Here’s how to fix it.

The issue is caused by the box you are SSHing into trying to resolve your IP address to a DNS name for the logs. We don’t really need this (especially for non-public servers), so we may as well disable it, drastically speeding up the time it takes for the box to prompt you for your password in the process (so drastic it becomes instant, in fact).

1. Login to your box as normal via SSH.
   .
2. Type in at the dollar prompt:

   $ sudo nano /etc/ssh/ssh_config

3. A text editor showing you the SSH server config will appear. Scroll right to the very bottom of the file.
   .
4. Now add the following line in (indent it to match the existing lines with a few spaces at the front):

   UseDNS no

5. Save your changes by pressing CTRL+X, then “Y” and then Enter.
   .
6. Restart the SSH server with the following command (won’t affect your existing session):

   $ sudo service ssh restart

7. Log out and test logging in via SSH again. This time the password prompt will appear immediately.
   .
8. Pat yourself on the back. You’re done.

This fix applies to any version of Ubuntu (or just about any distro, for that matter).

Why would you want to do this? Well, say for example you have a MythTV Frontend PC. For the most part, most Frontends are dedicated PC’s connected to a TV or projector that are not used for any other purpose, so technically there is really no need to have a hard-drive inside one as nothing new will ever be stored (all the media is streamed from the Backend server). There’s also the added bonus of less noise by not having a hard-drive installed.

Another good example of using a Diskless boot environment is for performing offline virus scans of Windows based PC’s in a safe environment that is not Windows, using tools like ClamAV. In a corporate environment, having a “normal” installation makes it easier to setup default settings that normally don’t suit booting up from an Ubuntu Live CD, such as corporate Proxy settings. Making a Diskless Boot setup is far easier than creating a customised Live CD in this instance.

So how exactly do you create a diskless booting PC?

Pre-requisites:

• First up, you need a working PXE server. If you haven’t built one yet, you can refer to my previous tutorial on how to build one.
• Sufficient space to store a typical Ubuntu Desktop install. I would recommend having at least 10GB free to allow space for both the OS and any additional apps you might install.
• A PC that you will using as your diskless machine that has the ability to boot from PXE.
• A temporary hard-drive for your diskless machine that we will be using to do our initial build on.
• A fast network. Diskless booting is only useful if your network can transfer data quickly. I would recommend that you use a gigabit network.
• It is recommended to have at least 2GB RAM on your diskless PC, primarily because we will not be running a swap file or swap partition (swapping over a network connection is just silly).
• Ubuntu 10.04.2 LTS was used for this tutorial, but you should be able to use just about any release of Ubuntu. Note: The version of Ubuntu on your PXE server and the version of Ubuntu you are turning into diskless boot does not have to be the same version, but I would recommend using 10.04.2 LTS or later.

What to do:

1. First up, install the temporary hard-disk into your PC that is going to become diskless. You can attach it in the traditional way or connect it via USB.
   .
2. Perform a normal installation of Ubuntu Desktop Edition onto the hard-drive of your PC.
   .
3. Once installed, update the system with any outstanding system updates.
   .
4. Install any additional apps you wish to utilise on this machine, such as MythTV.
   .
5. Once you are satisfied that your installation is setup the way you want it, we need to prepare the installation to be able to boot via PXE. Start by opening up a terminal and type in the following commands:

   $ cd ~
   $ sudo nano /etc/initramfs-tools/initramfs.conf

6. You will now be looking at a text configuration file in the Nano text editor. This is the configuration file that controls how the boot-time RAM disk image is created. We need to modify it so that we can create a PXE-ready version of the RAM disk image instead of the usual type. Scroll down until you find the configuration line that looks like:

   MODULES=most

   …and modify it so that is now looks like:

   MODULES=netboot

7. Scroll down a little further until you see the line:

   BOOT=local

   …and change it to read:

   BOOT=nfs

8. Press CTRL+X, then “Y” and then “Enter” to save your changes. You will be returned to the Terminal prompt.
   .
9. Now let’s create our custom RAM Disk image using the following command:

   $ sudo mkinitramfs -o ./initrd.img

10. Once that has finished, we also need an unmodified copy of the current kernel you are running:

    $ cp /boot/vmlinuz-`uname -r` ./vmlinuz

11. Now we need to prep the server. First up, we need to setup a way to transfer our temporary hard-disk contents to the PXE server, and we also need to ready how the PXE server will serve the data to the diskless workstation when booting. We will be using the Network FileSystem or NFS for short. On your Ubuntu PXE Server, open a terminal and type in the following:

    $ sudo apt-get install nfs-kernel-server

12. Once the NFS server is installed, we need to setup the directory that will store and serve the Diskless system’s files. Assuming you have all your PXE related data under /srv/tftp, let’s create the following for our Diskless station data:

    $ sudo mkdir -p /srv/nfs/disklessboot

13. Now let’s tell the NFS server about this new directory:

    $ sudo nano /etc/exports

14. In the text file that appears in the Nano text editor, scroll to the bottom of the file and add the following line:

    /srv/nfs/disklessboot          *(rw,no_root_squash,async)

    The above will allow read/write access to the path /srv/nfs/disklessboot on your PXE server from any incoming IP address.

15. Save your changes by pressing CTRL+X, then “Y” and then “Enter”.
    .
16. Now activate the new NFS share by using the following command:

    $ sudo exportfs -a

17. We’re nearly there! Now we just need to transfer the contents of our temporary workstation hard-drive to the new NFS share. On your soon-to-be-diskless PC (not the server), and assuming that your PXE server’s IP address is 192.168.0.10, type the following in a Terminal:

    $ mkdir /dev/shm/nfs
    $ sudo mount -t nfs -onolock 192.168.0.10:/srv/nfs/disklessboot /dev/shm/nfs

    The above will create a directory in your RAM disk to mount the PXE server’s NFS share on and the second command will mount the NFS share /srv/nfs/disklessboot in the directory you just created in your RAM disk locally, thus allowing you to now read and write data to that NFS share by using /dev/shm/nfs locally.

18. Now let’s copy the OS files from the temporary hard-disk to the PXE server’s NFS share:

    $ sudo cp -avx /. /dev/shm/nfs/.
    $ sudo cp -avx /dev/. /dev/shm/nfs/dev/.

    The above commands will take a minute or two to finish and will show you a giant list of files being copied.

19. We are now done with the soon-to-be-diskless PC, so shut it down and remove your temporary hard-drive.
    .
20. Go back to your server and type in the following at the Terminal prompt:

    $ sudo mkdir -p /srv/tftp/disklessboot
    $ sudo cp /srv/nfs/disklessboot/home/USERNAME/vmlinuz /srv/tftp/disklessboot/
    $ sudo cp /srv/nfs/disklessboot/home/USERNAME/initrd.img /srv/tftp/disklessboot/

    Note: Replace USERNAME above with whatever your username was on the temporary hard-disk setup, eg: “/home/jbloggs/”)

21. We now need to change the network setup of the diskless boot install so that it doesn’t try to configure itself with a new IP address on boot (since it’ll already have one when you do the initial PXE boot):

    $ sudo nano /srv/nfs/disklessboot/etc/network/interfaces

22. Look for a reference to “eth0″ which is your first ethernet adapter. It will generally be the last line in the file and may look like:

    auto eth0

    …or:

    iface eth0 inet dhcp

23. Once you locate it, comment out that line by putting a hash symbol at the front like so:

    #auto eth0

    …or:

    #iface eth0 inet dhcp

24. Now add a new line with the following:

    iface eth0 inet manual

25. Save your changes with CTRL+X, then “Y” and then “Enter”.
    .
26. Now we need to ensure that the devices mounted by your diskless booting PC do not try to mount physical disk devices for things like /home and the root filesystem. Type in the following command:

    $ sudo nano /srv/nfs/disklessboot/etc/fstab

27. Once in the text editor, DELETE EVERYTHING and replace it all with the following:

    proc            /proc           proc    defaults        0       0
    /dev/nfs        /               nfs     defaults        1       1
    none            /tmp            tmpfs   defaults        0       0
    none            /var/run        tmpfs   defaults        0       0
    none            /var/lock       tmpfs   defaults        0       0
    none            /var/tmp        tmpfs   defaults        0       0

28. Save your changes with CTRL+X, then “Y” and then “Enter”.
    .
29. That’s the Diskless setup ready to go. All that we now need to do is configure the PXE boot menu to give you an option to boot Diskless. To do this, edit your PXE boot menu configuration file. Assuming you called it /srv/tftp/mybootmenu.cfg, type in the following:

    $ sudo nano /srv/tftp/mybootmenu.cfg

30. In the text editor, scroll down to the bottom of your configuration file and add the following lines:

    label My Diskless Boot PC
       kernel disklessboot/vmlinuz
       append initrd=disklessboot/initrd.img root=/dev/nfs nfsroot=192.168.0.10:/srv/nfs/disklessboot ip=dhcp rw

    The above menu configuration will tell PXE to load the kernel and custom RAM disk image from the TFTP directory and then transfer control of the boot process to the OS files located in /srv/nfs/disklessboot which will be mounted on the diskless system at /dev/nfs as the root filesystem.

31. Save your changes by pressing CTRL+X, then “Y” and then “Enter”.
    .
32. Finally, make sure that the TFTP daemon can read the new files you added to it with the following command:

    $ sudo chmod 777 -R /srv/tftp

33. That’s it! You are now ready to boot! Turn on your now-diskless PC and tell it to boot from PXE. When the PXE boot menu appears, choose “My Diskless Boot PC” and watch in wonder and amazement as Ubuntu boots up without a hard-disk as if by magic!

NOTE: Whilst the Diskless PC will boot and operate much like any ordinary PC and in fact you can indeed perform future system updates this way, be aware that the kernel itself will NOT get automatically updated. Newer kernels will get installed to the NFS mount as you would expect, however the kernel actually boots from your TFTP directory, not the NFS share, so if you do update the kernel in future, you must create a new RAM disk image manually as per step 9 and copy both it and the new kernel file into your TFTP’s directory (in this case /srv/tftp/disklessboot/) otherwise your updated system will not boot anymore. Alternatively, simply allow all updates to be performed except for kernel updates.

Finally, hardware drivers: If you install the NVidia proprietary drivers for your Diskless Boot system, ensure that you don’t try to boot other systems that do not contain NVidia graphics hardware otherwise X will break. The reason for this is that the NVidia driver installation makes a number of changes to system-level libraries to work, effectively binding that system to NVidia graphics-based systems only. If you intend to use your Diskless setup on multiple hardware platforms, such as NVidia, Intel and ATi, do not install the proprietary NVidia (or ATi) driver.

Enjoy!

If you want to access your test site remotely, you need to know your public IP address, however your home ISP gives you a dynamic public IP address and every time you have a power failure, or reboot your router, you are assigned a brand new public IP address. This makes it very annoying if you are testing your site remotely.

Sure, you could use a Dynamic DNS service to keep track of when your public IP address changes, but what if you have a paranoid client who does not want to use even Dynamic DNS? How do you keep track of your new public IP without having to get to your internal network to read it each time?

What you need is a way to be able to have the system send you an email with your current public IP address so that there is no guess work involved. But how do we do this?

To determine your public IP, you need to have an external site to refer to which will record your public IP as connecting to it. You could use services such as What Is My IP or similar, but we don’t want to rely on an external service, nor do we want to hammer a third-party service with requests. Basically we just want to replicate the functionality of such a site as locally as possible and then use an email tool to send the information to you in a message.

Pre-requisites:

• Access to a web space outside of your working space with an IP or address that does not change. Generally every ISP out there will grant you some small quantity of complimentary web space to store some files. If you are working for clients like I am, you can use your own server if you like (after all, this blog is served from a fixed IP with a domain name attached to it).
• A PHP interpreter in your web space’s host. If you are using your own static host on Ubuntu, Apache 2 from a normal Ubuntu install be it Desktop or Server, then you should have the PHP interpreter installed by default. If not, then install the php5 package.
• If using your own static server, you will need a web server such as Apache installed on it to serve the PHP we’ll be creating.

Now let’s look at using a simple bit of PHP to obtain the address of the connecting host.

1. On your external static web space (not your local dynamic site), create a new text file somewhere readily available called getmypublicip.php or similar.
   .
2. In your favourite text editor, copy and paste in the following PHP code:

   <?php
   $ip = getenv("REMOTE_ADDR");
   echo $ip;
   ?>

3. Save your changes and exit the text editor.
   .
4. Test your new PHP file by referring to it from your ISP’s webspace, eg: If your ISP’s website was called myisp.com and you had a login called “noddy”, then the address to your webspace might be http://www.myisp.com/~noddy/getmypublicip.php, so type that in your web browser, which is usually the path most ISP’s use.
   .
5. When you go to that URL, the PHP on the remote side should parse the referring public IP address and echo it to the page. In this case, because you are referring yourself, you should see your own public IP address (let’s say it’s 74.125.237.84 for the purposes of this guide) printed at the top-left of the web browser page. If you ran this PHP locally, you would get the internal IP of the connecting host instead, eg: 192.168.0.100 or similar instead.
   .
6. Awesome, so now how do we email this newly obtained information to ourselves? Simple. We will make use of a neat little command line tool called sendemail (not to be confused with the venerable sendmail). It’s not installed in Ubuntu by default, so install it with:

   $ sudo apt-get install sendemail

7. Let’s create the script that will use sendemail. On your local test website, create a new text file outside of your web-accessible directories, eg: /home/noddy/getpublicip.sh using your favourite text editor.

   $ nano /home/noddy/getpublicip.sh

8. In the text editor, copy and paste the following in:

   #!/bin/bash
   wget -O- http://www.myisp.com/~noddy/getmypublicip.php 2>/dev/null |  sendemail -u "Client's public IP address" -f test@blah.com -t noddy@mycompany.com -s mail.mycompany.com

9. Save your changes and exit your text editor.
   .
10. Now test your script my manually running it using the following command:

    $ sh /home/noddy/getpublicip.sh

11. Check your personal email. You should have received an email sent by test@blah.com containing a single line showing your public IP address!
    .
12. Now we need to setup the script to be executed periodically. We will setup a cron schedule to run our new bash script as follows:

    $ crontab -e

13. Your text editor will appear. In this, add the following line on the end:

    0 * * * * sh /home/noddy/getpublicip.sh

14. Save your changes and exit your text editor.
    .
15. Now wait. On the next whole hour, you should receive an email in your mailbox containing a single line with an IP address – your public IP address! Now you need never know what your private server’s public IP address is if it changes in the middle of the day – on the hour you will be told what it is!
    .
16. If you’d like to receive notification on a more regular basis, modify the cron schedule. For example, the following change to step 13 will make your system run the script every 15 minutes:

    */15 * * * * sh /home/noddy/getpublicip.sh

17. That’s it! Pat yourself on the back. You’re done.

NOTE: Some ISP’s do not allow connections on dynamic IP’s to send periodic email on their mail server and may block it on the belief that your emails may be spam. Should that be the case, you should install a local mail daemon such as Postfix and change the sendemail arguments to direct their mail to localhost instead, and allow Postfix to send the notification email. In Ubuntu, you can install Postfix by simply installing the postfix package and electing the server to be an “Internet Site”. The default configuration is fine, but not secure. Ensure that no-one externally can access port 25 on your client’s server if you choose to go down this route.

Network booting the Live CD has obvious advantages – aside from booting faster than CD (especially on a gigabit network), it is indispensable as an emergency boot medium in a workplace environment, especially for broken Windows systems, and allows for Ubuntu effortless installations on netbook PC’s that don’t have optical drives and saves you having to have a USB stick handy.

Pre-requisites

• You need a working Linux PXE boot server. Doesn’t have to be Ubuntu, but it needs to be Linux. You cannot use a Windows PXE server.
• An Ubuntu 10.04 Live CD ISO or physical CD. Can be the 32-bit or 64-bit ISO, but you can also setup both of them at once!
• At least 700MB of drive space on your PXE server, more if you want to have more than one CD available.
• This tutorial was put together using Ubuntu Server 10.04 Lucid Lynx, but should work with all future releases and older versions to at least Ubuntu Server 8.04 Hardy Heron.

Getting it together

1. Login to your PXE server and mount the CD or ISO image (in this example we are copying the 32-bit disc). Assuming the CD is mounted at /media/cdrom, copy the CD’s files to your server as follows:

   $ sudo mkdir -p /srv/tftp/ubuntu-livecd-boot/i386
   $ sudo mkdir -p /srv/ubuntu-livecd/i386
   $ sudo cp -av /media/cdrom/* /srv/ubuntu-livecd/i386
   $ sudo cp -av /media/cdrom/.disk /srv/ubuntu-livecd/i386
   $ sudo cp -av /media/cdrom/casper/initrd.lz /srv/tftp/ubuntu-livecd-boot/i386
   $ sudo cp -av /media/cdrom/casper/vmlinuz /srv/tftp/ubuntu-livecd-boot/i386

1. (if you want to setup the 64-bit disc, then replace all instances of “i386″ with “amd64″, or you can setup both architectures by setting up both directories)
   .
2. Make sure the permissions of the files to be loaded by TFTP are correct:

   $ sudo chmod 777 -R /srv/tftp

3. The astute of you will have noticed that we have copied the Ubuntu CD outside of the TFTP directory and that we have made a separate copy of only two of the disc’s files inside the TFTP directory. Why is this? Well, the vmlinuz and initrd.lz files are the only files that TFTP will need to download to get started with the boot process. After that, we will use NFS to deliver the rest of the Live CD, so let’s set that up:

   $ sudo apt-get install nfs-kernel-server

   Note: I should point out that the files that are copied to the NFS share do not have to be on the PXE server. The NFS server can be any box as NFS is not related to PXE at all.

4. Once that is installed, we need to define an NFS share. This is done in the NFS exports file, so let’s get that into a text editor:

   $ sudo nano /etc/exports

5. Add the following line to the bottom of the file:

   # Ubuntu Live CD files for PXE booting
   /srv/ubuntu-livecd/i386        *(ro,async,no_root_squash,no_subtree_check)

   …if you are setting up the 64-bit version, replace “i386″ with “amd64″, or if setting up both architectures, list it as follows:

   # Ubuntu Live CD files for PXE booting
   /srv/ubuntu-livecd/i386        *(ro,async,no_root_squash,no_subtree_check)
   /srv/ubuntu-livecd/amd64       *(ro,async,no_root_squash,no_subtree_check)
   

   A breakdown of the parameters on the right is as follows:

   • The asterisk means “share with everyone on this network, regardless of who they are”.
   • The “ro” parameter means to share the data as read-only.
   • The “async” parameter allows the NFS server to reply before data is written to the share. Since it’s mounted as read-only anyway, the parameter is only there to keep NFS’ syntax happy.
   • The “no_root_squash” parameter means to allow the NFS client to use the mount as a root filesystem, otherwise it’s mounted as “nobody” instead of “root”. Since the LiveCD is essentially a diskless client, we need to be able to define the NFS mount as a root volume.
   • The “no_subtree_check” parameter helps to speed up transfers. Normally NFS will check to see if a requested file exists in an exported sub-directory. This slows things down, so turning this off means the only check that is made is that the requested file exists on the exported filesystem. Subtree checking can also cause issues when an open file is renamed, but since the export is read-only, this is irrelevant.
     .
6. Save your changes with CTRL+X, then “Y” and then Enter.
   .
7. Now have NFS re-read its export file and begin sharing the specified directories with:

   $ sudo exportfs -a

8. We’re nearly ready to rock and/or roll. All we need to do now is prepare the PXE boot menu to launch the Live CD for us. Assuming you followed by previous tutorial and your boot menu file is called mybootmenu.cfg:

   $ sudo nano /srv/tftp/mybootmenu.cfg

9. Assuming your NFS server’s IP address is 192.168.0.10, insert the following lines for a 32-bit Live CD entry:

   label Live CD 32-bit
       kernel ubuntu-livecd-boot/i386/vmlinuz
       append boot=casper netboot=nfs nfsroot=192.168.0.10:/srv/ubuntu-livecd/i386 initrd=ubuntu-livecd-boot/i386/initrd.lz -- splash quiet

   …and if you are doing 64-bit, you can replace or add as a separate menu option the following:

   label Live CD 64-bit
       kernel ubuntu-livecd-boot/amd64/vmlinuz
       append boot=casper netboot=nfs  nfsroot=192.168.0.10:/srv/ubuntu-livecd/amd64  initrd=ubuntu-livecd-boot/amd64/initrd.lz -- splash quiet
    

   The kernel line is the actual kernel that is loaded to run the session. The append line tells the boot process several things. First up, the casper directory contains the boot files (a SquashFS image in the case of the Live CD), the root of the NFS file system is located on the server with the IP 192.168.0.10 under the path /srv/ubuntu-livecd/i386 (or amd64) and that the image to fire up the RAM disk with is located under ubuntu-livecd/i386/initrd.lz and finally we have advised the boot process to suppress console messages and display the splash screen while loading. From this point on, the initrd.lz is extracted and will grab the SquashFS image from the casper directory via NFS, extract it and commence running it. All other files that the disc needs to do its thing (such as when installing Ubuntu from the Live environment) are also provided by the NFS share.

10. Save your changes by pressing CTRL+X, then “Y” and then Enter.
    .
11. You’re now ready to go – reboot your PXE workstation and you should now see an entry for your Live CD. Upon choosing it, your system should fire up with the Live environment! Once booted, you can even do an Ubuntu install to the PC you started on as though you’d booted from a CD or USB device!
    .
    

Enjoy!

Why would you want to boot a PC from the network? Well, it opens the door to booting diskless workstations, eg: Internet Cafe PC’s, or if you regularly install tens or hundreds of PC’s, you can start the installer on all those machines at once without needing to have individual boot/install media for each machine. You can even use Linux PXE for starting Microsoft Windows network installers and tools.

This article is going to show you how to setup a standard Ubuntu 10.04 Lucid Lynx Server to respond to a PXE boot request and present a boot menu ONLY. I will put practical applications such as installing Ubuntu over the network or booting a Live CD over the network into separate future articles.

Pre-requisites:

• A PC or virtual machine with an installation of Ubuntu Server on it. This tutorial was written using an Ubuntu Lucid 10.04 Server, but these instructions will work equally well on nearly any version of Ubuntu Server. This tutorial will not detail the initial build of a server as it is relatively straight forward.
• A DHCP server that allows you to specific PXE boot information. Most consumer routers will not give you these options. Suitable DHCP servers are the DHCP daemon on Ubuntu Server, third-party Linux router solutions such as Smoothwall or pfSense, and Windows Server among others.
• If your DHCP server is a dedicated network/firewall device that you do not wish to use as a file server to serve the network boot files, then you will need a separate PC to be a file server as well.
• Some free disk space. PXE booting take bugger-all space, but whatever you plan to serve from it will need space. If you plan to setup the Ubuntu Live CD to be bootable from PXE, you will need 700MB+ of hard-drive space on that server. You will need more than this if you wish to host things like multiple LiveCD’s such as both the 32-bit and 64-bit versions, or multiple different distributions.
• A PC workstation that has PXE boot capability. Any PC built in the last 10 years should definitely have this capability, though you may be required to enable it in BIOS. If you do not have a PC that can do this, you can use a virtual machine such as Virtualbox instead (you could have a virtual machine PXE boot off a virtual PXE boot server too! ).
• A copy of Ubuntu Server 10.04 that suits your server’s architecture.
• A copy of the Ubuntu ALTERNATE Install CD 10.04 that we need to get some PXE boot files from. Unfortunately the Live CD does NOT contain the files we need.

At the end of this exercise we will have a PXE server that will boot into a selection menu that will give us choices of things to do. We’re also going to pretty up the menu with a background image instead of just having plain boring text, and we’ll do this using some of the existing elements on the Ubuntu CD as most of the work has been done for you already!

Getting it together:

1. Login to your server.
   .
2. Let’s install the software we need:

   $ sudo apt-get install tftpd-hpa inetutils-inetd

   …this will install a Trivial FTP server which is essentially a super-simple FTP server plus the inetd daemon which will listen out for TFTP requests and direct them to the TFTP daemon.Before you ask, no you cannot use a regular FTP daemon like vsftpd or similar. It has to be a TFTP daemon. Beware: Ubuntu has two TFTP options in the repository – you must use the HPA version of the daemon as shown, as it handles large boot images while the other daemon does not. If you don’t use it, you will see boot errors.

3. By default Ubuntu sets up the TFTP daemon’s root directory to be /var/lib/tftpboot which may not suit your requirements. For the purposes of this tutorial, we will be changing this to /srv/tftp instead. To do this, we need to edit the /etc/inetd.conf file in a text editor:

   $ sudo nano /etc/inetd.conf

4. Scroll down to the bottom of the file and modify the tftp line (or add it if it’s missing) and substitute /var/lib/tftpboot bit on the end of that line with the path to your preferred directory:

   tftp    dgram   udp    wait    root    /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /srv/tftp

5. Save your changes by pressing CTRL+X and then “Y” and then Enter.
   .
6. Now we need to tell the Trivial FTP daemon where our TFTP root is. Open its config file with:

   $ sudo nano /etc/default/tftpd-hpa

7. Modify the TFTP_DIRECTORY line (usually the fourth line from the top) to be /srv/tftp:

   # /etc/default/tftpd-hpa
   
   TFTP_USERNAME="tftp"
   TFTP_DIRECTORY="/srv/tftp"
   TFTP_ADDRESS="0.0.0.0:69"
   TFTP_OPTIONS="--secure"

8. Save your changes by pressing CTRL+X and then “Y” and then Enter.
   .
9. Now we just need to restart the inetd and tftp services with:

   $ sudo service inetutils-inetd restart
   $ sudo service tftpd-hpa restart

10. So that’s the TFTP daemon ready to serve files to a PXE agent. Now we need to create the directory where we will be putting all our PXE goodness into:

    $ sudo mkdir -p /srv/tftp

11. We now need to copy some files off the Ubuntu Alternate Install CD that make up the PXE boot files and the menu config files. I will use the 32-bit disc in this example, though the files are the same on the 64-bit disc for this step. Insert the CD or mount the downloaded Ubuntu Alternate Install CD ISO. In this case I will assume you have a physical CD mounted at /media/cdrom.

    $ sudo cp /media/cdrom/install/netboot/pxelinux.0 /srv/tftp
    $ sudo mkdir -p /srv/tftp/ubuntu-installer/i386
    $ cd /media/cdrom/install/netboot/ubuntu-installer/i386
    $ sudo cp -R boot-screens /srv/tftp/ubuntu-installer/i386
    $ sudo cp initrd.gz linux /srv/tftp/ubuntu-installer/i386
    

    (if you’re using the 64-bit CD, substitute all instances of “i386″ above with “amd64″ instead.)
    .

12. Now we need to setup the initial PXE boot process:

    $ sudo mkdir /srv/tftp/pxelinux.cfg
    $ sudo nano /srv/tftp/pxelinux.cfg/default

13. You will now be looking at a blank text editor. In this, type the following:

    include mybootmenu.cfg
    default ubuntu-installer/i386/boot-screens/vesamenu.c32
    prompt 0
    timeout 100

    The timeout 100 line will provide a 10 second countdown before it automatically chooses the default PXE menu option when you boot into it. If you do not want a timeout, then change this to timeout 0 instead.

14. Press CTRL+X and then “Y” and then Enter to save your changes.
    .
15. Now let’s setup our actual boot menu that we’ll be choosing options from:

    $ sudo nano /srv/tftp/mybootmenu.cfg

16. Again you’ll be looking at a blank text editor. Type (or copy & paste) in the following. Indenting text is not important, but makes it more readable:

    menu hshift 13
    menu width 49
    menu margin 8
    menu title My Customised Network Boot Menu
    include ubuntu-installer/i386/boot-screens/stdmenu.cfg
    menu begin Cool options
        default myfirstoption
        label myfirstoption
            menu label This is a menu item
        label mysecondoption
            menu label This is another option
    menu end

17. Press CTRL+X, then press “Y” and then Enter to save your changes.
    .
18. Finally, we need to change the permissions of all files concerned because TFTP will not read any files unless they are set to full access:

    $ sudo chmod 777 -R /srv
    

19. That’s PXE server side ready to go. Now we need to tell PXE clients where to find the PXE boot server. If you are NOT using Ubuntu as your DHCP server, then skip to step 23, otherwise do the following:
     

    $ sudo nano /etc/dhcp3/dhcpd.conf

20. This opens up the DHCP config file into your text editor. Assuming your PXE server is at 192.168.0.10, scroll right to the very bottom of this file and add the following:

    next-server 192.168.0.10;
    filename "pxelinux.0";

    (note the semi-colon on the end)
    .

21. Press CTRL+X, then “Y” and then Enter to save your changes.
    .
22. Restart the DHCP daemon with:
     

    $ sudo service dhcp3-server restart

23. If you’re using a non-Ubuntu DHCP server, then look for any “network boot” options and specify the PXE boot server’s IP address and path to the pxelinux.0 file there. For example, in Smoothwall, you would go to Services->DHCP and then check the “Network boot enabled” checkbox, then specify “192.168.0.10″ (to suit our tutorial) into the “Boot server” box and “pxelinux.0″ in the “Boot filename” box and “/srv/tftp” in the “Root path” box.
    .
24. We should now be ready to try out our PXE boot server! On your test workstation or VM, enable booting off the network (in the case of PXE booting a Virtualbox VM, you must ensure that the network adapter is set to “bridged mode” instead of “NAT”) and fire away. You should first see your PC launch its PXE agent, looking for a DHCP server to tell it where the PXE server is:
    
    
25. If your PXE server is working, within a few seconds you will see your boot menu!.
    …and if you hit Enter on “cool options” you will now see a sub-menu showing your two options that we created.
    .
26. Well this is all well and good, but the menu currently doesn’t actually DO anything other than show us a bunch of options. How about we provide something, say the Memory Test application from the Ubuntu Alternate Install CD? Plus we’ll add an option to boot from the first HDD in your system. If your CD is still mounted on the server, then go back into the terminal you’ve been working in and copy over the MemTest app as follows:

    $ sudo cp /media/cdrom/install/mt86plus /srv/tftp

27. Now let’s add a menu entry for it:

    $ sudo nano /srv/tftp/mybootmenu.cfg

28. Modify the file so that it now looks like the following (add just the bolded lines):

    menu hshift 13
    menu width 49
    menu margin 8
    menu title My Customised Network Boot Menu
    include ubuntu-installer/i386/boot-screens/stdmenu.cfg
    label Boot from the first HDD
        localboot 0
    label Memory Tester
        kernel mt86plus
    menu begin Cool options
        default myfirstoption
        label myfirstoption
            menu label This is a menu item
        label mysecondoption
            menu label This is another option
    menu end

29. Save your changes and exit.
    .
30. Ensure the permissions of everything, including our newly copied files, have the correct permissions for TFTP to work:

    $ sudo chmod 777 -R /srv

31. Reboot your test PC via PXE and this time you will see your menu sports the new menu options at the top (you could have equally placed them at the bottom too):
    
32. Choosing “Memory Tester” from the menu will launch the MemTest app straight away, just like off the CD. But by now you are probably wondering “Aren’t we building off Ubuntu 10.04? Why does the menu have the old logo on it? Can we change it?” Sure, we can!
    .
33. The Ubuntu 10.04 installer CD has got the new Ubuntu logo, but for some reason it’s only saved as a PCX file which won’t work for the PXE boot menu. We can fix this by simply re-saving the PCX file as a PNG file. To start with, get a copy of the splash.pcx file from the /isolinux directory on the Ubuntu CD. This is the new Ubuntu logo that you normally see on the CD’s boot menu.
    .
34. Load this file into an image editor such as The GIMP and re-save it as a PNG file, eg: splash.png (of course there’s nothing stopping you from creating your own graphic either – just make sure it’s no greater than 640×480 in size and indexed down to 16 colours).
    .
    
35. Copy the re-saved image file into /srv/tftp/ubuntu-installer/i386/boot-screens and overwrite the original splash.png file.
    .
36. Ensure that the permissions of the newly added file is set correctly again with:

    $ sudo chmod 777 -R /srv

37. And when you reboot your PXE workstation again, your menu will now look like:
    

And there you have it. A working PXE server with menu!

I will document further uses of the PXE boot facility in future articles, including how to boot the Live CD environment without the CD or a USB key, setup the ability to use your local Ubuntu mirror as an installation source for new installs, how to launch tools like Clonezilla and also how to setup a diskless boot system that uses PXE to do a normal Ubuntu desktop boot directly off the network without a local hard-drive.

Stay tuned!

If your host is not configured with a Hosts file entry on your local PC or a DNS entry to associate a name with an IP address, Ubuntu can only use the IP address of that PC to communicate with it which means you have to remember what that IP address is with your feeble grey-matter in your head. Likewise, Ubuntu will not respond to a Windows PC pinging its NetBIOS name because Ubuntu does not use NetBIOS at all by default and so it will ignore such requests.

So how do we get Ubuntu to resolve NetBIOS names like Windows? And how can we allow Windows to ping Ubuntu like another Windows PC? Read on…

Let’s illustrate the problem first. You’ll need a Windows PC on your network to test this. For this article, the Ubuntu PC will be called “gordon” and the Windows PC will be called “alyx”.

On either PC, if you open a terminal or Command Line window and ping the opposing machine, eg:

$ ping alyx

or

C:\> ping gordon

You get an error stating that the host cannot be found. Now in the case of Windows, if you were to ping another Windows PC instead of an Ubuntu PC, you can ping its name with no problem.

Let’s sort this out, shall we?

Allowing Ubuntu to ping Windows NetBIOS names

Ubuntu is setup for Linux use, not Windows use, so we need to install a package that will allow Ubuntu to more readily mix in with Windows networks and use NetBIOS. This package is called “winbind”.

1. Open a terminal and type in the following at the terminal prompt:

   $ sudo apt-get install winbind

2. Once installed, we need to tell Ubuntu to use WINS (as provided by winbind) to resolve host names. Type in:

   $ sudo gedit /etc/nsswitch.conf

   …which will open the file into the Gnome Editor.
   .

3. Scroll down to the line that starts with “hosts:”. In Ubuntu Jaunty, it looks similar to this:

   hosts:          files mdns4_minimal [NOTFOUND=return] dns mdns4

4. Add the word “wins” on the end of this line such that is now looks like:

   hosts:          files mdns4_minimal [NOTFOUND=return] dns mdns4 wins

5. Save and exit the editor.
   .
6. Now let’s ping the name of our Windows box again.

   $ ping alyx

   …and it now resolves!
   .

7. Pat yourself on the back.

Allowing Windows to ping Ubuntu NetBIOS names

This is just one half of the equation. We now need to allow Windows to be able to ping Ubuntu PC’s using its NetBIOS name. This requires Ubuntu to recognise and respond to that request. We need to setup a server daemon to do this. In Ubuntu, this particular server daemon is called Samba.

1. Installing Samba is simplicity itself. Open a terminal and type in:

   $ sudo apt-get install samba

2. Once that has finished, your Ubuntu PC will automagically respond to all NetBIOS queries for its hostname straight away, and that’s not just from Windows machines, but other Ubuntu machines (configured with the “winbind” package) as well.
   .
3. Pat yourself on the back again.