Debian DHCP on Shared Network
 

Hi,

New to Linux but done alot of research...have the following problem

Using debain 5.0.4 with webmin 1.500 and dhcp3.

i have two subnets (192.168.1.0 & 192.168.2.0)at home.
The debain dhcp server has 1 NIC using eth0 and eth0:1 (virtual).

I have configured the dhcpd.conf file using webmin, with a shared network, given the two subents are on the same physical network. each subnet has a pool of about 20 addresses.

Problem is that when a host boots up which is under the 192.168.2.0 subnet (using its MAC add - no fixed ip assignment), it gets the associated DNS and gateway ip addresses for this subnet, however, the ip address is assigned from the 192.168.1.0 address pool.

I have checked the net extensivly finding all sorts of discussions and solutions. My problem occurs in the same manner irrespective of whether i connect a host direct to the dhcp server or via a (netgear) switch!

Please can someone advise on what settings i might be missing, or do i need to set up DNS/DDNS and/or a firewall/router before this works properly?

thanks jat

I think because request to DHCP server came from interface eth0, which is assigned to 192.168.1.0 network.

Thats what i understood as well, which is why i'm wondering other people must have a similar configuration. So how can ip requests for hosts tagged to the 192.168.2.0 subnet (as configured on the dhcp server using MAC) be recognised appropriately given i'm using the alias eth0:1 on the eth0 (subnet 192.168.1.0)? Is some kind of local port forwarding required or something (complete NooB!)

You can't. You should *NEVER* *EVER* use an alias with a different subnet. That's what vlans and "buying more nics" are for. when eth0:1 does a dhcp request, the request looks *EXACTLY* the same as from eth0, so no way to provide a different subnet. Additionally the server could not respond to it right even if it did look different, as it would only be listening on known valid local subnets without an additionall relay.