Comment 7 for bug 1314692

Here's a snippet from the leases:

lease 10.20.30.59 {
  starts 3 2014/04/30 15:41:39;
  ends 4 2014/05/01 03:41:39;
  cltt 3 2014/04/30 15:41:39;
  binding state active;
  next binding state free;
  rewind binding state free;
  hardware ethernet 00:1e:67:31:3b:eb;
  client-hostname "decathlete2";
}
lease 10.20.30.59 {
  starts 3 2014/04/30 15:41:43;
  ends 4 2014/05/01 03:41:43;
  cltt 3 2014/04/30 15:41:43;
  binding state active;
  next binding state free;
  rewind binding state free;
  hardware ethernet 00:1e:67:31:3b:eb;
  client-hostname "decathlete2";
}
lease 10.20.30.59 {
  starts 3 2014/04/30 15:41:50;
  ends 4 2014/05/01 03:41:50;
  cltt 3 2014/04/30 15:41:50;
  binding state active;
  next binding state free;
  rewind binding state free;
  hardware ethernet 00:1e:67:31:3b:eb;
  client-hostname "decathlete2";
}
lease 10.20.30.59 {
  starts 3 2014/04/30 15:42:00;
  ends 4 2014/05/01 03:42:00;
  cltt 3 2014/04/30 15:42:00;
  binding state active;
  next binding state free;
  rewind binding state free;
  hardware ethernet 00:1e:67:31:3b:eb;
  client-hostname "decathlete2";
}

Notice the lease times are overlapping.

Surely this is a bug in DHCPD? Why on earth is it writing a new lease out when one already exists? I guess it's just "renewing" it but perhaps we just need to make it more aggressively clean up old leases. (although having just read further it will only clean once per hour and that's a compiled-in setting)