[nsp] ppp address negotiation

[Adam Debus]

A few nights ago, we completed an upgrade to IOS 12.2(24) on our Cisco 7200
VXR NPE-400.

This router acts as the termination point for about 600 DSL customers
running PPPoA.

Now, when a customer signs on, if there is no static IP in the radius
response, it will assign the same address that the customer had before, even
if it is in use by another customer, or is not in the range from the "ip
local pool". I tried executing the command "no ppp ipcp accept-address" but
that seems to be the default, and has had no effect to how this is working.

This is what my Virtual-Template looks like:

interface Virtual-Template1
 ip unnumbered Loopback0
 ip route-cache flow
 no logging event link-status
 peer default ip address pool DSL
 ppp authentication pap
 ppp ipcp dns x.x.224.2 x.x.225.2
!
ip local pool DSL x.x.240.2 x.x.240.254
ip local pool DSL x.x.241.129 x.x.241.254

To give a real world example, a DSL has a static IP of x.x.242.1, after
removing that static IP from radius and performing a "clear int vi" on the
virtual-access interface, it logs out, logs back in, and receives the
address x.x.242.1 instead of one of the addresses from the local pool. We
also had an instance yesterday (which caused me to start looking at this)
where two customers were active with the same ip address. It wasn't even
checking to see if it was in use before assigning it.

We have checked modem configurations, and none of the modems are set to have
any address statically programmed, and rebooting the modem doesn't seem to
fix the issue either.

Does anyone have any thoughts or ideas as to what might be causing this and
how to fix it?

---
Adam Debus
Network Engineer, ReachONE Internet
adam at reachone.com