www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: Changed information for PR general/278
Date Wed, 16 Apr 1997 08:27:26 GMT
I looked closely at the benchmarks Joseph had done and the only conclusion
I could make was that he hadn't run them long enough to make me feel that
there was really any problem.  His running times were in the 1 to 2 second
range.  Apache 1.2 has the throttled child start thing whereas apache 1.1
doesn't have it... so if you're benchmarking in such a short amount of
time it's entirely likely you'll see that difference.  The tests that
1.1.3 was performing much better on involved 100 simultaneous clients, 1
request each... on a default server (start 5, minspare 5, maxspare 10). 
1.2 faired the same on the tests involving only 1 client w/100 requests.

While I would agree that 1.2's child mgmt stuff isn't the greatest... I
also don't think it's bad.  I'm succesfully running two servers with
HARD_SERVER_LIMIT=1024... the big problem I see at that number of children
is that mod_status takes too long :) 

Dean

On Tue, 15 Apr 1997, Marc Slemko wrote:

> What!??!  I would love to know what the heck fixed it...
> 
> On Tue, 15 Apr 1997 dgaudet@hyperreal.com wrote:
> 
> > Synopsis: 40-50% performance decrease relative to apache 1.1.3
> > 
> > State-Changed-From-To: feedback-closed
> > State-Changed-By: dgaudet
> > State-Changed-When: Tue Apr 15 18:12:41 PDT 1997
> > State-Changed-Why:
> > Joseph reports that 1.2b8 doesn't exhibit the same problems.
> > 
> > Dean
> > 
> 
> 


Mime
View raw message