I have seen a number of people doing this recently and being confused by the same thing..
The confusing part in this situation is what you call the "modem". These are usually modems, routers, dhcp/dns/firewall/etc as well. What this does, is effectively create a whole subnet (possibly multiple) that exists for no reason.
My preferred method, is to run the modem in "bridged" mode, assuming the modem supports it.
Bridging the modem will essentially make the modem simple maintain the hardware connection, and pass it on to the gateway machine, (i use gateway, since it serves many roles aside from router or firwall).
The gateway machine then creates a pppoe connection and authenticates it, with the ISP. Usually, this will create a new network interface "ppp0", or something similar, which uses the physical interface (eth0 in my case).
I find this to be the simplest way to set up a linux router/NAT/firewall/etc type machine.
Quote:
Finally I want to make sure I have DHCP server setup properly before I make the switch (I guess I could just make my computers static if it doesn't work) What can I do to make sure it is setup and running. In Webmin I see this: http://images.tuprox.com/dhcp.jpg
|
According to the screenshot, there is no subnets defined, which would indicate that no, DHCP is not set up properly. Also, setting static IP's on anything other than a DHCP server itself, is a bad idea (IMHO), it is better to assign static DHCP leases to hosts you want to guarantee will get a persistent address.
The second screenshot means very little to me, what are you trying to use bind for? A caching nameserver? or to host your own domains?
You may find this link beneficial.
http://www.linuxhomenetworking.com/