The key to changing web server identity is consistency. The trouble we went through to change the web server make may be useless if we leave the default Apache content around. The removal of the default content is equivalent to changing one’s clothes when going undercover. This action may be useful even if we do not intend to change the server identity. Applications often come with sample programs and, as a general rule, it is a good practice to remove them from production systems; they may contain vulnerabilities that may be exploited later.
Most of the default content is out of reach of the public, since we have built our Apache from scratch, changed the root folder of the web site, and did not include aliases for the manual and the icons. Just to be thorough, erase the following directories:
/usr/local/apache/manual (Apache 2 only)
You will probably want to keep the original /usr/local/apache/logs directory though the logs are stored in /var/www/logs. This is because many modules use the logs/ folder relative to the Apache installation directory to create temporary files. These modules usually offer directives to change the path they use, but some may not. The only remaining bit of default content is the error pages Apache displays when errors occur. These pages can be replaced with the help of theErrorDocumentdirective. Using one directive per error code, replace the error pages for all HTTP error codes. (A list of HTTP codes is given in Chapter 8; it can also be found at http://www.w3.org/ Protocols/rfc2616/rfc2616-sec10.html.)
An alternative to creating dozens of static pages is to create one intelligent script that retrieves the error code from Apache and uses it to display the appropriate message. A small bit of programming is required in this case, following guidance from the Apache documentation at http://httpd.apache.org/docs-2.0/custom-error.html.