Nagios causing signal 11 on apache.

Marc Powell marc at ena.com
Thu Feb 19 23:32:39 CET 2004


On Thursday, February 19, 2004 3:27 PM, Leif Nixon shared with us:

> Tim Uckun <tim at diligence.com> writes:
> 
> > > Hm, Apache threads are getting killed off by segmentation faults.
> > > That's bad. That's a bug in Apache, or possibly flaky memory. It's
> > > hard to see that Nagios could be involved, though.
> > 
> > If I turn off the nagios checking then the signal 11s go away. If I
> > turn it on again they come back. I am 100% convinced that nagios is
> > causing these signal 11s.
> 
> The nagios daemon (the bit that's doing the actual service checking,
> as opposed to the web page generating CGI stuff) doesn't have
> *anything* to do with the web server, so it seems... unlikely that it
> directly could cause Apache threads to die.  
> 
> However, if you do have a flaky memory bank in the high memory
> regions, the extra memory load caused by the nagios processes might
> push the Apache processes to higher memory locations, where they get
> hit by the bad memory. Just a shot in the dark, but I've seen
> stranger things.    
> 
> You don't mention what hardware you are running on, but if it's some
> sort of x86 box, I'd run memtest86 on it. 


Another, more direct relationship would be if he is using nagios to
check the availability or status of Apache on that box using check_http
or some other plugin in a manner that is causing apache to bork. Tim, is
that a possibility?

--
Marc


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id56&alloc_id438&op=click
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null





More information about the Users mailing list