From owner-freebsd-stable@FreeBSD.ORG Wed Feb 23 15:40:04 2011 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 9CF6B106566C for ; Wed, 23 Feb 2011 15:40:04 +0000 (UTC) (envelope-from feld@feld.me) Received: from mail-vx0-f182.google.com (mail-vx0-f182.google.com [209.85.220.182]) by mx1.freebsd.org (Postfix) with ESMTP id 5079F8FC0A for ; Wed, 23 Feb 2011 15:40:04 +0000 (UTC) Received: by vxc34 with SMTP id 34so2333396vxc.13 for ; Wed, 23 Feb 2011 07:40:03 -0800 (PST) Received: by 10.52.157.7 with SMTP id wi7mr6698200vdb.120.1298475427451; Wed, 23 Feb 2011 07:37:07 -0800 (PST) Received: from tech304 (supranet-tech.secure-on.net [66.170.8.18]) by mx.google.com with ESMTPS id e20sm1646027vbz.18.2011.02.23.07.37.06 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 23 Feb 2011 07:37:06 -0800 (PST) Content-Type: text/plain; charset=utf-8; format=flowed; delsp=yes To: freebsd-stable@freebsd.org References: <20110217115653.GH34314@home.opsec.eu> <20110217124515.GI34314@home.opsec.eu> <4D5D1D10.7010000@digsys.bg> <20110217180140.GL66849@dan.emsphone.com> <20110217204643.GJ34314@home.opsec.eu> <37576665@h30.sp.ipt.ru> <20110221161637.GR34314@home.opsec.eu> Date: Wed, 23 Feb 2011 09:37:05 -0600 MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: "Mark Felder" Message-ID: In-Reply-To: <20110221161637.GR34314@home.opsec.eu> User-Agent: Opera Mail/11.01 (FreeBSD) Subject: Re: 3TB disc and block alignment X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 23 Feb 2011 15:40:04 -0000 On Mon, 21 Feb 2011 10:16:37 -0600, Kurt Jaeger wrote: > Hi! > >> > Hmm, wasn't the issues with 3T drives, that they internally use >> > 4K blocks and emulate 512 and that therefore 8 block alignments >> > are an performance issue ? >> Hi guys, I'd just like to jump on this train real quick since I have a related question with alignment. I've been building new Virtual Machine templates lately because I recently learned that our SAN which we access over iSCSI uses 64K blocks and the default FreeBSD install starts a filesystem on sector 63 (512K sectors). This would be misaligned for most I/O transactions. As a result, I updated our internal documentation for FreeBSD standards to be 64K aligned by manually installing FreeBSD with GPT and making the FreeBSD-boot partition 64K in size: Fixit# gpart create -s gpt da0 Fixit# gpart add -t freebsd-boot -s 64K da0 Fixit# gpart add -t freebsd-swap -s 2G -b 2048 da0 Fixit# gpart add -t freebsd-ufs da0 Fixit# gpart bootcode -b /dist/boot/pmbr -p /dist/boot/gptboot -i 1 da0 At the time, the examples I found used -s 64K for the freebsd-boot partition which starts the next partition/filesystem on the next 64K block which would be aligned for our purposes. However, I have been seeing -s 1024K or larger which is also 64K aligned, but just larger. Am I shooting myself in the foot by not going ahead and aligning with -s 1024K or -s 2048K right now? Thanks for your opinions, Mark