On 04/05/2013 03:01 PM, Isabel Gancedo wrote:
Everything seems to be working fine; jobs run a maximum warp and the queue is becoming tiny.
Excellent. I see this line in the "New features" list from the Release notes:
The Job system was refactored to allow for different backing stores for queues as well as cross-wiki access to queues, among other things. The schema for the DB queue was changed to support better concurrency and reduce deadlock errors.
Perhaps that deserves more attention?
There is just one issue that I would like to mention: in my database (MySQL) pre-upgrade jobs have job_random set to 0 and do not seem to be picked up -not even when I use the option --type=replaceText.
I can repeat the replace text operations, so this is not big problem for me. However, if this is the normal behaviour for an upgrade, maybe it should be mentioned in the notes somewhere.
This may be a bug that was introduced during the refactoring. Could you file one?
https://bugzilla.wikimedia.org/enter_bug.cgi?product=MediaWiki&cc=mah@ev...
Thank you for your helpful comments!