Two Windows 2000 Servers with direct connection to Public Internet.
One server is DNS primary server with second W2K server (and 3 others elsewhere) are secondary servers.
There are a total of four domains, *one of which is an Active-Directory integrated domain*.
These two W2K servers are to be moved behind a firewall. The NICs that currently connect to the Public network will be moved to the new DMZ. In doing so their addresses will change from real public addresses to 10.x.x.x addresses.
My problem is NOT with the NAT that is necessary to forward DNS packets from outside to the DMZ NICs, but rather the contents of the Active Directory-Integrated DNS Zone that are updated by the servers themselves.
W2K insists on creating DNS A records for each IP address DNS is active on. That means I end up with a 10.x.x.x A record that corresponds to a NS name where I should have a public IP (that will be port-forwarded by the firewall to the DMZ IP address of the DNS server). If a remote client tries to connect to my DNS service using 10.x.x.x they will fail.
How do others do this? Were certainly not the first to move behind a firewall.
Do I throw another NIC in the box, configure it with a Public IP but NOT connect it to a network switch, activate DNS on that dummy NIC (so I get valid A records in DNS) and then &&.?? Perhaps combined with a generous dose of DnsAvoidRegisterRecords or something along those lines?!
Thanks in advance,