Running the Drupal cron job and flushing cache from the Linux cron

Working with Drupal 7 our default initial setup was running Drupal’s cron using the build-in cron mechanism (poorman’s cron). There are several downsides to this, mainly:

Cron will run only after a page hit. If there’s no traffic the cron is not triggered.
The addition of cron to a page load adds time and resources required for the request – slow response and might even exhaust resources.

So moving to the server’s cron was obvious.
First I tried the default setup as recommended and tried using wget / curl / lynx but it was not triggering the cron. You can when cron was last executed by going to the cron page in Drupal’s admin -> Configuration -> Cron.

After checking what was going on I found out that cron was running only for logged in users, not anonymous. So I had to emulate an admin user, run cron and watch for security holes which might grant admin rights to hackers. I came up with the following script:

<?php
if ($argv[1]=='sofij398uw8wf') {
	//set the working directory 
	chdir('/var/www/example.com/html/folder/');	
	define('DRUPAL_ROOT', getcwd());
	$_SERVER['REMOTE_ADDR'] = '127.0.0.1';
	$base_url = 'https://www.example.com/folder';
	include_once './includes/bootstrap.inc';
	drupal_bootstrap(DRUPAL_BOOTSTRAP_FULL);
	global $user;
	$original_user = $user; 
	$old_state = drupal_save_session(); 
	drupal_save_session(FALSE); 
	$user = user_load(1);
	// Take your action here where you pretend to be the user with UID = 1 (typically the admin) 
	// If your code fails, it's not a problem because the session will not be saved 
	drupal_cron_run();
	//drupal_flush_all_caches(); //in case we want to wipe the cache
	$user = $original_user; 
	drupal_save_session($old_state); 
	// From here on the $user is back to normal so it's OK for the session to be saved
}
?>

<?php

if ($argv[1]=='sofij398uw8wf') {

//set the working directory

chdir('/var/www/example.com/html/folder/');

define('DRUPAL_ROOT', getcwd());

$_SERVER['REMOTE_ADDR'] = '127.0.0.1';

$base_url = 'https://www.example.com/folder';

include_once './includes/bootstrap.inc';

drupal_bootstrap(DRUPAL_BOOTSTRAP_FULL);

global $user;

$original_user = $user;

$old_state = drupal_save_session();

drupal_save_session(FALSE);

$user = user_load(1);

// Take your action here where you pretend to be the user with UID = 1 (typically the admin)

// If your code fails, it's not a problem because the session will not be saved

drupal_cron_run();

//drupal_flush_all_caches(); //in case we want to wipe the cache

$user = $original_user;

drupal_save_session($old_state);

// From here on the $user is back to normal so it's OK for the session to be saved

}

A few notes about the above code:

I’m using $argv to read passed parameters since I’m running this through command line php:

php -q cronit.php sofij398uw8wf

1

php -q cronit.php sofij398uw8wf
The script didn’t execute using wget / curl (adjusted). It only works running in php as above.
The parameter used (‘sofij398uw8wf’) is used as a key to avoid unwanted visitors.
The user handling (switching to admin and back) is also a security measure to avoid granting admin permissions to unwanted visitors.
There is a line for flushing all cache:

drupal_flush_all_caches();

1

drupal_flush_all_caches();

it’s commented out. I recommend running it in its own cron job, separating between the two.

If you are less familiar with setting a cron job in Linux please read my post – Setting a cron job in Linux

For help setting up your cron jobs please comment below or contact me using the form to the right.

Leave a Reply Cancel reply