SunnyV
Contributor
Contributor

No DHCP on one NIC with multiple NIC desktop

Hi, hope that you can have a look at this issue!

I have built Windows XP Professional SP3 Desktop VM's and they are deployed as linked clones using NetApp.

The desktop has 2 NIC's. One is connected to an internal network (DVs) on the Cluster and the desktop get's DHCP from an internal Server on an internal network. The other is connected to an internal network (DVs), then connected physically to an external network, which has a DHCP server. See attached Doc from a Diagram.

When the desktop was deployed, it received an IP address and joined the Domain on the External network. That part worked fine. After some time, the desktop stopped receiving DHCP requests and now can not receive an IP address, even by forcing it (IPCONFIG /RENEW).

To fix this issue, I had to...

1. Shutdown the VM.

2. Remove the NIC that could not receive an IP Address from DHCP.

3. Add a new NIC and connect it to the external NIC again.

4. Restart the VM.

There is nothing in the OS log, so can not see why a working Virtual NIC would do this. We are using the VMXNET3. I did try to un-install the drive in the OS and restart to pick it up. This worked on one of the desktops, but quite a few others it did not.

I did not want to log a call on this, as we have a work around, but not really a real reason for why this is happening.

If anyone has seen this and can help, you help would be appreciated.

Thanks!

Sunny

0 Kudos
4 Replies
JasonV1
Enthusiast
Enthusiast

Have you tried running Wireshark within the VM to see what traffic occurs during the DHCP session?

The Wireshark Wiki goes into a little detail about DHCP traffic and has some sample captures from (I assume; I've never looked at them) a successful request/lease: http://wiki.wireshark.org/DHCP

Was my post helpful? If so please remember to award points by using the Helpful/Correct buttons. Twitter: http://twitter.com/jasonventresco Blog: http://vjason.com
0 Kudos
JasonV1
Enthusiast
Enthusiast

Also, I would change the MAC address on the NIC rather than removing/readding it to see if that "fixes" the problem. If changing the MAC address enables DHCP to function again the problem may be with the switches the VM hosts are plugged into (just a guess, maybe problems with ARP?).

I'm not sure if changing the MAC address within the OS is any different than changing it within the vSphere GUI; I would assume that they do the same thing.

Was my post helpful? If so please remember to award points by using the Helpful/Correct buttons. Twitter: http://twitter.com/jasonventresco Blog: http://vjason.com
0 Kudos
SunnyV
Contributor
Contributor

Thanks for your help!

I'll see if MAC-Address change helps. There's definitely something wrong with either the switch or the VNIC's, as it was working before. Also when I first provisioned it, some of the desktops had the same issue. The VM's that have just recently had the issue, were not the ones that originally had the issue. It's just seems so random, unfortunately!

0 Kudos
JasonV1
Enthusiast
Enthusiast

No problem.

I am far from a networking expert but the fact that no one else responded means that no one has (probably) seen this, or just that it is that rare. That makes me want to think the problem is elsewhere.

Also, if you have multiple vCenter servers make sure that they have unique ID's: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=102402... . The ID's are used (in part) to ensure unique MAC addresses are selected and if you have two identical ID's in an environment there is a *remote* possibility that you have duplicates. This is probably not the issue BUT it is one of those obscure things that has bitten people before.

Was my post helpful? If so please remember to award points by using the Helpful/Correct buttons. Twitter: http://twitter.com/jasonventresco Blog: http://vjason.com
0 Kudos