[Zope] Help: Zope is extremely slow
Mayers, Philip J
p.mayers@ic.ac.uk
Wed, 4 Apr 2001 20:22:49 +0100
I'm rather hoping that Postgres isn't the problem, but I can reliably hang
my Zope site using what I think should be fairly reasonable code. Could you
try something for me - make a ZSQL query, a PythonScript, and a DTML
document, and in the pythonscript do this:
print context.standard_html_header(context,context.REQUEST)
for v in context.SQL_query(param1='whatever'):
print context.dtml_document(context,REQUEST,
p1=v.p1,
p2=v.p2,
p3=v.p3,
<etc...>
p13=v.p13,
p14=v.p14,
p15=v.p15,
p16=v.p16)
print context.standard_html_footer(context,context.REQUEST)
return printed
The DTML document just has to contain anything: <P>Placeholder text</P>
Then call the PythonScript (access it's URL) as *fast* as you can (say as
fast a reasonable machine can generate the query). The more fields the SQL
query returns (and thus the more named parameters to the DTML document), the
easier it is to crash.
I get a mix of semi-hangs (child 4 threads dying), hangs (one of the 4 dies,
Zope locks up in handling a SIGSTOP as told by "strace -p pid -d", or
coredumps with sigsegv). And no, there are no hardware problems. This only
happens in very specific circumstances, but I can show you a page where it
happens all the time. *Something* is going on, and it's crippling my
development.
Regards,
Phil
+----------------------------------+
| Phil Mayers, Network Support |
| Centre for Computing Services |
| Imperial College |
+----------------------------------+
-----Original Message-----
From: Steve Drees [mailto:drees@the-bridge.net]
Sent: 04 April 2001 20:15
To: Mayers, Philip J; 'Daniel Rusch'; zope@zope.org
Subject: RE: [Zope] Help: Zope is extremely slow
> That sounds suspiciously like the "Zope dying" problem that a lot
> of people
> seem to be having. You aren't by any chance using Postgresql are you?
Postgres is not necessarily the problem.
Zope is solid as a rock at www.wejustrock.com and it is using postgresql.