failover with 3.1.0, daemon quietly dies

[Robert Gehr]

Hello there

I have been running 3.0.4 up to 3.0.6 without a hitch but 3.1.0 just
refuses to work.
I have two servers running a failover configuration. Here is upgrade
scenario.

Stopped the secondary server, stopped the primary server, installed
3.1.0 on both servers and started the primary and then the secondary again.

The processes come up all right but after a short while the process on
the secondary just quietly dies. All the logfile on the secondary shows
is the following:

--------------------------snip------------------------

Sep 11 16:05:12 bafs2 dhcpd: Internet Systems Consortium DHCP Server V3.1.0
Sep 11 16:05:12 bafs2 dhcpd: Copyright 2004-2007 Internet Systems
Consortium.
Sep 11 16:05:12 bafs2 dhcpd: All rights reserved.
Sep 11 16:05:12 bafs2 dhcpd: For info, please visit
http://www.isc.org/sw/dhcp/
Sep 11 16:05:12 bafs2 dhcpd: Wrote 0 deleted host decls to leases file.
Sep 11 16:05:12 bafs2 dhcpd: Wrote 0 new dynamic host decls to leases file.
Sep 11 16:05:12 bafs2 dhcpd: Wrote 462 leases to leases file.
Sep 11 16:05:12 bafs2 dhcpd: Listening on
LPF/bond0/00:02:a5:e7:38:eb/10.230/16
Sep 11 16:05:12 bafs2 dhcpd: Sending on
LPF/bond0/00:02:a5:e7:38:eb/10.230/16
Sep 11 16:05:12 bafs2 dhcpd: Sending on   Socket/fallback/fallback-net
Sep 11 16:05:12 bafs2 dhcpd: failover peer bafs2: I move from
communications-interrupted to startup
Sep 11 16:05:12 bafs2 dhcpd: Failover CONNECT from unknown: unknown server

------------------------snap-----------------------------

On the primary the log shows:

--------------------------snip------------------------

Sep 11 16:04:56 baadm1 dhcpd: Internet Systems Consortium DHCP Server V3.1.0
Sep 11 16:04:56 baadm1 dhcpd: Copyright 2004-2007 Internet Systems
Consortium.
Sep 11 16:04:56 baadm1 dhcpd: All rights reserved.
Sep 11 16:04:56 baadm1 dhcpd: For info, please visit
http://www.isc.org/sw/dhcp/
Sep 11 16:04:56 baadm1 dhcpd: Wrote 0 deleted host decls to leases file.
Sep 11 16:04:56 baadm1 dhcpd: Wrote 0 new dynamic host decls to leases file.
Sep 11 16:04:56 baadm1 dhcpd: Wrote 489 leases to leases file.
Sep 11 16:04:56 baadm1 dhcpd: Listening on
LPF/eth0/00:09:6b:ff:0e:92/10.230/16
Sep 11 16:04:56 baadm1 dhcpd: Sending on
LPF/eth0/00:09:6b:ff:0e:92/10.230/16
Sep 11 16:04:56 baadm1 dhcpd: Sending on   Socket/fallback/fallback-net
Sep 11 16:04:56 baadm1 dhcpd: failover peer baadm1: I move from
communications-interrupted to startup
Sep 11 16:05:11 baadm1 dhcpd: failover peer baadm1: I move from startup
to communications-interrupted
Sep 11 16:05:26 baadm1 dhcpd: failover: link startup timeout

------------------------snap-----------------------------

The config files are unchanged here is what the look like.

primary:

failover peer "baadm1" {
    primary;
    address 10.230.1.202;
    port 647;
    peer address 10.230.1.2;
    peer port 648;
    max-response-delay 120;
    max-unacked-updates 10;
    mclt 1800;
    split 128;
    load balance max seconds 3;
}

secondary:

failover peer "bafs2" {
    secondary;
    address 10.230.1.2;
    port 648;
    peer address 10.230.1.202;
    peer port 647;
    max-response-delay 120;
    max-unacked-updates 10;
    load balance max seconds 3;
}

What might be of interest is that the secondary is running a nic bonding
setup but this doesn't seem to be problem with version 3.0.x.
I also traced the traffic and there is definetly traffic going to an fro
between the machines before the secondary dies.
When I fire 3.0.6 up on both machines everything is fine again.

Any help is greatly appreciated.

Best regards

Rob