Issue with failover

Chris Adams cma at cmadams.net
Tue Sep 22 16:03:39 UTC 2015


Once upon a time, Chris Adams <cma at cmadams.net> said:
> I set up a couple of new servers running CentOS 7 with dhcp (CentOS
> package version 4.2.5-36.el7).  The servers are set up with failover
> enabled.  I'm seeing the message:
> 
> Sep 22 09:03:14 rad2 dhcpd: bind update on 173.233.178.172 from nu-z rejected: incoming update is less critical than outgoing update
> 
> for just about every update.  I see this referenced before in the list
> archives, but it seems to be blamed on configuration mistakes (but I
> don't see where anybody said what mistakes led to it).

Looking at the logs some more, I realized that the bigger problem may be
that both servers appear to be replying to the requests (with the same
IP at least):

Sep 22 09:03:14 rad1 dhcpd: DHCPREQUEST for 173.233.178.172 from f8:d0:ac:63:70:bd via 10.114.0.19
Sep 22 09:03:14 rad1 dhcpd: DHCPACK on 173.233.178.172 to f8:d0:ac:63:70:bd via 10.114.0.19
Sep 22 09:03:14 rad2 dhcpd: DHCPREQUEST for 173.233.178.172 from f8:d0:ac:63:70:bd via 10.114.0.19
Sep 22 09:03:14 rad2 dhcpd: DHCPACK on 173.233.178.172 to f8:d0:ac:63:70:bd via 10.114.0.19
Sep 22 09:03:14 rad2 dhcpd: bind update on 173.233.178.172 from nu-z rejected: incoming update is less critical than outgoing update

My understanding of the failover protocol was that when both servers are
operating in "normal" mode, only one should reply to a client.
-- 
Chris Adams <cma at cmadams.net>


More information about the dhcp-users mailing list