From owner-freebsd-fs@FreeBSD.ORG Sat Feb 20 12:25:55 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 CCF29106566C for ; Sat, 20 Feb 2010 12:25:55 +0000 (UTC) (envelope-from dimitry@andric.com) Received: from tensor.andric.com (cl-327.ede-01.nl.sixxs.net [IPv6:2001:7b8:2ff:146::2]) by mx1.freebsd.org (Postfix) with ESMTP id 959EC8FC0C for ; Sat, 20 Feb 2010 12:25:55 +0000 (UTC) Received: from [IPv6:2001:7b8:3a7:0:705d:318e:2cb0:5f81] (unknown [IPv6:2001:7b8:3a7:0:705d:318e:2cb0:5f81]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) by tensor.andric.com (Postfix) with ESMTPSA id D9DD25C59; Sat, 20 Feb 2010 13:25:52 +0100 (CET) Message-ID: <4B7FD4D2.7040802@andric.com> Date: Sat, 20 Feb 2010 13:25:54 +0100 From: Dimitry Andric User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.9.2.2pre) Gecko/20100218 Lanikai/3.1b1pre MIME-Version: 1.0 To: mailinglists References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-fs@freebsd.org Subject: Re: zfs on 4k sector disks 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: Sat, 20 Feb 2010 12:25:55 -0000 On 2010-02-20 10:47, mailinglists wrote: > This "Advanced Format" basically means the sectors size on disk is > changed to 4k in stead of 512k. Please note those disks still seem to expose 512B sectors to any OS; the 4kiB sectors are only used internally. And there seems to be no jumper to "fix" this behaviour... Unaligned writes can cause multiple read-write-modify operations, which are most likely bad for performance. > I'm wondering if freeBSD 8.0 using ZFS raidz and drives like these will > work properly, I've seen a lot of articles and postings about partitions > not being properly aligned with the physical disk layout and suffering > big performance hits. My question is; should I worry about these issues > or does ZFS already align the data on a 4k border? You can at least try to make sure your partitions are aligned to 4kiB. This should be easy enough to do with e.g. gpart(8). I see no specific alignment options in the zpool(1M) or zfs(1M) manpages, but there may be some trick that I am not aware of. :) With UFS, you might want to experiment with block and fragment sizes, to see what is most optimal. For example, a block size of 32kiB and fragment size of 4kiB, on a partition aligned to 4kiB.