[Zope] Will ZEO solve my problem?
Chris McDonough
chrism@digicool.com
Mon, 02 Jul 2001 08:27:48 -0400
Kyler,
You may want to make use of the detailed debug log (see the -M switch in
z2.py) in combination with the requestprofiler.py script that ships
within 2.4.0 and trunk series stuff in the "utilities" directory. There
is some talk about how to use the requestprofiler back a while ago in
this maillist, and searching should give it up.
- C
"Kyler B. Laird" wrote:
>
> [This is a terribly fuzzy question/situation. I'm
> looking for insight/suggestions, not guarantees.]
>
> We've been migrating everything we do to Zope for
> the past few months. We are just starting to get
> buy-in and funding for equipment. Already,
> however, we have lots of people using our Zope
> server which is hosted on a dual P-850 I was able
> to hijack for the purpose. (I found that the Sun
> E450 I had available was way too slow.)
>
> We're planning to split things up into a cluster
> that consists of multiple HTTPS proxies, some ZEO
> clients, a ZEO server and a hot ZEO server spare.
> Until then, I need to make the current system
> usable. (We have all of our Heads and Deans using
> ZWiki Pages to put together our strategic plan.)
>
> The problem we've been encountering is that every
> so often (several times a day) the server goes
> nuts and I end up restarting it. I can watch
> "top" and see a single python process take all of
> one CPU time for an extended period.
>
> I've been unable to determine from the logs (and
> the server-status of the HTTPS proxy) what's
> causing this to happen. Because we have so many
> applications running on this host, the traditional
> debugging methods I've seen don't seem to (easily)
> apply.
>
> Since I'm going to use ZEO eventually, I thought I
> would try it this week to see if it helps with
> this problem. At least I should be able to move
> the problem to either the ZEO server or client.
>
> If it ends up in the ZEO client (as I suspect it
> will), I can use the front-end proxy to route
> requests to other ZEO clients while I restart the
> one that's spinning. I can also restart the
> spinning ZEO client quickly.
>
> If it's in the ZEO server, I'm still screwed. I
> have been avoiding packing the database (so that I
> have the history and ability to undo some of the
> actions), so it takes awhile to start up.
>
> Any thoughts? Is there a way to tell exactly what
> a process is doing (in Zope/Python terms)? I just
> need a clue where to start looking.
>
> Thank you.
>
> --kyler
>
> Some versions:
>
> Mandrake 8.0
> Linux 2.4.4-ac11 (will change on reboot)
> version['ZDChart'] = '0.5.1b'
> version['MSWordMunger'] = '0.1.1'
> version['ZipFolder'] = '0.2.2'
> version['Formulator'] = '0.9.3'
> version['Distutils'] = '1.0.2'
> version['MXDateTime'] = '2.0.1'
> version['psycopg'] = '0.99.2'
> version['HTMLDocument'] = '0.2'
> version['PoPy'] = '2.0.2' #'3.0-beta1'
> version['ZPoPyDA'] = '1.2'
> version['ParsedXML'] = '1.1b1'
> version['PageTemplates'] = '1.3.2'
> version['ZTUtils'] = '1.0.0'
> version['TAL'] = '1.3.2'
> version['OpenLDAP']='1.2.11'
> version['Python']='1.5.2'
> version['Python_lib'] = string.join(string.split(version['Python'], '.')[:2], '.')
> version['PythonLDAP']='1.10alpha3'
> version['Zope']='2.3.3' # '2.3.3b1' # '2.3.2' # '2.3.2b2' # '2.3.2b1' # '2.3.1' # '2.3.0'
> version['ZopeLDAP']='1-1-0'
> version['ZWiki']='0.9.3' #'0.8.1'
> version['CMF']='1.1beta' # '1.0' #'1.0beta'
> version['DCOracle2'] = 'Beta3' # 'beta2'
>
> _______________________________________________
> Zope maillist - Zope@zope.org
> http://lists.zope.org/mailman/listinfo/zope
> ** No cross posts or HTML encoding! **
> (Related lists -
> http://lists.zope.org/mailman/listinfo/zope-announce
> http://lists.zope.org/mailman/listinfo/zope-dev )
--
Chris McDonough Digital Creations, Inc.
Publishers of Zope http://www.zope.org
""" Killing hundreds of birds with thousands of stones """