Previous: ?  Next: ?

In all of these documents, it is assumed that the master cluster control server (named "master")
is configured with its primary ethernet interface set to 192.168.2.1

The 192.168.0.0/255.255.0.0 address space is reserved for allocation by private internets.
See RFC1918.

We will use this for the cluster to greatly simply just about everything and keep us out
of the administrators' hair as well.

For instance, we can do proper domain name <TODO L>reverse mapping<TODO> using bind, which
will make connections more secure as well as faster to and among cluster nodes.
This also lets us run rsh wide open (using hosts.equiv) so mpi and pvm can make quick
connections, and the outside festering internet community of script kiddies is
kept reasonably at bay.  For one they are unable to route their packets our way at all, unless
they already have comprised a system/machine on the physical network we are on.  Not too likely, and
even if they get a macos machine, we will most likely reboot it into linux taking it off of the routable internet.

This requires a bit of <TODO L>trickery in /etc/dhclient.conf</TODO> on the slave nodes, so that they will only
use OUR DCHP server, and not the pre-existing DHCP server for routable addresses which MacOS will use when it boots.