>Any one seen this? Or have suggestions? Cisco was not very helpful last
tim
>e I called ("Just make two subnets - that IP problem is your problem
not Cl
>ean Access's") so I'm hesitant to open another TAC and spend my week on
the
>phone.

This probably won't be much help, but I would try this, especially if I
were using Microsoft DHCP. We had
a similar issue on our network and this what we did to resolve it. Other
than that, Like T-Rex in Dinosaur comics, I promise nothing.

When you see the phenomena, check DNS and see if the address is still
attached to some other PC, in both 
The forward and reverse zones. If not, immediately turn off or unplug
the Ethernet connection of the black holed computer.
Go to another machine and ping the affected address to further check if
the address has been double issued. Finally check the arp cache on
Your firewall/gateway for the problem machine and see what the mac
address for the IP  matches up to. If you have Cisco switches, you can
List mac-address-table and find what port the desired mac address is
attached to and hence its location. Not sure how to find it on
Other brands. Alternatively, if you needed an excuse to physically visit
every IP address on campus one by one, this is it :).
Once you find the  machine/device that has the same IP you may be able
to resolve why it has it, or why it hasn't let go, or why it isn't 
Properly in DNS or whatever. 

If you are using Mickysoft DHCP check in DNS (yes DNS) to see when you
scavenge records and how 
Stuff is refreshed, getting the scavenge value wrong causes bad DNS
references to hang around. The downside here, is I cannot recall for
sure 
How it should be. I THINK,(and I really stress I am unsure here so check
with somebody competent) that the scavenge period must be longer than
the refresh and expiration periods. 



Daniel Sichel, CCNP, MCSE,MCSA,MCTS (Windows 2008)
Network Engineer
Ponderosa Telephone (559) 868-6367