Some guest clients stuck in DHCP_REQD state

[Inayat Bunglawala]

Hi everyone,

We are regularly having issues with some (random) wireless guest users not being able to connect to our guest SSID while others are able to connect just fine. They get stuck in the DHCP_REQD state so do not even make it to the web authentication screen.  We are running WLC 4404 code 7.0.98 and using the WLC (two of them) as internal DHCP servers with a /23 scope divided between the two of them. I have checked the lease allocations and there are plenty of ip addresses spare.

A couple of days ago, all the users connecting via APs associated via one of the WLCs were unable to get an ip address and were stuck in the DHCP_REQD state all day and we ended up rebooting the controller which seemed to fix the problem temporarily but today some users have again been unable to connect. The users connected to the second controller all apppeared to be working just fine. The config on both WLCs appears to be identical.

I have seen some previous discussion about this being a known Cisco bug and which have recommended a code upgrade to v7.0.116 and v7.0.220 and I just really wanted to know your thoughts about whether a code upgrade is really necessary or a misconfiguration is the more likely issue.

Just to complicate matters we are in the process of tunnelling all our guest users to anchor controllers in the DMZ but at present this issue seems to affect both users who are connecting to local WLCs and those that are being tunnelled to the DMZ controllers.

Fianlly, I just want to add that we are also looking into installing external DHCP servers in the DMZ to cater to the guest users instead of using the internal DHCP server on the WLCs. I just wanted to know of your own experiences ie whether a code upgrade will resolve the issue for now.

[Inayat Bunglawala]

Just to add that we are also looking into installing external DHCP servers in the DMZ to cater to the guest users instead of using the internal DHCP server on the WLCs. I just wanted to know of your own experiences ie whether a code upgrade will resolve the issue for now.

[Viten Patel]

Inayat, the only way to be 100% sure that you are hitting that internal DHCP bug is to use a external dhcp server ...but reading the description I would say 99.9% you are hitting that bug and a upgrade would fix it..

[Scott Fella]

Viten,

What is the bug id for this. I remember this being an issue and just wanted I see what is affected.

Thanks

Sent from my iPhone

[Viten Patel]

http://tools.cisco.com/Support/BugToolKit/search/getBugDetails.do?method=fetchBugDetails&bugId=CSCth68708

[Inayat Bunglawala]

Thanks, Viten. We are looking at upgrading the WLC code to v7.0.220 asap. Meanwhile, the DHCP_REQD issue occurred again with all users connecting to one of our WLCs and I ended up disabling/re-enabling the guest SSID on that WLC a couple of times and that seemed to do the trick for now. Longer term, I think we may need to move to external DHCP servers in the DMZ.

[Inayat Bunglawala]

If anyone reads this thread in the future I would just like to say that I upgraded the WLC code to v7.0.220.0 ay the end of Jan 2012 and the DHCP_REQD issue disappeared and has not reappeared since. So, it looks like it was a bug in the previous code version.