From owner-freebsd-scsi Sat Apr 24 18:46:32 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from panzer.plutotech.com (panzer.plutotech.com [206.168.67.125]) by hub.freebsd.org (Postfix) with ESMTP id 988BB14D53 for ; Sat, 24 Apr 1999 18:46:30 -0700 (PDT) (envelope-from ken@panzer.plutotech.com) Received: (from ken@localhost) by panzer.plutotech.com (8.9.3/8.8.5) id TAA48692; Sat, 24 Apr 1999 19:46:17 -0600 (MDT) From: "Kenneth D. Merry" Message-Id: <199904250146.TAA48692@panzer.plutotech.com> Subject: Re: writing slower with SoftUpdates In-Reply-To: <199904222241.SAA96017@misha.cisco.com> from Mikhail Teterin at "Apr 22, 1999 6:41:37 pm" To: mi@aldan.algebra.com Date: Sat, 24 Apr 1999 19:46:17 -0600 (MDT) Cc: dnelson@emsphone.com, freebsd-scsi@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL43 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Mikhail Teterin wrote... > Dan Nelson once wrote (regarding iozone reporting ~10% slower results > with SoftUpdates, compared to without -- sync mount): > > > > WCE was ON on both drives: > > > > > > Checking SCSI drives for write-cache: > > > at scbus0 target 8 lun 0 (pass0,da0) > > > at scbus0 target 9 lun 0 (pass1,da1) > > > > I disabled WCE on both and re-ran the iozone tests. No noticable > > > changes from the previous numbers... What really bothers me, is the > > > fact that reading speed speed is consistently only 13-14Mb/s on one > > > disk, while it is 17-18 (same as writing) on the other. Can it be, > > > that one of the partitions I'm trying this on is /tmp (the other is > > > /var/tmp)? There are no files in either one of them and they are of > > > the same size and offset on two identical disks. > > > The different mountpoint shouldn't make any difference. I guess you > > could dump the bad-block table for both drives, and see if your slower > > disk has more relocated blocks than the faster one. "camcontrol > > defects -u 0 -f block" will dump the blocklist. > > (pass0:ahc0:0:8:0): READ DEFECT DATA(10). CDB: 37 0 0 0 0 0 0 fd e8 0 > (pass0:ahc0:0:8:0): RECOVERED ERROR asc:1c,0 > (pass0:ahc0:0:8:0): Defect list not found field replaceable unit: 18 sks:ca,2 Seagate drives generally don't support the block defect list format, and the error they spit out to tell you that isn't exactly intuitive. Try the physical sector format: camcontrol defects -u 0 -f phys -G To get the permanant defect list: camcontrol defects -u 0 -f phys -G To get both: camcontrol defects -u 0 -f phys -PG > > If you still can't find the problem, try asking in the freebsd-scsi > > mailinglist. > > CC-ed there. The machine is 350MHz PentumII with 128Mb of 100MHz RAM. > The disks are LVD being the only targets on the LVD outlet of Adaptec > 2940U2W. If I read your comments above correctly, it appears that you've already tried enabling/disabling write caching. That can certainly have an effect on things. One other thing to look at is where you are on the disk. I/O performance will be better on the outer diameter tracks. Ken -- Kenneth Merry ken@plutotech.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message