Dear All

We have a cluster of Linux web servers on an otherwise MS network.
The MS windows servers provide IP addresses to the XP clients (bear
with me - this is a Linux Networking issue).

We want to use System Imager to push images out to the web servers
when they need rebuilding. To make this work we need a dhcp server
configured to not only provide specific IP addresses to specific
servers (based on MAC address) but also to then point them in the
direction of the Image Server to both boot from and then install from.

The dhcp server (using DHCPD under Centos 4.0) is installed and
configured and works when the dhcpd server and web servers are on
their own network. We'd now like to add this network to the main
network and here's the problem - the Windows XP clients musn't be
allowed to retrieve an IP address from this particular DHCP server.
My question is -

If I configure the dhcpd server as shown below (eg using the a range
that only includes the four ip addresses and the correct host stanzas)
will it

a) only accept requests from the four web servers (what I hope will
happen)?

or

b) does this config merely mean that should a server with the
particular mac address request an ip address then it will receive the
specified address but will also dish out the four ip addresses to any
other clients on a first come first served basis?

Any help would be much appreciated.

Thanks

paul

subnet 10.80.16.0 netmask 255.255.248.0 {
range 10.80.18.202 10.80.18.205;
}


host webserver1 {
hardware ethernet 00:06:5B:EE:93:C2;
fixed-address 10.80.18.204;
option host-name "webserver1";
}

host webserver2 {
hardware ethernet 00:06:5B:EE:91:E9;
fixed-address 10.80.18.205;
option host-name "webserver2";
}

..... etc