From owner-freebsd-advocacy Tue Mar 21 18:43:59 2000 Delivered-To: freebsd-advocacy@freebsd.org Received: from peach.ocn.ne.jp (peach.ocn.ne.jp [210.145.254.87]) by hub.freebsd.org (Postfix) with ESMTP id 90A2A37B9D0 for ; Tue, 21 Mar 2000 18:43:56 -0800 (PST) (envelope-from dcs@newsguy.com) Received: from newsguy.com (p18-dn02kiryunisiki.gunma.ocn.ne.jp [211.0.245.83]) by peach.ocn.ne.jp (8.9.1a/OCN) with ESMTP id LAA27232; Wed, 22 Mar 2000 11:43:50 +0900 (JST) Message-ID: <38D82B85.29686551@newsguy.com> Date: Wed, 22 Mar 2000 11:10:13 +0900 From: "Daniel C. Sobral" X-Mailer: Mozilla 4.7 [en] (Win98; I) X-Accept-Language: en,pt-BR,ja MIME-Version: 1.0 To: Christian Weisgerber Cc: freebsd-advocacy@FreeBSD.ORG Subject: Re: 21st Century Unix - web serving References: <200003210130.KAA74668@daniel.sobral> <38D74CB3.DF3AA476@newsguy.com> <8b8tle$22nj$1@bigeye.rhein-neckar.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-advocacy@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Christian Weisgerber wrote: > > I wish somebody would put some substance to such anecdotal stories. > I'm currently quite close to a Linux box which gets slashdot-like > effects (basically caused by minor access spikes and a badly written > backend that causes the load to explode), and so far it seems to > hold up quite well. What's the swap (and general memory footprint) usage and i/o rate during peak? > And FreeBSD certainly doesn't work magic. A few months ago, I did > a simple test. Ten processes, each one allocated some memory and > ran in a loop doing nothing but continuously writing a byte to each > page of its chunk of memory. I chose the process count and memory > allocation to cover 1.5x the size of the real memory of the box. > When I started the test, the hard disk light lit up solidly and > for all pratical purposes the box ground to a halt. No more movement > under X11. No more switching back to a text console. No more network > login. I watched for some time with amusement and finally pressed > the reset button. Precisely what real-world conditions you planned to test with THAT? :-) Hey, open a socket in each process to another server, and write regularly to it. Then benchmark that in a FreeBSD against a Linux box. -- Daniel C. Sobral (8-DCS) dcs@newsguy.com dcs@freebsd.org dcs@zurichgnomes.bsdonspiracy.net One Unix to rule them all, One Resolver to find them, One IP to bring them all and in the zone bind them. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-advocacy" in the body of the message