From: Sean Chittenden <sean@...>
Date: 2002-04-15T04:20:09+09:00
Subject: Re: ruby-run in httpd.conf : WAS (Re: Switching from PHP to Ruby - Comments Please)

> > > Ok, just a few more questions still.
> > 
> > > What is the ruby-run needed for?
> > 
> > A better question would be, why is the RubyRequire needed?  If you
> > use RubyRequire in the httpd.conf, the parent apache process will
> > readin in apache/ruby-run.  The advantage of this comes from copy
> > on write memory pages.  So long as the child apache processes
> > don't write to the memory page where apache/ruby-run is located,
> > they will all share the same page of data which can dramatically
> > save on RAM consumption.  If you ommit the RubyRequire, I'm sure
> > the ruby-object or Apache::RubyRun handler will do the appropriate
> > require, however, the result of the require won't be cached or
> > "shared" among the other child processes.  Sooooo.....  after that
> > long winded response, the short response is, "memory management of
> > the webserver."
>
> On a related note, after setting this all up and testing it, I
> checked for a running ruby interpretor process, but did not find
> one. I kind of expected to see one running all the time with
> mod_ruby, unlike cgi or fcgi.

Nope.  The ruby interpreter is embedded inside of the httpd process
and not external to apache.  If you look at the size of the httpd
processes you should be able to see the difference there (and don't
forget to take into account shared versus resident memory).  -sc

-- 
Sean Chittenden