I am using the latest dev of jStats.
As soon as I activate jStats, I get this error in my log files for every single site that gets visited

PHP Fatal error: Call to undefined function lock_acquire() in .../includes/bootstrap.inc on line 405, referer: http://www.mydomain.com/node/[different nodes]

When jStats is deactivated, there are no errors.

Comments

DeFr’s picture

DeFr commented
Status: Active » Postponed (maintainer needs more info)
  1. What version of drupal core are you running ?
  2. Did you copy jstats.php to your webroot ? If you did, does the error still occur without copying the file ?
  3. Are you using Drupal standard's database cache backend, or are you using something else (Memcache, MongoDB, ... ?)

(For some reasons, it looks like jStats callback is trying to save something in the cache, but the jstats.php callback only loads things up to the database layer to get more performance, which means that it's failing)

MicS65’s picture

MicS65 commented
Status: Postponed (maintainer needs more info) » Active

> What version of drupal core are you running ?
I am using Drupal 7.12 / pressflow

> Did you copy jstats.php to your webroot ?
Yes

> If you did, does the error still occur without copying the file ?
I removed jstats.php from webroot: no more errors!
Copied it back: again errors!

> Are you using Drupal standard's database cache backend, or are you using something else (Memcache, MongoDB, ... ?)
I am using Drupal standard's database cache backend.
MySQL 5.1.61-0+squeeze1
no Memcache

Other information
Varnish -> Apache 2.2.16 (Debian)

DeFr’s picture

DeFr commented

Thanks for the additionnal informations, I've just reproduced the problem. What's going on is that the drupal_write_record in jstats.inc triggers drupal_get_complete_schema, and DrupalCacheArray::__destruct is trying to save it at the end of the request. Patch coming up shortly to fix it.

DeFr’s picture

DeFr commented
Status: Active » Fixed

Fixed in ad024b4ac8, thanks for the bug report !

There was no fatal error in the 6.x branch, but not loading all the .install file should give a notable performance boost, so I've also backported the change to the 6.x-1.x branch in a8e783a.

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.