From owner-freebsd-fs@FreeBSD.ORG Mon Jan 10 13:57:57 2005 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7D56B16A4CE; Mon, 10 Jan 2005 13:57:57 +0000 (GMT) Received: from obsecurity.dyndns.org (CPE0050040655c8-CM00111ae02aac.cpe.net.cable.rogers.com [69.199.47.57]) by mx1.FreeBSD.org (Postfix) with ESMTP id DCB2343D2D; Mon, 10 Jan 2005 13:57:54 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id 54A1951445; Mon, 10 Jan 2005 05:57:47 -0800 (PST) Date: Mon, 10 Jan 2005 05:57:47 -0800 From: Kris Kennaway To: Poul-Henning Kamp Message-ID: <20050110135747.GA44905@xor.obsecurity.org> References: <20050110100840.29845.qmail@web30301.mail.mud.yahoo.com> <41247.1105351904@critter.freebsd.dk> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="HlL+5n6rz5pIUxbD" Content-Disposition: inline In-Reply-To: <41247.1105351904@critter.freebsd.dk> User-Agent: Mutt/1.4.2.1i cc: FreeBsd-MailGrp cc: Scott Long Subject: Re: using mfs of size > 64Mb and system stability X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 10 Jan 2005 13:57:57 -0000 --HlL+5n6rz5pIUxbD Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Jan 10, 2005 at 11:11:44AM +0100, Poul-Henning Kamp wrote: > In message <20050110100840.29845.qmail@web30301.mail.mud.yahoo.com>, Bara= th S w > rites: > >Initially, I didn't go for swap based fs as I felt > >that the memory occupied will be from the swap area. > >As you are saying that the allocation will be from > >buffer/cache, I will test swap-mfs. >=20 > malloc backing should not be used for large disks. >=20 > If you _truly_ want to have a large disk which is memory backed, > you should consider using the "preload" backing as this will withdraw > the memory entirely from the kernels use. >=20 > In general, the benefit from using RAM disks is much smaller than > most people realize. I've found that using a swap-backed disk substantially cuts back on disk accesses for my purposes (package building, where everything that hits disk will be deleted again in a few minutes). It appears to give a reasonable performance boost, which I'm still trying to measure. Unfortunately, swap-backed md under 6.x deadlocks under some conditions which are being investigated. Kris --HlL+5n6rz5pIUxbD Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (FreeBSD) iD8DBQFB4onbWry0BWjoQKURAoRMAKDe4IyajL2SoKlKR+PvJBMS3gaeWwCg8vm8 Gq8hke7Io29IXodGpr1RQn4= =FgCB -----END PGP SIGNATURE----- --HlL+5n6rz5pIUxbD--