Phil's questions

I want my RewriteRule to not take effect if there is an actual file matching the surfer's request within the file system. So if someone goes to mysite.com/realfile.php?foo=bar, and realfile.php is a real file, the surfer gets served realfile.php?foo=bar. But if someone goes to mysite.com/nonexistentfile.php and nonexistentfile.php is not a real file, then the RewriteRule should interpret and pass the request to process.php

I have the following in my .htaccess:

 RewriteEngine on
 RewriteRule (.*)\.php$ process.php?id=$1 [QSA]

This works IF there is no real file in the file system matching the surfer's request. It does not work if the surfer enters realfile.php?foo=bar...in that case they hit the RewriteRule. It does work if they just do realfile.php, but I want query strings still get passed to the real file if it exists. I thought the local file system always took precedence over RewriteRule?

We have to store basic information about 3 million products. Currently the info is one 180 mb CSV which gets updated quarterly.

There will be about 30,000 queries per day, but the queries are just a very simple key value store. We only need to look up the product ID and display the rest of the information (which would all be in one record).

This is for the web, so fast performance is critical.

Should we use MySQL, even though we really don't need a relational database? Should we just generate 3 million static html files every quarter? Should we store a one line CSV for each product on something like Amazon S3 or Rackspace Cloud Files? What is the best way to do this?

We're using Amazon's 32-bit Linux AMI. What is the best way to automatically run a script after the OS is loaded (not on ssh login, just after it has booted up)?