When you review an article with PageTriage, it fires one request to set the reviewed status and then a 2nd request to refetch the data for the page. The first request invalidates the cache and the 2nd rebuilds it. On test.wiki it looks like the 2nd request is rebuilding the cache before the changes from the 1st request have propagated to the slave DBs. This problem doesn't seem to occur on en.wiki though.
Version: unspecified
Severity: normal
Benny suggested that we update the cache via the first request rather than invalidating it. That way the 2nd request is guaranteed to get the correct info.
(In reply to comment #0)
When you review an article with PageTriage, it fires one request to set the
reviewed status and then a 2nd request to refetch the data for the page.
If I understand you correctly you mean that API call to 'pagetriagelist' refetches the data for the page?
All that is needed to be done is update the cache with following fields:
'ptrp_reviewed', 'ptrp_reviewed_updated' and 'ptrp_last_reviewed_by'