Last modified: 2014-07-01 10:18:54 UTC
EventLogging's database contains tables that look like cruft and didn't get new data since some time [1]. If no one claims them by 2014-06-23, let's remove them as they get in the way when trying to sanitize data. [1] mysql:research@s1-analytics-slave.eqiad.wmnet [log]> SHOW TABLES FROM log LIKE '%\_1'; +----------------------------------------------+ | Tables_in_log (%\_1) | +----------------------------------------------+ | MultimediaViewerNetworkPerformance_7917896_1 | +----------------------------------------------+ 1 row in set (0.00 sec) mysql:research@s1-analytics-slave.eqiad.wmnet [log]> SHOW TABLES FROM log LIKE 'zz\_%'; +------------------------------------+ | Tables_in_log (zz\_%) | +------------------------------------+ | zz_MobileWebInfobox_6221064 | | zz_ModuleStorage_6356853 | | zz_NavigationTiming_7494934 | | zz_PageContentSaveComplete_5303086 | | zz_PageContentSaveComplete_5588433 | | zz_PageSaveTiming_5557427 | | zz_TimingData_7254808 | +------------------------------------+ 7 rows in set (0.00 sec)
Relayed issue to the analytics and eventlogging-alerts mailing lists: * http://lists.wikimedia.org/pipermail/analytics/2014-June/002233.html (Cannot link eventlogging-alerts post as that list is lacking archive: bug 66488)
The deadline set in the list-email has passed. So let's work towards removing those tables.
Since I lack permissions to delete them myself, I filed RT #7735 to have the tables deleted.
Tables have been dropped. Thanks Andrew and Sean!