From owner-freebsd-current@FreeBSD.ORG Tue Mar 14 02:33:32 2006 Return-Path: X-Original-To: freebsd-current@freebsd.org Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 888CF16A401 for ; Tue, 14 Mar 2006 02:33:32 +0000 (UTC) (envelope-from ahebert@pubnix.net) Received: from mail.pubnix.net (Mail.pubnix.net [192.172.250.1]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2840F43D45 for ; Tue, 14 Mar 2006 02:33:31 +0000 (GMT) (envelope-from ahebert@pubnix.net) Received: from [10.0.1.2] (aal.pubnix.net [64.235.216.13]) (authenticated bits=0) by mail.pubnix.net (8.13.4/8.13.4) with ESMTP id k2E2XU5h065351 for ; Mon, 13 Mar 2006 21:33:30 -0500 (EST) (envelope-from ahebert@pubnix.net) Message-ID: <44162B7A.4000302@pubnix.net> Date: Mon, 13 Mar 2006 21:33:30 -0500 From: Alain Hebert Organization: PubNIX, Inc. User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20060130 X-Accept-Language: en-us, en MIME-Version: 1.0 CC: freebsd-current@freebsd.org References: <4410F0BE.6050600@pubnix.net> <200603100604.48535.thierry@herbelot.com> <441112DB.9060907@pubnix.net> <20060310154305.GA29639@uk.tiscali.com> <4411D9D7.7060409@pubnix.net> In-Reply-To: <4411D9D7.7060409@pubnix.net> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Was: SOLVED: Re: HowTo?: Current -> RELENG_6_0 Patch for "semaphore timeout !! DANGER Will Robinson !!" issue with Promise Fasttrak Raid 0/1 Card X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: ahebert@pubnix.net List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 14 Mar 2006 02:33:32 -0000 Well no quite, The short form is does anybody has a working setup with a PDC20580 (Fasttrak 2300) and WD's drives? I think either the FT2300 is a joke or WD's are just not the right choice for real computing. (They might be good for gaming and windoze... Aint much productive for me) ----- The long form: With RELENG_6 WD1600JS are just making too many errors (every few minutes at heavy load). WD2500KS (which are suppose to be the RAID approuve ones make some errors (every hours at heavy load)). Effect Disk I/O stop for a few seconds, until the situation get resolved by the driver. FYI it should be the same with CURRENT. Didn't try it but I could. With RELENG_6_0 Got those Will Robinson Messages. Effect Disk I/O is just frozen and one must push the big red button. With RELENG_5_4 (which works with a PDC20378 with 2 x ST3160023AS) Raid just fail (at the first error) and randomly drop one of the 2 drive. The server still works. ------ What is heavy load? 4 iozone + a find / Alain Hebert wrote: > Thanks. > > I've got confused and taught RELENG_6 was 6.0-PRERELEASE. > ------ > > In this summary you'll find more details with 6.1-PRERELEASE and > the WD1600JS's that I'm using with the Fasttrak 2300 in RAID1. > > ------ > > A little thread about those drive: > > http://hardware.mcse.ms/archive12-2005-8-220010.html > > As you can see my drives are similar except for last number in > device model and that the firmware are totally different. > (A check with WDC show no way to upgrade/download firmware... argh!) > > Also they are capable of "automatic acoustic management" but its > disable... thus no timeout problem with raid controller. > > ----- ad4 > device model WDC WD1600JS-00MHB1 > serial number WD-WCANM2041663 > firmware revision 10.02E01 > ----- ad6 > device model WDC WD1600JS-00MHB0 > serial number WD-WCANM2124370 > firmware revision 02.01C03 > ----- > > # atacontrol cap ad4 > > Protocol Serial ATA v1.0 > Serial ATA II > device model WDC WD1600JS-00MHB1 > serial number WD-WCANM2041663 > firmware revision 10.02E01 > cylinders 16383 > heads 16 > sectors/track 63 > lba supported 268435455 sectors > lba48 supported 312581808 sectors > dma supported > overlap not supported > > Feature Support Enable Value Vendor > write cache yes yes > read ahead yes yes > Native Command Queuing (NCQ) no - 0/0x00 > Tagged Command Queuing (TCQ) no no 0/0x00 > SMART yes yes > microcode download yes yes > security yes no > power management yes yes > advanced power management no no 0/0x00 > automatic acoustic management yes no 254/0xFE > 128/0x80 > ----- > > Thore are the warnings that now come up instead of freezing like in > 6.0-RELEASE: > (only once, all at the same time, and early during this 2h of disk > stress test) > ad6: WARNING - SETFEATURES SET TRANSFER MODE taskqueue timeout - > completing request directly > ad6: WARNING - SETFEATURES SET TRANSFER MODE taskqueue timeout - > completing request directly > ad6: WARNING - SETFEATURES ENABLE RCACHE taskqueue timeout - > completing request directly > ad6: WARNING - SETFEATURES ENABLE WCACHE taskqueue timeout - > completing request directly > ad6: WARNING - SET_MULTI taskqueue timeout - completing request > directly > ad6: TIMEOUT - READ_DMA retrying (1 retry left) LBA=90088639 > > Looks like the driver may be changing the drive features > configuration for optimization during heavy I/O or maybe a bad > block... You'll let me know. > > Have fun... > > > Brian Candler wrote: > >> On Fri, Mar 10, 2006 at 12:47:07AM -0500, Alain Hebert wrote: >> >> >>> I looked/search everywhere for a thread. I just used the wrong >>> keywords it seems. >>> >>> I'll try -current tomorrow... I hope its stable enought for >>> production. >>> >> >> >> What you want is RELENG_6 (aka 6-STABLE). That gets you tracking the >> code >> base which will end up shortly as 6.1-RELEASE, not the bleeding edge >> -CURRENT. >> >> You can then continue tracking RELENG_6 and you'll incrementally get >> all the >> changes which will end up in 6.2, 6.3 etc. >> >> Regards, >> >> Brian. >> >> >> > -- Alain Hebert ahebert@pubnix.net PubNIX Inc. P.O. Box 175 Beaconsfield, Quebec H9W 5T7 tel 514-990-5911 http://www.pubnix.net fax 514-990-9443