https://bugzilla.wikimedia.org/show_bug.cgi?id=54197
Matthias Mullie <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Assignee|[email protected]. |[email protected] |org | --- Comment #3 from Matthias Mullie <[email protected]> --- Brain dump: Looking at e.g. feedback #755805 (https://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Cloud_computing/755805), this feedback entry is actually archived. DB data confirms: aft_archive = 1. Only the maintenance script, which has since been killed, can mark feedback as archived. Activity log confirms the last action was performed by Article Feedback v5. Where-condition, executed against slave, is: array( 'aft_archive = 0', "aft_archive_date <= '$now'" ). This means that this post should no longer appear when queried for "posts to be archived". I'd say that perhaps slave is/was lagging, but the wait is right there in the code, and I'll assume it works: wfWaitForSlaves( false, false, $wgArticleFeedbackv5Cluster ); Also, the script had been going at it for 10+ minutes (another feedback entry for even way longer) until it was killed, so that lag would have to have been pretty bad. This is a highly unlikely cause. It looks like the archiving kicked in right after a person manually un-archived the feedback. *ERROR #1*: archive date should be re-calculated after a person un-archives feedback (or nulled completely), not be the original archive date (which results in immediate re-archive). (probably unrelated) *ERROR #2*: maintenance script nulls aft_archive_date, but the subsequent update() call will fill out that value again. Have to run for now, will get these fixed and continue figuring out why exactly archive broke. -- You are receiving this mail because: You are the assignee for the bug. You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
