Odd lease exclusion in 4.1.1

Friesen, Don SSBC:EX Don.Friesen at gov.bc.ca
Thu May 27 18:00:33 UTC 2010


   I'm trying to figure out why 2 addresses in our pool range have no
record in the lease file.  We run using failover, and it was my
impression that the dhcpd.leases file would have one entry for every
address in the pool.  It isn't an issue right now, as we have lots of
free space, but it is an itch I'd like to scratch.  My shared network
definition looks like (I've trimmed out the options):

shared-network "10_22_165_0" {
   subnet 10.22.166.0 netmask 255.255.255.0 {
      pool {
         failover peer "DHCP2";
            range 10.22.166.2 10.22.166.167;
            range 10.22.166.169 10.22.166.253;
      }
   }
}

   The leases file has 249 entries, when I would expect 251.  The two
missing addresses are 10.22.166.183 and 10.22.166.184.

   The one significant aspect for these two addresses are that they were
forced free a few days ago.  We run a Remedy front end that a large
distributed number of users can access and administer their own little
chunk of our DHCP space... including forcing a lease out when a machine
is retired.  The front end will flag the address and the script will
build lease records, shut down the daemon, dump the artificial records
into the lease file, and restart the daemon.

   The artificial records look like:

lease 10.22.166.183 {
  binding state abandoned;
  next binding state free;
}
lease 10.22.166.184 {
  binding state abandoned;
  next binding state free;
}

   The daemon has been restarted a few times since then, with no records
relating to those IP addresses in either peer's lease file, and yet the
daemons have not introduced them.

   Have I chosen a poor binding state?  This appears to have worked
before hundreds of times.


Don Friesen


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20100527/9b538dec/attachment.html>


More information about the dhcp-users mailing list