Xlylith's world of words................

Reactivating

2011-07-24T05:37:00.001+07:00

It's been years since my last post. It seems that I can't blog when I don't put my heart in it.
Let's see if I can reignite the sparks.

How to Make an Offline Repository of Your Own Ubuntu System

2009-11-24T15:09:00.010+07:00

Introduction

Sometimes when you need to replicate an Ubuntu system to another computer, the usual step will be:

Install Ubuntu using Installer CD/ Live CD
Reboot
Connect to the Internet
Update
Installing additional packages that you need

However, if it occurs that you need to install to another computer at another place with uncertainty of the availability and/or reliability of the Internet connection over there, one would wonder if you could prepare those updates and additional packages yourself in an offline repository.

Assuming the additional packages is the same with your Ubuntu System, this howto will explain how to make such offline repository.

Creating Offline Repository

There are 4 steps to setting up a simple repository for yourself

Import the packages in a directory
Scan the packages and create a file apt-get update can read
Replicate the repository to portable media
Add a line to your sources.list pointing at your repository

Import Packages

In your Ubuntu System, the .deb files you have installed are stored in /var/cache/apt/archives/

To replicate these files, first create the directory you want to paste into. For this example, we'll use ~/mydebs.

mkdir ~/mydebs

Then, replicate the files using rsync

rsync -upt --progress /var/cache/apt/archives/*.deb ~/mydebs

Sometimes some files you have installed is not cached. If so, you need to retrieve them from the internet. First we need to generate the list of files to be downloaded.

sudo apt-get install $(dpkg -l|grep ^ii|awk '{print $2}') --reinstall --print-uris \

-y |awk '{print $1}'|sed "s/'//g"|grep tp:// > downlist

Then download it using wget

wget -ci downlist -P ~/mydebs

Scan Packages

To be included in apt-get process, we net to generate packages list of those .deb files.

First go into mydebs directory.

cd ~/mydebs

then generate Packages.gz using dpkg-scanpackages

dpkg-scanpackages . /dev/null | gzip -9c > Packages.gz

Replicate mydebs Directory to Portable Media

The repository can now be replicated to portable media, such as DVD-ROM or External Hard Drive. Assuming you are using external harddrive with volume name REPO:

rsync -a ~/mydebs /media/REPO/mydebs

Then, after finishing installing the new system, restore the directory to your new home folder

rsync -a /media/REPO/mydebs ~/mydebs

Adding mydebs Directory as Repository Source

To use these repository, these mydebs directory then must be included in the Repository Source file on your new system.

sudo su -pc 'echo deb file:$HOME/mydebs ./ >> /etc/apt/sources.list'

sudo apt-get update

Your repository is now ready to use

Counting Down on Intrepid Ibex

2008-10-28T14:24:00.003+07:00

I just can't wait :-D

Extracting Similarity Between Two (Text) Files

2008-10-27T10:52:00.002+07:00

When you need to compare difference between two files, you have diff.
When you need to find similarity between such files, as far as I know, you need to devise your own script.
This is my oneliner script for such need (assuming files to be compared named file-01 & file-02):

for ((i=1;i<=$(wc -l file-01|awk '{print $1}');i+=1)); do grep $(awk -v a=$i 'NR==a' file-01) file-02 ; done

Ugly, but working...

The Reason

2008-10-13T20:13:00.001+07:00

If I know the answer of "Why",
I believe I'll better understand the question of "How"

I need to know why I want to be rich, before seeking the way to be richer
I need to know why I want to be free, before breaking anything that hold me back
I need to know why I want to be happy, before pursuing the happiness itself

But, is it wise to seek the reason of being right, while I am still pursuing the very question of what is truth?

Reverse Proxy with Apache

2008-02-05T15:30:00.001+07:00

Sometimes, when we need to configure network devices within LAN (via http), we need to do it on site due to private addressing. While making the devices directly accessible from the Internet is a big no-no, there is another way to connect remotely and securely: Reverse HTTP Proxy (with Apache).

Apache has proxy-related modules to enable this operation. To get it working, you need to enable this modules:
- proxy
- proxy_http

Using Ubuntu 7.10, this modules can be loaded simply by typing:

sudo a2enmod proxy
sudo a2enmod proxy_http
sudo /etc/init.d/apache2 restart

To forward any http request to your internal address, you need to add the destination url into your virtual host configuration, i.e:

sudo nano /etc/apache2/sites-available/yourvirtualhostconfig

then add this line:

ProxyPass /your-public-url/ http://your-private-address

enable your site by typing:

sudo a2ensite yourvirtualhostconfig
sudo /etc/init.d/apache2/restart

By default, in Ubuntu 7.10 all proxy access are denied. Therefore, you
need to white list your address into the proxy.conf file

sudo nano /etc/apache2/mods-available/proxy.conf

change the config to allow your source address i.e:

<Proxy *>
AddDefaultCharset off
Order deny,allow
Deny from all
Allow from 203.153.240.197
</Proxy>

Then reload the configurations

sudo a2enmod proxy
sudo /etc/init.d/apache2 restart

That's it folks. Enjoy working on your devices remotely, securely.
(It is not so secure as the traffic is not encrypted. What I mean by
secure, is that I can hide the private address from any port scan.
Therefore, what is deemed secure by me, is not necessarily secure by
your definition. This is MY blog, so no complain!!!)

Connecting to XL 3G via Sony Ericsson W850i

2007-11-12T16:54:00.000+07:00

While it is not so newbie friendly, I never thought that configuring Internet connectivity using my cellular phone will be a straightforward process. All I did just edit the /wtc/wvdial.conf into this:

Init1 = ATZ
Init2 = ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0

Modem Type = USB Modem
Baud = 460800
Modem = /dev/ttyACM1
ISDN = 0
Phone = *99***1#
Phone1 = *99***2#
Password = proxl
Username = xlgprs

then, I connect by calling wvdial (with root privillege)

sudo wvdial

Just be very cautious with the cost as it is charged per KB!
Enjoy...

Let's Celebrate a Good Time

2007-09-18T13:43:00.000+07:00

Celebration Lyrics

Installing ZABBIX 1.4.1

2007-08-20T19:00:00.000+07:00

The new ZABBIX is getting better and better. Below are the steps I did to get
ZABBIX 1.4.1 up and running. (Assumed there is already Apache2, PHP and MySQL running)

1. First, get the source code.
$ wget

http://optusnet.dl.sourceforge.net/sourceforge/zabbix/zabbix-1.4.1.tar.gz

2. Extract the source code
$ tar -xvvzf zabbix-1.4.1.tar.gz

3. Prepare MySQL. by creating database for ZABBIX
$ mysql -u<username> -p<password>
mysql> create database zabbix;
mysql> grant all on zabbix.* to <username>@localhost identified
by '<mysqlpassword>';
mysql> quit;
$ cd <path-of-your-zabbix-source>/create/schema/
$ cat mysql.sql | mysql -u<username> -p<password> zabbix
$ cd ../data
$ cat data.sql | mysql -u<username> -p<password> zabbix
$ cat images_mysql.sql | mysql -u<username> -p<password> zabbix
$ cd ../..

4. Configure the source code, then proceed with install
$ ./configure --enable-server --enable-agent --with-mysql --with-net-snmp --with-libcurl
$ sudo make install

5. Create zabbix config files
$ sudo mkdir /etc/zabbix/
$ sudo cp
<path-of-your-zabbix-source>/misc/conf/zabbix_server.conf /etc/zabbix/
$ sudo nano /etc/zabbix/zabbix_server.conf

6. Create directories for the frontends
$ sudo mkdir /usr/local/zabbix
$ sudo chown -R <username.username> /usr/local/zabbix/
$ rsync -rvu <path-of-your-zabbix-source>/frontend/php /usr/local/zabbix/

7. Edit Apache2 config file to point your own ZABBIX url into the frontend's
path

8. Open a web browser, go to your ZABBIX url, then proceeds from there

9. To get ZABBIX up, you need to execute zabbix_server.

Done :-)

I Love You

2007-07-30T19:37:00.000+07:00

I Love You Lyrics

2007

2007-01-03T09:22:00.000+07:00

Selamat Tahun Baru 2007

Bimbo - Semalam di Malaysia

2006-12-22T10:56:00.000+07:00

aku pulang
dari rantau
bertahun-tahun di
negeri orang
oh, Malaysia

oh, dimana
kawan dulu, kawan dulu
yang sama berjuang
oh, Malaysia

kekasih hatipun
telah pula hilang
hilang tiada pesan
aduhai nasib
apakah daya
cita hampa
jiwaku merana
mana dinda

inilah kisahku
semalam di Malaysia
diri rasa sunyi
aduhai nasib
apakah daya
aku hanya
seorang pengembara
yang hina

It works!!!

2006-11-17T14:46:00.000+07:00

One thing I just realized after upgrading to Edgy, all hotkey in my desknote, which by the way is an ECS A907, works out of the box. No hacking, no tweaking, no user intervention. The volume manager key works, the "internet" key fires up Firefox and the "mail" key opens Evolution.

Now...if only I could make use of that f***in' flyin' windows key ...

Edgy upgrade fail on acpid

2006-11-16T20:32:00.001+07:00

I just upgraded my dapper into edgy. everything went fine but acpid and
everything that depends on it.
After several hours of googling... I found the answer from launchpad.
The workaround are:
1. Shutdown the acpid (I removed it instead)
2. sudo killall hald
3. sudo dpkg --configure -a
4. sudo apt-get install acpid

Perhaps 4 months is a bit too short for a release cycle of a distro as
massive as Ubuntu...

Mounting Remote Directories using SSHFS

2006-11-16T09:48:00.001+07:00

Due to limited harddisk capacity, I stored my edgy repository in an external USB hard drive.
To access it, I mount it using sshfs. Here is how I did it in Ubuntu 6.06.

$ sudo apt-get install sshfs

$ sudo modprobe fuse

$ mkdir /somewhere/localfolder/

$ sshfs -o allow_other user@remotehost:/somewhere/somefolder /somewhere/localfolder

To unmount it just use the normal umount command

$ sudo umount /somewher/localfolder

Now I have edgy repo, shall I start upgrading now... ?
I heard some bad rumors about upgrading dapper to edgy.
I think I need to do some backup first.

screen bash

2006-11-14T19:19:00.001+07:00

When executing application using screen, I found that if the application went bad, we couldn't find out the error unless we specificy to log it. My debmirror script sometimes fail and I don't know if it is fail unless I repeat the script again.

Hence, bash must go first :-)

It's also quite handy to hide whatever I am doing on the screen as I only need to press C-a d to detach it.

Lust : Feature or Bug

2006-11-13T11:28:00.001+07:00

Q: What is lust, a feature or a bug?
If it is a bug, then we are still in development stage and shouldn't have been released yet.
If it is a feature, then why some of us must go to hell because of it?

A: Well, probably deep inside our heart, there are lots of things that was meant to be patched by ourselves.

Hmm... a self-repairing apps...
Thank God we're not running Vista :p

pmount vs perl

2006-11-10T14:58:00.000+07:00

After modifying debmirror, i copied it into a folder in my mounted portable hard drive.
The remaining space is quite large, so I put dapper & edgy as the target and start executing the script.
*bam*
myscript: /usr/bin/perl: bad interpreter: Permission denied

what the h###?!

I am absolutely sure I have no problem with permissions as the same script also copied into the parent directory of
a running ubuntu local ftp repository, and it works.

After checking every possible permission problem I encountered, I finally came into mount permission.
I mounted the portable hd using pmount (it's a usb-connected external hd) and it turnout that the default parameter
used by pmount include a noexec option, therefore, when I mounted it using the classic mount, the script worked.

Now, one question remain...
How come a perl script is considered a binary executable, while bash script is not?
Hmm...

Blogging via mutt

2006-11-10T11:05:00.000+07:00

In blogspot, one can post his/her blog via mail. I like this feature because of two reasons :
1. The built in blogspot editor is terribly *cough*bloated*cough* slow
2. Due to aging hardware, opening firefox in my laptop takes zillion years
The problem is, the reason #2 is also the reason why I hate opening evolution in my laptop.

Few days ago, I installed mutt. The setup is not perfect yet, but I found myself quite enjoying it.
It loads extremely fast... that I only keep evolution just for reading html-formatted-mail sent to me.
I can also deal with junk mails much better :-)

Patching debmirror...

2006-11-09T21:01:00.000+07:00

I just realized that debmirror has a weakness in its design. Debmirror is designed to download the packages lists, selecting correct packages to be downloaded, then and only then, download the packages. However, the packages selection, is somewhat long and tedious. Even worse, there is no progress indicator to tell us how much time left for package selection process to start downloading those damn packages.

I wonder if I can modify it to make it faster.

* searching google with keywords: perl tutorial*

Free as in Freedom

2006-10-16T09:30:00.000+07:00

I just saw Ust. Quraish Shihab on MetroTV this morning saying that we are free to choose our religion. He also said that one must face the corresponding consequences of choosing such faith him/herself.
However, since apostasy in Islam is punishable by death, I doubt it is free enough to be called freedom.

It's not that I lost my faith. I just found it disturbing that such foremost muslim scholar unaware of such thing.

Shiver Me Timbers!

2006-09-19T17:59:00.000+07:00

apt-get moo

2006-09-11T16:37:00.000+07:00

$ apt-get moo


        (__)
        (oo)
  /------\/
 / |    ||
*  /\---/   ~~   ~~
...."Have you mooed today?"...

Partitioning USB Drive from Linux

2006-09-05T15:22:00.000+07:00

USB drive is indeed can be partitioned.
What you need to do is:
1. Unmount USB drive,
2. Disable auto-mounting
3. gparted /dev/sda

As I haven't been able to tame the automounting feature of Kubuntu, I used
Gparted-Live and partitioned from there. When I partitioned my USB drive in
Kubuntu, it persistently automounted the first partition (sda1) after it had
been created, even the whole partitioning process hadn't been finished yet.

Thus, it's the value of having a pocket-CD of Gparted-Live everywhere you go.

Now I have a 128MiB USB pen-drive partitioned into 80s MiB FAT16 and 30s MiB EXT2

And...

250 GiB USB Harddrive partitioned into 40s GiB FAT32, and two 100s GiB EXT3

Now ... I am backing-up everything :-D

Managing Wireless Equipment (802.11x) using ap-utils

2006-08-30T12:46:00.000+07:00

Being maintaining snmp-based nms for months, I've just found this tool few
hours ago. With ap-utils, I can manage my SmartBridge devices natively from
Linux. Ironically, I've been struggling with wine for years just to configure
my WiFi equipment, and even installing windows 98 via Qemu just to have the
wireless configuration software installed.

To install it in Ubuntu :

sudo apt-get install ap-utils

then

ap-config

The menu inside ap-utils is self-explanatory.
If only this tools can manage other proprietary wi-fi devices, (alvarion,
canopy, trango, etc) it will be much better.

QEMU in Ubuntu Dapper

2006-08-08T10:29:00.000+07:00

Should you need another OS emulated within your system, QEMU is the Free Software that do the job well.
In Ubuntu 6.06, you can have the version 0.8.0 packaged in the universe repo.

To use it, just install it using apt-get (or synaptic if you prefer):

$ sudo apt-get install qemu

then create the virtual hard drive using qemu-img

$ qemu-img create mydisk 3G

to install an OS into your emulated hard drive (using CD-ROM installer)

$ qemu -hda mydisk -cdrom /dev/cdrom -boot d

Qemu can be emulating at higher speed when it uses its accelerator module named Kqemu. However Kqemu is a proprietary software (although it'll cost 0 in price). To install the accelerator, you can use the method posted here

Unbloggable Months

2006-08-08T10:20:00.000+07:00

Well...two months without blogging...
I lost my desire to write these two months.
Mostly due to:

FIFA World Cup '06 Germany
(Somehow single) Parenting

Now, since I feel quite narcisist enough, it's time to write again :-D

Linux Only

2006-06-08T18:17:00.000+07:00

When my desknote crashed several days ago, I've tried to put another flavor of UNIX into it.
Yup, I tried to boot Belenix (Solaris-based) and PC-BSD (FreeBSD-based).
Both kernel got panic and dumped.
It seems that Linux is the only "free" kernel that will boot up my transmeta box.
(My Sempron 2100+ box has no problem at all)

Linux rules :-D

"Crippled" Repo

2006-06-08T18:01:00.000+07:00

After hours of trying and failing net-install Ubuntu 6.06 (Dapper Drake) from my local repository, I finally found out what's missing from it.
I did not include main/debian-installer in my debmirror script!
Thus, the repository is only good for installing packages into up & running Ubuntu system.
With less than 3 GB disk space remaining, I don't think I can afford to add main/debian-installer into my local repo.
Thus, having CD-installer ready is a must.

Proxy Traceroute

2006-05-12T14:35:00.000+07:00

When I need to ensure that an IP address is properly routed and reachable, I often use the traceroute service of www.nwtools.com. However when I have to query frequently, the slowness of firefox on my laptop is such a hinderance.
To keep myself from hammering down my laptop, I create shell script to do the task using elinks and sed. Here is my code:

#!/bin/sh
elinks -dump --no-references --no-numbering \ "http://www.nwtools.com/default.asp?prog=trace&host=$1"|sed '1,21d'|sed -n -e :a -e '1,7!{P;N;D;};N;ba'

I saved the script as proxytrace.sh and using it with the traceroute target as the only parameter.

example:
$ sh proxytrace.sh www.blahblahblah.com
or
$ sh proxytrace.sh 123.234.123.234

This way is much faster, really.

Intransitive Trust

2006-05-04T12:35:00.000+07:00

In those ol' days...there were Windows NT4.
I remember the pain of setting every domain to trust each other as the trust relationship between domain is one way and intransitive.
To get every domain connected in a sane way is a great challenge.
Especially if you put different version of windows into.

That memory suddenly came to mind...
As I feel so miserable when I've got so little trust from someone I love, while she got mine, fully.
Perhaps, I need to be wiretapped, equipped with gps, and monitored 24/7 as well?!
This is ridiculous and draining my sanity.

*Taking deep breath, holding it for several seconds, and release it slowly...repeated 3x*

Hello World! :-)

A Geek's Bedroom

2006-04-20T16:18:00.000+07:00

Kubuntu, with ATI RADEON 9250 Chipset, and TV-OUT, means Multimedia Playback in the Bedroom!
Now...if I can build MythTV on it...

Another Google April Fools' Joke

2006-04-04T12:32:00.000+07:00

I am always fascinated by their April1st Jokes
Just now they launched Google Romance
Just check it at http://www.google.com/romance/
Have fun

My sysadmin toolbox

2006-03-23T17:03:00.000+07:00

Inspired by linux.com :-)

For my day to day network administration, I use these apps frequently:

ssh,telnet,bash,coreutils,rsync,scp,elinks,wget,nano....if you don't know what
are these apps for, then probably you could skip this
*cough*article*cough* :-)
For serial console access, I use minicom. I love configuring my huawei box
with minicom using transparent konsole, it's so damn pretty.
For quick traffic monitoring via snmp poll, I use ifstat. If only I could
figure out how to parse the capture result into rrd files, I would make it
even more useful.
For more comperehensive network provisioning, JFFNMS is what I use. I just
wonder how to reduce its resource usage. I'll probably give nagios a shot
someday.

The others will be most of what GNOME & KDE offers....

100.000 km milestone

2006-03-23T15:40:00.000+07:00

The odometer of my motorbike has just passed 100.000 kms milestone :-)
The odometer itself, however, consists of 5 digit only,
hence, its returned to ZERO :-D

Time to sell the bike eh?

j/k

Merah Putih

2006-03-17T17:28:00.000+07:00

What makes a nation, nation?
no...I think the question should be ...
What keeps a nation, nation?

...

Are we really still a nation... or perhaps... we never was...?!

Aliases

2006-03-17T16:55:00.000+07:00

Going into bash, aliases make my life easier, really!
I kept adding these aliases in wherever .bashrc I have

someone@somewhere:~$ alias
alias acs='apt-cache search'
alias acso='apt-cache show'
alias dusk='du -s -k -c * | sort -rn'
alias greph='history | grep '
alias ls='ls --color=auto'
alias nsl='netstat -alnp --protocol=inet | grep -v CLOSE_WAIT | cut -c-6,21-94
| tail +2'
alias sadug='sudo apt-get dist-upgrade'
alias sapi='sudo apt-get install'
alias sapu='sudo apt-get update'
alias sapup='sudo apt-get upgrade'
someone@somewhere:~$

As you can see... it's very ubuntu, right? :-)
I think I need to add some more...
Any idea ?

Messing up with debmirror

2006-03-08T11:25:00.000+07:00

From debmirror manpages:

DEBMIRROR(1) User Contributed Perl Documentation DEBMIRROR(1)

NAME
debmirror - Debian partial mirror script, with ftp, http, hftp or rsync
and package pool support

SYNOPSIS
debmirror [options] mirrordir

DESCRIPTION
This program downloads and maintains a partial local Debian mirror. It
can mirror any combination of architectures, distributions, and sec‐
tions. Files are transferred by ftp, and package pools are fully sup‐
ported. It also does locking and updates trace files.

To support package pools, this program mirrors in three steps.

1. download Packages and Sources files
First it downloads all Packages and Sources files for the subset of
Debian it was instructed to get.

2. clean up unknown files
Any files and directories on the local mirror that are not in the
list are removed.

3. download everything else
The Packages and Sources files are scanned, to build up a list of
all the files they refer to. A few other miscellaneous files are
added to the list. Then the program makes sure that each file in
the list is present on the local mirror and is up-to-date, using
file size (and optionally md5sum) checks. Any necessary files are
downloaded.

I had never paid attention step two before I tried to add breezy-updates and breezy-security, into my existing Ubuntu Breezy repository. I executed debmirror without putting breezy in the dist parameter (instead, only breezy-updates and breezy-security), hence, I lost 11 Gigs of data.

Mental note: RTFM RTFM RTFM

Scripting Addiction

2006-02-28T13:58:00.000+07:00

Since I can do some bash scripting, I fell like I can't stop.
I even try to automate anything I usually do by scripting it.
Hmm... should I dive into python and made my own NMS?
I'm quite dissappointed with JFFNMS performance, btw.
Probably, I should take a look at ifstat source code and learn from it, and then make my own lightweight network monitoring system....
Hmm... scripting is really really really addictive....

A week with Fedora Core 4

2006-02-17T16:35:00.000+07:00

FYI, my laptop is now running Debian Sarge. It's just plain stable, albeit old system :-)
My PC, however, is running Kubuntu Dapper Drake, and with spare hdc (after cleaning up the pre-existing Ubuntu Breezy Badger) I decided to try another distro that is not debian-based.

Going into RPM world, Fedora is my first choice.

My first experience with GNU/Linux is with Red Hat 5 several years ago. Being a newbie, that was an unpleasant experience. I knew nothing about UNIX-like environment, getting annoyed with RH5 hardware detection, especially in Multimedia & Printing support, I finally mess with X and cannot log into GDM until one day my cousin did some fiddling with configuration files and fixed my X issue.
With those unpleasant memory, I begin installing FC-4.

Installation is graphical and intuitive. During the installation, I chose to manually partition my hard disk and customizing packages selection. I didn't found synaptic nor adept there, but hey, this was the time to use yum, right?

The installation ended with grub installation. Here, I made a mistake. I overwrote my Ubuntu's grub in MBR with Fedora's one. As Fedora did not recognize my Ubuntu installation, after reboot, I presented with no option to log into Ubuntu unless I manually edited the Grub.

The bootsplash is nice although not as informative as (K)Ubuntu's, and after passing my username & password into GDM, I was presented with GNOME 2.10. I had to say I much prefer the brown one than that bluecurve+clearlooks combo though :-)

After playing around for a while (and tasting Fedora's KDE 3.4) I started using yum.
I was thinking yum update is the equivalent of apt-get update when I noticed that it actually updated the system. I found that command is similar with apt-get update followed with apt-get upgrade.

As with apt, a proxy-server in the middle of the way messed up with yum. In the case of my Debian-based system, I changed the sources.list to make apt use ftp instead of http connection. In yum, I found it a bit tricky as it use mirrorlist. By default Fedora's yum using a mirror list to reach its mirror, defined in /etc/yum.repos.d/fedora*.repo (This is a good method, I wonder if Debianistas would adopt this). Since I couldn't find a way to tell yum to use ftp instead http, I created a directory named yum.custom.repo, copied the original .repo files there, made a script to grab those mirrorlist, and created the list locally.

#!/bin/sh
cat fedora.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora.mirrorlist
cat fedora-extras.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora-extras.mirrorlist
cat fedora-updates.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora-updates.mirrorlist
cat fedora-updates-testing.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora-updates-testing.mirrorlist
cat fedora-devel.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora-devel.mirrorlist
cat fedora-extras-devel.repo|grep mirrorlist|sed -e 's/mirrorlist=//' -e 's/$releasever/4/'|wget -i - -O -|grep ftp|sed 's/http:/ftp:/' >fedora-extras-devel.mirrorlist

Then I just edited the mirrorlist option in .repo files at /etc/yum.repos.d to point into those mirror list, e.g:

[base]
name=Fedora Core $releasever - $basearch - Base
#baseurl=http://download.fedora.redhat.com/pub/fedora/linux/core/$releasever/$basearch/os/
mirrorlist=file:///etc/yum.custom.repo/fedora.mirrorlist
enabled=1
gpgcheck=1
gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora

After yum update completed, I was left with a very up-to-date Fedora system, with surprisingly: a well-working KDE 3.5.

So far, I enjoyed my Fedora experience, and kept it in my system.

I still have space in my hard disk.
So... what will it be : Open SUSE, Mandriva, Gentoo or Slackware?

Merge PDF files with pdftk

2006-02-14T11:48:00.000+07:00

Having some of my manuals/documentations downloaded in several separated pdf files, I wanted to merge them into a single file.

First, I installed pdftk

# apt-get install pdftk

Then, I merged the files into a single file

$ pdftk file1.pdf file2.pdf file3.pdf cat output file123.pdf

When the files to be merged is a lot, I edited the file list in a text editor (or, when I am in a good mood, using ls & awk & sed) then combine it with the script found here

$ cat stdin_pdf.sh
#!/bin/sh
# combine PDFs; pass filenames in on stdin;
# pass output to stdout
read input_filenames
pdftk $input_filenames cat output -

$ cat filelist|./stdin_pdf.sh>mergedfile.pdf

So far, I am satisfied with the result.
However, I am still looking for the method to modify the hyperlink to suit the new file, since the previously scattered files actually hyperlinked to each other. RTFM is in progress :-)

Size of Ubuntu Repository

2006-02-09T10:47:00.000+07:00

When I tried to create local repository of Ubuntu, I was haunted by an unanswered question: How big will it be?
Googling and diving into ubuntuforums.org gave me no solution.
Then, a friend in Ubuntu-id mailing list gave me the answer: calculate the size from every packages listed in Package.gz.

This is a sample (ugly) script I used to get main repository size calculated.

$ wget http://id.archive.ubuntu.com/ubuntu/dists/breezy/main/binary-i386/Packages.gz -O - |zcat|sed 's/Installed-Size://'|grep Size|awk '/Size:/ {pkgsize+=$2} END {print "package size = "pkgsize" Bytes"}'

The size of other repositories can be calculated by simply changing the path of Packages.gz

Building Ubuntu Local Repository (CD)

2006-02-07T14:00:00.000+07:00

I've just read from this site about creating Ubuntu DVD Repositories. I thought I need to create such local repo, but in CDs, as I don't have DVD burner.

First, I created local copy from main repository using debmirror.
In this case, I am using mirror from CBN.

$ debmirror --nosource -m --passive --host=ubuntu.cbn.net.id --root=Ubuntu/ --method=http--progress --dist=breezy --section=main --arch=i386 ubuntu/ --ignore-release-gpg

Creating destination folders for partitioning

$ mkdir ubuntu-cd

Then I partitioned those files using debpartial

$ debpartial --nosource --dirprefix=ubuntu --section=main --dist=breezy --size=CD74 ubuntu/ ubuntu-cd/

After the packages lists had been generated, I created the symlink to the corresponding packages using debcopy script. It can be downloaded from here.

$ ruby debcopy -l ubuntu/ ubuntu-cd/ubuntu0
$ ruby debcopy -l ubuntu/ ubuntu-cd/ubuntu1
$ ruby debcopy -l ubuntu/ ubuntu-cd/ubuntu2
$ ruby debcopy -l ubuntu/ ubuntu-cd/ubuntu3
$ ruby debcopy -l ubuntu/ ubuntu-cd/ubuntu4

The partitioned folders then converted into iso image files, and burnt.

$ mkisofs -f -J -r -o ubuntu-cd-main-0.iso ubuntu-cd/ubuntu0
$ mkisofs -f -J -r -o ubuntu-cd-main-1.iso ubuntu-cd/ubuntu1
$ mkisofs -f -J -r -o ubuntu-cd-main-2.iso ubuntu-cd/ubuntu2
$ mkisofs -f -J -r -o ubuntu-cd-main-3.iso ubuntu-cd/ubuntu3
$ mkisofs -f -J -r -o ubuntu-cd-main-4.iso ubuntu-cd/ubuntu4

$ sudo cdrecord -v dev=/dev/cdrw speed=32 driveropts=burnfree -dao -eject ubuntu-cd-main-0.iso
$ sudo cdrecord -v dev=/dev/cdrw speed=32 driveropts=burnfree -dao -eject ubuntu-cd-main-1.iso
$ sudo cdrecord -v dev=/dev/cdrw speed=32 driveropts=burnfree -dao -eject ubuntu-cd-main-2.iso
$ sudo cdrecord -v dev=/dev/cdrw speed=32 driveropts=burnfree -dao -eject ubuntu-cd-main-3.iso
$ sudo cdrecord -v dev=/dev/cdrw speed=32 driveropts=burnfree -dao -eject ubuntu-cd-main-4.iso

The method to create repo-cd for universe & multiverse repository is similar.
I Just changed the apropriate parameter in debmirror & debpartial.
Now I have access to ubuntu repository even when I am not connected to the net. Right now, I am currently trying to create repo from breezy-updates & security. Hopefully, I will have it mirrored before it changes (again) :-)

Background Processing

2006-01-25T15:34:00.000+07:00

I am managing a headless server (GNU/Linux) for provisioning network under my supervision. I remotely manage the server via ssh and using it as fileserver as well. Therefore, linux iso images from many distros are downloaded and stored there

One thing I love from this scheme is that I can download any iso files from the net and left the downloading process in background using wget, such as:

$ wget -b http://someurl/someisofile.iso

However, when the source server get busy, downloading from http or ftp is not a good idea. Hence, bittorrent comes in handy.
Using combination of screen and btdownloadcurses.py, the command looks like this:

$ screen btdownloadcurses sometorrentfiles.torrent

I can left the download process in background by pressing Ctrl-A followed by D.
Whenever I want to check the download progress, I can list the screen process by typing

$ screen -ls

Then, to get the screen-ed process back, I just type

$ screen -dr name-of-the-process

Now, I'm downloading every distro I can get for free.
I just have to make sure I'm not clogging my company's bandwidth, otherwise, I'll be in trouble :-D

Enlightened Ubuntu

2006-01-03T18:12:00.000+07:00

I scrapped my Xubuntu two weeks ago to be replaced by Vector Linux 5.1 Standard. From what I read, VL is a very good distro for old hardware, thus, perfect for my ECS A907 desknote.

Having it installed, I immediately know what I missed from Xu/Ku/Ubuntu: excelent package management, good configuration utility and strong community support. It's not that Vector Linux is bad. It runs very fast! However, I don't feel convenient with its Slackware-style and decided to reinstall my ubuntu.

Now the problem is that both Ubuntu & Kubuntu are too slow, and having tried VL, I want something faster than Xubuntu. Hence, I decided to have Enlightenment clean-installed.

So far I made it :-)

I just install ubuntu server install, then after first login I did:

$ sudo apt-get install linux-686 x-window-system-core enlightenment eterm enlightenment-themes esound menu e16keyedit e16menuedit

I also grabbed several tools I need such as minicom, mc, abiword, gnumeric, xmms, firefox, sylpheed, epiphany-browser (for light browsing) and the three apps I need from KDE: Gwenview, Konqueror and Ark.

I am still tweaking my system, but overall, I love E16. Light yet Elegant.
I just can't wait for E17 :-)

Split Personality

2006-01-03T17:48:00.000+07:00

Wow, finally...
First post of the year :-)
It seems I am not narcissist enough to write regularly last year.
However, It doesn't mean that I had nothing to write about.
I just couldn't write it here since I would like to keep several things private.
Now, since I have splitted my dark-self into other pensieve, I think there will be more writings here...probably :-)

Anyway...
HAPPY NEW YEAR!!!

Microsoft signs as Wembley backer

2005-10-21T11:28:00.000+07:00

Being a fan of English Premiership League, and a linux user, the news of MS becoming a Wembley backer quickly get my attention. From TFA, MS will support Wembley with its technology. I wonder what it means. Does it mean that any information service provided in Wembley will be run in MS Platform? Will any information can only be accessed trough MS tools?
If that's what MS pursues in, I think it's a smart marketing manouver since Wembley is a very Iconic stadium for all football fans.
I just hope that Mark Shuttleworth or any other open source tycoons(?!) can find a counter of such tactic. After all, football fans are also human being, right? :-D

Ubuntu Linux 5.10 (Breezy Badger) At A Glance

2005-10-15T20:19:00.000+07:00

Breezy has been officially released!!!.
Now, I believe it's the perfect time for me to get a good grasp with the newest version of my favorite GNU/Linux distribution.
This is the specification of my system:

AMD Sempron 2200
Asrock Motherboard K7S41
DDRAM 512 MB
2x40GB Maxtor Hard Disk
Pioneer DVD-ROM
on board sound & vga

I. Installation
The text-based installation is similar with Hoary's. It's not eye-candy like Xandros' or Linspire's installer, but it definitely gives you more control.
There is a new feature in Breezy installer : installing into free space reclaimed by shrinking existing partition. It's a nice feature for those new to linux and want too keep his OS by dual-boot. However, I choose to customise the partitioning. My 1st HD contains my main system: Ubuntu 5.04 (Hoary) and I do not want to mess with it, so I divide my 2nd HD into root and home, then install Breezy into them.

II. First Boot
The first installation phase is smooth, and after grub installed into my 2nd HD, it reboots. This first boot, however, is not smooth. I have to manually edit the Grub configuration file myself, before the installation goes into the second phase. After additional packages were installed and configured, The new splash screen shows up, and bring me into the familiar ubuntu login screen.

III. Logging in
Finally, I got the taste of Gnome 2.12. I notice some improvements in nautilus, evolution and app-install. However, what I really like is these new features: smeg (menu editor), boot-admin & disks-admin. I know I can edit menu manually, tinkering with menu.lst, or mount those hard disks manually, but having these tools will make my Gnome more humanly, just like Ubuntu slogan: Linux for Human Being.

III. Applications
I don't see any big difference with hoary, since I am not a software developer (those new gcc compiler will remain under the hood forever). One thing I'd like to try is JFFNMS 0.8.2. I use ubuntu at work as nms, so it's my chance to try it out. I'll write about this later.
I would also try Ubuntu as headless server. I feel my desktop performance much below hoary. I think it's because of GNOME 2.12 (My XFCE & E16 run well). It takes some moment for most apps to appears. Perhaps I should test Kubuntu to confirm this.

IV. Conclusion
Wow, it's too early to conclude anything since I've just tried it for two days!
Anyway, I like breezy, but the low performance of this release bothers me much. I even scrap my standard breezy installation on my old laptop to replace it with XFCE-Ubuntu(!). I think I'll keep monitoring the ubuntuforum to seek resolutions of these issues.

Double Choco Latte or Request Tracker

2005-10-06T10:15:00.000+07:00

After hundreds of unarchived troubleshooting effort, I felt that I needed a method to keep those records logged. This idea brought me into this solution: trouble ticketing system. I liked the idea of having all the troubleshooting progress monitored and logged, simply by entering the information from web.

Crawling the web for any related open source project brought me some...well, actualy plenty of choices:

Bluetail Ticket Tracker (BTT)
Bugin
Bugzilla
Double Choco Latte (DCL)
GNATS (aka PRMS)
HelpDesk
JitterBug
Job Control System (JCS)
OpenTicket
Open Track
PHP HelpDesk
PHPSAT
ProManager
Request Tracker (RT)
ReqNG
Roundup
Scarab
Teacup
TouxDoux
WebCall
WHUPS
wreq

Quick look at my Ubuntu repositories reduced them to two: DCL & RT

So, it was the time to choose........hmm..............nevermind, I just picked both.

Being quite a coffee geek, I went first with DCL. After wrestling with postgresql, and set it up, I felt that this animal was so buggy and relatively unmaintained that I had to abandon it.

The setup of RT was done somewhat easier, as its documentation spread over the web. I felt that RT was much more mature since it had been around for longer period. However, I felt the administration was a bit kludgy. I also disliked the user interface style. Rather than customizing it, I decided to continue my journey in searching trouble ticketing system apropriate to my work.

So...here I am....
Back with manualy logged troubleshooting,.... with pen and papers scattered troughout my desk.
ROFLMAO

MERDEKA!!!

2005-08-15T10:34:00.000+07:00

Telat sehari...... tapi..........

MERDEKA !!!

Semoga dengan berjalannya waktu....kita dapat memerdekakan diri kita pula,
dari penjajahan ekonomi, budaya dan teknologi.

Mumpung temanya tentang kemerdekaan, dengan ini saya proklamasikan kemerdekaan komputer saya dari jeratan virus, spyware dan software2 terlarang (baca:bajakan) :-D

Merdekalah selalu Indonesiaku, Merdekalah pula komputerku...

MERDEKA!!!

-Written in Ubuntu Linux 5.04, using Mozilla Firefox Web Browser-

Instant Pasta

2005-08-09T17:37:00.000+07:00

I like Italian food. However, I don't fancy preparing it myself. So far, all I need to do when I want some pizza or pasta is going to the nearest Pizza Hut franchise or pick up the phone and make orders. That's before I found Pazto instant pasta and LaFonte instant spaghetti.

The preparation was quick and easy (just like preparing instant noodle). Boil a glass of water, pour the ingredients, stir well, wait for several minutes and voila! I had a bowl of hot pasta :-D

Now I'm going to try it with hot water instead of boiled water. If I succeed, I'll be having hot pasta for lunch, at work .... hmmmm........ yummy

Digital Posession

2005-08-09T12:09:00.000+07:00

When we bought a digital media, do we really OWN it? In my opinion, "buying digital media" is an absurd phrase. Why? Because when we bought a gum, whether we chew it, tear it, spit it or even swallow it, we are free to do so.... as it is ours. The essential meaning of posession should be the freedom to do anything on/at/with the posessed object. And that's not the case with digital media.

"Buying" digital media means the we're bonded with some "EULA-ish" agreement, which prohibits us to do anything the author disapprove. We cannot really own such media because the original author only shared some portion of media posession to us. Hence, we are only granted some right of use.

However, could such agreement really be enforced? Could anyone prohibits any others to do anything with their digital media? Due to the nature of digital media, it will never possible to fully prevent any replication and modification, and any attempt to do so will fail miserably. On the other hand, it is now possible to digitize many kind of things. Perhaps, in the future, even ourself can be digitized as well.

Maybe we can start rethinking the idea of copyright.
Just maybe............