From Debian Clusters

This project will walk through installing and configuring all the software necessary for a cluster. It starts with the assumption that you already have the hardware that you're going to be using, and that you have already installed Debian – preferably the testing version used at the time of this writing, Lenny – on the machines you'll be using. (If you want some background on this, see the Basic Linux Skills page for pointers to resources.)

The information on the website is laid out in approximately the same order I followed in configuring my own cluster. I'll make reference to my cluster throughout the project and use examples from it (more information in Network Topology). Not all the information may be applicable to you, or you may need to tweak it to fit your needs.

The first steps in setting up a cluster are to configure the firewall, services, head node, and worker node image. In this particular example, I'll first be building the firewall and services on the same machine, then working on another machine to turn into the head node plus file services. Then I'll be working on the worker node image, and then sending that image out to the rest of the worker nodes.

Before getting started, there a few things to be decided on:

• First and foremost, the IP range and subnet mask for the internal network. (192.168/16, 172.16/12, and 10/8 are set aside for private networks and are non-routable. These are very good choices.)
• Second, the hostnames for the systems will come into play very early on. It's a good idea to have a plan about these.
• Third, the IPs for the head node, firewall, and any services machines also need to be known early on. (See the first point.)

For more information, continue on to Network Topology.