Config problems, "error some other host already uses"

They both seem to be set correctly. But it is difficult to tell. This is why. ifconfig -a doesn't show the IP addresses at all. Also when the computer starts it displays one address 10.100.1.14 for eth0 and another address 64.72.*.* (The stars are real numbers) for the eth1 address. I am getting the error message for both of them now.

Also, the GUI that comes with Redhat shows the proper ip addresses. It just can't save it properly because of the error.

Let me give some more background.

The server is behind a physical firewall with one-to-one NAT set up. (The ip address for eth0 was 10.100.1.14 with the public ipaddress assigned to it.)

eth1 had an ip address 10.100.1.11 but nothing useful was mapped to it. The firewall has a WAN, LAN, and DMZ ports. I was assigned to move the email services from one server to this server so I wanted to start using eth1. I also wanted to hook up one connection to the DMZ.

This is "briefly" what happened.
I was using an ssh client connected to the eth0 address. I typed i"fconfig eth1 <public ip address>" and hit enter by accident when I was trying to look at the man pages for the syntax and other options. At that point my ssh connection was disconnect and I lost all network access to the server.

At a shell prompt I tried "service network restart". The network cards will be shutdown "OK" but then fail to restart. The error again is.  "Error some other host already uses this IP address" and it will state the respective IP address. It would also say "2000-2002 Netfilter..." or something like that, which I haven't seen before.

When I type "ifconfig -a" the three adapters lo, eth0, and eth1 will display but the eth* does not show any IP addresses.

Here is a funny thing that happened. I restarted the computer. On the screen it showed the "login:" prompt. I disconnected the cable from the DMZ port and plugged it into the switch. On the screen it said "e100: eth0 NIC Link is down" (I don't remember exactly) but then on the next line it said "e100: eth0 NIC Link is up 100 Mbps Full duplex."  What is that all about? Is that normal for something like that to display when I haven't logged in and the symptoms above still exist?

After a few hours of using the GUI, "netconfig", and interfaces I was getting no where so I decided to not use the DMZ and assign the cards ip address of 10.100.1.15 and 10.100.1.16. I then changed, or my coworker, then changed the firewall settings to map the public IP address to these two addresses.

The server is working fine with the NAT set up. But I would still like to know I had this problem and if this problem is going to come up next time I try to restart my network service. I would still like to get the DMZ set up but am scared to change the settings and restart the network just to get the same error.

I'll change the question a little.
Could smbd and nmbd configuration make the network service think the addresses are already taken?

What are the steps necessary to setup one card on the LAN and one card on the DMZ?  

I still don't know what setting was being kept in memory or why it happened. I don't what to lose 300 points withoug a definite solution. Oh, well, have 150 points each for your suggestions.

Thanks. I wish I could have helped you more.

Paul