X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C7334D.5E15D14C@onstor-exch02.onstor.net>; Mon, 8 Jan 2007 09:49:38 -0800
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C7334D.5E15D14C"
Content-class: urn:content-classes:message
Subject: /usr/bin/install the problem
Date: Mon, 8 Jan 2007 09:49:38 -0800
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E0A90F2@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: /usr/bin/install the problem
thread-index: AcczSqTG+wOnVBo6TSWQsxObAXBhMA==
From: "Larry Scheer" <larry.scheer@onstor.com>
To: "Andy Sharp" <andy.sharp@onstor.com>
Cc: "Tim Gardner" <tim.gardner@onstor.com>,
	"Jay Michlin" <jay.michlin@onstor.com>

This is a multi-part message in MIME format.

------_=_NextPart_001_01C7334D.5E15D14C
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Andy,
   Over the weekend I ran continuous tests of upgrade and upgrade =
simulations using cp, tar, and /usr/bin/install with and without the -S =
flag. I also removed the -S (safe option) flag to the install command in =
the ugrade program and tested it. When using the install command I see =
file corruption about 10-20% of the time. It usually occurs when the =
system starts using swap space and nears the limits of memory.

I would like your opinion on how to proceed.=20

Changing the upgrade program to use cp -p might be the quickest thing =
but some work would need to be done to make sure the permissions are the =
desired ones when the tar file is created, but this shouldn't be hard to =
do.

I could download the latest version of /usr/bin/install from the BSD =
distribution and build it into our version of OpenBSD but there is no =
guarantee this will change anything. Actually I have a version of 3.8 =
OpenBSD in one of my directories today that I can use to replace the =
version we are using. I can build this and test it pretty quickly.

The other option is diving into the sources to figure out where and why =
the corruption is occurring. This would most likely need your expertise =
with the kernel and memory management to figure out and fix. This is =
most likely the "most correct" approach because we would now fully =
understand the issue but may also be time consuming.

I could pursue the other two approaches while you investigate the BSD =
code.

Let me know what you think.

Larry=20



------_=_NextPart_001_01C7334D.5E15D14C
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 =
6.5.7650.28">
<TITLE>/usr/bin/install the problem</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->

<P><FONT SIZE=3D2>Andy,<BR>
&nbsp;&nbsp; Over the weekend I ran continuous tests of upgrade and =
upgrade simulations using cp, tar, and /usr/bin/install with and without =
the -S flag. I also removed the -S (safe option) flag to the install =
command in the ugrade program and tested it. When using the install =
command I see file corruption about 10-20% of the time. It usually =
occurs when the system starts using swap space and nears the limits of =
memory.<BR>
<BR>
I would like your opinion on how to proceed.<BR>
<BR>
Changing the upgrade program to use cp -p might be the quickest thing =
but some work would need to be done to make sure the permissions are the =
desired ones when the tar file is created, but this shouldn't be hard to =
do.<BR>
<BR>
I could download the latest version of /usr/bin/install from the BSD =
distribution and build it into our version of OpenBSD but there is no =
guarantee this will change anything. Actually I have a version of 3.8 =
OpenBSD in one of my directories today that I can use to replace the =
version we are using. I can build this and test it pretty quickly.<BR>
<BR>
The other option is diving into the sources to figure out where and why =
the corruption is occurring. This would most likely need your expertise =
with the kernel and memory management to figure out and fix. This is =
most likely the &quot;most correct&quot; approach because we would now =
fully understand the issue but may also be time consuming.<BR>
<BR>
I could pursue the other two approaches while you investigate the BSD =
code.<BR>
<BR>
Let me know what you think.<BR>
<BR>
Larry<BR>
<BR>
<BR>
</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C7334D.5E15D14C--
