Linux Lab

In this post, we weill share how to quick bring up a Ruby On Rails development server running on NodeJS on an Ubuntu Server.

First, install Ruby:

$ sudo apt-get install ruby

Then, install gems:

$ sudo apt-get install rubygems

Now, install NodeJS:

$ sudo apt-get install nodejs

Finally, install Ruby On Rails:

$ sudo gem install rails

In time, Git is always a good idea:

$ sudo apt-get install git

In this post, I will share some commands that will convert endline characters from DOS to Linux.

On DOS (and Windows) there are two special characters that denotes the end of a line, these are the New Line (\n) and the Carriage Return (\r).

On Linux, there is only one special character, the New Line (\n).

On MacOS, the character is the Carriage Return (\r).

A classic lack of standard...

Anyway, if you are on a Linux system and is having problems with files or scripts that uses the DOS (\n\r) end of line characters, you can use a stylish shell line to convert it using vim:

$ for file in *.txt; do vim -b $file -c "update" -c "e ++ff=dos" -c "setlocal ff=unix" -c "wq"; done;

Or, as I just discovered, there are a simpler program that do this: dos2unix

$ dos2unix file.txt

Internationalization is the process of designing a software application so that it can be adapted to various languages and regions without engineering changes. Internationalization is also referred as i18n, cause it holds 18 letters between the first i and the last n.

Nice article about SAS, Near Line SAS and SATA disks on:

In this post, we will share a simple way of creating a filesystem.

Most of the time, you will always be added by the your Linux distribution setup, that provides a cool graphic interface to create the disks partitions and filesystems.

However, you might need to add disks to a running server, to increase a directory size or fit a new software needs.

Here goes an overview of the process:

Identify the disk device file under /dev

Uses fdisk to check/create partitions using the MS-DOS partition table

Uses mkfs to create a filesystem

Uses e2label to optionally add a label to the filesystem

First, check the /dev filesystem for the disk device. If you are not sure about it, check the kernel ring through the dmesg command. In this post, we will assume that the device file is /dev/sdd.

Second, uses fdisk to create a single primary partition. According to your needs, you might create more than one partition, including logical ones.

# fdisk /dev/sdd Command (m for help): p ... Command (m for help): n ... Command (m for help): w

The commands above prints the current partition table (p), creates a new partition (n) and writes the partition table (w). The fdisk dialog is very simple and clear, you will not have problems to answer his its questions.

Third, uses the mkfs to create a new filesystem. The main argument here is "-t" that will define the type of the filesystem. In this post, we will create a ext3 filesystem:

# mkfs -t ext3 /dev/sdd1

Last, you can optionally add a label to your new filesystem, using the e2label program. Labels might be useful when you do not want to record devices filenames.

# e2label /dev/sdd1 LINUXLAB

In this post, I will share a simple way of resetting the root user password within a Linux system. Although enough, you can find another techniques around the Internet. This technique is simple and worked for me.

Here goes an overview:

Boot the system in single user runlevel (1)

Uses the shell to reset the password

The first step is to forget the root password. You should not do it on purpose, of course.

Second, you will need physical access to the system, as there are no network in runlevel 1 for default. This will not work for remote servers, like those in Amazon AWS.

Third, reboot the system and on the boot loader (probably Grub), select the kernel which you want to boot and press "a". This will allow you to append arguments to the kernel initialization. Add the "single" argument to boot in runlevel 1 (single user). The kernel command will look like this:

ro root=LABEL=/ single

Press enter and watch the system boot.

Fourth, use the passwd program to reset the root password:

passwd root

Finally, reboot the system again. You will be able to login with the new root password.

In this post, I will share how to define a new hardware address (aka mac address) to a network interface within a Linux System. This might be useful if you need to replace a broken network interface and keep all running smoothly.

By definition, a Linux system will give new network interfaces higher numbers, as eth1, eth2, eth3 and so on. This can be a huge problem if you used eth0 within scripts or configuration files.

To keep the old interface number (eth0) when you install a new interface, edit the /etc/udev/rules.d/70-persistent-net.rules file and change the value of the ATTR{address} attribute to the new network hardware address.

Save and reboot the server.

In this post, I will share how to setup some swap space on a Linux system. The swap, or Virtual Memory, might be essential on high load systems.

To learn more about Virtual Memory, take a look at:

http://simple.wikipedia.org/wiki/Virtual_memory

Step 1: Get a swap partition

First of all, you will need a clean partition. Use the fdisk tool to create a partition. As example, lets use the /dev/sdb disk:

# fdisk /dev/sdb

Do not forget to mark the partition as a swap partition. Use the fdisk t command and set the partition type id to 82.

To confirm, use the fdisk p command, the output might be something like:

Device Boot      Start     End     Blocks    Id  System /dev/sdb1        1         522     4192933+  82  Linux swap / Solaris

Step 2: Create a swap filesystem

Now, use the mkswap tool to create a swap filesystem in your brand new partition. If you like some sophistication, you can also add a label to the partition. Adding a label, will allow you to reference this partition using the label instead the device file, something like a symbolic link:

# mkswap -L SWAP /dev/sdb1

Step 3: Edit the /etc/fstab file

Naturally, every system administrator wants to get things done when the system boots. We can use the /etc/fstab file to tell the kernel to turn on swap every time the system boots. Add these lines to the end of the file:

LABEL=SWAP    swap    swap    defaults     0    0

Be careful! If you mess with this file, your system will not boot and you will need a recovery media.

Step 4: Turn the swap on

Now, turn on the swap:

# swapon -a

This is the same command that will be executed by the kernel on boot time.

Step 5: Check if the swap is active

Last, check if the swap is active using one of the following commands:

# cat /proc/swaps

Click on the title to checkout an awesome RFC that will give you the basics (and more) about network security policies.

The original URL is:

Click on the post title to check some benefits and examples of using SELinux.

The original URL is:

In this post I will share two simple commands that allows you to dump and/or reload a Bind DNS server cache. This is useful when you need to do some troubleshooting in the server.

First, sending the SIGINT signal to the named process forces it to dump its cache:

# kill -INT $(cat /var/run/named/pid)

Finally, sending the SIGHUP signal to the named process forces it to reload (clean) the cache:

Click on this post title to read another great chapter in Red Hat Enterprise Linux documentation, this time about virtualization of Linux servers.

Here is the URL:

Click on the post title to check a great chapter about clustered linux systems, available at the Red Hat Enterprise Linux official documentation.

The original URL is:

In this post, I will share a useful command that I use to copy files through the network using a pipeline with tar, gunzip and ssh. The power of this command lies on speed, as the copy is done through a "compressed ssh tunnel", which is very useful when you need to handle files through servers very quickly.

$ tar czvf - /home/me | ssh me@somewhere "tar xzvf -"

In this post, I will share a simple solution for a simple problem (more a decision) that I got in the past with my virtual development Linux server. It is very specific I know, but if you were lucky enough to get here, keep read on.

I am a network and system administrator and I have a sane problem: I like to develop when I have free time. I think in web applications, CSS layouts, these kind of stuff. Actually, CSS is very useful for network administrators, it learns you to think in a cascade and pluggable way and be more organized.

So, I had some free time and decided to download VirtualBox and create a virtual (Linux for sure) development server using Linux, Apache, MySQL and PHP. As I do in real life, I first created the server then started thinking about how to configure the network. Essentially, VirtualBox uses six modes of networking, which is flexible enough for most cases.

The first model was to use a bridged network mode, so I get as close as possible from a "real" environment. I set the network interface in this mode and after some seconds I was able to get an IP address from my home router. At that moment, it was good enough. I added an entry in my hosts file and started working.

The first issue raised when my home Internet connection got down and discovered that my virtual development server was highly dependent of a real network environment, I was using the bridged mode. First problem: I need a mode that works offline.

A few weeks later, I was in my mother-in-law house and then got some free time. I was bored and thinking on a few improvements for a personal project. Knowing that I could not work offline, I asked my brother-in-law the wireless key and started all over. Second problem: people tends to use different network addresses than you.

Putting simple, I got two problems: I need a mode that works offline and on different networks, so I do not lose the mobility of my notebook.

I took a good read on the VirtualBox Networking Documentation to understand the networking modes. I kept this in my mind and then, when I was in a bus I got a simple solution:

To use two network interfaces. eth0 uses the Host-only mode, where we create a virtual network between the virtual machine and the host (my notebook). This solved my offline issue. eth1 uses the Bridged mode, which allows me to connect to any real network, gets a real IP address and reach the Internet.

In this post, I will do a quick talk about guest shares on a Samba file server. Keep up to, it will not take more than 5 minutes.

Putting simple, a guest share is a share which one do not need to provide user credentials to access, as the name implies, guests are welcome.

Although simple in theory, these guest shares might become a little more complex, according to your specific scenario.

Talking about configuration, there are two main options that affects how guest shares works.

guest account: defines which Linux system user account will be used for guest access. Here, you can use the nobody user, create a guest user or whatever. This option goes on the global section

guest ok: defines whether the share will be a guest share. Use a boolean here, like true, 1 or yes. This options goes on the share section

First, sets the guest user account:

[global] guest account = nobody

Now a simple guest share:

[public] path = /var/www/public comment = This is a public share guest ok = yes # You also can use public = yes

With this configuration, everyone in your network will be able to read/write on the www share - please do not do this in a production environment. Obviously, the nobody user account should have the proper permissions on the system directory, in this case, the /var/www was changed to user and group nobody.

Another cool configuration is the readonly share. In this share, the files and directories should be readable for everyone but writable for only a few, usually a group. Let's take a look at the configuration for a readonly share.

[readonly] path = /var/www comment = This is a readonly share guest ok = yes write list = @webdev

A little more of spice...

[readonly] path = /var/www comment = This is a spicy readonly share guest ok = yes write list = @webdev force group = +webdev create mask = 0775 directory mask = 0775

The force group = +webdev line is very interesting. It will force the files and directory groups to webdev, if the accessing user belongs to this group. This is nice for group shares.

The create mask and directory mask simple defines a Unix mask that will be ANDed with the file effective mask, calculated from the DOS mapping to Unix mask. Putting simple, using these options you can assure that the files and directories will have, AT LEAST, these Unix modes.