NGINX rewrite rules for PATH_INFO and SCRIPT_NAME

We use nginx as our web server on all our servers here at Em Space as it does a great job of serving web sites with very little overhead, essential on low memory virtual servers.

We have a stock virtual host file which serves Drupal sites very nicely (I'll post that sometime) but Josh needed to install a PHP based OAI2 harvester application which uses the server variable PATH_INFO for its query handling. Our default Drupal based rewrite and fastcgi rules didn't work, so we set about figuring out how to fix them for this site.

This is what we came up with:

server {
    listen 80;
    server_name doaih.urbits.com;
 
    access_log	/home/swin/oaih/dev/logs/access.log;
    error_log /home/swin/oaih/dev/logs/error.log;
 
    root   /home/swin/oaih/dev/public;
    index  index.php index.html index.htm;
 
    location / {
        if (!-e $request_filename) {
            rewrite ^/(.*)$ /index.php/$1 last;
        }
    }
 
    location ~ .php$ {
        fastcgi_pass   127.0.0.1:9000;
        fastcgi_index  index.php;
        fastcgi_param  SCRIPT_FILENAME  $document_root$fastcgi_script_name;
 
        include /etc/nginx/fastcgi_params;
    }
 
    location ~ \.php($|/) {
 
        set $script     $uri;
        set $path_info  "";
 
        if ($uri ~ "^(.+\.php)(/.+)") {
            set $script     $1;
            set $path_info  $2;
        }
 
        fastcgi_pass   127.0.0.1:9000;
        fastcgi_index  index.php;
        fastcgi_param  SCRIPT_FILENAME    $document_root$script;
        fastcgi_param  SCRIPT_NAME        $script;
        fastcgi_param  PATH_INFO          $path_info;
 
        fastcgi_param  QUERY_STRING       $query_string;
        fastcgi_param  REQUEST_METHOD     $request_method;
        fastcgi_param  CONTENT_TYPE       $content_type;
        fastcgi_param  CONTENT_LENGTH     $content_length;
 
        fastcgi_param  REQUEST_URI        $request_uri;
        fastcgi_param  DOCUMENT_URI       $document_uri;
        fastcgi_param  DOCUMENT_ROOT      $document_root;
        fastcgi_param  SERVER_PROTOCOL    $server_protocol;
 
        fastcgi_param  GATEWAY_INTERFACE  CGI/1.1;
        fastcgi_param  SERVER_SOFTWARE    nginx;
 
        fastcgi_param  REMOTE_ADDR        $remote_addr;
        fastcgi_param  REMOTE_PORT        $remote_port;
        fastcgi_param  SERVER_ADDR        $server_addr;
        fastcgi_param  SERVER_PORT        $server_port;
        fastcgi_param  SERVER_NAME        $server_name;
        fastcgi_param  REDIRECT_STATUS    200;
 
    }
 
}

The first location directive sets up the base options while the second location directive matches a request to index.php on its own and passes it to fastcgi to handle. This was needed as requests for the domain without index.php were not being handled correctly by the next block.

The key section in the whole config file for dealing with our problem is this:

    location ~ \.php($|/) {
 
        set $script     $uri;
        set $path_info  "";
 
        if ($uri ~ "^(.+\.php)(/.+)") {
            set $script     $1;
            set $path_info  $2;
        }
 
        fastcgi_pass   127.0.0.1:9000;
        fastcgi_index  index.php;
        fastcgi_param  SCRIPT_FILENAME    $document_root$script;
        fastcgi_param  SCRIPT_NAME        $script;
        fastcgi_param  PATH_INFO          $path_info;
    }

This location directive matches any PHP requests not previously matched and sets the $script variable to the full uri. It also empties the $path_info variable. The if statement matches any requests like the following: and puts index.php into the $script variable and /foo into the $path_info variable.

$script and $path_info are inserted into the server variables by the fastcgi parameters SCRIPT_NAME and PATH_INFO and allow PHP to properly process the requests.