Just searching through the site, and I have a similar problem. We have configured a bunch of wireless APs in our municipal transit garage. The transit busses have been fitted with 802.11G modules (Lantronix) and are connecting to the APs to download schedule info, etc. It is using WPA2 AES PSK, and it is working fine from a functionality POV.

The issue is that the people in transit responsible for this system are running an Expect-style script to collect stats from the APs on a scheduled basis, and ocassionally a show dot11 ass detail will show MACs of busses (with a 0.0.0.0 address) that have left the shop and been on the road for hours.

Just wondering why they would not be timed out of the association table, and if there are any settings I might be able to tweak to fix this?

The APs are a mix of 1232 and 1242, running a number of different IOS versions. I'm not even sure this is a Cisco issue, just looking for ideas.

Any thought are appreciated. TIA

look in the event viewer on the windows pc and see if there are any DHCP related logs during the wireless auth period. It looks like those are all intel nics there is a debugging utility built into the intel supplicant, try turning it on while loggin in and see if by any chance the pc is getting a 0.0.0.0 address prior to getting its routable address.

its probaly not related but I noticed the other day while looking at the cisco vpn interface in windows under "network connections" that the ip address configured is 0.0.0.0, see if the systems that are showing up in the log all have something in common like a VPN client and in the system even viewer, compare it with other systems that do not show up as 0.0.0.0 and work fine.