Unclear on behavior of dhcp server during recover wait period

[Nomad]

I've setup dhcp-3.1.0 on two linux machines and everything appears to be
working.  The primary has a split of 255 so it's the preferred server.
 Almost had a heart attack the first time I reloaded the primary and I saw
all the:

... not responding (recover wait)

messages.  I did some additional reading and realized this was normal as the
server being reloaded will wait the duration of the MCLT before it comes
back up.  So I lowered the MCLT to lower the length of my heart problems
after a reload.

We have a monitoring application that requests an address from the server
every few minutes and if it receives an offer, considers the service to be
up.   For all our networks we have helper-addresses configured on our
routers to foward dhcp requests.

My question is this:  This application sends a request to each specific
server.  After the new server install, we've been receiving outage notices
after the dhcp service is reloaded and is in recover wait state.  Since this
is a request to the specific server that is down and it's in recover wait
state, it would seem that it does not "send" the request to it's peer.  This
would seem to mimic the behavior that acquired it's address from this server
and attempts to renew.

Is there something I'm missing in what the "recover wait" server is doing?
 I know it's waiting for the expiration of the MCLT but shouldn't clients be
getting redirected (if that makes any sense) to the peer?

Any information woudl be appreciated.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20080108/4bf7d216/attachment.html>