Return-Path: owner-new-httpd Received: by taz.hyperreal.com (8.6.12/8.6.5) id MAA11144; Tue, 25 Jul 1995 12:02:26 -0700 Received: from gw0.telebase.com by taz.hyperreal.com (8.6.12/8.6.5) with ESMTP id MAA11137; Tue, 25 Jul 1995 12:02:23 -0700 Received: from wormhole.telebase.com by gw0.telebase.com id PAA27868 for ; Tue, 25 Jul 1995 15:19:51 -0400 Received: from khan.telebase.com (chuck@khan.telebase.com [192.132.57.215]) by wormhole.telebase.com (8.6.12/8.6.9.1) with ESMTP id PAA18236 for ; Tue, 25 Jul 1995 15:14:32 -0400 Received: (from chuck@localhost) by khan.telebase.com (8.6.12/8.6.9.1) id PAA24599 for new-httpd@hyperreal.com; Tue, 25 Jul 1995 15:05:30 -0400 From: Chuck Murcko Message-Id: <199507251905.PAA24599@telebase.com.> Subject: Re: Redirect in .htaccess files? To: new-httpd@hyperreal.com Date: Tue, 25 Jul 1995 15:05:29 -0400 (EDT) In-Reply-To: <9507251817.AA19715@volterra> from "Robert S. Thau" at Jul 25, 95 02:17:32 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1868 Sender: owner-new-httpd@apache.org Precedence: bulk Reply-To: new-httpd@apache.org Robert S. Thau liltingly intones: > > [ discussion of server dieoff and why I picked an insanely low > hard limit ] > > However, I am somewhat loath to let people set AbsMaxServers (or > whatever) to 7 and think they've done themselves a favor. As it is, > they've still got the source code, and they can screw themselves if > they really want to, but we haven't made it easy for them. > I agree, actually. I hadn't taken the time to think about the sizes of real requests for that example. I also think that the absolute max is probably not a good thing for casual tweaking, and putting it in the source is a way to make it harder to config up a dead dawg. Folks with heavy hardware and traffic should be able to bump it up. (Our machine here starts with 50 children, and actually does rise some from there to a stable level, typically 65-70). The nature of sent data has a lot of say in the way the server's configured, though, so the word's still caveat configurator. And even the best httpd config loses when the underlying machine config is not up to the job. My first stress tests, for instance, only used 11 or so servers because they were sending many small data items. I never got close to my listen backlog for that reason. It took more clients requesting larger chunks of data to do that, and the config profile became radically different, both for httpd and the OS. It seems a real quality rating for an httpd is not just how many hits a day it takes, but also how big the transactions are, and what kind of transaction mix is involved. BTW, BSDI and Solaris both seem stable under load at this time. chuck Chuck Murcko Telebase Systems, Inc. Wayne PA chuck@telebase.com And now, on a lighter note: Aphorism, n.: A concise, clever statement. Afterism, n.: A concise, clever statement you don't think of until too late. -- James Alexander Thom