[Nagiosplug-help] Division by zero error in Win2K.

[David Maynard]

I don't suppose that the system is running Win2K Advanced Server and/or
clustering is it?  We have one instance where  the nsclient test for the
Paging File utilization intermittently logs an error and causes one of the
clustered services to hang for a few minutes.  The error only occurs on the
server where one particular cluster group is active.  If you move that
cluster group to a different server, you start getting the error on the
other server.  The error is very intermittent.  It happens about twice per
week on a test that runs every 5 minutes.  In this case, I am tempted to
blame Win2K Advanced Server and/or its interaction of PDH.dll.  For now, I
just pulled the Paging File test and the problem has not returned.

These systems are actually running a 1.0.6.X version of nsclient, but I
thought it was worth mentioning since some of the Win2K performance counter
problems we have encountered seem to be more impacted by loosely related
events such as installing service packs or other new software packages.
They tend to manifest themselves through nsclient though.

If I had to guess, I would guess that the Win2K performance data for the CPU
has been mangled on that server.  The Microsoft knowledge base has several
articles that talk about how to restore mangled performance data for things
like DNS and other Internet services.  You might try searching there to see
if it is possible to restore the base OS performance info.  I believe it is
possible, but the instructions might read something like "Fix the base info,
then reinstall the performance info for every application and service that
you have ever installed on the server."

--
  David P. Maynard
  OutServ.net, Inc. -- Coordinated IT Operations Solutions
  Email: dmaynard at outserv.net, Tel: +1 512 977 8918, Fax: +1 512 977 0986
--

-----Original Message-----
From: Patrick Benson (TT) [mailto:patrick.benson at tradingtechnologies.com] 
Sent: Wednesday, February 12, 2003 8:06 AM
To: 'nagiosplug-help at lists.sourceforge.net'
Subject: [Nagiosplug-help] Division by zero error in Win2K.


I have been upgrading several Win2K and NT4 servers from PNSClient version
1.0.6.2 to version 1.0.7.1. Out of sixty upgrades, one system is now
generating a "Division by zero" application error every five minutes, and it
sites NSClient as the source with an Event ID: 2. The original install was
uneventful, and uninstalling and reinstalling the PNSClient did not help.
The MS knowledge base was a bust as to what might be causing this.

It was suggested that I check the counters.defs file for the correct
Language ID, but the language ID is correct. I even stopped the NetSaint
service, copied and pasted the files for the client from a working system
onto the server, and restarted the service but the error still continues.

This error appears to coincide with the Nagios CPU check, which we have
scheduled to occur every five minutes. There is nothing unusual about this
server compared to the other servers that were upgraded, and this error did
not occur when running the 1.0.6.2 client.

Has anyone had a similar experience, or knows of a fix? Please let me know.
Thanks!


> -----------------------------------------------------------------
> TRADING TECHNOLOGIES
> TT NET Technician
> Patrick B. Benson
> patrick.benson at tradingtechnologies.com
> Internal Operations  +1 312-476-1017
> -----------------------------------------------------------------
>