Hi Mark,
On Sat, Jul 13, 2013 at 11:35:56AM +0200, Mark Janssen wrote:
> On Fri, Jul 12, 2013 at 5:57 PM, Lukas Tribus <luky...@hotmail.com> wrote:
>
> > Hi Mark,
> >
> > > epoll_wait(0, {}, 200, 0) = 0
> > > (repeated 10-15 times)
> >
> > A few questions:
> >
> > Can you reproduce this without the health checks?
> >
>
> I haven't tried yet... and can't really test this currently
>
>
> > Do you have the possibility to run without nbproc?
> >
>
> We were running without nbproc, and saw the 100% cpu load (though this was
> still on dev18). We switched to nbproc=7 to work-around the 100% cpu load,
> as I was told that the sites weren't responsive at the 100% cpu load times
> (though I couldn't reproduce this)
>
> How often do you see the cpu-peak and how long exactly does it take
> > to come back to your normal load?
> >
>
> A couple of times per day.... for a few minutes each, the cacti-cpu graph
> will show the peaks clearly.
>
> I do have more strace's now, from the peak-cpu times and when they divert
> back to normal. (Though the files are big... 800+ MB)
Well it can be useful anyway. If you have enough bandwidth, I would
appreciate a URL from which to download them. That will give me more
information about the surrounding context (eg: only happens when a
connection fails, etc...)
Thanks,
Willy
