The first step in configuring a DHCP server is to install one of the DHCP servers from the ports tree. The one I recommend and use is isc-dhcp3, which is in /usr/ports/net/isc-dhcp3.

Once that port is installed, visit /usr/local/etc, because you're going to make a configuration file named dhcpd.conf in that directory. Below is a skeleton configuration file.

Initial DHCP Configuration

The syntax of the dhcpd.conf file is very familiar if you are a C/C++ programmer. Sections are broken up by braces, and lines are terminated by semi-colons.

shared-network DHCP-NET {
	option domain-name "example.net";
	option netbios-name-servers "foo.example.net";
	option nis-domain "example.net";
	use-host-decl-names on;
	default-lease-time 600;
	max-lease-time 3600;
	subnet 192.168.0.0 netmask 255.255.255.0 {
		option broadcast-address 192.168.0.255;
		option domain-name-servers 192.168.0.1;
		option subnet-mask 255.255.255.0;
		option routers 192.168.0.1;
		range 192.168.0.32 192.168.0.254;
	}
}

This defines the necessary options for the 192.168.0.0 network with a netmask of 255.255.255.0 to allow dynamic configuration of hosts on the network.

Here's a quick rundown of the options present:

option domain-name

This is the domain name that the machines on the network will be part of.

 

option netbios-name-servers

This option describes the NetBIOS Browse Master for the network. This is only useful if there are Windows 9X/NT/2000 machines on the network.

 

option nis-domain

This is the domain name for the NIS master server. This is only useful if NIS is used on the network.

 

use-host-decl-names

This, when set to ON, allows static hosts to acquire their hostname from the DHCP server given their static entry (more on static entries below).

 

default-lease-time

This is the time, in seconds, until a client's DHCP lease can expire. I personally set this to a lower value so any changes I make will take effect faster. Keep in mind that doing so increases network traffic.

 

max-lease-time

This is the maximum lease time a system can "hold" a DHCP address.

 

subnet IP netmask MASK

This defines a subnet that the server is going to be providing DHCP services for. It takes IP, which is the IP address of the network, and MASK, which is the netmask for the network.

The following options sit inside a subnet definition:

option broadcast-address

This is the broadcast IP address for the subnet.

 

option domain-name-servers

This is the IP address of the DNS server(s) for the network. If you have more than one, then you can separate them with commas.

 

option subnet-mask

This is the netmask for the network in question.

 

option routers

The default gateway for the network goes here.

 

range START END

This defines the range of IP addresses that can be used for the dynamic configuration pool. Every IP address in this range will be given dynamically to hosts that do not have a static entry.

Static Host Configuration

You can define hosts that get configured by DHCP, but will always get the same IP address. To do this, add an entry for every host in the shared-network section, like so:

shared-network DHCP-NET {
	option domain-name "example.net";
	option netbios-name-servers "foo.example.net";
	option nis-domain "example-net";
	use-host-decl-names on;
	default-lease-time 600;
	max-lease-time 3600;
	subnet 192.168.0.0 netmask 255.255.255.0 {
		option broadcast-address 192.168.0.255;
		option domain-name-servers 192.168.0.1;
		option subnet-mask 255.255.255.0;
		option routers 192.168.0.1;
		range 192.168.0.32 192.168.0.254;
	}
	host foobar {
		option host-name "foobar.example.net";
		hardware ethernet 00:a0:d2:18:84:49;
		fixed-address foobar.example.net;
	}

This will create a static entry for host foobar.example.net. Note that foobar.example.net needs to exist in the DNS server for this to work.

Here are the options available:

option host-name

The name of the client, as listed in the DNS records for the network.

 

hardware ethernet

The hardware's MAC address. This is used by DHCP to determine the host before it assigns it an IP address.

 

fixed-address

This is the address or hostname that the given host will receive when it comes up and asks for a DHCP lease.

Once you set up the config file, you can start dhcpd from a script in /usr/local/etc/rc.d. Here's the startup script I use:

#!/bin/sh
case "$1" in
  start)
    if [ -x /usr/local/sbin/dhcpd ] && \
      [ -f /usr/local/etc/dhcpd.conf ]
      then echo -n " dhcpd"
      /usr/local/sbin/dhcpd -q -cf /usr/local/etc/dhcpd.conf fxp0
      exit 0
      fi
    exit 1
    ;;
  stop)
    ;;
  *)
    echo "usage: `basename $0` (start|stop)" >&2
    exit 64
    ;;

Be sure to replace the interface (fxp0) with the network interface for your system.

You can test your configuration at this point by running:

/usr/local/sbin/dhcpd -cf /usr/local/etc/dhcpd.conf interface

Replace interface with the network interface that dhcpd will be listening on.

Also, if you get any configuration errors, be sure that your semi-colons and braces are all there because dhcpd is really picky about them.

- Chris

Return to the February 2001 Issue