From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 25 Oct 2002 10:30:07 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 25 Oct 2002 10:30:07 -0400 Received: from draal.physics.wisc.edu ([128.104.137.82]:10181 "EHLO draal.physics.wisc.edu") by vger.kernel.org with ESMTP id ; Fri, 25 Oct 2002 10:29:37 -0400 Date: Fri, 25 Oct 2002 09:35:42 -0500 From: Bob McElrath To: linux-kernel@vger.kernel.org Subject: process load and swapping Message-ID: <20021025143541.GC13482@draal.physics.wisc.edu> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="S1BNGpv0yoYahz37" Content-Disposition: inline User-Agent: Mutt/1.4i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org --S1BNGpv0yoYahz37 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable I have been experiencing the following behavior on my alpha with many kernel versions (2.4 series, latest is 2.4.19 with Ingo's O(1) scheduler patch). I'm running apache and zope, and when I load a zope page, the load jumps to ~12 and the computer starts skipping. That is, about once every 2s the music (xmms) stops for ~0.5s and then comes back. This repeats every 2s until zope is finished creating the page (which takes sometimes longer than a minute). When the computer "skips", not only does the music stop, but the mouse stops moving (i.e. I move it continuously and it pauses for 0.5s every 2s). This happens both at the console in X and on virtual terminals (ssh'ed from another computer). It's as if it's not servicing interrupts for 0.5s. Or does scheduling not occur when the kernel decides to load N pages from swap, no matter how large N is? This is totally due to swap use. If I turn off all swap ('swapoff -a'), the response of zope is speedy. This behavior has happened with all kernel versions I have used, at least since 2.4.17. So, any ideas? I know there have been arguments about the 'swappiness' of this kernel, but the load going to 12 and interrupts not being serviced is totally quantitative and totally repeatable... Cheers, -- Bob Bob McElrath Univ. of Wisconsin at Madison, Department of Physics "The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."=20 -- Nietzsche --S1BNGpv0yoYahz37 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAj25Vr0ACgkQjwioWRGe9K0IQQCfVoAtKpixoM+KkcHr5pauKgjQ DJcAoORods/oWPNkxxSE0oR50B9c3vjB =xHTG -----END PGP SIGNATURE----- --S1BNGpv0yoYahz37--