From owner-freebsd-fs@FreeBSD.ORG Wed May 12 22:02:22 2010 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id A968A106564A for ; Wed, 12 May 2010 22:02:22 +0000 (UTC) (envelope-from pjd@garage.freebsd.pl) Received: from mail.garage.freebsd.pl (chello089077043238.chello.pl [89.77.43.238]) by mx1.freebsd.org (Postfix) with ESMTP id ECF928FC08 for ; Wed, 12 May 2010 22:02:21 +0000 (UTC) Received: by mail.garage.freebsd.pl (Postfix, from userid 65534) id CA29945E8E; Thu, 13 May 2010 00:02:20 +0200 (CEST) Received: from localhost (chello089077043238.chello.pl [89.77.43.238]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.garage.freebsd.pl (Postfix) with ESMTP id 3150E45DF4; Thu, 13 May 2010 00:02:15 +0200 (CEST) Date: Thu, 13 May 2010 00:02:06 +0200 From: Pawel Jakub Dawidek To: "Alexander V. Ribchansky" Message-ID: <20100512220206.GB2154@garage.freebsd.pl> References: <4BE95F1F.5090009@zk.informjust.ua> <20100511223813.GB3044@garage.freebsd.pl> <4BEAA0EB.3010407@zk.informjust.ua> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="oLBj+sq0vYjzfsbl" Content-Disposition: inline In-Reply-To: <4BEAA0EB.3010407@zk.informjust.ua> User-Agent: Mutt/1.4.2.3i X-PGP-Key-URL: http://people.freebsd.org/~pjd/pjd.asc X-OS: FreeBSD 9.0-CURRENT amd64 X-Spam-Checker-Version: SpamAssassin 3.0.4 (2005-06-05) on mail.garage.freebsd.pl X-Spam-Level: X-Spam-Status: No, score=-0.6 required=4.5 tests=BAYES_00,RCVD_IN_SORBS_DUL autolearn=no version=3.0.4 Cc: freebsd-fs@freebsd.org Subject: Re: Freebsd 8.0 kmem map too small X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 12 May 2010 22:02:22 -0000 --oLBj+sq0vYjzfsbl Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, May 12, 2010 at 03:36:59PM +0300, Alexander V. Ribchansky wrote: > I apply you patch on today's STABLE. with 2Gb of ram > m.kmem_size_max: 1073741824 > vm.kmem_size_min: 0 > vm.kmem_size: 1073741824 >=20 > hw.physmem: 2125430784 > hw.usermem: 1291554816 > hw.realmem: 2146959360 >=20 > vfs.zfs.arc_max: 268435456 >=20 > it survive two sequent buildworld & buildkernel. Wired memory is 791M.=20 > Without UMA Wired memory was around 500M on THE SAME settings. > Could you please explain as for nOOb, what we get with UMA, and why it=20 > uses MORE Wired memory? Is it for better perfomance or for what? [...] Well, UMA doesn't really free memory until it is asked by the pageout daemon when we are running out of physical pages. In general UMA allows to achive better performance than regular malloc(9), but I wasn't the one who decided to switch to UMA - it was Kip Macy. I'm not sure what tests he performed and what performance benefits he measured. > [...] What is=20 > your recomended tuning for i386 FreeBSD-8 on 2Gb ram typical desktop? Well, I was running i386 kernel on my laptop with 2.5GB for a few years without single 'kmem_map too small' panic. I had KVA_PAGES set to 512 in my kernel config and vm.kmem_size set to around 1GB (the biggest possible value I could find). --=20 Pawel Jakub Dawidek http://www.wheelsystems.com pjd@FreeBSD.org http://www.FreeBSD.org FreeBSD committer Am I Evil? Yes, I Am! --oLBj+sq0vYjzfsbl Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (FreeBSD) iEYEARECAAYFAkvrJV0ACgkQForvXbEpPzSJswCfcUrhTGZrZ1z2Qq3gSSkFFDex jFQAnAnWPOql5WOwH2Ox14dzt2VtUaWI =ZicN -----END PGP SIGNATURE----- --oLBj+sq0vYjzfsbl--