NSClient++ Help (#1) - Error (#961) - Message List

Can no one help me??? :(

I encounter the same problem on each Windows 2008 Server R2 x64 machine I tested this far.

On 2003 (x86 and x64) NSClient++ works fine.

In Testmode, I see the line

e tem\PDHCollector.cpp:143 Failed to query performance counters: \238(_total)\6: PdhGetFormattedCounterValue? failed {format: 1024}: -2147481642: A counter with a negative denominator value was detected.

every few seconds.

I already tried to delete %systemroot%\System32\wbem\Repository as mentioned in http://www.nsclient.org/nscp/discussion/message/2434 but the problem remains.

Additional Info:

I get the error with NSClient++ 0.4.0.170 and 0.39

Just tried rebuilding the Performance Counter Indexes using LodCtr?.exe /R, but to no avail.

Using NSClient++ 0.4.0.170, I get the error several times per second just starting the service, without querys from the icinga server.

This is a "known issue". I will try to see if I can track it down during summer as I sporadically get it here as well.

I will also for 0.4.2 rewrite the "windows check subsystem" making it more modern but that is a "long way off" (read fall/winter)...

Sorry I don't have happier news but it is a difficult to track down issue as it seldom happens on my machines...

It may be, that I have found a solution to that problem.

In every installation where I had that problem, I had disabled installing the firewall excepption (since we disable the windows firewall by default, firewalling is done on our firewalls).

When I installed the client on a testmachine I was to lazy to change all those setting and installed the default way, and it worked.

I have already reinstalled th client on one of the prblematic systems and with the firewall exception enabled, everything works.

I don't know, if that is the global solution to the problem, but at it may be worth a try.

For me the counters are working without an errormsg... The only thing i recognized was, that counters like "%Disk Time" or "Avg. disk queue length" return only 0 as value.. meanwhile my values in the perfmon are about 0,02 or 0,015.. and if i then want to add crit and warn values (like warn=80 and crit=90 or sth like that) nagios does not care... the state is always green seems like it is ignoring float values?

I think this is "random" which is why it is sometimes working and sometimes not...

I did (in 0.4.0.172) a check if the error occurs and a wait for half a second and a retry... hopefully this will improve the situation...

Hmm, that sounds very similar to the bug I reported here: ​http://nsclient.org/nscp/ticket/523

Well, im using checkcounter with nsca under NSClient 4.0.170, everything seems to work fine but it cant be possible that most of the counter i check return 0? got no errors in the nsclient.log neither in the nagios.log/syslog

if i then check counter like processor time, disk time, its always like :

nagios.log:

CURRENT SERVICE STATE: virtualpc;disktime_check;OK;HARD;1;OK: \LogicalDisk(_Total)\% Disk Time: 0

but if i add a counter like processor IO Write Bytes/sec.. it seems that the check is most of times 0, but at least sending some correct data?!

nagios.log:

CURRENT SERVICE STATE: virtualpc;io_check;CRITICAL;HARD;1;CRITICAL: \Process(_Total)\IO Write Bytes/sec:1.03891e+007 >critical

*EDIT:

I now saw that Counters Like DiskTime? or Processor Time are giving back values.. In an hour i got 2 Values over 0... is that possible?