[Nagiosplug-help] strange check_dhcp behavior

[Subhendu Ghosh]

On Wed, 10 Mar 2004, SCHAUSBERGER Christian wrote:

> 
> here´s the problem:
> 
> I´m monitoring 2 dhcp servers as a cluster. Nagios checks them correctly
> and most of the time everything is fine. But every now an then (once or
> twice per day) Nagios produces a false alert, because one of the dhcp
> servers is not responding.  So i wrote a little script that run´s the
> check_dhcp command permanently and logs the output. There i found out
> that around 5 - 10 percent of the requests don´t get an dhcpoffer. So i
> looked at the dhcp-servers: every single nagios-request appears in the
> logs and for every request an offer has been sent. So it is not a
> network problem. Next I sniffed a little bit of network traffic to see
> if the dhcp offers actually arive at the nagios host. And they do. For
> every request there is an offer logged by tcpdump. Is it possible that
> the plugin somehow misses some of the offers?
> 

This may be an issue of check_dhcp running simultaneously for each of the 
dhcp servers and clobbering the socket. It shouldn't happen, but...

Can you find any overlap between the request-offer intervals for the 2
checks...  specially when the offers are being lost.

--
-sg