dhcpd in failover not reusing expired leases

David W. Hankins David_Hankins at isc.org
Tue Oct 14 15:03:56 UTC 2008


On Tue, Oct 14, 2008 at 05:25:58AM -0400, Robert Blayzor wrote:
> If we restart one or both of the servers, it seems things balance out, at 
> least for a while, but then we just end up back at this point in a few 
> days.

'expired' leases (and 'reset', 'released') are what I call
"transitional" states.  They live in that state only until the
peer agrees (sends back an ack), at which time they go straight to
'free' after the ack (no additional bndupd).

I think it is suspicious that you've got an expired/free thing going
on.  If the lease was free at one time, then active, then expired, the
servers should have it down as expired/active or expired/expired.

I would suspect something is causing the failover channel to
disconnect and hold it down, so the peer misses the binding update
to active as well as the expired update.  But then it seems to
reconnect on a restart?

So that's a very strange symptom.

> any ideas?

Grep logs for failover state changes.  See if they're going to
comms-interrupted and holding there.

Check your clocks; being too far out of synch may cause updates to be
refused.

Beyond that, no idea other than to try 3.0.7 or 3.1.1.  I don't know
offhand if we touched failover between 3.0.5 and 3.0.7, you could
check the relnotes.

-- 
Ash bugud-gul durbatuluk agh burzum-ishi krimpatul.
Why settle for the lesser evil?	 https://secure.isc.org/store/t-shirt/
-- 
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