From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay2.corp.sgi.com [137.38.102.29]) by oss.sgi.com (Postfix) with ESMTP id BB5CA7CA0 for ; Tue, 9 Aug 2016 10:45:53 -0500 (CDT) Received: from cuda.sgi.com (cuda2.sgi.com [192.48.176.25]) by relay2.corp.sgi.com (Postfix) with ESMTP id 893D4304062 for ; Tue, 9 Aug 2016 08:45:50 -0700 (PDT) Received: from nm15-vm4.bullet.mail.sg3.yahoo.com (nm15-vm4.bullet.mail.sg3.yahoo.com [106.10.149.51]) by cuda.sgi.com with ESMTP id RJ6nhR0pfrFBEbEf (version=TLSv1.2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128 verify=NO) for ; Tue, 09 Aug 2016 08:45:48 -0700 (PDT) Date: Tue, 9 Aug 2016 15:43:03 +0000 (UTC) From: Gim Leong Chin Message-ID: <291675510.10842010.1470757383533.JavaMail.yahoo@mail.yahoo.com> In-Reply-To: <7faf941c-2188-589d-b624-9fedb22dee40@telefonica.net> References: <5787852A.7030900@st-andrews.ac.uk> <20160808161132.1d76eb5c@harpe.intellique.com> <1470670691.8958.11.camel@filmlight.ltd.uk> <20160808174427.24d8272c@harpe.intellique.com> <592673683.10446004.1470715375850.JavaMail.yahoo@mail.yahoo.com> <7faf941c-2188-589d-b624-9fedb22dee40@telefonica.net> Subject: Re: Advice needed with file system corruption MIME-Version: 1.0 Reply-To: Gim Leong Chin List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: multipart/mixed; boundary="===============8451113948269555530==" Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: "Carlos E. R." , XFS mail list --===============8451113948269555530== Content-Type: multipart/alternative; boundary="----=_Part_10842009_715577767.1470757383530" ------=_Part_10842009_715577767.1470757383530 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On 2016-08-09 06:02, Gim Leong Chin wrote: >> Drives connected to RAID controllers with battery backed cache should >> have their caches "disabled" (they are really set to write through mode >> instead).=C2=A0 By the way, I found out in lab testing that 7200 RPM SAT= A >> drives suffer a big performance loss when doing sequential writes in >> cache write through mode.< > If you disable the disk internal cache, as a consequence you also > disable the disk internal write optimizations. It has to be much slower > at writing. It seems to me obvious. > --=20 > Cheers / Saludos, =C2=A0> =C2=A0 =C2=A0=C2=A0=C2=A0 Carlos E. R. =C2=A0> =C2=A0 =C2=A0=C2=A0=C2=A0 (from 13.1 x86_64 "Bottle" at Telcontar) The drop in sequential write data rate for 3.5" 7200 RPM SATA drives was ar= ound 50%, I cannot remember the exact numbers, that is not obvious to me. As a reminder, the drive cache is really set to write through mode, it is n= ot possible to disable the cache, as an application engineer from HGST told= me, so the drive internal write optimizations are still there, just that t= he IO command is reported to be completed only when the data has been write= n to the drive platter. 10k and 15k RPM SAS drives connected to LSI Internal RAID controllers have = their drive cache "disabled" automatically, I wonder how much is the data r= ate drop compared to drive cache "enabled", considering that LSI IR control= lers do not have cache. GL ------=_Part_10842009_715577767.1470757383530 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
On 2016-08-09 06:02= , Gim Leong Chin wrote:



=
>> Drives connected to RAID controllers with batte= ry backed cache should
>> have their caches "disabl= ed" (they are really set to write through mode
>> i= nstead).  By the way, I found out in lab testing that 7200 RPM SATA>> drives suffer a big performance loss when doing se= quential writes in
>> cache write through mode.<=

> If you disable the disk internal cache, as a = consequence you also
> disable the disk internal write= optimizations. It has to be much slower
> at writing.= It seems to me obvious.

> --
> Cheers / Saludos,

&= nbsp;>       Carlos E. R.
 >     &= nbsp; (from 13.1 x86_64 "Bottle" at Telcontar)


The drop in sequential write data rate for 3.5" 7200 RPM SATA d= rives was around 50%, I cannot remember the exact numbers, that is not obvi= ous to me.

As a reminder, the drive cache is really= set to write through mode, it is not possible to disable the cache, as an = application engineer from HGST told me, so the drive internal write optimiz= ations are still there, just that the IO command is reported to be complete= d only when the data has been writen to the drive platter.

10k and 15k RPM SAS drives connected to LSI Internal RAID controlle= rs have their drive cache "disabled" automatically, I wonder how much is th= e data rate drop compared to drive cache "enabled", considering that LSI IR= controllers do not have cache.

GL
=
------=_Part_10842009_715577767.1470757383530-- --===============8451113948269555530== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs --===============8451113948269555530==--