Details
-
Bug
-
Resolution: Fixed
-
High
-
4.7.0, 5.0, 5.1, 5.2
-
eZ Publish 4.7 EE with eZDFS cluster
Description
Hi,
We have noticed a cache problem on eZDFS environments when you run:
php bin/php/ezcache.php --clear-id=ezjscore-packer
In that case the expired flag index dfs table isn't set to "1". It stays at 0.
Executing the ezcache.php with the --purge option does seems to work as intended (files are cleared).
Further info:
Enabling kernel-clustering debug condition and executing the ezcache script with debug yelds the following info:
Debug: (<kernel-clustering>) dfs::ctor( 'var/ezdemo_site/cache/public' ) Debug: (<kernel-clustering>) dfs::delete( 'var/ezdemo_site/cache/public' ) Notice: (cluster::mysql::_delete(var/ezdemo_site/cache/public)::_begin[0 rows, 0.000 ms] query number per page:1) BEGIN Notice: (cluster::mysql::_query[0 rows, 0.001 ms] query number per page:2) UPDATE ezdfsfile SET mtime=-ABS(mtime), expired=1 WHERE name_hash=MD5('var/ezdemo_site/cache/public') Notice: (cluster::mysql::_delete(var/ezdemo_site/cache/public)::_commit[0 rows, 0.000 ms] query number per page:3) COMMIT
