Ladsgroup added a comment.

  In T205045#5288620 <https://phabricator.wikimedia.org/T205045#5288620>, 
@jcrespo wrote:
  
  > Yes, we were worried about the frequency of it (it was the most common 
DB-related error for some time) because of the performance implications, not 
that it happens sometimes.
  
  
  I agree, but is the current state acceptable or do you think it should be 
less frequent?
  
  > I am guessing that this is due to multiple edits happening at the same time 
about the same subject/item/related item?
  
  It's actually a site-wide issue. User X, edits article foo and make it to 
redo the usage tracking again so the job tries to update it while User Y edits 
article bar and make it to do the similar thing.
  
  > Sadly there is limited capabilities to make the jobqueue smarter, I think. 
Maybe there could be strict partitioning on ranges to prevent collision? I 
don't know.
  
  Reiterating how the thing works actually gave me the idea that we can be a 
little bit smarter on what we inject to the database and avoid injecting things 
that hasn't changed. I need to double check the code but I think it removes all 
usages and injects them again.

TASK DETAIL
  https://phabricator.wikimedia.org/T205045

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Ladsgroup
Cc: Michael, Lydia_Pintscher, jcrespo, Ladsgroup, hoo, Addshore, Marostegui, 
Aklapper, Krinkle, darthmon_wmde, Nandana, Lahi, Gq86, GoranSMilovanovic, 
QZanden, LawExplorer, _jensen, rosalieper, Jonas, Wikidata-bugs, aude, 
Jdforrester-WMF, Mbch331, Jay8g, Krenair
_______________________________________________
Wikidata-bugs mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to