From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751612Ab1JLAZ6 (ORCPT ); Tue, 11 Oct 2011 20:25:58 -0400 Received: from tn-76-7-162-101.sta.embarqhsd.net ([76.7.162.101]:51435 "EHLO animx.eu.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750959Ab1JLAZ5 (ORCPT ); Tue, 11 Oct 2011 20:25:57 -0400 Date: Tue, 11 Oct 2011 20:25:56 -0400 From: Wakko Warner To: linux-kernel@vger.kernel.org Subject: Slow write performance with mpt2sas Message-ID: <20111012002556.GA32495@animx.eu.org> Mail-Followup-To: linux-kernel@vger.kernel.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org I've noticed that my SAS9212-4i4e is having write issues. I have a raid5 of 4 2tb seagate green drives and a raid5 of 3 300gb seagate 15k sas drives. Write performance on the green drives is fairly poor. The riad4 of 4 disks has a write performance of about 6-10mb/sec. My onboard sas controller (1068e chip IT mode) does not have this issue (I'm using raid6 on 8 of the same model green drives). Write performance on the 15k drives in raid 5 was 140mb/sec (4gb volume). Here's some information from lsi's sas2flash util: Controller : SAS2008(B1) PCI Address : 00:03:00:00 SAS Address : 500605b-0-01d0-ab00 NVDATA Version (Default) : 0a.03.00.02 NVDATA Version (Persistent) : 0a.03.00.02 Firmware Product ID : 0x2713 Firmware Version : 11.00.00.00 NVDATA Vendor : LSI NVDATA Product ID : SAS9212-4i4e BIOS Version : 07.21.00.00 UEFI BSD Version : N/A FCODE Version : N/A Board Name : SAS9212-4i4e Board Assembly : H3-25134-00A Board Tracer Number : P131430410 I'm using stock kernel 2.6.39 compiled for 64-bit. This sas controller has IR firmware. Relevent information from dmesg: [ 1.585459] mpt2sas version 08.100.00.00 loaded [ 1.585564] scsi0 : Fusion MPT SAS Host [ 1.585772] mpt2sas 0000:03:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16 [ 1.585846] mpt2sas 0000:03:00.0: setting latency timer to 64 [ 1.585849] mpt2sas0: 64 BIT PCI BUS DMA ADDRESSING SUPPORTED, total mem (12301640 kB) [ 1.585966] mpt2sas0: IO-APIC enabled: IRQ 16 [ 1.586027] mpt2sas0: iomem(0x00000000fac3c000), mapped(0xffffc90000020000), size(16384) [ 1.586110] mpt2sas0: ioport(0x000000000000a000), size(256) [ 1.872511] mpt2sas0: sending diag reset !! [ 2.355485] Refined TSC clocksource calibration: 2133.333 MHz. [ 2.355561] Switching to clocksource tsc [ 2.974166] mpt2sas0: diag reset: SUCCESS [ 3.121510] mpt2sas0: Allocated physical memory: size(3999 kB) [ 3.121574] mpt2sas0: Current Controller Queue Depth(1753), Max Controller Queue Depth(2000) [ 3.121659] mpt2sas0: Scatter Gather Elements per IO(128) [ 3.353750] mpt2sas0: LSISAS2008: FWVersion(11.00.00.00), ChipRevision(0x02), BiosVersion(07.21.00.00) [ 3.353848] mpt2sas0: Protocol=(Initiator), Capabilities=(Raid,TLR,EEDP,Snapshot Buffer,Diag Trace Buffer,Task Set Full,NCQ) [ 3.354367] mpt2sas0: sending port enable !! [ 4.888265] mpt2sas0: host_add: handle(0x0001), sas_addr(0x500605b001d0ab00), phys(8) [ 10.998203] mpt2sas0: port enable: SUCCESS Any suggestions? PS. I am on the list. I do prefer courtesy CC. -- Microsoft has beaten Volkswagen's world record. Volkswagen only created 22 million bugs.