Can someone help me troubleshoot a Server 2008 DHCP issue?

zrockstar

Sorry, I don't know if this is the right place to put this, but I am trying to setup a Server 2008 virtual box in GNS3 as a DHCP server for some Cisco labs. I have a two router setup connected through serial, each with a different subnet off of the FA0/0 interface. The Win Server is off of R1 10.1.1.0 and across the WAN link the client is on R2 10.1.3.0. I know the DHCP server is working because clients on 10.1.1.0 can get addresses, but 10.1.3.0 clients can't. I have setup an ip-helper address to the DHCP server on the FA0/0 of the 10.1.3.0 network, and have setup a DHCP scope on the server for 10.1.3.0. I have Wiresharked the whole thing and have seen the DHCP request traverse the whole network and come out of R1 as a unicast for 10.1.1.253 (server IP) from 10.1.3.254 (FA0/0 on R2). I don't know a whole lot about Server 2008, but it appears that the packet is choking there. I see no increment in the DHCP statistics, so it looks like the server is dropping the packet before processing it. Any ideas, hidden tricks, or steps with Server that need to be taken to get this to work? I know this is probably difficult to troubleshoot over a paragraph on a forum, so props to anyone able to point me in the right direction.

emerald_octane

Depending on how you have server 2008 installed you may need to have promiscious mode turned on (if it's in a VM).

You've double checked your scope, made sure it's authorized and all that jazz.

Install network monitor on your win server ( Download: Microsoft Network Monitor 3.4 - Microsoft Download Center - Download Details ) and capture packets. Look for traffic from your router to the server for DHCP. If you see a DHCP response from packet cap and then an error of some sort directly afterwards that indicates a loop that needs to be eliminated because the dhcp is seeing responses from itself (or something like that).

Honestly though win server config is mindblowingly easy in my experience any DHCP issues I had were directly related to my router config, probably because my ciscofu is alot weaker than my windoze micro

ptilsen

Just as a sanity check, can you assign a secondary IP to the server and plug it directly into the L2 network that 10.1.3.0 sits on? That will tell you for sure whether the issue lies with the routers/IP helper or with the server config.

That being said, chances are it is not the server config. Unless you have not authorized or activated the 10.1.3.0 scope or have the wrong subnet mask, or something silly like that, it's pretty hard to misconfigure DHCP on Windows.

You could run wireshark on both server and client, and see the whole DORA process (or however far it gets) on both sides. That may help point you in the right direction. Otherwise, you're going to want to packet capture the routers themselves to see where the problem lies (between server and R1, R1 and R2, or R2 and client).

Zartanasaurus

Can the DHCP server ping the fa0/0 interface of R2?

zrockstar

Hey guys, well thank you for your feedback and trying to help out, but I feel like a total bonehead. Well, I guess I am 100% new to server (classes for it start next month), so it's to be expected, but since pings worked as suggested by Zarta and moving it to the .3.0 network failed, I knew it had to be something in the server. Turns out since I had recently installed an AD role on the server, and the DHCP server role had not been authorized on the AD and the .1.0 clients were still operating on valid leases. So note to the other server noobs out there...AUTHORIZE YOUR SERVER ROLES IN YOUR AD!!!! LOL, thanks again, problem solved.

Zartanasaurus

You tricked us! I was going to ask if the DHCP server was authorized, but you said the other subnet was getting addresses. The other "easy" thing to look for would be that you activated the new scope.