Well, I have been running it for 9 hours now and I have had no orphans!=20
Setting reaper_frequency =3D 1 did it. Also my number of nagios process=20
running at one time has dropped significantly, from usually near 200 to=20
between 1 and 40 with an average of about 15.=20
I wonder if some shared memory limit or pipe limit was met in linux, wher=
e=20
within some 10 second period the right (or wrong!) number of checks was=20
run and something went over the threshold, making them pile up until=20
nagios killed them (?) and called them orphans. I saw this buildup happe=
n=20
where more and more processes would get older until most of them were old=
,=20
(nagios had only 200 slots to work with), and soon after they all went=20
away. Thoughts?
Thanks to everyone who helped with suggestions!
--Geoff
On Wednesday 12 March 2003 08:48, Geoff Lovett wrote:
> That's a good thought. I was thinking that's what it looked like. I'l=
l
> try adjusting the reaper_frequency and let you know.
>
> Thanks,
> Geoff
>
> On Wednesday 12 March 2003 02:35, Maljaars, Eelco wrote:
> > Hi,
> >
> > I had a similar problem with tons of Nagios processes laying around
> > pretty much doing nothing.
> >
> > After some trial and error, I recuded the reaper_frequency from 10
> > seconds to 1 second. The reaper frequency determines how often the
> > main nagios proces collects returned data from childs (I think). Mayb=
e
> > the processes you see are just waiting to deliver their results to th=
e
> > main nagios proces.
> >
> >
> > Eelco Maljaars
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: geoff.lovett@lightningsource.com