, or segfault).  In addition normal users don't have
full system privileges.  Beyond that, we have a great many macro languages
available and few systems have the same configuration enough that a virus
can be sure of one or another feature being present. Having to make decisions
makes such "invaders" large - larger invaders are more easily spotted, or
may set off other defenses.  So while in theory it's not impossible for a
Linux virus to exist, it's much harder.
</BLOCKQUOTE>
<BLOCKQUOTE>
The main case I know of was basically a research virus - it could only
spread if the system's user also did a few things to improve his ability
to access the system as root when working remotely.  Very few people do
that, or even want to.
</BLOCKQUOTE>
<BLOCKQUOTE>
We have much more to fear from crackers trying to generate these failures
deliberately, than from viruses trying to invade our systems automatically.
</BLOCKQUOTE>
<BLOCKQUOTE>
However, it's worth noting that LILO is a master boot record - it looks
different, but it's still an mbr, so any virus you catch in a dual boot
system that attacks the mbr, will attack your LILO.  That it's code "coming
from Linux" won't save it.  It does have a few defenses, but it's not very
big.  Many other bootloaders exist too, and if you're living in a virus
rich environment you might want to use one that specifically has some
antivirus features.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
if so will you please tell me in more detail
the responcible differences between fat and inodes tables ?
</STRONG></P>
<P><STRONG>
needing yr help
<br>regards
<br>narender
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, it's not the responsible thing, but it's a fair question.
</BLOCKQUOTE>
<BLOCKQUOTE>
FAT is a table at the beginning of the disk, which divides the disk up into
"clusters" and marks how each cluster is used. (There's actually two tables,
so that there is a safe copy in case of problems, but normally, they contain
the exact same data.)
</BLOCKQUOTE>
<BLOCKQUOTE>
inodes contain a small amount of information (called metadata) about the
things they point to, and the things they point to can be put anywhere on
the disk, because part of the metadata says where that is.  We have a
different way of keeping track of what disk space is still free to allocate.
For more about this, study about the "superblock" since we do have things
that affect how many inodes we can use, and so on, as options when we format
a disk under linux.
</BLOCKQUOTE>
<BLOCKQUOTE>
So it is simplest to say that the difference is that FAT directly represents
the disk, but Linux' system indirectly represents the disk.
</BLOCKQUOTE>
<!-- end 14 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/15"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 15 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Installing Linux without cdrom</H3>


<p><strong>From Berry Vos
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<!-- sig -->

<P><STRONG>
Hello,
</STRONG></P>
<P><STRONG>
The other day I bought a second-hand computer. I already had a
laptop-computer with <A HREF="http://www.suse.com/">SuSE</A> linux installed, which is running perfectly. I
bought the new pc to install linux to experiment a bit without messing up
my primary computer in case I do something wrong. The problem is that the
new pc has no cdrom. Also, neither the new pc, nor my laptop has a network
card. I wonder if it is possible to connect the laptop with the pc with a
null-modem cable and install linux off of the laptop on the pc. Is this
possible? If so, how?
</STRONG></P>
<P><STRONG>
I hope you can help me.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
The process is thoroughly described on the last page of the PPP-HOWTO
under "Using PPP across a null modem (direct serial) connection". You will
need an actual null modem cable ("LapLink cable"), detailed in the
Hardware Book (see the "hwb" package), in "ca_Nullmodem9to9.html". Don't
just buy a regular "9-to-9 serial cable" from Radio Shack; it won't work. A
lot of what are called "null modem cables" won't either. Buy a LapLink, or
build one; if you can handle a soldering iron, it takes about 15 minutes.
</BLOCKQUOTE>
<BLOCKQUOTE>
The one thing you want to be aware of is that, at best, you'll be pumping
the data across at just over 11kB/s. To put that in perspective, copying
the  installation CD to your laptop (this is not necessary for the
installation, but it'll give you an idea) will take approximately 16
<em>hours</em>. I've done a <A HREF="http://www.debian.org/">Debian</A> installation by pumping across the base install
files (24MB, about 40 minutes) and running them, then setting up 'apt' to
FTP the individual files from the source machine via the serial link. I
let it run overnight; by the time I woke up the next morning, it was all
done.
</BLOCKQUOTE>
<BLOCKQUOTE>
Not that this is a big deal - it wasn't to me, anyway - but you want to be
aware of the time scale involved. &lt;grin&gt; I do love my serial link. It lets
me get out of buying a PCMCIA NIC for a laptop that's on its last legs
anyway, and 99% of the files are small enough that speed isn't really an
issue.
</BLOCKQUOTE>

<!-- end 15 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/16"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 16 -->
<H3 align="left"><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>Installing RedHat 7.0 and a driver for the Chipset Cirrus CL-GD5436</H3>


<p align="right"><strong>Answer By Wilf
<br></strong></p>

<!-- sig -->

<BLOCKQUOTE>
Hello and Good'ay,
</BLOCKQUOTE>
<BLOCKQUOTE>
Installing RedHat 7.0 was nothing special though programmes seem to run a bit
faster than under RedHat 6.0... but:
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQuote>
Althoug RH 7.0 does recognize the graphics chipset Cirrus CL-GD5436 (on a
Compaq Deskpro 2000) it does not provide the correct driver as RH 6.0
does (though after a long and despairing search I found the information that
the driver was no longer included in "XF86_SVGA-3.3.6-33.rpm" shipped with RH
7.0). The mouse pointer under X (I use WindowMaker) was displayed as a
barrecode found on products, windows overlapped and were redrawn in a rather
chaotical and randomly manner. Other Window Managers (<A HREF="http://www.kde.org/">KDE</A> or <A HREF="http://www.gnome.org/">GNOME</A>) did not
display it any better.
</BLOCKQuote></BLOCKQUOTE>
<BLOCKQUOTE>
Falling back to RH 6.0 or coping with this particular problem? I did both, but
used some (perhaps not necessary force) to install RH 7.0 all the same.
</BLOCKQUOTE>
<BLOCKQUOTE>
I first installed RH 7.0 and configured graphics display with default values.
After installing and BEFORE calling up a graphical display with "startx" I
force-installed the driver from RH 6.0 with:
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQUOTE><CODE>
rpm -i XF86_SVGA-3.3.3.1-49.rpm --force
</CODE></BLOCKQUOTE></BLOCKQUOTE>
<BLOCKQUOTE>
and -much to my delight and would ya believe it, matey?- all possible mode of
graphical displays (640x480, 800x600 and 1024x768, all 16bits colour) now work
as expected: very good indeed!
</BLOCKQUOTE>
<BLOCKQUOTE>
In the event that this procedure could have been done otherwise, I'd appreciate
a comment.
</BLOCKQUOTE>
<BLOCKQUOTE>
Yours Linuxely,
</BLOCKQUOTE>
<BLOCKQUOTE>
Wilf
</BLOCKQUOTE>
<BLOCKQUOTE>
P.S. I take it that section "2 cent tips" rather means "2 US &amp;cent; Tips" or are you
talking EURO &amp;cent;? Mind you, it's just a thought ... 
<IMG SRC="../gx/dennis/smily.gif" ALT=";-)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        > 
[Heather] Well, the usage comes from an idiom for "putting in our two cents
worth" -- cheerfully offered advice, not always wanted, sometimes useful.
So, if you have a local idiom that expresses the same concept but uses a
different chunk of change, then you know the exchange rate 
<IMG SRC="../gx/dennis/smily.gif" ALT=";P" 
		height="24" width="20" align="middle">
Our friends
in the UK can surely use Tuppence Tips.
</BLOCKQUOTE>
<BLOCKQUOTE>
Forcing the old series rpm into place in this fashion is just the right
thing here.  You should be safely able to apply X parts from 6.2
... XFree86-3.3.6-20.i386.rpm ... as an upgrade (<TT>-U</TT>) to the 3.3.3 you
successfully installed, if you like.
</BLOCKQUOTE>

<!-- end 16 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/17"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 17 -->
<H3 align="left"><img src="../gx/dennis/bbubble.gif" 
	height="50" width="60" alt="(!) " border="0"
	>cd-writing mini-howto</H3>


<p align="right"><strong>Answer By Chris Coyle
<br></strong></p>

<!-- sig -->

<BLOCKQUOTE>
I found the "CD-Writing with an ATAPI CDR Mini-HOWTO"
(<A HREF="http://www.linuxgazette.com/issue57/stoddard.html"
	>http://www.linuxgazette.com/issue57/stoddard.html</A>)
very helpful.  Thank you.
</BLOCKQUOTE>
<BLOCKQUOTE>
Here are a couple of suggestions which other readers
who are interested in the same subject may find useful.
</BLOCKQUOTE>
<BLOCKQUOTE><ol>
<li> (very minor) I think it should be <TT>/etc/modules.conf</TT>
not <TT>/etc/conf.modules</TT>
<li> I just discovered that the ide-scsi module
in kernel 2.2.17 (from RH rpms I just DL'ed),
either has a big problem, or else it is significantly
incompatible with previous kernels.
</ol>
</BLOCKQUOTE>
<BLOCKQUOTE>
Here's what happened:
</BLOCKQUOTE>
<BLOCKQUOTE>
I DL'ed and installed the RH rpms for kernel 2.2.17-14.
These were recommended in a security advisory.
I kept my previous kernel (2.2.16) installed just in case,
adding a new section to <TT>/etc/lilo.conf</TT> by copying the
previous one, mutatis mutandis.  Then I ran lilo and rebooted.
At first everything appeared to be OK with the new kernel,
but then I tried to mount my cd-rom and it failed, giving
the message
</BLOCKQuote>
<BLOCKQUOTE><pre>
mount: wrong fs type, bad option, bad superblock
on /dev/cdrom or too many mounted filesystems
</pre></BLOCKQUOTE>
<BLOCKQUOTE>
While I was searching for the cause of this I remembered
that I had set up my cd-recorder to use ide-scsi.
</BLOCKQUOTE>
<BLOCKQUOTE>
My regular cd-rom reader is hdc and the cd-recorder is hdd.
Following the directions in your mini-howto, I had inserted
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQUOTE><code>
append="hdd=ide-scsi"
</code></BLOCKQUOTE></BLOCKQUOTE>
<BLOCKQUOTE>
and this I had copied faithfully into the new 2.2.17 section.
When I removed it and rebooted, I found I could mount the
cd-rom again.  Then I put the line back in and rebooted.
</BLOCKQUOTE>
<BLOCKQUOTE>
This time I looked at what scsi devices were detected.
Eureka!  By looking at dmesg and also by using
"cdrecord -scanbus" I discovered that the ide-scsi module
had taken over both hdc and hdd, even though I requested
only hdd.  I asked for help on comp.os.linux.misc and
within hours someone else confirmed the same thing, namely
</BLOCKQUOTE>
<BLOCKQUOTE>
"...if you have two devices on an IDE channel,
and one of them is under ide-scsi emulation,
it's better to treat both of them as if they
were under ide-scsi emulation.
</BLOCKQUOTE>
<BLOCKQUOTE>
I don't know if this is due to a an error or a design
change, but the work-around was quite straightforward.
</BLOCKQUOTE>
<BLOCKQUOTE>
The only tricky bit was that I wanted to be able to
boot 2.2.16 so I had to devise a way to make both kernels
boot up in a state where they could use the same devices
and configuration files.  My solution is as follows:
</BLOCKQUOTE>
<BLOCKQUOTE><ol type="a">
<li> Change the lines in <TT>/etc/lilo.conf</TT> to
<BLOCKQUOTE>
append="hdc=ide-scsi hdd=ide-scsi"
</BLOCKQUOTE>
in both kernel sections.
<li> Move the <TT>/dev/cdrom</TT> link from hdc to scd0.
<li> Change the scsi configuration for the cd-recorder
in <TT>/etc/cdrecord.conf</TT> to 0,1,0 (since it is now the
second scsi host).
</ol></BLOCKQUOTE>
<BLOCKQUOTE>
After all that I am finally back to the point where I
can mount the cd-rom and use the cd-recorder, with either
2.2.16 or 2.2.17 kernel.
</BLOCKQUOTE>

<!-- end 17 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/18"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 18 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Reading the logs</H3>


<p><strong>From Andrew
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<!-- sig -->

<!-- ::
Reading the logs
~~~~~~~~~~~~~~~~
:: -->
<P><STRONG>
Hello Mr Answer Guy,
</STRONG></P>
<P><STRONG>
While i'm here i'm going to get my 2
cents worth &amp;, so throw a few questions at you ( hehe that's funny since
you offer your knowledge for nicks). I'll get in now before you decide
to go commercial  
<IMG SRC="../gx/dennis/smily.gif" ALT="8^)" 
		height="24" width="20" align="middle">..
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Some of us are consultants, for those who enjoy directly working with a
linux guru, or to get guaranteed an answer of some sort - TAG gets a lot
more mail than anybody can really answer, and complicated or non linux
things often get ignored.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Running Redhat 6.1
1./ 1st thing is as soon as i decide to start logging Kernel logs
to <TT>/var/log/kernel</TT> via syslog.conf i get the following
</STRONG></P>

<pre><strong>Mar 28 14:20:12 echelon kernel: klogd 1.3-3, log source = /proc/kmsg started.
Mar 28 14:20:12 echelon kernel: Inspecting /boot/System.map-2.2.12-20
Mar 28 14:20:12 echelon kernel: Loaded 6865 symbols from /boot/System.map-2.2.12-20.
Mar 28 14:20:12 echelon kernel: Symbols match kernel version 2.2.12.
Mar 28 14:20:12 echelon kernel: Loaded 168 symbols from 12 modules.
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
That part's normal...
</BLOCKQUOTE>

<Pre><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Mar 28 14:20:12 echelon kernel: VFS: Disk change detected on device ide1(22,64)
Mar 28 14:20:44 echelon last message repeated 17 times
Mar 28 14:21:46 echelon last message repeated 31 times
Mar 28 14:22:47 echelon last message repeated 30 times
Mar 28 14:23:49 echelon last message repeated 31 times
Mar 28 14:24:51 echelon last message repeated 31 times
Mar 28 14:25:52 echelon last message repeated 30 times
</pre></strong>
<P><STRONG>
(What does this mean???)
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Uh, that it's gone crazy thinking there's a disk change when there's not.
Ide1 is your second IDE chain, so maybe your CDrom, or an ls-120 bay.
</BLOCKQUOTE>
<BLOCKQUOTE>
Removable media bays have either optical or mechanical sensors to detect
that new media has arrived ... enough dust particles can screw up either
one.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I have included my syslog.conf . Do you have any idea how i can stop this
ocurring?? I thought it had something to do with having multiple things
pointing to the same place
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, if you have two devices on your second IDE chain, check that they
aren't both set to master, or both set to slave, in their jumpers.  It's
only a guess but if the BIOS let them get this far in such state, the
kernel could be confused who was talking, and have assumed it was a disk
change.
</BLOCKQUOTE>
<BLOCKQUOTE>
But I'd do a shutdown and try a clean air cannister anyway, it doesn't
hurt.  Don't forget to cover your mouth, there are usually a lot more
dust bunnies than I expect when I do this.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
2./   Should I be concerned with this . I get it continually in my logs
</STRONG></P>
<Pre><STRONG>
Mar 28 12:01:02 echelon sendmail[25388]: f2S212W25388: forward /home/Users/andrew/.forward.eziekiel: World writable directory
Mar 28 12:01:02 echelon sendmail[25388]: f2S212W25388: forward /home/Users/andrew/.forward: World writable directory
</STRONG></Pre>
<P><STRONG>
I mean obviously if i am to receive mail this would need to be writable
from ,as it says the world. I am right in thinking that aren't I ??
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
No, what this is saying is, since your home directory <TT>/home/Users/andrew</TT>
turns out to be world writable, anybody else who ever logged into your
system could change your .forward.  That's a security problem, some utter
stranger could get your mail, and the kind folks at sendmail got tired
of people claiming that such lossages (whether pranks or malicious) were
some sort of bug in sendmail.  So, it checks.
</BLOCKQUOTE>
<BLOCKQUOTE>
You should either fix your home from being world writable (after all,
your other stuff is vulnerable too) or, you can set the DONT_BLAME_SENDMAIL
feature in sendmail, and it will stop checking for silly things like these.
And your own fault if it breaks wickedly because of weird permissions.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
There are so many questions I have when it comes to Linux.
</STRONG></P>
<P><STRONG>
3./  When I shut down X I might see these errors. They don't mean that
</STRONG></P>
<P><STRONG>
much but I would love to know how to fix then . These are found
in .xsession-errors
</STRONG></P>
<pre><strong>
xscreensaver-command: no screensaver is running on display :0.0
Xlib: connection to ":0.0" refused by server
Xlib: Client is not authorized to connect to Server
xscreensaver: Can't open display: :0
xscreensaver: initial effective uid/gid was root/root (0/0)
xscreensaver: running as nobody/nobody (99/99)
rm: cannot remove `/root/.gnome//gmc-aoiM8A': No such file or directory
subshell.c: couldn't get terminal settings: Inappropriate ioctl for device
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
When you <EM>shut down</EM> X numerous things will lose their server connections.
If the xscreensaver stuff is happening during startup of X you probably
have to fix your .Xauthority or something.
</BLOCKQUOTE>
<BLOCKQUOTE>
rm not being able to remove absent files, that's not a bug, it's just being
noisy.
</BLOCKQUOTE>
<BLOCKQUOTE>
Usually apps that use ioctls recover from ioctl glitches, since ioctls
are so "close to the bare metal" they behave differently on a lot of systems.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
4./ When I start a ppp session via <tt>ifup ppp0</tt> I get the following
</STRONG></P>
<P><STRONG>
command not found but then I kicks in anyhow &amp; dials up without problem.
<br>Wish I could fix that strange one
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Your chatscript probably tells it to run an apps which is not installed
on your system.  The ppp documentation is hug, but most of the control
files are plain text under <TT>/etc/ppp</TT> or <TT>/etc/chatscripts</TT>
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
5./  I think snort is a great program but it still throws some false alarms
I constantly see info I don't need to like the following
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, I don't use snort so I can't explain its stuff. 
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Then the like of this error
</STRONG></P>
<Pre><STRONG>
Mar 27 01:15:20 echelon pam_console[11450]: can't find device or X11 socket to examine for 1.
</strong></pre>
<P><STRONG>
Can you suggest a book that gets away from the obvious within Linux &amp;
helps with questions that aren't as common like the last one for example..
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
 X however, uses a
special breed of networking internal to your box, called "UNIX domain
sockets".  So that's the kind of socket it's talking about looking for.
What sort of examination it wanted to do I still can't say.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thankyou
</STRONG></P>
<P><STRONG>
Andrew
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Hope that helped.  There are lots of Linux books, but I'm used to
recommending towards a less technical crowd.  Some linux-y things you
were asking about above are not very linux specific, so good UNIX books
can help too.
</BLOCKQUOTE>
<BLOCKQUOTE>
Jim Dennis wrote a nice book "Linux System Administration" from New Riders,
but it's more an explanation of planning and things to do in being a
daily sysadmin, not "how to read syslogs".  Mr. Sobell's "Hands On Linux"
is good for getting people to swimming level in the Linux icy seas, but
again, it's more about doing things, and less about logs reading.
</BLOCKQUOTE>
<BLOCKQUOTE>
Not that I'm trying to discourgae you!  If more sysadmins cared a bit
what the messages their logs contain really mean, I think many systems
would be healthier.  I just don't know a book that's the kind of reference
you're thinking of.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello Heather,
</STRONG></P>
<P><STRONG>
Wow you were right on the money with these kernel
errors. I have just added a removable harddrive to this computer so i'll
look into the jumper setting..Thanx
</STRONG></P>
<P><STRONG>
The one i'm not to sure about though is the sendmail part. My permisions for
lets say my account/user directory is as follows
</STRONG></P>

<pre><strong>drwxr-xr-x   28 andrew   users        4096 Mar 29 12:52 andrew
</strong></pre>
<P><STRONG>
What permissions would you suggest here &amp; for my other users ???
</STRONG></P>
<P><STRONG>
Thanks  agian
</STRONG></P>
<P><STRONG>
Andrew
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Your home directory looks okay, maybe you should see if any directories
further up the chain are world writable.
</BLOCKQUOTE>
<BLOCKQUOTE>
The really security conscious person might have one group per user, and
reserve use of the group named "users" that contains normal accounts, for
things for all the people to use, so that they can avoid world writable
directories at all.  Unfortunately directories and files can only belong
to one group at a time.  And it's a little odd to make your home world
<EM>readable</EM> too, but not uncommon, and in a private system, not so much
of a big deal.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello Heather,
</STRONG></P>
<p><strong>
Just a quick message to again say thankyou very much for your prompt
email reply. Un fortunately my friends &amp; collegues are more windows based so
i cant call on to many people for help when Linux hiccups..
</STRONG></P>
<p><strong>
Being able to ask people like you these strange types of questions help sooo
much
</STRONG></P>
<p><strong>
Cheers
<br>Andrew
</STRONG></P>

<!-- end 18 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/19"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 19 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>So many users, So few POP accounts</H3>


<p><strong>From Thomas Nyman
</strong></p> 
<p align="right"><strong>Answered By Mike Orr
<br></strong></p>

<!-- sig -->

<!-- ::
So many users, So few POP accounts
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:: -->
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I have recently entered the magical world of Linux (<A HREF="http://www.redhat.com/">Red Hat</A> 7.0). In the
You see I would like to configure my linux machine so
that it polls a couple of pop accounts via a dialup ISP, and the distributes
any mail to users on the local network. In my mind a reasonable request. I
understand that sendmail and fetchmail can be used in this respect (although
sendmail "sends" mail and does not collect it). I have so far been inable to
find out exactly what I need to configure (besides fetchmail) to do this. I
have also tried to configure sendmail to no avail, it keeps complaining that
I have not set a que and have not set a mailbox...but try finding a how-to
that tells you how to setup a que and a mailbox locally - I cant do it. I
have also tried to instal qmail. I've downloaded a tar.gz file. Unpacked it
with gunzip and the run the tar -xvf comman on it. SO fall all looks fine. I
have then followed the install instructions and goes well untill I reach the
part of the install instruktions that instructs me to "make setup check",
many attempts have been made but qmail simply will not understand the
instruction..hence I cant continue the installation....ah but I digress...the
point is I want to collect popmail from different pop accounts and distribute
it to either eudora och outlookexpress on windows machines..can I do
this..and if so whaich programs do I need to configure and how???
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
Fetchmail works by popping the mail down, changing the envelope-to
address and passing it on to the local mail-transfer program for final
delivery.  So the first step is to get a working mail-transfer program.
This can be sendmail, qmail, exim, postfix, smail, etc.
</BLOCKQUOTE>
<BLOCKQUOTE>
The next step is to set up your .fetchmailrc.  Assuming all the mail
from each pop account is going to a single user, you can use a configuration
like this:
</BLOCKQUOTE>

<blockquote><pre>poll pop.my-isp.net
	proto pop3
	user bob there with password XXXXX is bobby here

poll pop.my-other-isp.net
	proto pop3
	user frederick there with password YYYYY is fritz here
</pre></blockquote>
<BLOCKQUOTE>
Now, each time fetchmail runs, bobby and fritz will find their pop mail in
their Unix mailbox.  You would then need to make that mailbox visible to
Eudora or Outlook Express somehow, but that's another issue.
</BLOCKQUOTE>
<BLOCKQUOTE>
If your mail transport agent seems to be working but popped mail is still
being lost, use fetchmail's -v flag to determine whether fetchmail is
generating the correct recipient address and whether the mail transfer
agent is accepting the message.
</BLOCKQUOTE>
<BLOCKQUOTE>
If you wish to distribute mail from a single pop account to several
Unix accounts, it's more complicated.  You could have
fetchmail deliver it all to a single account which then uses procmail to
distribute it (e.g., according to a special prefix in the subject).
Or you could use uucp instead of pop/fetchmail.  Uucp was designed for the
"my site has multiple users but I only have one ISP account" problem, but
pop was not.  Pop was designed assuming each user would have their own
mailbox at the ISP.  However, finding an ISP that supports uucp nowadays
is difficult, they may want a higher price for it, the configuration
would be more complicated, and it would probably work best if you had your
own domain.
</BLOCKQUOTE>

<!-- end 19 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/20"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 20 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>script</H3>


<p><strong>From Paul Wilts  
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello there, I am hoping you can help me out. I am writing a script. I have
a file that has two columns. one column with numbers and one column with
names. This file stores users disk-usage/user-name ie: 50000   paul. I would
like to run a script/command that would look into the file and if a user is
over a certain number , that number/numbers along with the user/users name
is copied from that file and put into a different file. I have tried almost
everything I know, which is limited, but have not had any success.
Thank you for your help
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Well, you don't say what it is that you tried, or what language the script
is in, but I'll take a flyer in a "bash" one-liner. If we have a file
called "quotas" that looks like this:
</BLOCKQUOTE>

<blockquote><pre>5       joe
7       jim
12      jack
10      jeff
20      jose
1       jerry
3       jenny
8       jamal
6       jude
</pre></blockquote>
<BLOCKQUOTE>
and we want only those users whose numbers exceed, say, 7, then we might
do something like this:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>while read a b; do [ $a -gt 7 ] &amp;&amp; echo $a $b; done &lt; quotas
</font></code></blockquote>
<BLOCKQUOTE>
What we've done here is read in each line and load the strings into two
variables, $a and $b. We then check to see if $a is greater than our
target number, and echo both of them if it is.
</BLOCKQUOTE>
<BLOCKQUOTE>
Note that the whitespace between the names and the numbers is ignored by
'read'; I only put it in to demonstrate how clever "bash" is about stuff
like that. 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>
<BLOCKQUOTE>
You could also do it in Perl -
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>perl -wane 'print if $F[0] &gt; 7' quotas
</font></code></blockquote>
<BLOCKQUOTE>
- split each line into an array, print if the 1st member of the array
(arrays are indexed starting from 0) is greater than the target.
</BLOCKQUOTE>
<BLOCKQUOTE>
That should save lots of wear and tear on your fingers. 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thanks very much for your help. Yes I was using Bash. I tried using the test
and expr. What would you suggest for a good web site that would also be a
good reference for information on scripts.  Once again thanks.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Heh. I might have a suggestion.
</BLOCKQUOTE>
<BLOCKQUOTE>
A while ago, I wrote a 6-part series right here in Linux Gazette called
"Introduction to Shell Scripting". It's been translated into 7 languages,
and is used in at least two college courses. It was intended as a basic
text - don't expect to be introduced into The Deepest Mysteries - but I
believe that it's a very good start for anyone trying to learn shell
scripting, and should get you up to basic competence in short order.
</BLOCKQUOTE>
<BLOCKQUOTE>
Take a look at <EM>LG</EM> issues 52-55 and 57-58 or
&lt;a href="<A HREF="http://www.linuxgazette.com/search.html"
	>http://www.linuxgazette.com/search.html</A>"&gt;search for my last
name (Okopnik)&lt;/a&gt;, since one of the articles got misnamed in the
e-mail shuffle.
</BLOCKQUOTE>

<!-- end 20 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/21"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 21 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Linux Box on windows</H3>


<p><strong>From Uri Rado
</strong></p> 
<p align="right"><strong>Answered By Heather Stern, Breen Mullins
<br></strong></p>

<!-- sig -->

<P><STRONG>
How can I install a linux windows on windows?????
<br>Thanks!!!!!!!!
</STRONG></P>
<P><STRONG>
Uri Rado.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Hmm, such a simple question, so many ways to interpret it.
</BLOCKQUOTE>
<BLOCKQUOTE><ul>
<li> "a linux windows"
<BLOCKQUOTE>
The usual graphical system for Linux is called the X Window System (no
plural) or often, just X.  The usual flavor of that we use is "Xfree86"...
even if we're on Alphas and some other hardware, but since you're using
Windows(tm) you're pretty likely to be using 80x86 compatible stuff.
</BLOCKQUOTE>
<BLOCKQUOTE>
The part that makes it look impressive is called a "window manager" and
many of them, even the really plain ones, support themes, though not the
same theme files from Windows without a little help.
</BLOCKQUOTE>
<li> "on windows"
<BLOCKQUOTE><ol>
<li> a linux windows "inside" windows ... to run like a windows app?
<blockquote>
One possible way is to use VMware for Windows, and install Linux into
the child volume that it would create for you.  This can be a little
tricky to setup, but may not be too bad, and allows you to run Windows
and Linux things at the same time.  Yes, we here in The Answer Gang are
more likely to do it the other way, running Windows inside of Linux, but
it's your choice to make, not ours 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</blockquote>
<li> a linux windows "inside" a windows partition, but to boot into
from windows.
<blockquote>
There are several distributions (many of them describe themselves as a
"loopback loading" linux) whose install does not repartition your drive,
just uses up a bunch of space inside of your C: ... I have little experience
with them personally so I can't tell which one is the best. (PhatLinux
got a lot of attention at one point :D)  All I can say is that Linux Weekly
News' "Distributions" section (<A HREF="http://www.lwn.net"
	>http://www.lwn.net</A>) has the best listing,
and you'll want to look at several web pages before picking one, if this
is what you wanted.
</blockquote>
<li> I'm using windows, how do I install linux without wiping my windows?
<blockquote>
Many of the commercial distros promise to do this, and they have varying
degrees of success... the most frightening part is the part about resizing
your drive space so Linux can use the rest.
</BLOCKQUOTE>
<BLOCKQUOTE>
Don't feel bad, it's frightening to us old hands too - and it should be.
There was never a better time to make a backup!
</BLOCKQUOTE>
<BLOCKQUOTE>
Partition Magic is reeeeally popular for doing that, in fact, so popular
that it is what a couple of the big Linux distros use.  But, I have heard
of a few rare cases where it didn't do things right - and if you have
Windows ME you desperately, importantly, don't-take-maybe-for-an-answer,
<EM>need</EM> to use the newest version, or it won't work.
</BLOCKQUOTE>
<BLOCKQUOTE>
FIPS used to be the handiest, but it's a DOS app.  It's also quite old
and unmaintained.  Basically if you have a new enough version of Windows
that you're not sure how to boot from a floppy into a DOS prompt... or
that you *can't* ... you shouldn't use FIPS anymore.   Its docs <EM>are</EM>
still valuable reading though.
</BLOCKQUOTE>
<BLOCKQUOTE>
GNU Parted is the current good tool to use if you come up in Linux.  I'm
not sure if it's seen any problems with those WinME partitions or not,
but it is being actively developed, at least...
</BLOCKQUOTE>
<BLOCKQUOTE>
The most careful way is to do your own defragging and backups, use any one
of probably several available tools to cut your C: into a C: and D:, and
then, start your selected Linux distro installer.  The space assigned to
be your D: can safely be chopped up into parts for Linux (for ordinary
flavors of Linux) or used as the resting place for Linux parts (for
distros which use UMSDOS, live in FAT filesystems directly, or use
"loopback loading".)
</BLOCKQUOTE>
<BLOCKQUOTE>
If you have an IDE drive your C: will almost certainly be <TT>/dev/hda1</TT> when
mentioned in the Linux tools, and should be left alone and absolutely NOT
formatted during your installation of Linux.  (For SCSI systems, <TT>/dev/sda1.</TT>)
</BLOCKQUOTE>
<li> I just want to make my windows box look like Linux so my nerdy friend
will stop bugging me
<BLOCKQUOTE>
(Hey, it's april coming up, isn't it? 
<IMG SRC="../gx/dennis/smily.gif" ALT=";P" 
		height="24" width="20" align="bottom">) 
 Well, it might be a little hard
to make it look like K desktop, but there are apps for virtual workspaces
in MSwin, and I <EM>bet</EM> someone has a hack to replace that "Start" logo with
you own thing, so with the right backdrop, a kinda boring enough cursor,
and maybe being able to tune the bevels so they look flatter (can it do
that naturally?  Been a while since I looked) ... plus some effort with
an ico editor turning some popular "gnome" and "K" icons into MSwin icons...
</BLOCKQUOTE>
<BLOCKQUOTE><DL><DT>
Maybe a student's project "icoutils" could help you out here:
<DD><A HREF="http://www.student.lu.se/~nbi98oli/src/icoutils-0.12.0.tar.gz"
	>http://www.student.lu.se/~nbi98oli/src/icoutils-0.12.0.tar.gz</A>
</DL></BLOCKQUOTE>
<BLOCKQUOTE>
You'd have to be a little bit programmer instead of artist though, he says
turning images back into .ico's is "not yet implemented".  But it was
just updated this week, so let's encourage him, okay?
</BLOCKQUOTE>
<BLOCKQUOTE>
Anyways don't forget to use Windows' features to make the icons 48x48 size
and not to lose pixels by getting stuck with 32x32 "old style" .ico or it
won't look very linux-y...
</BLOCKQUOTE>
<BLOCKQUOTE>
Hopefully, your buds would get the hint that by applying apropos Windows
tricks, that you are nerdy enough in your own right, in the environment
you prefer.
</BLOCKQUOTE>
</ol></blockquote>
</ul></BLOCKQUOTE>
<p>[about adding Linux to an existing Windows install]</p>
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Breen]
WinME apparently added a new and bizarre way of reporting cylinder
numbers on large drives (the physical cylinder modulo 1024 or some
such) which confused the dickens out of Parted. I don't know if
the fix is in the latest released version but it has been reported
on the parted list.
</BLOCKQUOTE>
<BLOCKQUOTE>
Make very very sure that you're using the latest version of whatever
tool you're using if you've got WinME anywhere near your box.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        >
Thanks!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
</strong></p>
<p><strong>
Uri Rado.
</strong></p>

<!-- end 21 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/22"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 22 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Take a Breath!</H3>


<p><strong>From Wolf
</strong></p> 
<p align="right"><strong>Answered By Jim Dennis, Heather Stern
<br></strong></p>

<!-- sig -->

<!-- ::
Take a Breath!
~~~~~~~~~~~~~~
:: -->
<P><STRONG>
Answer Guy,
</STRONG></P>
<P><STRONG>
I just recently read that message from Rik, having trouble with
some bad clusters under Windows. Anyway, I used to run a Win98
system, and I experienced the same problems, even if they might not
be related to Rik's (just bad hardware or mis-partitioning, I
assume). I had my HD partitioned into one primary and one extended
partition, with approximately 20 gig each (HD is a Samsung 40.8
gig). Then, I used Partition Commander to make 4 out of the primary
partition: One FAT32 (14 gig), and 3 FAT16 (2 gig each); and on the
extended partition I put a 12 gig FAT32 and 8 gig for Linux,
pre-formatted with Partition Commander. Now I know, that Linux
doesn't like anything above 1024 cyl., but I always assumed, that
it's translated in such a way, that only 1024 are reported (or am I
wrong?). Installing Linux on that last partition was a bold
failure. First, I never got it to boot from the HD, no matter if I
used Loadlin or (tried) LILO, in the MBR or on the partition. When
it booted after the CD install, fsck found a load of errors, and
they all seemed to be beyond 32 gig. So I deleted the last
partition and reformatted it as FAT32, which seemed to succeed
without errors; even with bad clusters checking on. Letting a disk
utility have it's way with it later, revealed again a bunch of bad
clusters, and again. above the 32 gig limit. Not sure, if that's an
OS or a hard error? Right now, I'm going to low-level format and
re-partition the drive, then assign Linux all the space (this
version doesn't like partitions picked from the middle of the
drive). Hopefully it boots. Or the HD craps out altogether, there's
still warranty on it... Had anyone reporting similar stuff with a
big HD like mine is? Windoze is good for all kind of surprising
crap, but I need it for development...
</STRONG></P>
<P><STRONG>
Thx,
</STRONG></P>
<P><STRONG><DL><DT>
Wolf
</DL></STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [JimD]
Wolf,  I promise, no one was going to interrupt you.  [This whole
messages was all in one line of text!]  Perhaps a few paragraph
breaks would have also helped.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Heh. The advantage of email is that folks can use paragraphs, commas
and periods - but still get their whole say before anyone <EM>gets</EM> to
interrupt.
</BLOCKQUOTE>
<BLOCKQUOTE>
To give the short form of the answer -- it's not an OS error at all.
Bootloaders come before the OS, whether DOS or Linux.  It's a dependency
on firmware features - the BIOS on your system either does, or doesn't,
have 1024 cylinder problems.  If your BIOS doesn't have the boundary, the
bootloader still has to make <EM>different</EM> calls to ask about later areas
of the disk, because this is a newer feature, and is tacked on to the
BIOS design.
</BLOCKQUOTE>
<BLOCKQUOTE>
I can just about guarantee that the final 8 Gb on a 40 Gb drive would
be above that boundary!
</BLOCKQUOTE>
<BLOCKQUOTE>
LILO has an old keyword to beat this boundary (linear) and a new way (LBA)
but this drive is so large you may need an even more special call (LBA32).
Linear basically asks for cylinder/head/sector stuff.  The LBA flavors
tell LILO to make the new BIOS calls.
</BLOCKQUOTE>
<BLOCKQUOTE>
You mentioned the drive but not the motherboard - assuming it's modern
enough, I'd try the LBA32 keyword in <TT>/etc/lilo.conf.</TT>  (On a line alone.)
If that doesn't work I'd probably use Loadlin, throw an icon for its
correct command line on my desktop (looking like Tux of course) and forget
worrying about it.  Just remember to copy your kernel to the right place
on your FAT drive that loadlin expects, whenever you decide to update your
kernel.
</BLOCKQUOTE>
<BLOCKQUOTE>
Unfortunately, that you're seeing the drive poorly when the 8 Gb are FAT
partitioned or ext2 partitioned, implies that I may be wrong about how
modern it is.  Or, the drive has memorized something poor about how to
present itself to us.  (Yes, drives have had their own brains for a while,
that means they get to be artificially stupid sometimes too.)  So, I'd
check if your motherboard manufacturer has a BIOS revision, because it
may help some.  And during low level format I'd be really extra picky
about looking through the options, in case something leaps out as meaning
"My motherboard is so stupid I can't even see all 40 Gb.  Just give me 32
of it.  Thanks."  Or as we'd say, ack!  No thanks!
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [JimD]
I'm responding to your message to dispel the misconception that
you've repeated here.
</BLOCKQUOTE>
<BLOCKQUOTE>
Linux has no problem at all with anything past the 1024th cylinder.
The Linux kernel can handle any commodity drive.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
vmlinuz doesn't care, no more than Windows should care how big its C: is -
by the time you're this far, you're in protected mode, and not using the
BIOS directly anymore.  It's fdisk that gets all the headaches.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [JimD]
However, the <EM>bootloader</EM> (LILO) has traditionally been constrained
by the level of support offered by your system BIOS (or the lack
thereof).  Once you get the Linux kernel "bootstrapped" (loaded into
memory and running) then it can easily handle just about any
arrangement of partitions.  LILO has to ask the BIOS to locate and
load the specific device and blocks in which the kernel (and any
initial RAM disk) are stored.  So the BIOS must support calls to
access these devices.  If the BIOS only supports calls to handle
the first 1024 cylinders of a device (a common constraint several
years ago) then we have to locate the kernel (and our RAM disk)
within those 1024 cylinders.  Alternatively we can use a different
bootloader (syslinux off of a floppy, or Zipdisk, etc; grub,
<TT>LOADLIN.EXE</TT>, etc).
</BLOCKQUOTE>
<BLOCKQUOTE>
Now over the years that have been several different workarounds
to this problem.  First we note that SCSI drives have normally
not been afflicted with these limitations (since they don't
emulate the old WD-1003 controller interface; they have their
own BIOS extensions which provide the necessary support through
the "INT 13" calls).  Also we note that this problem is specific
to the PC (so it's never been a problem on Macs, SPARCs or the
many other platforms that Linux supports).
</BLOCKQUOTE>
<BLOCKQUOTE>
Also most IDE drives, though they mostly emulate the ST-506
interface, mostly as implemented by Western Digital's old 1003
chipset; they will perform their own "autotranslation" internally
translating "virtual head" addresses into larger cylinder numbers.
Later these drives dropped all pretense of using cylinder/head/sector
(CHS) co-ordinates and used a technique called LBA (linear block
addressing).  That basically means that any block request it gets
(which comes in the form of a cylinder/head/sector triplet) is
translated into a single number (multiple the three numbers in the
triplet together) and that is fetch according to the drive's own
indexing and mapping.  BIOS' then started supporting LBA which
overcame the 8Gb limit.
</BLOCKQUOTE>
<BLOCKQUOTE>
Meanwhile the latest versions of LILO support the necessary BIOS
call extensions to boot from any cylinder on any IDE drive; so
long at the PC BIOS in question also supports the extension.
</BLOCKQUOTE>
<BLOCKQUOTE>
I've described alternatives to LILO <EM>MANY</EM> times in our column.
Since this is almost always an issue of installing Linux unto a
system that already has a copy of MS-DOS (or any of it's
ilk, Win '9x, OS/2, etc) then it's usually easiest to configure
you system to boot into MS-DOS and to run a program called
<TT>LOADLIN.EXE</TT> to load your Linux kernel.  Because MS-DOS	is
being booted from "the first" partition on the first or second
drive (the only supported configuration) and because it has
access to the "rest" of the drive (with it's device drivers, and
various Microsoft supplied extensions) then LOADLIN can load
any kernel that MS-DOS can "see."
</BLOCKQUOTE>
<BLOCKQUOTE>
Anyway, this issue is old and obsolete.  Please reconsider before
you repeat this misconception any further.  This is not a "Linux"
problem.  It is a PC problem which has been faced (and addressed)
by many Linux users *because Linux doesn't impose arbitrary
constraints on how you configure your filesystems*.  Linux doesn't
make you install it in the first drive or the first partition,
etc.  Unfortunately with the freedom has common the confusing
choices that have caused so many questions among Linux users.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Having choices available, means actually having to make choices.  It's
a tough job sometimes but I vastly prefer it to the alternative.
</BLOCKQUOTE>
<BLOCKQUOTE>
This is the part where I rant about how if we improve the documentation
enough at least these can be <EM>informed</EM> choices.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [JimD]
LILO and the related questions are confusing to converts from
MS-DOS	and Windows, and they are just as confusing for old hand
UNIX users coming from RISC platforms, and for converts from
the old SCO and other PC UNIX platforms.
</BLOCKQUOTE>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Yup.  There are other bootloaders around too, which are easier for some
but each has their own new flavors of problem.  And not all of them can
get over this 1024 thing, which is to say, they actually expect the BIOS
to be helpful.  Can't always trust that.  Welcome to the PC.
</BLOCKQUOTE>

<!-- end 22 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/23"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 23 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>about the adaptation.</H3>


<p><strong>From Meltem YAGLI  
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<!-- sig -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello,
I am the researh assistant in Eastern Mediterranean University, and i am doing
master in computer engineering department. About the adaptation of linux with
other operating systems, i have a problem. if you can help me, i will be very
happy.
</STRONG></P>
<P><STRONG>
I have a program that is written in C and my operating system is linux now.
This program has been done before on dos operating system, (it includes
stdio.h, stdlib.h, conio.h, etc..). so, if i want to run this program, linux
can not find conio.h. (there is no such library file in linux)
Could you help me for this. i wonder, is there any correspondingfile in linux
that can do the functions of conio.h? The only chance isto include this
corresponding file to my program.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
&lt;Smile&gt; I certainly hope that it's not the <EM>only</EM> chance. If you take a
look at the very top of "conio.h" on a DOS machine, you'll probably see
something like the following (this is from the "conio.h" that came with
the old Borland Turbo-C):
</BLOCKQUOTE>

<blockquote><pre>/*	conio.h

	Direct MSDOS console input/output.

        Copyright (c) Borland International 1987,1988,1990
	All Rights Reserved.
*/
</pre></blockquote>
<BLOCKQUOTE>
That's why there's none in the standard C 'include's for Linux: it's a
DOS-specific library! Now, I've done very little C programming in the last
few years - mostly just little quick things - so I haven't had to deal with
any fancy console stuff, and no need for anything like "conio.h". If I did,
the library that is commonly used in Linux for console I/O is "curses.h".
Take a look at libncurses5-dev; you'll have to do a bit of rewriting, since
Linux handles console I/O differently from DOS, but it shouldn't be too
bad.
</BLOCKQUOTE>
<BLOCKQUOTE>
Good luck!
</BLOCKQUOTE>

<!-- end 23 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/24"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 24 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Changing the "login-sequence" in Linux?</H3>


<p><strong>From Alf Kato Brandal   
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik 
<br></strong></p>

<!-- sig -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Hello! We are working on a student-project where we want to get a Perl-script
to log us in to Linux. Do anyone know how we can do this? Please do not just
answer that we need to use the "Expect" module in Perl, cause we don't
understand how we can use this. Is there any other way?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
Erm... let's run this by again. You don't understand how to do something
in a certain language which you don't know... so your solution is to
switch to another, <EM>more complex</EM> language which you don't know. Does this
make any sense to you?
</BLOCKQUOTE>
<BLOCKQUOTE>
Read the "expect" man pages and documentation. It is probably the best
tool for what you describe, since it is written for exactly this purpose.
Put in a little effort, and you'll get back the result you're looking for.
</BLOCKQUOTE>

<!-- end 24 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/25"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 25 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>Linux, X, Dell Video Card</H3>


<p><strong>From Glenn Martyna   
</strong></p> 
<p align="right"><strong>Answered By Ben Okopnik, Daniel S. Washko  
<br></strong></p>

<!-- sig -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I have loaded <A HREF="http://www.redhat.com/">Red Hat</A> Linux 7.0 on my new Dell computer,
(4MB Integrated AGP Video Card,i810 chips set)  however, X windows
will not start. I tried loading Xfree's 3.3.6 instead of higher
level of X present in the Red Hat release.  Now I get an error
that says &quot;_X11TransSocketUNIXConnext&quot;  failure.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Ben]
The Evil TransSocket Error 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle"> simply means that your X server can't connect
to your video hardware. Usually, it means that you've chosen the wrong
server. I'm not even going to speculate on the problems that you may have
created by installing XFree 3.3.6 over a 4.0 installation (at least that's
how I'm reading it), but I'm certainly not brave enough to try something
like that.
</BLOCKQUOTE>
<BLOCKQUOTE>
Also, if I remember correctly, there was either no or minimal support for
the i810 in 3.3.6, but there was in 4.0. You may want to reinstall, this
time with 4.0, and see what happens.
</BLOCKQUOTE>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Daniel]
I have a Dell Optiplex 110 running RedHat 7.0.  It has the Intel 810
chipset and required this driver:
<A HREF="http://appsr.intel.com/scripts-df/filter_results.asp?strOSs=39&strTypes=DRV%2CUTL&ProductID=178&OSFullName=Linux*&submit=Go%21"
	>http://appsr.intel.com/scripts-df/filter_results.asp?strOSs=39&amp;strTypes=DRV%2CUTL&amp;ProductID=178&amp;OSFullName=Linux*&amp;submit=Go%21</A>
</BLOCKQUOTE>
<BLOCKQUOTE>
I hope this link works for you, if not, it is the first selection if you
do a google search on "intel 810 and linux."  The driver works with
XFree 3.3.6 and up.
</BLOCKQUOTE>

<!-- end 25 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/26"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 26 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>sendmail</H3>


<p><strong>From Kenneth Moad
</strong></p> 
<p align="right"><strong>Answered By  Jonathan Markevich, Dan Wilder 
<br></strong></p>

<!-- sig -->

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I am trying to have sendmail send the contents of a file to an email
address. I want to do this from the command line though.
</STRONG></P>
<P><STRONG>
I think the command is something like [sendmail -t &lt;&lt;fff] but that does
not seem to be working correctly.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Jonathan]
I like to use a MUA rather than a MTA for this... I prefer mutt (of course).
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>mutt -s "Here's the file" -a ~/procmail.log root@localhost
</font></code></blockquote>

<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Dan]
On the other hand, sendmail is possibly more suitable for scripting
applications, such as automatic email notifications of irregularities
in logs. It offers much better portablility in scripts, and better
control over headers.  For example, using sendmail you have direct
control over "From:" headers, which can be something of a trick
with various MUAs.  If needed, you can script-generate MIME
attachments, secure in the knowledge that they won't be mangled
by an MUA that thinks it knows more about what you want than you do.
</BLOCKQUOTE>
<BLOCKQUOTE>
Most MTAs offer a "sendmail" binary with at least some command line
compatibility.  I've used Smail, Exim, and Postfix in preference to
sendmail these last ten years, and the following works just fine with
all of them.  Most likely, it also works with Qmail, MMDF, and
anything else that attempts to offer some sendmail compatibility.
</BLOCKQUOTE>
<BLOCKQUOTE>
"sendmail" may not be on your path.  Try
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>  which sendmail
</font></code></blockquote>
<BLOCKQUOTE>
and if it doesn't get you anything, use the full pathname.  The
usual location these days is <TT>/usr/sbin/sendmail.</TT>  On older systems
try <TT>/usr/lib/sendmail.</TT>  If that doesn't work, try "locate sendmail".
</BLOCKQUOTE>
<BLOCKQUOTE>
To use "sendmail -t" you put the headers in the source document,
with no intervening blank lines, then an empty line, then your
email text.  For example (drop the indents in the real thing):
</BLOCKQUOTE>

<blockquote><pre>From: me@someplace.com
To: nobody@noplace.you.know.com
Subject: email test

This is a test.  If this was a real email you would
have been asked to read it.  This is only a test
</pre></blockquote>
<BLOCKQUOTE>
The "&lt;&lt;" construct you mention is a so-called "here" document.
The above example, in the context of such, would look like
(again, delete any indents or "&gt;" quoting):
</BLOCKQUOTE>

<blockquote><pre>/usr/sbin/sendmail -t &lt;&lt;fff
From: me@someplace.com
To: nobody@noplace.you.know.com
Subject: email test

This is a test.  If this was a real email you would
have been asked to read it.  This is only a test.
fff
</pre></blockquote>
<BLOCKQUOTE>
This can be very handy for scripts, as the shell expands
shell variables that may appear inline.  So:
</BLOCKQUOTE>

<blockquote><pre>WHAT="small armadillo"
/usr/sbin/sendmail -t &lt;&lt;fff
From: me@someplace.com
To: nobody@noplace.you.know.com
Subject: email ${WHAT}

This is a ${WHAT}.  If this was a real email you would
have been asked to read it.  This is only a ${WHAT}.
fff
</pre></blockquote>
<BLOCKQUOTE>
will expand what was previously "test" as "small armadillo".
</BLOCKQUOTE>
<BLOCKQUOTE>
To use the contents of a separate file, say, a file called
"fff", use
</BLOCKQUOTE>

<blockquote><pre>/usr/sbin/sendmail -t &lt;fff
</pre></blockquote>
<BLOCKQUOTE>
note the single "&lt;".  The contents of the file need to
be the same as in the "here" document, no blank lines before
the end of the headers, headers including at least "From: ",
"To: ", and "Subject: ", then an empty line, then the
body of the email.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        >
Thank you very much for the help! I decided I will use the "&lt;" instead
of the "&lt;&lt;" in my script thanks to your email. You also gave me a couple
of other ideas too.
</strong></p>
<p><strong>
shade
</strong></p>

<!-- end 26 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/27"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 27 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>about a stubborn mount error</H3>


<p><strong>From Gabriel Florit
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<!-- sig -->

<P><STRONG>
Dear Linux Gazette,
</STRONG></P>
<P><STRONG>
I come to you hoping that I might finally solve this problem. I have
extensively searched newsletters and IRC sessions, but nothing. Most
users give up after an hour or so, telling me they have no clue. I hope
you do... 
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="middle">
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Interesting.  What suggestions have they offered that didn't work?
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I am running RH7. I have two hard drives, a 10G and a 40G. The 10G is
the master one, no partitions, and it is where i have my win98 system.
The 40G is divided into four, where I have a swap, two linux natives, and
a dos partition, as storage for the win98 system. Now, when I am in Win,
I see both C (the 10G) and D(the dos partition of the 40G). But when I am
in linux, i only see hda1, that is, the 10G drive. sfdisk -l tells me that
the dos partition in the 40G drive is hdb7, but when i try to mount it
using
</STRONG></P>

<code><strong><font color="#000033"><br>mount -t vfat /dev/hdb7 /mnt/win
</font></strong></code>
<P><STRONG>
or
</STRONG></P>

<code><strong><font color="#000033"><br>mount -t msdos /dve/hdb7 /mnt/win
</font></strong></code>
<P><STRONG>
i get an error that says
</STRONG></P>

<pre><strong>mount: wrong fs type, bad option, bad superblock on /dev/hdb7,
       or too many mounted file systems
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Let's investigate each of the three points.
</BLOCKQUOTE>
<BLOCKQUOTE>
You're certainly able to see the rest of <TT>/dev/hdbN</TT>, otherwise, your complaint
would be about Windows making Linux not work.
</BLOCKQUOTE>
<BLOCKQUOTE>
Wrong fs type:
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQuote>
There are numerous partition types usable by Windows these days.  You mention
that sfdisk -l says <TT>/dev/hdb7</TT> is your dos partition, but not which type that
it is.
</BLOCKQuote></BLOCKQUOTE>
<BLOCKQUOTE>
I've been hearing that WindowsME has slightly tweaked their partition type;
this gave both Partition Magic and parted fits. So... while in most cases
we here at the Gang wouldn't care... which flavor of Windows do you have,
and do you have any Security updates or service packs?  How did you make
the D: drive?
</BLOCKQUOTE>
<BLOCKQUOTE>
Anyways I assume that you have the msdos and vfat filesystem support
properly installed since you say that getting <TT>/dev/hda1</TT> mounted isn't a
problem.  So my first guess would be that <TT>/dev/hda1</TT> and <TT>/dev/hdb7</TT> are
different flavors of DOS partition.
</BLOCKQUOTE>
<BLOCKQUOTE>
Bad option:
</BLOCKQUOTE>
<BLOCKQUOTE><BLOCKQuote>
Your command lines looked okay to me.  Assuming <TT>/mnt/win</TT> exists.
</BLOCKQuote></BLOCKQUOTE>
<BLOCKQUOTE>
Bad superblock on <TT>/dev/hdb7:</TT>
</BLOCKQUOTE>
<BLOCKQUOTE>
<BLOCKQUOTE>
Well, I suppose there might be something subtle that really is wrong with
your D: and Linux is just being extra super duper cautious.  So perhaps
you have run the Windows disk checker with all the "yes check everything
thoroughly" options turned on.  (As opposed to their normal mode, where
they skip time consuming things like looking for bad spots on the drive.)
</BLOCKQUOTE>
</BLOCKQUOTE>
<BLOCKQUOTE>
just curious, does <TT>/dev/hdb7</TT> straddle the 1024 cylinder boundary?  I've
never heard of mount caring about that, but, it is in the middle of a huge
drive, so...
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I have created the win dir in the mnt dir. (lots of people seem to ask me
that).
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
<IMG SRC="../gx/dennis/smily.gif" ALT=":)" 
		height="24" width="20" align="bottom">   I would have created <TT>/mnt/c</TT>, <TT>/mnt/d</TT> ... but that's just because, if my
client is a serious dual booter, they continue to think of the windows parts
as "drive letters" so this is good for keeping them from getting mixed up.
(Simple enough: once it's mounted, it's a drive letter.)  So, I often use
<TT>/mnt/a</TT> for floppy access forced to vfat fs, in case I have any trouble with
a DOS floppy.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I up2dated everything but the kernel, as suggested.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
So you have the current stock kernel for RH7, which version is that?  I
think it weird that you weren't given suggestions to rebuild a kernel and
leave everything else alone - only the kernel, its modules, and mount
should have anything to do with your problem.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Still nothing. The odd thing is that I can access the hdb7 from windows.
I can even write to it. But in Linux, RH7 using <A HREF="http://www.gnome.org/">GNOME</A>, I can't.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
GNOME has nothing to do with it... or it shouldn't.  Have your tried logging
into a plain text console as root to do this?
</BLOCKQUOTE>
<BLOCKQUOTE>
If you come up in "linux single"  (by typing that at your boot: prompt) you
should be in the same state that the mounting mechanism from <TT>/etc/fstab</TT> is
in when doing its original mounts.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I have asked many different linux users. None can help me. Hope you have an
idea of what's going on.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, first we have to discover what it is, then maybe we can figure out why
it's doing it.  If it's the mount command at fault, we'll have to look in
mount's sources for its maintainer.
</BLOCKQUOTE>
<BLOCKQUOTE>
Hmm, here's an idea, if you have spare space on one of your other partitions
equal or greater than the complaining partition, you can make a binary copy
of it to a file:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>	dd if=/dev/hdb7 of=/usr/local/bigspace/D-driving-me-crazy
</font></code></blockquote>
<BLOCKQUOTE>
Yes, this will take a while.  Might want to add bs=1024 or even bs=4096 on
the end so it will grab things in chunks.  I think that should work even if
the partition image isn't an exact multiple of the blocksize... but one of
the Gang who plays more with dd than I do should comment on that.
</BLOCKQUOTE>
<BLOCKQUOTE>
Then, you can ask file if it looks like what you think it is:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>	file /usr/local/bigspace/D-driving-me-crazy
</font></code></blockquote>
<BLOCKQUOTE>
And if it agrees that it's a filesystem image, then try to loopback mount
the file:
</BLOCKQUOTE>

<blockquote><code><font color="#000033"><br>	mount -o loop -t vfat /usr/local/bigspace/D-driving-me-crazy
</font></code></blockquote>
<BLOCKQUOTE>
...which is <EM>not</EM> a good solution to your problem, but would pinpoint that
mount can, or cannot, mount this flavor of DOS partition.  If it can't,
then factors to consider are its size, and what type it really is;  having
got this far it probably wouldn't be a cylinder problem, since the image is
at a new location.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Regards,
</STRONG></P>
<P><STRONG>
Gabriel Florit
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, let us know if these thoughts shed any light on the matter!
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Dear Linux Gazette,
<br>or The Gang,
</strong></p>
<p><strong>
Thanks very much for your prompt response! I will follow your advice and let you know as soon as possible.
</strong></p>
</strong></p>
<p><strong>
Cheers,
<br>Gabriel Florit
<br>(the guy with the mount problem)
</strong></p>

<!-- end 27 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/28"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 28 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>How do I choose?</H3>


<p><strong>From Serge Wargnies
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<!-- sig -->

<!-- ::
How do I choose?
~~~~~~~~~~~~~~~~
:: -->
<P><STRONG>
Hi,
</STRONG></P>
<P><STRONG>
I may confess that I am coming from another platform and wants to migrate to
the open world ...
</STRONG></P>
<P><STRONG>
My question which distribution choosing ... I am lost between caldera,
redhat and other SuSe ...?
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Nope, not a stupid question at all.  You actually want to take a look at
the differences before picking one;  sounds wise to me.
</BLOCKQUOTE>
<BLOCKQUOTE>
We have had this question fairly recently and discussed it in some detail.
Basically, you need to know what kind of things you want to use your Linux
for, and what things various distros are aiming to be -- then, you can pick
one that is trying to head your direction, and you have a much better chance
of picking usefully.
</BLOCKQUOTE>
<BLOCKQUOTE>
The Gang expounded on this in Issue 60, "Best Linux Distro for a Newbie...?"
(<A HREF="../issue60/lg_answer60.html#tag/4"
	>http://www.linuxgazette.com/issue60/lg_answer60.html#tag/4</A>) and I hope
you'll find that answer useful too.  If these aren't quite enough, let us
know what you're thinking of, and we'll try to help out a bit more.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thanks very much in advance...
Regards
Serge Wargnies
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Welcome to the world of Linux, Serge.  I hope you'll find your first forays
here pleasant.
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Thank for you answer, I read the Gazette but I am still a bit confused so
...
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
I have cc'd back in the <A HREF="mailto:linux-questions-only@ssc.com"
	>linux-questions-only@ssc.com</A> address, so the rest of The Answer Gang
can see, and reply if they also have more comments.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I am coming from the Windows world, started as a developer on Windows 2, 3,
3.1, 9X and NT/2000. I also did a lot on the system side as well as acquired
a certain knowledge on Databases ...
</STRONG></P>
<P><STRONG>
As you mentioned, it depends what you want to do with it. I don't search for
a new graphical environment, I look for acquiring some new knowledge on a
growing system that - if I am not that wrong - is quite close to UNIX.
Because as my experience should tell that I am not too much into console
mode only, I want something with a GUI which is not available on my home PC
at this time...or I am wrong...???
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Well, you can take a look at a bunch of the screenshots over at <A HREF="http://www.libranet.com/">LibraNet</A>,
because they give a very clear sense of what the K desktop looks like.
The Gnome desktop also looks very similar.  (<A HREF="http://www.libranet.com"
	>http://www.libranet.com</A>)
</BLOCKQUOTE>
<BLOCKQUOTE>
If those are close enough to a usable GUI for you then you can probably
do okay with most of the nicer distributions, and the next concern would
be making sure that you have a decently safe bet on a clean install,
followed by an interest in good access to developer tools.
</BLOCKQUOTE>
<BLOCKQUOTE>
If that's too different from the GUI you enjoy, there's a Window Manager
named fvwm95 that's designed, as you might guess, to be a <EM>really</EM> close
match.  That means the task bar acts the same, for example.  There will
still be slight differences.
</BLOCKQUOTE>
<BLOCKQUOTE>
Once you start to get used to a few applications, you can play with loading
a few other Window Managers and see if you like some of the others;  many
have interesting extra features.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I plan to learn about the environment but once it is installed, I don't want
to spend 10 weeks - I am not often at home - to have the PC installed...This
has to be done in one shot. I will learn from the system after...as well as
starting to do some development - porting application from Windows to
LINUX/UNIX ....
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
So you probably want to try following the Willows Software Twin API
(<A HREF="http://www.willows.com"
	>http://www.willows.com</A>).  Or stick very closely to GTK+ and follow the
same style that The Gimp did, since it has a win32 version as well as
a linux one you get a successful example to study the similarities and
differences.  GTK has its own site, <A HREF="http://www.gtk.org"
	>http://www.gtk.org</A> and so does the
Gimp, <A HREF="http://www.gimp.org"
	>http://www.gimp.org</A>
</BLOCKQUOTE>
<BLOCKQUOTE>
...and if having to run an occasional, but possibly well behaved Windows
binary on your Linux is interesting, you'll also want to keep an eye on
the <A HREF="http://www.winehq.com/">WINE</A> project (<A HREF="http://www.winehq.com"
	>http://www.winehq.com</A>) which is trying to provide a
support layer for win32 binaries to be run directly within Linux and a
few other OS'.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
So what can I do, doctor?
</STRONG></P>
<P><STRONG>
Serge Wargnies
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
Okay, so we want to get you into a developer-friendly install, but not one
that expects you to be a guru during the installation itself.
</BLOCKQUOTE>
<BLOCKQUOTE>
Most packages available out there are available in at least one of 4 states:
source tarballs (you get to build it yourself; if you're lucky that's only
3 commands, not very hard, and listed in the README or INSTALL textfile for
the package), Redhat style rpm files, rpm files for non-redhat derivitive
systems (like <A HREF="http://www.suse.com/">SuSE</A> or <A HREF="http://www.turbolinux.com/">TurboLinux</A>), and <A HREF="http://www.debian.org/">Debian</A> packages.  Mandrake and some
other Redhat derivitives can share Redhat style rpms.  <A HREF="http://www.stormix.com/">Stormix</A>, <A HREF="http://linux.corel.com/">Corel</A>Linux,
Libranet, Progeny, and Debian itself all can share deb files.
</BLOCKQUOTE>
<BLOCKQUOTE>
If you decide that a debian based system works for you, then I'd definitely
choose either Libranet or Corel, instead of "the real debian" installer,
because in your case, you're not a Linux expert;  even though the Debian
installer is ok, the boost in helpfulness that these commercial distros
provide during the install would be extremely worthwhile for you.
</BLOCKQUOTE>
<BLOCKQUOTE>
If you're afraid of repartitioning but still want an fairly easy install of
a "big name" Linux distro, consider BigSlack... the version of ZipSlack
(<A HREF="http://www.slackware.com/zipslack"
	>http://www.slackware.com/zipslack</A>) that includes X and Gnome, but can be
installed directly into a FAT filesystem just using PKunzip.  <A HREF="http://www.slackware.org/">Slackware</A>
has been around a long time and is well known as being friendly for people
who like to work directly with source code.
</BLOCKQUOTE>
<BLOCKQUOTE>
If you decide that because there are lots of Redhat-style packages out there,
you need a redhat compatible system, I guess Mandrake would be worth a try.
Make sure to get a really recent version or buy it direct tho, because they
had some bugs during install that they fixed recently, and you wouldn't want
to get nailed by one just because the local store had a dusty copy.
</BLOCKQUOTE>
<BLOCKQUOTE>
I notice you're not in the U.S. so if English isn't your native language,
maybe there's a localized variant that would be handy for you.  Linux Weekly
News lists a whole bunch of them (<A HREF="http://www.lwn.net"
	>http://www.lwn.net</A>) in its Distributions
sidebar.  Some of the major distros support many languages too.
</BLOCKQUOTE>
<BLOCKQUOTE>
Backing up the system in its current state is a good idea, not so much
because of the risk (well, yes, there's some, not horridly bad) but because
now is a good time to decide what's important and not on your machine;  it
will be good to have if there's any sort of trouble, not just linux install
issues.  For example, a power outage right when you've almost got things
humming :(
</BLOCKQUOTE>
<BLOCKQUOTE>
Let us know if you need more!
</BLOCKQUOTE>

<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
        HEIGHT="28" WIDTH="50" BORDER="0"
        >

Thanks very much fir the answer, I guess you have summarized the situation
pretty well ...
</strong></p>
<p><strong>
I will follow the links...
</strong></p>
<p><strong>
Regards
<br>Serge Wargnies
</strong></p>

<!-- end 28 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/29"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 29 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>I was wondering</H3>


<p><strong>From andrew
</strong></p> 
<p align="right"><strong>Answered By Mike Orr
<br></strong></p>

<!-- sig -->

<P><STRONG>
i see a number of suspicious files in my proc directory.For
example there is a directory that is called 6 &amp; when i look in this folder i
see i number of files eg
</STRONG></P>

<pre><strong>[root@echelon 6]# ls -la
ls: exe: Permission denied
ls: root: Permission denied
ls: cwd: Permission denied
total 0
dr-xr-xr-x    3 root     root            0 Mar 26 14:28 .
dr-xr-xr-x   89 root     root            0 Mar 26 07:32 ..
-r--r--r--    1 root     root            0 Mar 26 14:29 cmdline
lrwx------    1 root     root            0 Mar 26 14:29 cwd
-r--------    1 root     root            0 Mar 26 14:29 environ
lrwx------    1 root     root            0 Mar 26 14:29 exe
dr-x------    2 root     root            0 Mar 26 14:29 fd
pr--r--r--    1 root     root            0 Mar 26 14:29 maps
-rw-------    1 root     root            0 Mar 26 14:29 mem
lrwx------    1 root     root            0 Mar 26 14:29 root
-r--r--r--    1 root     root            0 Mar 26 14:29 stat
-r--r--r--    1 root     root            0 Mar 26 14:29 statm
-r--r--r--    1 root     root            0 Mar 26 14:29 status
</strong></pre>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
This is normal.  See "man proc".
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Notice the permission denied on those 3 files. Why is this if i am root??.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
I get this error when I'm not root but not if I am root.  The three
"files" are symbolic links to other directories.  So it would depend what
the permissions of those "other" directories are.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
I
cant delete them or change anything about them. What would you suggest?? I
mean they are links to other files so why can i just unlink them.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
You shouldn't try to change or unlink them.  The directory will disappear
when process 6 dies.
</BLOCKQUOTE>
<BLOCKQUOTE>
To see for yourself that nothing funny is going on, run "umount <TT>/proc</TT>"
as root.  (If you get a "Device Busy" error, it probably means some
process has its current directory inside <TT>/proc.</TT>  You cannot unmount
a filesystem if somebody's current directory is inside it.)
The <TT>/proc</TT> directory should
be empty now.  Run "mount <TT>/proc</TT>" or "mount -t proc proc <TT>/proc</TT>" and
the "files" should reappear.
</BLOCKQUOTE>
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Also as a side note do you have any idea that when im in shell within this
directory that those 3 files are flashing??
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Mike]
That's part of the color configuration of the 'ls' command.  Usually,
flashing means it's a dead symbolic link (a link pointing to a
nonexistent file).  If it's inside <TT>/proc</TT>, I would assume the kernel
knows what it's doing and not worry about it.
</BLOCKQUOTE>

<!-- end 29 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<A NAME="tag/cardboard"><HR WIDTH="75%" ALIGN="center"></A>
<!-- begin 31 -->
<H3 align="left"><img src="../gx/dennis/qbubble.gif" 
	height="50" width="60" alt="(?) " border="0"
	>More observations of a cardboard box</H3>


<p><strong>From Randjbarnhart
</strong></p> 
<p align="right"><strong>Answered By Heather Stern
<br></strong></p>

<!-- sig -->

<!-- ::
More observations of a cardboard box
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:: -->
<P><STRONG><IMG SRC="../gx/dennis/qbub.gif" ALT="(?)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	>
Maybe the lady who asked about the cardboard box is a spanish teacher.  I
recently got a version of Don Quixote that talks about him fixing up old
knight accessories with cardboard.  Since Cervantes wrote the book in the
1500's, I was wondering when cardboard was first used.  I know that when my
students read this, someone is going to ask about it.  Maybe the word
"carton" meant a thin type of board used for crates or something. The
footnote says cardboard but it has to be wrong --don't you think so?  Anyway,
sorry for boring you.  Just wanted to express empathy for the cardboard box
lady.
</STRONG></P>
<BLOCKQUOTE><IMG SRC="../gx/dennis/bbub.gif" ALT="(!)"
	HEIGHT="28" WIDTH="50" BORDER="0"
	> [Heather]
The cardboard box is now infamous in <EM>LG</EM> (as a reference from
<a href="../issue52/tag/22.html">Issue 52</a>)
</BLOCKQUOTE>
<BLOCKQUOTE>
Our only previous mention of cardboard before that had been to describe
that chroot (while imperfect) was better than someone's attempt to keep his
users safely trapped in their home directories.
</BLOCKQUOTE>
<BLOCKQUOTE><DL><DT>
There's a more complete history of packaging than we found last time at:
<DD><A HREF="http://www.ag.ohio-state.edu/~ohioline/cd-fact/0133.html"
	>http://www.ag.ohio-state.edu/~ohioline/cd-fact/0133.html</A>
</DL></BLOCKQUOTE>

<!-- end 31 -->
<P> <hr> </p>
<!-- *** BEGIN copyright *** -->
<H5 align="center">This page edited and maintained by the Editors
        of <I>Linux Gazette</I>
<a href="http://www.linuxgazette.com/copying.html"
        >Copyright &copy;</a> 2001
<BR>Published in issue 65 of <I>Linux Gazette</I> April 2001</H5>
<H6 ALIGN="center">HTML script maintained by
        <A HREF="mailto:star@starshine.org">Heather Stern</a> of
        Starshine Technical Services,
        <A HREF="http://www.starshine.org/">http://www.starshine.org/</A>
</H6>
<!-- *** END copyright *** -->
<!-- ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: -->
</BODY></HTML>

<center>
<H1><A NAME="tips"><IMG ALIGN=MIDDLE ALT="" SRC="../gx/twocent.jpg">
More 2&cent; Tips!</A></H1> <BR>
<!-- BEGIN tips -->

Send Linux Tips and Tricks to <A HREF="mailto:gazette@ssc.com">gazette@ssc.com</A></center>
<UL>
<!-- index_text begins -->
<li><A HREF="#tips/1"
	><strong>Re: Graphics Programming for Printing / Faxing (Issue 60)</strong></a>
<li><A HREF="#tips/2"
	><strong>Alliance Pro-Motion driver</strong></a>
<li><A HREF="#tips/3"
	><strong>How to avoid launching Midnight Commander by accident</strong></a>
<li><A HREF="#tips/4"
	><strong>Gazette, I.55, Answer: Missing Root Password</strong></a>
<li><A HREF="#tips/5"
	><strong>SNMP Tool for networking (re: March tips)</strong></a>
<li><A HREF="#tips/6"
	><strong>distro version upgrade? (slackware)</strong></a>
<li><A HREF="#tips/7"
	><strong>2-cent Tip: Cleaning up after Netscape</strong></a>
<li><A HREF="#tips/8"
	><strong>Regarding backups [issue64 tag/28.html]</strong></a>
<li><A HREF="#tips/9"
	><strong>Modules cannot load with kernel recompile</strong></a>
<li><A HREF="#tips/10"
	><strong>RE: Linux PPP route question</strong></a>
<li><A HREF="#tips/11"
	><strong>"Interrupt for Linux" question from S. Auejai</strong></a>
<li><A HREF="#tips/12"
	><strong>2ct tip - Removing temp files</strong></a>
<li><A HREF="#tips/14"
	><strong>Linux RedHat question</strong></a>
<li><A HREF="#tips/16"
	><strong>Question on stty</strong></a>
<li><A HREF="#tips/18"
	><strong>inode related question</strong></a>

<li><A HREF="#tips/20"
	></a>HELLO --or--
<br><A HREF="#tips/20"
	><strong>Protecting web pages</strong></a>

<li><A HREF="#tips/21"
	><strong>SSH article</strong></a>
<li><A HREF="#tips/23"
	><strong>Linux commands</strong></a>
<li><A HREF="#tips/24"
	><strong>How write a selfextracting sh script ?</strong></a>
<li><A HREF="#tips/27"
	><strong>Searching for a text revisioning tool</strong></a>
<li><A HREF="#tips/29"
	><strong>2.4.2 and loop devices</strong></a>
<li><A HREF="#tips/30"
	><strong>Re your Fortran answer (tag 15, iss 64)</strong></a>
<li><A HREF="#tips/31"
	><strong>Agenda Computing Challenges Palm</strong></a>
<li><A HREF="#tips/32"
	><strong>Mailbag #62; Memory mystery</strong></a>
<li><A HREF="#tips/34"
	><strong>mcad</strong></a>
<!-- index_text ends -->
</UL>
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/1"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Re: Graphics Programming for Printing / Faxing (Issue 60)</FONT></H3>
Tue, 27 Feb 2001 16:32:59 -0800
<BR>Anthony Greene <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20%5Bagreene%40pobox.com%3A%2520Re%3A%2520Graphics%2520Programming%2520for%2520Printing%2520%5C%2520Faxing%2520(Issue%252060)%5D">(The Answer Gang)</a>


<p>Re: Graphics Programming for Printing <TT>/</TT> Faxing 
(<a href="../issue60/lg_answer60.html#tag/11">Issue 60</a>)</p>
<P>
The quick and easy way for a Perl programmer to do convert data to faxable
invoices/reports is to output the data as HTML, convert it to Postscript
using html2ps &lt;<A HREF="http://www.tdb.uu.se/~jan/html2ps.html"
	>http://www.tdb.uu.se/~jan/html2ps.html</A>&gt;, then fax the
result using <tt>efax</tt> or <tt>mgetty+sendfax</tt>.
</P>

<!-- end 1 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/2"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Alliance Pro-Motion driver</FONT></H3>
Sun, 25 May 1997 23:23:08 -0400
<BR>Ralph E Bugg<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20"> (buggr from sssnet.com )</a>

<P>
There was a letter to you from an unidentified person looking for a
driver for an Alliance Pro-Motion video card.
</P>
<strong>
<P>
"....Anyways, I would just run Linux but my problem is that Xwindows
doesn't have advanced support for my video card, so the best I can get
is
640x480x16colors and I just can't deal with that. Maybe I'm spoiled. The
guy I wrote on the Xwin development team told me that they were working
on better support for my card, though. (Aliance Pro-Motion). ...."
</P>
</strong>
<P>
If he goes to <A HREF="http://www.alsc.com"
	>http://www.alsc.com</A> and follows the path to tech support,
he will find a SVGA driver (no source code though) for X-windows.  I am
using an NEC Ready 9618 system which uses one of the Alliance chips on
the mother board.  It took a LOT of fiddling with the configuration file
but it will work at higher resolutions @ 256 colors.
</P>
<P>
Hope you can pass this on to him.
</P>
<P>
Thanks,  Ralph Bugg.
</P>

<!-- end 2 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/3"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">How to avoid launching Midnight Commander by accident</FONT></H3>
Mon, 26 Feb 2001 10:31:51 -0500
<BR>Allan Peda<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20How%2520to%2520avoid%2520launching%2520Midnight%2520Commander%2520by%2520accident"> (apeda from linkshare.com)</a>


<!-- sig -->

<P>
I've typed "mc foo bar" one time too many when I really meant to type
"mv foo bar".  Removing Midnight commander
is not an option, because that breaks some file exploror type GUI
utilities, so I cooked up a bash script to double confirn that
I wanted to type what I (probably mis-)typed :
</P>
<p align="center">See attached script 
	<a href="misc/tips/mc.bash.txt">mc.bash.txt</a></p>

<!-- end 3 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/4"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Gazette, I.55, Answer: Missing Root Password</FONT></H3>
Wed, 28 Feb 2001 17:23:18 +0100 (CET)
<BR>Johannes Kaiser<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Gazette,%2520I.55,%2520Answer%3A%2520Missing%2520Root%2520Password"> (uehj from rz.uni-karlsruhe.de)</a>


<!-- sig -->

<P>
It should be easy to get in if you use LILO. At the boot prompt, type in
the name of your boot image (you can find that out by hitting the
"tab" key twice), followed by the word single. For a normal redhat
installation, typing "linux single" should do. You also can append
"init=/bin/sh" instead of "single", that leaves remounting your root
filesystem rw to you.
</P>

<!-- end 4 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/5"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">SNMP Tool for networking (re: March tips)</FONT></H3>
Thu, 1 Mar 2001 17:02:10 +0100
<BR>Casas Bouza, Robert<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%202%2520Tip%2520answer"> (robert.casas from puig.es)</a>


<!-- sig -->

<P>
Hi!
</P>
<P>
About the question done by Antonio Sidona (<a href="../issue64/lg_tips64.html#tips/7">looking for a SNMP tool for networking</a>, tips) on you March 2001 
issue, we have tried netsaint
(<a href="http://www.netsaint.org/">www.netsaint.org</a>). 
It's a great tool, although needs to be configured
properly, but you can monitor any system that support SNMP or not. A funny
thing is that we HAVE HP OpenView installed, but you need license per
console, and NetSaint can be installed on a Web Server and accessed through
a browser. We actually used them on a complementary basis.
</P>
<P>
Robert Casas
</P>

<!-- sig -->


<!-- end 5 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/6"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">distro version upgrade? (slackware)</FONT></H3>
Thu, 01 Mar 2001 13:45:43 -0800
<BR>Michael Moore<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20distro%2520version%2520upgrade%3F%2520(slackware)"> (michael_moore from csnw.com )</a>


<!-- sig -->

<P>
Dan Blazek wrote:
</P>
<P><STRONG>
Hi,
I think I'm running <A HREF="http://www.slackware.org/">Slackware</A> 2.2 (kernel is 2.0.27 for sure
anyway). Is there some kind of cluster or patch bundle I can download to
upgrade my box. Like a single package I can install to at least jump up
to slackware 3? And if there is.. can you please tell me where to find
it, and if there is there a special way to install it? Or am I going to
be stuck installing a new image?
</STRONG></P>
<P>
Heather wrote:
</P>
<BLOCKQuote>
<P><STRONG>
I thought there wasn't one, but rarely say so without looking. And what
do you know, I found:
</STRONG></P>
<P><STRONG><DL><DT>
slackUp - The Slackware Auto-Upgrade Utility
<DD><A HREF="http://xfactor.itec.yorku.ca/~xconsole/download.html"
	>http://xfactor.itec.yorku.ca/~xconsole/download.html</A>
</DL></STRONG></P>
<P><STRONG>
You should read its readme yourself, to check that it can handle your
version. If it can't. get involved with the authors ... they haven't
updated it in almost a year (or at least the webpage) and you may spark
an entirely new round of development for the project.
</STRONG></P>
</BLOCKQuote>
<P>
David Cantrell, one of the Slackware staff members, has also made a pretty
comprehensive Slackware upgrade utility, autoslack.  While this is not a
supported Slackware project, David's involvement probably means it is
likely to work well with their site.  You can find it on their unsupported
projects server at <A HREF="http://zuul.slackware.com"
	>http://zuul.slackware.com</A>
</P>
<P>
-Michael
</P>

<!-- end 6 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/7"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">2-cent Tip: Cleaning up after Netscape</FONT></H3>
Thu, 1 Mar 2001 17:51:39 -0500
<BR>Ben Okopnik <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%202-cent%2520Tip%3A%2520Cleaning%2520up%2520after%2520Netscape">(The Answer Gang)</a>

<P>
Linux is a wonderfully reliable OS: even the software that runs under it
is reliable. X Windows runs reliably. Midnight Commander is reliable.
Even Netscape Communicator crashes reliably.
</P>
<P>
Ooops...
</P>
<P>
Netscape is a nice piece of software, in that it supports everything (and
then some) that a modern "fancy" browser should support. Unfortunately,
the rate at which it goes down brings to mind expressions about hookers on
payday - and in my experience, it's been this way from day one. Not only
that, it tends to leave behind hung copies of itself (which makes the
processor load shoot right up into the red) and lockfiles that create error
messages the next time you try to start it up.
</P>
<P>
A few months ago, tired of having to clean up the random garbage, I created
this script. If Netscape has crashed, or is simply frozen, it will take
care of everything. Nowadays, it's my automatic response to a Netscape
crash. &lt;sigh&gt; I'm getting awfully familiar with typing "notscape"...
</P>

<P align="center">See attached script <a href="misc/tips/notscape.bash.txt">notscape.bash.txt</a></p>


<!-- end 7 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/8"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Regarding backups [http://www.linuxgazette.com/issue64/tag/28.html]</FONT></H3>
Thu, 1 Mar 2001 19:02:35 -0500
<BR>David Jao<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Regarding%2520backups%2520%5Bhttp%3A%5C%5Cwww.linuxgazette.com%5Cissue64%5Ctag%5C28.html%5D"> (scythe from dominia.org)</a>


<!-- sig -->

<P>
Hi guys,
</P>
<P>
This is in response to Bruce Harada's message at
</P>
<P>
<A HREF="../issue64/tag/28.html"
	>http://www.linuxgazette.com/issue64/tag/28.html</A>
</P>
<P>
I would have preferred to contact him directly but I could not find an
email address for him on the page.
</P>
<P>
Using gzip on backup files 2GB in size is a <EM>really</EM> bad idea, since if
the compressed file gets corrupted at any point, then everything
occuring after the point of corruption will be unrecoverable.
</P>
<P>
Of course if hard drives are perfectly reliable then corruption is no
problem, but if that were the case then you wouldn't be doing backups
anyway.
</P>
<P>
In general, compressing large backups is almost never worth it because
of the reliability issues. If one must use compression, bzip2 is a
better choice, since it uses 900kB blocks and corruption would only
affect an individual data block.
</P>
<P>
-David
</P>

<!-- sig -->


<!-- end 8 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/9"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Modules cannot load with kernel recompile</FONT></H3>
Thu, 01 Mar 2001 22:39:33 -0500
<BR>Tom Walsh<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Modules%2520cannot%2520load%2520with%2520kernel%2520recompile"> (tom from cyberiansoftware.com)</a>


<!-- sig -->

<P>
Regarding '<A HREF="../issue64/tag/16.html"
	>http://www.linuxgazette.com/issue64/tag/16.html</A>',  I use
'make install' myself, saves you the step of copying the image to <TT>/boot</TT>
and forgetting to run lilo.
</P>
<P>
--
Tom Walsh 
</P>

<!-- sig -->


<!-- end 9 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/10"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">RE: Linux PPP route question</FONT></H3>
Fri, 02 Mar 2001 14:06:07 -0600
<BR>Brian Finn <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20RE%3A%2520Linux%2520PPP%2520route%2520question"> (nacmsw from airmail.net)</a>


<!-- sig -->

<P>
Hi,
</P>
<P>
I found a dial-on-demand package for Linux called Diald. I think
it may help alleviate your PPP problems. You can find it at:
</P>
<P><BLOCKQuote>
<A HREF="http://diald.sourceforge.net"
	>http://diald.sourceforge.net</A>
</BLOCKQuote></P>
<P>
Hope this helps!
<br>Brian Finn
</P>

<!-- sig -->


<!-- end 10 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/11"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">"Interrupt for Linux" question from S. Auejai</FONT></H3>
Mon, 05 Mar 2001 12:09:38 -0600
<BR>Bill McConnaughey<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20%2522Interrupt%2520for%2520Linux%2522%2520question%2520from%2520S.%2520Auejai"> (mcconnau from biochem.wustl.edu)</a>


<!-- sig -->

<P>
I found Alessandro Rubini's book, Linux Device Drivers, published by
O'Reilly and Associates, very helpful in getting started on writing device
drivers (including interrupt handlers).
</P>

<!-- sig -->


<!-- end 11 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/12"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">2ct tip - Removing temp files</FONT></H3>
Tue, 06 Mar 2001 20:58:25 -0800
<BR>forsberg <a href="mailto:linux-questions-only@ssc.com&Subject=%20Re%3A%20%5BLG%2065%5D%202ct%2520tip%2520-%2520Removing%2520temp%2520files"> (forsberg from adnc.com)</a>


<!-- sig -->

<P>
When writing a program that uses temporary files on a
UNIX/Linux system it is convenient to use a feature of
UNIX. Create the temporary file, then remove it (i.e.
<TT>unlink()</TT> ) without closing the file.
</P>
<pre>
fd = fopen("/tmp/somefilename.tmp",...);
unlink("/tmp/somefilename.tmp");
.
. Use temp file
.
fclose(fd) or exit();
</pre>
<P>
Then you can read and write to this file during
the existance of this process. The temp file will
not be removed until a close statement on the
file descriptor or the program terminates. Only
then will the kernel remove the file. Use this
technique to guarantee that all temp files are
cleaned up if your program crashes.
</P>
<P>
Bruce Forsberg
</P>

<!-- sig -->


<!-- end 12 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/14"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Linux RedHat question</FONT></H3>
Wed, 28 Mar 2001 12:18:34 -0800
<BR>Ray Hanes <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Linux%2520RedHat%2520question">(high_tech_hanes from yahoo.com)</a>


<P><STRONG>
I saw your page and don't know if your still  actively
maintaining it and answering questions but in case you  are.     I'm
trying to find a variable for what version of  <A HREF="http://www.redhat.com/">Red Hat</A> is running.
If there is no variable for it from the system then  how can I get a
script file to detect the Distribution Verson assign it to a
variable?
</STRONG></P>
<P>
Hi Ray --
</P>
<P>
On a default RedHat install, the file <TT>/etc/redhat-release</TT>
contains the version. Most RedHat installs leave that
file there. (I always delete it because the existence of that
file causes the rc.local script to overwrite <TT>/etc/issue</TT> at bootup.)
</P>
<P>
Hope this helps -
<br>--
<br>Breen Mullins
</P>

<!-- sig -->


<!-- end 14 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/16"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Question on stty</FONT></H3>
Tue, 20 Mar 2001 11:46:06 -0800
<BR>Iris Louie <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Question%2520on%2520stty">(IHo from altera.com)</a>


<P><STRONG>
I have to type in stty erase "backspace" each time I log in.  How can I get
set it as apart of the default stty setting?
</STRONG></P>
<P>
Put the command in your ~/.bashrc file or whatever file your shell
reads at startup. -- Mike
</P>

<!-- end 16 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/18"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">inode related question</FONT></H3>
Fri, 16 Mar 2001 09:57:20 -0800
<BR>HCL Amritsar <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20inode%2520related%2520question"> (narenderpk from usa.net, tag from ssc.com)</a>


<P><STRONG>
in unix file system if inode of current directory is known .explain how to
find the inode of the file <TT>../file1.</TT>
</STRONG></P>

<pre>$ ls -i ../joey/.bashrc
 407098 ../joey/.bashrc
</pre>
<p>-- Mike</p>


<!-- end 18 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/20"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Protecting web pages</FONT></H3>
Mon, 26 Mar 2001 10:05:00 -0800
<BR>Doranda L Martin <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Protecting%2520web%2520pages">(anonymous)</a>


<!-- sig -->

<!-- ::
Protecting web pages
~~~~~~~~~~~~~~~~~~~~
:: -->
<P><STRONG>
Hello,
<br>my name is D and i have a web page.  I have a question.  I would like to
know how to put an entry box in my web page.  Actually i am trying to hav
it so that you must have a password to get to certain parts of my web
page, basically the table where my poems are and then have a way to make
them have to enter a password to look at the poems if someone
accidentally got to the table.  i would like:
</STRONG></P>
<P><STRONG><BLOCKQuote>
box 1:  their email address
<br>box 2: password
<br>submit
</BLOCKQuote></STRONG></P>
<P><STRONG>
please help, if you could send me codes or somewhere to go or anything it
would be a great help
</STRONG></P>
<P>
If your web server is <A HREF="http://www.apache.org/">Apache</A> and it has been configured to support (1)
HTTP Basic Authentication, and (2) .htaccess files, do the following:
</P>
<ol>
<li> Use the htpasswd
program to create a password file.  (This is not the UNIX
password file; for security, you should use different passwords than
your login passwords.)

<li> Create a file called .htaccess in the highest-level directory you
wish to protect.  The file should contain:
</ol>

<blockquote><pre>AuthName "Poems"
AuthType Basic
AuthUserFile /path/to/htpasswd/file
require valid-user
</pre></blockquote>
<P>
Now, when the user tries to access anything in or under that directory,
the browser will prompt her to type her "Poems" username/password.  If
she does not type it correctly, she'll get an "Unauthorized" error.
</P>
<P>
Your Apache configuration file must "AllowOverride AuthConfig" for
either the entire site or the portion of the site you're concerned
about.
</P>
<P>
See the Apache documentation:
<A HREF="http://httpd.apache.org/docs/mod/mod_auth.html"
	>http://httpd.apache.org/docs/mod/mod_auth.html</A> and
<A HREF="http://httpd.apache.org/docs/mod/core.html#allowoverride"
	>http://httpd.apache.org/docs/mod/core.html#allowoverride</A>
<br>-- Mike
</P>

<!-- end 20 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/21"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">SSH article</FONT></H3>
Tue, 6 Mar 2001 14:55:51 -0800
<BR>Bryan Henderson <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20SSH%2520article">(bryanh from giraffe-data.com)</a>

<P>
In the article on <a href="../issue64/dellomodarme.html">ssh, scp, and sftp</a>
in the March issue, there is an
important area that isn't covered: client/server compatibility.
</P>
<P>
If you're just doing a basic ssh (to get a remote shell), you're
using a standard SSH protocol and any program named "ssh" is likely
to work with any remote system that offers a service it calls "ssh."
</P>
<P>
But scp and sftp are not standard protocols.  If you run the scp
program from openssh against a remote system that's running an
original ssh server, it will not work.  (And when I learned this the
hard way, it was very hard indeed:  the error message isn't "this
server doesn't implement this scp protocol."  It is, for reasons
that took a day of debugging to figure out, "invalid file descriptor"!
</P>
<P>
--
Bryan Henderson     
</P>

<p><em>This was also forwarded along to the author of that article for
comment, but we got no reply by press time. -- Heather</em></p>

<!-- end 21 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/23"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Linux commands</FONT></H3>
Wed, 14 Mar 2001 09:13:48 -0500
<BR>katja.andren <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Linux%2520commands"> (katja.andren from spray.se)</a>


<!-- sig -->

<P><STRONG>
Hi!
</STRONG></P>
<P><STRONG>
I'm new Linuxuser (Redhat ver.) and I'm loking for a summery of commands,
"Linux version of DOS-commands". Do you have any good tips on where I can
find it?
</STRONG></P>
<P>
As it happens, such a thing exists. The summary, as well as a lot of other
useful tips for those who are used to DOS or Windows, are all included in
the DOS-Win-to-Linux-HOWTO. Take a look at "<TT>/usr/doc/HOWTO</TT>" (if you have
them installed on your system - if you don't, you should!), or
&lt;<A HREF="http://www.linuxdoc.org&gt"
	>http://www.linuxdoc.org&gt</A>; for the latest version. -- Ben
</P>

<!-- end 23 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/24"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">How write a selfextracting sh script ?</FONT></H3>
Thu, 15 Mar 2001 07:35:49 +0100
<BR>Josep Torra Valles<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20How%2520write%2520a%2520selfextracting%2520sh%2520script%2520%3F"> (jtorra from campus.uoc.es)</a>


<P><STRONG>
I would like to know how write a selfextracting sh script with a
tar.gz(source code of my program) to be installed, and after
it's extraction I need run make in order to compile and finish
the installation.
</STRONG></P>
<P><STRONG>
Thanks in advance
</STRONG></P>
<P>
Strange as this may sound, about a year ago, I wrote a shell script that
does exactly that - including automatically running "make" or another
program to process the files. I even packaged it as a tarball, with
documentation, configuration files, and even a man page... but I never
released it. Why? &lt;shrug&gt; There are a lot of tangled issues, including the
fact that this mechanism can be easily misused for malicious purposes. On
the other hand, so can anything that you download off the Web and execute
without checking it out first. Whatever, your e-mail here has spurred me
to go ahead and make it public: you can download "SFX" from my site, as
&lt;<A HREF="http://www.geocities.com/ben-fuzzybear/sfx-0.9.4.tgz&gt"
	>http://www.geocities.com/ben-fuzzybear/sfx-0.9.4.tgz&gt</A>;. If you run it
without any options, it'll tell you how to create files that will
self-extract <EM>and</EM> compile, all in one shot. I also took some trouble with
the documentation; the "method" files are a pretty cool way to specify
action after extraction, and you can always create your own.
</P>
<P>
I'd really appreciate feedback from anyone who ends up using SFX; if
there's enough interest, I'll rewrite it, possibly in C or Perl.
</P>

<!-- sig -->


<!-- sig -->


<!-- end 25 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/27"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Searching for a text revisioning tool</FONT></H3>
Sun, Mar 11, 2001 04:48:16PM +0100
<BR>Peter Paluch <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Searching%2520for%2520a%2520text%2520revisioning%2520tool"> (peterp from frcatel.fri.utc.sk)</a>


<!-- sig -->
<P><STRONG>
Hi,
<br>=-=
</STRONG></P>
<P><STRONG>
I often do revisions and checks of articles and text documents that my
colleagues wrote, and under Linux I miss the ability of MS Word97 and above
which allowed me to do revisions very conveniently. Under "revisioning" I
understand writing several marks and suggestions for the author to the
revisioned document, striking-out whole words or sentences and replacing
them with new ones.
</STRONG></P>
<P><STRONG>
I'm thus searching for a Linux document revisioning tool. It would be lovely
if the tool worked with XML. Do you know anything that could help me?
(Please notice that CVS is not what I need.)
</STRONG></P>
<P><STRONG>
Thanks a lot in forward.
</STRONG></P>
<P>
Have you taken a look at WordPerfect 8 for Linux? I don't have it
installed on my current machine, but I seem to remember seeing some kind
of revision-type stuff in the menus. -- Ben
</P>


<!-- end 27 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/29"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">2.4.2 and loop devices</FONT></H3>
Tue, 13 Mar 2001 22:42:11 -0800
<BR>David Ellement <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%202.4.2%2520and%2520loop%2520devices">(david.ellement from home.com)</a>


<!-- sig -->

<P><STRONG>
I've recently compiled the 2.4.2 kernel (under RH 7.0).  It seems I
can no longer run any commands the interact with the block loop
devices: mkbootdisk, mkinitrd, mke2fs <TT>/dev/loop*</TT>, mount -o, ...
If I run one of them, they hang at mke2fs <TT>/dev/loop;</TT> if I try to
halt the system afterward, it hangs trying to shutdown the
file-systems.
</STRONG></P>
<P><STRONG>
I've tried to compile with loop device support as a built-in, and as
a module (and lsmod show it loaded).  What am I missing?
</STRONG></P>
<p><em>... but he managed to discover for himself ...</em></p>
<P>
The 2.4.2 kernel has a bug which caused a deadlock for loop devices.
It is fixed in the 2.4.3-pre2 and later patches.
</P>
<p><em>Thanks for passing us the Tip, David! -- Heather</em></p>

<!-- sig -->


<!-- end 29 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/30"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Re your Fortran answer (tag 15, iss 64)</FONT></H3>
Tue, 13 Mar 2001 17:04:00 +0000 (GMT)
<BR>duncan <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Re%2520your%2520Fortran%2520answer%2520(tag%252015,%2520iss%252064)"> (D.C.Martin.2000 from Cranfield.ac.uk)</a>


<!-- sig -->

<P>
I read with interest about how g77 works. I plan on using it when I get a
chance.  The questioner would probably find it useful to check out
www.fortran.com - it has links to many different Fortran products,
services, and benchmark tests, and a lot of what is on there is relevant
to/directly aimed at linux users.  Many compilers seem to be aimed
squarely at the linux market.  I guess that is because of the popularity
of Beowulf type clusters, but it's nice to know that even where almost
everything is (visibly) written in C there is still room for Fortran.
<br>Hope this helps
<br>Cheers
<br>Duncan
<br>ps TAG is great.  Keep it up.
</P>

<!-- sig -->


<!-- end 30 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/31"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Agenda Computing Challenges Palm</FONT></H3>
Thu, 15 Mar 2001 16:34:11 -0800 (PST)
<BR>Heather<a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20%5Bangela%40agendacomputing.com%3A%2520Agenda%2520Computing%2520Challenges%2520Palm%5D"> (The Editor Gal)</a>

<P><STRONG>
Is this press release true?  Can somebody summarize how far the Linux-on-PDAs
projects have gotten?
</STRONG></P>
<P>
Handhelds.org has a great deal of information about putting Linux onto PDAs.
Transvirtual's PocketLinux (their penguin is very cute - his whole tummy is
a pocket protector) runs on iPaq, VTech's Helio, maybe others by now.  The
pocketlinux has to be put on by having a dev environment on another box, but
this is no different than the first fellow who forcefed Linux onto his laptop
across its plink cable or ethernet crossover.  The result is operational
without an external bootstrap, but varies in usability.
</P>
<P>
Certainly some complete OS bigots have tried to put Linux on their Palms.
</P>
<P>
Agenda <EM>may</EM> be the first to actually sell a PDA preloaded with Linux, and
not designed for some other OS first, though.
</P>
<P>
And, their Linux environment has the usual PDA features, rather than trying
to be X or a terminal. -- Heather
</P>
<P><STRONG>
----- Forwarded message from Agenda Computing -----
</STRONG></P>
<P><STRONG>
Subject: Agenda Computing Challenges Palm
Date: Fri,  9 Mar 2001 19:49:05 -0800 (PST)
</STRONG></P>
<p>The complete text of their Press Release can be found at
	<a href="http://www.agendacomputing.com/about/press20010309.html"
		>http://www.agendacomputing.com/about/press20010309.html</a>
</p>

<!-- sig -->

<!-- end 31 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/32"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">Mailbag #62; Memory mystery</FONT></H3>
Tue, 20 Mar 2001 12:35:52 +0100
<BR>Frode Lillerud <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20Mailbag%2520%252362%3B%2520Memory%2520mystery">(frode.lillerud from c2i.net)</a>


<!-- sig -->

<P>
I know that Abit had a similar problem with their BH6 motherboard, Linux
wouldn't show RAM over 64MB.
They solved it by releasing a BIOS patch.
</P>
<P>
Yours sincerely
<br>Frode Lillerud, Norway
</P>

<!-- sig -->


<!-- end 32 -->
<!--     .~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.~~.     -->
<P> <A NAME="tips/34"><HR WIDTH="75%" ALIGN="center"></A> <P>
<H3><IMG ALIGN=BOTTOM ALT="" SRC="../gx/lil2cent.gif">
<FONT COLOR="navy">mcad</FONT></H3>
Mon, 26 Mar 2001 09:29:15 -0800 (PST)
<BR>Heather <a href="mailto:linux-questions-only@ssc.com&subject=%20Re%3A%20%5BLG%2065%5D%20mcad"> (The Editor Gal)</a>


<!-- sig -->

<P><STRONG>
Hello.  I keep seeing the term "mechanical CAD", but am not sure of it's
actual meaning.  What is mechanical CAD and what differentiates it from CAD?
Thanks.
RES
</STRONG></P>
<P>
This isn't really a question about Linux, but I'll toss in a potshot.
</P>
<P>
There are absolutely piles of CAD software available for Linux.  Most of
it appears to be for cirvuitboard description.  That's not terribly useful
for developing instructions to send to a metal lathe so a part can be cut.
And both of these are very different from architectural CAD for designing
building layouts.
</P>
<P>
I would guess that by saying "mechanical CAD" one could easily note that
you meant the second kind.
</P>

<!-- sig -->


<!-- end 34 -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Opera - a lightweight browser for Linux</font></H1>
<H4>By <a href="mailto:matthiasarndt@gmx.net">Matthias Arndt</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<H2>Table Of Contents</H2>
<UL>
<LI><A HREF="#intro">Introduction</A></LI>
<LI><A HREF="#layout">Screenlayout &amp; Look'n'Feel</A></LI>
<LI><A HREF="#bookmarks">Bookmark handling</A></LI>
<LI><A HREF="#browsing">Quality of browsing - what can Opera render?</A></LI>
<LI><A HREF="#compare">Opera in comparison to Netscape</A></LI>
<LI><A HREF="#custom">Customization</A></LI>
<LI><A HREF="#technical">Technical Notes &amp;  Downloading</A></LI>
<LI><A HREF="#install">Installation</A></LI>
<LI><A HREF="#conclusion">Conclusion</A></LI>
</UL>

<A NAME="intro">
<H2>Introduction</H2>
<P>
Currently, the Linux community lacks a stable and fast web browser.
</P>
<P>
Of course, there's Netscape but it's neither fast nor 100% stable. Netscape will crash sometimes,
especially when you're downloading large files over a slow internet link.
(I will refer to Mozilla as Netscape here because it still is a "Netscape"-like browser.)
</P>
<P>
There are several alternatives out there but they all lack features that are required 
by most (multimedia) webpages like Java, Javascript, Frame support, tables, CSS and even Flash.
My personal opinion is that most of this is just trash, not really needed for a decent 
website except tables.
</P>
<P>
Ported from the Windows world, <A HREF="http://www.opera.com/" TARGET="_blank">Opera</A> 
seems to fill the gap.  It's still not a full replacement for Netscape on the Linux 
platform, but it's very close to reaching this goal.
</P>
<P>
This article focuses on the advantages and disadvantages of Opera, its
concepts, and finally a comparison to Netscape.
</P>

<A NAME="layout">
<H2>Screenlayout &amp; Look'n'Feel</H2>
<P>
A picture is always a good starting point. Click on the link below to see
a screenshot of Opera.
</P>
<P>
<A HREF="misc/arndt/mainframe.jpg">Opera - as seen when started, browsing Slashdot</A> &nbsp;[237 KB]
</P>
<P>
If you're used to the Windows version of Opera, you'll recognize that the screen layout is 
the same as in the Windows version.
</P>
<P>
The first thing that you'll recognize is a somewhat large banner containing advertisements.
Opera is a commercial product so the try-before-you-buy version has a banner there. 
This banner can be controlled to show specific advertisements. But I do not recommend that 
because you'll lose some of your limited anonymity on the net.
</P>
<P>
The navigation buttons are familiar and most of them work the same way as in Netscape. 
However you'll notice the lack of a <EM>STOP</EM> button to cancel a transfer.
</P>
<P>
Opera uses some sort of multiple document interface. Contrary to Netscape, all document windows 
will be opened in the mainframe as subframes. 
You can choose between either a full view of a single document or you can see several subframes 
be open at the same time.
</P>
<P>
Just take a look at the two examples below....
</P>
<P>
<A HREF="misc/arndt/2frames.jpg">
Opera, two documents open at the same time, one is shown fullscreen</A> &nbsp; [210 KB]
</P>
<P>
<A HREF="misc/arndt/2frames-mdi.jpg">
Opera, two documents open at the same time, both shown in seperate subframes</A> &nbsp; [196 KB]
</P>
<P>
You'll notice that you have seperate buttons for each open document. This allows you to 
operate Opera even if theres no windowmanager running. Additional: frame switching is much 
faster compared to Netscape.
</P>
<P>
A very nice feature, borrowed from the MS IE, is the ability to switch to a real fullscreen mode.
Press <B>F11</B> and the current browser frame will be shown fullscreen, really filling the 
full screen, not only the subframe of the Opera window.
</P>
<P>
Disadvantages in this mode are that you can neither switch to other subframes nor use the 
forward and backward buttons.
</P>
<P>
You can customize the look'n'feel of Opera to the extend. Much more than Netscape.
Display of documents is controlled by CSS. You may either use one from a web site on the net 
or supply your own.
You can select a whole file to do this or you can customize color, font and size of the various
objects as headers, paragraphs, etc.
</P>
<P>
Take a look at the customization dialog below:
</P>
<P>
<A HREF="misc/arndt/preferences-css.jpg">
customization dialog for applying personal CSS</A> &nbsp; [49 KB]
<P>
You cannot customize that all in Netscape. Pretty cool feature.
</P>

<A NAME="bookmarks">
<H2>Bookmark Handling</H2>
<P>
Bookmark handling is very good in Opera. You can import your Netscape bookmarks and 
your KDE shotcuts (if you have some).
A feature, imported from the Windows version, is the possibility to import MS IE 
bookmarks as well. But I guess, no real Linux user has a need for that, IMHO.
One drawback is that the import is readonly. That's somewhat limiting but acceptable.
</P>
<P>
You can switch between a view with bookmarks or without. The default layout is close to the MS IE ones.
</P>
<P>
<A HREF="misc/arndt/bookmarks.jpg">
Opera, with imported Netscape bookmarks open</A> &nbsp; [246 KB]
</P>
<P>
Note that import of foreign bookmarks is done automatically.
</P>

<A NAME="browsing">
<H2>Quality of browsing - what can Opera render?</H2>
<P>
Well, Opera renders almost any decent HTML code. Tables, Frames, CSS, all are no problem. 
In that way it has the full quality of Netscape, not found in many alternative browsers 
on the Linux platform.
</P>
<P>
In particular, the CSS support is even better than in Netscape.
Just compare a site relying completely on CSS and optimized for MS IE. View this page 
with the MS IE, with Netscape and finally with Opera.<BR>
As far as I have noticed so far, Opera's output is closer to the MS IE than Netscape. But 
that may be subjective.
</P>
<P>
Opera has three or four features that it does not do properly:
<UL>
<LI>Java Support - not implemented yet</LI>
<LI>Javascript - well works for most things but not for all</LI>
<LI>Flash - no support for this</LI>
<LI>no Netscape plugins</LI>
</UL>
But IMHO no one really needs Javascript, this one is really annoying. Flash is for multimedia 
freaks and Java, well if you need this, you can still use Netscape.
</P>
<P>
Opera is able to display PNG pictures, a feature not supported by most alternative browsers.
</P>

<A NAME="compare">
<H2>Opera in comparison to Netscape</H2>
<P>
Take a look at the following table, then select for yourself.
</P>
<P>
<SMALL>I haven't used the Konqueror from KDE2 yet - so I cannot give you an comparative overview of that one with Opera.</SMALL>
</P>
<TABLE WIDTH="100%">
<TR BGCOLOR="#FFFFEE">
<TD>
</TD>
<TD>
Opera
</TD>
<TD>
Netscape
</TD>
</TR>

<TR>
<TD WIDTH="28%">
Cost
</TD>
<TD>
free - but advertisement is shown
</TD>
<TD>
free - for non-commercial use
</TD>
</TR>

<TR>
<TD>
Size
</TD>
<TD>
average (statically linked)<BR>
small (dynamic linked)
</TD>
<TD>
big (statically linked only)
</TD>
</TR>

<TR>
<TD>
Speed
</TD>
<TD>
startup is fast<BR>
document loading and rendering is fast
</TD>
<TD>
startup is very slow, even with much RAM<BR>
document loading and rendering is average
</TR>

<TR>
<TD>
Rendering quality of text (compared to the Windows versions)
</TD>
<TD>
average (at least with my font settings) 
</TD>
<TD>
average to bad (depends on the font sizes in CSS and fonts used)
</TD>
</TR>

<TR>
<TD>
Table support
</TD>
<TD>
yes
</TD>
<TD>
yes
</TD>
</TR>

<TR>
<TD>
Frames support
</TD>
<TD>
yes
</TD>
<TD>
yes
</TD>
</TR>

<TR>
<TD>
JavaScript support
</TD>
<TD>
yes but incomplete
</TD>
<TD>
yes
</TD>
</TR>

<TR>
<TD>
Java support
</TD>
<TD>
no - but seems to be planned for the future
</TD>
<TD>
yes
</TD>
</TR>

<TR>
<TD>
CSS support
</TD>
<TD>
yes
</TD>
<TD>
yes but incomplete
</TD>
</TR>

<TR>
<TD>
Stability
</TD>
<TD>
rather good<BR>
sometimes crashes without a reason (at least on my system)
</TD>
<TD>
average to good
</TD>
</TR>
</TABLE>

<A NAME="custom">
<H2>Customization</H2>
<P>
Opera is very customizable. You can select your own CSS style sheets to use, define
shortcuts to search engines.
</P>
<P>
You can choose the Identity string as well. Using this feature you can claim that 
your're using the MS IE or Netscape, instead of Opera.
</P>
<P>
This might be useful on some sites that require the use of one of the big browsers out there.
</P>
<P>
Last but not least, the screenlayout, the positons of navigation and status bars, can be customized.
You can even select to show an advanced navigation bar instead of the small default one.
</P>

<A NAME="technical">
<H2>Technical Notes &amp;  Downloading</H2>
<P>
Opera uses the QT2.2 library. However, it runs nicely without KDE.
</P>
<P>
Opera is available at <A HREF="http://www.opera.com/" TARGET="_blank">www.opera.com</A>.
</P>
<P>
You can choose between tar.gz, deb and rpm packages. These come either statically linked or 
dynamically linked.
</P>
<P>
A version for PowerPC Linux is available as well.
</P>
<P>
I suggest using the statically linked version. Although the packages are bigger, it is more 
likely that Opera runs.
</P>

<A NAME="install">
<H2>Installation</H2>
<P>
Installation of Opera is easy.
</P>
<P>
The tar.gz archives come along with an install skript.
</P>
<P>
Just unpack the Opera archive to a temporary place and run <EM>install.sh</EM> in the directory,
</P>
<P>
I have no experience with the deb or rpm packages of Opera.
Due to the nature of these formats, I suppose that both just install Opera and you can use it 
out of the box afterwards.  
</P>

<A NAME="conclusion">
<H2>Conclusion</H2>
<P>
Opera is a fast and lightweight web browser. It has very good features and is able to render
almost 90% of the webpages out there.  
</P>
<P>
There are still some features missing or incomplete. At least, today, Opera is still not
ready to be used standalone if you want Javascript, Java and multimedia stuff like Flash.
But if you can live without these, you'll find that 
Opera <EM>can be</EM> a 100% replacement for Netscape. 
</P>
<P>
Give it a try. Opera has many nice features not seen before in the Linux environment.
</P>




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Matthias Arndt.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Your Own Home Domain With ADSL</font></H1>
<H4>By <a href="mailto:rayxtra@hotmail.com">Ray Chan</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->



<P> <EM>Note: Domain names and IP numbers in this article have been changed.
I have no connection with myfakedomain.com and myhome.net--please do not send
questions or complaints to them.</EM>

<P>
<B>Acknowledgement</b><P>
This article is a walk through the steps I did  
to host my own domain name at home.
It is not a guide or tutorial about how to set up and host your domain. 
There are already lots of HOW-TOs
and tutorials on that topic. However, this artice provides working example for your reference, and I've also included
 URLs to some really useful web sites. <P>
<B>Background</b><P>
In late 2000, when everyone were talking or already using broadband, I was
still using my Hayes 28.8kbps modem to surf the net. My reason is simple, none of the broadband 
provider provides fix I.P. address although they did provide unlimited usage plan. I have a few 
domains name registered and hosting at some ISP. The service of the web hosting companies are 
limiting to html, perl cgi, pop server and maybe mod_rewrite. They never provide SMTP, MySQL, PHP4.
 whatever useful or at a really high price. That's why I'm looking for a broadband provider 
 willing to provides fix I.P. so that I can host my own web site and run whatever I want.<P>
 
Thanks god. At Jan 2001, one of the broadband provider at my area annouced that they 
will provides fix I.P. with extra cost. It is really expensive but hey that's what I need. I'm willing 
to pay for any services that fit my needs. On the other hand, I can save a lot of butts from web hosting
company where my domain names currently located. Why not dynamic I.P.? Yes dynamic I.P. may also do the same 
using some tricks with dynamic DNS as provided by no-ip, DynDNS... etc. but it is too annonying and 
not really good if you are going to host your own email server. <P>

<B>Planning the Network</b><P>

OK I subscribed to the broadband service finally. It takes two weeks to arrange a technical guy to
install the splitter and ADSL modem. Actually I can do it myself but they don't want me to. Anyway this 
is a good time to build the network and prepair for the
high speed connection. Before actually building the network, it is better to think about the topology
first. I make use of my spare old hardware and spent some money to build two linux box.
One linux box will be the baston host running Apache web server, ftp server, email server and MySQL 
database server. The baston host will act as an exterior router routing traffic between the internet
and the intranet. The other linux box will be the Intranet server hosting internal application and
data. The intranet box will act as an interior router. Someone asked, why two linux box? Well, 
for security reason of course. Please refer to your technical books about firewalling
for details explaination. Figure 1 shows the network diagram of my home network.<P>

<IMG SRC="misc/chan/raynet.jpg" VSPACE="20" BORDER="2"><P>

Since I got only one fixed IP, I'm not going to run any high traffic web site.
Only one baston host may do the job well, since it is a basic and simple
network. It is the solution for me, not neccessary for everyone who are reading
this article. Again, think about your own plan.<P>

<B>Building the network</b><P>
I downloaded and installed RedHat 7.0 to both of the linux boxes. Choose your own packages that sounds 
interest to you. It is fine for you to use other distribution. However, there were some essential 
components required in order to setup an internet server. Please refer to the HOW-TO at linuxdocs.org.
Again this is not a tutorial. I strongly suggest the following HOW-TOs for this section:<P>
<LI><A HREF="http://linuxdocs.org/HOWTOs/ISP-Setup-RedHat.html">ISP-Setup-ReadHat</a>
<LI><A HREF="http://linuxdocs.org/HOWTOs/DSL-HOWTO/index.html">DSL HOWTO for Linux</a>
<P>

And the following mini-HOWTOs:<P>
<LI><A HREF="http://linuxdocs.org/HOWTOs/mini/Domain.html">Setting Up Your New Domain Mini-HOWTO</a>
<LI><A HREF="http://linuxdocs.org/HOWTOs/mini/Home-Network-mini-HOWTO.html">Home-Network-mini-HOWTO</a>
<LI><A HREF="http://linuxdocs.org/HOWTOs/mini/IP-Subnetworking.html">IP-Subnetworking</a>
<P>
If you know nothing about what linux can do, you must read '<A HREF="http://www.linuxdoc.org/HOWTO/Networking-Overview-HOWTO.html">The Linux Networking Overview HOWTO</a>'.<P>

<B>Secure the baston host by packet filtering firewall using ipchains</b><P>
Ok now I got RedHat installed but the linux boxes were not protected yet. I need to setup firewall
and routing table in order to protect the linux machines and forwarding packets from Internal network 
to extranet network. This is a really big job for home user, and me too. I did a lot of search at 
freshmeat.net, google and sourceforge. I tried a lot of free firewalling 
scripts and none of them provides good security and hard to modify. Yes I'm lazy to write my own
filtering and routing rules. You are lucky. I found a really good firewall scripts @ <A HREF="http://iceberg.als.cx/">ICEBERG</a>.
Their scripts are easy to modify and setup all the routing. I run their scripts on both of my linux
machines and then I'm free to do other tasks now. Thanks again ICEBERG. Following is a list of useful 
documentation regarding firewalling and packet forwarding:<P>
<LI><A HREF="http://www.linuxdoc.org/HOWTO/Firewall-HOWTO.html">Firewall-HOWTO</a><BR>
<LI><A HREF="http://www.linuxdoc.org/HOWTO/IP-Masquerade-HOWTO.html">IP-Masquerade-HOWTO</a><BR>
<LI><A HREF="http://www.linuxdoc.org/HOWTO/IPCHAINS-HOWTO.html">IPCHAINS-HOWTO</a><P>
If you wanna use Napster behind the firewall, you should read <A HREF="http://www.linuxdoc.org/HOWTO/mini/IPMasquerading+Napster.html">IPMasquerading+Napster mini-HOWTO</a><BR>
<P>
<B>Setup External DNS Server at baston host</b><P>
Although I'll use <A HREF="http://www.hn.org">HAMMER NODE</a> to host the DNS entry for my domain name, a working caching only nameserver is still required 
to run the linux box. Configuration files were shown below:<P>
<A HREF="misc/chan/baston/named.boot.txt">/etc/named.boot</a><BR>
<A HREF="misc/chan/baston/named.conf.txt">/etc/named.conf</a><BR>
<A HREF="misc/chan/baston/named.ca.txt">/var/named/named.ca</a><BR>
<A HREF="misc/chan/baston/named.local.txt">/var/named/named.local</a><BR>
<A HREF="misc/chan/baston/named.myfakedomain.com.txt">/var/named/named.myfakedomain.com</a><BR>
<A HREF="misc/chan/baston/named.myhome.net.txt">/var/named/named.myhome.net</a><BR>
<A HREF="misc/chan/baston/named.rev.3.txt">/var/named/named.rev.3</a><BR>
<A HREF="misc/chan/baston/named.rev.2.txt">/var/named/named.rev.2</a><BR>
<P>

<B>Connecting to the ADSL modem</b><P>
Connecting the ADSL modem under linux is easy, just download the RPM of <A HREF="http://www.roaringpenguin.com/pppoe/">RP-PPPOE</a> from Roaring Penguin 
Software Inc, install it and then run the adsl-setup, that's all. As easy as an window machine.<P>

<B>Migrating domain name to baston host</b><P>
At this moment, the web server does not seems working yet. I fixed it by adding the line below
to the /etc/httpd/conf/httpd.conf file:<P>
ServerName www.myfakedomain.com (for baston host)<BR>
ServerName www.myhome.net (for Intranet Server)<P>
The web servers on both linux were up and running after a reboot. Now what's next? I started my favourite
browser Netscape and did a search on my favourite search engine Google for a Free DNS server. Finally I 
reach <A HREF="http://www.hn.org">HAMMER NODE</a>. I was lucky that I could reached hn.org. They provides free services for both dynamic I.P. and 
static I.P. user. They have good and easy to use UI and manages to provides both reliable and stable service. I created a virtual
domain mappings accounts and have the configuration like this:<P>
<FORM>
<TABLE BORDER=2 CELLSPACING=0 CELLPADDING=3>
    <TR><TD>Rec FQDN</TD><TD>Rec Type</TD><TD>Rec Value</TD><TD>DynDNS</TD><TD>MX Pref</TD><TD>Commands</TD></TR>

 <TR><TD>myfakedomain.com<TD>NS</TD><TD>ns1.hn.org</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16131">

 <TR><TD>myfakedomain.com</TD><TD>NS</TD><TD>aux1.hn.org</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16132">

 <TR><TD>www.myfakedomain.com</TD><TD>CNAME</TD><TD>myfakedomain.com</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16134">

 <TR><TD>myfakedomain.com</TD><TD>A</TD><TD>202.xxx.xxx.xxx</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16135">
 <TR><TD>mail.myfakedomain.com</TD><TD>MX</TD><TD>202.xxx.xxx.xxx</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16136">
 <TR><TD>ns.myfakedomain.com</TD><TD>NS</TD><TD>myfakedomain.com</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="16138">
 <TR><TD>mail.myfakedomain.com</TD><TD>CNAME</TD><TD>myfakedomain.com</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="18787">
 <TR><TD>ns.myfakedomain.com</TD><TD>CNAME</TD><TD>myfakedomain.com</TD><TD>0</TD><TD>0</TD><TD><INPUT TYPE=BUTTON VALUE="Delete"></TD></TR>
 <INPUT TYPE=HIDDEN NAME="dom_m" VALUE="2750"><INPUT TYPE=HIDDEN NAME="d_rr_idno" VALUE="18823">
</FORM>
   </TABLE>
<P>
After setup the DNS account from hn.org, I change the DNS entry, both of the primary and secondary server to the DNS server
provided by hn.org from the domain registration company (usually register.com or whatever). It may take some times to get
the DNS entry refresh. 
<P>
Wonderful! Now the DNS entry was refreshed and all request to www.myfakedomain.com will forward to my baston host. That's simple huh?
Thanks for the great work of hn.org. For details about how to setup DNS entries, please refer to <A HREF="http://www.linuxdoc.org/HOWTO/DNS-HOWTO.html">DNS-HOWTO</a>.<P>
Because the machine connected to ADSL modem provide services for the public, that mean it will be accessed by anyone who have Internet
access from anywhere. I need to restrict the access of various tcpd services for this machine for security reason. I edited the file 
/etc/hosts.allow and /etc/hosts.deny accordingly:<P>
/etc/hosts.allow<P>
ALL: 127.0.0.1<BR>
in.telnetd: 192.168.2.2<BR>
in.ftpd: 192.168.2.2<BR>
sshd: 192.168.2.2 203.xxx.xxx.xxx
<P>
/etc/hosts.deny<P>
ALL: ALL : spawn (echo Attempt from %h %a to %d at `date` | tee -a /xxx/xxx/tcp.deny.log | mail my@email.com )
<P>
As shown from the above configuration files, all machines from internal network can telnet, ftp, ssh and 
sftp to the baston host. The address 203.xxx.xxx.xxx is the I.P. address of my office machine which is allowed
to remote login to the baston host using ssh and transfer file to the baston host using sftp. Telnet and ftp to
the baston host will never allow from machine outside the internal network because user name and password is 
 transmit in plaintext format. It may be captured by hacker easily. HTTPD is not included in the above
configuration file because HTTPD is not under controlled of INETD. <P>

<B>Connect to the baston host safely using SSH</b><P>
Telnet and FTP is allowed to connect to the baston host from the internal network. SSH and SFTP must be used to 
connect from external network. Refer to the article '<A HREF="../issue61/dellomodarme.html">Using ssh</A>' from <I>Linux Gazette</I> about how
to setup and usage of SSH. You must install and running SSHD in order to support SSH. SFTP can be download from 
<A HREF="http://enigma.xbill.org/sftp/">http://enigma.xbill.org/sftp/</a>. SFTP is easy to use and install, please refer to the readme from the web site.<P>

<B>Setup the Intranet Server</b><P>
In order to protect the internal network, I 
disable all access from external network to my internal network:<P>

/etc/hosts.allow<P>

ALL: LOCAL 192.168.1.2 192.168.1.7<P>

/etc/hosts.deny<P>

ALL: ALL : spawn (echo Attempt from %h %a to %d at `date` | tee -a /xxx/xxx/tcp.deny.log | mail my@email.com )<P>

An email will be sent to my mailbox in case there are any activities attempt to connect to any prohibited services
to both of my linux server.<P>

As shown from figure 1, all internal machines have a host name. You can use whatever
host name and domain name for your internal network even the domain name is already 
registered at NIC, however, special care must be taken when setting up your own internal
DNS server.<P>

<B>Setting up intranet DNS server - named</b><P>
Again, please refer to the HOWTO or technical books about how to setup a DNS server.
Following shows my configuration files of the DNS server running at the Intranet 
server:<P>
<A HREF="misc/chan/intraserver/named.boot.txt">/etc/named.boot</a><BR>
<A HREF="misc/chan/intraserver/named.conf.txt">/etc/named.conf</a><BR>
<A HREF="misc/chan/intraserver/named.ca.txt">/var/named/named.ca</a><BR>
<A HREF="misc/chan/intraserver/named.local.txt">/var/named/named.local</a><BR>
<A HREF="misc/chan/intraserver/named.myhome.net.txt">/var/named/named.myhome.net</a><BR>
<A HREF="misc/chan/intraserver/named.rev.1.txt">/var/named/named.rev.1</a><BR>
<A HREF="misc/chan/intraserver/named.rev.2.txt">/var/named/named.rev.2</a><BR>
<P>
<B>More security issues</b><P>
Hackers are arounding you, only firewalling with packet filtering and 
controlling services access from hosts.allow/hosts.deny are never enough.
A few security holes may discover everyday. You should subscribes to 
corresponding mailing list and upgrade your linux constantly. A few more articles and
software about security is good and worth to introduce:<P>
<LI><A HREf="../issue46/pollman.html">Security for the Home Network LG #46</a><BR>
<LI><A HREF="http://www.linux-firewall-tools.com/linux/">Linux Firewall and Security Site</a><BR>
<LI><A HREF="http://users.dhp.com/~whisper/mason/">Mason - the automated firewall builder for Linux</a><BR>
<LI><A HREF="http://www.astaro.com">Astaro AG (Great firewall linux distribution with web interface)</a><BR>
<LI><A HREF="http://www.ethereal.com/">The Ethereal Network Analyzer</a><BR>
<LI><A HREF="http://www.nessus.org/">Nessus - The Security Scanner</a><BR>
<LI><A HREF="http://www.stunnel.org/">Stunnel - Universal SSL Wrapper</a><BR>
<P>
<B>How about POP3 and SMTP server?</b><P>
POP3, as same as TELNET and FTP, transfer username and password in plaintext and is considered
insecure. SPOP maybe setup to encrypt POP data. However, I don't want to store my personal email 
in any machine outside internal network including my office's workstation. So I'm not going to
setup POP3 in the baston host. The reason not to allow SMTP because relaying mail is dangerous
because spammer will make use of your relayed SMTP server to send their hateful spam mails.
On the other hands, setting up a non-relayed SMTP server for yourself is meaningless because you
cannot send mail from your SMTP server outside the network. I
can simply login to my baston host using ssh and run pine to check and reply my message in a 
secure way.<P>
<B>Subdomain for web server</b><P>
Wow, everything working now. I can host my web server, email server and ftp server at my home linux
box. It rocks! Now I need a subdomain resume.myfakedomain.com to host my online resume. Just add
the following lines to the /etc/httpd/conf/httpd.conf handles all the magic:<P>

RewriteEngine on<BR>
## Ignore www.myfakedomain.com<BR>
RewriteCond %{HTTP_HOST}  !^www\.myfakedomain\.com [NC]<BR>
## A directory with the name of the subdomain must exist<BR>
RewriteCond %{DOCUMENT_ROOT}/%1 -d<BR>
## Add the requested hostname to the URI<BR>
## [C] means that the next Rewrite Rules uses this<BR>
RewriteRule ^(.+) %{HTTP_HOST}/$1 [C]<BR>
## Translate abc.myfakedomain.com/foo to myfakedomain.com/abc/foo<BR>
RewriteRule ^([a-z-]+)\.myfakedomain\.com/?(.*)$ http://www.myfakedomain.com/$1/$2 [L]<BR>

<P>
<B>Other useful configuration files</b><P>
/etc/hosts (baston host)<P>
<PRE>
127.0.0.1	localhost.localdomain 	localhost
192.168.2.1	router.myhome.net	router
192.168.2.2	gateway.myhome.net	gateway
202.xxx.xxx.xxx	www.myfakedomain.com	www
</pre>
<P>
/etc/hosts	(intranet gateway)<P>
<PRE>
127.0.0.1	localhost.localdomain 	localhost
192.168.1.1	server.myhome.net	server
192.168.1.2	devel.myhome.net 	devel
192.168.1.3	php.myhome.net	php
192.168.1.4	asp.myhome.net	asp
192.168.1.7	be.myhome.net	be
192.168.2.1	router.myhome.net	router
192.168.2.2	gateway.myhome.net	gateway
</pre><P>
/etc/resolv.conf	(baston host)<P>
<PRE>
search myfakedomain.com
nameserver	127.0.0.1
</pre><P>
/etc/resolv.conf	(intranet gateway)<P>
<PRE>
search	myhome.net
nameserver	127.0.0.1
</pre><P>
<B>Network Card Setting</b><P>
Ethernet port setting:<P>
<IMG SRC="misc/chan/serverport.jpg" BORDER="2">
<P>

More network configuration files:<P>
<A HREF="misc/chan/baston/network.txt">/etc/sysconfig/network</a> (baston host)<BR>
<A HREF="misc/chan/baston/ifcfg-eth0.txt">/etc/sysconfig/network-scripts/ifcfg-eth0</a> (baston host)<BR>
<A HREF="misc/chan/baston/ifcfg-eth1.txt">/etc/sysconfig/network-scripts/ifcfg-eth1</a> (baston host)<P>

<A HREF="misc/chan/intraserver/network.txt">/etc/sysconfig/network</a> (Intranet gateway)<BR>
<A HREF="misc/chan/intraserver/ifcfg-eth0.txt">/etc/sysconfig/network-scripts/ifcfg-eth0</a> (Intranet gateway)<BR>
<A HREF="misc/chan/intraserver/ifcfg-eth1.txt">/etc/sysconfig/network-scripts/ifcfg-eth1</a> (Intranet gateway)<P>
<A HREF="misc/chan/rc.local.txt">/etc/rc.d/rc.local</a> (Both of the Baston host and Intranet gateway)<BR>
<P>
<B>TCP/IP setting summary</b><P>
<TABLE border=1>
	<TR><TD colspan=2><B>Baston host</b></td></tr>
	<TR><TD>Default Gateway:</td><TD>ppp0</td></tr>
	<TR><TD>Nameserver:</td><TD>127.0.0.1</td></tr>
	<TR><TD colspan=2>&nbsp;</td></tr>
	<TR><TD>Network interface:</td><TD>eth0</td></tr>
	<TR><TD>I.P. Address:</td><TD>192.168.3.1</td></tr>
	<TR><TD>Subnet mask:</td><TD>255.255.255.0</td></tr>
	<TR><TD colspan=2>&nbsp;</td></tr>
	<TR><TD>Network interface:</td><TD>eth1</td></tr>
	<TR><TD>I.P. Address:</td><TD>192.168.2.1</td></tr>
	<TR><TD>Subnet mask:</td><TD>255.255.255.0</td></tr>
</table>
<P>
<TABLE border=1>
	<TR><TD colspan=2><B>Intranet Server</b></td></tr>
	<TR><TD>Default Gateway:</td><TD>192.168.2.1</td></tr>
	<TR><TD>Nameserver:</td><TD>127.0.0.1</td></tr>
	<TR><TD colspan=2>&nbsp;</td></tr>
	<TR><TD>Network interface:</td><TD>eth0</td></tr>
	<TR><TD>I.P. Address:</td><TD>192.168.1.1</td></tr>
	<TR><TD>Subnet mask:</td><TD>255.255.255.0</td></tr>
	<TR><TD colspan=2>&nbsp;</td></tr>
	<TR><TD>Network interface:</td><TD>eth1</td></tr>
	<TR><TD>I.P. Address:</td><TD>192.168.2.2</td></tr>
	<TR><TD>Subnet mask:</td><TD>255.255.255.0</td></tr>
</table>
<P>
<TABLE border=1>
	<TR><TD colspan=2><B>Workstations from Internal Network</b></td></tr>
	<TR><TD>Default Gateway:</td><TD>192.168.1.1</td></tr>
	<TR><TD>Nameserver:</td><TD>192.168.1.1</td></tr>
	<TR><TD colspan=2>&nbsp;</td></tr>
	<TR><TD>Network interface:</td><TD>eth0</td></tr>
	<TR><TD>I.P. Address:</td><TD>192.168.1.X</td></tr>
	<TR><TD>Subnet mask:</td><TD>255.255.255.0</td></tr>
</table>
<P>
<B>Further setup and reading</b><P>
What if you want to access your internal machine running 
windowsz from the other network while maintaining security through the firewall? 
The answer is using 
Virtual Private Network (<A HREF="http://whatis.techtarget.com/WhatIs_Definition_Page/0,4152,213324,00.html">VPN</a>) technology. Linux do 
support VPN in recent version. More details can be find
at <A HREF="http://linuxdocs.org/HOWTOs/mini/VPN.html">VPN HOWTO</a>.
If you have more than one domains and want to host at the same baston host, you may require special setting for your apache web server and sendmail
server. The next version of this article will include the walkthrough of the VPN and virtual domain setup.<P>

If you have any suggestions or comments regarding this document, please feel
free to contact me at <A HREF="mailto:rayxtra@hotmail.com">rayxtra@hotmail.com</A>.




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Ray Chan.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">HelpDex</font></H1>
<H4>By <a href="mailto:shane_collinge@yahoo.com">Shane Collinge</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<P> The first cartoon is a reference to the
<A HREF="http://www.cobalt.com/">Cobalt Qube</A>, a network server
appliance that runs on Linux.

<IMG ALT="forJon.jpg" SRC="misc/collinge/forJon.jpg" 
	WIDTH="750" HEIGHT="303">
<BR CLEAR="all">

<IMG ALT="accessgone.jpg" SRC="misc/collinge/accessgone.jpg" 
	WIDTH="750" HEIGHT="303">
<BR CLEAR="all">

<IMG ALT="name.jpg" SRC="misc/collinge/name.jpg" 
	WIDTH="750" HEIGHT="303">
<BR CLEAR="all">

<IMG ALT="0x88.jpg" SRC="misc/collinge/0x88.jpg" 
	WIDTH="750" HEIGHT="303">
<BR CLEAR="all">

<P> Shane's cartoon archive is available on his web site,
<A HREF="http://www.ShaneCollinge.com/">http://www.ShaneCollinge.com/</A>.


<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Shane Collinge.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Interview with Linux Today's Marty Pitts</font></H1>
<H4>By <a href="mailto:frc@linux.com.br">Fernando Ribeiro Correa</a> and
Marcos Martins Manh&atilde;es<BR>
Originally published at <A HREF="http://www.olinux.com.br/Interviews">OLinux</A></H4>
</center>
<P> <HR> <P>  

<!-- END header -->



<EM>Enjoy this interview with Marty Pitts, Managing Editor at 
<A HREF="http://www.linuxtoday.com/">Linux Today</A>. He
talks about Linux Today's evolution and the growth of its main subject, Linux
operating systems.</EM>

      <p>
        <b>OLinux: Please introduce yourself. (career, education,
        hobbies, personal and professional achievements).</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      My name is Marty Pitts.I worked in the nuclear industry for
      13 years before joining Linux Today, in jobs ranging from
      Purchasing Agent to Network Admin. I like to ski in the
      winter, hike and camp in the summer and read SciFi in
      between.I also like to play around with the latest Linux
      distros.</p>

      <p>
        <b>OLinux: How long have you been working and what are your
        responsibilities at Linux Today?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      When I became interested in using Linux at work, I started
      looking for a information about Linux online.One of the
      resources I came across was Linux Today. I liked that it was
      updated hourly. When I found news that they did not have, I
      started using their contrib form.After several months, the
      site owners: Dave and Dwight, asked if I would be interested
      in working as a volunteer on the site. Having become a Linux
      news junkie, I jumped at the chance.</p>

      <p>In the summer of 1999, Dave sent me an email asking what
      my employment situation was.It just so happened that at my
      current job, my boss of several year had just turned in his
      notice to quit.It was a good opportunity to think about a
      career change.How many people actually get a chance to work
      at what they love?</p>

      <p>I started working for Dave and Dwight in September of 1999
      full time as the Managing Editor.About a month later, Dave
      and Dwight sold the Linux Today web properties, which
      included 
      <A HREF="http://LinuxPR.com">LinuxPR.com</A>, to 
      <A HREF="http://internet.com">internet.com</A>. I have been as a full
      time employee of internet.com ever since.</p>

      <p>
        <b>OLinux: How's the site organized? Give us an idea of how
        the Linux Today works. How many people are involved?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      For the whole channel, which includes 14 web sites, there are
      approximately 9 full time editors and programmers.</p>

      <p>Right now there are two full time people who work on Linux
      Today, myself and Michael Hall.We also take care of LinuxPR
      and a couple of other sites in the Linux/Open Source
      channel.</p>

      <p>Michael lives on the east coast of the US, and I live in
      Washington state on the west coast.So naturally we break up
      the day, with Michael covering the first part of the day and
      then I come online later with a couple hours of overlap.</p>

      <p>
        <b>OLinux: Can you describe Linux Today evolution since it
        began?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      Dave and Dwight were the ones that came up with the idea for
      Linux Today and they are the ones who successfully executed
      that idea.They were successful enough to attract the
      attention of internet.com.</p>

      <p>It started as a labor of love for Dave and Dwight.They
      wanted to provide a resource that people could use to find
      out what was going in the Linux/Open Source world.They
      started the site on September 30, 1998. A year later, they
      had both quit their daytime jobs to work full time on the
      site, they had been able to hire a full time editor, and they
      had posted over 10,000 stories. Currently we are right at
      34,000 stories posted, just on Linux Today.</p>

      <p>After the sale of the site to internet.com, somethings
      changed and others remained surprisingly the same.Dave chose
      to leave and pursue other goals, Dwight stayed on and we
      worked to keep the site going. To replace Dave, who had done
      most of the site programming, Paul Ferris was hired.</p>

      <p>Paul, a great guy, started working on the programming side
      but still found time to write his column: Rant Mode Equals
      One.Currently we are using the second iteration of the site
      code, which Paul wrote, and we are about to roll out the
      third iteration of the code.It will provide increased
      flexibility so that the code will be able to be used across a
      variety of different sites, each with its own unique
      requirements.</p>

      <p>What stayed the same, during the transition, was the
      direction and focus of Linux Today. We were told to keep
      doing what we had been doing that had made Linux Today a
      popular site, which was a relief.</p>

      <p>Today we have a lot more original content than we used to.
      In addition, our focus is on making the whole Linux/Open
      Source channel work together well.</p>

      <p>
        <b>OLinux: Are there companies sponsoring or maintaining
        Linux Today?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      Since Linux Today is owned by internet.com, they are the ones
      who pay for the maintenance of this and the other sites.</p>

      <p>
        <b>OLinux: Is there any central control to avoid redundancy
        and improve editorial efforts?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      Yes. We have, as part of the backend to Linux Today, an
      Editorial Board that keeps track of who is working on what
      stories. In addition, we use email extensively, plus we have
      an IRC channel for quick communication.</p>

      <p>In spite of that, we occasionally will have a duplicate
      story go up.Which is why, sometimes you will see a message
      that says, 'This story has been unposted.'</p>

      <p>
        <b>OLinux: How difficult is to present good content day by
        day? Besides the users' contribution, do you have any other
        content resource (agencies like Reuters, etc.)?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      Early in the week, Monday and Tuesday, it is usually very
      easy to find content to post.As the week progresses though,
      it can be a struggle to find good content and resist posting
      something that is just a rehash of a story that has already
      been covered.Weekends are more difficult since there is
      usually no news from the traditional sources. Since we like
      to have time off as well, we break up the weekend between the
      editors and we also future post some items so that they show
      up over a regular period of time.This way, we are able to
      take a break and our readers can find some fresh content.</p>

      <p>Our readers are a very valuable source of content.Without
      them providing links and suggestions, Linux Today would not
      be were it is today.</p>

      <p>We are able to find some relevant content elsewhere within
      the internet.com properties, which we use when available.</p>

      <p>
        <b>OLinux: How do you see Linux Today in the Open Source
        world? What's the best contribution Linux Today has been
        giving to the Linux community during its existence?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      We see Linux Today as the place to stop if you want to know
      what is going on in the Linux/Open Source community today. We
      search out the events and news and bring it to one place so
      that our readers don't have to spend the time doing that
      search for themselves.</p>

      <p>Through the forums and story talkbacks, we help to
      facilitate discussions within the community and give our
      readers a place to react to the news of the day.</p>

      <p>I believe that Linux Today's greatest contribution is that
      we are able to raise the awareness of our readers about the
      events, good and bad, that are happening within and to the
      Linux/Open Source community.</p>

      <p>
        <b>OLinux: What are the new features being developed for
        Linux Today? Can you detail the main currently
        projects?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      We are about to roll out redesigned site software that will
      provide a greater flexibility and robustness to all the sites
      on the channel.</p>

      <p>
        <b>OLinux: What is your opinion about the growth of Linux
        in the enterprises? What about desktops, do you have a
        projection for the future?</b>
      </p>

      <p>
      <b>Marty Pitts:</b>

      From my experience, Linux is infused into the enterprise
      deeper than anyone suspects. When a problem can be solved
      without having to ask for a new budget item, the guys/gals on
      the front lines will use what works. I see the projections by
      companies like Gartner and IDC and I have to laugh. They
      don't know how to properly measure the revolution that is
      taking place under their noses. Their methodology can't
      account for stealth deployments.</p>

      <p>The desktop is there already. Ease of use and graphical
      tools have come a long way in just the past year or two. I
      use Linux as my work environment, and for many like myself,
      Linux is already there.Just look at what we have available to
      us, DVD decoding and playback capability, the latest video,
      sound and networking hardware. The environments available are
      amazing as well. Even though I don't use KDE or GNOME (I use
      a pure 'Enlightenment' desktop), I have both of them on my
      system and use their apps.</p>




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Fernando Ribeiro Correa.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Internet Printing - Another Way</font></H1>
<H4>By <a href="mailto:grahjenk@au1.ibm.com">Graham Jenkins</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<h2>The Problem</h2>

<p>You are doing some work on your home PC, connected to your favorite ISP - and
you decide you want to print a Word document on the high-speed color printer
at your office.
That printer is connected to the corporate LAN, but you can't talk to it
using LPR or IPP because it is hidden behind the corporate firewall.</p>

<p>You could perform a print-to-file operation, then email the file to somebody
at your office, and get them to send it to the printer. But there are a few
steps here - and it gets more complicated if there is a restriction on the
length of email messages which can be passed through one of the servers
along the way. You will then have to perform some sort of file-split operation
and send the individual parts.</p>

<h2>Client Software</h2>

<p>The people who make <a href="http://www.brother.com">Brother</a> printers thought
of all this, and developed a set of Windows printer drivers. These enable
users to print directly to a designated email address. The print-job is
automatically split into parts if necessary, and each part is base64-encoded
prior to transmission. Users can also nominate an address for email
confirmation.</p> 

<p>These Windows printer drivers (for Windows 95/98, and for Windows
NT-4.0/2000) can be downloaded from the Brother website.</p>

<h2>Printer Capabilities</h2>

<p>What the Brother people expect users to do their printing on is, of course,
a Brother printer - specifically, in this instance, one equipped with a
network card able to accept, decode and re-assemble mail messages directed
to it.</p>

<p>But what if you wish to print on a printer from another manufacturer?</p>

<h2>Doing it in Software</h2>

<p>My first stab at this was a Korn-shell program to which appropriate
incoming mail items were piped via a sendmail alias.  The program used
'awk' to extract information such as job and part number, then decoded
each such item into an appropriately named file in a designated directory.</p>

<p>After receiving a part, the program marked it as "complete", then
set an anti-simultaneity lock and went through a procedure to determine
if all necessary parts had been received in full. If they had, it
concatenated them in sequence, piped the result to the nominated printer,
and deleted them.</p>

<p>It was then that I started thinking: "What if there isn't enough room
to store all the parts for all the jobs which may currently be arriving?"
And: "How do the Brother people do it on a network card?"</p>

<h2>Doing it Without Local Storage</h2>

<p>The answer to my second question is: "They use a POP3 server!". The
components of each job stay on that server until the network card determines
that all necessary parts are available, at which stage it sucks them down
and decodes them in sequence, sending the output to the printer mechanism,
and requests their deletion from the server.</p>

<p>So here's how it can be done on a Linux machine. The program has been
written in Perl so that the NET::POP3 module can be used for easy access to
a POP3 server. It has been tested on both NetBSD and Solaris machines, so it
should work almost anywhere; all you'll have to change are the location of
the Perl interpreter, the name used for 'awk', and the format of the 'lpr'
command.
<A HREF="misc/jenkins/BIPprint.pl.txt">[Text version of this listing.]</A>
</p>

<pre>
#!/usr/bin/perl -w
# @(#) BIPprint.pl      Acquires Brother-Internet-Print files from POP3 server
#                       and passes them to designated printer(s). Small-memory
#                       version.  Intended for invocation via inittab entry.
#                       Graham Jenkins, IBM GSA, Feb. 2001. Rev'd: 17 Mar. 2001.

use strict;
use File::Basename;
use Net::POP3;
use Date::Manip;
use IO::File;
my $host="bronzeback.in.telstra.com.au";        # Same host and password for
my $pass="MySecret";                            # each printer.
my $limit=30*1024*1024;                         # Maximum bytes per print job.
my ($printer,$awkprog);
defined($ARGV[0]) || die "Usage: ", basename($0). " printer1 [ printer2 ..]\n";
open(LOG,"|/usr/bin/logger -p local7.info -t ".basename($0)); autoflush LOG 1;
$awkprog="";                            
while (&lt;DATA&gt;) {$awkprog = $awkprog . $_};      # Build awk program for later,
while (1) {                                     # then loop forever, processing 
  sleep 30;                                     # all printers in each pass, and
  foreach $printer (@ARGV) {process($printer);} # sleeping for 30 seconds
}                                               # between each pass.

sub process {
  my ($flag,$i,$j,$k,$l,$m,$allparts,$user,$pop,@field,@part,$count,$top15,
      $msgdate,$parsdate,$notify,$reply,%slot,$fh);
  $user = $_[0];
  $pop = Net::POP3-&gt;new($host);                 # Login to POP3 server and get
  $count = $pop-&gt;login($user,$pass) ;           # header plus 1st 15 lines
  $count = -1 if ! defined ($count) ;           # of each message. Use apop
  for ($i = 1; $i &lt;= $count; $i++ ) {           # instead of login if server
    $top15=$pop-&gt;top($i,15) ;                   # supports it.
    if ($top15) {                       
      $msgdate = ""; $notify="None"; $reply="";
      for ($j = 0; $j &lt; 99; $j++ ) {
        if (@$top15[$j]) {                      # Use arrival-date on POP3
          if($msgdate eq "") {                  # server to ascertain age of
            (@field) = split(/;/,@$top15[$j]);  # message; if it is stale,
            if ( defined($field[1])) {          # delete it and loop for next.
              $parsdate=&amp;ParseDate($field[1]);  # (Search for semi-colon
              if( $parsdate ) {                 # followed by valid date.)
                $msgdate="Y";
                if(&amp;Date_Cmp($parsdate, &amp;DateCalc("today","-3 days") ) lt 0 ) {
                  print LOG "Stale msg: $user $parsdate\n";
                  $pop-&gt;delete($i);
                  goto I;                       # If POP3 server does
                }                               # automatic message expiration
              }                                 # this entire section can be
            }                                   # omitted.
          }
          (@field) = split(/=/, @$top15[$j]);
          if ( defined($field[0]) ) {   
            if ($field[0] eq "BRO-NOTIFY") {chomp $field[1];$notify=$field[1];}
            if ($field[0] eq "BRO-REPLY")  {chomp $field[1];$reply =$field[1];}
            if ( $field[0] eq "BRO-PARTIAL" ) { # Comment above line to
              ( @part )=split("/", $field[1]);  # prevent mail notification.
              chomp $part[1];           
            }
            if ( $field[0] eq "BRO-UID" ) {     # Determine print-job and part
              chomp $field[1];                  # thereof contained in message.
              $slot{$field[1]."=".$part[0]} = $i ;
              $allparts = "Y";                  # As we see each message, check
              for ($k=1;$k&lt;=$part[1];$k++) {    # whether we have all parts.
                $allparts = "N" if ! defined($slot{$field[1]."=".$k}) ; 
              }
              if ( $allparts eq "Y" ) {         # Print and delete all parts.
                print LOG "$field[1] $part[1] =&gt; $user\n";
                if(($notify ne "None") &amp;&amp; ($reply ne "")) {system 
                  "echo Print Job Received, $part[1] pcs|Mail -s$user $reply";}
                $fh=new IO::File "|awk '{$awkprog}' Limit=$limit |lpr -P $user";
                for ($k = 1;$k&lt;=$part[1];$k++) {
                  $pop-&gt;get($slot{$field[1]."=".$k},$fh) ;
                  $pop-&gt;delete($slot{$field[1]."=".$k}) ;
                }                               # If there is enough filespace,
                $fh-&gt;close;                     # pipe awk output thru gzip to
              }                                 # a temporary file, then print
            }                                   # it and delete all parts; this
          }                                     # caters for connection failure.
        }                       
      }                                         # The awk program here-under
    }                                           # is used to extract parts from
I:}                                             # a file containing multiple
  $pop-&gt;quit() if ($count &gt;= 0);                # parts and feed each of them
}                                               # through a decoder to stdout.
__DATA__
if( Flag == 2 ) {
    Size=Size+length
    if(length == 0) { Flag=0; close("mmencode -u 2&gt;/dev/null") }
    else if(Size&lt;=Limit*4/3) print $0 |"mmencode -u 2&gt;/dev/null" }
  if( Flag == 1 ) if(length == 0) Flag=2
  if( Flag == 0 ) if($1 ~ /^Content-Transfer-Enc/) if($NF == "base64") Flag=1
</pre>

<h2>Program Walk-Through</h2>

<p>The program builds a small 'awk' program for later use; then, for each
printer declared on it's command line, it accesses a mailbox of the same
name and examines each message therein. If a message is stale, it is
deleted. Otherwise the contents of some Brother-specific lines are
extracted; these indicate
whether email notification is required, and which part of which job
is contained in the message.</p>

<p>If, during examination of a message, it is determined that all the parts
of its corresponding job have been seen in the mailbox, an email
notification is generated if required, and the parts are
extracted in sequence and piped via the 'awk' program (which decodes each
part as it arrives) to an appropriate printer command. Each part is deleted
as soon as it has been processed in this manner.</p>

<p>Ideally, we should wait
until success (or other) notification of print submission was obtained
before performing the email and deletion tasks; however, as noted in the
listing, this requires some local storage. In a like vein, whilst the
Brother client software allows selection of email notification for
several different conditions, we send notification of job submission unless
"None" has been selected.</p>

<h2>Concluding Remarks</h2>

<p>This program contains a password, so it should be readable only by the user
who will execute it. No special privileges are required for execution, and your
entry for it in /etc/inittab should look something like:
<pre>bi:345:respawn:su - nobody -c "/usr/local/bin/BIPprint.pl lp1 lp2 &gt;/dev/null 2&gt;&amp;1"</pre>

<p>If you have read this far, you are probably saying: "OK, so the program
doesn't need much local storage - but it sends its output to a print
spooler!  How bad is that?"  If the size of your spool area is of
concern, you can use something like 'netcat' or 'hpnpout' to send the job
directly to a printer port instead of spooling it. Or you may be able to
pipe your job through an FTP connection to your printer. If you do
bypass the spooler in this fashion, you should use a separate instance
of the program for each printer.</p>

<p>It's not rocket science, and there's no user-authentication or
content-encryption. But it may make your life a little easier. Enjoy!</p>



<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Graham Jenkins.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Parallel Processing on Linux with PVM and MPI</font></H1>
<H4>By <a href="mailto:jurahul@hotmail.com">Rahul U. Joshi</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<EM>This article aims to provide an introduction to PVM and MPI, two widely used
software systems for implementing parallel message passing programs. They enable us to use a group of heterogeneous UNIX/LINUX computers connected by a network as a single machine for solving a large problem.</EM>

<H2><A NAME="s1">1. Introduction to Parallel Processing</A></H2>

<P>Parallel processing is a form of computing in which a number of activities are
carried out concurrently so that the effective time required to solve the
problem is reduced. In the previous days, parallel processing was used for such
thing as large scale simulations (e.g. molecular simulations, simulation of the
explosion of an atomic bomb etc), solving large number crunching and data
processing problems (e.g. compiling the census data) etc. However, as the cost
of hardware is decreasing rapidly, parallel processing is being uses more and
more in routine tasks. Multiple processor servers have been in existence for
a long time. Parallel processing is also used in your own PC too. For example,
a graphics processor working along with the main processor to render graphics
on your monitor is also a form of parallel processing.
<P>
<P>However, apart from the hardware facilities for parallel processing, some
software support too is required so that we can run the programs in parallel
and coordinate their execution. Such a coordination is necessary due to the
dependencies of the parallel programs on one other. This will become 
clearer when we work through an example. The most widely used method to achieve
such coordination is <EM> message passing </EM> in which the programs coordinate
their execution and in general communicate with each other by passing <EM>message's</EM> to one other. So, for example, a program may tell another program,
``Ok! Here is the intermediate result you need to proceed.'' If all this sounds
too abstract, lets proceed with a very simple example.
<H2><A NAME="s2">2. A Very Simple Problem</A></H2>

<P>In this section, we will consider a very simple problem and consider how we
can use parallel processing to speed up its execution. The problem is to find
the sum of a list of integers stored in an array. Let us say that there are
100 integers stored in an array say <CODE>items</CODE>. Now, how do we parallelize
this program? That is, we must first find out a way in which this problem can
be solved by a number of programs working concurrently. Many a times, due to
<EM> data dependencies</EM>, parallelization becomes a difficult problem. For
example, if you want to evaluate <EM>(a + b) * c</EM>, which involves two
operations, we cannot do them concurrently, the addition must be done before
the multiplication. Fortunately, for the problem that we have chosen, 
parallelization is easy. Suppose that 4 program or processors will be working
simultaneously to solve the addition problem. Then the simplest strategy would
be to break the array <CODE>items</CODE> into 4 parts and have each program process
one part. Thus the parallelization of the problem is as follows:
<OL>
<LI> Four programs say P0, P1, P2 and P3 will solve the problem.</LI>
<LI> P0 will find the sum of array elements <CODE>items[0]</CODE> to 
<CODE>items[24]</CODE>. Similarly, P1 will find the sum of <CODE>items[25]</CODE>
to <CODE>items[49]</CODE>, P2 <CODE>items[50]</CODE> to <CODE>items[74]</CODE> and
P3 <CODE>items[75]</CODE> to <CODE>items[99]</CODE>.</LI>
<LI> After these programs have executed, there must be some other program to
find the sum of the 4 results obtained and give the final answer. Also,
the elements of the array <CODE>items</CODE> are not known to the programs
P0 to P3 and hence some program must tell these programs the values
of the elements. Thus, apart from P0 to P3, we will require one more program
that distributes data, collects results and coordinates execution.
We call such a program as <EM>master</EM> and the programs P0 to P3 as
<EM>slaves</EM> and this organization as the <EM>master - slave paradigm</EM>.</LI>
</OL>
<P>With this organization in mind, let us write the algorithms for the master
and the slave programs.

<HR NOSHADE>
<PRE>
/* Algorithm for the master program */
initialize the array `items'.

/* send data to the slaves */
for i = 0 to 3
    Send items[25*i] to items[25*(i+1)-1] to slave Pi
end for

/* collect the results from the slaves */
for i = 0 to 3
    Receive the result from slave Pi in result[i]
end for

/* calculate the final result */
sum = 0
for i = 0 to 3
    sum = sum + result[i]
end for

print sum
</PRE>
<HR NOSHADE>
The algorithm for the slave can be written as follows.
<HR NOSHADE>
<PRE>
/* Algorithm for the slave program */

Receive 25 elements from the master in some array say `items'

/* calculate intermediate result */
sum = 0
for i = 0 to 24
    sum = sum + items[i]
end for

send `sum' as the intermediate result to the master
</PRE>
<HR NOSHADE>

<H2><A NAME="s3">3. Implementing with PVM</A></H2>

<P>Now that the basic algorithm has been designed, let us now consider how we
can implement it. What hardware shall we run this program on? Clearly, very
few of us have access to special machines designed to run parallel programs.
However, no special hardware requirements are there in order to implement this
program. A single computer or a group of interconnected computers will do,
thanks to PVM, a software system that enables us to use interconnected computers
for parallel program execution. PVM stands for Parallel Virtual Machine. It
enables you to create number of programs or processes that run concurrently on same or different machines
and provided functions with which you can pass messages among the processes for
coordination. Even if you have a single computer, PVM will work on it, although
there will be no ``real'' parallel processing as such. However, for learning
purpose, that should be fine. Later on I will describe how to do ``real''
parallel processing using the PVM.
<P>In order to use the PVM system, you need to install the PVM software on your
Linux system. In case you are using Red Hat Linux, then the RPM package for
PVM is included on the CD, so that you can install it as you normally install
other packages. Assuming that you have installed PVM system on your machine,
create the following directories(s) in your home directory: 
<CODE>~/pvm3/bin/LINUX/</CODE>. Why ? Because PVM requires that some of the 
executables you create be copied in this directory. Once you have done this,
your setup is ready. Test this by giving the command <CODE>pvm</CODE> on the
prompt. This will start the <EM>PVM Console</EM> from which you can give
commands to the PVM system and query status information. If everything is set
OK, you will see the <CODE>pvm&gt;</CODE> prompt. Here enter the command <CODE>conf</CODE>.
The output should look something like this.
<PRE>
pvm&gt; conf
conf
1 host, 1 data format
                    HOST     DTID     ARCH   SPEED       DSIG
               joshicomp    40000    LINUX    1000 0x00408841
</PRE>
<P>What does this mean? The PVM System allows you to consider a group of 
interconnected LINUX system to be viewed as a ``virtual'' computer having much
higher computing capacity than the individual machines. Thus, PVM will
distribute the processes among a number of computers. However, by default, PVM
considers that only the host that you are working on is to be included in the
PVM machine, i.e. all processes you create will be scheduled to run on the
same host. The <CODE>conf</CODE> command shows what hosts or nodes are in the
PVM. Currently, there is only one. Later on, we will see how to add more hosts.
Presently, exit the PVM console by giving the command <CODE>halt</CODE>
<P>
<P>
<H2>3.1 A Demonstration Program</H2>

<P>Now that you are ensured that the PVM system has been properly installed,
let us see how to write the programs. Programs for the PVM system can be
written in both FORTRAN and C. We will be using the C language. To use the
PVM system, you include some calls to the PVM functions in your C program
along with the other statements and link the PVM library with your programs.
To get you started with PVM, let us write a simple program in which there
will be a master and a slave. The master will send the slave some string,
which the slave will convert to upper case and send back to the master. The
master and the slave programs are given as follows. To compile the programs,
give the command <CODE>make -f makefile.demo</CODE>. 

<P> <A HREF="misc/joshi/Pvmmpi.tgz">[Click here for a tar file containing the
program listings.]</A>

<HR NOSHADE>
<PRE>
      1 /* -------------------------------------------------------------------- *
      2  * master_pvm.c                                                         *
      3  *                                                                      *
      4  * This is the master program for the simple PVM demonstration.         *
      5  * -------------------------------------------------------------------- */
      6 #include &lt;stdio.h&gt;
      7 #include &lt;stdlib.h&gt;
      8 #include &lt;pvm3.h&gt;           /* declares PVM constants and functions */
      9 #include &lt;string.h&gt;
        
     10 int main()
     11 {
     12     int mytid;              /* our task ID          */
     13     int slave_tid;          /* task ID of the slave */
     14     int result;
     15     char message[] = "hello pvm";
     16     
     17     /* enroll ourselves into the PVM system and get our ID */
     18     mytid = pvm_mytid();
        
     19     /* spawn the slave */
     20     result = pvm_spawn("slave_pvm", (char**)0, PvmTaskDefault, 
     21                         "", 1, &amp;slave_tid);
        
     22     /* check if the slave was spawned successfully          */
     23     if(result != 1)
     24     {
     25         fprintf(stderr, "Error: Cannot spawn slave.\n");
        
     26         /* clean up and exit from the PVM system            */
     27         pvm_exit();
     28         exit(EXIT_FAILURE);
     29     }
        
     30     /* initialize the data buffer to send data to slave     */
     31     pvm_initsend(PvmDataDefault);
        
     32     /* ``pack'' the string into the data buffer             */
     33     pvm_pkstr(message);
        
     34     /* send the string to the slave with a message tag of 0 */
     35     pvm_send(slave_tid, 0);
        
     36     /* wait and receive the result string from the slave    */
     37     pvm_recv(slave_tid, 0);
        
     38     
     39     /* ``unpack'' the result from the slave                 */
     40     pvm_upkstr(message);
        
     41     /* show the result from the slave                       */
     42     printf("Data from the slave : %s\n", message);
        
     43     /* clean up and exit from the PVM system                */
     44     pvm_exit();
     45     
     46     exit(EXIT_SUCCESS);
     47 } /* end main() */
        
     48 /* end master_pvm.c */
</PRE>
<HR NOSHADE>
<PRE>
      1 /* -------------------------------------------------------------------- *
      2  * slave_pvm.c                                                          *
      3  *                                                                      *
      4  * This is the slave program for the simple PVM demonstration           *
      5  * -------------------------------------------------------------------- */
      6 #include &lt;stdio.h&gt;
      7 #include &lt;ctype.h&gt;
      8 #include &lt;stdlib.h&gt;
      9 #include &lt;pvm3.h&gt;
        
     10 #define MSG_LEN     20
     11 void convert_to_upper(char*);
        
     12 int main()
     13 {
     14     int mytid;
     15     int parent_tid;
     16     char message[MSG_LEN];
        
     17     /* enroll ourselves into the PVM system         */
     18     mytid = pvm_mytid();
        
     19     /* get the task ID of the master                */
     20     parent_tid = pvm_parent();
        
     21     /* receive the original string from master      */
     22     pvm_recv(parent_tid, 0);
     23     pvm_upkstr(message);
        
     24     /* convert the string to upper case             */
     25     convert_to_upper(message);
        
     26     /* send the converted string to the master      */
     27     pvm_initsend(PvmDataDefault);
        
     28     pvm_pkstr(message);
     29     pvm_send(parent_tid, 0);
        
     30     /* clean up and exit from the PVM system        */
     31     pvm_exit();
     32     
     33     exit(EXIT_SUCCESS);
     34 } /* end main() */
        
     35 /* function to convert the given string into upper case */
     36 void convert_to_upper(char* str)
     37 {
     38     while(*str != '\0')
     39     {
     40         *str = toupper(*str);
     41         str++;
     42     }
     43 } /* end convert_to_upper() */
        
     44 /* end slave_pvm.c */
</PRE>
<HR NOSHADE>
<PRE>
      1 # Make file for the demo PVM program
        
      2 .SILENT :
      3 # paths fro PVM include files and libraries
      4 INCDIR=-I/usr/share/pvm3/include
      5 LIBDIR=-L/usr/share/pvm3/lib/LINUX
        
      6 # link the PVM library
      7 LIBS=-lpvm3
      8 CFLAGS=-Wall
      9 CC=gcc
     10 TARGET=all
        
     11 # this is where the PVM executables go
     12 PVM_HOME=$(HOME)/pvm3/bin/LINUX
        
     13 all : $(PVM_HOME)/master_pvm $(PVM_HOME)/slave_pvm
        
     14 $(PVM_HOME)/master_pvm : master_pvm.c
     15     $(CC) -o $(PVM_HOME)/master_pvm master_pvm.c $(CFLAGS) $(LIBS) \
     16           $(INCDIR) $(LIBDIR)
        
     17 $(PVM_HOME)/slave_pvm : slave_pvm.c
     18     $(CC) -o $(PVM_HOME)/slave_pvm slave_pvm.c $(CFLAGS) $(LIBS) \
     19           $(INCDIR) $(LIBDIR)
</PRE>
<HR NOSHADE>
<P>
<P>Once your programs have been
compiled, you must copy them into the <CODE>~/pvm3/bin/LINUX</CODE> directory. 
(The makefile does it by default). Now to run the programs, you must first
start the PVM system. To do this give the command <CODE>pvm</CODE> to start the
PVM Console. Now at the <CODE>pvm&gt;</CODE> prompt, type <CODE>quit</CODE>. The output
will be as follows:
<PRE>
pvm&gt; quit
quit

Console: exit handler called
pvmd still running.
</PRE>

Notice the last line, indicating that the PVM daemon (<CODE>pvmd</CODE>) is still
running. To run the PVM programs, you need to run the PVM daemon which manages
the exchange of messages and that what we are doing here. Once the PVM daemon
is running, you can run the program by the following commands:
<PRE>
[rahul@joshicomp rahul]$ cd ~/pvm3/bin/LINUX/
[rahul@joshicomp LINUX]$ ./master_pvm
Data from the slave : HELLO PVM
[rahul@joshicomp LINUX]$
</PRE>
<P>
<P>Notice that the string is now in upper case as expected.
<P>
<H2>3.2 Explanation of the program</H2>

<P>In this section, we will see exactly how this program works. First of all
to use PVM function, you need to include a header file <CODE>pvm3.h</CODE> in
your programs. This is done in line 8 of <CODE>master_pvm.c</CODE> and in
line 9  of <CODE>slave_pvm.c</CODE>. Also when compiling the programs, you need
to link it with the PVM library. This is done by specifying the <CODE>-lpvm3</CODE>
option to the compiler, as done in line 7 of <CODE>makefile.demo</CODE>. Also, you
need to specify to the compiler the paths of header and library files, as
is done on lines 4 and 5 of the makefile.
<P>
<P>In the master program, we first get the <EM>task ID</EM> of the master by 
calling the PVM function <CODE>pvm_mytid()</CODE>. The
PVM system assigns each process a unique 32 bit integer called as its <EM>task
ID</EM> in the same way as Linux assigns each process a process ID. The task
ID helps us identify the process with which we need to communicate. However,
the master does not uses its task ID (stored in <CODE>mytid</CODE>) ever. Our
intention here is just to call the function <CODE>pvm_mytid()</CODE>. This
function enrolls the process into the PVM system and generates a unique task
ID for the process. If we do not explicitly enroll the process, PVM 
automatically enrolls our process on the first call to any PVM function. Next
we use <CODE>pvm_spawn()</CODE> to create the slave process. The first parameter,
<CODE>"slave_pvm"</CODE> is the name of the executable for the slave. The second
parameter is the arguments that you wish to the pass to the slaves (similar to 
<CODE>argv</CODE> in normal C). Since we do not want to send any arguments, we
set this value to 0. The third parameter is a flag with which we can control
how and where PVM starts the slave. Since we have only a single machine, we
set this flag to <CODE>PvmTaskDefault</CODE>, specifying PVM to use default
criteria while spawning the slave. The fourth parameter is the name of the
host or the architecture on which we wish to run the program and here it is
kept empty. It is used to specify the host or the architecture when the flag
is other than <CODE>PvmTaskDefault</CODE>.The fifth parameter specifies the number 
of slaves to spawn and the sixth parameter is a pointer to an array in which 
the IDs of the slaves will be returned. This function returns the number of 
slaves actually spawned which we check for correctness.
<P>
<P>A message in PVM consists of basically two parts, the data and a <EM>tag</EM>
that identifies the type of the message. The tag helps us distinguish between
different messages. For example, in the addition example, which we are going
to implement, suppose that you are expecting that each slave will send to the 
master an integer which is the sum of the elements it added. It is also
quite possible that some slave may encounter some error and may want to send
the master an integer which indicates the error code. How does the master
distinguish whether an integer it received from the slave is an intermediate
result or an error code? This is where tags come in picture. You may assign
the message for intermediate result a tag say <CODE>MSG_RESULT</CODE> which you
will <CODE>#define</CODE> in some header file and a tag say <CODE>MSG_ERROR</CODE> for
the message indicating error. The master will then look at the message tags
to decide whether the message contains intermediate result or error.
<P>
<P>To send a message, you first need to ``initialize'' the send buffer. This is
done by calling the <CODE>pvm_initsend()</CODE> function. The parameter to
this function specifies the ``encoding'' scheme to be used. When we want to
exchange data between machines with different architectures (like say between
a Pentium machine and a SPARC Workstation) then we need to encode the data at
the sending end and decode at the receiving end so that data is properly
delivered. The parameter to <CODE>pvm_initsend()</CODE> specifies the encoding
scheme to be used. The value <CODE>PvmDataDefault</CODE> specifies an encoding
scheme which enables data to be safely exchanged between heterogeneous
architectures. Once the buffer has been initialized, we need to put data into
the buffer and encode it. In our case, the data is a string, so we use the
function <CODE>pvm_pkstr()</CODE> to ``pack'' i.e. encode and put the data into
the buffer. If we had to send an integer, there is a different function
<CODE>pvm_pkint()</CODE>. Similarly, there are functions for other data types.
Once the data is packed, we call <CODE>pvm_send()</CODE> to send the message. 
The first argument is the ID of the process to which the message is to be sent
and the second argument is the message tag. Since there is only one type of 
message here, we set the tag to 0. 
<P>
<P>Once the data is sent to the slave, the slave will process it and return it
to the master as we shall see. So we now call <CODE>pvm_recv()</CODE> to receive
the data from the slave. Again, the parameters are the task ID from which
the message is expected and the tag of the expected message. If the desired
message has not yet been sent, this function waits and does not return. Thus,
in effect, the master is now waiting for the slave to process the data. Once
the message arrives, the data is still in the receive buffer. It needs to be
``unpacked'' i.e decoded to get the original message. This decoding is done
by the <CODE>pvm_upkstr()</CODE> function. We then display the processes string.
<P>
<P>Before the PVM program exits, it must tell the PVM system that it is leaving
the PVM system so that resources occupied by the process can be released. This
is done by calling the <CODE>pvm_exit()</CODE> function. After that, the master
exits.
<P>
<P>The slave program is easy to understand. First it finds the task ID of the
master (which is also its parent as the master spawned the slave) by calling
the function <CODE>pvm_parent()</CODE>. It then receives the message string from
the master, converts it to uppercase and send the resulting string to the
master.
<H2>3.3 The Addition Program</H2>

<P>Now that you know some basics of a PVM program, let us implement the addition
algorithm we developed using PVM. There will be one master and 4 slaves. The
master will first spawn 4 slaves and send each one their part of data. The 
slaves will add the data and send the results to the master. Thus, two
types of messages are exchanged, one when the master send data to slaves, for
which we will use the tag <CODE>MSG_DATA</CODE> and the other when the slaves
send results to master, for which we will use the tag <CODE>MSG_RESULT</CODE>.
The rest is simple. The master and the slave programs are given below.
<HR NOSHADE>
<PRE>
      1 /* -------------------------------------------------------------------- *
      2  * common.h                                                             *
      3  *                                                                      *
      4  * This header file defines some common constants.                      *
      5  * -------------------------------------------------------------------- */
      6 #ifndef COMMON_H
      7 #define COMMON_H
    
      8 #define NUM_SLAVES      4                   /* number of slaves     */
      9 #define SIZE            100                 /* size of total data   */
     10 #define DATA_SIZE       (SIZE/NUM_SLAVES)   /* size for each slave  */
    
     11 #endif
     12 /* end common.h */
</PRE>
<HR NOSHADE>
<PRE>
      1 /* -------------------------------------------------------------------- *
      2  * tags.h                                                               *
      3  *                                                                      *
      4  * This header file defines the tags that will be used for messages.    *
      5  * -------------------------------------------------------------------- */
      6 #ifndef TAGS_H
      7 #define TAGS_H
    
      8 #define MSG_DATA            101     /* data from master to slave    */
      9 #define MSG_RESULT          102     /* result from slave to master  */
    
     10 #endif
    
     11 /* end tags.h */
</PRE>
<HR NOSHADE>
<PRE>
  1 /* -------------------------------------------------------------------- *
  2  * master_add.c                                                         *
  3  *                                                                      *
  4  * Master program for adding the elements of an array by using PVM      *
  5  * -------------------------------------------------------------------- */
  6 #include &lt;stdio.h&gt;
  7 #include &lt;stdlib.h&gt;
  8 #include &lt;pvm3.h&gt;           /* PVM constants and declarations   */
  9 #include "tags.h"           /* tags for messages                */
 10 #include "common.h"         /* common constants                 */
    
 11 int get_slave_no(int*, int);
    
 12 int main()
 13 {
 14     int mytid;
 15     int slaves[NUM_SLAVES]; /* array to store the task IDs of slaves    */
 16     int items[SIZE];        /* data to be processes                     */
 17     int result, i, sum;
 18     int results[NUM_SLAVES];    /* results from the slaves              */
    
 19     /* enroll into the PVM system   */
 20     mytid = pvm_mytid();
    
 21     /* initialize the array `items' */
 22     for(i = 0; i &lt; SIZE; i++)
 23         items[i] = i;
    
 24     /* spawn the slaves             */
 25     result = pvm_spawn("slave_add", (char**)0, PvmTaskDefault,
 26                        "", NUM_SLAVES, slaves);
    
 27     /* check if proper number of slaves are spawned     */
 28     if(result != NUM_SLAVES)
 29     {
 30         fprintf(stderr, "Error: Cannot spawn slaves.\n");
 31         pvm_exit();
 32         exit(EXIT_FAILURE);
 33     }
    
 34     /* distribute the data among the slaves     */
 35     for(i = 0; i &lt; NUM_SLAVES; i++)
 36     {
 37         pvm_initsend(PvmDataDefault);
 38         pvm_pkint(items + i*DATA_SIZE, DATA_SIZE, 1);
 39         pvm_send(slaves[i], MSG_DATA);
 40     }
    
 41     /* receive the results from the slaves      */
 42     for(i = 0; i &lt; NUM_SLAVES; i++)
 43     {
 44         int bufid, bytes, type, source;
 45         int slave_no;
 46         
 47         /* receive message from any of the slaves       */
 48         bufid = pvm_recv(-1, MSG_RESULT);
    
 49         /* get information about the message            */
 50         pvm_bufinfo(bufid, &amp;bytes, &amp;type, &amp;source);
 51         
 52         /* get the slave number that sent the message   */
 53         slave_no = get_slave_no(slaves, source);
    
 54         /* unpack the results at appropriate position   */
 55         pvm_upkint(results + slave_no, 1, 1);
 56     }
    
 57     /* find the final result            */
 58     sum = 0;
 59     for(i = 0; i &lt; NUM_SLAVES; i++)
 60         sum += results[i];
    
 61     printf("The sum is %d\n", sum);
    
 62     /* clean up and exit from the PVM system    */
 63     pvm_exit();
    
 64     exit(EXIT_SUCCESS);
 65 } /* end main() */
 66         
 67 /* function to return the slave number of a slave given its task ID */
 68 int get_slave_no(int* slaves, int task_id)
 69 {
 70     int i;
    
 71     for(i = 0; i &lt; NUM_SLAVES; i++)
 72         if(slaves[i] == task_id)
 73             return i;
    
 74     return -1;
 75 } /* end get_slave_no() */
    
 76 /* end master_add.c */

</PRE>
<HR NOSHADE>
<PRE>
  1 /* -------------------------------------------------------------------- *
  2  * slave_add.c                                                          *
  3  *                                                                      *
  4  * Slave program for adding elements of an array using PVM              *
  5  * -------------------------------------------------------------------- */
  6 #include &lt;stdlib.h&gt;
  7 #include &lt;pvm3.h&gt;
  8 #include "tags.h"
  9 #include "common.h"
    
 10 int main()
 11 {
 12     int mytid, parent_tid;
 13     int items[DATA_SIZE];           /* data sent by the master  */
 14     int sum, i;
 15     
 16     /* enroll into the PVM system       */
 17     mytid = pvm_mytid();
    
 18     /* get the task ID of the master    */
 19     parent_tid = pvm_parent();
    
 20     /* receive the data from the master */
 21     pvm_recv(parent_tid, MSG_DATA);
 22     pvm_upkint(items, DATA_SIZE, 1);
    
 23     /* find the sum of the elements     */
 24     sum = 0;
 25     for(i = 0; i &lt; DATA_SIZE; i++)
 26         sum = sum + items[i];
    
 27     /* send the result to the master    */
 28     pvm_initsend(PvmDataDefault);
 29     pvm_pkint(&amp;sum, 1, 1);
 30     pvm_send(parent_tid, MSG_RESULT);
    
 31     /* clean up and exit from PVM       */
 32     pvm_exit();
 33     
 34     exit(EXIT_SUCCESS);
 35 } /* end main() */

</PRE>
<HR NOSHADE>
<PRE>
  1 # Make file for the PVM program for addition - makefile.add
    
  2 .SILENT :
  3 # paths fro PVM include files and libraries
  4 INCDIR=-I/usr/share/pvm3/include
  5 LIBDIR=-L/usr/share/pvm3/lib/LINUX
    
  6 # link the PVM library
  7 LIBS=-lpvm3
  8 CFLAGS=-Wall
  9 CC=gcc
 10 TARGET=all
    
 11 # this is where the PVM executables go
 12 PVM_HOME=$(HOME)/pvm3/bin/LINUX
    
 13 all : $(PVM_HOME)/master_add $(PVM_HOME)/slave_add
    
 14 $(PVM_HOME)/master_add : master_add.c common.h tags.h
 15     $(CC) -o $(PVM_HOME)/master_add master_add.c $(CFLAGS) $(LIBS) \
 16           $(INCDIR) $(LIBDIR)
 17   
 18 $(PVM_HOME)/slave_add : slave_add.c common.h tags.h
 19     $(CC) -o $(PVM_HOME)/slave_add slave_add.c $(CFLAGS) $(LIBS) \
 20          $(INCDIR) $(LIBDIR)
</PRE>
<HR NOSHADE>
<P>
<P>Let us consider the slave program first, because it is simple. The slave
receives the 25 array elements from the master in the array <CODE>items</CODE>,
finds their sum and sends the result to the master with the message tag
as <CODE>MSG_RESULT</CODE>. Now consider the master. We define an array 
<CODE>slaves</CODE> of size <CODE>NUM_SLAVES</CODE> which will store the task ID's of
the slaves spawned by the parent. There is another array <CODE>results</CODE> in
which the results from the slaves are stored. The master first initializes
the array <CODE>items</CODE> and then spawns the slaves. After that it distributes
the data among the slaves. In the call to <CODE>pvm_pkint()</CODE> on line 38,
the first parameter is the pointer to the array in which the integers are
stored, the second is the number of integers to pack and the third is the
``stride.'' Stride means how many elements to skip when packing. When it is 1,
consecutive elements are packed. When it is 2, PVM will skip 2 elements when
packing with the result that all even numbered elements (0, 2, 4 ...) will
be packed. Here we keep its value as 1. 
<P>
<P>Once the data has been distributed among the slaves, the master has to wait
till the slaves return the intermediate results. One possibility when
accepting the results is that the master will first collect the result from 
slave 0 (i.e slave whose task ID is stored in <CODE>slave[0]</CODE>), then from
slave 1 and so on. However, this may not be an efficient approach. For example,
it may be that slave 0 is working on a slower machine than slaves 1, 2 and 3.
In that case, since the master is waiting from slave 0, the results from
slaves 1, 2 and 3 are yet to be collected even though the calculations are
completed. In this case it may be fine, but consider the situation in which
the slave, when finished doing one job is given another job. In that case, we
would like to give a slave its next job immediately after it has completed its
current job. Thus, the master must be in a position to respond messages from
any of the slaves. This is what is being done here.
<P>In the call to <CODE>pvm_recv()</CODE> on line 48, we know that the first 
parameter is the task ID of the message source. If this value is kept -1, it 
signifies a <EM>wild card</EM> i.e. messages from any process with message tag
<CODE>MSG_RESULT</CODE> will be received by the master. The received message
along with some control information is stored in a buffer called as <EM>active
receive buffer</EM>. The call returns a unique ID for this buffer. Now, we
want to know who is the sender of the message so that we can assign the message
data to the appropriate element of the array <CODE>results</CODE>. The function
<CODE>pvm_bufinfo()</CODE> returns information about the message in the buffer,
such as the message tag, the number of bytes and the senders task ID. Once we
have the senders task ID, we set the appropriate element of the <CODE>results</CODE>
array to the integer sent by the slave. The rest of the program should be
easy to understand.
<P>
<H2>3.4 Working with PVM</H2>

<P>In case you are interested, you can think of some problems for which you can
write parallel programs. Many a times, due to bugs etc., you may need to clean
up the state of the things before starting. The PVM Console provides with
the command <CODE>halt</CODE> that kills the PVM daemon. Then all the PVM processes
will halt or you can halt them with the Linux <CODE>kill</CODE> command. In case
you have a network of Linux machines interconnected by say a LAN, then you
can also do ``real'' parallel processing. First of all, install PVM on all the
hosts you wish to use and then use the <CODE>add</CODE> command in the PVM Console
to add hosts to the virtual machine. Then PVM will schedule some of the 
processes to run on these hosts, so that real parallel processing is achieved.
<P>
<H2><A NAME="s4">4. Implementing with MPI</A></H2>

<P>We have seen in the previous section the implementation of the addition
program using the PVM. Now let us consider another approach that can be
used in developing parallel programs. This approach is using the MPI
library. MPI stands for <EM>Message Passing Interface</EM>. It is a standard
developed to enable us to write portable message passing applications. It
provides functions for exchanging messages and many other activities as well.
It must be noted that unlike PVM which is a software system, MPI is a standard,
so that many implementations of the MPI standard exist. We will use an
implementation of MPI called LAM which stands for <EM>Local Area Multicomputer</EM>. It is also available on the Red Hat Linux CD as an RPM package, so
installation may not be a problem.
<P>
<P>After you have installed the RPM package, go to the <CODE>/usr/boot</CODE>
directory and create a file named <CODE>conf.lam</CODE> and type in a single line 
in it: <CODE>lamd $inet_topo</CODE>. The same directory will also have a file
named <CODE>bhost.def</CODE> else create it and type in a single line in it:
<CODE>localhost</CODE>. Now to test whether everything is working correctly,
type at the prompt, <CODE>lamboot</CODE>. You will get the following response:
<PRE>
[rahul@joshicomp boot]$ lamboot

LAM 6.3.1/MPI 2 C++/ROMIO - University of Notre Dame

[rahul@joshicomp boot]$
</PRE>
<P>
<P>If the output indicates an error, then there is some problem with the 
installation, either follow the above steps or see the <EM>lamboot(1)</EM>
manual page for troubleshooting.
<P>
<P>Assuming that LAM/MPI is properly installed on your system, let us again
write a small demonstration program for MPI. 
<P>
<H2>4.1 A Demonstration MPI Program</H2>

<P>We will again write a simple master - slave program in which we are supposed to
evaluate the expression <EM>(a + b) * (c - d)</EM>. The master will read the
values of <EM>a, b, c,</EM> and <EM>d</EM> from the user and one slave will
calculate <EM>(a + b)</EM> and the other one will calculate <EM>(c - d)</EM>.
The program is as follows.
<P>
<HR NOSHADE>
<PRE>
  1 /* -------------------------------------------------------------------- *
  2  * mpi_demo.c                                                           *
  3  *                                                                      *
  4  * A simple MPI demonstration program to evaluate an expression.        *
  5  * -------------------------------------------------------------------- */
  6 #include &lt;stdio.h&gt;
  7 #include &lt;stdlib.h&gt;
  8 #include &lt;lam/mpi.h&gt;            /* for MPI constants and functions      */
    
  9 #define MSG_DATA        100     /* message from master to slaves        */
 10 #define MSG_RESULT      101     /* message from slave to master         */
    
 11 #define MASTER          0       /* rank of master                       */
 12 #define SLAVE_1         1       /* rank of first slave                  */
 13 #define SLAVE_2         2       /* rank of second slave                 */
    
 14 /* functions to handle the tasks of master, and the two slaves          */
 15 void master(void);
 16 void slave_1(void);
 17 void slave_2(void);
    
 18 int main(int argc, char** argv)
 19 {
 20     int myrank, size;
 21     
 22     /* initialize the MPI system                                        */
 23     MPI_Init(&amp;argc, &amp;argv);
    
 24     /* get the size of the communicator i.e. number of processes        */
 25     MPI_Comm_size(MPI_COMM_WORLD, &amp;size);
    
 26     /* check for proper number of processes                             */
 27     if(size != 3)
 28     {
 29         fprintf(stderr, "Error: Three copies of the program should be run.\n");
 30         MPI_Finalize();
 31         exit(EXIT_FAILURE);
 32     }
 33     
 34     /* get the rank of the process                                      */
 35     MPI_Comm_rank(MPI_COMM_WORLD, &amp;myrank);
    
 36     /* perform the tasks according to the rank                          */
 37     if(myrank == MASTER)
 38         master();
 39     else if(myrank == SLAVE_1)
 40         slave_1();
 41     else
 42         slave_2();
    
 43     /* clean up and exit from the MPI system                            */
 44     MPI_Finalize();
    
 45     exit(EXIT_SUCCESS);
 46 } /* end main() */
    
 47 /* function to carry out the masters tasks          */
 48 void master(void)
 49 {
 50     int a, b, c, d;
 51     int buf[2];
 52     int result1, result2;
 53     MPI_Status status;
    
 54     printf("Enter the values of a, b, c, and d: ");
 55     scanf("%d %d %d %d", &amp;a, &amp;b, &amp;c, &amp;d);
    
 56     /* send a and b to the first slave              */
 57     buf[0] = a;
 58     buf[1] = b;
 59     MPI_Send(buf, 2, MPI_INT, SLAVE_1, MSG_DATA, MPI_COMM_WORLD);
    
 60     /* send c and d to the secons slave             */
 61     buf[0] = c;
 62     buf[1] = d;
 63     MPI_Send(buf, 2, MPI_INT, SLAVE_2, MSG_DATA, MPI_COMM_WORLD);
    
 64     /* receive results from the slaves              */
 65     MPI_Recv(&amp;result1, 1, MPI_INT, SLAVE_1, MSG_RESULT, 
 66              MPI_COMM_WORLD, &amp;status);
 67     MPI_Recv(&amp;result2, 1, MPI_INT, SLAVE_2, MSG_RESULT, 
 68              MPI_COMM_WORLD, &amp;status);
    
 69     /* final result                                 */
 70     printf("Value of (a + b) * (c - d) is %d\n", result1 * result2);
 71 } /* end master() */
    
 72 /* function to carry out the tasks of the first slave       */
 73 void slave_1(void)
 74 {
 75     int buf[2];
 76     int result;
 77     MPI_Status status;
 78     
 79     /* receive the two values from the master       */ 
 80     MPI_Recv(buf, 2, MPI_INT, MASTER, MSG_DATA, MPI_COMM_WORLD, &amp;status);
 81     
 82     /* find a + b                                   */
 83     result = buf[0] + buf[1];
    
 84     /* send result to the master                    */
 85     MPI_Send(&amp;result, 1, MPI_INT, MASTER, MSG_RESULT, MPI_COMM_WORLD);
 86 } /* end slave_1() */
    
 87 /* function to carry out the tasks of the second slave      */
 88 void slave_2(void)
 89 {
 90     int buf[2];
 91     int result;
 92     MPI_Status status;
 93     
 94     /* receive the two values from the master       */
 95     MPI_Recv(buf, 2, MPI_INT, MASTER, MSG_DATA, MPI_COMM_WORLD, &amp;status);
 96     
 97     /* find c - d                                   */
 98     result = buf[0] - buf[1];
    
 99     /* send result to master                        */
100     MPI_Send(&amp;result, 1, MPI_INT, MASTER, MSG_RESULT, MPI_COMM_WORLD);
101 } /* end slave_2() */
    
102 /* end mpi_demo.c */
</PRE>
<HR NOSHADE>
<PRE>
  1 # Makefile for MPI demo program - makefile.mpidemo
  2 .SILENT:
  3 CFLAGS=-I/usr/include/lam -L/usr/lib/lam
  4 CC=mpicc
    
  5 mpi_demo : mpi_demo.c
  6     $(CC) $(CFLAGS) mpi_demo.c -o mpi_demo
</PRE>
<HR NOSHADE>
    
<P>To compile this program, give the command <CODE>make -f makefile.mpidemo</CODE>.
Once you have compiled the program, to run the program you first need to 
``start'' or ``boot'' the Local Area Multicomputer system. This is done with
the <CODE>lamboot</CODE> command. After that, to run the program by giving the
following command: <CODE>mpirun -np 3 mpi_demo</CODE>.
<PRE>
[rahul@joshicomp parallel]$ lamboot

LAM 6.3.1/MPI 2 C++/ROMIO - University of Notre Dame

[rahul@joshicomp parallel]$ mpirun -np 3 mpi_demo
Enter the values of a, b, c, and d: 1 2 3 4
Value of (a + b) * (c - d) is -3
[rahul@joshicomp parallel]$
</PRE>
<P>
<H2>4.2 Explanation of the Program</H2>

<P>To use the MPI system and functions, you first need to include the header
file <CODE>mpi.h</CODE> as is done in line 8. In case of PVM, different processes
are identified with their task ID's. In case of MPI, the MPI system assigns
each process a unique integer called as its <EM>rank</EM> beginning with 0.
The rank is used to identify a process and communicate with it. Secondly,
each process is a member of some <EM>communicator</EM>. A communicator can
be thought of as a group of processes that may exchange messages with each
other. By default, every process is a member of the communicator called
<CODE>MPI_COMM_WORLD</CODE>. Although we can create new communicators, this leads
to an unnecessary increase in complexity, so we suffice ourselves by using the
<CODE>MPI_COMM_WORLD</CODE> communicator.
<P>
<P>Any MPI program must first call the <CODE>MPI_Init()</CODE> function. This function
is used by the process to enter the MPI system and also do any specific
initialization required by the system. Next, we get the size of the 
<CODE>MPI_COMM_WORLD</CODE> communicator i.e. the number of processes in it using the <code>MPI_Comm_size()</code> function. The
first parameter is the communicator and the second is a pointer to an integer
in which the size will be returned. Here, we need exactly 3 processes, one 
master and two slaves. After that, we get the rank by calling 
<CODE>MPI_Comm_rank()</CODE>. The three processes will have ranks 0, 1 and 2. All
these processes are essentially identical i.e. there is no inherent 
master - slave relationship between them. So it is up to us to decide who will
be the master and who will be the slaves. We choose rank 0 as master and ranks
1 and 2 as slaves. It can also be seen that we have included the code for both
the master and the two slaves in the same program. Depending upon the rank,
we choose to execute the appropriate function. Note that there is no spawning
of processes as in PVM, and as we shall see, we choose to decide the number
of process to be spawned from a command line argument rather than the program
spawning slaves. Once the execution is finished,
we must call the <CODE>MPI_Finalize()</CODE> function to perform final clean up.
<P>
<P>Let us now consider the master function. After reading the values of a, b, c, 
and d from the user, the master must send a and b to slave 1 and c and d to
slave 2. Instead of sending the variables individually, we choose to pack them
up in an array and send the array of 2 integers instead. It is always better
to pack up the data you want to send into a single message rather than to send
a number of messages for individual data items, this saves the communication
overhead involved in passing the messages. Once the buffer is ready, unlike PVM,
we do not need to pack or encode the data, MPI will manage these details
internally. So we can directly call the <CODE>MPI_Send()</CODE> function to send
the data. The first parameter (line 59) is the address of the buffer, the
second one the number of elements in the message, the third is a specification
of the data type of the buffer, which here is <CODE>MPI_INT</CODE> specifying that
the buffer is an array of integers. Next comes the rank of the process to which
we want to send the message. Here it is <CODE>SLAVE_1</CODE> (#defined as 1).
Next is the <EM>message tag</EM> similar to that in case of PVM. Final parameter
is the communicator of which the receiver is a member, which in this case, is
<CODE>MPI_COMM_WORLD</CODE>.
<P>
<P>Once the data is distributed among the slaves, the master must wait for the
slaves to send the results. For simplicity, we first collect the message from
the slave 1 and then from slave 2. To receive a message, we use the 
<CODE>MPI_Recv()</CODE> function. Again, packing and decoding is handled by MPI
internally. The first argument (line 65) is the address of the buffer in which
to receive the data. The second is the size of the buffer in terms of the
number of elements, which in this case is 1. Next is the data type, which is
<CODE>MPI_INT</CODE> here. Next three parameters specify the rank of the source of
the message, the tag of the expected message and the communicator of which the
source is the member. The final argument is a pointer to a structure of type
<CODE>MPI_Status</CODE> in which some status information will be returned (however,
we ignore this information). Now that you know about the basic MPI terms,
the <CODE>slave_1()</CODE> and <CODE>slave_2()</CODE> functions should be clear.
<P>
<P>
<P>In this program, the code for the master as well as the slaves was in the same
executable file. Later on we will see how we can execute multiple executables.
From the makefile, we see that to compile the MPI program, a wrapper program
<CODE>mpicc</CODE> is provided which links the required libraries automatically.
To run the program, use the <CODE>mpirun -np 3 mpi_demo</CODE> command after 
booting the LAM. Here we specify LAM to create 3 processes, one master and two
slaves.
<P>
<H2>4.3 The Addition Program Again</H2>

<P>Let us now re implement the addition program that we designed before using MPI.
Here we will also show you how to execute separate programs in MPI. When we
use a single executable in the MPI program, we call it <EM>Single Program
Multiple Data (SPMD)</EM> application. When two or more executables are
involved, we call it <EM>Multiple Program Multiple Data (MPMD)</EM> application.
With LAM, MPMD programs are executed with the help of an <EM>application
schema</EM>. But before that, let us see the source of the master and the slave
programs.
<HR NOSHADE>
<PRE>
  1 /* -------------------------------------------------------------------- *
  2  * master_mpi.c                                                         *
  3  *                                                                      *
  4  * Master program for adding the elements of an array using MPI         *
  5  * -------------------------------------------------------------------- */
  6 #include &lt;stdio.h&gt;
  7 #include &lt;stdlib.h&gt;
  8 #include &lt;lam/mpi.h&gt;        /* MPI constants and functions              */
  9 #include "tags.h"           /* tags for different messages              */
 10 #include "common.h"         /* common constants                         */
    
 11 int main(int argc, char** argv)
 12 {
 13     int size, i, sum;
 14     int items[SIZE];
 15     int results[NUM_SLAVES];
 16     MPI_Status status;
    
 17     /* initlalize the MPI System                */
 18     MPI_Init(&amp;argc, &amp;argv);
    
 19     /* check for proper number of processes     */
 20     MPI_Comm_size(MPI_COMM_WORLD, &amp;size);
    
 21     if(size != 5)
 22     {
 23         fprintf(stderr, "Error: Need exactly five processes.\n");
 24         MPI_Finalize();
 25         exit(EXIT_FAILURE);
 26     }
    
 27     /* initialize the `items' array             */
 28     for(i = 0; i &lt; SIZE; i++)
 29         items[i] = i;
    
 30     /* distribute the data among the slaves     */
 31     for(i = 0; i &lt; NUM_SLAVES; i++)
 32         MPI_Send(items + i*DATA_SIZE, DATA_SIZE, MPI_INT, i + 1,
 33                  MSG_DATA, MPI_COMM_WORLD);
    
 34     /* collect the results from the slaves      */
 35     for(i = 0; i &lt; NUM_SLAVES; i++)
 36     {
 37         int result;
 38         
 39         MPI_Recv(&amp;result, 1, MPI_INT, MPI_ANY_SOURCE, MSG_RESULT,
 40                  MPI_COMM_WORLD, &amp;status);
 41         results[status.MPI_SOURCE - 1] = result;
 42     }
    
 43     /* find the final answer                    */
 44     sum = 0;
 45     for(i = 0; i &lt; NUM_SLAVES; i++)
 46         sum = sum + results[i];
    
 47     printf("The sum is %d\n", sum);
    
 48     /* clean up and exit the MPI system         */
 49     MPI_Finalize();
    
 50     exit(EXIT_SUCCESS);
 51 } /* and main() */
    
 52 /* end master_mpi.c */
</PRE>
<HR NOSHADE>
<PRE>
  1 /* -------------------------------------------------------------------- *
  2  * slave_mpi.c                                                          *
  3  *                                                                      *
  4  * Slave program for adding array elements using MPI.                   *
  5  * -------------------------------------------------------------------- */
  6 #include &lt;stdio.h&gt;
  7 #include &lt;stdlib.h&gt;
  8 #include &lt;lam/mpi.h&gt;        /* MPI functions and constants  */
  9 #include "tags.h"           /* message tags                 */
 10 #include "common.h"         /* common constants             */
    
 11 #define MASTER  0           /* rank of the master           */
    
 12 int main(int argc, char** argv)
 13 {
 14     int items[DATA_SIZE];
 15     int size, sum, i;
 16     MPI_Status status;
    
 17     /* initialize the MPI system            */
 18     MPI_Init(&amp;argc, &amp;argv);
    
 19     /* check for proper number of processes */
 20     MPI_Comm_size(MPI_COMM_WORLD, &amp;size);
    
 21     if(size != 5)
 22     {
 23         fprintf(stderr, "Error: Need exactly five processes.\n");
 24         MPI_Finalize();
 25         exit(EXIT_FAILURE);
 26     }
    
 27     /* receive data from the master         */
 28     MPI_Recv(items, DATA_SIZE, MPI_INT, MASTER, MSG_DATA,
 29              MPI_COMM_WORLD, &amp;status);
    
 30     /* find the sum                         */
 31     sum = 0;
 32     for(i = 0; i &lt; DATA_SIZE; i++)
 33         sum = sum + items[i];
    
 34     /* send the result to the master        */
 35     MPI_Send(&amp;sum, 1, MPI_INT, MASTER, MSG_RESULT, MPI_COMM_WORLD);
    
 36     /* clean up and exit MPI system         */
 37     MPI_Finalize();
    
 38     exit(EXIT_SUCCESS);
 39 } /* end main() */
    
 40 /* end slave_mpi.c */
</PRE>
<HR NOSHADE>
<PRE>
  1 # Makefile for MPI addition program - makefile.mpiadd
  2 .SILENT:
  3 CFLAGS=-I/usr/include/lam  -L/usr/lib/lam
  4 CC=mpicc
    
  5 all : master_mpi slave_mpi
    
  6 master_mpi : master_mpi.c common.h tags.h
  7     $(CC) $(CFLAGS) master_mpi.c -o master_mpi
    
  8 slave_mpi : slave_mpi.c common.h tags.h
  9     $(CC) $(CFLAGS) slave_mpi.c -o slave_mpi
</PRE>
<HR NOSHADE>
<P>
<P>To compile the programs, type <CODE>make -f makefile.mpiadd</CODE>. (The
files <code>common.h</code> and <code>tags.h</code> are the same as used for the PVM program.)
This will create the <CODE>master_mpi</CODE> and <CODE>slave_mpi</CODE> executables. Now how do
we tell MPI to run both these executables. This is where <EM>application
schema file</EM> comes in. The application schema file specifies the executables
to be run, the nodes on which to run and the number of copies of the executable
to run. Create a new file <CODE>add.schema</CODE> and type in it the following
lines:
<PRE>
# Application schema for the addition program using MPI
n0 master_mpi
n0 -np 4 slave_mpi
</PRE>
<P>This file specifies that MPI should start 1 copy of the master (which will have
rank 0) and 4 copies of slaves on the node n0, i.e. the local node. You can
specify many more parameters in this schema file like command line arguments 
etc., see the manual page <EM>appschema(1)</EM>. Once the schema file is ready,
you can run the programs as follows:
<PRE>
[rahul@joshicomp parallel]$ lamboot

LAM 6.3.1/MPI 2 C++/ROMIO - University of Notre Dame

[rahul@joshicomp parallel]$ mpirun add.schema
The sum is 4950
[rahul@joshicomp parallel]$
</PRE>
<P>
<P>Much of the program should be easy to understand. On line 39, when receiving
intermediate results from the slaves, we specify the source as 
<CODE>MPI_ANY_SOURCE</CODE>, since we want to respond to slaves in the order in
which they complete the calculations, as discussed earlier. In this case, the
<CODE>status</CODE> structure contains the actual source in the field 
<CODE>MPI_SOURCE</CODE>. We use this information to set the appropriate element from
the <CODE>results</CODE> array to the intermediate result received. 
<P>In case you have a network of interconnected computers, you can make programs
run on many computers by suitably modifying the application schema file. Instead
of specifying n0 as the host, specify the name of the host and the number of
processes you wish to schedule on that host. For more information about this,
see the manual pages and the references. 
<P>
<H2><A NAME="s5">5. Conclusion</A></H2>

<P>We have seen how to write parallel programs using the PVM and MPI libraries.
Since there libraries are available on many platforms and these are the defacto
standards used for implementing parallel programs, programs written with PVM
or MPI will run with little or no modification on large scale machines, if the
need arises. What we have basically concentrated on in this article is the
<EM>point to point</EM> communication functions provides by these libraries and
their use in message passing. Apart from these facilities, both PVM and MPI
provide a number of advanced features such as <EM>collective communication 
(broadcasting or multicasting), process groups and group management, reduction
functions etc.</EM> You are welcome to explore these advanced features. These
public domain softwares enable us to use a network of computers as a single large
computer, so in case you have some such large problem to solve, you may consider
using a network at your college or office. You will have to refer to the books
given below for the exact details of how such a setup may be established.
Many tutorials as well as books are available to help you. Below is a list of 
the material I referred.
<P>
<OL>
<LI><EM>PVM: Parallel Virtual Machine - A User's Guide and Tutorial for
Networked Parallel Computing</EM>, Al Geist, Adam Beguelin, 
Jack Dongarra, Robert Manchek, Weicheng Jiang and Vaidy Sunderam,
MIT Press. Available at 
<A HREF="http://www.netlib.org">www.netlib.org</A></LI>
<LI> <EM>MPI: The Complete Reference</EM>, Marc Snir, Steve Otto, 
Steven Huss-Lederman, David Waker and Jack Dongarra, MIT Press.
Available at 
<A HREF="http://www.netlib.org">www.netlib.org</A>.</LI>
<LI>  <EM>RS/6000 SP: Practical MPI Programming</EM>,Yukiya Aoyama and Jan 
Nakano, International Techical Support Organization, IBM Corporation,
<A HREF="http://www.redbooks.ibm.com">www.redbooks.ibm.com</A>.
</LI>
<LI> <EM>A Beginner's Guide to PVM Parallel Virtual Machine</EM>, Clay 
Breshears and Asim YarKhan, Joint 
Institute of Computational Science, University of Tennessee, USA.
<A HREF="http://www-jics.cs.utk.edu/PVM/pvm/_guide.html">www-jics.cs.utk.edu/PVM/pvm/_guide.html</A>.</LI>
<LI> <EM>PVM: An Introduction to Parallel Virtual Machine</EM>, Emily Angerer 
Crawford, Office of Information Technology, High Performance Computing,
<A HREF="http://www.hpc.gatech.edu/seminar/pvm.html">www.hpc.gatech.edu/seminar/pvm.html</A>.</LI>
</OL>

<P>
<H2>6. Acknowlegements</H2>
<P>
I would like to thank my project guide <em>Dr. Uday Khedker</em> for his
encouragement and help. I would like to thank the <em>Center for Developement
of Advanced Computing</em> for allowing me to run the MPI and PVM programs on
the PARAM Supercomputer and <em>Dr. Anabarsu</em> for guiding me during the
implementation.





<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Rahul U. Joshi.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Web Portals Bank on Open-Source Infrastructure</font></H1>
<H4>By <a href="mailto:ned@greatbridge.com">Ned Lilly</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<i>Wireless
Developer Network (WDN) and GeoCommunity say sayonara to the database big
boys.</i>

<p>When a single company formed to operate two web portals
for online communities, they turned to the biggest names in the business to
build their technology infrastructure. 
<A HREF="http://www.wirelessdevnet.com">Wireless Developer Network (WDN)</A>
for wireless communications professionals, and its sister site, 
<A HREF="http://www.geoComm.com">GeoCommunity</A>
for geographic information specialists, used hosting services to license the
use of Microsoft and Oracle respectively. The two sites were up and running in
no time, offering virtual homes to thousands of professionals who needed
industry news, software downloads, product reviews and live chats. But while
the portals seemed a success on the surface, disaster lurked not far below.</p>

<p>Within months, WDN ran into performance problems with
SQL Server and security issues with MS IIS. Oracle 8i worked like a dream for
the GeoCommunity, but the licensing fees threatened to crush the small
company. As the portal expanded capacity and users, it knew Oracle's
aggressive pricing structures would cut even deeper into its slim profit
margin. That's when the company's technical staff began to push a radical
concept: ditch the big boys in favor of a single open-source technology
platform for both web portals.</p>

<p>The web portals had to be able to serve web pages up
24/7 without any crashes or service interruptions. The developers wanted to go
with Linux because their experiences told them it was a stable and reliable
platform for the web. They also knew firsthand that Apache web servers were a
superior product, faster, more scalable and easier to configure, and optimized
for the Linux platform.</p>

<p>The technical staff had worked in both open-source and
proprietary environments, and had come to believe open source was the more
secure choice for web-based applications. In their view, open-source
technologies grew up on the web, while most proprietary applications were
later adapted to it. Throughout the open-source development process,
developers drill down on security and performance issues for web applications.
With hundreds of developers and users testing and tweaking the programs,
security holes are often caught and corrected at an early stage of the
development process.</p>

<p>WDN's leadership quickly bought into the idea.
However, many managers believe that "nothing good can be free" and
subscribe to the common myth that open-source products lack professional
technical support. But this company's senior management team was quickly
sold on their developers' enthusiasm and positive experiences with open
source.</p>

<p>In fact, managers skittish about defying convention
need only look around for evidence of the proliferation of open-source
applications and tools in business. Linux powers an estimated 36 percent of
Internet-connected servers today, while Apache web servers are on about 61
percent of public web sites, according to the Internet research firm, Netcraft.
Industry analysts at Forrester Research recently identified open source as a
powerful growing trend in business with potential to radically reshape the
software industry by 2004.</p>

<p>Yet the database market, the core of web-based
businesses today, still remains firmly in the grip of proprietary vendors such
as Oracle, Microsoft and IBM. But in recent years, open-source databases such
as PostgreSQL and MySQL have evolved to the point where they're beginning to
compete with the propriety giants in performance and functionality. They're
attracting skilled development and user communities, as well as enterprise
business users across a wide range of industries, and are rising up to
challenge the proprietary status quo in the competitive database market.</p>

<H2>The heart of the portals</H2>

<p>For WDN and GeoCommunity, the most difficult decision
related to its technology infrastructure was its choice of the right database
management system. The portals needed a system that was scalable and
functional enough to handle thousands of visitors each month and power scores
of dynamic applications, including e-commerce.</p>

<p>The portals tested two of the most widely used open
source databases, 
<A HREF="http://www.mysql.org/">MySQL</A> and 
<A HREF="http://www.postgresql.org/">PostgreSQL</A>. While MySQL was simple to
configure and use, it lacked the transaction support and scalability that the
company needed to run their highly interactive sites. MySQL has attracted a
large user base, but the staff thought it seemed more suitable for
lower-traffic web sites. The staff also ran rigorous tests on PostgreSQL, a
heavy-duty object-relational database. It withstood the barrage of tests
without flinching, supporting advanced features during heavy simulated
transactions very well.</p>

<p>After downloading their selection of open-source
applications, including the Red Hat Linux operating system, Apache web server
and PostgreSQL database, the technical staff configured the system in less
than a half hour. The portals run 12 servers, with PostgreSQL powering dynamic
applications such as book sales, message boards and mailing lists. The new
system was up and running in minutes, with no interruptions, and neither web
portal has since crashed or lost data.</p>

<H2>Confronting the FUD (fear, uncertainty and doubt) factor</H2>

<p>Yet there are good reasons why open-source technologies
were once the exclusive domain of skilled hackers and expert users. In the
past, these applications purposefully lacked the bells and whistles of their
proprietary competition and were difficult for the less advanced user even to
install. Open-source programs have become much more user-friendly over time
because independent developers have begun to pay greater attention to
improving tools, additional features and perhaps most importantly,
documentation.</p>

<p>For WDN and GeoCommunity, the decision to migrate from a
proprietary to an open-source system was less difficult than for most
traditional businesses. The web portals employ technical staff with experience
in both environments. At every level, the company embraced the idea of
adopting a more flexible, and less financially draining, open-source
alternative. They understood the open-source development model and bought into
its underlying philosophy. Just as importantly, they possessed the technical
skills to confront many of the issues that could arise in an open-source
platform. In fact, with access to their new system's code, they could now
even modify their software's features to better fit the company's needs.

<p>Many e-businesses like these web portals, along with
brick and mortar retailers that are moving into e-commerce, have similar
needs, but lack the background and technical expertise to easily integrate
open-source technologies or migrate to a fully open-source platform. These
businesses simply want web sites that their customers and vendors can use
without difficulty. They need database-enabled web applications with 24/7
availability that won't crash or lose data--even with thousands of daily
transactions. They want a site that always works, convenient ways for
customers to buy their products, and secure methods through which to bring in
their money. Because database applications are so critical to their mission,
many businesses adopt well-known proprietary systems, feeling confident these
companies will deliver quality and reliability. Yet the rising costs, the
uncertain economy, and in some cases, the surprisingly unpredictable
performance of commercial applications, all have begun to spark greater
interest in open-source technologies today.

<P>Still, these businesses are understandably skeptical
about open source. They're used to the proprietary business model, and
can't quite fathom why good software applications would be available to
download for free from the Internet. The fact that these applications are not
owned by a corporation causes suspicion and concern; if no single vendor owns
it, people assume the software is not secure, powerful or reliable, and that
it lacks accessible support and services. And the idea of thousands of
independent developers around the world collaborating to create free software
strikes many business managers as chaotic, which makes them even more
reluctant to trust the results.

<P>Slowly, these businesses are becoming educated about the
open-source development model, which evolved not to make money, but to produce
functional software efficiently. They're finding that while the development
process varies for each open-source application today, the best projects most
often attract a global community of highly skilled developers. And it's
becoming clearer that these systematic meritocracies encourage rigorous
testing and rapid development rates, and result in fewer bugs and security
holes and more frequent releases of new and improved features.

<P>A growing number of e-businesses such as WDN and the GeoCommunity are
building their businesses on open-source platforms. These web portals have
found that the software's fast-evolving development cycles, its lower costs,
and its customizability make it ideal in their high-growth, quick-changing
industries. The lower overall cost of open-source software is attractive to
small and mid-sized businesses like these, who often have to spend
thousands--even hundreds of thousands--of dollars on purchasing or licensing
proprietary applications alone. 

<P>Another
important advantage is that its code is open and modifiable. The open-source
model rests on the belief that software develops faster and better when its
source code is accessible to all skilled developers. Mature open-source
technologies such as Apache, Linux, and PostgreSQL have thrived under the
principles of open collaboration. Similarly, businesses that employ open
source technologies can benefit both from open source's accelerated
development model and free access to its internal code, which enables them to
modify the code as needed. Open-source technology is highly conducive to
innovation, and ensures that most of the applications it produces improve
continuously and quickly. Companies that use it usually find that their
software programs evolve as quickly as their businesses do.

<H2>Why isn't everyone using open source?</H2>

<P>The perceived lack of professional support services
for open-source software remains the stumbling block to its widespread use in
business and industry. Business managers want to be able to call a service
center when problems arise. In the past, those experiencing problems with 
open-source applications could send out an e-mail and usually within hours
receive the right solution from the developers themselves. These informal
networks of technical support can provide the highest possible levels of
support, but they are not always immediately available, nor can they scale to
meet the growing demands.

<p>The issue of technical support was important to WDN and
GeoCommunity because they knew they would occasionally need a technical safety
net. They purposely chose applications with strong development communities in
order to get the help they need directly from their web sites. Their technical
staff concedes that companies without their own in-house technical staff need
more comprehensive support. It's the one issue that continues to scare
managers away from open source.</p>

<p>Fortunately, entrepreneurs always rush to fill a vacuum.
Red Hat was one of the first companies to provide support for the Linux
operating system, and now a slew of other companies are springing up to
provide support, training and consulting services for some of the best open
source database applications. The open-source support gap is shrinking fast,
which is good news for emerging companies in need of an affordable platform
for their growing business.</p>

<P>In
the meantime, the WDN and GeoCommunity remain satisfied with their open-source
decision, and the reliability, error reporting, community support, clean
designs and strict adherence to industry standards that came with it. It's
also refreshing that with open source, there can be no effort to force
"lock-in" or add proprietary hooks that will prevent transition to other
products in the future. As one senior staff member said, "It's truly been
a liberating experience to use good products that were designed simply to meet
a need--not to further a corporate agenda."




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Ned Lilly.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Finding my computer at home from the outside</font></H1>
<H4>By <a href="mailto:articles@gnujobs.com">Mark Nielsen</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<ol>
<li>
<a href="#Introduction">Introduction</a></li>
<li><a href="#perl">Perl script uploading ip address.</a></li>
<li><a href="#apache">Webpage and perl script on remote computer.</a></li>
<li> <a href="#cron"> Cron job I run in the background.</a></li>
<li>
<a href="#Conclusion">Conclusion</a></li>
<li>
<a href="#REF">References</a></li>
</ol>


<h3>
<a NAME="Introduction"></a>Introduction</h3>
The purpose of this article is to make it so I can find my computer at
home when I am traveling around the Bay Area doing computer work, recruiting, 
and volunteer work. Most of the time, I am busy traveling around, 
although I am able to work from home half the time now. 
My computer at home uses a Ricochet modem. The dumb people who promised
me a good DSL connection and a satellite connection where I live were
a bunch of morons. The max DSL I could get would be 144k (which I found out
AFTER I moved in), which is pointless
when I already have a Ricochet modem at 128k. Plus, I am facing the wrong way
for a satellite connection. Whatever you do, make sure the morons who sell
you their apartments have it in the contract that you are promised certain
speed connections to the internet, or you can break the contract with no
penalty. As soon as it is worth, I am moving. For now, I am stuck with
a dial-up connection, which isn't bad most of the time. 
<p>
Some people can have static DSL connections, which takes of the problem I 
have, which is my ip address to the internet changes each time I dial up. 
I used to email myself the ip address, parse out the data, and put it on a
webpage. I have a better solution now. I use ssh to transfer a file to my 
remote web server once an hour. 

<h3>
<a NAME="ssh"></a>
Setting up ssh. 
</h3>
The version of ssh I am using is 1.2.27. I should be using OpenSSH, but for
now, I am using commercial ssh. 
<p>
We need to make it so we can transfer files securely from my computer at
home to the remote computer. We use the ssh-keygen program (which comes
with ssh). Here is a paragraph from the manpage for ssh.
<BLOCKQUOTE>
       Ssh  implements  the RSA authentication protocol automati
       cally.  The user creates his/her RSA key pair  by  running
       ssh-keygen(1).   This stores the private key in .ssh/iden
       tity and the public key in .ssh/identity.pub in the user's
       home  directory.   The  user  should  then  copy the iden
       tity.pub to .ssh/authorized_keys in his/her home directory
       on  the  remote  machine  (the authorized_keys file corre
       sponds to the conventional .rhosts file, and has  one  key
       per line, though the lines can be very long).  After this,
       the user can log in  without  giving  the  password.   RSA
       authentication is much more secure than rhosts authentica
       tion.
</BLOCKQUOTE>

So I ran "ssh-keygen" as a user on my computer at home. 
Then I transferred the ".ssh/identity.pub" file on my computer at home to 
the remote computer as ".ssh/authorized_keys" for the user "web1"
on the remote computer. This makes it so I can login in from home to my 
remote computer without having to use a password. This can also be used
to transfer files. 
<pre>
rsync -e ssh -av /home/test1/IP.txt web1@somecomputer.com:public_html/IP.txt
</pre>

<a NAME="perl"></a>
<h3>Perl script uploading ip address.
</h3>

Here is the perl script I use to upload the ip address. You should
change values of the usernames and remote computer address. 
<a href="misc/nielsen/Remote_Website.pl.txt">[Text version of this script.]</a>
<pre>
#!/usr/bin/perl

use strict;

  ### Run ifconfig and store the data in the @Temp list. 
my @Temp = `/sbin/ifconfig`;

  #### Search for ppp
my $Search = "ppp";
  ### If you are looking for the ip address of your ethernet card, 
  ### uncomment the next line;
# $Search = "eth0";

  ### Make the line we find the ip address blank initially.
my $Match_Line = "";
my $Match_Device = "no";

  ## Search through the lines, if we find a match, save the lines until
  ## we find a blank line. 

foreach my $Line (@Temp)
  {
    ### If we have a match, abort. 
  if ($Match_Line ne "")   {@Temp = ();}
    ### else, see if we can find a match at the beginning of line;
  elsif ($Line =~ /^$Search/) {$Match_Device = "yes";}
    ### else, if we found the device, and we find the line we are looking for
  elsif (($Match_Device eq "yes") &amp;&amp; ($Line =~ /^ +inet/)) 
    {$Match_Line = $Line;}  
  }

  ## If our $Match_Line is not blank, split it and get the ip address.
my $IP = "";
if ($Match_Line ne "") 
   {
    ### Get rid of stuff before addr:
   my ($Junk,$Good) = split(/addr\:/, $Match_Line,2);
    ### Get rid of stuff after the first space
   my ($Good,$Junk) = split(/ /, $Good,2);
   $IP = $Good;
   }

  ## If $IP is not blank, we have something. Save to file and transfer file
  ## to remote site. 
  ### Please don't use the /tmp to store this file, but some other location.
if ($IP ne "")
  {  
  open(FILE,"&gt;/tmp/IP.txt");
  print FILE "$IP\n";
  close FILE;
  system ('rsync -av -e ssh /tmp/IP.txt web1@somecomputer.com:public_html/IP.txt');
  }
   ### Else, we should send ourselves an email, or do something
   ### to let us know it didn't work. This is left as an exercise.
else {}
</pre>

<h3>
<a NAME="apache"></a>
Webpage and perl script on remote computer.
</h3>
On the remote computer storing the ip address, we need to detect if
it is an hour old. If it is less than an hour old, we should print out
an error message. So I use this perl script. I name it 
"/home/web1/public_html/IP.pl".
<A HREF="misc/nielsen/IP.pl.txt">[Text version of this listing.]</A>

<pre>
#!/usr/bin/perl

use strict;

print "Content-type: text/html\n\n\n\n";

my $File = "/home/web1/public_html/IP.txt";
open(FILE,"/home/web1/public_html/IP.txt");
my $Line = &lt;FILE&gt;;
chomp $Line;
close FILE;

my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
   $atime,$mtime,$ctime,$blksize,$blocks)
     = stat($File);
my $time = time();

print "&lt;br&gt; Last known ip address was $Line\n";
print qq(&lt;br&gt; To transfer to the website, 
     &lt;a href="http://$Line"&gt;click here&lt;/a&gt;\n);

my $Diff = $time - $mtime;
if ($Diff &gt; 4000) 
  {
  print "&lt;p&gt;ERROR: The ip address should have been updated once an hour, 
  but 4000 seconds has past since the last update.
  &lt;br&gt; $time - $mtime = $Diff \n";
  }
</pre>

<p>
You may want to consider moving this perl script into the normal cgi-bin 
directory of your web server. Otherwise, here is a dangerous example of how to 
make it so you can run perl scripts from a user's directory. THIS IS DANGEROUS!
If your web server allows any user to execute a perl script, 
that person can get the web server to do anything they want. 
<p>To make it so you can execute perl scripts on your web server, 
<pre>

&lt;Directory /home/*/public_html&gt;
   ## Options All is reduntant with some of the other options. 
    Options All Indexes FollowSymLinks MultiViews ExecCGI Includes 
    AllowOverride All
    Order allow,deny
    Allow from all
&lt;/Directory&gt;

   #### This requires several perl apache modules
 &lt;Files *.pl&gt;
 SetHandler perl-script
 PerlHandler Apache::OutputChain Apache::SSIChain Apache::Registry 
 PerlSendHeader On
 Options ExecCGI
 &lt;/Files&gt;
</pre>


<h3>
<a NAME="cron"></a>The Cron entry to make it run nightly</h3>
Put this in your crontab on the remote server using the "crontab -e" command.

<pre>
#/bin/sh

  ### Download every two hours
1 * * * *   /www/Cron/Remote_Website.pl &gt;&gt; /www/Cron/out  2&gt;&amp;1  
</pre>

<h3>
<a NAME="Conclusion"></a>Conclusion</h3>
I know people are probably doing the same thing in different ways. 
I like this solution because the files are transferred securely. 
This makes it so people can't see what I am transferring over the
internet. So that nobody can get to the file, we should password
protect the webpage and perl script that display the ip address. 

<h3>
<a NAME="REF"></a>References</h3>

<ol>
<li> <a href="http://www.ssh.com/products/ssh/download.html">ssh</a>
<li> <a href="http://www.openssh.com/ftp.html">OpenSSH</a>
<li> <a href="http://httpd.apache.org/dist/">Apache</a>
<li>
If this article
changes, it will be available here
<a href="http://www.gnujobs.com/Articles/17/Remote_Website.html">
http://www.gnujobs.com/Articles/17/Remote_Website.html</a></li>
</ol>

<p>
<i> Mark works as an independent consultant donating time to causes like
GNUJobs.com, writing articles, writing free software, and working 
as a volunteer at <a href="http://www.eastmont.net">eastmont.net</a>.</i>






<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Mark Nielsen.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Learning Perl, part 3</font></H1>
<H4>By <a href="mailto:ben-fuzzybear@yahoo.com">Ben Okopnik</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<br><tt>The trouble with teaching Perl as a first computer language is
that your students won't appreciate it till they start learning their second.
The trouble with teaching Perl as a second language is that there's no
single suitable first language to go in front.</tt>
<br><tt>&nbsp;-- Larry Wall</tt>
<p><tt>When they say that Perl is a `glue language', what they really mean
is that it is good for cleaning up after the mistakes of other programs.</tt>
<br><tt>&nbsp;-- Mark-Jason Dominus in comp.lang.perl.misc</tt>
<br>&nbsp;
<br>&nbsp;
<p><b>Overview</b>
<p>This month, we'll look at Perl's conditional and looping constructs,
and look at a few scripts that use them. We will also explore how they
work with Perl's variables, and take a quick look at capturing user input.
Once you understand this part, I suggest hacking out a couple of experimental
scripts and playing with them; sure, you'll make mistakes - but from this
point on, you'll actually need to supplement your reading by getting down
and dirty. If you don't play, you can't win...
<br>&nbsp;
<br>&nbsp;
<p><b>Conditionals</b>
<p>Here are the conditional statements that Perl uses; nothing particularly
unusual, if you're used to conditionals in other languages. Perl checks
if the condition is true or false, and branches the execution based on
that.
<p>
<HR NOSHADE WIDTH="100%">
<br><tt>if&nbsp;&nbsp;&nbsp; ( traffic_light_is_red ) {&nbsp;&nbsp;&nbsp;&nbsp;
# If condition 1 is true, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; stop;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Action 1</tt>
<br><tt>}</tt>
<br><tt>elsif ( traffic_light_is_yellow ) {&nbsp; # If condition 2 is true,
do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; hit_the_gas;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Action 2</tt>
<br><tt>}</tt>
<br><tt>else&nbsp; {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# In all other cases, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; proceed_with_caution;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Action 3</tt>
<br><tt>}</tt>
<p>Note that the "elsif" clause isn't required; neither is the "else".
Also, note that "else" is the 'catch-all option': if the light is anything
except red or yellow - burned out, just got knocked down by accident, etc.
- the action is '<tt>proceed_with_caution</tt>'.
<p>Unlike C, even single actions must be enclosed in a block (defined by
the curly brackets):
<p><tt>if ( $tomato eq "red" )&nbsp;&nbsp; print "Ripe.\n";&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# WRONG!</tt>
<br><tt>if ( $tomato eq "red" ) { print "Ripe.\n"; }&nbsp;&nbsp;&nbsp;
# Right</tt>
<p>
<HR NOSHADE WIDTH="100%">
<p><tt>unless ( $blarg == $foo ) {&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# If condition 1 is false, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "Unequal!.\n";&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Action 1</tt>
<br><tt>}</tt>
<br><tt>else&nbsp;&nbsp; {&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Otherwise, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "They're equal.\n";&nbsp;&nbsp;&nbsp;&nbsp;
# Action 2</tt>
<br><tt>}</tt>
<p>Pretty obvious. It may help to think of "unless" as the "if not" conditional.
Once again, the "else" is optional. No, there's no such thing as "elseunless".
:)
<br>&nbsp;
<br>&nbsp;
<p><b>Loops</b>
<p>Ah, wonderful loops. These are the things that make actions happen,
as many times as we want, based on a condition. You might even say that
this loops are the main reasons for computers in general, their main use
as the tool that they are: precise repetitive work. Here are the three
most
<br>common types of loops under Perl:
<p>
<HR NOSHADE WIDTH="100%">
<br><tt>while ( $cat eq "away" ) {&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# While cond. 1 is true, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "The mice will play.\n";&nbsp;&nbsp;
# Action 1</tt>
<br><tt>}</tt>
<p>
<HR NOSHADE WIDTH="100%">
<br><tt>until ( $time &gt; 1159 ) {&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# While cond. 1 is false, do</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp; print "It's morning.\n"&nbsp;&nbsp;&nbsp;
# Action 1</tt>
<br><tt>}</tt>
<p>
<HR NOSHADE WIDTH="100%">
<br>The "for" loop can be implemented in two different ways - one is like
the "for" loop in C:
<p><tt>for ( $n = 99; $n &gt; 0; $n-- ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp; print "$n bottles of beer on the wall, $n bottles
of beer,";</tt>
<br><tt>&nbsp;&nbsp;&nbsp; ...</tt>
<br><tt>}</tt>
<p>In this case, we set $n to an initial value (99), decrement it by 1
each time we go through the loop, and check to make sure that it's greater
than 0. If it's not, we exit the loop.
<p>The second method, somewhat like the Clipper, FoxPro, etc. "foreach"
loops, is by far the most common:
<p><tt>foreach $n ( 0..1000 ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "Day $n on this
deserted island. So far, I've had ";</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print $n * 100, " bananas.
I hope I'm rescued soon.\n";</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ...</tt>
<br><tt>}</tt>
<p>It can also be used this way:
<p><tt>for ( 0..1000 ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp; print "Day $_ on this deserted island. So far,
I've had ";</tt>
<br><tt>&nbsp;&nbsp;&nbsp; print $_ * 100, " bananas. I hope I'm rescued
soon.\n";</tt>
<br><tt>&nbsp;&nbsp;&nbsp; ...</tt>
<br><tt>}</tt>
<p>Our old friend, the "$_" (explained in 
<A HREF="../issue64/okopnik.html">the previous part of this series</A>.)
He does indeed come in handy. Note that "foreach" is just an alias for
"for", and they can be used interchangeably.
<p>
<HR NOSHADE WIDTH="100%">
<p>All of the above conditionals and loops can also be used as single-statement
modifiers, as well:
<p><tt>print "This is line $_ of 50.\n" for ( 1..50 );</tt>
<p>The above will print 50 lines, numbered in an obvious way.
<p><tt>print "I've found him!" if /Waldo/;</tt>
<p>The above line will be printed if the default buffer ($_) contains a
match for "Waldo".
<br>&nbsp;
<p>An interesting fact that combines well with loops and conditionals is
that empty variables in Perl return a null value - which is "false". This
is perfect for checking them out:
<p><tt>print if $_;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Prints $_ if it contains anything</tt>

<P> The next example shows that a zero value is also false:
<PRE>
print "5280 is true.\n" if 5280;   # This will print.
print "0 is true.\n" if 0;         # This won't print.
</PRE>

<P> Here's an example with a list:

<p><tt>while ( @a ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print pop @a;&nbsp;&nbsp;&nbsp;&nbsp;
# "Pop" the last value off @a and print it</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; $count =&nbsp; @a;&nbsp;&nbsp;&nbsp;&nbsp;
# Get the number of elements in @a</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print $count, " elements left in
\@a.\n";</tt>
<br><tt>}</tt>
<p>When the last element has been popped off, the loop will end.
<p><tt>unless ( %hash ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; %hash = ( 'first' =&gt;&nbsp;
'Mighty Joe',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'last'&nbsp; =&gt;&nbsp; 'Young',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'type'&nbsp; =&gt;&nbsp; 'gorilla',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'from'&nbsp; =&gt;&nbsp; 'Pangani Mountains',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'born'&nbsp; =&gt;&nbsp; '1949',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'Mom'&nbsp;&nbsp; =&gt;&nbsp; 'Jill',</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
'Dad'&nbsp;&nbsp; =&gt;&nbsp; 'Gregg'</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; );</tt>
<br><tt>}</tt>
<p>If "%hash" is empty, we populate it with some initial values.
<br>&nbsp;
<p>The range operator, which we've used a couple of times so far, is a
useful widget: it allows you to specify a range of numbers or letters.
Note that the ranges have to be of the same 'kind' - if you specify ('a'..'Z')
or ('A'..'z'), the output will not be what you expect. Also, you cannot
specify ('z'..'a'); that won't work either. However, there is an easy way
to do that:
<p><tt>foreach $letter ( reverse 'a'..'z' ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp; print "$letter\n";</tt>
<br><tt>}</tt>
<p>It will also properly increment "letter lists":
<p><tt>for ( 'aa'..'zz' ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp; print "$_ ";&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Will print "aa ab ac ... zx zy zz"</tt>
<br><tt>}</tt>
<br>&nbsp;
<br>&nbsp;
<p><b>User Input</b>
<p>Capturing keyboard input, or input from STDIN in general - such as the
lines piped to the input of our script via something like
<p><tt>cat file | perl_script</tt>
<p>&nbsp;- is easy; it's what Perl's "diamond operator" is for.
<br>&nbsp;
<p><tt>while ( &lt;&gt; ) {&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; # Capture
all keyboard or piped input</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# Print each line as long as input exists</tt>
<br><tt>}</tt>
<p>The above works exactly like "cat" - it will print all input piped to
it, will "cat" a file if it's run with the filename used as an argument,
and will accept (and echo) user input until you hit Ctrl-D or Ctrl-C. It
can also be written this way:
<p><tt>print while &lt;&gt;;</tt>
<p>for a more "Perlish" syntax. Note that "&lt;&gt;" and "&lt;STDIN&gt;" are
related but not equivalent:
<p><tt>print while &lt;STDIN&gt;;</tt>
<p>will respond to keyboard and piped input, but <b>will not</b> print
the contents of a file supplied as an argument. I've never found a situation
where I needed that kind of functionality, so I simply use "&lt;&gt;".
<p>If you want to assign user input to a variable, Perl also makes that
easy - but there's a bit of a trap built in of which you need to be aware:
<p><tt>$answer = &lt;&gt;;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; # Get
the input, assign it to the variable</tt>
<br><tt>if&nbsp;&nbsp;&nbsp; ( $answer eq "y" ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "Yes\n";</tt>
<br><tt>}</tt>
<br><tt>elsif ( $answer eq "n" ) {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "No\n";</tt>
<br><tt>}</tt>
<br><tt>else {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "No idea!\n";</tt>
<br><tt>}</tt>
<p>The above script will always print "No idea!" Hmm... it <i>looks</i>
right; what could be the problem?
<p>The problem is that Perl captures <b>everything</b> that you give it.
So, when you type "y", what's the next key you hit? "Enter", that's what!
So, the variable stored in $answer is NOT "y", it's "y\n" - the answer
<i>and</i> the linefeed. How do we deal with that? Perl, of course, has
a function - one you should <b>always</b> use when getting user input:
<p><tt>chomp ( $answer = &lt;&gt; );</tt>
<p>"chomp" will remove the linefeed, or "end-of-line" character, from the
string to which it is applied. It will also remove EOLs from every element
of an array which it receives as an argument. The old Perl4 version, "chop",
removed the last character from a scalar (or from the elements of the array)
no matter what it was; it's still available if you should need it for that
purpose, but for taking user input, use "chomp" (also known, via Perl's
error messages, as the "safe chop").
<br>&nbsp;
<br>&nbsp;
<p><b>Exercises For The Mind</b>
<p>Try building a couple of scripts, just for your own education and entertainment:
<p>A script that takes a number as input, and prints "Hello!" that many
times. As a bonus, check the input for illegal (non-numeric) characters
(hint: use //, the match operator.)
<p>A script that takes the current hour (0-23) as input and says "Good
morning", "Dobriy den'", "Guten Abend", or "Buenas noches" as a result.
&lt;grin&gt;
<p>If you come up with something particularly clever, don't hesitate to
send it to me for the next part of this series: you'll get the credit for
writing it, I'll happily dissect it for you, and we'll both become micro-famous
and retire to Belize on the proceeds. &lt;laugh&gt;
<p>Don't forget: your shebang line should always contain "-w". If you don't
ask Perl to help you with your mistakes, you'll be wasting a lot of time.
Let the computer do the hard work!
<br>&nbsp;
<p><tt>#!/usr/bin/perl -w</tt>
<br><tt>print "See you next month!"</tt>
<br>&nbsp;
<p>Ben Okopnik
<br><tt>perl -we'print reverse split//,"rekcah lreP rehtona tsuJ"'</tt>
<br>
<HR NOSHADE WIDTH="100%"><tt>References:</tt><tt></tt>
<p><tt>Relevant Perl man pages (available on any pro-Perl-y configured</tt>
<br><tt>system):</tt><tt></tt>
<p><tt>perl&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - overview&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
perlfaq&nbsp;&nbsp; - Perl FAQ</tt>
<br><tt>perltoc&nbsp;&nbsp; - doc TOC&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
perldata&nbsp; - data structures</tt>
<br><tt>perlsyn&nbsp;&nbsp; - syntax&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
perlop&nbsp;&nbsp;&nbsp; - operators/precedence</tt>
<br><tt>perlrun&nbsp;&nbsp; - execution&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
perlfunc&nbsp; - builtin functions</tt>
<br><tt>perltrap&nbsp; - traps for the unwary&nbsp; perlstyle - style guide</tt><tt></tt>
<p><tt>"perldoc", "perldoc -q" and "perldoc -f"</tt>




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Ben Okopnik.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">So You Like Color !!!<BR>
(The mysterious ^[[ characters)</font></H1>
<H4>By <a href="mailto:p_padala@yahoo.com">Pradeep Padala</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




Have you ever redirected the output of a curses program with colors and wondered
what those mysterious ^[[ are? Did you
ever try to produce colors with a printf without using curses? If the 
answer to either of these questions is yes, read on... 
</P>
This article attempts to explain those mysterious characters that one finds in 
the output of a curses program which produces colors. Later on, we extend
the concept to produce colors with a mere printf.

</P>
<H3> Terminal Codes </H3> 
<P>
In the olden days of teletype terminals, terminals were away from 
computers and were connected to them through serial cables. The terminals could
be configured by sending a series of bytes to each of them. All the capabilities of
terminals could be accessed through these series of bytes which are usually called escape
sequences because they start with an escape(0x1B) character. Even today with vt100
emulation, we can send escape sequences to the emulator and it will have the
same effect on the terminal window.  Hence, in order to print
color, we merely echo a control code.
</P>
Type this on your console.
<PRE>
	echo "^[[0;31;40mIn Color"
</PRE>
<P>
The first character is an escape character, which looks like two characters
^ and [. To be able to print that you have to press CTRL+V and then the ESC key. All the others are
normal printable characters.
You see the string "In Color" in red. It stays that way and to revert back type
this
</P>
<PRE>
	echo "^[[0;37;40m"
</PRE>
<P>
As you can see it's pretty easy to set color and reset it back. There are 
a myriad of escape sequences with which you can do a lot of things like moving
the cursor, resetting the terminal etc..
</P>
<H3> The Color Code: &nbsp;&nbsp;&nbsp;&nbsp;&lt;ESC&gt;[{attr};{fg};{bg}m </H3>
<P>
I'll explain the escape sequence to produce colors. The sequence to be printed
or echoed to the terminal is 
<PRE>
	&lt;ESC&gt;[{attr};{fg};{bg}m
</PRE>
<P>
The first character is ESC which has to be printed by pressing CTRL+V and then 
ESC on the Linux console or in xterm, konsole, kvt, etc.  ("CTRL+V ESC" is also
the way to embed an escape character in a document in vim.)  Then {attr}, {fg},
{bg} have to be replaced with the correct value to get the corresponding
effect. attr is the attribute like blinking or underlined etc.. fg and bg are
foreground and background colors
respectively. You don't have to put braces around the number. Just writing the  
number will suffice. 
</P>
<P>
{attr} is one of following
<PRE>
	0	Reset All Attributes (return to normal mode)
	1	Bright (Usually turns on BOLD)
	2 	Dim
	3	Underline
	5	Blink
	7 	Reverse
	8	Hidden
</PRE>
{fg} is one of the following
<PRE>
	30	Black
	31	Red
	32	Green
	33	Yellow
	34	Blue
	35	Magenta
	36	Cyan
	37	White
</PRE>
{bg} is one of the following
<PRE>
	40	Black
	41	Red
	42	Green
	43	Yellow
	44	Blue
	45	Magenta
	46	Cyan
	47	White
</PRE>
<P>
So to get a blinking line with Blue foreground and Green background, the combination to be used should
be
<PRE>	
echo "^[[5;34;42mIn color"
</PRE>
which actually is very ugly. :-)  Revert back with 
<PRE>
echo "^[0;37;40m"
</PRE>


<H3> With printf() </H3>
<P>
What if you want to use this functionality in a C program? Simple! Before you printf
something print this escape sequence to produce it in the desired color. I 
have written a small routine <CODE>textcolor()</CODE> which does this 
automatically. You can use it in your programs along with the #define constants. 
The text version of this program is <A HREF=misc/padala/color.c.txt> here </A>
</P>
<H3><CODE>textcolor()</CODE> </H3>
<TABLE BGCOLOR="#C5C5C5" WIDTH=450>
<TR> <TD>
<PRE>
#include &lt;stdio.h&gt;

#define RESET		0
#define BRIGHT 		1
#define DIM		2
#define UNDERLINE 	3
#define BLINK		4
#define REVERSE		7
#define HIDDEN		8

#define BLACK 		0
#define RED		1
#define GREEN		2
#define YELLOW		3
#define BLUE		4
#define MAGENTA		5
#define CYAN		6
#define	WHITE		7

void textcolor(int attr, int fg, int bg);
int main()
{	textcolor(BRIGHT, RED, BLACK);	
	printf("In color\n");
	textcolor(RESET, WHITE, BLACK);	
	return 0;
}

void textcolor(int attr, int fg, int bg)
{	char command[13];

	/* Command is the control command to the terminal */
	sprintf(command, "%c[%d;%d;%dm", 0x1B, attr, fg + 30, bg + 40);
	printf("%s", command);
}
</PRE>
</TD>
</TABLE>
<P>
The <CODE>textcolor()</CODE> is modeled against the Turbo C API function. You 
call the function to set the color and then print with a <CODE>sprintf()</CODE>
(a function used in Turbo C to produce console output in color). 
</P>
<H3> A Demo of colors</H3>
<TABLE BGCOLOR="#C5C5C5" WIDTH=450>
<TR> <TD>
<PRE>
#include &lt;stdio.h&gt;

#define RESET		0
#define BRIGHT 		1
#define DIM		2
#define UNDERLINE 	3
#define BLINK		4
#define REVERSE		7
#define HIDDEN		8

#define BLACK 		0
#define RED		1
#define GREEN		2
#define YELLOW		3
#define BLUE		4
#define MAGENTA		5
#define CYAN		6
#define	WHITE		7

#define ARRAY_SIZE(a) (sizeof(a) / sizeof(a[0]))

char *attrs[] = {"NORMAL", "BRIGHT", "DIM", "UNDERLINE", "BLINK",
		 "REVERSE", "HIDDEN", "EXIT"};
char *colors[] = {"BLACK", "RED", "GREEN", "YELLOW", "BLUE", "MAGENTA",
		 "CYAN", "WHITE", "EXIT"};
void textcolor(int attr, int fg, int bg);
int print_menu(char *array[], int n_options, char *title);
int main()
{	int attr, fg, bg;
	int attr_size, colors_size;
	
	attr_size = ARRAY_SIZE(attrs);
	colors_size = ARRAY_SIZE(colors);
	while(1)
	{	printf("\n");
		attr = print_menu(attrs, attr_size, "Choose the attr you want:");
		if(attr == attr_size - 1)
			break;
		fg = print_menu(colors, colors_size, "Choose the foreground you want:");
		if(attr == colors_size - 1)
			break;
		bg = print_menu(colors, colors_size, "Choose the background you want:");
		if(attr == colors_size - 1)
			break;
		printf("\n");
		textcolor(attr, fg, bg);	
		printf("This is what you get if you use the combination %s attribute %s foreground and %s
 background", attrs[attr], colors[fg], colors[bg]);
		textcolor(RESET, WHITE, BLACK);
		system("clear");
	}
	return 0;
}

int print_menu(char *array[], int n_options, char *title)
{	int choice, i;
	for(i = 0;i &lt; n_options; ++i)
		printf("%d.%s\n", i, array[i]);
	printf("%s", title);
	scanf("%d", &amp;choice);
	return choice;
}		
void textcolor(int attr, int fg, int bg)
{	char command[13];

	/* Command is the control command to the terminal */
	sprintf(command, "%c[%d;%d;%dm", 0x1B, attr, fg + 30, bg + 40);
	printf("%s", command);
}
</PRE>
</TD>
</TABLE>
<P> This program asks the user to play with attributes and colors and shows 
a string in that color. I usually use it to find out the best combination
of colors for my GUIs. Text version of above program is <A HREF=misc/padala/demo.c.txt> here </A>.
</P>
<H3> The Catch </H3>
<P>
Then what's the catch? If producing color is so easy, why do people waste their
time
writing huge programs in curses, which in turn query terminfo in a
complex way? As we know, there are many terminals with very few capabilities
and terminals which don't recognize these escape codes or need different codes
to achieve the same effect. So if you want a portable program which would run
on any terminal with the same (or reduced) functionality, you should use
curses. Curses uses terminfo to find the correct codes to
accomplish the task in style. Terminfo is a big database
which contains information about the various functionalities of different
terminals.
</P>
<P>
But if you just want to write a simple program which produces color on a Linux 
console or xterm window, you can just use the escape sequences above to do
it easily. The Linux console mostly emulates vt100, so it recognizes these
escape sequences. 
</P>
<H3> With tput </H3>
<P>
But there is a way to query the terminfo database and do the work. tput is the
command which queries the database and executes the functionality you specify.
The two capabilities setf and setb are useful to set foreground and background
colors. Use this to set foreground color to red and background color to green.
</P>
<PRE>
	tput setf 4	# tput setf {fg color number}
	tput setb 2	# tput setb {bg color number}
</PRE>
<P>
This can be used in shell scripts where you want. See the tput manual page for
additional capabilities of tput.
The terminfo manpages contain a lot of information
regarding terminal capabilities - how to get and set their values and
more.  There are two terminfo manpages.  "man 5 terminfo" describes
the terminfo database.  
"man 3ncurses terminfo" describes the C functions that use the database.

<P> These are the color numbers to be passed as arguments to "tput setf" and
"tput setb".
</P>
<PRE>
	0	Black
	1	Red
	2	Green
	3	Yellow
	4	Blue
	5	Magenta
	6	Cyan
	7	White
</PRE>
<P>
Have fun !!!
</P>

<H3> References </H3>
<UL>
<LI>The <A HREF=http://www.linuxdoc.org/HOWTO/Text-Terminal-HOWTO.html>
	Text-Terminal-HOWTO </A> </LI>
<LI>Man pages for <A HREF=http://linux.ctyme.com/man/man2488.htm>tput</A> 
and <A HREF=http://linux.ctyme.com/man/man2346.htm> terminfo<A>. </LI>
</UL>




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Pradeep Padala.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Book Review: <EM>Networking Printing</EM></font></H1>
<H4>By <a href="mailto:dpuryear@usa.net">Dustin Puryear</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<p>Network Printing
<br>O'Reilly and Associates
<br>October 2000
<br>ISBN 0-596-00038-3
<br>$34.95

<p>There are few applications so beneficial, pervasive,
and--oftentimes--complex as network printing. Network printing is beneficial
because it reduces the number of printers required for an organization.
Allowing users to print to a limited set of printers, rather than requiring
a printer for each user realizes an obvious reduction in capital cost.
This also equates to a savings in space requirements and power consumption.
(These are two often overlooked but important factors.) The reason network
printing is so pervasive is directly a result of the benefit of network
printing--it reduces cost, both in terms of capital outlays and maintenance.

<p>Unfortunately, network printing can also be quite
complex. This is especially true for heterogeneous networks. In a heterogeneous
network, not only do administrators need to worry about printers and print
servers speaking the same lingo, but also that each device is actually
using the same network layer protocols (i.e., TCP/IP). Even when a network
is homogenous there can be difficulties, especially in large organizations
where printers number in the hundreds or thousands.

<p>In order to combat this complexity, and with it's
resulting rise in cost and overhead, an administrator needs a solid set
of documentation and a framework from which to grow. O'Reilly has attempted
to satisfy just this need with their release of "Network Printing," by
Todd Radermacher and Matthew Gast. "Network Printing," published in October
of 2000, provides a step-by-step guide for building an infrastructure to
support network printing in heterogeneous networks (and, by extension,
homogenous ones as well).

<p>So what exactly does Radermacher and Gast, the authors,
bring to the table? Both Radermacher and Gast have several years of experience
in the computer industry. They also both have a very readable writing style,
and consistently speak to the reader in the first person. (This helps to
engage the reader in the material, and often leads to more readable technical
literature.) Now, on to the book!

<p>In Chapter 1, "A Brief History of Printing and Publishing",
Network Printing begins with a general introduction to printing in general.
By "in general" I mean the entire field of printing, and not just network
printing. The authors give a quick overview of the history of printing,
including the introduction of such notables as papyrus scrolls and the
Linotype. Personally, I feel this type of material is usually best left
to the history books, but you may disagree.

<p>The second chapter, "Printer Languages," progresses
to the more relevant topic of page-description language. A page-description
language is the <i>lingua franca</i> used by a print server and a printer.
Common examples, and ones that are covered in the book, are Adobe's PostScript
and Hewlett-Packard's Print Command Language (PCL). All in all, the authors
do a good job of summarizing these languages. However, if you are looking
for in-depth coverage, you will need to go elsewhere.

<p>Chapters 3, 4, and 5 concern three popular UNIX print
systems currently in use: BSD, SysV, and LPRng. The emphasis of the book
is on using UNIX as the central print server platform for an organization,
so the concentration on these systems is important. (However, I would have
liked to see more focus on NT print servers.) Special attention is paid
to print filters, which form the core of the UNIX print process.

<p>In Part II, "Front-End Interfaces to UNIX queues," the authors begin with
the requisite chapter on Samba. Chapter 6, "Connecting Windows to UNIX Servers:
Let's Samba", describes deploying Samba on UNIX machines so that the servers
can interface with Windows networks. Certainly, this book is not the end-all
for documentation relating to Samba and it's various configuration options, but
Gast and Radermacher cover it in enough detail to get the reader up and
running.

<p>After the coverage on Samba and Windows environments,
the authors move to a more underserved support issue in many books stressing
UNIX solutions: integration with Macintosh and NetWare networks. In Chapters
7 and 8 the authors cover netatalk and ncpfs, respectively. Similar to
the Samba chapter, the authors'
main focus here is to educate the reader about the aspects of the support
software relating to printing.

<p>In Part III, "Administration," Radermacher and Gast
enter into one the more crucial aspects of network printing--effectively
and efficiently administering the system. At this point the authors assume
you have the knowledge to implement network printing for the various networks
covered, and they move to making the system not only effective but also
efficient.

<p>In Chapter 9, "Using SNMP to Manager Networked Printers,"
the authors demonstrate how to use SNMP to monitor and control your printer
infrastructure. Of note is their good overview of SNMP, and review of SNMP
agents, such as MRTG. Not the strongest chapter in the book, but more than
sufficient.

<p>Next, in Chapter 10, "Using Boot Servers for Basic
Printer Configuration," and Chapter 11, "Centralized Configuration with
LDAP," the emphasis is on methods for maintaining a centralized configuration
for all of the network printers. In small to medium networks these chapters
may not be truly useful, but for large installations, centralized configuration
is vital. The chapter on LDAP is especially informative, and offers several
insights.

<p>Finally, in Chapter 12, "Accounting, Security, and
Performance," the authors tie many loose ends left from earlier chapters.
The main point of this chapter is demonstrating the use of scripts for
accounting and monitoring and tuning server performance. The section on
security is rather small unfortunately, and I would have liked to see more
detail. Alas, it was not forthcoming.

<p>In conclusion, I think this was a rather well done
book. The authors did an excellent job of keeping a rather boring subject
(for most of us at least) somewhat upbeat. I also was quite happy to see
several rather keen insights, especially the use of LDAP to pull configurations
to print servers. If you are a network administrator that is not afraid
of Linux or UNIX and need to better organize and control your printer infrastructure
then this is an excellent resource.





<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Dustin Puryear.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<!-- END header -->




<table COLS=2 WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td NOSAVE>
<center>
<h1>
<img SRC="misc/steffler/makingSmalltalk.png" height=53 width=417>&nbsp;</h1></center>

<center>
<h3>
OO Thinking&nbsp;</h3></center>

<p>&nbsp;
<br>&nbsp;
<br>&nbsp;
<br>&nbsp;
<br>&nbsp;
<center>
<p>By <a href="mailto:jagwar@magma.ca">Jason Steffler</a></center>
</td>

<td WIDTH="200" NOSAVE>
<center><img SRC="misc/steffler/penguinInBalloon.gif" height=230 width=200></center>
</td>
</tr>
</table>

<div align=right>
<hr WIDTH="100%">
<p><b>Article #5 - Apr 2001</b></div>

<h2>
<a NAME="abstract"></a>Abstract</h2>
<a href="http://www.magma.ca/~jagwar/makingSmalltalkForwardingPage.html"></a>
<br>&nbsp;&nbsp;&nbsp; For those who haven't read the previous articles,
be sure to read <font size=+1><a href="http://www.linuxgazette.com/issue60/steffler.html#statementOfPurpose">the
statement of purpose</a></font> first.&nbsp; This month, we're going to
discuss OO thinking.&nbsp; For those looking to read the whole series locally
or information about upcoming articles, you can <font size=+1><a href="http://www.magma.ca/~jagwar/makingSmalltalkForwardingPage.html">check
the MST page</a></font>.&nbsp; For those looking for further information
on learning Squeak, here are <a href="http://minnow.cc.gatech.edu/squeak/377">some
good resources</a>.
<br>&nbsp;&nbsp;&nbsp; This is the last planned article for this series.&nbsp;
The reader interest has been high enough for me to continue with the next
series, but unfortunately my available writing time has quickly dwindled
:-( as my wife nears her due date :-)&nbsp; So this will be the last regular
article at least for a while.
<h2>
Quote of the day</h2>
Reason never changed a man's opinion which by reason he never acquired.
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -- Mark Twain
<h2>
OO Thinking</h2>
&nbsp;&nbsp;&nbsp; If you're just getting into OO from another programming
background, you'll soon realize that it requires a change in the way that
you think, the way you approach problems, and (IMHO) how much fun you're
having.&nbsp; This month, we go over some things to keep in mind when doing
OO programming.
<br>&nbsp;
<h3>
Breaking Linear Thinking</h3>
&nbsp;&nbsp;&nbsp; This is the first hurdle I've seen many people trip
over.&nbsp; They're so used to programs with a main() routine of some sort,
that when they first dip their toes into the Smalltalk pool they're frightened
off by not being able to find a linear beginning, middle, and end of something.&nbsp;
Realize that Smalltalk is about working with a group of collaborating objects.&nbsp;
To be sure, you will need some entry point to your code/application, however
it likely be in the form of opening your starting window, then saving/stripping
your image.
<br>&nbsp;&nbsp;&nbsp; Thinking of problems in terms of nouns and verbs
(objects and responsibilities) is a more natural way of thinking, and often
leads to a much different decomposition of the problem than functional
decomposition.&nbsp; Try to identify which objects are inherit to the problem,
which objects need to involved to help out, then think of the most basic
responsibilities and distribute them appropriately across the objects.
<br>&nbsp;&nbsp;&nbsp; This leads us to our next item:&nbsp; OO programming
lends itself well to iterative development.&nbsp; It's a natural activity
to define the basic objects, then start adding basic relationships and
responsibilities.&nbsp; If you find something doesn't fit right, then shift
the responsibility elsewhere.&nbsp; Flesh out your objects and responsibilities
over time.
<br>&nbsp;&nbsp;&nbsp; Try to use short methods to help maximize reuse
and maintainabilty.&nbsp; If you find yourself writing 100 line methods,
then you're still thinking linearly.&nbsp; The average method length varies
depending on whom you ask, but it should be short - somewhere around 8
statements or so.&nbsp; Of course, there's always exceptions to any rule
- this is just a rule of thumb.
<h3>
Decision Making vs Commanding</h3>
&nbsp;&nbsp;&nbsp; This is what I often think is the most fundamental difference
between OO programming and procedural programming.&nbsp; In procedural
programming it's common to do things in terms of decision making.&nbsp;
You do things like:
<ul>
<li>
<b>if this, then that, else that</b></li>

<ul>
<li>
For example, if data is an integer and user input is a float, then convert
the float to an integer to add</li>
</ul>

<li>
<b>for i = 1 to i = maxRange do this unless i > maxBounds, and if early
break condition is met then break out of loop</b></li>

<li>
<b>1 + 2 * 3 = ?</b></li>

<ul>
<li>
This example uses operator precedence, which is something that most languages
have.&nbsp; The statement is evaluated as: (1 + (2 * 3)) = 7.&nbsp; But
to determine precedence the language <i>needs to make a decision</i> which
operator to precede.</li>
</ul>
</ul>
&nbsp;&nbsp;&nbsp; A common problem that arises from decision making programming
is that you have similar decision making being done in several parts of
a program.&nbsp; Then when requirements or needs inevitably change, there
are many different spots that you need to update/modify your program to
update all the decision making spots.
<p>&nbsp;&nbsp;&nbsp; In OO programming, it's more common to do things
in terms of commanding.&nbsp; You command (or ask if you're polite) objects
to do things.&nbsp; If the object shouldn't do something, or should do
something differently, then it should know that.&nbsp; Since you ask different
objects the same thing, and they respond as each of them should, there's
no decision making.&nbsp; You do things like:
<ul>
<li>
<b>object doSomething</b></li>

<ul>
<li>
For example, it doesn't matter that you're adding a float to an int, the
float object knows how to add floats to itself, how to add ints to itself,
how to add fractions to itself.</li>
</ul>

<li>
<b>aCollection do: [:eachElement | eachElement doSomething]</b></li>

<ul>
<li>
Notice how there are no bounds checking - a collection object already knows
how to do that and does it for you.</li>
</ul>

<li>
<b>1 + 2 * 3 = ?</b></li>

<ul>
<li>
In this example, remember integers are objects too in Smalltalk (part of
the pure OO nature of Smalltalk).&nbsp; So we're asking the object 1 to
add itself to 2, then the resulting object to multiply itself by 3.&nbsp;
Hence, the statement is evaluated as: ((1 + 2) * 3) = 9.</li>

<li>
As a side note, it's funny how often I've seen some of my C++ or Java coworkers
flee from Smalltalk because this doesn't make sense to them.&nbsp; They
still haven't entirely made the shift to OO thinking.</li>
</ul>
</ul>

<h3>
Don't Sweat the Details</h3>
&nbsp;&nbsp;&nbsp; I once heard Alan Knight remark that you know somebody
is starting to get Smalltalk if they answer the question:&nbsp; "How does
Transcript show: 'HELLO WORLD' work?" with: "I don't care".&nbsp;&nbsp;
A common theme among Smalltalk newbies is a need to know exactly how everything
works, and step through all the methods of the objects from the library
that they use.&nbsp; This is related to linear thinking, in that you need
to understand how a linear path flows to determine how it broke down the
road.&nbsp; If you find yourself sweating the details of the class library,
then you're probably still in linear thinking mode.
<br>&nbsp;&nbsp;&nbsp; A related theme is that Smalltalk lends itself to
top-down coding.&nbsp; Put off work as long as possible and put off decisions
as long as possible - abstract and stub out reponsibilities if you can.&nbsp;
It's a powerful feeling to define even a trivial system that works, then
keep it working as you add real meat to it.&nbsp; You're most often in
a state of things working.
<h3>
Simplification by Encapsulation</h3>
&nbsp;&nbsp;&nbsp; Try and group data together with appropriate operations
in an object.&nbsp; If you're acting directly on an object's data in some
manner, then you're breaking encapsulation.&nbsp; If you're doing something
like:&nbsp; anObject aDataAttribute aPartOfAttribute doSomething, then
you're breaking encapsulation.
<br>&nbsp;&nbsp;&nbsp; A nice example of encapsulation is the looping noted
above.&nbsp; The collection class knows how many elements it has, and how
to loop over its elements, and you're not concerned with bounds checking
nor should you be.
<h3>
Reuse</h3>
Opportunities for reuse abound, and not just from the usual place of inheritence
<h4>
...through the class library</h4>
&nbsp;&nbsp;&nbsp; Before coding something, browse the class library to
see if it's already been done for you.&nbsp; Reinventing the wheel is definately
non-OO and wastes time.
<br>&nbsp;&nbsp;&nbsp; Another rule of thumb for knowing when you're getting
Smalltalk is the proportion of time you spend browsing the class library
to the proportion of time you spend coding.&nbsp; As you gain experience
and familiarity with the library, your proportion of time will go down,
but for a beginner you should expect to spend the majority of your time
browsing the library and the minority of your time coding.
<br>&nbsp;&nbsp;&nbsp; An appropriate remark I once heard (sorry, don't
remember the source), during a LOC metrics flame war is that Smalltalkers
should be measured by the LOC they <u>don't write</u>, as they're saving
time and maintentance costs by reusing the class library.
<h4>
...through goodies</h4>
&nbsp;&nbsp;&nbsp; Smalltalk has a rich history and a great user community.&nbsp;
There may be a freeware or opensource goodie out there that will satisfy
your needs.&nbsp; Have a look at the <a href="http://st-www.cs.uiuc.edu">UIUC
repository</a>, or search the web or ask the newsgroups for goodies.
<h4>
...by approrpiate responsibilities</h4>
&nbsp;&nbsp;&nbsp; If it isn't your reponsibility, then don't do it (or
redo it).&nbsp; Conversely, avoid responsibilities as much as possible
(only take the appropriate responsibilities).&nbsp; By trying to stick
to only appropriate responsibilities, then you're more likely to reuse
responsibilities elsewhere in the system.
<br>&nbsp;&nbsp;&nbsp; For example, don't have the responsibility to login
to your application in your client's login GUI (a bad practice in general),
if you later have a web GUI, then you need to either copy the login logic
to your web GUI, or factor out the login code to a reusable object.
<h4>
...through inheritence</h4>
&nbsp;&nbsp;&nbsp; Now we finally get to reuse through inheritence.&nbsp;
I leave this for last, as reuse through inheritence has been (IMHO) overhyped
and often overabused with needlessly deep class hierarchies that complicate
maintenance.
<br>&nbsp;&nbsp;&nbsp; For example, if you're writing a hospital system
you'd probably want to reuse a <b>Person</b>'s characteristics of <b>firstName</b>,
<b>lastName</b>,
and <b>socialSecurityNumber</b> by making subclasses of
<b>Doctor</b> and
<b>Patient</b>.
<h3>
Distributing responsibilities</h3>
&nbsp;&nbsp;&nbsp; Watch out for bloated parts of system - you can see
this if you're drawing your system out and your diagram looks like an octopus.&nbsp;
This is a sign that there are too many responsibilities on one object,
and that object is going to get harder to maintain as it bloats.&nbsp;
You should try and have groups of peer objects collaborating.
<br>&nbsp;&nbsp;&nbsp; Another warning sign is using a 'manager' object.&nbsp;
Again, there are perfectly good times and uses for a manager object, and
it can be difficult to determine if you're abusing a manager object.&nbsp;
I like to use a rule of thumb I heard from Alan Knight: object managers
should be like real world managers:&nbsp; they should not do any real work
- they should facilitate or manage interactions between other objects.
<h2>
<a NAME="sweetSqueak"></a>A Sweet Squeak</h2>
This month's sweet squeak is the release of Squeak 3.0! :-)&nbsp; To be
generic as possible, this description covers the scenario where you want
to run Squeak in Windoze or Linux.&nbsp;&nbsp; For this simple path install,
on Linux you will need root priviledges.&nbsp; <i>(Note: you can install
without root privilidges if you're familiar with updating your paths, I'm
not going to cover that topic in this simple guide)</i>
<br>&nbsp;
<h3>
Step 1:&nbsp; Downloading Squeak 3.0</h3>
Go to the FTP site: <b>ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/3.0</b>
and download:
<ul>
<li>
<b>Squeak3.0-win.zip</b>, includes:</li>

<ul>
<li>
<b>Squeak.exe</b>, the virtual machine (only good for Windoze, we'll need
to compile a VM for linux)</li>

<li>
<b>Squeak3.0.image</b>, (can use this on linux or Windoze)</li>

<li>
<b>SqueakV3.sources</b>, (can use this on linux or Windoze)</li>
</ul>

<li>
<b>Squeak-3.0pre2.tar.gz</b></li>

<ul>
<li>
Source files for compiling the linux VM</li>
</ul>
</ul>

<h3>
Step 2: Set a base directory to run squeak from</h3>
Assumes your Windoze mount point is <b>/windoze</b>, change for your system.
<p><i>Note:&nbsp; if you don't have/want to run dual boot, just change
your install location to be whatever you desire, for example: ~myuserid/squeak3,
and delete the unnecessary files: <b>NPSqueak.dll, Squeak.exe, SqueakFFIPrims.dll.</b></i>
<blockquote>
<li>
Make a <b>/windoze/squeak3 </b>directory</li>

<li>
Unzip the <b>Squeak3.1-win.zip</b> file into a <b>/windoze/squeak3</b>
directory.</li>
</blockquote>

<h3>
Step 3: Installing VM for linux</h3>
This is a <i>very easy</i> thing to do - even if you've never programmed
or compiled anything in your life before.&nbsp; Here are the steps:
<ol>
<li>
Unzip <b>Squeak-3.0pre2.tar.gz</b> to wherever (be sure to unzip with directories,
this unzips into a <b>Squeak-3.0 </b>directory)</li>

<li>
<b>cd</b> to where you unzipped the sources.&nbsp; (The <b>BUILD.UnixSqueak</b>
is a quick-n-easy guide from which these steps were condensed from)</li>

<li>
<b>mkdir build</b></li>

<li>
<b>cd build</b></li>

<li>
<b>../src/unix/configure --bindir="/windoze/squeak3"</b></li>

<li>
<b>make</b></li>

<li>
<b>make install</b>&nbsp; (NOTE:&nbsp; here is where you'll need root privilidges
with the default install, as stuff is copied to /usr/lib, /usr/man, etc)</li>

<ol>
<li>
Here, you're going to get a couple of errors (unless you're installing
to a Linux location), as you can't make links on a Windoze file system</li>

<li>
Copy the referenced files to your <b>/windoze/squeak3</b> directory:</li>

<li>
<b>cp /usr/lib/squeak/3.0/squeak /windoze/squeak3</b></li>

<li>
<b>cp /usr/lib/squeak/3.0/inisqueak /windoze/squeak3</b></li>
</ol>
</ol>

<h3>
<b>Step 4: Start Squeak :-)</b></h3>

<ul>
<li>
<b>cd /windoze/squeak3</b></li>

<li>
<b>squeak Squeak3.0final.image</b></li>
</ul>
...I'll leave starting up Squeak in Windoze as an exercise for the reader&nbsp;
;-)
<h2>
Quick tour</h2>
&nbsp;&nbsp;&nbsp; When I started up the image for the first time, I was
pleasently surprised that the default GUI to come up is the newer morphic
GUI (as opposed to the older MVC GUI that was mentioned in Article 1).&nbsp;
For the read-along folks, you'll see (click on the below half size images
for full size images):
<p>The entry screen.&nbsp; The Squeak logo in the top right is an xeyes
type of app, where the eyes follow the mouse.
<br>&nbsp;&nbsp;&nbsp;&nbsp;<a href="misc/steffler/squeak3-screen1.png"><img SRC="misc/steffler/squeak3-screen1-halfSize.png" height=249 width=400></a>
<p>If you put the mouse curor over the logo, you'll notice the pop-up balloon
help is enabled.
<br>&nbsp;&nbsp;&nbsp;&nbsp;<a href="misc/steffler/squeak3-screen2.png"><img SRC="misc/steffler/squeak3-screen2-halfSize.png" height=249 width=400></a>
<p>If you click on the project at the bottom right of the screen, it'll
zoom to full screen size as you enter it.
<br>&nbsp;&nbsp;&nbsp;&nbsp;<a href="misc/steffler/squeak3-screen3.png"><img SRC="misc/steffler/squeak3-screen3-halfSize.png" height=249 width=400></a>
<p>And finally, lets click on the music project to have a look.
<br>&nbsp;&nbsp;&nbsp;&nbsp;<a href="misc/steffler/squeak3-screen4.png"><img SRC="misc/steffler/squeak3-screen4-halfSize.png" height=249 width=400></a>
<br>&nbsp;
<h2>
<a NAME="lookingForward"></a>Looking forward</h2>
&nbsp;&nbsp;&nbsp; Alas, there will be no immediate looking forward due
to my time constraints.&nbsp;&nbsp; The next series I was planning covers
some basic programming basics like:&nbsp; unit testing (SUnit), source
code management (change sets and SCAN), an object tour of commonly used
objects, control structures, and Squeaklets.
<p>&nbsp;&nbsp;&nbsp; In the meantime though, I <i>highly recommend</i>
downloading v3.0 of Squeak (noted below) and trying out the STP goodies
as your first goodie exploration.&nbsp; They're available from:&nbsp; <a href="http://www.create.ucsb.edu/squeak/STP12.html">http://www.create.ucsb.edu/squeak/STP12.html</a>
<p>&nbsp;&nbsp;&nbsp; I've enjoyed learning about Squeak over the past
few months, and I hope you've enjoyed the series.
<hr WIDTH="100%">
<h2>
<a NAME="personSampleCode"></a><b>Smalltalk Code</b></h2>
Somebody pointed out to me that the ScopedBrowser used in Article 4 doesn't
work properly in Squeak v3.0, so here's an<b> <a href="misc/steffler/MakingSmalltalk-Article5.st">updated
version</a></b>.
<br>Note:&nbsp; I noticed that SUnit is now included as part of the base
image now, so I've included some programmatic unit tests.&nbsp; After loading
the code, if you wish to run the unit tests, do: <b>TestModel openAsMorph</b>,
then click the <b>Run</b> button.&nbsp; You'll notice 8 windows pop up
and close, and there shouldn't be any errors listed in the error pane.




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Jason Steffler.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">A Private Home Network</font></H1>
<H4>By <a href="mailto:jstumpel from zonnet dot nl">Jan Stumpel</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<h3>
1. Introduction</h3>
Until recently I paid little attention to the security of my home network
for the following reasons:
<ul>
<li>
A dial-up home installation will not attract much attention from crackers.</li>

<li>
Linux is safe anyhow, compared to MS Windows.</li>

<li>
The people who put together my Linux distribution have surely taken
security into account.</li>

<li>
I took some security measures (hosts.allow, hosts.deny, ipchains) following
the examples in the HOWTO's.</li>

<li>
I don't understand this security business anyway.</li>
</ul>
Common human psychology suggests that this is how many ordinary Linux users
think. In my case, unfortunately, all these points turned out to be pure
wishful thinking, apart from the last one. <i>Because of&nbsp; </i>the
last one.
<p>How did I find this out? In order to prepare for the happy day in the
future when permanent, high-speed connections to the Internet will be offered
in my area, I decided it was a good idea to start investigating security
issues. The results were shocking.
<p>The first shock came from looking at my long-neglected <tt>/var/log/syslog*
</tt>files.
A few 'refused connect from' entries. One 'connect from' to ftp which apparently
succeeded. Oops. Dial-up Internet users are not overlooked by the crackers
after all. And my security is not bullet-proof. Better to spend some time
really looking at security. And to try to <i>understand</i> something of
it this time. So this meant reading books, FAQ's, HOWTO's, and a lot of
articles on the Web; and doing some experiments.
<p>This is the result of my investigations. Mind: I am not an expert, but
just an amateur, a home user trying to make things work. Nothing of this
comes with any guarantee.
<h3>
2. The system</h3>
I have a very simple home network with two machines:
<ul>
<li>
<b>earth</b> is a Win 95 machine without printer or modem.</li>

<li>
<b>heaven</b> runs Debian Linux 2.1 (with various upgrades). It runs exim
(for local mail service and for sending mail <i>to</i> the outside world),
qpopper (pop3 server for use by <b>earth</b>), and samba (to provide file
sharing and printing to <b>earth</b>). <b>heaven</b> connects on-demand
to the ISP, opening a modem (ppp) link. Mail <i>from</i> the outside world
is collected by means of fetchmail.</li>
</ul>
This local network uses one of the IP address ranges reserved for private
networks, 192.168.1.0/24. <b>heaven</b> is 192.168.1.1, <b>earth</b> is
192.168.1.2.
<p>The contents of <tt>/etc/hosts</tt> on <b>heaven</b>, and <tt>c:\windows\hosts</tt>
on <b>earth</b>, is:
<blockquote><tt>127.0.0.1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
localhost</tt>
<br><tt>192.168.1.1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
heaven.my.home&nbsp; heaven</tt>
<br><tt>192.168.1.2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
earth.my.home&nbsp;&nbsp; earth</tt></blockquote>
This shows that my network uses the domain name <tt>my.home</tt>. This
name is unregistered and meant only for local use. Mail to the outside
will have its 'message from' and 'envelope from' addresses translated (in
the <a href="#jwexim">July</a> and <a href="#jwsmtp">September</a> issues,
1999, of <i>LG</i> I described how to do this).
<h3>
3. 'Private' home networks</h3>
My notion of security is to have a private network. By this I mean a network
that provides <i>no public functions</i>. It does not serve WWW pages or
files. You cannot telnet into it. It does not even <i>listen</i> to anything
coming in from the outside. If anyone comes knocking, there is simply <i>no
response</i>. This idea was recently put forward by <i>Sander Plomp</i>,
whose <a href="#fortress1">articles</a> at rootprompt.org provided much
of the inspiration for this piece.
<p>A LAN which is not connected to the Internet is a private network by
definition. Unplug your modem, and you make your network private. But that
is a not the kind of private network that I mean. I want to use the Net,
send and receive mail, browse the Web, download files, etc. I just do not
want anyone from the outside to <i>enter</i> my network.
<p>Linux systems generally <i>aren't</i> private networks. By default,
the installation procedure of a Linux distribution sets up all sort of
nice network services (like telnet, ftp, finger, etc.) which are accessible
by anyone in the world, protected (if at all) only by a password. Also,
Microsoft Windows home LAN's are generally
<i>not</i> private. Connect
two Win95 computers and enable 'file sharing', and the whole world can
share your files while your Internet connection is up.
<p>To make non-private networks safe, various techniques are used; passwords
of course, and also other security techniques that have been discussed
in the Linux Gazette many times, like tcpd (alias tcp wrappers) and kernel-level
packet filtering (with ipchains as the user interface). These techniques
give some privacy to a system which is essentially
<i>public</i>. They
are like guards at the door, put there to keep out unwanted characters,
while letting the desirable customers in. But why should there be doors
<i>at
all</i>? I don't <i>want</i> any customers. My network is <i>private</i>!
<p>If we have servers running, reachable by the outside world, we always
have to worry about having made some configuration mistake which can be
exploited. Also, server programs often have bugs in them which offer openings
to crackers. Only recently one was discovered in <tt>named</tt>. OK, it
was later patched, so people who got the newest version of <tt>named</tt>
do not have to worry anymore about <i>this</i> bug. But what about the
next one? Better not to to have any doors at all!
<p>If you really <i>want</i> to enable services (mp3 distribution, or whatever)
for use by computers outside your home, you must study more advanced security
techniques. But if you simply want a <i>private</i> network for the home,
read on.
<h3>
4. How safe <i>is</i> your network?</h3>
To test the safety of a home network, you can have it 'scanned' from the
outside, for instance by <i><a href="http://www.sdesign.com/securitytest">Secure
Design</a></i>. Click 'scan me now' and 'basic scan'. Try it both from
the gateway machine and from the other machines on your network. When I
did this, I got the second shock. It was embarrassing. A long list showing
possible entry points into my system: Samba shares, telnet, the print service,
X, the mail server, ftp, finger, etc. I had some rudimentary safety measures
in place, so the system was somewhat protected from serious intrusion (I
hope). But I consider it a breach of my privacy that the outside world
even <i>knows</i> that I have a mail server (let alone letting them break
into it). These services were meant for use within my home network. They
are not the business of the outside world in any way.
<p>So: have your system scanned! Apart from Secure Design, other scan services
exist, e.g. <i><a href="https://grc.com/x/ne.dll?bh0bkyd2">Shields Up!</a></i>,
<i><a href="http://www.dslreports.com/scan">DSL
Reports</a></i>,
<i><a href="http://scan.sygatetech.com">Sygate Online
Services</a></i>, and many others. A whole lot of them can be found at
an Austrian site,
<i><a href="http://www.home.pages.at/heaven/sec0185.htm">Sicherheit
im Kabelnetzwerk</a></i> ('Security in the Cable Network'; there is also
an English version with almost the same information). Use several scan
services. Print the results. What this 'scanning' actually means will hopefully
become clearer in the course of this article.
<p>You can also 'scan' your system yourself by calling (after <tt>su</tt>-ing
to <tt>root</tt>) <tt>netstat -pan --inet</tt>. Use a wide xterm window
when doing this, because the output consists of rather long lines. Programs
which have 0.0.0.0 in the 'Local Address' column are visible to the whole
world!
<h3>
5. Servers and clients</h3>
The distinction between servers and clients is not always clear to users.
If you want to use ftp, for instance (getting files from, and putting files
into, another computer) you use an ftp <i>client program</i> to connect
to the other computer. If that is all you want to do with ftp, the client
program is all you need. An ftp <i>server</i> is only needed if you want
to allow others to get files from, or put them into, <i>your</i> computer.
Similarly with telnet: a client program for your <i>own</i> use, a server
program for <i>other</i> peoples' use. The client and server programs are
completely different and have different names; for instance <tt>/usr/bin/telnet</tt>
for the telnet client, <tt>/usr/sbin/in.telnetd</tt> for the server. For
novices, the distinction is not always clear. If a Linux setup program
asks 'shall I set up the ftp server?', users may think 'well eh.. yes,
I certainly want to use ftp, so go ahead'. Often you are not even asked,
and an ftp server is installed by default.
<p>One way of creating a private network is not to install servers at all,
just clients. But that is too simple if you have a network at home connecting
two or more computers. <i>Inside</i> your network you want to telnet from
one machine to another, you want to run an internal mail service, etc.
In other words you can't do without servers.
<p>What servers do is <i>listen.</i> They listen for a signal that says:
<i>I
want your service</i>. The signal (at least for the TCP-based services)
is a special IP packet, called a SYN packet, that enters your computer
and specifies the number of a service. For instance, the number of the
telnet service (that the
<tt>in.telnetd</tt> program, if it is running,
listens to) is 23. These numbers are usually called 'port numbers'. If
the <tt>in.telnetd</tt> program is <i>not</i> running, no one listens to
SYN packets with the number 23. So, as they say, 'port 23 is closed'.
<p>Ports do not exist by themselves, like little doors in your computer
that you can open or close. A port is open if a server listens to it. Otherwise
it is closed. A TCP port comes into existence if there is a program which
listens to it, and if not, it does not exist!
<p>How do the SYN packets get into your computer? In the case of <b>heaven</b>,
a packet can get 'into' it in three different ways:
<ul>
<li>
packets sent from <b>earth</b> come in via the Ethernet card (also called
the eth0 interface); they are addressed to the fixed IP address 192.168.1.1.</li>

<li>
packets from the outside world come in through the ppp link, or ppp0 interface;
that also has an IP address, but is is not fixed. At every Internet session
the ISP hands out a 'dynamic' address, valid for this session only.</li>

<li>
packets can also be sent from <b>heaven</b> itself, addressed to itself
as it were. This way of sending packets is often used for testing; the
packets are addressed to the so-called loopback interface with address
127.0.0.1. The name <b>localhost</b> refers to this loopback interface,
while the name <b>heaven</b> refers to 192.168.1.1. (This is a rather important
point: names and IP addresses refer to <i>interfaces</i>, not to computers,
although in daily usage this distinction is often blurred.)</li>
</ul>
Now the key point is that servers normally listen for any packets with
'their' port number, <i>no matter which way they enter the system</i>.
If we want to make a private network, offering no services to the outside
world, we must somehow change this.
<p>It would be nice if all the server programs available on Linux systems
had options specifying <i>which</i> interfaces they will listen to. In
that case you could just tell all your servers never to listen to the ppp
line, and you'd be all set. Hardly any security measures would be needed
at all (tcpd, firewalls, etc.); you would only use them 'for good measure',
as an extra precaution. Maybe this will happen at some time in the future,
but at the moment only a few server programs have this (including the important
cases of exim and samba). So we have to do several things to make our network
private:
<ol>
<li>
it never hurts to follow the commonly-heard advice to 'close unneeded ports',
in other words not to run servers that you do not need.</li>

<li>
for services that have the option, make them listen to internal interfaces
(eth0 and, if necessary, loopback) only.</li>

<li>
the 'super-server' <tt>inetd</tt> (which is used for 'waking up' a lot
of different servers on a Linux system) should be replaced by <b><tt><a href="http://www.xinetd.org">xinetd</a></tt></b>,
which has the option to listen to internal interfaces only. <i>NOTE</i>:
apparently, Red Hat Linux 7.0 installs <tt>xinetd</tt> by default.</li>

<li>
unsafe servers which cannot run from <tt>xinetd</tt> (like the print server
<tt>lpd)</tt>
should, where possible, be replaced.</li>

<li>
for the remaining difficult cases you need a firewall that blocks SYN packets
from the outside. This could also be used for blocking unwanted UDP and
ICMP access.</li>

<li>
for 'advanced' security, a few other possibilities suggest themselves.
One is <i>not</i> to use IP masquerading and forwarding on your network.</li>
</ol>

<h3>
6. Removing unneeded services</h3>

<h4>
6.1 Unnecessary <tt>inetd</tt> services</h4>
First, there are some services which run 'from inetd.conf'. Almost all
Linux systems have a 'super-server' called <tt>inetd</tt>, whose job it
is to listen to a lot of ports at the same time, and to 'wake up' services
when needed. However, it will also wake up services which you do
<i>not
</i>need.
<p>Examples of unneeded services are:
<ul>
<li>
the ftp server. I do not plan to serve any files to the outside world,
and internally on my network I can use Samba (and smbfs) to transfer files.
Not having an ftp <i>server</i> in no way stops you from running an ftp
<i>client</i>
when you want to, and to exchange files with the outside world. But most
distributions, including Debian, install an ftp server by default.</li>

<li>
the portmapper and anything related to RPC calls (same reason). The portmapper
is used for allowing remote procedure calls. Basically you need this if
you use NFS. But I do not use NFS (Samba, which I need anyway because there
is a Win95 machine on the network, provides enough file sharing facilities).
So anything related to the portmapper and RPC can be commented out from
inetd.conf.</li>

<li>
finger and ident. About the usefulness or otherwise of 'ident', opinions
seem to be divided. I removed it, and did not suffer any ill effects.</li>

<li>
several other obviously unnecessary things that can be started up from
<tt>inetd.conf</tt> (like <tt>saft</tt>).</li>

<li>
several services in <tt>inetd.conf</tt> which are used only for testing
networks: <tt>echo</tt>, <tt>chargen</tt> (pronounced <i>kargen,</i> 'character
generator'), <tt>discard</tt>, <tt>daytime</tt>, and <tt>time</tt>. The
last two (in case you are worried) have nothing to do with <i>timekeeping</i>
on your system; they are just services which will tell your system's time
to others. You don't need them, nor any of the other 'test' services.</li>
</ul>
These services can all be disabled by commenting out (putting a # character
at the beginning of) the corresponding lines in <tt>/etc/inetd.conf</tt>,
and restarting <tt>inetd</tt> (in Debian: <tt>/etc/init.d/inetd restart</tt>).
<h4>
6.2 Other unnecessary services</h4>
If a service is not woken up by <tt>inetd</tt>, it runs independently,
as a 'daemon' or background program.&nbsp; Among the daemons which you
may not need, apart from the portmapper (if run as a daemon) is a local
nameserver (<tt>named</tt>, pronounced <i>name-dee</i>). There is no reason
why a small home network should run such a thing. <tt>/etc/hosts</tt> and
<tt>C:\windows\hosts</tt>
files on your machines, and the addresses of your ISP's name servers in
<tt>/etc/resolv.conf</tt>
will enable address lookup on your network.
<p>There is usually a command to prevent a service from starting automatically
upon boot-up, by removing it from the start-up directories; for instance
I got rid of a <i>tamagotchi server</i>, automatically installed by Debian
2.1, by calling
<blockquote><tt>update-rc.d -f /etc/init.d/tama remove</tt></blockquote>

<h3>
7. Securing wanted services: non-<tt>inetd</tt></h3>
Now we tackle services that we <i>do</i> want to use, but do not want to
be visible to the outside world. Often there is some configuration option
that will keep the service 'private'. Examples follow.
<h4>
7.1 X</h4>
X was designed as a network-oriented window system, but in fact the network
features are never used in most setups, and present a security risk. You
can eliminate X's network capabilities by starting X with a command-line
option: <tt>startx -- -nolisten tcp</tt>. Secure Design now no longer reports
that 'X11 is open'.&nbsp; To make this permanent, you can make an alias
for <tt>startx</tt> in<tt> ~/.bashrc, /etc/profile</tt>, or some other
good location, like this:
<blockquote><tt>alias startx="startx&nbsp; --&nbsp; -nolisten tcp"</tt></blockquote>
The <tt>-nolisten tcp</tt> command should really be in one of the X11 resource
files, but so far I haven't found out which one. The 'alias' approach works
in any case. To test, run (as root) <tt>netstat -pan --inet</tt>. X should
no longer be mentioned. Of course it would be nicer if we could keep X's
network abilities for the local network, only blocking them against outside
access, but I couldn't find a way to do that.
<h4>
7.2 Samba</h4>
In a Debian system, the configuration file for Samba is <tt>/etc/samba/smb.conf</tt>
(on other systems, it may be <tt>/etc/smb.conf</tt>).<tt> </tt>When installing
Samba, I chose 'let Samba run as daemons'; it did not work properly from
<tt>inetd</tt>.
Any lines referring to netbios (which is what Samba uses) in <tt>/etc/inetd.conf</tt>
must therefore be commented out. Then in <tt>/etc/samba/smb.conf</tt>,
section <tt>[global]</tt>, I added
<blockquote><tt>&nbsp;bind interfaces only = True</tt>
<br><tt>&nbsp;interfaces = 192.168.1.1</tt></blockquote>
After <tt>/etc/init.d/samba restart</tt>, the Samba daemons only listen
to our home LAN. They are no longer visible to the outside world. Check
with <tt>netstat -pan --inet</tt>, and by having the system scanned.
<h4>
7.3 Exim</h4>
Exim is the mail server (or Mail Transport Agent, MTA) on my system. You
may have something else (like sendmail or postfix) but then the same principle
applies: your private mail agent should <i>not listen</i> to the outside
world. People who send mail to you, and to the other users in your home,
send it to mail accounts at the ISP (or to several mail accounts at different
ISP's). You retrieve it from there using, e.g., <tt>fetchmail</tt>. People
cannot send mail to your network <i>directly</i>.
<p>Exim turns out to have an option <tt>local_interfaces</tt> (which goes
into the <tt>MAIN CONFIGURATION</tt> section of <tt>/etc/exim.conf</tt>).
This is a list of (IP addresses of) interfaces that exim will listen to.
This only works when exim runs as a daemon, independent of <tt>inetd</tt>.
To set this up:
<ul>
<li>
In <tt>/etc/exim.conf</tt>, in the <tt>MAIN CONFIGURATION</tt> section,
enter a line:</li>

<br><tt>local_interfaces = 192.168.1.1:127.0.0.1</tt>
<br>(apart from the local net, also loopback must be specified, otherwise
fetchmail won't work; or you must call <tt>fetchmail -S <i>yourmachinename</i></tt>).
<li>
Comment out the <tt>smtp</tt> line in <tt>/etc/inetd.conf</tt> .</li>

<li>
In /<tt>etc/init.d/exim</tt> comment out the line <tt>exit 0</tt> near
the beginning, just after the line <tt>#usually this is disabled and exim
runs from /etc/inetd.conf</tt>. This causes exim to run as a daemon after
you call <tt>/etc/init.d/exim start</tt>. Letting exim run as a daemon
means that you have to call <tt>/etc/init.d/exim restart</tt> after every
change to <tt>exim.conf</tt>.</li>
</ul>
Letting exim run as a daemon means that it consumes some cycles and some
memory all the time. But as a bonus, exim's <tt>RETRY CONFIGURATION</tt>
now works properly as well, which it never did when running under <tt>inetd</tt>.
<h4>
7.4 Junkbuster</h4>
Junkbuster is an http proxy server which you can configure to keep out
ads and other unwanted stuff. It works very well. In a Debian system it
listens to port 5865, in other systems to port 8000. This is set in the
file
<tt>/etc/junkbuster/config</tt>. By default, junkbuster listens to
<i>all</i>
interfaces (in other words, to the whole world). However, you can set in
the config file
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp; listen-address&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
192.168.1.1:5865</tt>
<p>and now only machines on our own network can connect to it (including
the gateway machine that junkbuster runs on, <b>heaven</b> in this example,
provided <tt>heaven</tt> is entered in the Netscape 'Preferences/Advanced/Proxies'
menu, not <tt>localhost</tt>).
<h4>
7.5 Other (non-<tt>inetd</tt>) services</h4>
The above examples are <i>only</i> examples. If your system runs other
services outside <tt>inetd</tt>, check their documentation for ways to
make them private. For instance, it appears that<i> sendmail</i> can be
made to listen only to the local network by means of
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp; 0 DaemonPortOptions=Addr=192.168.1.1</tt>
<p>in the <tt>sendmail.cf</tt> file. I did not try this.
<h4>
7.6 Remaining problem cases, like <tt>lpd</tt></h4>
<tt>lpd</tt> remains a problem. It cannot be made to listen to the internal
network only. Basically, it should be replaced by something safe. <a href="#fortress1">Sander
Plomp</a> recommends replacing it by <b><tt><a href="http://pdq.sourceforge.net/">pdq</a></tt></b>.
I've been too lazy to do this yet, but it certainly needs attention in
the near future.
<p>Other problem cases may remain: servers which you need in your own network
but which cannot be made private, and for which private alternatives do
not exist. I have this problem with <tt>cannaserver</tt>, a system for
inputting Japanese characters from the keyboard. Such services must be
screened from the outside world by means of a packet-filtering firewall.
See section 10 of this article.
<h3>
8. Masking the <tt>inetd</tt> services through <tt>xinetd</tt></h3>
By now the list of visible services on the system, according to the Secure
Design scan, has become:
<ul>
<li>
telnet</li>

<li>
pop3</li>

<li>
lpd (the print system)</li>
</ul>
Much fewer than there used to be, but still far too much. Telnet and pop3
are started by inetd/tcpd and thus are secured by <tt>hosts.allow</tt>
and <tt>hosts.deny</tt>, but I'm not sure that this protection is 100%
cracker-proof. <tt>lpd</tt> remains totally 'open', as far as I can guess,
unsecured. It cannot, I think, be started from inetd.
<h4>
8.1 Replacing <tt>inetd</tt></h4>
Security for a home network really requires replacing <tt>inetd</tt> by
something which can distinguish between requests for service from the <i>local
network</i> and from the <i>outside</i>. Plomp recommends <tt>tcpserver</tt>;
I tried
<tt>xinetd</tt>. First kill
<tt>inetd</tt>, then install <tt>xinetd.
</tt><i>Important:</i>
The Debian script <tt>/etc/init.d/xinetd</tt> not only starts the <tt>xinetd</tt>
daemon by itself, but also the portmapper. We do not need/want the portmapper,
which is used for RPC calls and NFS, which we do not use. So anything related
to the portmapper in <tt>/etc/init.d/xinetd</tt> must be commented out
(# at the beginning of the line).
<p>One way to configure xinetd for telnet and pop3 is to put in <tt>/etc/xinetd.conf</tt>:
<p><tt>defaults</tt>
<br><tt>&nbsp;&nbsp;&nbsp; {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; instances = 10</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; log_type = SYSLOG daemon</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; log_on_success += DURATION HOST
USERID</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; log_on_failure += HOST</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; interface = 192.168.1.1</tt>
<br><tt>&nbsp;&nbsp;&nbsp; }</tt>
<p><tt>service telnet</tt>
<br><tt>&nbsp;&nbsp;&nbsp; {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; socket_type = stream</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; wait&nbsp; = no</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; user&nbsp; = root</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; server = /usr/sbin/in.telnetd</tt>
<br><tt>&nbsp;&nbsp;&nbsp; }</tt>
<p><tt>service pop-3</tt>
<br><tt>&nbsp;&nbsp;&nbsp; {</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; socket_type = stream</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; wait&nbsp; = no</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; user&nbsp; = root</tt>
<br><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; server = /usr/sbin/in.qpopper</tt>
<br><tt>&nbsp;&nbsp;&nbsp; }</tt>
<p>So apart from a general 'defaults' section which specifies the interface,
there is a separate section for each service that you want to run. Although
the format is completely different, the <i>data</i> for the various sections
can be found in your existing <tt>inetd.conf</tt>. See also <tt>man xinetd.conf</tt>.
<p>I started <tt>xinetd</tt> and verified that is is now possible to telnet
to <b>heaven</b> both from
<b>heaven</b> itself and from <b>earth</b>.
However, Secure Design no longer reports that my system has open telnet
and pop3 ports! Success! <i>NOTE</i>: from my own machine, <tt>telnet heaven</tt>
succeeds, but <tt>telnet localhost</tt> does not. <tt>xinetd</tt> can only
bind to one interface; in this case 192.168.1.1, not at the same time to
<tt>localhost</tt>,
which is the loopback interface (127.0.0.1).
<p>By now, all other services in <tt>/etc/inetd.conf</tt> have been commented
out. Therefore <tt>inetd</tt> no longer does anything and we can get rid
of it in the boot-up scripts. In Debian, it goes like this:
<blockquote><tt>update-rc.d -f /etc/init.d/inetd remove</tt></blockquote>
Its place is taken by <tt>xinetd</tt>:
<blockquote><tt>update-rc.d xinetd defaults</tt></blockquote>
OK; another step towards security successfully taken.
<p>The output of <tt>netstat -pan --inet</tt> is now something like:
<p><tt><font size=-1>heaven:~# netstat -pan --inet</font></tt>
<br><tt><font size=-1>Active Internet connections (servers and established)</font></tt>
<br><tt><font size=-1>Proto&nbsp; Local Address&nbsp;&nbsp;&nbsp;&nbsp;
Foreign Address&nbsp; State&nbsp;&nbsp; PID/Program name</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 127.0.0.1:25&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 11391/exim</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 192.168.1.1:25&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 11391/exim</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 192.168.1.1:139&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 10761/smbd</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 192.168.1.1:5865&nbsp; 0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
LISTEN&nbsp; 1670/junkbuster</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 192.168.1.1:110&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 161/xinetd</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 192.168.1.1:23&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 161/xinetd</font></tt>
<br><tt><font size=-1>tcp&nbsp;&nbsp;&nbsp; 0.0.0.0:515&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LISTEN&nbsp; 148/lpd
MAIN</font></tt>
<br><tt><font size=-1>udp&nbsp;&nbsp;&nbsp; 192.168.1.1:138&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
10759/nmbd</font></tt>
<br><tt><font size=-1>udp&nbsp;&nbsp;&nbsp; 192.168.1.1:137&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
10759/nmbd</font></tt>
<br><tt><font size=-1>udp&nbsp;&nbsp;&nbsp; 0.0.0.0:138&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
10759/nmbd</font></tt>
<br><tt><font size=-1>udp&nbsp;&nbsp;&nbsp; 0.0.0.0:137&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
10759/nmbd</font></tt>
<br><tt><font size=-1>raw&nbsp;&nbsp;&nbsp; 0.0.0.0:1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
-</font></tt>
<br><tt><font size=-1>raw&nbsp;&nbsp;&nbsp; 0.0.0.0:6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
0.0.0.0:*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
-</font></tt>
<p>Almost all services now listen to a local interface. The print system
is the exception: it listens to address 0.0.0.0 (i.e., everywhere) on port
515. Sure enough, if the system is scanned now, only port 515 is reported
as 'open'. In fact some Windows-oriented scan services will report your
system as totally 'closed', because they do not scan port 515.
<h3>
9. What about IP Masquerading?</h3>
For ages I have used IP Masquerading, to give the Windows box in my home
access to the Internet. Or so I thought. When recently, in the course of
my investigations into system safety, I switched off Masquerading, the
Windows box could use the Internet as before.
<p>What happened? Simply that <b>earth</b>, the Windows machine, only does
three Internet-related things:
<ol>
<li>
e-mail, which does not require <b>earth</b> to connect to the Internet,
only to the smtp and pop3 servers running on <b>heaven</b>;</li>

<li>
Web browsing; for this, <b>earth</b> connects to junkbuster on <b>heaven</b>,
again not directly to the Internet;</li>

<li>
fetching the outside mail from the ISP; for this, <b>earth</b>'s user telnets
to <b>heaven</b> and runs fetchmail from there.</li>
</ol>
So, ever since I installed junkbuster about six months ago, <b>earth</b>
has never approached the Internet directly, and Masquerading is now superfluous.
I had not realized this. Inadvertently I had created a 'proxying firewall'.
This means that Masquerading can - and must - now simply be switched off.
This has several advantages:
<ul>
<li>
Simplicity: ipchains (see the next section) no longer has a FORWARD chain,
so we don't have to worry about it. We do not need to set up DNS on <b>earth</b>
(entering nameserver addresses).</li>

<li>
Security: if ever some Trojan becomes established on <b>earth</b>, it will
not be able to contact its evil accomplices through the Internet.</li>
</ul>
The only 'downside' is that <b>earth</b> is now restricted to e-mail and
http only. No ping and telnet to the outside world, no ftp, Real Audio,
chat, etc. But for the moment mail and http is all that's required. If
other services become necessary on <b>earth</b>, I suppose I shall have
to install proxies for them on <b>heaven</b>.
<p>To switch IP masquerading and forwarding off, in Debian, you do
<ul>
<li>
change the first line in <tt>/etc/network/options</tt> so it says <tt>ip_forward=no</tt></li>

<li>
disable forwarding in the kernel by means of&nbsp; <tt>echo 0 > /proc/sys/net/ipv4/ip_forward</tt></li>

<li>
remove any special commands for Masquerading, like <tt>ipchains -A forward
-s 192.168.1.0/24 -j MASQ</tt> from your startup scripts, ip-up script,
or wherever you had them.</li>
</ul>

<h3>
10. Closing the last doors</h3>
Let's recapitulate: by <i>eliminating</i> services, <i>reconfiguring</i>
other services so they don't listen to the gateway interface, by <i>wrapping</i>
others inside xinetd, and by turning off Masquerading, we have created
a system which is already quite secure
<i>without a firewall</i>. Now it
is time to add the final touch: we build a packet-filtering firewall around
the system using ipchains. This should be the last step, not the first.
<p>One often reads the advice to configure ipchains in such a way that
'everything is blocked by default', and then to make exceptions for the
things that you want to allow. Theoretically this may be the right thing,
but in practice it leads to much frustration. If everything is blocked,
your system will basically not work. You are more or less groping in the
dark when it comes to deciding what you have to allow. So I allow everything
by default, and then add restrictions one by one. If the system breaks
(e.g. no ping, or no Web page viewing) the last restriction has been too
drastic, and must be undone. Setting the default policy of a chain to DENY
or REJECT can then (again) be the last step, not the first.
<p>I started by taking down the firewall (<tt>ipchains -F</tt>) and then
running a simple firewall script with one rule:
<p><tt>#!/bin/sh</tt>
<br><tt># simple firewall</tt>
<p><tt>ipchains -F input</tt>
<br><tt>ipchains -P input ACCEPT</tt>
<br><tt>ipchains -A input -i ppp0 -p TCP --syn -j DENY -l</tt>
<p>This blocks SYN packets coming from the outside interface, enhancing
the privacy of the system very considerably. Nobody from the outside can
start a connection; outside scan services report that the site is completely
closed (some now even call it 'stealthed' or 'invisible'). But we can add
more restrictions. That is, we can use <i>more general </i>DENY/REJECT
rules, and <i>more specific</i> ACCEPT rules.
<p>Before you add restrictions, it is useful to do some experiments. You
can make ipchains-type rules which let packets through while logging them
(<tt>-j ACCEPT -l</tt>). So even if (like me) you do not really
<i>know</i>
which packets to block, you can see what is going on 'normally' by keeping
a window open with <tt>tail -f /var/log/syslog</tt> in it. Then afterwards
you can make rules to block packets which are not 'normal'. I strongly
advise you to do your <i>own</i> experiments, and to make rules based on
your <i>own</i> understanding.
<p>After a few such experiments,<i> my</i> firewall script in <tt>/etc/ppp/ip-up.d</tt>
looks as follows. This assumes you have no nameserver running, but have
the addresses of TWO nameservers provided by your ISP in <tt>/etc/resolv/conf</tt>.
Mind the important <i>backquotes</i> (<tt>`</tt>)! They may disappear if
you cut-and-paste from this page.
<p><tt><font size=-1>#!/bin/sh</font></tt>
<br><tt><font size=-1># A slightly more complicated firewall</font></tt>
<p><tt><font size=-1># Find external name server addresses</font></tt>
<br><tt><font size=-1>ns="`grep nameserver /etc/resolv.conf | awk '{print
$2}'`"</font></tt>
<br><tt><font size=-1>nameserver1="`echo $ns | sed -e 's/ .*//'`"</font></tt>
<br><tt><font size=-1>nameserver2="`echo $ns | sed -e 's/.* //'`"</font></tt>
<p><tt><font size=-1># Set up INPUT rules</font></tt>
<br><tt><font size=-1>ipchains -F input</font></tt>
<br><tt><font size=-1>ipchains -P input ACCEPT</font></tt>
<p><tt><font size=-1># Block outside input from reserved address ranges</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -s 10.0.0.0/8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
-j DENY</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -s 172.16.0.0/12&nbsp;&nbsp;
-j DENY</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -s 192.168.0.0/16&nbsp;
-j DENY</font></tt>
<p><tt><font size=-1># Block TCP connections from the outside</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -p TCP --syn -j DENY -l</font></tt>
<p><tt><font size=-1># Block all UDP except nameserver replies</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -p UDP -s $nameserver1
53 -j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -p UDP -s $nameserver2
53 -j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -p UDP -j DENY -l</font></tt>
<p><tt><font size=-1># Allow (for now) but log all ICMP</font></tt>
<br><tt><font size=-1>ipchains -A input -i ppp0 -p ICMP -j ACCEPT -l</font></tt>
<p><tt><font size=-1># From local net, allow only packets to us and broadcasts</font></tt>
<br><tt><font size=-1># Forwarding is off, other packets won't go anywhere,
but</font></tt>
<br><tt><font size=-1># now we can log them to detect illegal activity
on our net</font></tt>
<br><tt><font size=-1>ipchains -A input -i eth0 -d 192.168.1.1&nbsp;&nbsp;
-j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A input -i eth0 -d 192.168.1.255 -j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A input -i eth0 -j REJECT -l</font></tt>
<p><tt><font size=-1># Set up OUTPUT rules</font></tt>
<br><tt><font size=-1>ipchains -F output</font></tt>
<br><tt><font size=-1>ipchains -P output ACCEPT</font></tt>
<p><tt><font size=-1># Don't send packets out to reserved address ranges</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -d 10.0.0.0/8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
-j REJECT</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -d 172.16.0.0/12&nbsp;&nbsp;
-j REJECT</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -d 192.168.0.0/16&nbsp;
-j REJECT</font></tt>
<p><tt><font size=-1># Block all UDP except nameserver requests</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -p UDP -d $nameserver1
53 -j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -p UDP -d $nameserver2
53 -j ACCEPT</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -p UDP -j REJECT -l</font></tt>
<p><tt><font size=-1># Allow (for now) ICMP to the outside, but log</font></tt>
<br><tt><font size=-1>ipchains -A output -i ppp0 -p ICMP -j ACCEPT -l</font></tt>
<p><tt><font size=-1># We do not have FORWARD rules; forwarding is off</font></tt>
<p>Such a firewall (which you should adapt to your personal tastes and
needs) will provide an extra 'shell' around the system. But basically,
the security of your system should not <i>depend</i> on the firewall; if
only because firewalls are complicated things, and it is far too easy to
make mistakes with them. Many other things can be done first to ensure
the privacy of your network.
<h3>
11. References and further reading</h3>

<ol>
<li>
<a NAME="fortress1"></a><i><a href="http://rootprompt.org/article.php3?article=903">Amateur
Fortress Building in Linux, part 1</a></i>, by Sander Plomp (rootprompt.org)</li>

<li>
<a NAME="fortress2"></a><i><a href="http://rootprompt.org/article.php3?article=931">Amateur
Fortress Building in Linux, part 2</a></i>, by Sander Plomp (rootprompt.org)</li>

<li>
<a NAME="rwls"></a><i><a href="http://www.realworldlinuxsecurity.com">Real
World Linux Security</a></i>, by Bob Toxen (Prentice-Hall, 2001).</li>

<li>
<i><a href="http://logi.cc/linux/reject_or_deny.html">What is the difference
between REJECT and DENY?</a></i> (Linux@home)</li>

<li>
<i><a href="http://logi.cc/linux/ipchains-log-format.html">Ipchains log
format</a></i> (Linux@home). For understanding what you see while running
<tt>tail
-f /var/log/syslog</tt>.</li>

<li>
<i><a href="http://www.isi.edu/in-notes/iana/assignments/icmp-parameters">ICMP
Type numbers</a></i> (IANA)</li>

<li>
<a NAME="jwexim"></a><i><a href="../issue43/stumpel.html">Setting
Up Mail for a Home Network Using Exim</a></i> (Linux Gazette, July 1999)</li>

<li>
<a NAME="jwsmtp"></a><i><a href="../issue45/stumpel.html">Experiments
with SMTP</a></i> (Linux Gazette, September 1999)</li>
</ol>




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Jan Stumpel.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <HR> <P> 
<!--===================================================================-->

<center>
<H1><font color="maroon">Speeding Up Your Net Browsing with PDNSD Domain Name Caching</font></H1>
<H4>By <a href="mailto:sunil_tt@yahoo.com">Sunil Thomas Thonikuzhiyil</a></H4>
</center>
<P> <HR> <P>  

<!-- END header -->




<p>1. Where to find this document
<br>2. About PDNSD
<br>3. Installation
<br>4. Sample configuration file
<br>5. Tweaking configuration files
<br>6. FAQs
<br>7. Credits
<h2>
1. Where to find this document</h2>

<p><br><a href="http://geocities.com/sunil_tt/pdnsd.txt">http://geocities.com/sunil_tt/pdnsd.txt</a>
<h2>
2. About PDNSD.</h2>
DNS is is the Domain Name System. DNS converts machine names to the IP
addresses that all machines on the net have. Name serving on Unix is commonly
done by a program called named. This is a part of the ``BIND'' package
which is coordinated by Paul Vixie for The Internet Software Consortium.
<p>PDNSD is a caching DNS proxy server. Unlike BIND, it saves the RAM cache
to a file and the same is read up by PDNSD for the next Dial-Up session.
BIND when acting as a cacheing nameserver on your local Dial-Up machine
stores/caches the name to number translation data in your RAM. This is
not written back to the hard-disk upon disconnection 'coz it is not intended
for a user/site who/which is not always connected to the Net.
<p>PDNSD can be configured to speed up Net Surfing on a Dial-Up connection.
Since DNS resolution is referenced from the cached file, time is not wasted
on the name to number lookup. This in turn speeds up the name to number
translation, which actually accelerates your surfing.
<p>PDNSD is distributed under the GNU/GPL and is available for download
at: <A HREF="http://home.t-online.de/home/Moestl/">http://home.t-online.de/home/Moestl/</A>
<p>Redhat RPMS are at:
<p><a href="http://home.t-online.de/home/Moestl/">http://home.t-online.de/home/Moestl/</a>
<p>Debian DEBS are at:
<p><a href="ftp://ftp.debian.org/debian/pool/main/p/pdnsd/pdnsd_1.1.2.a-2_i386.deb">ftp://ftp.debian.org/debian/pool/main/p/pdnsd/pdnsd_1.1.2.a-2_i386.deb</a>
<h2>
3. Installation.</h2>
Download pdnsd-&lt;version&gt;.tar.gz from the above source.
<p>Decompress and untar using
<pre>tar zxvf pdnsd-&lt;version&gt;.tar.gz</pre>
Change directory to pdnsd-&lt;version&gt; and type
<pre>$ ./configure</pre>
Configure script accepts a number of parameters, see manual.txt file supplied
with the PDNSD source. Command line parameters --prefix and --with-distribution
are interesting.
<pre>$ ./configure --help</pre>
will list all options
<p>I am assuming that you have not specified any command line options.
Makefile generated by configure will have the following defaults:
(It is worth taking a look at the generated Makefile)
<p>Default installion directory for PDNSD is /usr/local (can be changed
with --prefix option to configure).
Default location of the PDNSD cache is /var/cache/pdnsd.
PDNSD configuration file pdnsd.conf will be found in /etc.
<p>Now type:
<pre>$ make</pre>
This will compile pdnsd. I did not face any problem compiling it on both
Debian 2.2 and Redhat 6.1. Next su to root, as installation requires root
previleges. Then type:
<pre># make install</pre>
This step will do the following (quoted from pdnsd manual.txt):

<OL>
<LI>Copies pdnsd to $(prefix)/sbin/ <P>
<LI>Copies pdnsd-ctl to $(prefix)/sbin/ <P>
<LI>Copies docs/pdnsd.conf (a sample configuration) to /etc/ (and backs
up /etc/pdnsd.conf to /etc/pdnsd.conf.old). If you have an /etc/pdnsd.conf.old
you do not want to be overwritten, save it to another place/name before
doing 'make install' <P>
<LI>Creates your cache directory if it is not there. After installation,
you should check the file permissions and edit /etc/pdnsd.conf to fit your
needs . If you use the run_as option, please make sure that your cache
directory is owned by the user you specified with this option! Please note
that the permission issue has been fixed as of the latest releases. Now
/usr/local/sbin will contain two binaries pdnsd and pdnsd-ctl. 'pdnsd'
is the proxy DNS daemon and 'pdnsd-ctl' is a program to control the cache. <P>
</OL>

<p>The cache is located at /var/cache/pdnsd/pdnsd.cache.The cache file
size will be 4 initially and will grow as and when you browse.This aspect
of the /etc/pdnsd.conf viz. 'perm_cache=&lt;value&gt;;'. By default it is
set as 512(KB).Increase it according to your judgement and a safe value
would be 2048(KB) for a machine having 64MB RAM. The cache file size will
be 4 bytes initially and will grow as and when you browse. Cache growth will be
observed only after a reboot or after restart of the PDNSD daemon. This is due to the fact that PDNSD saves the RAM cache upon exit only.
<p>PDNSD must be started up each time you boot the system . For this, you
have to install start up scripts. rc folder of the source distribution
contains startup scripts for Redhat, SuSE and Debian. I have not tested
the SuSE scripts.
<p>Do the following depending on your distribution.
<h3>
3.a) Debian GNU/Linux.</h3>
Copy pdnsd-{version}/src/rc/Debian/pdnsd to /etc/init.d and type update-rc.d
Stop bind if you have it installed on your system. Edit /etc/resolv.conf
and add the following.
<pre>
      nameserver 127.0.0.1
 </pre>
<p>Comment out entries for all other name servers. Start pdnsd by typing
/etc/init.d/pdnsd start. Test pdnsd by typing nslookup. On my system it
diplays:
<pre>
     Default Server: debian
     Address: 127.0.0.1
     &gt;
</pre>
<p>Stop pdnsd by typing /etc/init.d/pdnsd stop.
<br>Fire up your editor and add a line like this to the end of your /etc/hosts
file:
<pre>
     127.0.0.2    testhost
</pre>
<p>Save the file and start pdnsd once again. Type nslookup. Inside nslookup
type 'testhost'.
<pre>
    &gt; testhost
      Server: debian
      Address: 127.0.0.1
      Non-authoritative answer:
      Name: testhost
      Address: 127.0.0.2
</pre>
If this answer is obtained it shows that your pdnsd is working (remember
to remove the last line from /etc/hosts)
<h3>
3.b) Redhat Linux</h3>

<p><br>Copy pdnsd-{version}/src/rc/Redhat/pdnsd to /etc/rc.d/init.d Stop
bind if you have it installed on your system. Edit /etc/resolv.conf and
add the following
 <pre>
    nameserver 127.0.0.1
 </pre>
<p>Comment out entries for all other name servers. Start pdnsd by typing
/etc/rc.d/init.d/pdnsd start. Test pdnsd by typing nslookup. On my system
it diplays
<pre>
    Default Server: Redhat
    Address: 127.0.0.1
    &gt;
</pre>

<p>Stop pdnsd by typing /etc/rc.d/init.d/pdnsd stop.
<br>Fire up your editor and add a line like this to the end of your /etc/hosts
file.
<pre> 
    127.0.0.2    testhost
</pre>
<p>Save the file and again start pdnsd. Type nslookup. Inside nslookup
type 'testhost'.
<pre>
  &gt; testhost
  Server: Redhat
  Address: 127.0.0.1
  Non-authoritative answer:
  Name: testhost
  Address: 127.0.0.2
 </pre>
<p>If this answer is obtained it shows that your pdnsd is working (remember
to remove the last line from /etc/hosts)
<h2>
4. Sample configuration file.</h2>
My pdnsd.conf looks like this
<HR NOSHADE>
<pre>
global {
 perm_cache=2048;
 cache_dir="/var/cache/pdnsd";
 max_ttl=204800;
 run_as="nobody";
 paranoid=on;
 server_port=53;
 server_ip="127.0.0.1";
}
server {
 ip="202.54.6.5";
 timeout=260;
 interval=900;
 uptest=none;
 ping_timeout=500;
 purge_cache=off;
 caching=on;
}
server {
 ip="202.54.1.30";
 timeout=260;
 interval=900;
 uptest=none;
 ping_timeout=500;
 purge_cache=off;
 caching=on;
}
server {
 ip="202.9.128.6";
 timeout=260;
 interval=900;
 uptest=none;
 ping_timeout=500;
 purge_cache=off;
 caching=on;
}
source {
 ttl=86400;
 owner="localhost.";
 serve_aliases=on;
 file="/etc/hosts";
}
/*
rr {
 ttl=86400;
 owner="localhost.";
 name="localhost.";
 a="127.0.0.1";
 soa="localhost.","root.localhost.",42,86400,900,86400,86400;
}
rr {
 ttl=86400;
 owner="localhost.";
 name="1.0.0.127.in-addr.arpa.";
 ptr="localhost.";
 soa="localhost.","root.localhost.",42,86400,900,86400,86400;
} */</pre>
<HR NOSHADE>


<p>This is a sample working configuration (DNS servers are of VSNL
an Indian ISP).You must edit servers section of pdnsd.conf to suit
your needs.(Fill in DNS servers of your ISP aginst IP entry). Start
PDNSD once more and connect to the Internet. Type nslookup and do a query
for say, yahoo.com. The server will respond somthing like:
<pre>&gt; yahoo.com
Server: debian
Address: 127.0.0.1
Non-authoritative answer:
Name: yahoo.com
Addresses: 204.71.200.245</pre>
Stop PDNSD and disconnect from the Internet. Start PDNSD again and query
for yahoo.com through nslookup. If you are geting the same answer as above
, fine have a coffee and relax. Else if, there is something wrong and ....??
<h2>
5. Tweaking configuration files.</h2>
If you are using BIND as your primary nameserver, one can very well make
PDNSD the secondary one. But here you have Catch-22 situation, on which
local IP and port would you make PDNSD listen ? Look at ragOO's pdnsd.conf
file and named.conf file:
<p>[pdnsd.conf]
<pre>global {
perm_cache=2048;
cache_dir="/var/cache/pdnsd";
max_ttl=604800;
run_as="nobody";
paranoid=off;
server_port=53
server_ip="127.0.0.2";
}</pre>
[named.conf--relevant section only]
<pre>options {
directory "/var/cache/bind";
forward first;
forwarders {127.0.0.2;202.54.6.1;202.54.1.30};
};</pre>
ragOO's GNU/Linux machine has local (lo) IP addresses from 127.0.0.1 to
8.This is the same in all GNU/Linux systems and one has the option to specify
127.0.0.2 to be the alternate local server. PDNSD listens on Port 53 and
note that 127.0.0.2 is the first forwarder in named.conf. This means that
your machine/BIND looks up the PDNSD cached records for a number match
of the address you/the client program has requested, if not there then
it queries the DNS resolvers of your ISP; in order it goes.
<br>
<br>
<h2>
6. FAQs.</h2>
The follwing question and answer is from the correspondence I had with
Thoams Meostl author of pdnsd.
<p>Q. I had some problem with your default installation. The cache was
not growing. It was stuck at 4 bytes. I changed permissions to 'nobody'
and it started growing. Probably a problem with my configuration. Will
you please let me know the correct file permissions for /var/cache/pdnsd
and /var/cache/pdnsd/pdnsd.cache ?
<p>A. The best thing is to give the user who runs pdnsd write permissions
to the cache directory (and of course to the cache file): chown &lt;user&gt;
/var/cache/pdnsd chmod 0700 /var/cache/pdnsd chown &lt;user&gt; /var/cache/pdnsd.cache
chmod 0600 /var/cache/pdnsd/pdnsd.cache
<p>Where the permissions can of course be more liberal, if you want. The
ones given are the minimum required permissions. The default permissions
"make install" sets on the files are also OK. The only important thing
is to chown the file. Normally, "make install" should also chown the cache
file (maybe a bug? If it didn't for you, please drop me a mail).
<h2>
7. Credits.</h2>
Thanks to the author of this nifty utility, Thomas Moestl for clarifying
certain points and doubts. He made me a better user of PDNSD :-) !
<p>Thanks to <a href="mailto:vmathew@eth.net"> Manoj Victor Mathew</a> 
and <a href="mailto:ragu@vsnl.com"> Raghavendra Bhat</a> (ragOO)
for mentioning about 'pdnsd' during one of the ILUG-Cochin meets. ragOO
edited and modified the draft heavily and encouraged me to keep on modifying
the draft.
<p>Last but not the least, to all users of this elegant program who may
have found this rant useful. Enjoy....;




<!-- *** BEGIN copyright *** -->
<P> <hr> <!-- P --> 
<H5 ALIGN=center>

Copyright &copy; 2001, Sunil Thomas Thonikuzhiyil.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR> 
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


<H4 ALIGN="center">
"Linux Gazette...<I>making Linux just a little more fun!</I>"
</H4>

<P> <hr> <P> 

<H1><font color="maroon">The Back Page</font></H1>
<ul>
<li><a HREF="#authors">About This Month's Authors</a>
<li><a HREF="#notlinux">Not Linux</a>
</ul>

<a name="authors"></a>
<P> <HR> <P> 
<!--======================================================================-->

<center><H3><font color="maroon">About This Month's Authors</font></H3></center>

<P> <HR> <P> 
<!--======================================================================-->

<!-- BEGIN bio -->
<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Shane Collinge</H4>
Part computer programmer, part cartoonist, part Mars Bar. At night, he runs
around in a pair of colorful tights fighting criminals. During the day... well,
he just runs around.  He eats when he's hungry and sleeps when he's sleepy.                                           

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Fernando Correa</H4>
Fernando is a computer analyst just about to finish his 
graduation at Federal University of Rio de Janeiro. Now, he has built
with his staff the best 
<A HREF="http://www.olinux.com.br">Linux portal</A> in Brazil and have further 
plans to improve services and content for their Internet users.

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Rahul Joshi</H4>
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
I am a final year computer engineering student from the Government College of
Engineering, Pune, India. I have been using Linux for about 2 years. I have
also contributed to the <em>Linux Documentation Project</em> by maintaining the
<em>Linux Swap Space mini HOWTO</em>. I was introduced to PVM and MPI during
my final year project and I have implemented some brute force programs of our
project on the <em>PARAM 1000 Supercomputer</em> at the <em>Center for 
Developement of Advanced Computing</em>, University of Pune, using both the
PVM and MPI Libraries.



<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Ned Lilly</H4>
Ned Lilly is vice president of hacker relations for <A
HREF="http://greatbridge.com/">Great Bridge</A>, a company formed to promote,
market and provide professional support services for 
<A HREF="http://www.postgresql.org/">PostgreSQL</A>, the open
source database, and other open source business solutions. He can be reached at
<A HREF="mailto:ned@greatbridge.com">ned@greatbridge.com</A>.

<p> 
<h4><img align=bottom alt="" src="../gx/note.gif">Mark Nielsen</h4>
Mark works at ZING 
(<A HREF="http://www.genericbooks.com">www.genericbooks.com</A>) and 
<A HREF="http://GNUJobs.com">GNUJobs.com</A>. Previously,
Mark founded <A HREF="http://www.gnujobs.com">The Computer Underground</A>.
Mark works on non-profit and volunteer projects which promote free literature
and software. To make a living, he recruits people for GNU related jobs and
also provides solutions for web/database problems using Linux, FreeBSD, Apache,
Zope, Perl, Python, and PostgreSQL.

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Ben Okopnik</H4>
A cyberjack-of-all-trades, Ben wanders the world in his 38' sailboat, building
networks and hacking on hardware and software whenever he runs out of cruising
money. He's been playing and working with computers since the Elder Days
(anybody remember the Elf II?), and isn't about to stop any time soon.


<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Pradeep Padala</H4>
I am a software engineer at Hughes Software Systems. I love hacking and adore
Linux. I graduated this year with a B.E (equivalent to B.S) in Computer Science
and Engineering. My intersets include solving puzzles and playing board games. I
can be reached through 
<A HREF="mailto:p_padala@yahoo.com">p_padala@yahoo.com</A> or
<A HREF="http://pradeeppadala.homestead.com">http://pradeeppadala.homestead.com</A>.

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Dustin Puryear</H4>
Dustin is a professional working in the Information Systems
industry. He is author of "Integrate Linux Solutions into Your Windows
Network," as well as numerous articles for both paper- and online
publications. He may be reached at 
<A HREF="mailto:dpuryear@usa.net">dpuryear@usa.net</A>.

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Jason Steffler</H4>
<a href = "http://www.magma.ca/~jagwar">Jason</a> is a Software Architect for
McHugh Software International.&nbsp;
His computer related interests include: OO programming &amp;
design, Smalltalking, the peopleware aspects of software, and noodl'n around
with Linux.

<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Jan Stumpel</H4>
Jan lives in Oegstgeest, The Netherlands.  He has been
a Linux user since 1995. At the moment he is trying to get a
Debian installation just right.


<P> 
<H4><IMG ALIGN=BOTTOM ALT="" SRC="../gx/note.gif">Sunil Thomas Thonikuzhiyil</H4>
I teach Computer Science at College of Applied Sciences Calicut,India.
I have been hooked on linux since 1996. I have a Masters in Computer
Science from Cochin University. I am interested in all sorts of operating
systems. In my free time I love to listen to Indian classical music.

<!-- END bio -->

<a name="notlinux"></a>
<P> <hr> <P> 
<!--====================================================================-->

<center><H3><font color="maroon">Not Linux</font></H3></center>

<P> <HR> <P> 
<!--======================================================================-->
<P>

<!-- *** BEGIN Not Linux *** -->

<H3>World of Spam</H3>

<H4> Some of the funnier spams found in the <EM>Gazette</EM> mailbox.</H4>

<P> Why not OWN your very own profit center gazette?

<P> Design it to advertise the program(s) you want
to see placed up front!  Tired of advertising for
others?  gazette, don't settle for a small part of the pie...

<P> gazette, YOU can have it ALL!!
gazette, receive a monthly check for referrals generated from
your very own profit center!  Allow the system to auto-
matically send out a personalized message on your behalf...

<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->


<P> I have a domain name for sale that will be exceptionally good for your
business: www.midwifes.com.

<P> Your business will be seen as being more successful. There is prestige in
owing a Dot Com Domain.

<P>  Your business will be seen as being more creditable. A Dot Com domain gives
your business a more established feel.

<BLOCKQUOTE><EM>
	[Doesn't he know that midwives is spelled with a "v"?  Or is it
intentionally misspelled because www.midwives.com is already taken?]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<BLOCKQUOTE><EM>
	[This one isn't spam but it's funny anyway.]
</EM></BLOCKQUOTE>

<P> Indonesia's First Penguin Meets Kangaroos

<P> Trabas, the first Indonesian Penguin, will meet the Kangaroos at The Linux
Business Expo & Open Source Conference on 7-9 March 2001 at Sydney Convention
&amp; Exhibition Center, Australia.

<BLOCKQUOTE><EM>
	[<A HREF="http://www.trabas.com">Trabas</A> makes an Internet Account
Management &amp; Billing System (IAMBS) for ISPs and other Internet-based
service providers, as well as other software products.]
</EM></BLOCKQUOTE>




<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> You are receiving this E-mail because  you signed up for Big Daddy up dates
and you showed interest in motorcycles or motorcycle products.
If you are looking for an E-Zine for bikers look no further!


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> Do you have an interest in Affordable Personal Alcohol Detectors???
Please visit our Drunk Driving Prevention Center at...

<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> This is not considered SPAM.

<BLOCKQUOTE><EM>
	[By whom?]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> You and I have not met, but because you?re a respected business professional
with an interest in improving employee productivity, I would like to offer you
a free preview of the Professional Selling SkillMap(tm).


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> I found your address on a site about wine and spirits, cigar and good
living.  X is a virtual club for all those interested in wine in both a
professional and personal capacity.  You too can be among our 6992 members to
receive our free weekly bulletin.


<BLOCKQUOTE><EM>
	[And you too can be among the 6048 who receive an lg-announce message
every month.]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> We are responding to your request for FREE analysis of your site...
We feel there is very substantial potential to promote your site on the
Internet...

<P> Please REPLY to this email and include your full name, telephone number
and URL.

<BLOCKQUOTE><EM>
	[What?!  They are responding to </EM>us<EM> and they're asking for our
URL?  I thought they already had it....]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> FIRST, I MUST SOLICIT YOUR CONFIDENCE IN THIS TRANSACTION...  YOU HAVE
BEEN RECOMMENDED BY AN ASSOCIATE WHO ASSURED ME IN CONFIDENCE OF YOUR ABILITY
AND RELIABILITY TO PROSECUTE A TRANSACTION OF GREAT MAGNITUDE INVOLVING A
PENDING BUSINESS TRANSACTION REQUIRING MAXIMUM CONFIDENCE.

<P> WE ARE TOP OFFICIALS OF THE FEDERAL GOVERNMENT CONTRACT REVIEW PANEL WHO
ARE INTERESTED IN IMPORTATION OF GOODS INTO OUR COUNTRY WITH FUNDS WHICH ARE
PRESENTLY TRAPPED IN [West African country].  

<P> THE SOURCE OF THIS FUND IS AS FOLLOWS : DURING THE  REGIME OF OUR LATE HEAD
OF STATE, [name], THE GOVERNMENT OFFICIALS SET UP COMPANIES AND AWARDED
THEMSELVES CONTRACTS WHICH WERE GROSSLY OVER-INVOICED IN VARIOUS MINISTRIES...
WE HAVE IDENTIFIED A LOT OF INFLATED CONTRACT FUNDS WHICH ARE 
PRESENTLY FLOATING IN THE CENTRAL BANK OF [that country].

<P> HOWEVER, DUE TO OUR POSITION AS CIVIL SERVANTS AND MEMBERS OF THIS PANEL,
WE CANNOT ACQUIRE THIS MONEY IN OUR NAMES. I HAVE THEREFORE, BEEN DELEGATED AS
A MATTER OF TRUST BY MY COLLEAGUES OF THE PANEL TO LOOK FOR AN OVERSEAS PARTNER
INTO WHOSE ACCOUNT THE SUM OF US$25,000,000.00 (TWENTY  FIVE MILLION UNITED
STATES DOLLARS) WILL BE PAID BY TELEGRAPHIC TRANSFER. HENCE WE ARE WRITING YOU
THIS LETTER. 
   
<P> WE HAVE AGREED TO SHARE THE MONEY THUS:<BR>
1.  70% FOR US (THE OFFICIALS)<BR>
2.  20% FOR THE FOREIGN PARTNER (YOU)<BR>
3.  20% TO BE USED IN SETTLING TAXATION AND ALL LOCAL AND FOREIGN EXPENSES. 
   
<P> IT IS FROM THIS 70% THAT WE WISH TO COMMENCE THE IMPORTATION BUSINESS. 

<P> PLEASE NOTE THAT THIS TRANSACTION IS 100% SAFE AND WE HOPE THAT THE FUNDS 
CAN ARRIVE YOUR ACCOUNT IN LATEST TEN (10) BANKING DAYS FROM THE DATE OF 
RECIEPT OF THE FOLLOWING INFORMATION . A SUITABLE NAME AND BANK ACCOUNT 
INTO WHICH THE FUNDS CAN BE PAID. 

<P> THE ABOVE INFORMATION WILL ENABLE US WRITE LETTERS OF CLAIM AND JOB
DESCRIPTION RESPECTIVELY. THIS WAY WE WILL USE YOUR COMPANY'S NAME TO APPLY FOR
PAYMENTS AND RE-AWARD THE CONTRACT IN YOUR COMPANY NAME. 

<BLOCKQUOTE><EM>
	[Now let me get this straight.  You expect the total taxes and fees
to add up to only 20%?  But the US government will want 50% or more in taxes;
where will the other 30% come from?  
And 70 + 20 + 20 adds up to 110%, not 100%.
Where will <EM>that</EM> money come from?  And I'm supposed to put my 
business' reputation on the line for you?  What do you want to import anyway,
drugs?  And finally, whose money is this anyway?  If it was government 
overpayment then the money belongs to the taxpayers of your country and should
be returned to them.  It's not there for you skim the pork off your civil
service job and get rich on the backs of your countrymen.]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> Dear Friend and Future Millionaire...


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> Mystery Shoppers Needed!
GET PAID to shop at your favorite stores...


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> My name is [name] and I came upon your site and think
there is a GREAT opportunity for us to partner. At [site]
you can create your own branded/private label travel website with
your own banners, logo's, custom design, and graphics for FREE.

<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<PRE>
begin 644 Happy99.exe
M35I0``(````$``\`__\``+@`````````0``:````````````````````````
M``````````````````````$``+H0``X?M`G-(;@!3,TAD)!4:&ES('!R;V=R
M86T@;75S="!B92!R=6X@=6YD97(@5VEN,S(-"B0W```````````````````` 
</PRE>

<BLOCKQUOTE><EM>
	[This one got Your Editor </EM>really<EM> riled up and he wrote, 
"F*** you, [name].  We don't need no stinkin' viruses!"  Then Ben calmed me down
by reminding me that it was probably sent without the querent's knowledge.    
<P> In any case, I can't believe the idiodicy of uuencoding a Windows virus.
Most Windows machines don't have a uudecoder installed.  Or does MS Outlook
uudecode?"
]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->

<P> Dear webmaster,<BR>
Your name was given to me by a colleague who thought
you would be interested in this special opportunity since you are in the bulk
email/internet marketing business.

<BLOCKQUOTE><EM>
	[I am?]
</EM></BLOCKQUOTE>


<HR WIDTH="40%" ALIGN="center"> <!-- ************************************** -->


<BLOCKQUOTE><EM>
	[Unlike the other messages above, which are bona fide e-mails I 
received, this is one <A HREF="mailto:rory@ssc.com">Rory Krause</A> and I made
up.]
</EM></BLOCKQUOTE>

<P> Looking for a desktop OS to go with your Linux servers?  How about
Microsoft Windows?  Your office staff will love the familiar user
interface.  Your tech-support people will no longer have to answer
the question, "What's that funny window key on the keyboard for?"  And   
best yet, it's is Samba-comaptible!!








<HR NOSHADE WIDTH="80%"> <!-- ************************************** -->

<P> Happy Linuxing!

<P> Michael Orr<br>
Editor, <A HREF="http://www.linuxgazette.com/"><i>Linux Gazette</i></A>, <A
HREF="mailto:gazette@ssc.com">gazette@ssc.com</a>
<BR CLEAR="all">

<!-- *** END Not Linux *** -->

<!-- *** BEGIN copyright *** -->
<P> <hr> <P> 
<H5 ALIGN=center>
Copyright &copy; 2001, the Editors of <I>Linux Gazette</I>.<BR>
Copying license <A HREF="../copying.html">http://www.linuxgazette.com/copying.html</A><BR>
Published in Issue 65 of <i>Linux Gazette</i>, April 2001</H5>
<!-- *** END copyright *** -->


</BODY></HTML>