Ok, we have a straight NT network here, exchange, tcp/ip, 75 clients all running NT4 Workstation SP6. We’ve got RAS enabled on a machine, (known as BLC022) so our managers can dial up and synchronize their exchange inboxes with their laptops (win2K)… Last week, I set up a test environment, to test an ERP package upgrade. We set up a pdc and three clients, but had the 10mbps switch connected to the patch panel as we needed to drop 100gb dbs on there from the lan. Unfortunately, I forgot that this would created a big collision domain. I had set up dhcp on the test pdc, to assign addresses from 192.168.5.1 to 192.168.5.5 for the three clients. As you all may foresee, as soon as we plugged the switch into the patch panel, it issued IP’s tothe workstations on the network (subnet 192.168.0.1 – 192.168.0.254). BLC022 (the RAS machine) grabbed the ip 192.168.5.2.. and, screwed up all the dial up connections. I disabled dhcp on the test pdc, and in the end, just chucked them all on the same subnet just to make things easy for us (why didn’t I do this earlier?). I flushed the 192.168.5.x ip from the RAS machine by disabling & re-enabling the PPP adapters, but, now, everytime anyone dials up to the RAS, it assigns them a 192.168.5.x IP. Where is this coming from? Ipconfig on BLC022 (RAS machine) shows 192.168.0.104, after countless /renew /release. I can’t put the 192.168.5.x range in the excluded ip’s on the current dhcp because it’s not within the scope. But whenever someone dials up, and you run ipconfig /all, it ALWAYS comes up with 192.168.5.x. Even more confusing is that we tried to assign statics to the laptops in the 192.168.0.x range, but that still doesn’t work. You can’t ping the exchange server when you dial up. You can establish a connection, but nothing else.
Anybody have any ideas whatsoever? We’re going to reload win2k on the laptops if we can’t think of anything else.