[geeks] Need expert opinion -- resolved traffic issue

[velociraptor]

> > Hmmm...this is Solaris.  I was speculating on the cause of a
> > huge decrease in traffic (24K hits/day to 4K hits/day across 4
> > servers), but neither of those results would explain it.  If the
> > httpd died, a) the load balancers would ship it to the other
> > servers, and b) our alarm would go off.  Even if "b" failed, "a"
> > would still happen.
> >
> >Httpd visits.  On a particular URL we went from 24K hits to 4K hits.

In case anyone was wondering, it was not the cronjob rolling the
logs.  It was an upgrade of SSL that I did which caused a problem
with the non-SSL httpd.  Apache was compiled with the mod_ssl 
switch, but with the libraries dynamically linked.  So, on the SSL 
side, I was able to create links and it just worked.  On the non-SSL 
side of httpd, though, it created some kind of odd bug on the httpd
to Weblogic backend that WL coming back to the httpd proxy was 
tickling.

For this particular URL, we were generating ~11 "Abort (6)" errors
per minute.  The others were generating maybe 1 per hour until put
under load.

Even if I had done testing in our staging environment with httpd 
in addition to what I did with the SSL sites, I doubt I would have
found this doing the manual testing that I did, as the app that 
was causing the 11 errors a minute was recently switched to be 
using the "portal" function of WL.

Looks like now I am going to have to talk the bosses into some
type of software that will generate loads in staging so we can
perform our own stress tests.  And I'll have to learn it as well...
in my copious spare time.

=Nadine=