I have a general question about how IPs are allocated when using DHCP and mask requests.  It appears the IPs are allocated in a sequential order regardless of what kind of pool is set up.  IE a local pool, dhcp network pool, or even a dhcp host pool.

example config:

R1(server) <---> R2(client)

!R1 config

ip local pool r2 192.168.12.2

interface Serial1/0
 ip address 192.168.12.1 255.255.255.252
 encapsulation ppp
 peer default ip address pool r2
 ppp ipcp mask 255.255.255.252
 serial restart-delay 0

!R2 config

ip dhcp pool r2
 import all
 origin ipcp

interface Serial1/0
 ip address pool r2
 encapsulation ppp
 ppp ipcp mask request
 ppp ipcp route default
 serial restart-delay 0

!Here is the output after negotiation takes place:

R2(config-if)#do sh ip int br | i up
Serial1/0                  192.168.12.1    YES manual up                    up    

R1#sh ip local pool

 Pool                     Begin           End             Free  In use   Blocked
 r2                       192.168.12.2    192.168.12.2       0       1       0

The above shows R1 (server) allocates a single IP for the pool 192.168.12.2, and when showing the output for the pool it lists the address in use. However when R2 allocates an IP, it seems to do it based on the mask/network assigned instead of from a given pool and actually assigns the 1st IP available from the network and ignores the pool all together.  Above you can see R2(client) is assigned 192.168.12.1.  Is there a way to adjust the behavior of pool assignment?

The above uses a local pool, but I have tried this using DHCP pools on R1 also with the same result.  See attached for debug ppp negotiation.