Potential edge case with page-level caching.

For anonymous users, 'cache_get' is domain specific. That is, the entire URL is used as the CID.

SELECT data, created, headers, expire FROM cache_page WHERE cid = 'http://ken.test/drupal-5.3/?q=node/22'

This might create a condition where multiple cached versions of the same page exist, but only one gets updated. We may have to manually clear out the cache using hook_nodeapi.

Investigating.

Comments

agentrickard’s picture

Priority: Normal » Critical
agentrickard’s picture

Status: Active » Closed (fixed)

This appears to be a non-issue.

Drupal calls cache_clear_all() on node save, which should wipe the page cache appropriately.