I believe you can stop this just by changing the PCGI_PUBLISHER variable in your Zope.cgi to a directory, /dev/null or something. We do this for all our site, using the mod_pcgi apache module. It works fairly well... at least well enough for us not to consider changing yet. Michael Olivier wrote:
At 07:33 PM 3/23/2001 -0800, sam gendler wrote:
I haven't used Zope to any significant degree in over a year. When I was last using Zope regularly, FastCGI still had some stability issues, which made it difficult to use in a production environment. However, I have always felt that FastCGI was the better solution when choosing between proxy, pcgi and fastcgi, from a theoretical standpoint. Can someone comment as to the efficacy of using FastCGI to implement an interface between my HTTP enabled application and current version of Zope, or should I use one of the other solutions.
Hi Sam and others --
I was about to ask the same question myself. I'm also wondering about switching solutions to FastCGI or proxy. I read on the zope web site that zope.org is using the proxy approach. Can someone in-the-know please shed some light?
I'm using PCGI (2.0a4) with ZPublisher and while it's worked reasonably well for very small traffic levels, I have recently run into an issue where lots of extra pcgi_publisher.py processes get started until the system grinds to a halt. I found two old postings from others (John Eikenberry, Craig Allen) about this problem, including this one:
http://classic.zope.org/pipermail/zope/1999-April/003926.html
... and am wondering if it's been solved and if there's a later release (Zope 2.3 stable has pcgi 2.0a4 still).
Thanks! Michael
-- John Eikenberry [jae@kavi.com] ______________________________________________________________ "A society that will trade a little liberty for a little order will deserve neither and lose both." --B. Franklin