Re: [Pnp4nagios-users] Nagios hanging when performance processing enabled

[Hendrik B. <an...@pr...>]

Hi Eric,

Eric F Crist schrieb:
> Hey folks,
> 
> I think this may be a Nagios problem, but need to start at the bottom  
> of the totem.  As I mentioned in an earlier post, I've upgraded to  
> 0.4.7 and I'm running Nagios 2.10.  When I change  
> process_performance_data=1 from 0 and restart Nagios, the nagios  
> process begins to chew 100% of my processor as soon as it's done  
> calling the perfparse perl script provided by pnp4nagios.  I've set  
> log level to 100 in the perl script and there is absolutely NO useful  
> data I can discern indicating a problem or failure in pnp4nagios.
> 
First: There is no loglevel 100 within PNP. AFAIR a value of 3 is the 
maximum.

How many Services are you checking with nagios? If you are monitoring 
more than 2000 Services at a average check_interval=300 seconds: try 
pnp4nagios in bulk mode since it's a killing feature.
Nagios will execute the perfdata script after every single service check.

Some maths: 2000 Checks in 300 seconds == 6,66 Nagios check results per 
seconds.
With enabled performance data processing: 13,33 actions (reaping results 
+ handling perfdata)


> Has anyone else had a problem?  If I disable  
> process_performance_data=0 from 1, nagios hums along fine.  FWIW, the  
> hung process is a child of the nagios daemon, which looks like it may  
> be a check.  Could it be the perl script isn't exiting correctly or  
> something along those lines?
> 

The process_perdata.pl is setting a timeout by itself. So I guess it 
shouldn't hanging around more than 10 seconds (default timeout).
But without deeper debugging it is just guessing around.

Regards,
Hendrik