[nycbug-talk] Odd behavior on FreeBSD 6.3 box

Jesse Callaway bonsaime at gmail.com
Fri Jan 30 01:23:52 EST 2009

On Fri, Jan 30, 2009 at 1:08 AM, Matt Juszczak <matt at atopia.net> wrote:
> I have a simple webserver/mysql box that usually works fine.  But tonight,
> I was seeing load averages in the 80's and 90's, incredibly high I/O wait,
> and perl in the top of the processlist using 80-90% of CPU.  Seemed to be
> spamassassin related, but I also had a ton of apache processes running.
> I'm still looking to see if perhaps a website was being hammered, but in
> the meantime I noticed that I was getting this repeatedly (about once a
> second) in my http-access log:
> ::1 - - [30/Jan/2009:05:52:23 +0000] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache/2.2.9 (FreeBSD) mod_ssl/2.2.9 OpenSSL/0.9.7e-p1 DAV/2 PHP/5.2.6
> with Suhosin-Patch mod_perl/2.0.4 Perl/v5.8.8 (internal dummy connection)"
> Does anyone know what that is (other than the fact that its a loopback
> dummy connection)?  It seems to have stopped since I restarted postfix and
> apache.
> Thanks for any thoughts...
> -Matt
> _______________________________________________
> talk mailing list
> talk at lists.nycbug.org
> http://lists.nycbug.org/mailman/listinfo/talk

Apache does this when it's trying to kill off threads. I forget
exactly what brings you to this condition, but check out the MaxSpare,
MinSpare, and all that stuff. Run 'httpd -M' i think to find out
whether you are MPM prefork or the other other one, the knobs mean
different things depending on which module is implemented. mod_perl
shouldn't be invoking /usr/bin/perl as far as I know, so this and the
Apache thing might be not directly related.... check that the cgi
module isn't being used for anything just to clear that out.


More information about the talk mailing list