WLC 2106 DHCP Relay - IP Conflict

[pdisse]

I have a Cisco 2106 WLC in DHCP Relay mode. For some reason, all of the clients that attach to the network via WiFi get an IP Conflict notification, even though the IP address is not in conflict (they're all unique, and the next available on the DHCP server). I can't figure out if the problem is related to my Windows 2008 DHCP server, or the WLC. They are all using the same subnet (192.168.48.xxx). I'm not even sure how to begin tracing the problem.

I'm sure I will need to provide more info, which I'm willing to do. I'm just not sure where to start. I really appreciate any help given. We are a small non-profit, and it's great to have the forums available as a resource.

[George Stefanick · Accepted solution]

Hello,

When i have issues like this myself. I would extended the vlan to a wired port. Plug in your laptop into the switch and see what happens.

So the client is saying there is a conflict ? What type of NIC and are they ALL doing it ?

[pdisse]

I'll try plugging into the switch.

ALL clients on the Wifi get an IP conflict, yet have a unique address. It doesn't matter what IP they get. It's always a different IP, always an address that is available if I look on the DHCP server. It's very strange. It's pulling addresses from the Windows 2008 DHCP server, and it's pulling the next free address in line, but all of the clients on the Wifi still get an IP Conflict message. My allowed range is 192.168.48.50 - 200, and it always pulls the next free address inside of that range, and it does this properly(x.x.x.50 is taken by another client, DHCP relay will assign x.x.x.51 yet still get a conflict). All Wifi clients just keep getting the message that there is an IP conflict.

It's very strange, because it IS working properly. The users can all use the network normally, share files/printers, surf the internet, even join the AD domain as a new PC over Wifi, etc. They simply keep getting a message that there is an IP conflict. If you disconnect any of the clients, and ping the address they just had, you get no response. There's nothing else sharing the client's IP. Why the conflict?

[Tiago Antunes]

Hi,

When you say "They are all using the same subnet (192.168.48.xxx).", do you mean that the WLC and windows server have an IP address on this subnet also?

If yes, please make sure that you are excluding from the DHCP pool any static address you have configured, like the WLC interface, the windows server ip address.

HTH,

Tiago

-----

- if this helped youe please rate it so others can check this discussion -

[pdisse]

Yep, the WLC and the Windows DHCP server are all on 192.168.48.xxx.

The WLC is beneath the range the server starts assigning. The DHCP server starts assigning addresses at 192.168.48.50. The WLC has 19 and 20.

[George Stefanick]

So the wireless icon is stating conflict ? What is alarming you of the issue?

[pdisse]

See attached Error.jpg. Any connecting client gets a notification like this, whether it's a Windows 7, Windows Vista, or Windows XP machine. It happens on every client that connects. There are often three or more users connected, and I tell them to simply ignore the problem, as for all intents and purposes, everything is working fine, they are just presented with this error. The network and services all function as they should, which I would imagine they would not if there were truly an IP conflict.

Clicking to dismiss the notification will cause to come back every 30 seconds or so until you disconnect.

[George Stefanick]

Are you using DHCP override on the WLAN?

WLAN->SELECT WLAN->ADVANCE TAB... DHCP Server Override enabled?

[pdisse]

So, the issue has been solved. Looking in Event Manager for any of the clients showed that, no matter what IP the client PC had, there was a single MAC address that was seen to be in conflict. Once client would be 192.168.48.75, another was 192.168.48.110, and yet another at 192.168.48.148, but they ALL listed to be in conflict (in Event Manager) with the same MAC address.

I went into DHCP and found the machine. Apparently, the previous IT guy had put an additional PCI NIC in one of the desktops, and plugged one into a 100Mb switch and another into a 1000Mb switch, then teamed them with the Broadcom utility. In DHCP, despite there being a network team, both cards were pulling an IP address for this desktop through DHCP.

I unplugged and disabled one of the NICs and somehow the problem went away. I have two of the clients connected and still no error message. It seems that the whole time I had correctly configured the 2106. In addition, the person who sits at this particular desktop has remarked how much faster his machine is running.

Is anyone able to explain how this is possible? Shouldn't ALL of the network clients, not just the ones going through the WiFi's DHCP relay have gotten this message? Why would they have gotten this message at all, simply from a misconfigured desktop? Why would a misconfigured desktop cause only the clients connected to the 2106 to get the error, yet all of the other (over 100) network clients to be fine. Very, very strange.

[George Stefanick]

Paul,

I always say, a new day will bring new challenges. This is a good one and excellent troubleshooting on your part, good job. I would say we may have to sniff the wired into the controller to see why. Perhaps one of our Cisco friends here can provide some insight ...