From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Andrewartha Subject: High iowait and poor performance - ata_piix Date: Mon, 15 Aug 2005 17:34:25 +0800 Message-ID: <430061A1.2040806@daa.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from quoll.daa.com.au ([203.17.251.1]:31402 "EHLO quoll.daa.com.au") by vger.kernel.org with ESMTP id S932377AbVHOJen (ORCPT ); Mon, 15 Aug 2005 05:34:43 -0400 Received: from [203.17.251.34] (narbalek.daa.com.au [203.17.251.34]) by quoll.daa.com.au (Postfix) with ESMTP id 5AA9D3F8028 for ; Mon, 15 Aug 2005 17:34:27 +0800 (WST) Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: linux-ide@vger.kernel.org Hi, I've got a Proliant DL360 G4 (dual 3.4GHz Xeon, 1GB ram), which has the Intel E7520 chipset and uses ata_piix. I'm getting some very high iowait - 6-10% while idle, and 40-70% while processing mail (which is postfix+dpsam/postgresql). I've tried the ubuntu packages for 2.6.{10,11,12} with no change. I also noticed some odd results with bonnie++ - it's slower writing per block than per character (although block writing uses a lot less cpu): 2.6.11 on a Seagate ST3160827AS (160GB 7200.7): Version 1.03 ------Sequential Output------ --Sequential Input- --Random- -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP quoll 4G 22016 42 20982 7 15462 3 32869 56 39013 4 155.9 0 2.6.12 on a Maxtor 6Y080M0 (80GB DiamondPlus 9): Version 1.03 ------Sequential Output------ --Sequential Input- --Random- -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP quoll-maxtor 4G 30381 53 28041 5 16912 3 35664 61 57015 7 108.3 0 The Seagate results were done on a lvm on raid 1, while the Maxtor was just on an lvm, and neither disk was under any significant load while testing (mail was turned off). For reference, a 200GB Seagate 7200.7 on a dual opteron 242 with 2GB ram and the promise SX8 controller (using the driver from the promise website): Version 1.03 ------Sequential Output------ --Sequential Input- --Random- -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP martello-seagate 4G 40457 96 58846 28 18211 6 29935 64 62569 8 199.3 0 Unfortunately there's no option in the BIOS to use AHCI mode (if it's even possible on this chipset), and it's 1RU system so I can't use another controller. Any suggestions on working out what's wrong? -- James Andrewartha Systems Administrator Data Analysis Australia Pty Ltd