From owner-freebsd-fs@FreeBSD.ORG Sat Aug 27 07:21:24 2011 Return-Path: Delivered-To: fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B4FCC106564A for ; Sat, 27 Aug 2011 07:21:24 +0000 (UTC) (envelope-from lev@FreeBSD.org) Received: from onlyone.friendlyhosting.spb.ru (onlyone.friendlyhosting.spb.ru [IPv6:2a01:4f8:131:60a2::2]) by mx1.freebsd.org (Postfix) with ESMTP id 52A6D8FC12 for ; Sat, 27 Aug 2011 07:21:24 +0000 (UTC) Received: from lion.home.serebryakov.spb.ru (unknown [IPv6:2001:470:923f:1:b1b7:d4b2:b3b3:a68b]) (Authenticated sender: lev@serebryakov.spb.ru) by onlyone.friendlyhosting.spb.ru (Postfix) with ESMTPA id 5DBFF4AC31; Sat, 27 Aug 2011 11:21:22 +0400 (MSD) Date: Sat, 27 Aug 2011 11:21:16 +0400 From: Lev Serebryakov Organization: FreeBSD Project X-Priority: 3 (Normal) Message-ID: <758608837.20110827112116@serebryakov.spb.ru> To: Kirk McKusick In-Reply-To: <201108262052.p7QKqpen039191@chez.mckusick.com> References: <1963980291.20110826232758@serebryakov.spb.ru> <201108262052.p7QKqpen039191@chez.mckusick.com> MIME-Version: 1.0 Content-Type: text/plain; charset=windows-1251 Content-Transfer-Encoding: quoted-printable Cc: fs@freebsd.org Subject: Re: Strange behaviour of UFS2+SU FS on FreeBSD 8-Stable: dreadful perofrmance for old data, excellent for new. X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: lev@FreeBSD.org List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 27 Aug 2011 07:21:24 -0000 Hello, Kirk. You wrote 27 =E0=E2=E3=F3=F1=F2=E0 2011 =E3., 0:52:51: > Given how much bigger your new filesystem has grown, copying out to > another filesystem should not be necessary. However, it will likely It is theoretically different strategies: (a) copy from FS to same FS, mand remove old copy later, force OS to choose new place for data. (b) copy out, remove, copy in allow OS to use same place. > be quicker to copy out to another filesystem as you will be using > two spindles instead of seeking back and forth on one. But before > doing all that work, try copying one file to ensure that you get the > expected speedup. Yep, three biggest files were speed up significantly. I'm going to investigate alter, why it is ony ~180MiB/s, when theoretically it should be about (90*4) 360MiB/s linear read, and whom to blame: UFS or geom_raid5 or both :) --=20 // Black Lion AKA Lev Serebryakov