Re: SUEXEC (but why one instance)
David Birnbaum <[email protected]>
| Newsgroups | gmane.comp.web.fastcgi.devel |
|---|---|
| Message-ID | <[email protected]> |
Michael, (and Rob), One thing I note in that FAQ entry - PHP now runs in both FastCGI and non-FastCGI mode without a problem, so that FAQ entry should be updated so there's only one binary created by default. No need for two anymore. David. ----- On Mon, 29 Nov 2004, David Birnbaum wrote: > Michael, > > PHP uses its own process manager, so it won't work in any sort of normal > FastCGI sense, as best as I can figure. Here's what worked for me: > > http://www.fastcgi.com/docs/faq.html#PHP > > David. > > ----- > > On Tue, 30 Nov 2004, Michael wrote: > >> Apologies for my previous set of e-mails (as I just keep developing new >> conclusions on this one) >> >> By removing the 'FastCgiWrapper /usr/local/apache2/bin/suexec', all seems >> to be >> well. I suppose suexec is stripping more than I want it to in this case. >> I now >> know where the problem lies and will investigate further. >> >> With Suexec out of the way, the variables are being passed to my program >> through >> the environment, so I know now that the setting is taking effect. However, >> I'm >> curious as to why, despite sending 20,000 request at the Apache web server >> with >> a concurrency of 35, I still only have one responder process. >> >> httpd 0:01 /usr/local/apache2/bin/fcgi- >> httpd 0:21 /var/www/cgi-sys/php >> >> Shouldn't there be more of them being scaled up to the load? Only one for >> this >> server? My fastCGI config is as follows: >> >> FastCgiConfig -restart-delay 3 -restart -initial-env >> PHP_FCGI_MAX_REQUESTS=100000 -initial-env TESTING=1 -maxClassProcesses 100 >> -updateInterval 5 -killInterval 5 >> >> Shouldn't this be starting up as many as needed, up to 100 php >> interpreters? >> >> -M >> >> > ___________________________________ fastcgi-developers mailing list http://fastcgi.com/fastcgi-developers/