failover with 3.1.0, daemon quietly dies

David W. Hankins David_Hankins at
Wed Sep 12 15:59:07 UTC 2007

On Wed, Sep 12, 2007 at 10:28:39AM +0200, Robert Gehr wrote:
> 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

> Sep 11 16:05:11 baadm1 dhcpd: failover peer baadm1: I move from startup
> to communications-interrupted

'bafs2' and 'baadm1' do not match.  I think the 'CONNECT from unknown'
log message could be improved - we get the remote system's peer name,
we can log that it doesn't match any configured name.


In 3.1.0, we synched the failover implementation up to draft level
12 (coming short in a few details).  One of the details we complied
with was that failover relationships be identified by name - not by
peer IP address as was done in 3.0.x.

This means that two systems with the same peer IP addresses can have
multiple failover relationships (and then multiple TCP sockets) if
you choose.  It also means that we'll be able to add failover
authentication soon (the peer name is a sort of 'username' half of
the authentication).

But it also means that the failover peer 'name' fields have to match
on both sides (which they were supposed to all along).

Shutdown one server.  Edit dhcpd.conf to the remote system's peer
name.  Edit dhcpd.leases and search for 'failover peer "name" state',
and change it as well.

Ash bugud-gul durbatuluk agh burzum-ishi krimpatul.
Why settle for the lesser evil?
David W. Hankins	"If you don't do it right the first time,
Software Engineer		     you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins

More information about the dhcp-users mailing list