drupal cache_set / cache_get expiration

Question

I have a function:

function _dc_election_races_get_articles($search_term)
{
    $search_url = 'http://feeds.example.com/vote_up/search.php?search='.rawurlencode($search_term);

    if ($cached = cache_get($search_url, 'cache'))
    {
        $results = $cached->data;
    }
    else
    {
        $results = json_decode(file_get_contents($search_url));
        cache_set($search_url, $results, 'cache', 60*60*3); //3 hours
    }

    return $results;
}

I basically want the cache to expire after 3 hours and re-fetch the data from the url. The documentation on drupal made it seem like that it might not be removed from cache after 3 hours. Will cron cleanup caches?

Answer 1

If you start chasing the code in cache_get you will eventually end up in DrupalDatabaseCache::getMultiple. If you look at the code, you will see a call to DrupalDatabaseCache::garbageCollection at the top. This will handle cache expiration before cron runs.

As part of system_cron you will see that cache_clear_all get called with NULL as the first parameter which

$cid: If set, the cache ID to delete. Otherwise, all cache entries that can expire are deleted.

This cleans up the cache tables every time cron runs.

The Drupal 6 plumbing is different for pruning the cache (start reading cache_get, but the end result is essentially the same.

TL;DR, your cache TTL will always be honored when you get values, it's just the cache table cleanup that may not happen exactly when the TTL is up.