X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C75F3A.7AC8162C@onstor-exch02.onstor.net>; Mon, 5 Mar 2007 08:25:17 -0700
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C75F3A.7AC8162C"
Content-class: urn:content-classes:message
Subject: RE: mount -sync ?
Date: Mon, 5 Mar 2007 08:25:17 -0700
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04EF30A65@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: mount -sync ?
Thread-Index: AcdfNr+OYv3nwWLQQ1GwdvAzfpCMjQAAhWGk
References: <BB375AF679D4A34E9CA8DFA650E2B04E02A956B2@onstor-exch02.onstor.net>
From: "Huy Duong" <huy.duong@onstor.com>
To: "Steffen Thuemmel" <steffen.thuemmel@onstor.com>,
	"dl-se" <dl-se@onstor.com>,
	"dl-cstech" <dl-cstech@onstor.com>

This is a multi-part message in MIME format.

------_=_NextPart_001_01C75F3A.7AC8162C
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Normally, an NFS client delays sending application write requests, =
allowing application processing to overlap with NFS write operations. An =
NFS client only causes an application to wait for writes to complete =
when the application closes or flushes a file. When a client sends write =
operations synchronously, however, the client causes applications to =
wait for each write operation to complete at the server. This results in =
much lower performance.

I am not familiar with RMAN, however what I know is that for databases =
we generally recommend using the nointr flag, but can anyone confirm =
this? This will allow NFS operations to NOT be interrupted while waiting =
for a response from the server.=20
Does this slow down performance?
=20

-----Original Message-----
From: Steffen Thuemmel
Sent: Mon 3/5/2007 6:58 AM
To: dl-se; dl-cstech
Subject: mount -sync ?
=20
I've been called at a customer site to investigate slow performance.

It turned out to be the mount option "sync" which dropped the =
performance. Using sync on a RH 4 client allows a read performance of =
about 2.5 MB/s. Without that option, one client was around 60 MB/s, with =
two clients I could max one IP port at about 120 MB/s.

=20

I know that "sync" impacts the performance, but I could not imagine that =
the penalty is that high.

Here my questions:

-          Does the "sync" option has any impact on the server (the =
BobCat) ?

-          Is the performance penalty introduced by "sync" dependened on =
the client ? Do you have any experience with other clients ?

=20

The customer needs to use "sync" because they want to backup ORACLE DBs =
using  RMAN to a NFS share. RMAN in ORA 10 requires "sync" mounted =
Filesystems.  Do you have any suggestions for workarounds or =
alternatives ?

=20

Thanks for your help,

=20

St.

=20

Steffen Thuemmel=20

Manager Systems Engineering

Central Europe

=20

ONStor

Schleussner Str. 42

D-63263 Neu-Isenburg

Germany

=20

telf.       +49 6102 884 84-0

mobil.     +49 173 673 3434

mail.       steffen.thuemmel@onstor.com =
<mailto:steffen.thuemmel@onstor.com>=20

web.       www.onstor.com <http://www.onstor.com/>=20

=20



------_=_NextPart_001_01C75F3A.7AC8162C
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
6.5.7652.24">
<TITLE>RE: mount -sync ?</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->

<P><FONT SIZE=3D2>Normally, an NFS client delays sending application =
write requests, allowing application processing to overlap with NFS =
write operations. An NFS client only causes an application to wait for =
writes to complete when the application closes or flushes a file. When a =
client sends write operations synchronously, however, the client causes =
applications to wait for each write operation to complete at the server. =
This results in much lower performance.<BR>
<BR>
I am not familiar with RMAN, however what I know is that for databases =
we generally recommend using the nointr flag, but can anyone confirm =
this? This will allow NFS operations to NOT be interrupted while waiting =
for a response from the server.<BR>
Does this slow down performance?<BR>
<BR>
<BR>
-----Original Message-----<BR>
From: Steffen Thuemmel<BR>
Sent: Mon 3/5/2007 6:58 AM<BR>
To: dl-se; dl-cstech<BR>
Subject: mount -sync ?<BR>
<BR>
I've been called at a customer site to investigate slow performance.<BR>
<BR>
It turned out to be the mount option &quot;sync&quot; which dropped the =
performance. Using sync on a RH 4 client allows a read performance of =
about 2.5 MB/s. Without that option, one client was around 60 MB/s, with =
two clients I could max one IP port at about 120 MB/s.<BR>
<BR>
<BR>
<BR>
I know that &quot;sync&quot; impacts the performance, but I could not =
imagine that the penalty is that high.<BR>
<BR>
Here my questions:<BR>
<BR>
-&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Does the =
&quot;sync&quot; option has any impact on the server (the BobCat) ?<BR>
<BR>
-&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Is the =
performance penalty introduced by &quot;sync&quot; dependened on the =
client ? Do you have any experience with other clients ?<BR>
<BR>
<BR>
<BR>
The customer needs to use &quot;sync&quot; because they want to backup =
ORACLE DBs using&nbsp; RMAN to a NFS share. RMAN in ORA 10 requires =
&quot;sync&quot; mounted Filesystems.&nbsp; Do you have any suggestions =
for workarounds or alternatives ?<BR>
<BR>
<BR>
<BR>
Thanks for your help,<BR>
<BR>
<BR>
<BR>
St.<BR>
<BR>
<BR>
<BR>
Steffen Thuemmel<BR>
<BR>
Manager Systems Engineering<BR>
<BR>
Central Europe<BR>
<BR>
<BR>
<BR>
ONStor<BR>
<BR>
Schleussner Str. 42<BR>
<BR>
D-63263 Neu-Isenburg<BR>
<BR>
Germany<BR>
<BR>
<BR>
<BR>
telf.&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; +49 6102 884 84-0<BR>
<BR>
mobil.&nbsp;&nbsp;&nbsp;&nbsp; +49 173 673 3434<BR>
<BR>
mail.&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; steffen.thuemmel@onstor.com =
&lt;<A =
HREF=3D"mailto:steffen.thuemmel@onstor.com">mailto:steffen.thuemmel@onsto=
r.com</A>&gt;<BR>
<BR>
web.&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; www.onstor.com &lt;<A =
HREF=3D"http://www.onstor.com/">http://www.onstor.com/</A>&gt;<BR>
<BR>
<BR>
<BR>
<BR>
</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C75F3A.7AC8162C--
