Building a system to easily manage 100's of sites

My very first steps towards multi-tenancy

[le_gber]

Hi guys,

I am currently tinkering with the following idea:

I want to create a "core" site with a dozen pages that can be re-purposed and customized.

For each re-purposed site I am then thinking of including or requiring the page dynamically if a local copy of the page doesn't exist.

The url would look like:
www.repurposed.com/index.php?dir=about&page=index

if local /about/index.php file doesn't exists
include core file
if local /about/index.php file exists
include local copy

(openbasedir and proper validation of the $dir and $page using switch(){case:} is assumed)

Following on that idea, I also thought of doing a php include in the .js file for script used across all those sites (like carousel) using something like:


<FilesMatch "\.(js)$">
SetHandler application/x-httpd-php
</FilesMatch>

Do you guys see any reasons why I shouldn't do any of the above? Any flaws in my plan? Should I use a database instead to hold all the files? Am I right in thinking that including files is faster and put less stress on the server (but maybe not safer)?

[DrDoc]

I assume duplicate content is not a concern?

[le_gber]

Hi DrDoc, no it isn't, those pages would be hidden from SE and the site would have more "unique" pages.

[DrDoc]

Personally, I'm a proponent for database driven sites. But redirects would work as well ... You can simply check if the file exists (

-f

), or else load a different file internally.

I've done that on image servers to load a "no-image" image.

[phranque]

if your solution is to have the php script do a HTTP GET Request to gather content for your reponse, that's going to be slow for the visitor.

a file include should be no problem.

[le_gber]

@DrDoc, by doing a file exist (-f) do you mean in htaccess rather than PHP?
@phranque, the HTTP GET is only used to know which page to serve, does this still slow things down? I was thinking of mode_rewriting everything to make the URL look nicer like:

www.repurposed.com/index.php?dir=about&page=index
to
www.repurposed.com/about/index.php

[phranque]

the HTTP GET is only used to know which page to serve, does this still slow things down?

if your php script is doing a separate roundtrip to another server before responding, yes.

www.repurposed.com/about/index.php

use the DIrectoryIndex directive to set the default directory index document to index.php and use this url instead:

www.repurposed.com/about/

[le_gber]

if your php script is doing a separate roundtrip to another server before responding, yes.

- it wouldn't, the "core" site and the "repurposed" site would be reference with paths from Apache /home/

use the DIrectoryIndex directive to set the default directory index document to index.php and use this url instead:
www.repurposed.com/about/

that's my rewrite at the moment, will it simply ignore the second variable if not defined?

RewriteEngine On
RewriteRule ^([^/]*)/([^/]*)\.htm$ /index.php?section=$1&amp;page=$2 [L]

[phranque]

RewriteEngine On
RewriteRule ^([^/]*)/([^/]*)\.htm$ /index.php?section=$1&amp;page=$2 [L]

you don't need to use an html entity for anything in the url here.

RewriteEngine On
RewriteRule ^([^/]*)/([^/]*)\.htm$ /index.php?section=$1&page=$2 [L]

the only way to match the pattern and get an undefined $2 is to request something like:
www.example.com/something/.htm

but yes - that would internally rewrite to:
/index.php?section=something&page=

[le_gber]

ha, yes I forgot, I was trying to "hide" PHP by having .htm extensions and enabling PHP to run on htm files :)

Thanks for your help.