asked on

php web cache routine not working quite right

I have some php code, that was developed with help from experts here. The intent of the code was to allow me to use php includes in my web pages to combine common blocks of code which might change from time to time. If I change the individual block/module of HTML (include file), the affecetd source files will auto-adjust and pickup the new HTML. The code is supposed to build an exploded (post include) of itself in a directory (./cache). If the user requests a file on my website, the cache is checked against the dependent modules and if no module is newer, the file is served from the cache. The idea was to reduce the overhead of rebuilding the files with every access.

So, all this works great, most of the time. Problem is occasionally I get the wrong file built with the right filename. i.e. cache/example1.php will end up containing the actual HTML for sample1.php (some different file).

Also, another issue, and perhaps related, when I look in my cache folder, I get a lot of files that should not be there at all. It appears to be limited to files I have linked on teh pages, like pdf files (e.g "myfile1.pdf") . The pdf should not go into the cache folder at all. But, the cache will get apparently random entries that may relate to those linked files. They will have filenames like "myfile1.pdf.php".

I don't know what casues the erroneous behavior, I was hoping you guys could help. I've pasted a sample code sections from one of my source php files below:

This is running on an Apache server FWIW.

########### HEADER ##############
<?php        
// $reqfilename = basename($_SERVER['REQUEST_URI'], ".php");

// Uncomment below to make cache names be unique and allow dup file names across dir's
// $reqfilename = md5($_SERVER['SCRIPT_FILENAME']);
$reqfilename = basename($_SERVER['REQUEST_URI'], ".php");

$sendfile = $_SERVER['SCRIPT_FILENAME'];

$cachefile = "cache/".$reqfilename.".php";

$dependent1 = "includes/body/member.php";
$dependent2 = "includes/navbar.php";

// Serve from the cache if it is the same age or younger than the last 
// modification time of the included file (includes/$reqfilename)

  if (file_exists($cachefile) 
	&& (filemtime($sendfile) < filemtime($cachefile))
	&& (filemtime($dependent1) < filemtime($cachefile))
	&& (filemtime($dependent2) < filemtime($cachefile))
	) {  

    include($cachefile);
    echo "<!-- RequestName: $sendfile -->\n";
    echo "<!-- CacheName: $cachefile -->\n";
    echo "<!-- Cached ".date('F d Y H:i', filemtime($cachefile))."-->\n";
    exit;        
  }

//else
// start the output buffer adn "real" HTML       
ob_start(); 
?>

############### typical includes ####################
<?php include("includes/navbar.php"); ?> 	    
<?php include("includes/body/member.php"); ?>


#####################3 footer #######################3
<?php    
    // PHP, open the cache file for writing
    $fp = fopen($cachefile, 'w');

    // save the contents of output buffer to the file        
    fwrite($fp, ob_get_contents());

    // close the file        
    fclose($fp);

    // Send the output to the browser        
    ob_end_flush();
?>

Open in new window

Ray Paseur

Honestly, unless you can show that the dynamic scripts are so slow that the web site suffers performance problems, I would just abandon the whole cache design pattern. It's really only useful when it is implemented at the level of an opcode cache, or when the scripts that build the pages take very long times to run.

What was the reason for this cache scheme in the first place? Did you have benchmarks that showed the site was suffering bad performance from too much traffic?