Thin Clients Booting over a Wireless Bridge
One of the requirements for a thin client is that it has a PXE-bootable system. Normally, PXE is part of your network card BIOS, but if your card doesn't support it, you can get an ISO image of Etherboot with PXE support from ROM-o-matic (see Resources). Looking at the server with, for example, ten clients, it should have plenty of hard disk space (100GB), plenty of RAM (at least 1GB) and a modern CPU (such as an AMD64 3200).
The following is a five-step how-to guide on setting up an Edubuntu thin-client network over a fixed network.
1. Prepare the server.
In our network, we used the standard standalone configuration. From the command line:
sudo apt-get install ltsp-server-standalone
You may need to edit /etc/ltsp/dhcpd.conf if you change the default IP range for the clients. By default, it's configured for a server at 192.168.0.1 serving PXE clients.
Our network wasn't behind a firewall, but if yours is, you need to open TFTP, NFS and DHCP. To do this, edit /etc/hosts.allow, and limit access for portmap, rpc.mountd, rpc.statd and in.tftpd to the local network:
portmap: 192.168.0.0/24 rpc.mountd: 192.168.0.0/24 rpc.statd: 192.168.0.0/24 in.tftpd: 192.168.0.0/24
Restart all the services by executing the following commands:
sudo invoke-rc.d nfs-kernel-server restart sudo invoke-rc.d nfs-common restart sudo invoke-rc.d portmap restart
2. Build the client's runtime environment.
While connected to the Internet, issue the command:
If you're not connected to the Internet and have Edubuntu on CD, use:
sudo ltsp-build-client --mirror file:///cdrom
Remember to copy sources.list from the server into the chroot.
3. Configure your SSH keys.
To configure your SSH server and keys, do the following:
sudo apt-get install openssh-server sudo ltsp-update-sshkeys
4. Start DHCP.
You should now be ready to start your DHCP server:
sudo invoke-rc.d dhcp3-server start
If all is going well, you should be ready to start your thin client.
5. Boot the thin client.
Make sure the client is connected to the same network as your server.
Power on the client, and if all goes well, you should see a nice XDMCP graphical login dialog.
Once the thin-client network was up and running correctly, we added a wireless bridge into our network. In our network, a number of thin clients are located on a single hub, which is separated from the boot server by an IEEE 802.11 wireless bridge. It's not an unrealistic scenario; a situation such as this may arise in a corporate setting or university. For example, if a group of thin clients is located in a different or temporary building that does not have access to the main network, a simple and elegant solution would be to have a wireless link between the clients and the server. Here is a mini-guide in getting the bridge running so that the clients can boot over the bridge:
Connect the server to the LAN port of the access point. Using this LAN connection, access the Web configuration interface of the access point, and configure it to broadcast an SSID on a unique channel. Ensure that it is in Infrastructure mode (not ad hoc mode). Save these settings and disconnect the server from the access point, leaving it powered on.
Now, connect the server to the wireless node. Using its Web interface, connect to the wireless network advertised by the access point. Again, make sure the node connects to the access point in Infrastructure mode.
Finally, connect the thin client to the access point. If there are several thin clients connected to a single hub, connect the access point to this hub.
We found ad hoc mode unsuitable for two reasons. First, most wireless devices limit ad hoc connection speeds to 11Mbps, which would put the network under severe strain to boot even one client. Second, while in ad hoc mode, the wireless nodes we were using would assume the Media Access Control (MAC) address of the computer that last accessed its Web interface (using Ethernet) as its own Wireless LAN MAC. This made the nodes suitable for connecting a single computer to a wireless network, but not for bridging traffic destined to more than one machine. This detail was found only after much sleuthing and led to a range of sporadic and often unreproducible errors in our system.
The wireless devices will form an Open Systems Interconnection (OSI) layer 2 bridge between the server and the thin clients. In other words, all packets received by the wireless devices on their Ethernet interfaces will be forwarded over the wireless network and retransmitted on the Ethernet adapter of the other wireless device. The bridge is transparent to both the clients and the server; neither has any knowledge that the bridge is in place.
For administration of the thin clients and network, we used the Webmin program. Webmin comprises a Web front end and a number of CGI scripts, which directly update system configuration files. As it is Web-based, administration can be performed from any part of the network by simply using a Web browser to log in to the server. The graphical interface greatly simplifies tasks, such as adding and removing thin clients from the network or changing the location of the image file to be transferred at boot time. The alternative is to edit several configuration files by hand and restart all dæmon programs manually.
- Nmap—Not Just for Evil!
- Resurrecting the Armadillo
- High-Availability Storage with HA-LVM
- March 2015 Issue of Linux Journal: System Administration
- Real-Time Rogue Wireless Access Point Detection with the Raspberry Pi
- DNSMasq, the Pint-Sized Super Dæmon!
- Localhost DNS Cache
- Days Between Dates: the Counting
- The Usability of GNOME
- Linux for Astronomers