blog dot kam

Recently I have written a quick “how to” on restoring, formatting a USB flash drive. The “how to” had a lot of hits from different places since then, mostly from GOOGLE, of course, but some from independent bloggers, like my blogging self.

One of such places was www.newlinuxuser.com. Although the guide “saved them” (welcome :)) they had a very constructive critique that one thing the guide missed was how to actually find which device to restore / format.

Hence I decided to write a new little 2 step “how to” that will teach you just that.

Step 1. System Log is your friend, listen to what it has to say…

Imagine you close your eyes, and someone puts an apple in your mouth - would you be able to identify what the heck was put into your mouth? For most people the answer would be “yes”. ( If you’d like to experiment, feel free :). The thing is once you bite on that apple, your brain goes through many lines of code (given that we are written in C), finds that match, and reports:

"The object in your mouth was identified as an Apple - we've had that before, I just talked to the stomach, it knows how to digest it."

You would think that your Linux/Unix system is any different? Well, not really.

Right after you insert a USB device into a slot, Linux/Unix will try to read, and identify it. While doing that, it will also assign it to a particular device as “/dev/particular-device”. This device is exactly the information we need, so we can talk to it, and mount it.

Although most people would approach it with running a dmesg, and look at the output, I prefer a more natural OS way to look at things - we’ll look directly in the eye of a System Log!

Insert your USB drive into a slot, and type:

tail -f /var/log/messages

You will see the output similar to:

Dec  5 14:53:19 your-hostname kernel: [81585.308993] usb 4-1: new full speed USB device using uhci_hcd and address 3
Dec  5 14:53:19 your-hostname kernel: [81585.456757] usb 4-1: not running at top speed; connect to a high speed hub
Dec  5 14:53:19 your-hostname kernel: [81585.484884] usb 4-1: configuration #1 chosen from 1 choice
Dec  5 14:53:19 your-hostname kernel: [81585.498817] scsi6 : SCSI emulation for USB Mass Storage devices
Dec  5 14:53:24 your-hostname kernel: [81590.514870] scsi 6:0:0:0: Direct-Access     USB 2.0  USB Flash Drive  0.00 PQ: 0 ANSI: 2
Dec  5 14:53:24 your-hostname kernel: [81590.519874] sd 6:0:0:0: [sdb] 15794175 512-byte hardware sectors (8087 MB)
Dec  5 14:53:24 your-hostname kernel: [81590.522834] sd 6:0:0:0: [sdb] Write Protect is off
Dec  5 14:53:24 your-hostname kernel: [81590.534817] sd 6:0:0:0: [sdb] 15794175 512-byte hardware sectors (8087 MB)
Dec  5 14:53:24 your-hostname kernel: [81590.537814] sd 6:0:0:0: [sdb] Write Protect is off
 
>>>>  Dec  5 14:53:25 your-hostname kernel: [81590.537888]  sdb: sdb1  <---- GOT YOU!
 
Dec  5 14:53:25 your-hostname kernel: [81590.654848] sd 6:0:0:0: [sdb] Attached SCSI removable disk

Note that the USB drive was “connected”, or associated with sdb device

[81590.654848] sd 6:0:0:0: [sdb] 15794175 512-byte hardware sectors (8087 MB)

and more precisely, with sdb1 device

[81590.537888]  sdb: sdb1

And that means we can talk to it! The full name of the guy would be “/dev/sdb1″.

Now let’s greet our friend. Say: “Hi /dev/sdb1″!

Step 2. Mount USB drive’s device to the File System.

Just an extra step, in case you need to mount it. If you can’t, and would like to format it, so you can mount it afterwards, read this.

To mount the drive enter this:

sudo mount -t vfat /dev/sdb1 /media/usbdrive/

where “/dev/sdb1″ is the name of the device, we found in the step above. “/media/usbdrive/” is the directory that we are going to mount it to. Make sure this directory exists (otherwise create it “sudo mkdir /media/usbdrive/”). And “-t vfat” is asking your Linux/Unix OS to mount this device as a “vfat” (FAT16, FAT32) device.

Many, if not most, USB devices are VFAT, however if you have an NTFS USB hard drive, for example, you can mount it by entering:

sudo mount -t ntfs-3g /dev/sdb1 /media/usbdrive/ -o force

“sudo” in above couple commands comes from mostly Ubuntu way to “run command as a super user”. If you have any other flavor of Linux/Unix, you may want to just run it as a “root” user.

Eat more apples, and good luck!

It is quite frustrating when you have a wireless card built in, but you cannot use it - don’t you agree? That seems to be the case for wifi cards that are served by “iwl3945” driver. One of such examples can be “Intel Corporation PRO/Wireless 3945ABG” card.

Before, ( e.g. Feisty and earlier ) “ipw3945” driver was used instead, and worked beautifully with Intel cards. However the active development of this driver has stopped a couple of years ago, which means that security risks that were identified in the last couple of years were not patched. Therefore Ubuntu community switched to a more recent and supported “iwl3945” driver.

However that created a problem with NetworkManager that is used as a default network user interface in (Gnome) Ubuntu. It appears that in order “to be compatible with NetworkManager, a wireless driver must support both hardware and software scanning. Currently, hardware scanning is faster and more reliable and so is recommended for use with NetworkManager“. But unfortunately “iwl3945″ driver does not support hardware scanning very well, however it is a default behaviour that NetworkManager expects.

But usually, and that is what I love Ubuntu for, if there is a problem, there is 99.9999% a solution to this problem. So this how to will help you solve it, and enjoy your Intel Wireless card with Ubuntu, despite of inconsiderate (in this case) NetworkManager.

1. Remove the bad guy

sudo apt-get remove network-manager

2. Disable “iwl3945″ hardware scan

sudo vi /etc/modrobe.d/iwl3945

enter:

	alias wlan0 iwl3945
	options iwl3945 disable_hw_scan=1   # <-- enables software scanning

and save it (:wq).

At this you might want to restart or (optinal) just re-insert the module like this:

sudo modprobe -r iwl3945
sudo modprobe iwl3945

3. Install the good guy

Here you’ll install “the good guy” - his/her name is WICD:

Add wicd’s repository to Ubuntu sources:

sudo vi /etc/apt/sources.list

	 deb http://apt.wicd.net intrepid extras

(if you have Hardy Heron, then add “deb http://apt.wicd.net hardy extras” instead)

Update repositories

   sudo apt-get update

Install wicd “the normal Ubuntu way”:

   sudo apt-get install wicd

4. Run the good guy

   wicd-client

or restart, it should run automatically.

Happy wifying!

While wireless signal is good, the network is dead - why is that? Many reasons, of course. You can spend time to figure out the reason and then try to fix, which is a good approach, but requires some time. Or you can restart the network to see if it resolves the issue, and if it does - forget that the problem ever existed.

However the way to restart a wireless network is not always “black and white”. Sometimes it is possible using GUI, and sometimes by typing something that means “network service restart”. And yes these ways are “clean”, but have a drawback - they rely on operating system to do what it suppose to do and restart the network.

However, OS does not always behave (yes, Linux does not always behave, along with Mac, and Windows, and “any” OS.. ). But here is a sure way to restart it - you would need to get down to the driver level though - to be less OS (or distribution) specific. But I’ll guide you through, don’t worry..

So, the signal is full/good/strong:

but there is no network:

$ ping -c 4 google.com
ping: unknown host google.com

First thing to do is to see what wireless card you are using:

$ lspci | grep -i network
08:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02)

In my case it is Intel 3945ABG. Next, check what driver is used for this card. I did a simple google search, and saw that the driver is “ipw”something.

Let’s see what ipw-like modules/drivers are currently running/loaded:

$ modprobe -l | grep ipw
/lib/modules/2.6.20-17-generic/kernel/ubuntu/wireless/ipw3945/ipw3945.ko
/lib/modules/2.6.20-17-generic/kernel/drivers/usb/serial/ipw.ko
/lib/modules/2.6.20-17-generic/kernel/drivers/net/wireless/ipw2200.ko
/lib/modules/2.6.20-17-generic/kernel/drivers/net/wireless/ipw2100.ko

Here it is “ipw3945″. Let’s kill it (-r stands for “remove”):

$ sudo modprobe -r ipw3945

Let’s start it back up:

$ sudo modprobe ipw3945

Checking connectivity:

$ ping -c 4 google.com
 
PING google.com (64.233.187.99) 56(84) bytes of data.
64 bytes from jc-in-f99.google.com (64.233.187.99): icmp_seq=1 ttl=238 time=43.3 ms
64 bytes from jc-in-f99.google.com (64.233.187.99): icmp_seq=2 ttl=238 time=28.9 ms
64 bytes from jc-in-f99.google.com (64.233.187.99): icmp_seq=3 ttl=238 time=27.7 ms
64 bytes from jc-in-f99.google.com (64.233.187.99): icmp_seq=4 ttl=238 time=34.7 ms
 
--- google.com ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3005ms
rtt min/avg/max/mdev = 27.742/33.685/43.323/6.165 ms

Perfect!

On the client site when working on Java (or should I say JEE) projects, the application servers are most of the time “rotate” between these four: JBoss, Websphere, Weblogic, and someties Geronimo. However there are some clients who have pretty tight contracts with Sun, and in these cases the Sun Application Server is used.

I think now days in development world in general, well except maybe .NET part of it, Ubuntu and/or Mac OS slowly become OSs of choice for developers. Sun app server however is not the most used app server out there, and since its installation could be a bit non-straightforward, here are four simple steps on how to install it on Ubuntu box:

Step 1. Download the “.bin” form sun:

“Sun Java System Application Server”

http://java.sun.com/j2ee/1.4/download.html (I needed 8.2, but there are more recent ones)

Step 2. Change permissions, to make it runnable:

chmod 744 sjsas_pe-8_2-linux.bin

Step 3. If you just run it:

./sjsas_pe-8_2-linux.bin

It is going to complaint that it is missing a standard c++ library:

./sjsas_pe-8_2-linux.bin: error while loading shared libraries: libstdc++-libc6.2-2.so.3: cannot open shared object file: No such file or directory

Having done some “apt-cache” searches ( apt-cache search libstdc++ ), found that Ubuntu has “libstdc++.so.6″ in /usr/lib.
From /usr/lib run:

 sudo ln -s libstdc++.so.6 libstdc++-libc6.2-2.so.3

Step 4. Run it now, it will install Sun App Server successfully! * **


* Do not run installation as root - it will fail (for most sun app servers versions)

** If you use any form of Beryl (or some Compiz’es), disable it, or reload window manager as a “Gnome Manager”. This is due to the fact the the installer is written in Swing, and Swing does not get along too well with some display managers.

It is quite a simple change but it makes an Ubuntu laptop to boot 2-3 times faster! What takes the most time during boot in Ubuntu (and many other Linux distros) are DHCP discovery and DHCP timeout that are run by networking startup scripts. This of course can be solved with tweaking the timeout or just configuring a static IP, however this solution is not all that elegant, just because “What if the system needs this particular timeout?” or “What if it is a laptop, and it’s IP cannot be static?”

The idea to solve this is simply to take an advantage of concurrency. So instead of all other boot scripts to wait on the networking script(s) to finish its discovering and “timeouting” during a system boot, these network script(s) could be started in parallel, as separate child processes. This will speed up boot time, since the system should no longer wait on the “networking” to finish.

One thing to understand here, is that networking is not taken out of the boot process - it stills belongs to it (it is its child), and it would still run, and would do its important job - the difference is it would do it in parallel.

Here are two easy steps on howto archive this in Ubuntu (it can be applied to pretty much any Linux/Unix distribution, however the scrips/locations will/might be different) :

Step 1. Find network startup scripts:

cd /etc
 
user@host:/etc$ sudo find . | grep network
./rcS.d/S40networking
./init.d/networking
./network
./network/.interfaces.swp
./network/if-post-down.d
./network/if-post-down.d/avahi-daemon
./network/if-post-down.d/wireless-tools
./network/if-post-down.d/wpasupplicant
./network/interfaces
./network/if-up.d
./network/if-up.d/ntp
./network/if-up.d/clamav-freshclam-ifupdown
./network/if-up.d/ntpdate
./network/if-up.d/avahi-daemon
./network/if-up.d/sendmail
./network/if-up.d/avahi-autoipd
./network/if-up.d/wpasupplicant
./network/if-up.d/mountnfs
./network/if-pre-up.d
./network/if-pre-up.d/wireless-tools
./network/if-pre-up.d/wpasupplicant
./network/if-down.d
./network/if-down.d/clamav-freshclam-ifupdown
./network/if-down.d/sendmail
./network/if-down.d/avahi-autoipd
./network/if-down.d/wpasupplicant
./networks
user@host:/etc$ ll ./rcS.d/S40networking
lrwxrwxrwx 1 root root 20 2007-05-20 18:48 ./rcS.d/S40networking -> ../init.d/networking

FOUND IT: In this case the netwoking script that runs on startup is “../init.d/networking”

Step 2. Extract “start” case into a separate method:

user@host:/etc$ sudo vi ../init.d/networking

here is a “start case” in the original file:

case "$1" in
start)
	log_action_begin_msg "Configuring network interfaces"
        type usplash_write >/dev/null 2>/dev/null && usplash_write "TIMEOUT 120" || true
	if [ "$VERBOSE" != no ]; then
	    if ifup -a; then
		log_action_end_msg $?
	    else
		log_action_end_msg $?
	    fi
	else
	    if ifup -a >/dev/null 2>&1; then
		log_action_end_msg $?
	    else
		log_action_end_msg $?
	    fi
	fi
        type usplash_write >/dev/null 2>/dev/null && usplash_write "TIMEOUT 15" || true
	;;

let’s modify it by extracting the “start case” into a separate method so it can be run as a background process on the start up:

here is a “start case” in the modified file:

# adding this method so it can be run as a background process on the start up
start_on_boot () {
 
	log_action_begin_msg "Configuring network interfaces"
        type usplash_write >/dev/null 2>/dev/null && usplash_write "TIMEOUT 120" || true
	if [ "$VERBOSE" != no ]; then
	    if ifup -a; then
		log_action_end_msg $?
	    else
		log_action_end_msg $?
	    fi
	else
	    if ifup -a >/dev/null 2>&1; then
		log_action_end_msg $?
	    else
		log_action_end_msg $?
	    fi
	fi
        type usplash_write >/dev/null 2>/dev/null && usplash_write "TIMEOUT 15" || true
}
 
case "$1" in
 
start)
 
	# start networking on the background, so it does not slow down the boot time
	start_on_boot &
 
	;;

Save the file, reboot and enjoy

DONE

Interesting related articles:

making ubuntu boot in 19 seconds
understanding bash fork bomb

Ever heard about rails.vim project? “Accept no imitations: rails.vim is the one true Vim plugin for syntax highlighing, easy navigation, and script invocation for all your Ruby on Rails applications, transparently and unobtrusively” says creator Tim Pope.

The very good “rails.vim” guide can be found here or just by reading project’s vimdoc here.

PROBLEM:

There is however one gotcha for Ubuntu Hardy Heron lovers (or other modern Linux distros). Using rails.vim would result in VIM “segmentation fault”crashes similar to:

Vim: Caught deadly signal ABRT
Vim: Finished.
Aborted

REASON:

This is due to the fact that a packaged VIM that comes from some Linux distros repositories has old patches.

SOLUTION:

One of possible solutions would be to download most current VIM sources from http://www.vim.org/sources.php and compile/install it manually. Below is how it is done on Ubuntu (but should be pretty similar on any Linux distro):

1. Get vim sources:

wget ftp://ftp.vim.org/pub/vim/unix/vim-7.1.tar.bz2

(where ‘7.1′ is the current VIM version at the moment of writing)

2. Unpack it

tar -xvjf vim-7.1.tar.bz2

3. Install terminal libraries (vim needs them to compile correctly)

sudo apt-get install libncurses5-dev

4. Configure / Compile / Install

./configure --with-features=huge
make
sudo make install

5. Point your system to newly compiled VIM:

sudo rm /etc/alternatives/vi
sudo rm /etc/alternatives/vim
sudo rm /etc/alternatives/vimdiff
 
sudo ln -s /usr/local/bin/vim /etc/alternatives/vi
sudo ln -s /usr/local/bin/vim /etc/alternatives/vim
sudo ln -s /usr/local/bin/vimdiff /etc/alternatives/vimdiff

DONE

Frequently I need to access my LinuxMCE machine over VNC, due to different reasons (when I am not home, when my Wii Remote batteries died, etc..). Usually SSH solves most of my needs, but sometimes it is very nice to control my linuxMCE box over VNC.

Here I will show you several very easy steps on how to configure your server/client to be able to tunnel VNC traffic through SSH, and be able to control your server in the GUI way This approach will work on most systems (not LinuxMCE specific) although here I chose Ubuntu as an “example OS”.

Before we begin, I would assume that:

• The server already has ssh server installed, if not:

  sudo apt-get install openssh-server

• If connecting from the outside (like from work to home box), your router’s firewall forwards port 22 to your server’s IP

Here is how it is done:

Server Side Configuration

1. Installing VNC server, here I chose to install x11vnc, but it could be pretty much any VNC server of your choice (TightVNC, etc..):

sudo apt-get install x11vnc

2. Now let’s finish our server configuration by running the VNC server:

x11vnc

by default it is going to run without a password, and on port 5900

Client Side Configuration

1. In order to connect to the VNC server, we need to have a VNC viewer installed:

sudo apt-get install xvncviewer

* again, it can be pretty much any VNC viewer of your choice

2. Now the most interesting part of this whole process - enter this command:

ssh -L 5900:127.0.0.1:5900 yourusername@yourserver.com

this tells your system to tunnel all the traffic from the port 5900 on “yourserver.com” to the local (client’s) port 5900 via SSH. Simple, and yet very powerful - a “magic one-liner”

3. Fire up another shell on the client side and enter:

xvncviewer 127.0.0.1

do you see a cute colorful square screen? Do you recognize your server’s Desktop?

Done

- Go to the LinuxMCE web admin
- From the upper menu, go to “Advanced -> Network -> Firewall rules”

add this three rules:

tcp  	5900 to 5900  	5900  	192.168.80.1             port_forward
tcp 	5900 to 5900 	0 	0                        core_input
tcp  	22 to 22  	0  	0                        core_input

- The first rule allows external network (Internet) to access the VNC server on the Core (LinuxMCE server)
- The second rule allows anybody inside the home network ( since it is not wise to open 5900 port on your router to the whole world ) to have their 5900 port to be forwarded to the Core.
- The third rule (I think you already have it) allows anybody to access LinuxMCE server via port 22 (SSH)

Be careful about the rules above, as I assume that your LinuxMCE server is behind another firewall, cause you really do not want to open 5900 port to the public