[Zope] Report on Solaris Zope performance available

oliver.erlewein@sqs.de oliver.erlewein@sqs.de
Thu, 21 Mar 2002 15:03:25 +0100


This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C1D0E1.2AFFC9F0
Content-Type: text/plain;
	charset="iso-8859-1"

(possible repost. I sent from oerlewein@gmx.de but that didn't reach the
list!)

Hi Tony

I saw the report yesterday and read it immediately. It is an EXCELLENT
summary of whats been discussed the last year or so! The explanations are
VERY
understanable and allow for easy catching up on the subject. I'm very glad
that
the report has been written.

At this point (although it's a little late - diddn't know that you were
writing up a document) I'd like to offer my time as review partner if there
is
something more coming on threading. 

There are some points in the document that I don't quite agree to. I'll
adress these as soon as I have time for it (I'll have to do that on my
private
time).

This morning I reconfigured my testcomputer with the infos from the report
and I still have the same problem. Could you add the exact compile options
for
Python and the setup information for Zope?

I tried this morning to compile with SUN Forte (cc). Solaris 2.8-sparc,
Python 2.1.2:

export LD_LIBRARY_PATH=/usr/lib/lwp
export CC='cc -Xc -mt -lpthread -R /usr/lib/lwp'
./configure --with-threads --without-gcc --with-cxx=CC
make
make install

There were sone errors that came up on the *Splitter* files. There you just
have to delete some unimportant comments (or ^M's) and then all goes well. 

"Recompiled" Zope (2.4.4b1) with python wo_pcgi.py

I still have the "single process" thread. I'll have another go with gcc
today or tomorrow. I'll post the results. I'll compile with:

export LD_LIBRARY_PATH=/usr/lib/lwp
./configure --with-threads
make
make install

****
What the decision for or against Zope is concerned I don't quite understand
the problem. The question here is Solaris against Linux, FreeBSD,...
The Zope yes/no question is something completely different. The chain of
argunents used for the decision makers must be clear on that! These TWO
subjects
only mingle when the company/varsity is fixed on Solaris as only OS.
****

Regards Oliver Erlewein

------_=_NextPart_001_01C1D0E1.2AFFC9F0
Content-Type: text/html;
	charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2653.12">
<TITLE>Re: [Zope] Report on Solaris Zope performance available</TITLE>
</HEAD>
<BODY>

<P><FONT SIZE=2>(possible repost. I sent from oerlewein@gmx.de but that didn't reach the list!)</FONT>
</P>

<P><FONT SIZE=2>Hi Tony</FONT>
</P>

<P><FONT SIZE=2>I saw the report yesterday and read it immediately. It is an EXCELLENT</FONT>
<BR><FONT SIZE=2>summary of whats been discussed the last year or so! The explanations are</FONT>
<BR><FONT SIZE=2>VERY</FONT>
<BR><FONT SIZE=2>understanable and allow for easy catching up on the subject. I'm very glad</FONT>
<BR><FONT SIZE=2>that</FONT>
<BR><FONT SIZE=2>the report has been written.</FONT>
</P>

<P><FONT SIZE=2>At this point (although it's a little late - diddn't know that you were</FONT>
<BR><FONT SIZE=2>writing up a document) I'd like to offer my time as review partner if there</FONT>
<BR><FONT SIZE=2>is</FONT>
<BR><FONT SIZE=2>something more coming on threading. </FONT>
</P>

<P><FONT SIZE=2>There are some points in the document that I don't quite agree to. I'll</FONT>
<BR><FONT SIZE=2>adress these as soon as I have time for it (I'll have to do that on my</FONT>
<BR><FONT SIZE=2>private</FONT>
<BR><FONT SIZE=2>time).</FONT>
</P>

<P><FONT SIZE=2>This morning I reconfigured my testcomputer with the infos from the report</FONT>
<BR><FONT SIZE=2>and I still have the same problem. Could you add the exact compile options</FONT>
<BR><FONT SIZE=2>for</FONT>
<BR><FONT SIZE=2>Python and the setup information for Zope?</FONT>
</P>

<P><FONT SIZE=2>I tried this morning to compile with SUN Forte (cc). Solaris 2.8-sparc,</FONT>
<BR><FONT SIZE=2>Python 2.1.2:</FONT>
</P>

<P><FONT SIZE=2>export LD_LIBRARY_PATH=/usr/lib/lwp</FONT>
<BR><FONT SIZE=2>export CC='cc -Xc -mt -lpthread -R /usr/lib/lwp'</FONT>
<BR><FONT SIZE=2>./configure --with-threads --without-gcc --with-cxx=CC</FONT>
<BR><FONT SIZE=2>make</FONT>
<BR><FONT SIZE=2>make install</FONT>
</P>

<P><FONT SIZE=2>There were sone errors that came up on the *Splitter* files. There you just</FONT>
<BR><FONT SIZE=2>have to delete some unimportant comments (or ^M's) and then all goes well. </FONT>
</P>

<P><FONT SIZE=2>&quot;Recompiled&quot; Zope (2.4.4b1) with python wo_pcgi.py</FONT>
</P>

<P><FONT SIZE=2>I still have the &quot;single process&quot; thread. I'll have another go with gcc</FONT>
<BR><FONT SIZE=2>today or tomorrow. I'll post the results. I'll compile with:</FONT>
</P>

<P><FONT SIZE=2>export LD_LIBRARY_PATH=/usr/lib/lwp</FONT>
<BR><FONT SIZE=2>./configure --with-threads</FONT>
<BR><FONT SIZE=2>make</FONT>
<BR><FONT SIZE=2>make install</FONT>
</P>

<P><FONT SIZE=2>****</FONT>
<BR><FONT SIZE=2>What the decision for or against Zope is concerned I don't quite understand</FONT>
<BR><FONT SIZE=2>the problem. The question here is Solaris against Linux, FreeBSD,...</FONT>
<BR><FONT SIZE=2>The Zope yes/no question is something completely different. The chain of</FONT>
<BR><FONT SIZE=2>argunents used for the decision makers must be clear on that! These TWO</FONT>
<BR><FONT SIZE=2>subjects</FONT>
<BR><FONT SIZE=2>only mingle when the company/varsity is fixed on Solaris as only OS.</FONT>
<BR><FONT SIZE=2>****</FONT>
</P>

<P><FONT SIZE=2>Regards Oliver Erlewein</FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C1D0E1.2AFFC9F0--