recovery without communication problem

Serveur-Faucon Surveillance SrvFaucon at
Fri Dec 1 21:55:03 UTC 2006

This problem is now resolved, but I want to do some phylosophy if it should be normal.
One of the dhcp server had to be reinstall, so we did. When putting back the dhcpd.conf and dhcpd.failover (or dhcpd.master in the documentation), I accidentally put different ports in the dhcpd.failover file. So instead of port 647, the correct one, it was port 519, the bad one.
(IPs where change to protect the innocent)
---- (dhcp2 server)
failover peer "redondance01" {
         port 519;
         peer address;
         peer port 519;
         max-response-delay 20;
         max-unacked-updates 10;
         mclt 1800;
         split 128;
         load balance max seconds 3;
So everything look fine on the startup of the dhcp2 server. But, eventually, I looked the logs with "tail -f /var/log/messages"
=On the first server (good port)Dec  1 15:56:17 dhcp1 dhcpd: DHCPDISCOVER from 00:06:29:29:aa:aa via peer holds all free leases
=On the second server (the new one with the bad port)Dec  1 15:56:17 dhcp2 dhcpd: DHCPDISCOVER from 00:06:29:29:aa:aa via not responding (recovering)
=In the logs...=On the first server (good port)
Dec  1 14:39:10 dhcp2 dhcpd: failover peer redondance01: I move from recover to s
Dec  1 14:39:23 dhcpd2 dhcpd: failover peer redondance01: I move from startup to r

Those server are with ISC DHCP Server V3.0.4

So it started alright, but always stay in recover mode. Even if the other server would not respond. Is this normal behavour?
Alexandre Racine - Gardien Virtuel - Sécurité Informatique
Montréal, Québec, Canada

More information about the dhcp-users mailing list