dhcpd in failover not reusing expired leases

Robert Blayzor rblayzor.bulk at inoc.net
Tue Oct 14 09:25:58 UTC 2008


I'm noticing a strange behavior of isc-dhcpd (3.0.5) setup in  
failover.  Everything appears to work fine, but after some time both  
servers start saying they're out of free leases.  If I check the  
actual DHCP database I see tons of IP's that are currently not in use  
in the following state:


On the primary:

lease 64.x.x.x {
   starts 1 2008/10/13 14:16:00;
   ends 1 2008/10/13 14:26:00;
   tstp 1 2008/10/13 16:21:00;
   tsfp 4 2008/10/09 23:40:16;
   cltt 1 2008/10/13 14:16:00;
   binding state expired;
   next binding state free;
   hardware ethernet 00:1e:90:31:ed:e4;
   uid "\001\000\036\2201\355\344";
   option agent.remote-id 1:1:0:0:40:16:24:1:10:c:0:8a;
   client-hostname "YOUR-C30BE43EA5";
}

on the secondary:

lease 64.x.x.x {
   starts 4 2008/10/09 20:40:16;
   ends 4 2008/10/09 22:40:16;
   tstp 4 2008/10/09 22:40:16;
   tsfp 4 2008/10/09 23:40:16;
   atsfp 4 2008/10/09 23:40:16;
   cltt 4 2008/10/09 20:40:16;
   binding state free;
   hardware ethernet 00:40:ca:a2:96:ba;
   uid "\001\000@\312\242\226\272";
}




These IP's are currently not active and should be reused?  At least I  
think they would, but in out DHCP logs we see:

Oct 14 09:23:11 dhcp0 dhcpd: DHCPDISCOVER from 00:1c:df:22:37:60 via  
64.x.x.x: network 64.x.x/24: no free leases
Oct 14 09:23:11 dhco1 dhcpd: DHCPDISCOVER from 00:1c:df:22:37:60 via  
64.x.x.x: peer holds all free leases


If we restart one or both of the servers, it seems things balance out,  
at least for a while, but then we just end up back at this point in a  
few days.

any ideas?

-- 
Robert Blayzor, BOFH
INOC, LLC
rblayzor at inoc.net
http://www.inoc.net/~rblayzor/





More information about the dhcp-users mailing list