Switching from Single DHCP Server to Failover

Glenn Satchell Glenn.Satchell at uniq.com.au
Wed Apr 19 11:53:29 UTC 2006


>To: dhcp-users at isc.org
>Subject: Switching from Single DHCP Server to Failover
>Date: Tue, 18 Apr 2006 15:24:03 -0500
>From: Martin McCormick <martin at dc.cis.okstate.edu>
>
>If one is changing from a single DHCP server to a failover pair and
>wants to keep dhcp alive during the change, is it necessary to start
>the primary dhcp server in partner down mode?  The documentation for
>dhcpd.conf states that if both servers have never communicated, there
>will be a dead time equal to mclt before both servers go in to normal
>operation.  That much dead time in our world usually raises a few
>eyebrows so I would like to try to keep the primary talking so that
>clients don't start asking, "What's wrong?"
>
>	Some of the examples I have seen have
>  mclt 1800;
>while others show 1 hour or 3600 seconds.  Either way, that's too long
>to be down if we can help it.

Assuming your current server is to become the primary. Add the failover
commands to the configuration and restart the server. Start the
secondary. The above scenario only occurs if the secindary has already
issued leases. If it has an empty lease file the first thing that
happens is that the secondary requests a copy of all the leases and
then both server switch to normal mode and are able to hand out leases
immediately. At least that's been my experience.

>	One other question:
>
>	In each pool declaration as well as in the primary or
>secondary statements for dhcpd.conf, there is a line like:
>
>failover peer "abc" {
>
>	Am I correct in understanding that this string can be anything
>inside the double quotes so long as both peers contain the same
>string?

Yes. But making it descriptive helps with the log entries.

>	Many thanks.
>
>Martin McCormick WB5AGZ  Stillwater, OK 
>Systems Engineer
>OSU Information Technology Department Network Operations Group

regards,
-glenn



More information about the dhcp-users mailing list