Failover peer ignores some requests

[David W. Hankins]

On Mon, May 08, 2006 at 02:25:29PM +0200, Sebastian Hagedorn wrote:
> May  8 13:50:06 noc2 dhcpd: pool 88af2f0 134.95.128/24 total 27  free 5 
> backup 0  lts 2
> 
> Even that is unclear to me. Doesn't "free 5" mean that *this* peer has 5 
> free leases? Anyway, what's even more baffling is the behavior of the other 

Well, yes, "free 5" means there are 5 free leases...but what you're
missing out on is that the secondary can only allocate leases in the
'backup' state.

Only the primary can allocate leases in the 'free' state.

> May  8 13:54:19 lvr13 dhcpd: pool 89be580 134.95.128/24 total 27  free 4 
> backup 1  lts -1

It's pretty clear that your lease state databases are out of synch.

I'd upgrade to 3.0.4 (or wait until Jason completes testing and
distributes a package).

Before or after the upgrade (or both) you'll also need to fault the
secondary's dhcpd.leases (stop dhcpd, rm it and touch a new one, start
dhcpd).  The servers will transfer the lease database.  I should remind
you the secondary will not answer queries for a time equal to MCLT - this
is normal.

-- 
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