Steven D. Morrey wrote:
> Hi Everyone,
>
> I was wondering if anyone has had a chance to run the event profiler on their system and see what comes up?
I don't have a permanent test setup here to be able to run it
continuously to get some idea, but I plan to setup one sometime in the
next month, I need to purchase a desktop. I'm keen to test this out, I
just haven't gotten a good chance.
> The reason I ask is that after running the event profiler for a while here, I've noticed, that at least on my setup, upwards of 70% of the time is spent in the reaper.
> Admittedly this could be due to a misconfiguration on my part, but changing the service reaper frequency from 3 to 15 made things worse.
> My Used/High/Total Check Result Buffers: are 264 / 622 / 4096 respectively.
> My avg time per event for the reaper is 3.26 seconds it went up to 10 seconds when the frequency was set to 15.
> My hunch is it probably would have been longer but there is the reaper bailout after a certain amount of time.
>
> However I'm using DNX and DNX writes directly to the circular results buffer, so the question is...
> What are the odds this is a problem with lock contention, or is there a more likely candidate.
>
I have no idea about this, does our implementation of locking block at
the moment?
> Also I'd like to extend the event profiling metrics to find out for sure what's consuming all thing time, but before I do I was hoping someone more versed in the internals of Nagios could point me in the direction of what would be some good next things to measure.
> Even so, if someone could please run the event profiler on their system and see if your getting similar results that would be a HUGE help.
> Thanks in advance!
>
I'm beginning to explore the insides, when I'm comfortable, I could help
more with this, for now though all I can do is run it for you when I get
a decent test environment setup.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]