...making Linux just a little more fun!

November 2007 (#144):


This month's answers created by:

[ Sayantini Ghosh, Amit Kumar Saha, Ben Okopnik, Kapil Hari Paranjape, Karl-Heinz Herrmann, René Pfeiffer, Minh Van Nguyen, Neil Youngman, Paul Sephton, Rick Moen, Suramya Tomar, Thomas Adam ]
...and you, our readers!

2-Cent Tips

2 cent tip: juxtapose image and text in LaTeX

Minh Nguyen [nguyenminh2 at gmail.com]

Mon, 8 Oct 2007 10:05:37 +1000

Greetings LG readers,

When using LaTeX, there are times when you might want to place an image and a short paragraph side by side. Here's an illustration of what I'm referring to:

   |image goes here      | short paragraph     |
   |                     | goes here           |
   |                     |                     |
   |                     |                     |
   |                     |                     |
   |                     |                     |
   |                     |                     |
   |                     |                     |
In that case, try using the minipage environment. The following snippet of LaTeX macro can be used to implement the above illustration:

   \end{minipage} \qquad\qquad
   A short paragraph of text goes here.
Play around with the various options to get your desired effect.

Regards Minh Van Nguyen

2 cent tip: boot Slackware 12.0 using grub

Minh Nguyen [nguyenminh2 at gmail.com]

Mon, 8 Oct 2007 09:58:55 +1000

Greetings LG readers,

The default boot loader in Slackware is LILO. On some platforms, such as the IBM ThinkPad R40, my experience is that LILO can take quite some time before it boots Slackware 12.0. You can add the option "compact" to your /etc/lilo.conf to speed up the process, but this is not guaranteed to work on all platforms. An alternative is to use the GRUB boot loader.

If you boot Slackware 12.0 with GNU GRUB 0.97, you might see the following boot message on your screen:


> Testing root filesystem
> * ERROR: Root partition has already been mounted read-write. Cannot check!
> For filesystem checking to work properly, your system must initially mount
> the root partition as read only. Please modify your kernel with 'rdev' so that
> it does this. If you're booting with LILO, add a line:
>    read-only
> to the Linux section in your /etc/lilo.conf and type 'lilo' to reinstall it.
> If you boot from a kernel on a floppy disk, put it in the drive and type:
>    rdev -R /dev/fd0 1
> If you boot from a bootdisk, or with Loadlin, you can add the 'ro' flag.
> This will fix the problem AND eliminate this annoying message. :^)
> Press ENTER to continue.

The boot process hangs until you actually press the ENTER button on your keyboard. To get rid of this boot message and the pause, try the GRUB boot option "ro", or add this option to your /boot/grub/menu.lst. For example, here is a snippet of my /boot/grub/menu.lst:


> # Booting Linux distribution Slackware 12.0
> title Slackware 12.0
>     root (hd0,2)
>     kernel /boot/vmlinuz root=/dev/hda3 vga=0x31a resume=/dev/hda2 splash=verbose showopts ro

Regards Minh Van Nguyen

2 Cent tip: Finding your Kernel using Grub

Martin J Hooper [martinjh at blueyonder.co.uk]

Mon, 08 Oct 2007 15:16:39 +0100

If you have done something such as adding a new partition that changes all your partition numbers so that grub can't find the right kernel do this at a grub prompt:

find <kernel>

Grub will then spit out the fully qualified path to the kernel in question and allow you to edit your menu.lst to suit.

[ Thread continues here (5 messages/4.99kB) ]

Our Mailbag

article "A Question Of Rounding" in issue #143

Ben Okopnik [ben at linuxgazette.net]

Wed, 3 Oct 2007 07:36:22 -0500

On Wed, Oct 03, 2007 at 12:36:53PM +0200, Pierre Habouzit wrote:

>   Hi,
>   I would like to report that the article "A Question Of Rounding" in
> your issue #143 is completely misleading, because its author doesn't
> understand how floating point works. Hence you published an article that
> is particularly wrong.

Thanks for your opinion; I've forwarded your response to the author.

>   I'm sorry, but this article is very wrong, and give false informations
> on a matter that isn't very well understood by many programmers, hence I
> beg you to remove this article, for the sake of the teacher that already
> have to fight against enough preconceived ideas about ieee 754 numbers
> already.

Sorry, that's not in the cards - but we'll be happy to publish your email in the next Mailbag.

I understood, even before I approved the article for publication, that a lot of people had rather strong feelings and opinions on this issue; i.e., the author getting flamed when he tried to file a bug report on this was a bit of a clue. Those opinions, however, don't make him wrong: whatever other evils can be ascribed to Micr0s0ft, their approach to IEEE-754 agrees with his - and is used by the majority of programmers in the world. That's not a guarantee that they (or he) are right - but it certainly implies that his argument stands on firm ground and has merit.

You are, of course, welcome to write an article that presents your viewpoint. If it meets our requirements and guidelines (http://linuxgazette.net/faq/author.html), I'd be happy to publish it.

* Ben Okopnik * Editor-in-Chief, Linux Gazette * http://LinuxGazette.NET *

[ Thread continues here (36 messages/102.86kB) ]

rsync not working when trying to update LG mirror

Suramya Tomar [security at suramya.com]

Wed, 03 Oct 2007 22:33:18 +0530

Hey everyone, I maintain a mirror for LG at my site and recently it was brought to my attention that the mirror had gone out of date. I had the update script set up as a cron job so didn't monitor it. But apparently it has been failing with the following errors:

u38576182:~/public_html/suramya.com/linux/gazette > ./sync_mirror.sh rsync: failed to connect to linuxgazette.net: Connection refused rsync error: error in socket IO (code 10) at clientserver.c(97) rsync: failed to connect to linuxgazette.net: Connection refused rsync error: error in socket IO (code 10) at clientserver.c(97)

The sync script contains the following code:

u38576182:~/public_html/suramya.com/linux/gazette > cat sync_mirror.sh RSYNC_RSH=/usr/bin/ssh export RSYNC_RSH rsync -avz --exclude /ftpfiles/ linuxgazette.net::lg-all www_root rsync -avz linuxgazette.net::lg-ftp ftpfiles

This only happens on this one particular system. When I try it on my local system it runs without issues. Any idea why the lingazette.net is rejecting the rsync connection? If I try to ssh to linuxgazette.net from the server I get a password prompt.

This server has rsync version 2.5.6cvs protocol version 26 installed. Its a shared hosting account so I don't have root access on the server.

Any help would be appreciated

Thanks, Suramya

Name : Suramya Tomar
Homepage URL: http://www.suramya.com

[ Thread continues here (16 messages/21.87kB) ]

480% CPU?

Neil Youngman [ny at youngman.org.uk]

Mon, 22 Oct 2007 13:07:52 +0100

Would anyone care to hazard a guess why top is claiming that the top 3 process are between them using 480% CPU on a dual core box?


[ Thread continues here (6 messages/5.22kB) ]

problem with usb support

nishith datta [nkdiitd2002 at yahoo.com]

Sat, 20 Oct 2007 11:59:07 -0700 (PDT)


I got a hp pavillion dv1000 series laptop and it's loaded with Redhat

I am unable to get the usb support working on it.

I checked out the lsmod and it shows the uhci and ohci drivers inserted.

I however donot see any hid driver or usbcore module. I did a slocate on my system but didnot find them either. I definitely chose to have these supports when compiling my kernel.

What am I doing wrong ?


[ Thread continues here (2 messages/1.70kB) ]

using smp kernel, get 100% cpu usage one one cpu without any real load on the system

jim ruxton [cinetron at passport.ca]

Mon, 08 Oct 2007 14:21:21 -0400

Hi I'm on a P4 running an smp kernel ie.

uname -a : Linux jims-laptop 2.6.20-16-386 #2 Sun Sep 23 19:47:10 UTC 2007 i686 GNU/Linux

For some reason occasionaly one of my cpu's runs away and starts showing 100% cpu usage, however top only shows the program using the most cpu at 7% or less. Where is that 93% going? This is driving me crazy. I have to reboot the machine to get it back to normal. I've tried Fedora and Kubuntu. Though Kubuntu appeared more stable it still happened. Played around with changing max_cstate without success.I've just downgraded to a 386 kernel to see if that helps. This isn't a real solution however. Do you have any suggestions? I've asked around on forums and irc channels without luck. Thanks. Jim

output of cat /proc/cpuinfo :

processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 3.06GHz
stepping : 9
cpu MHz : 3059.352
cache size : 512 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 1
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
bogomips : 6123.68
clflush size : 64

[ Thread continues here (13 messages/29.23kB) ]


Terry T [timburwa at gmail.com]

Mon, 8 Oct 2007 09:16:18 +0200

Hie I am new UNIX.I want to copy 30 files with different names using the following command.

ftp -i -s:filename > logfilename.log

The command works well.

My problem is to type the same command 30 times for each file name. How do I transfer all the 30 files at the same time?

[ Thread continues here (7 messages/6.50kB) ]

A couple of Perl questions...

Jimmy ORegan [joregan at gmail.com]

Tue, 9 Oct 2007 19:20:41 +0100

I have a couple of scripts that almost work, and I was wondering if anyone (Ben? :) could tell me why...

First, I want to convert a list of tags in the IPA PAN's corpus format (subst:pl:dat:f) to Apertium's tag format (n.f.pl.dat). I have this:

use warnings;
use strict;
# tags to replace
my %terms = qw(n nt pri p1 sec p2 ter p3 subst n);
while (<>)
	my @in = split/:/;
	my @out = map { ($terms{$_} ne "") ? $terms{$_} : $_ } @in;
	if ($#out > 3) {
		my $type = $out[3];
		$out[3] = $out[2];
		$out[2] = $out[1];
		$out[1] = $type;
	print join '.', @out;

That's broken, because it only works for tag sets which have more than 4 entries, but changing the if to "($#out >= 3)" gives me this: ".sg.nomxxs.m3" from "xxs:sg:nom:m3". I also get a lot of warnings:

Use of uninitialized value in string ne at foo.pl line 11, <> line 1085.
Use of uninitialized value in string ne at foo.pl line 11, <> line 1086.

Next, I have a list of names extracted from a Polish morphology dictionary[1] that I'm trying to convert to a list of word stems and endings. I have this, which works (aside from a couple of errors):

use warnings;
use strict;
use String::Diff qw/diff_fully/;
use Data::Dumper;
	my $endings = $_;
	my @a = split/, /;
	my $stem = find_stem(@a);
	$endings =~ s/$stem//g;
	print $stem;
	if ($endings =~ /?/) {print ":n.f:";}
	elsif ($endings =~ /owie/) {print ":n.m1:";}
	else {print ":n.??:";}
	print $endings . "\n";
sub test()
	my $test = "Adam, Adama, Adaemie, Adamowi, Adamem, Adamach, Adamami, Adamom";
	my @t = split/, /, $test;
	print find_stem(@t);
	print "\n";
sub find_stem()
	my @in = @_;
	my ($r, $l, $cur, $last);
	my $i=0;
	while ($i<($#in))
		($r, $l) = diff_fully($in[$i], $in[$i+1]);
		$cur = $r->[0]->[1];
		$last = $cur if (!$last);
		if ($cur ne $last) {
			($r, $l) = diff_fully($last, $cur);
			$last = $r->[0]->[1];
	return $last;

but if I change the end of the while() to this:

	else {print ":n.??:";}
	my @ends = split/, /, $endings;
	$endings = join(',', at ends);
	print $endings . "\n";

to sort the endings, it... doesn't. What am I missing?

[1] "S?ownik alternatywny", under the GPL: http://www.kurnik.pl/slownik/odmiany/

[ Thread continues here (5 messages/16.04kB) ]

Please help me :(...

Lin, Hong [hlin at devry.edu]

Mon, 29 Oct 2007 18:09:11 -0500


I have a wired problem and I hope I can get some help in your place.

I received a "Forbidden / You don't have permission to access /~user/index.html on this server." Error while I try to display my web page.

The OP is Fedora 7 Apache is 2.2.6

Permissions of all directories and files are set to rwxr-xr-x leading from /home all the way to the files inside public_html.

I have modified the httpd.conf file to make sure it looks for the /~user/public_html directory.

I did not touch any other file or area. If I "killall httpd" and run /usr/sbin/httpd, then the index.html under the /~user/public_html displays. That means that it seems worked for my purpose.

However, after I run "/etc/init.d/httpd restart", it displayed stopping httpd [ok] starting httpd [ok], but I will not be able to see my index.html file under the /~user/public_html

The system index.html (the testing page) always worked.

The wired thing is after I taking out all the #lines within the /etc/init.d/httpd file, I can use "/etc/init.d/httpd restart" to make it work.

However, when the machine reboots, it does not work again. I have tried to put "/etc/init.d/httpd restart" at end of the rc.local to force the web serve stop and start at the boot. The server will stop and then start, but it does not display my web page.

I have also tried to put "/etc/init.d/httpd stop" and "/usr/sbin/httpd" at end of rc.local. It does not display my web either.

The S85httpd is linked with .../init.d/httpd

I have tried with "/usr/sbin/apachectl start", it does not work either manually or inside the rc.local

Here is the summary:

Manually /usr/sbin/httpd worked (after I killed all the httpd process)

Manually /etc/init.d/httpd worked (only after I took out all the #lines, and I don't know why it matters)

Apachectl never worked.

The machine boot up does not work.

May you help me?

Thanks in advance.


[ Thread continues here (5 messages/11.51kB) ]

Copyright & Copyleft

Ramanathan Muthaiah [rus.cahimb at gmail.com]

Wed, 3 Oct 2007 00:05:16 +0530

Just wondering, why all the GNU/Linux man pages refer to copyright only and do not have no reference / brief statement regd "copyleft".

Quicky browsing through the "http://www.gnu.org/copyleft/copyleft.html", I did understand that this concept of copyleft is incorporated by applying copyright and then adding terms of distribution.

To quote from the above URL,

"To copyleft a program, we first state that it is copyrighted; then we add distribution terms, which are a legal instrument that gives everyone the rights to use, modify, and redistribute the program's code or any program derived from it but only if the distribution terms are unchanged . . . "

Any thoughts ?


[ Thread continues here (3 messages/3.03kB) ]

Thoughts on "Mercurial"

Amit Kumar Saha [amitsaha.in at gmail.com]

Sun, 28 Oct 2007 12:09:51 +0530

Hi all!

I am working on an article titled "Distributed Version Control with Mercurial".

Basically, I am a newbie to Mercurial, so the article will be a result of learning, and putting it in a simple, as-i-did-it way.

1. Now, I would like to know thoughts, experiences on Mercurial from those of you who have already used it.

2. If you have made the switch from CVS/SVN to mercurial, then why did you do it?

3. Automated tools you use to facilitate CVS/SVN repos to Mercurial.

All comments, insights are welcome.

Thanks, Amit

Amit Kumar Saha
*NetBeans Community Docs
Contribution Coordinator*
me blogs@ http://amitksaha.blogspot.com

Version control for /etc

Kapil Hari Paranjape [kapil at imsc.res.in]

Fri, 26 Oct 2007 08:53:02 +0530


I was looking at version control mechanisms to handle /etc on the machines here. If people on TAG have used such systems I would appreciate feedback.

CVS: seems to be the classic solution. Cons: People say it is old and unmaintained code which is "end-of-life".

Mercurial: One of the modern VC systems considered "notable" on Rick Moen's knowledge base. One difficulty with "hg" is that it insists on the "distributed" model. Putting the version control history outside /etc (a la CVS) would require convoluted mounts.

GIT: Another modern VC system (though not "notable" as per Rick's kb). It is rather similar to Mercurial in many ways. One difference is that one can use the environment variable GIT_DIR to point to a different directory for storing VC history.

Some reasons to keep VC history outside /etc: 1. This way one can easily check for "cruft" without adding an explicit "ignore" for ".hg" or ".git" ... 2. Uses less space in "/etc". 3. Can keep the history on an "archival" disk safe from potential corruption.

Any thoughts/suggestions by people on TAG are welcome as usual!


Kapil. --

[ Thread continues here (15 messages/30.26kB) ]

Question on how to block a ssh host from being used as a Socks proxy

Suramya Tomar [security at suramya.com]

Fri, 19 Oct 2007 04:21:16 +0530

Hey Everyone, I have been using a SOCKs proxy via SSH (using port tunneling [1]) to browse the net from unsecure locations and it works great.

However I have noticed that when I connect to certain hosts I am unable to use the connection as a SOCKS proxy and I was wondering how these hosts were configured to do this. It seems like a good feature to have on servers that I configure. Are there any disadvantages to this setup that I am missing?

I have tried looking for a solution online but I guess I am not asking the right questions because I didn't find anything useful. So any idea's/suggestions on what/where to look?

Thanks in advance.

- Suramya

[1] To set up a SOCKS proxy using SSH from a windows system follow these steps:

Open PuTTY. You should be greeted with a configuration screen. First, you will enter the hostname or IP address of the SSH server. Type in a name for your connection settings in the box below ?Saved Sessions?, and click the Save button.

Now you need to look at the tree of options to the left; expand the SSH tree, and select ?Tunnels?. Enter 4567 (or any port number above 1024) in the Source Port area, and click the Dynamic radio button to select it. Leave the Destination field blank, and click ?Add?.

Now go back to the Session tree (very top of the left section), and save again. You will be prompted to enter a username, which is the username of your shell account. Type that in, hit enter, and then type in your password when it prompts you.

In your browser change the proxy setting to localhost and the port you used earlier and you can browse the net safely.

Name : Suramya Tomar
Homepage URL: http://www.suramya.com

[ Thread continues here (6 messages/9.88kB) ]

Apertium: an open source machine translation system

Jimmy ORegan [joregan at gmail.com]

Mon, 8 Oct 2007 17:05:38 +0100

On 30/09/2007, Jimmy O'Regan <joregan at gmail.com> wrote:

> Apertium (http://xixona.dlsi.ua.es/apertium-www/) is an open source
> machine translation system.


> Polish is still[1] my main language of interest, and I'm working on a
> configuration for the morphological analyzer.

The start of my Polish-English module was added to SVN today (http://apertium.svn.sourceforge.net/viewvc/apertium/apertium-en-pl/) and can be tested here: http://xixona.dlsi.ua.es/testing/index.php

It's nothing more than a tiny inflecting dictionary at the moment, but I'm working on it :)

Talkback: Discuss this article with The Answer Gang

Published in Issue 144 of Linux Gazette, November 2007



[ In reference to "Mailbag" in LG#143 ]

Etienne Lorrain [etienne_lorrain at yahoo.fr]

Wed, 3 Oct 2007 15:13:03 +0200 (CEST)

Clock problem

>Hi , my name is Jimmy,can anyone help me fix the time on my PC.  I
>change the Battery so many time ,I mean the new CMOS Battery ,but my
>time is still not read correct

Check the board jumper which reset the CMOS configuration and your motherboard documentation - this jumper shall not be let in the "reset" position while operating the computer...

Just a guess.


[ In reference to "A Question Of Rounding" in LG#143 ]

Mauro Orlandini - IASF/Bologna [orlandini at iasfbo.inaf.it]

Tue, 2 Oct 2007 09:53:23 +0200 (CEST)

Talking about rounding, I almost went crazy to find out why a perl program I was writing did not give the right result. Here is the code:

$f = 4.95;
$F = (4.8+5.1)/2;
$F_spr = sprintf("%.10f", (4.8+5.1)/2);
printf "     f: %32.30f\n", $f;
printf "     F: %32.30f\n", $F;
printf " F_spr: %32.30f\n", $F_spr;
if ($f == $F) { print "\n f equals F!\n\n"; }

The test fails even if $f and $F are the same! I solved by using $F_spr instead of $F, but it took me a day to find it out... 8-(

Ciao, Mauro

  _^_                                                                    _^_
( _ )------------------------------------------------------------------( _ )
  | / | Mauro Orlandini          Email: orlandini at iasfbo.inaf.it           | \ |
  | / | INAF/IASF Bologna        Voice: +39-051-639-8667                   | \ |
  | / | Via Gobetti 101          Fax:   +39-051-639-8723                   | \ |
  | / | 40129 Bologna - Italy    WWW:   http://www.iasfbo.inaf.it/~mauro/  | \ |
  | / |--------------------------------------------------------------------| \ |
  | / | Today's quote:                                                     | \ |

[ Thread continues here (6 messages/16.09kB) ]

Talkback: 143/anonymous.html

[ In reference to "Linux Console Scrollback" in LG#143 ]

Andre Ferreira [andre.ferreira at safebootbrasil.com.br]

Thu, 18 Oct 2007 10:20:23 -0300


I have a suggestion for the article about the Linux Console Scrollback that appears in Linux Gazette #143 (October 2007).

He talks about the console scoll back function is not enough to see the kernel messages, also he talks about see the message log that don't have everything.

My suggestion is to mention the "kernel ring buffer" that can be accessed and manipulated using the command "dmesg".

Of course change the size of the scroll back buffer in console has other uses than to see the output of the kernel boot, but as he gives this as an example I think is useful to talk about the kernel ring buffer.

Best regards,


Talkback: issue45/lg_tips45.html

[ In reference to "/lg_tips45.html" in LG#issue45 ]

Gary Dale [garydale at torfree.net]

Wed, 17 Oct 2007 19:33:16 -0400

While the coverage of this topic was generally thorough, there was one important point that was missed. If you want to use tar to directly split archives, you cannot use compression. If you want a compressed archive, you must compress it first then use split to break it into appropriate chunks.

However, if you do things this way, you must be aware that each chunk must be good for you to be able to rejoin them later. For example, if you were to split an archive into IG segments to back up to a UDF filesystem (until recent kernels there was a 1G limit on UDF files) such as a DVD-RAM, you must verify the copies to ensure you will be able to restore the original archive later. If one of the segments develops a bad sector, you may lose the entire archive.

Talkback: 135/misc/lg/diagnosing_sata_problems.html

[ In reference to "Linux Console Scrollback" in LG#135 ]

Neil Youngman [ny at youngman.org.uk]

Thu, 25 Oct 2007 22:02:09 +0100

On Thursday 25 October 2007 20:18, you wrote:

> Hi!
> I read your post in
> http://linuxgazette.net/135/misc/lg/diagnosing_sata_problems.html
> Did you get your sata drive work?
> I've got these kind of problems:
> http://ubuntuforums.org/showthread.php?t=591288
> Can you give any advices?
> -Juho from Finland

I solved my problems by replacing the SATA card. I'm afraid I can't offer much advice. If it was working before you upgraded and you haven't touched the hardware then it sounds like software. I'd try a few different live CDs (Knoppix, Mint, etc), see which ones work and look at whether there are obvious differences, such as modules loaded. That might offer some clues.

I've Cced the Answer Gang to see if they can offer any more useful advice.


Talkback: 108/bilbrey.html

[ In reference to "Using a Non-Default GUI (in RHEL and kin) " in LG#108 ]

Brian Bilbrey [bilbrey at orbdesigns.com]

Wed, 31 Oct 2007 18:29:19 -0400

On Wed, Oct 31, 2007 at 12:15:46PM -0700, nishith datta wrote:

>    hi brian,
>    Sorry if I am bothering you with this email.
>    This is in connection with your article about winmans in linuxGazette. It
>    was very informative and good. I was struggling with installing fluxbox on
>    RHEL 4.0 and I finally could do it with help of your article.
>    Just that I have not understood what is the purpose of  the following
>    files /etc/X11/gdm/Sessions/fluxbox and the
>    /etc/X11/dm/Sessions/fluxbox.desktop.
>    Both have a exec command and I have entered the fluxbox binary only in the
>    fluxbox.desktop file . It is working fine . What is the fluxbox file in
>    gdm/Sessions dir for ?
>    my files look like this :-
>       /etc/X11/gdm/Sessions/fluxbox file
>        #!/bin/bash
>        exec /etc/X11/xdm/Xsession FluxBox
>        /etc/X11/dm/Sessions/fluxbox.desktop file
>        [Desktop Entry]
>        Encoding=UTF-8
>        Name=FluxBox
>        Comment=This session logs you into fluxbox
>        Exec=/opt/fluxbox-1.0.0/src/fluxbox
>        Icon=
>        Type=Application
>    I hate to bother you . Hope it is alright and you will help me out in
>    understanding things better.

No problem, Nishith. You're referring to this article, I think:


I would hazard to guess that /etc/X11/dm is the default display manager directory, where by default I mean "the place where Red Hat expects to find display manager session stuff". Then there's directories for the assorted actual display managers: xdm, kdm, gdm. That's for stock X, KDE, and Gnome, respectively. But in each of those login manager screens, you can select which window manager you want to use for that (and optionally future) X sessions. So in each display manager configuration directory, there are session files designed to work with that particular display manager, for each installed Window Manager. So, for instance, if you installed RHEL4 in kitchen sink mode, with both KDE and Gnome goo, then you're likely to find these:

/etc/X11/ dm xdm gdm kdm

Then, when you install a non-stock RPM of, say, fluxbox, you'll likely find separate files for that window manager to define a session under each of those DMs. That they're different (which seems to be your confusion) is a function of the fact that each is a configuration file for a different display manager.

But all of the others eventually refer back to files in /etc/X11/dm/Sessions for the setup and startup of the window manager. In one Red Hat box I can touch, I find /etc/X11/gdm/Sessions/GNOME. That invokes /etc/X11/xdm/Xsession with an argument of "gnome". Looking at Xsession, I see that ... it leads elsewhere entirely: /usr/share/switchdesk.

[ ... ]

[ Thread continues here (1 message/4.29kB) ]

Talkback: 141/misc/lg/backup_software_strategies.html

[ In reference to "Linux Console Scrollback" in LG#141 ]

Kapil Hari Paranjape [kapil at imsc.res.in]

Mon, 22 Oct 2007 11:05:33 +0530


There was a discussion a while ago about Backup Strategies.

The enclosed URL points to some pre-strategic thoughts on Backup. I have not yet used this thinking to formulate a proper Backup plan that is consistent with it.


I have not yet submitted it as an article for LG since the actual strategy is missing --- what good is theory without a practical formula?!


Kapil. --

"A Question Of Rounding" in issue #143

Paul Sephton [paul at inet.co.za]

Sat, 06 Oct 2007 19:53:57 +0200

Hello, all

In order to clarify a few points, set some issues to rest, and update the listing of the source code that appeared in "A Question Of Rounding", the following may be of interest.

First of all, I have to state unequivocally that there is no bug. There never was a bug. The printf() family of functions completely follow the letter of both the C99 and IEEE specifications (at least for GLibC). The matter has been completely explored and put to rest.

The issues brought to light in the article are, however very pertinent to those of us who make a living producing code. The listing below addresses some of the real life issues some of us face when having to deal with floating point arithmetic.

Please refer to the discussion which follows for an explanation as to why printf() behaves the way it does, and what this code does to address some of the problems I highlighted in my article. The discussion is also a summary of some of the many points raised in the bug report discussion, and the information provided to me during that process by a few very smart people.

[ ... ]

[ Thread continues here (1 message/11.46kB) ]

Talkback: Discuss this article with The Answer Gang

Published in Issue 144 of Linux Gazette, November 2007


By Howard Dyckoff

News in General

Ubuntu 7.10 "Gutsy Gibbon" Released

As promised, Ubuntu 7.10 server and desktop were released on 10/18. It offers new server and desktop features, including better hardware and driver support, especially for printers and wireless devices.

Mark ShuttleWorth, CEO of Cannonical, noted that "...this is the 7th release of Ubuntu and the thing we are most proud of... we were with a day of schedule on all releases." Shuttleworth added, "We are delivering what we consider to be an enterprise-class operating system, and we expect these will be used immediately in production environments."

Among the innovations in this edition are "tracker", a hard drive indexing system with the ability to search the entire desktop, "I believe we are the first distro to deliver this capability," Shuttleworth said.

The server package adds more standardized configuration templates, to speed deployments and aid rollout scalability. Configurations are included for Web servers, databases, LAMP applications, file and print servers, etc.

This is a non-LTS release of Ubuntu. The next release, targeted for April 2008, will be a Long-Term-Support [LTS] release for those on support contracts. LTS releases are planned for annual or biannual release.

This latest Ubuntu version also includes the following new features:

Ubuntu Links

Laptop, thin client, and server:

To coincide with the Ubuntu 7.10 launch, Canonical Ltd. announced updates to Edubuntu, Kubuntu, and Xubuntu derivatives, including advanced thin client capabilities and a KDE 4 beta tech preview. Additional information is available at http://www.ubuntu.com/news/ubuntu-family710.

OpenAjax Alliance Promotes InteropFest, Security for Mashups

At the September AjaxWorld conference, the OpenAjax Alliance, which is dedicated to open and interoperable Ajax-based Web technologies, revealed new initiatives for secure mashups and mobile AJAX.

AJAX is the technology behind the increasingly popular "mashup", a Web site or application that combines content from more than one source into an integrated experience. As AJAX and mashups continue to gain widespread acceptance under the Web 2.0 umbrella, it is critical for organizations to understand these technologies and to avoid possible problems by adhering to best practices.

The alliance prepared a new white paper titled "AJAX and Mashup Security", itemizing the ways in which AJAX applications could be attacked, and providing a set of best practice techniques to address each vulnerability. Available at www.openajax.org, the white paper represents the collaborative efforts of AJAX security experts, and was a joint effort with the Marketing Working Group.

"ICEsoft has long recognized that security for enterprise-class applications is a critical requirement," said Robert Lepack, VP of Marketing for ICEsoft Technologies. "We view the publication of the Open Ajax white paper 'AJAX and Mashup Security' to be an important step in the ongoing need to both educate customers on the potential security risks of AJAX applications, and the best practices described in the paper to be a key step toward developing much needed standards."

In addition to a strong focus on security, the OpenAjax Alliance held InteropFest 1.0, which is the final integration testing phase of OpenAjax Hub 1.0. OpenAjax Hub is a small JavaScript library that allows multiple AJAX toolkits to work together on the same page. The central feature is a publish/subscribe event manager, which enables loose assembly and integration of AJAX components. OpenAjax Alliance will deliver both an open specification and a reference open source implementation. Standards are the key to interoperability, and allow the true possibilities of AJAX and Web 2.0 to be realized.

"To further advance the AJAX ecosystem, OpenAjax Alliance members together are developing a standard way to describe AJAX controls and their programmatic interfaces so that it becomes easier for developers to use AJAX libraries with development tools," said Kevin Hakman, director, TIBCO Software Inc. and Chair of the Alliance IDE Working Group. "We're on pace to have an AJAX control description specification ready for early 2008."

With finalization activities on OpenAjax Hub 1.0, the alliance has begun work on OpenAjax Hub 1.1, which will add support for secure mashups and to enable mediated Comet-style client-server messaging. As with OpenAjax Hub 1.1, the alliance will deliver both a specification and a commercial-quality open source reference implementation. The secure mashup features of OpenAjax Hub 1.1 will isolate mashup components in secure "sandboxes" and use the OpenAjax Hub's publish/subscribe features to achieve mediated cross-component messaging.

Also a part of InteropFest 1.0, interoperability certificates were awarded to the following member organizations for their participation: 24SevenOffice, Apache XAP, Dojo Foundation, ILOG, Getahead, IT Mill, Lightstreamer, Microsoft, Nexaweb, Open Link, Open Spot, Software AG, and Tibco. The interoperability event required integration of an organization's AJAX toolkit with the OpenAjax Hub and at least one other AJAX component, where cross-component messaging is accomplished using the OpenAjax Hub. Interestingly, Microsoft's ASP.NET AJAX, formerly called Atlas, passed the interoperability tests: http://www.internetnews.com/dev-news/article.php/3701966

About 90 companies have joined the OpenAjax Alliance since it was formed in 1990. The mashup security white paper is available at http://www.openajax.org/whitepapers/Introducing%20Ajax%20and%20OpenAjax.html

Sun Opens More Open Source

Sun held a developer summit of sorts, Oct. 15th, on its Santa Clara campus, for many of Open Source projects it supports. Besides a report on Java and the OpenJDK project, Sun engineers who are principals in the Postgres project and their Glassfish application server project also reported on progress.

Mark Reinhold, a member of the OpenJDK project's governing board, confirmed that more of Java is now open source, with only about 4% still requiring what he called "binary plugs" for sound and image drivers. He explained that Sun still does not have the rights to make that code available, but is pushing the code owners to release the necessary items or face a "name and shame" initiative from Sun and other contributors to the OpenJDK.

The OpenJDK interim governance board was formed at JavaOne, last spring, and has met in teleconference twice and face-to-face only once before the Sun Summit. According to Reinhold, the plan is to have a draft constitution written by this December, and to get it ratified by Spring 2008. As this independent infrastructure is being built up, Sun is providing leadership and support. Reinhold said that Sun would mostly act as "a benevolent dictator", until OpenJDK could run on its own.

Version 2 of the Glassfish application server, based on Java 5, was released in mid-September. It featured substantial performance improvements, especially in clustering, and billed as enterprise-ready. Sun will develop this Glassfish release as a product for its customers.

GlassFish Version 3 is already under development, and will be a major change in project architecture. It will become fully modular, based on a 100kB kernel, and will also feature multiple class-loaders. The micro-kernel and JVM will have extensions to efficiently run major scripting languages, such as PHP and Ruby.

Josh Berkus, both a long time Sun engineer and a long time contributor to the PostgreSQL project, discussed the coming version of Postgres, as well as its history going back to the 1980s at UC Berkeley.

There are an estimated 15 million "embedded" users of Postgres and over 200 active developers. Postgres is now distributed with Solaris & OpenSolaris.

The presentations concluded with an Open Source Worldwide Panel led by Ian Murdock and attended by developers from Europe, India, and Latin America, including Suvendu Ray and Bruno Souza, among others. One issue discussed was the twin problem of localization and de-localization: not only efficient translation of language, but also mechanisms for sharing back the work of developers collaborating in their local languages. A video of the panel is available from Podtech here:
Open Source Worldwide Panel.

On October 22-23, in parallel with InterOp NY, Sun will host the Start-up Camp event in New York City. The goal of Start-up Camp is to bring entrepreneurs and vendors together to network, share advice, and learn from the experiences of other startups. The event will continue to be a face-to-face, collaborative event, with the agenda determined during the event. Start-up University will kick off Start-up Camp, and provide an opportunity for attendees to learn about various topics in an educational format.

Gphone due in November

The long-rumored Google answer to the Apple iPhone will make its debut in early November, probably on November 5th, according to blogs and Google watch sites. It will be a data-driven, Linux mobile device hosting Google desktop gadgets and Google Web apps.

Unlike Apple, which tried to restrict developer access to its proprietary iPhone platform, Google has been building developer communities for its varied applications, and will probably leverage that support to enhance the new platform. The Google Gears framework allows developers to create widgets and gadgets that can be intermittently connected. And Google already has a partner relationship with Samsung for putting Gmail and Google Search on 3G phones. Google also recently purchased the Finnish social-networking company Jaiku, which helps "friends" find each other via their mobile devices. Perhaps this was the missing piece to Google's vision.

What is uncertain is whether Google will partner with the wireless carriers or compete with them. And will it partner with one or more music services? Also to be determined is if there will be an ad-supported version of the messaging and voice services. Maybe 2008 will be the year of free Mobile Linux.


VMworld 2007 Winners Announced

The Best of VMworld Awards, presented by SearchServerVirtualization.com, recognize the best products shown during VMworld 2007. The overall Best of Show award was given to IBM and VMware, jointly, for the recently announced IBM System x 3950 M2 with VMware's new embedded hypervisor, ESX Server 3i.

In this first year for the Best of VMworld 2007 awards, judges awarded a Gold award and Finalist awards in each of the six award categories. Go here for complete details:


The judges also awarded a New Technology award to Onaro, Inc. for VM Insight, which monitors and manages both physical and virtual machines; to Marathon Technologies Corp. for its everRun FT for XenEnterprise; and to InovaWave, Inc. for the InovaWave VirtualOctane for ESX Server. In addition, Green Storage to AMD for its quad-core processor technology.



NY Technology Forum
November 1-2, New York, NY

2007, November 3-9, Washington, D.C.
at the Hyatt Regency Crystal City.

Interop Berlin, Nov 6-8, 2007

Executive Forum 2007
November 7-8, Millennium Broadway Hotel, New York City.

QCon San Francisco conference
Nov 5-9, 2007, Westin Market Street
Discount code: devtownstation_qconsf2007

OpenWorld San Francisco
November 11-15, 2007,
San Francisco

Supercomputing 2007
Nov 9-16, Tampa, FL

MemCon Tokyo
Nov. 13-14, 2007
Attendance is free for industry professionals

Certicom ECC Conference 2007
November 13-15, Four Seasons Hotel, Toronto
(focused on the use of Suite B crypto algorithms in the enterprise)

Mobile Internet World
November 13-15, 2007, Hynes Center, Boston

IT Governance Compliance Conference
November 14-16, 2007, Boston, Mass.

Large Installation System Administrators (LISA) Conference
November 14-15, 2007, Dallas, Texas, USA

Gartner Identity & Access Management Summit
14-16 November 2007, Los Angeles, CA

Gartner 26th Annual Data Center Conference
November 27-30, Las Vegas, NV


Agile Development Practices Conference
December 3-6, 2007, Shingle Creek Resort, Orlando, FL

Gartner Enterprise Architect Summit
December 5-7, Las Vegas, NV


SPIE Photonics West 2008
19 - 24 January,
San Jose, CA

MacWorld Conference and Expo
January 14-18, San Francisco


PostgreSQL beta 1 for version 8.3

A new version of the venerable Postgres database went into beta in October. Among the features in the new version are:

... and many others. See the release notes here for a more complete list of new features.

In September, EnterpriseDB announced a new version of its EnterpriseDB Postgres, the professional-grade distribution of the open source PostgreSQL database. The new version of EnterpriseDB Postgres is based on Postgres 8.2.5.

The latest release includes a MySQL-to-PostgreSQL Migration Toolkit and a Procedural Language Debugger. The Migration Toolkit, previously available as a proprietary component of EnterpriseDB Advanced Server, the company's flagship commercial database offering, has been re-released under the Artistic License, an OSI-approved open source license.

OpenOffice.org 2.3 Released

The new OpenOffice includes new features in addition to improvements to stability, performance, Microsoft Office compatibility, and accessibility. Some new features are available as free extensions that increase OpenOffice.org functionality.

New Chart: OpenOffice.org 2.3 introduces an all new chart with a new chart wizard for complex charts. The 3D charts include several new charts types, for example:

Also improved is the data source handling where data ranges for columns and rows labels can be separated from the data source range for the chart values, and the ability to select different x-values for different series. There is enhanced automatic logarithmic scaling. Finally, there are performance improvements and an enhanced import and export of Microsoft Office charts.

OpenOffice.org 2.3 is available for download here.

New Feature: OpenOffice.org Extensions are independent add-ons to enhance OpenOffice.org functionality. Sample extensions are:

Sun Report Builder: Now availible for OpenOffice.org 2.3 is the Sun Report Builder for creating stylish and smart database reports. The flexible report editor can define group and page headers as well as group and page footers and even calculation fields are available to accomplish complex database reports.

The Sun Report Builder uses the Pentaho Reporting Flow Engine of Pentaho BI. This Extension is available for free on Sun Report Builder.

Send and receive faxes with the OpenOffice.org eFax Extension:

Replace your fax machine with the eFax extension for OpenOffice.org, an online fax service that eliminates the need for a fax machine. You can continue to use your existing fax number or use a new number for free. When someone faxes to your selected number, the fax is displayed in the eFax Messenger solution on your computer, or the fax is converted to a file that is emailed to you as an attachment.

eFax is a service of J2global; for more information see the eFax web page. This Extension is available for free on OpenOffice.org eFax Extension.

Mandriva Linux 2008.0 Now Out

The 21st release of Mandriva Linux was released in October and includes three editions: "One", "Powerpack", and "Free". Versions "One" and "Free" can be downloaded free of charge from official Mandriva mirrors and via BitTorrent (visit this page for downloads). "Powerpack" is a commercial edition available by purchase from the Mandriva Store. For information on the differences between the editions, visit Mandriva's Choosing the right edition page.

A recent development snapshot of KDE 4 is available as a preview in Mandriva Linux 2008, and NTFS write support is also included. The Mandriva Linux 2008 Errata list known issues with 2008 and how to fix or work around them.

KDE 4 Beta 3 is out, Fedora 8 coming soon

This late-stage beta was available in mid-Oct. and also marks a freeze of the KDE Development Platform. New educational features and enhancements are included with further polishing of the KDE codebase. The aim of the 4.0 release is to put foundations in place for future innovations.

However, KDE 4.0 will probably not be ready for the first release of Fedora 8 in early November. The October Fedora 8 beta included a completely free and open source Java environment called IcedTea, which is derived from the OpenJDK project.

Elpicx 1.1 helps prep for LPI exam

The elpicx 1.1 live DVD was released in mid-October. Elpicx is a Knoppix and CentOS-based Linux system that helps students prepare for the Linux Professional Institute (LPI) certification exam, by providing test emulators and a number of LPI reference cards, study notes, preparation guides, and exam exercises.

Version 1.1 is based on Knoppix 5.1.1 and CentOS 4.3. Visit the project's home page to read the complete release announcement.

MEPIS 7 and its Anti-X due by November

MEPIS 7 RC Beta 5 was released in late September, and another RC is expected around November 10. Then it is on its way toward final release of version 7, at the end of November or early December.

The 7.0 RC 1 release of "antiX", a lightweight derivative of MEPIS, was out in late September. AntiX is built and maintained by a MEPIS community member as a free version of MEPIS for very old 32-bit PC hardware. The antiX Web site is at antix.mepis.org, and an antiX forum is hosted at www.mepislovers.org

AntiX 7.0 RC 1 is like AntiX 7.0 Beta2, except that the look and the packages have been updated to be in sync with MEPIS 7.0 Beta 4. There are also fixes for reported bugs that are unique to AntiX.

AntiX is designed to work on computers with as little as 64MB of RAM and Pentium II or equivalent AMD processors. ISO images and deltas are available in the "testing" subdirectory at the MEPIS Subscriber's Site and at the MEPIS public mirrors.


ASUS Launches Low Cost Eee PC

ASUSTeK will be offering a new low-cost laptop PCs to consumers for as little $250, by year end. These laptops are targeted at non-power-users in markets that are nearing saturation. They will function as personal as opposed to business computers, and may be adopted by teens and tweeners.

Called Eee PCs, the new line of notebooks will initially be available in English and classical Chinese versions, the latter for sale in Taiwan. Other languages may be supported in 2009. Prototypes were shown at Intel IDF in San Francisco.

These low-cost PCs will run both open-source Linux and Windows (but that was a change for the prototypes shown in SF). Windows versions of the computer would cost only about $50 more than Linux versions, about $300, suggesting that Microsoft had offered the Windows operating systems at a big discount in order to play on the Eee. But this still should provide a Linux incentive to consumers.

The Eee will be first laptop to carry the ASUSTeK brand, rather than being OEMed for other companies.

Zpower innovates Laptop Batteries

A safer, lighter, smaller and recyclable battery for computers and phones is poised for production.

ZPower, formerly known as Zinc Matrix Power, has developed advanced, rechargeable silver-zinc batteries. Initially targeted for notebook computers, cell phones, and MP3 players, ZPower batteries currently have 30% greater energy density than traditional lithium-ion batteries. Intel is interested in promoting the technology, and has made a 10% equity investment in the firm.

ZPower showed its wares at the recent Intel IDF conference in SF. This included working silver-zinc batteries for laptops that were half the thickness of current lithium-ion batteries. When in full production in 2008, these batteries will get 20-30% more battery life on a charge. Production goals for 2009 are 20-30% weight reductions and a target of 30-50% greater battery life.

In addition to the performance benefits, ZPower batteries are inherently safer. The technology features a water-based chemistry that, and is not flammable or toxic. The battery contains no lithium or flammable liquids and is therefore free from the problems of thermal runaway, fire, and danger of explosion. Silver-zinc batteries are not subject to airline restrictions.

The new batteries also environmentally friendly. The primary elements used to produce the batteries can be 100 percent recycled and re-used. The raw materials recovered in the recycling process of silver-zinc batteries are the same quality as those that went into the creation of the battery. The company plans to offer cash incentives to consumers recycling the batteries.

Dr. Ross Dueber, a scientist at ZPower, told Linux Gazette that the first production silver-zinc batteries will be about 10mm thick but about the same weight as current batteries. Weight reduction will come from better use of the materials and manufacturing efficiencies. The first manufacturing partner will be Tyco Electric. ZPower is currently working with other leading manufacturers of notebook computers and cell phones to incorporate silver-zinc technology in next generation products.

ZPower won Intel's "Technology Innovation Accelerated" (TIA) Award in the Mobility category at the 2006 Fall IDF conference. The company was also recently named a "GoingGreen 100" winner by AlwaysOn.

Troy Renken, Vice President of Product Planning and Electronics at ZPower Inc. was a featured speaker in the Extended Battery Life panel at the Intel Developer Forum (IDF) in Taipei, Taiwan on October 16th. The panel, titled "Realizing the Vision of All Day and Beyond Battery Life for Mobile PCs," discussed industry initiatives underway to extend battery life.

IBM and Linden Labs Partner on Open 3D Standards for Collaboration

IBM and Linden Lab, creator of the virtual world Second Life (www.secondlife.com), will work with a broad community of partners to drive open standards and interoperability to enable avatars -- the online persona of visitors to these online worlds -- to move from one virtual world to another with ease and support applications of virtual world technology for business and society in commerce, collaboration, education and more.

As more enterprises and consumers explore the 3D Internet, the ecosystem of virtual world hosts, application providers, and IT vendors need to offer a variety of standards-based solutions in order to meet end user requirements. To support this, IBM and Linden Lab are committed to exploring the interoperability of virtual world platforms and technologies, and plan to work with industry-wide efforts to further expand the capabilities of virtual worlds.

"We have built the Second Life Grid as part of the evolution of the Internet," said Ginsu Yoon, vice-president, Business Affairs, Linden Lab. "Linden and IBM shares a vision that interoperability is key to the continued expansion of the 3D Internet, and that this tighter integration will benefit the entire industry. Our open source development of interoperable formats and protocols will accelerate the growth and adoption of all virtual worlds."

IBM and Linden Lab plan to work together on issues concerning the integration of virtual worlds with the current Web; driving security-rich transactions of virtual goods and services; working with the industry to enable interoperability between various virtual worlds; and building more stability and high quality of service into virtual world platforms. These are expected to be key characteristics facing organizations that want to take advantage of virtual worlds for commerce, collaboration, education and other business applications.

More specifically, IBM and Linden Lab plan to collaborate on:

IBM has hosted discussions on virtual world interoperability, the role of standards, and the potential of forming an industry-wide consortium open to all. Linden Lab has formed an Architecture Working Group that describes the road-map for the development of the Second Life Grid. This open collaboration with the community allows users of Second Life to help define the direction of an interoperable, Internet-scale architecture.

For more information about the Second Life Grid, visit http://secondlifegrid.net/.

IBM Introduces its "Virtualization-Ready" System X Server

IBM has previewed the fourth generation of its virtualization chipset technology, X4, to be available in a high-end, scalable server leveraging the latest in quad-core processing technology from Intel. The System x3950 M2 server will debut with a new embedded hypervisor capability, enabling clients to easily deploy virtualized server applications right out of the rack.

X4 chipsets mark significant advances in performance, availability, and processing efficiencies for the IBM System x line of servers. X4 will enable x86 server configurations to fuel virtualization on high-end systems. Several other new features will allow clients to easily adopt virtualization.

The new server will be ready for virtualization right out of the box by eliminating software setup and installation time. An internal USB interface will accommodate chip-based or "embedded" virtualization software preloaded on a 4GB USB flash storage device. The new system offers double the memory slot capacity. Also, four times the amount of memory can be hosted on a single chassis compared to the previous system, enabling more virtualization workloads.

IBM has developed and released three generations of X-Architecture chipsets since 1997, and remains the only top-tier vendor in the industry to incorporate its own chipset in Intel-based servers. The third generation chipset, X3, introduced in 2005, was optimized for server consolidation and enterprise application software.

Sun Intros Smallest Enterprise-Class Four-Socket Quad-Core Server

In late September, Sun Microsystems introduced its first quad-core x64 systems, which deliver up to twice the expandability and compute power as other servers at half the size. The Sun Fire X4450 and Sun Fire X4150 servers, powered by Quad-Core Intel Xeon processors, solve critical problems in the datacenter by offering more performance, higher density, and better power efficiency than do competing systems. Both servers also can run Solaris, Linux, Windows, or VMware.

The Sun Fire X4450 server is smallest four-socket 16-way rackmount server in a 2U form factor, and saves as much as 50 percent of the energy consumption tof other servers, resulting in lower energy and cooling costs. The Sun Fire X4150 server, powered by the Quad-Core Intel Xeon processor 5300 series, is a two-socket 1U system with up to twice the memory, internal storage, and networking connectivity as competitive two-socket 1U servers. With more than 1 terabyte of high-performance internal disk storage, it is a good solution for database and other disk-intensive applications.

The Sun Fire X4150 server has entry-level pricing starting at $2,995. The Sun Fire X4450 has entry-level pricing starting at $8,895. For more information on the Sun Fire X4450 and X4150 servers, please visit: http://www.sun.com/X4450 and http://www.sun.com/X4150

A video of the server announcement is available here: http://sunfeedroom.sun.com/linking/index.jsp?skin=oneclip&fr_story=FRsupt215388&hl=false

--and all video segments here:

http://sunfeedroom.sun.com [...]

At the announcement, Pat Gelsinger, Senior Vice-President and General Manager for Intel's Digital Enterprise Group, spoke about Intel-planned developments at the 45 nm level. Sun also announced new Solaris features, including a graphical installer, and enhancements to its famous D-Trace, including integration with its NetBean Studio developer tools.

Talkback: Discuss this article with The Answer Gang

Bio picture

Howard Dyckoff is a long term IT professional with primary experience at Fortune 100 and 200 firms. Before his IT career, he worked for Aviation Week and Space Technology magazine and before that used to edit SkyCom, a newsletter for astronomers and rocketeers. He hails from the Republic of Brooklyn [and Polytechnic Institute] and now, after several trips to Himalayan mountain tops, resides in the SF Bay Area with a large book collection and several pet rocks.

Howard maintains the Technology-Events blog at blogspot.com from which he contributes the Events listing for Linux Gazette. Visit the blog to preview some of the next month's NewsBytes Events.

Copyright © 2007, Howard Dyckoff. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007

Virtualization made Easy

By Edgar Howell


Recently, I finally found the time to check out something that had originally been brought to my attention by an announcement in a local publication early this year: Innotek's VirtualBox. In short, this is an outstanding product that deserves consideration by anybody in any way interested in virtualization.


For a very long time, VMWare has been the name in virtualization. A year or so ago, I played with their VMPlayer and what it could do was very impressive, but an OS under it was noticeably slower than natively on hardware. My understanding was that it could emulate any environment, but at a considerable price.

Xen has been the "other" contender for a while. When I was experimenting with it, it seemed to address only servers. And required modifications to both operating systems involved.

VirtualBox, as I understand it, is in a sense between the two. Ignoring the optional "guest additions", it needs no alteration to either OS. Rather, it provides a layer of software that makes available one particular machine configuration and can trap "dangerous" instructions and patch them on-the-fly such that the OS in the VM can execute without prior modification. Due to the patching, this only needs to be done once. In my experience, this seems to work quite well.


Typically, the term "host" refers to a combination of an operating system and a PC or other computer (i.e., software + hardware), as in the name of a GNU/Linux host in a network. But in the context of virtualization:
• "host" is the OS immediately interacting with the hardware under which, in this case, VirtualBox has been installed;
• "guest" is the OS installed in this environment and accessing the virtual hardware made available by VirtualBox running under control of the host.

The VirtualBox documentation talks about 2 states, "powered off" and "saved". This is a simplification that might cause some grief for those not used to dual-boot:
• "Saved" is likely the state of greatest interest. When you tell VirtualBox to save the machine state, it is much like making use of the swap partition to save the current state of an OS to speed up later re-boot after shutdown.
• "Powered off" has precisely the same effect as pulling the power plug on hardware: buffers will not be flushed, you almost certainly will lose data, on re-boot you will go through the fsck procedure. This may make sense in some situations as when you so mess up an installation that you want to start over anyhow.

On those -- likely rare -- occasions when you do not want to save the current machine state of the guest OS, in all probability, you will want to use the shutdown you would use normally, whether via GUI or root: "shutdown -h now". This leads to the state "powered off", but without loss of data.


This is professional-quality software with corresponding documentation. Installation is extremely straight-forward. This is true of both VirtualBox itself and installing virtual machines under it. It works as described in the documentation, and is quite easy to use. No need for comments, other than that you should be sure to allow enough room in the partition where the host resides.

Here are some statistics from a notebook:
Host partitions:

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sda12            27774776  21330516   5033356  81% /
tmpfs                   518276         0    518276   0% /lib/init/rw
udev                     10240       116     10124   2% /dev
tmpfs                   518276         0    518276   0% /dev/shm
/dev/sda9              4814936    109892   4460456   3% /shareEXT2
total 14488176

Guest machines available in the root partition:

-rw------- 1 web web   13634048 2007-09-06 11:30 Borland.vdi
-rw------- 1 web web 1844457984 2007-09-18 10:50 Debian-4.0a.vdi
-rw------- 1 web web 2926592512 2007-09-13 17:17 Debian-4.0.vdi
-rw------- 1 web web 2346725888 2007-09-26 14:19 Debian-4.0-xfce.vdi
-rw------- 1 web web 2905618944 2007-09-08 20:36 Fedora-Core-6.vdi
-rw------- 1 web web      12800 2007-09-09 08:08 Knoppix-5.1.1.vdi
-rw------- 1 web web 2422223360 2007-09-05 12:04 Kubuntu.vdi
-rw------- 1 web web 2775593472 2007-09-10 12:58 Mandriva.vdi

By the way, the above are just "virtual drives" and do not reflect, for example, the Debian Xfce and Knoppix ISO's also taking up space in an almost 30GB partition. Based on past experience I gave each machine 3GB to be used dynamically as needed, not taken and formatted up front.

A Debian Example

Here are a couple of screen-shots of an XFCE-based Debian installation under VirtualBox.

This is what the display shows after using a VM as it is started by VirtualBox.

You can see a bit of the VirtualBox window, the Debian guest itself, and to the extreme left and right the task bars of the Debian host, which I moved there only because of the proportions of the notebook display.

And here it is full-screen.

Not a great deal of difference: on a notebook at least, full-screen doesn't do much more than hide everything other than the VM of interest and the task bars of the host. In particular, the VirtualBox task bar with the buttons to shut down the VM are no longer visible.

Guest Additions

The so-called "Guest Additions" install code in the guest OS to enhance behavior. Innotek recommends installing them, and I concur.

The major enhancements are:
• Coordination of time on VM and host
While there is no absolute necessity of having the VM reflect the correct time of day -- indeed, the IBM-DOS/Borland VM shown on the VirtualBox window in the first screen-shot above can't even display the correct year! -- this seems desirable.
• Sharing directories between VM and host
The VirtualBox documentation refers to "shared folders", but they are directories and very useful if you need to move any data from one environment to another.
• Elimination of "mouse capture"
Natively a VirtualBox VM will not let the mouse move the cursor outside of the area it occupies on the display. If you have more than one VM active on the display at one time, this can be a feature (overridden with the "host" key as documented). I prefer to use the workspaces or desktops or whatever the dickens the GUI-VTs are called to house the one or two VMs simultaneously active (as on the bottom part of the left-most Debian host task bar, above).

Remote Access

Remote access is very easy, and could be a way of obtaining lots more use from old notebooks or PCs. All you really need is a host machine with enough power to support a couple of users, pretty much any PC by now. You can either start a virtual machine via the GUI on the host or use ssh -- you need remote access anyhow -- as in the following from a PC under Knoppix via the LAN to the notebook:

Once the virtual machine is running (a matter of a few seconds, if it had previously been ended with "save state"), you can logon to it with the command seen in the tiny window at the bottom:

At this point, response times are only limited by the network. Internet access is available if the VM has it via the host. The only caveat is that it is very easy to forget to shut down the VM properly; terminating all connections to the VM without doing this leads to the state "aborted", possible loss of data, and the inevitable fsck delay on re-boot.


With no claim to expertise in security matters I hesitate to offer suggestions, only the following comments.

To the extent one trusts Innotek or has looked at the source code, VirtualBox seems to improve security. With NAT to the outside world it should only recognize responses -- no connection requests -- from outside the host; pretty much the behavior of a paranoid firewall.

Since VirtualBox supports snapshots, once set up it is possible to re-boot the virtual machine from a snapshot, thereby eliminating any possible changes to the environment about as effectively as using a live CD such as Knoppix to browse the Internet.

To further improve security one can remove lots of software from the host. After all, it is only functioning as an intermediary, doesn't really need a browser, for example.

However, keep in mind the reason that IP-Cop refuses to share a hard disk with any other OS: the more software on the hardware, the greater the risk of failure somewhere.

And if someone somehow should be able to break into your host with sufficient privileges, he could easily send himself all of your VM's to be looked at off-line at his leisure!

As usual, your call.

Use in Practice

Since initially installing VirtualBox with a Debian 4.0 host and among others a Debian 4.0 guest early in August, I have done easily 80 - 90% of my Internet access in this environment on a notebook. On rare occasion there has been disk I/O that the host attributed to VirtualBox, sometimes going on for a minute or more. But the "Internet experience" hasn't been noticeably impacted. I am still happy with response times for access to BBC and CNN.

The bit of Internet I have been doing outside of this environment is with Mandriva booted natively from another partition for e-mail. It quite literally was trivial to copy my bookmarks from Firefox there to the "shared folder" partition and then from there into the Iceweasel (Debian, don't ask) directory in a VM. And across the network to the other machine, of course.

For what it is worth, the "Borland" guest seen in the VirtualBox window above was created with floppy disks on a PC and sent over the LAN to the notebook which has no floppy drive!


Shortly after starting to experiment with VirtualBox, I mentioned it to a colleague who had begun to try Linux out. Recently he said that it worked well under Wimp/OS, where he has been playing with Ubuntu and Knoppix.

He pointed out that once the Guest Additions had been installed, time co-ordination is less costly than under VMWare (which he has been using for quite some time) where this is supposed to be a known issue.

Perhaps more interesting, he mentioned that RealPlayer under VirtualBox retains synchronization between sound and film.


Among the things VirtualBox supports but which I haven't had a chance to investigate are:
• Bridge on host at the same time as NAT
VirtualBox shows 4 network adapters available. So it should be possible to connect the guest to the LAN in addition to its use of NAT to take advantage of whatever connectivity the host has.
• CUPS to a network printer
Likely related, it would be very useful to be able to access a network printer.
• Access to USB
This is available, just described as dangerous, for obvious reasons, and thus not high on my priority list.
• Remote access to USB
Again, available, and should be quite useful in the appropriate environment but at the moment shared folders are sufficient (and see above).

Guest Problems

The command 'df' doesn't work with shared folders. This is really minor since the needed information can be obtained from the host.

vi can't write to a shared directory without error message, seems to work but not to be trusted (write error in swap file). It can change a file to zero length. I now copy a file locally for modification and then copy it back to make it available to other environments.

cp works EXT2-to-EXT2 but not EXT2-to-DOS. It isn't clear whether this is even a bug. Who uses DOS?

Under Mozilla <Ctrl>+ doesn't increase font size but using the mouse View|Text size|Increase is OK. And <Alt>-VZI works, so what is there to complain about?

Pros and Cons

Using snapshots you can produce "checkpoints" with which to recover from experiments -- teach yourself by trial-and-error at your own pace.

Shared folders can be extremely confusing, particularly after re-boot of guest or host, at least during testing. Re-boot of the host requires that the device subsequently be mounted on the mount-point again if not done permanently on the host in /etc/fstab. On re-boot of the guest if a shared folder was defined with the parameter "-transient" it will have to be defined again on the host and subsequently mounted on the guest. Adding it to /etc/fstab on the guest works fine, as long as "-transient" wasn't used on the host. In short: avoid "-transient".

The VirtualBox title bar on the VM pushes the "screen" down too far to get to Debian's "bottom" task bar. This is restored by full-screen mode. Full-screen mode makes more sense anyhow, unless you insist on several virtual machines on one physical display -- in which case, good luck!

On the notebook, at least, full-screen mode merely hides the VirtualBox framework; it doesn't increase the size of the window (VirtualBox seems to retain length and width ratio of display). Is this a bug or a feature?

The VBoxManage reference is very nice but it would be far nicer if the sub-commands were repeated where the command is discussed -- that would avoid much paging back and forth.

As good as it is, VirtualBox is not the perfect test-bed in all cases. Due to the one machine configuration it makes available to a VM, some possibilities are precluded. For example, both the notebook and PC used here have WLAN access to the Internet. But the NAT of VirtualBox effectively turns the host into a firewall. From a VM there is no way to learn how to manage a WLAN, how to establish a connection with the access point.

Apparently there are situations where VirtualBox fails miserably. Fedora Core 2 on both PC and Notebook was so outrageously slow that it is totally worthless, both during installation and afterwards. Yeah, I know -- look at the screen-shot -- FC 6 is OK and who would want to use 2. But something can really screw response times up.

And not every attempt to create a virtual machine was successful. Perhaps I just didn't have enough patience, but when the system monitor shows the CPU pegged over a considerable period of time and "ps aux" verifies it in a VT on the host...

A Non-VirtualBox VM

As mentioned above this package has several other features that I haven't been able to check out just yet. But the ability to use VMs from another environment seemed very interesting. So I found a VM on the Internet that had been produced with VMWare.

After unzipping the file and moving the VMDK into the VDI directory and defining a VM with it as virtual drive, this is what resulted:

Subjective Evaluation

VirtualBox has vastly exceeded any expectations I might have had. Indeed, there are a few rough edges yet. But this is a very mature product, an outstanding sand-box in which to experiment and learn.

It would be very interesting to hear about experience of hard-core gamers and people wanting to consolidate servers.

In any case to my mind there is no reason for anybody to be afraid of testing the Linux water, be it someone interested in a different flavor of Linux or considering changing other loyalties risk-free.

Talkback: Discuss this article with The Answer Gang

Bio picture Edgar is a consultant in the Cologne/Bonn area in Germany. His day job involves helping a customer with payroll, maintaining ancient IBM Assembler programs, some occasional COBOL, and otherwise using QMF, PL/1 and DB/2 under MVS.

(Note: mail that does not contain "linuxgazette" in the subject will be rejected.)

Copyright © 2007, Edgar Howell. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007

3D programming with Vpython

By Maxin B. John

"3-Dimensional Programming for Ordinary Mortals"- claims the Vpython project website. Vpython is a python module for 3 Dimensional scientific programming, specifically useful for the people with a Physics background.

Based On A True Story

Almost 3 years back, the Department of Computer Applications , Government Engineering College, Thrissur, India decided to conduct their annual National Seminal Seminar and Programming Contest, known as Renaissance - 2005, in the month of November. The people for Linux (which includes me) convinced the Professors to make the Renaissance 2005 a celebration of Free/Open Source . Our decision was to make it a true celebration of Linux. Almost all the programs were on Linux based technologies.

We conducted regular meetings to ensure the quality of the program. In one such meetings, we faced a small but striking question from one of the participants: How are we going to manage the Introductory Multimedia Presentation? In our college, the "Intro Presentation" was the way to show off the artistic and technical talent of the hosting Department. Most of the other Departments were using "Flash"-based eyecandy presentations as their intro. But we have decided not to depend on "not so open" technologies, and roll one for our own. But we couldn't reach a conclusion on the alternative technology at that time.

Here Comes Vpython

Whenever we had a technical doubt, we always had a door to knock at. Out team has met our beloved guru "Mr. Pramode CE" at his home. On hearing our "little bit strange" requirement, we heard what we wanted to hear. "You people can try Vpython for this purpose. You may integrate the music with your application using gamepy if you want to utilize the multimedia capabilities of python.".

A 3D Hello World

The visual module can be downloaded from http://vpython.org. It depends on Numpy, gtkglarea and boost libraries. But for Debian or Ubuntu user,it is a simple one liner:

apt-get install python-visual

You install the gamepy module if you want to play music with your presentations. Of course, gamepy can do a lot more than simply playing music!. It can be installed by using apt.

apt-get install gamepy

Python is famous for it's simplicity and elegance. The vpython "hello world" program is as follows:

from visual.text import *

# At present, VPython supports only numbers and uppercase characters. Other characters will be displayed as *
# Specifying the Title of the window
scene.title = "Hello World"
# Here goes the hello world text
text(pos=(0,3,0), string='HELLO WORLD', color=color.orange, depth=0.3, justify='center')

When you run his program as "python helloworld.py", your output will look like this

Now, let's try to 3D sine curve. You will be able to move this curve using your mouse.

from visual import *

This will print the sin curve 
scene.title = "Sin Curve"
scene.center = vector(0,0,0)

# using a suitable 'box' as x- axis 
xaxis = box(length= 20, height=0.2, width= 0.5, color=color.blue)

#creating the sine curve object
sinecurve = curve( color = color.red, radius=0.2)
dt = 0.1

for t in arange(0,10,dt):
dydt = vector( t,sin(t), 0 );
sinecurve.append( pos=dydt, color=(1,0,0) )
rate( 500 )

If you want to try a 'real application' like clock, let's make a funny 3D clock.

from visual.text import *
import time

scene.title = "3D Clock"
while 1:
	cur_time = time.localtime()
	time_string = str(cur_time[3]) +": "+ str(cur_time[4]) + ": "+ str(cur_time[5])
	timer = text(pos=(-3,0,-2), string=time_string, color= color.red, depth=0.5 )

Exploring some details of 3D vision

The stereo vision provides visual perception in 3D. Each human eye captures its own view, and the two separate images are sent on to the brain for processing. When the two images arrive simultaneously in the brain, they are united into a single picture. The resulting image is a three-dimensional stereo picture. This binocular disparity is interpreted by the brain as depth.

Vpython has the ability to generate real 3D objects. The redblue stereo vision can be enabled in vpython by setting scene.stero='redblue' in programs. We need the 'redblue' goggles to view the 3D extravaganza.

A Bit of Physics with Vpython

The main purpose of the visual module is to popularise the visualizations in Physics. So let's analyse the secrets of gravitational pull of our earth using the very famous equation Vt = V0 + at, where Vt is the velocity at time t, V0 is the initial velocity, and a is the acceleration (in the case of a falling body, acceleration is g, the acceleration due to gravity, i.e., 9.8m/s2).

Also , V= d/t , ie velocity is the rate of change of displacement. Putting all these together:

from visual import *

#A floor is an instance of box object with attributes like length, height, width etc.
floor = box(length=4, height=0.5, width=4, color=color.blue)

# A ball is a spherical object with attributes like position,radius, color etc..
ball = sphere(pos=(0,4,0),radius=1, color=color.red)

#Ball moves in the y axis 
ball.velocity = vector(0,-1,0)

# small change in time 
dt = 0.01

while 1:
    #setting the rate of animation speed
    # Change the position of ball based on the velocity on the y axis 
    ball.pos = ball.pos + ball.velocity*dt
    if ball.y < 1:
        ball.velocity.y = -ball.velocity.y
        ball.velocity.y = ball.velocity.y - 9.8*dt

Replace acceleration due to gravity (here 9.8) with the acceleration due to gravity in the Mars, to see how the ball bounces in Mars. Later, we can modify the value to see how the ball will bounce in other planets. Finding the acceleration due to gravity for other planets is left as an exercise for the reader.

From Mars to the Past

Finally, the days of Renaissance arrived. The day before the day of inauguration, I paid a visit to the local toy market to buy the "raw material" for the redblue stereovision glass. I bought 25 red colored goggles and 25 blue colored goggles from the local toy shop. It cost me only around 2 Dollars. At the hostel, we created 50 redblue stereovision glasses by keeping the red glass in the right frame and blue glass in the left frame.

After the inauguration function, I conducted an interactive session on Vpython with the help of a projector. My friends distributed the "redblue stereovision goggles" to the audience. After the session, I showed the audience the Real 3D Multimedia Introduction of the Renaissance.

from visual.text import *
import time

#importing pygame to play the background music :)
#import pygame
#Uncomment it if you have the redblue goggles

scene.title = "Renaissance"
#scene.fullscreen = 1
scene.fov = 0.001
scene.range = 0

# Uncomment this if you need to play the background music 

def intro():
	Title= text(pos=(0,3,0), string='MCA PROUDLY PRESENTS', color=color.red, depth=0.3, justify='center')
	for i in range(20):
		scene.range = i
	scene.range = 0
	Header= text(pos=(0,3,0), string='RENAISSANCE 2005', color=color.yellow, depth=0.3, justify='center')
	for i in range(20):
		scene.fov = 3
		scene.range = i
	# Now play with colors
	Header.reshape(color= color.cyan)
	Header.reshape(color= color.blue)
	Header.reshape(color= color.green)
	Header.reshape(color= color.red)
	# Now let's delete the Header
	scene.range = 10
	scene.fov = 0.2
	Body= text(pos=(0,3,0), string='A CELEBRATION OF LINUX ', color=color.red, depth=0.3, justify='center')
        #Here I am not adding the rest of the code as it just shows the schedule of that days programs
# Invoking intro()
if __name__ == '__main__':


Renaissance 2005 became one of the most successful programs ever conducted in GEC. But the most important thing I have felt was we had real fun. Our team has succeeded in showing others how fun linux can be.


The stereoscopic vision details are explained in http://en.wikipedia.org/wiki/Stereopsis. The reference manual for Vpython can be found at http://www.vpython.org/webdoc/visual/index.html.

Talkback: Discuss this article with The Answer Gang

Bio picture

I am an ardent fan of GNU/Linux from India. I admire the power, stability and flexibility offered by Linux. I must thank my guru, Mr. Pramode C. E., for introducing me to the fascinating world of Linux.

I have completed my Masters in Computer Applications from Govt. Engineering College, Thrissur (Kerala, India) and am presently working at Ushus Technologies, Thiruvananthapuram, India, as a Software Engineer.

In my spare time, you can find me fiddling with Free Software, free diving into the depths of Ashtamudi Lake, and practicing Yoga. My other areas of interest include Python, device drivers, and embedded systems.

Copyright © 2007, Maxin B. John. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007

Slow Receivers in a Distributed Data Management System

By Sudhir Menon

Slow receivers explained

A slow receiver is a node in the distributed system that cannot process incoming messages due to limited network bandwidth, CPU, I/O, or a combination of these issues. In all cases, the slow receiver either fails to pick up data from its incoming network buffers causing the system to bottleneck or fails to send application or protocol level acknowledgements which would allow the sender to proceed.

Slow receivers represent a performance problem in a distributed system. When using TCP or multicast, the presence of a slow receiver causes other members of the distributed system to slow down, and in extreme cases bring system throughput to a complete standstill.

In connection-oriented protocols like TCP, the sender has to copy data into its kernel buffers and send it out to each receiver individually. The send completes only when the data has been delivered to the receiver's kernel socket buffers. If the receiver's socket buffers are full, the send blocks until the buffers become available, slowing down the performance of other receivers who cannot receive messages from this sender because the sender is blocked trying to send a message to this slow receiver.

In connectionless protocols like reliable multicast, the sender sends the data out onto the multicast network by copying it out once onto the Ethernet card and then broadcasting it out on the network with the appropriate time-to-live parameters. The sender is not bogged down by receivers at the network buffer level.

Protocol reliability is achieved by having the sender maintain a buffer of sent messages and waiting to receive ACK messages from the receivers that they have received a particular message. The senders' buffer is the limiting factor when it comes to re-transmissions to receivers who cannot pick up data from the receiver buffers fast enough and then request the sender to re-transmit the lost data. Even in this case, one can see that senders end up spending CPU cycles and memory resources tending to slow receivers and thereby bogging down the system throughput. Slow receivers are often referred to as "crybaby" receivers in network parlance.

Slow Receivers and Cache Consistency

The ability to receive and process every piece of relevant data is critical to the functioning of a distributed cache. It is assumed that the messages coming in are relevant to the receiver and in order to maintain cache consistency, it is essential to make attempts to process the incoming data and provide some cache consistency guarantees to the consuming application.

At the same time, this desire to receive and process every message can result in a system that runs at the speed of the slowest consumer - clearly something that most distributed applications would not want to tolerate.

The solution is to define the consistency level that the cache elements within an application need and then provide a solution that deals with receiver slowdown. However, before looking at solutions, let us consider the situations that result in a slow receiver.

Detecting a slow receiver

For every message that is sent from a sender to a receiver, the sender maintains some stats on the average time to completion. When the time to completion stat starts showing an upward trend and breaches a threshold, the sender flags that receiver as a slow one. This sort of detection works well in connection-oriented environments where the sender and receiver share a connection.

In connectionless environments, the sender has to maintain stats on the number of retransmission requests made by the receiver, and when that crosses a certain threshold, tag the receiver as a slow receiver.

A third class of slow receiver detection is not really detection. Instead, a slow receiver, upon failing to keep up with the rest of the system or finding excessive use of memory in its application announces itself as a slow receiver, allowing the rest of the system to activate policies that have been configured for slow receivers.

Each member of the distributed system has stats that allow the member to detect that it is entering into slow receiver mode and can be configured with policies to deal with the situation.

Dealing with Slow Receivers

When it comes to slow receivers, there is no "one size fits all" policy that works (that works well anyway). The options that the system has once it encounters a slow receiver depend on its data consistency policy. What this implies is that a node has set certain data consistency expectations with other system members. These expectations play a major role in deciding how the member will be dealt with once it goes into slow receiver mode.

The slow receiver can choose to drop data, fire data loss notifications to the application, and catch up if the problem was temporary. This implies that not every update coming into the system has to be processed in order, and that if the application needs to fetch data from the cache, it will be fetched from other nodes on demand.

The slow receiver can send out a notification to other nodes stating that it is unable to accept any data until further notice. The remaining nodes would then ignore the member until they received a notice that the member was again open for business. Cache misses on other nodes would not be directed to this node, and data on the slow receiver would be considered suspect for the rest of the system, even though the local cache on the slow receiver would continue to serve the application and clients that it was attached to.

The system can quarantine the slow receiver thus isolating the rest of the system from the ill effects of the slow receiver. The senders could consider, store, and forward models for updates to that slow receiver. Applying interleaved updates from multiple publishers would become an issue in a system where all publishers were equal peers. In a single publisher system for a given piece of information, this would work well.

Another option is to have the notion of data ownership. This allows the slow receiver to apply updates from the owner of the data, without worrying about updates from other nodes.

A less desirable option is for the system to do nothing and run at the speed of the slow receiver. If the problem is temporary, the slow receiver comes out of that mode and the performance of the system improves.

Thus the options for dealing with slow receivers come down to the following:

Slow Receiver Support in an Enterprise Data Fabric (EDF)

In the previous section, we discussed a problem scenario in a distributed data management system. An Enterprise Data Fabric (EDF) provides mechanisms to detect slow receivers in a distributed system by collecting stats on network activities in the system; in addition, since the EDF is an active data management platform, it can be configured to make decisions on slow receivers in real-time. These decisions can be based on the applications sharing data in the data fabric and the need for data consistency across multiple applications. It can also be based on roles played by different applications in the data fabric and the criticality of getting data to the applications in the event of slow receiver behavior in the system.


A distributed data management system is a complex entity and deploying one in a production environment requires careful planning and analysis. Since we are dealing with temporal data and data consistency, it is important to have a good understanding of the network environment in which the application operates.

Every distributed system needs to have policies for dealing with slow receivers in the system. These policies have to be crafted keeping in mind the load characteristics of the system, data consistency guarantees, data loss notifications, and the system throughput requirements. Tuning the network to meet system objectives including throughput and latency has to be a part of the overall system design when you consider deploying an Enterprise Data Fabric.

Up-front capacity planning to ensure that hardware resources like network bandwidth, network partitioning, CPU, memory, and I/O characteristics of the nodes that participate in the distributed system will go a long way in avoiding unnecessary slowdowns and glitches in overall system performance. It is also important to understand the congestion characteristics of the network to ensure that the system as a whole is geared to deal with bursty traffic and temporary unavailability. Planning system redundancy, disk usage, and number of applications/instances that compete for resources on a system are factors that help prevent slow receiver problems and result in a smooth-running system.

It is also a very good idea to ask what support your distributed data management vendor has in their offering to deal with slow receivers. When it comes to dealing with slow receivers in a distributed data fabric, it is a question of "when" rather than "if."

Talkback: Discuss this article with The Answer Gang


Sudhir Menon, Director of Engineering, GemStone Systems

With over 17 years of cutting edge software experience with marquee firms like Gemstone, Intel, EDS and CenterSpan communications, Sudhir Menon is one of the key architects for the Gemfire Enterprise Data Fabric. Sudhir is the Director of Engineering for GemStone Systems, Inc. and works closely with various development teams (both onsite and offshore) working on the Gemfire Enterprise Data Fabric. His expertise in distributed data management spans multiple languages (Java, C++ and .NET) and multiple platforms and he has architected and developed network stacks for the last 10+ years. At Centerspan communications, he was one of the key architects who built the largest secure peer to peer content distribution platform over the internet.

Copyright © 2007, Sudhir Menon. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007

FlickOff: Escaping the Clutches of Web 2.0

By Ben Okopnik

The Trap

Some months ago, while reading a Web site that contained suggestions for "streamlining your life", I signed up with Flickr.com, so I could have the photos from my cameraphone immediately posted to the Web. There were a number of benefits to doing so: I could photograph the label of a good brand of socks (or beans) that I didn't want to buy right then but wouldn't be able to remember when I did want them; I could snap a pic of something I wanted to research later without filling up my phone's memory - in short, I would now have a visual "memo pad" in addition to the written one that I keep around. I did chafe a bit under the restrictions of having to do this via a third party (Flickr) whose interests weren't exactly aligned with my own: as the price of this "free" service, I had to stare at the advertisements that, more and more, infested their pages; I had to make my way through their clunky, Java-scripted interface, which often created problems over my occasionally-slow connection; and, perhaps most annoying of all, I had to put up with their subtle-but-pervasive nagging, a feeling of being a cheapskate unless I bought their "premium" service. Hell, I was only using one of their features, and wasn't interested in anything else they had to offer! However, the overall convenience (slightly) outweighed the pain - so I stuck with it.

The Last Straw

Then, Yahoo bought Flickr - and added their password-and-nag system on top of the one already in place. That, I decided, was that. Enough, and Just Too Damned Much, in fact. I also recalled a conversation that Rick Moen and I had here, in LG's Answer Gang.

Every time I decline to use some third-party, usually proprietary
service that I can reasonably do with Linux and open source on my own
systems, instead, I am protecting and promoting the perception by
onlookers that Linux and open source are the right solution -- not to
mention improving my own competence.  Internet services are supposed to
be the Linux community's core competency: What sort of message does it
send to go moving our affairs onto other people's third-party services
as a convenience, in an area where we're supposed to be the leaders?
 -- Rick Moen

Rick had it exactly right. I gritted my teeth, set my shoulders, and growled... in short, I did all those things that a hero does in a movie before taking on the Bad Guys, blowing up a few buildings in slow motion, and Getting the Girl. I mean, if you're going to set out to do something this important, you might as well do it right - right? :) It seemed a little excessive for a programming project, but sometimes, you just need that little bit of extra motivation.

The Daring Escape

Cheating Checking Freshmeat and SourceForge didn't bring any positive results, so I decided to create my own "image upload" system, one that was more suited to my needs. I'd been using a piece of photo-gallery software called "LiveFrame" to arrange and display my photos; it was a bit old, and the author had moved on to other things, but I liked the way it worked. Most of all, I liked its transparency in configuration and operation: the images were all kept in subdirectories named after a given category, the image headers and descriptions were listed in a single per-category configuration file, and the gallery program itself was written in my favorite language, Perl. Anything I didn't like, I could easily tweak - and since I was going to use Perl to write this new app, integrating the two might become an eventual goal.

I went through the LiveFrame code to make sure that everything worked OK and fixed a few minor bits; then, I settled down to writing my program, which I decided to call FlickOff. The interesting part was not so much the program itself - although that took a bit of work - but the network-interactive nature of the process: I decided to change not only where the images went, but how they were handled as well. E.g., since cameraphones aren't really professional photo apparatus, there'd be a number of times when I'd want to edit the image before sending it up to the Web. Some of the images would be private (I hacked LiveFrame to create categories that were invisible unless you knew their names.) In short, the process now looked like this:

Step 1:
Take pic using cameraphone; add a subject line; email to self

Step 2 (automated):
When the email arrives at my laptop, anything from my phone's address is
intercepted and sent to FlickOff. The subject line is parsed to determine
the appropriate gallery - anything without one is sent to the 'unsorted'
gallery; otherwise, the first word is the gallery name and the rest of the
line is the image description. The local copy of LiveFrame then receives
the properly-resized images and the description info is added to the config

Step 3:
I vet/adjust/process the images and the descriptions as desired, then run a
script to synchronize the local LiveFrame repository with the "remote"
(Web-based) installation, which makes the images available to the world.

In short, I've replicated the functionality that I wanted, got rid of the annoyances, and am continuing to refine the process as time goes on. At this point, I feel like FlickOff is nicely stable and useful to anyone else who'd like this kind of a system, so here's the setup procedure:


Note that FlickOff was only tested with T-Mobile and Cingular/AT&T phones; if you have a problem with your particular setup, don't hesitate to e-mail me with the details (please make sure to include the photo-message that made it fail) and I'll do my best to adapt FlickOff to handle those as well. Happy Linuxing to all!

Talkback: Discuss this article with The Answer Gang


Ben is the Editor-in-Chief for Linux Gazette and a member of The Answer Gang.

Ben was born in Moscow, Russia in 1962. He became interested in electricity at the tender age of six, promptly demonstrated it by sticking a fork into a socket and starting a fire, and has been falling down technological mineshafts ever since. He has been working with computers since the Elder Days, when they had to be built by soldering parts onto printed circuit boards and programs had to fit into 4k of memory. He would gladly pay good money to any psychologist who can cure him of the recurrent nightmares.

His subsequent experiences include creating software in nearly a dozen languages, network and database maintenance during the approach of a hurricane, and writing articles for publications ranging from sailing magazines to technological journals. After a seven-year Atlantic/Caribbean cruise under sail and passages up and down the East coast of the US, he is currently anchored in St. Augustine, Florida. He works as a technical instructor for Sun Microsystems and a private Open Source consultant/Web developer. His current set of hobbies includes flying, yoga, martial arts, motorcycles, writing, and Roman history; his Palm Pilot is crammed full of alarms, many of which contain exclamation points.

He has been working with Linux since 1997, and credits it with his complete loss of interest in waging nuclear warfare on parts of the Pacific Northwest.

Copyright © 2007, Ben Okopnik. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007


By Javier Malonda

The Ecol comic strip is written for escomposlinux.org (ECOL), the web site that supports es.comp.os.linux, the Spanish USENET newsgroup for Linux. The strips are drawn in Spanish and then translated to English by the author.

These images are scaled down to minimize horizontal scrolling.


Click here to see the full-sized image

All Ecol cartoons are at tira.escomposlinux.org (Spanish), comic.escomposlinux.org (English) and http://tira.puntbarra.com/ (Catalan). The Catalan version is translated by the people who run the site; only a few episodes are currently available.

These cartoons are copyright Javier Malonda. They may be copied, linked or distributed by any means. However, you may not distribute modifications. If you link to a cartoon, please notify Javier, who would appreciate hearing from you.

Talkback: Discuss this article with The Answer Gang

Copyright © 2007, Javier Malonda. Released under the Open Publication License unless otherwise noted in the body of the article. Linux Gazette is not produced, sponsored, or endorsed by its prior host, SSC, Inc.

Published in Issue 144 of Linux Gazette, November 2007

The Linux Launderette

Dinesh has Tagged you! :)

Rick Moen [rick at linuxmafia.com]

Wed, 24 Oct 2007 10:30:22 -0700

Quoting Dinesh Sikhwal (disikh at yahoo.com):

>    Dinesh S, 21
>    Dinesh has added you as a friend on Tagged.
>    Is Dinesh your friend?

Functionally speaking, Dinesh is a spammer.

>   [imgsrv.php?uid=5382936315&imgn=1&imgt=1&iw=320&ih=170&iy=53&is=35&nfr
>    =3]
>    [2]Click here to block all emails from Tagged
>    P.O. Box 193152 San Francisco, CA 94119-3152

No, I really don't think so: I'd rather just globally block all mail at all of my MTAs from the entirety of "taggedmail.com".

I'll also throw in "tagged.com" while I'm at it, as this "social networking" business appears to be more than usually pestilential.

http://labnol.blogspot.com/2007/03/how-to-block-tagged-mail-latest-e-mail.html says:

Tagged.com describe themselves as a "teen social networking destination on the web" but it is currently the largest source of email spam in our inbox.

Every other minute, there's a new email message from Tagged.com with a subject "[Your Friend Name] has tagged you! :)".

There's a link at the bottom of the email message that says "Click to unsubscribe" but even when you unsubscribe, "TaggedMail Invites" will continue to flood your inbox.

Goodbye, Dinesh. Goodbye, Tagged.com.

[ Thread continues here (2 messages/6.29kB) ]

The whole Internet on one page

Ben Okopnik [ben at linuxgazette.net]

Sun, 21 Oct 2007 19:26:32 -0400

Some people have interesting ideas. :)


* Ben Okopnik * Editor-in-Chief, Linux Gazette * http://LinuxGazette.NET *

[OT] Subliminal

Jimmy ORegan [joregan at gmail.com]

Mon, 1 Oct 2007 13:03:47 +0100

My Dad was looking for some subliminal MP3s (he's a hippy...) - for this site (http://www.freesubliminals.com/) specifically - when he came across this: "Learn Polish Language SUBLIMINAL Instant MP3 Download" (http://cgi.ebay.com/Learn-Polish-Language-SUBLIMINAL-Instant-MP3-Download_W0QQitemZ150166054708QQihZ005QQcategoryZ3144QQcmdZViewItem)

So, he called me to the computer. I thought the most amusing part of the blurb was this:

"No dangerous pills Safe & Natural No dangerous diets No nauseating side effects No dangerous exercise"

Wow! I didn't realise that language learning had become so dangerous!

This also came up in his search: "Subliminal Recording Software - Make your own subliminal cd's and MP3's" (http://subliminalrecorder.com/). The blurb for that had this: "Create Subliminal Ringtones!". Like... umm... what? "Answer the phone?"

Talkback: Discuss this article with The Answer Gang

Published in Issue 144 of Linux Gazette, November 2007