Roan Kattouw wrote: > On Tue, Mar 20, 2012 at 7:12 AM, MZMcBride <z...@mzmcbride.com> wrote: >> Some people have suggested that the current API behavior is intentional. >> That is, that having different servers return the same error code is better >> than having different servers return an error code or not. I think this is >> flawed logic due to the problems that it presents (scripts unable to get >> past the error code), but it's definitely something that needs investigation >> for the future. > > You're confusing Apache servers and database servers. You're hitting > different Apache servers at random, those are the srv numbers. Each > Apache requests the current lag from all the DB servers serving that > wiki, computes the maximum (in this case db36 wins easily) and checks > that against your maxlag parameter. Obviously the highest lag is > greater than your maxlag setting, so it throws an error and bails. > > Had it continued, or had you not set a maxlag parameter, you would > have hit a DB server at random (except that it wouldn't have let you > hit db36 because it was lagged to hell and back). So yeah /normally/ > you hit DB servers at random and different servers might respond > differently (or be lagged to different degrees), but in this > particular case it was always the same DB server returning the same > lag value. Nothing strange going on here, this is how the maxlag > parameter works.
Thank you for this explanation. It was very helpful. I had some follow-up questions about what to do to mitigate this issue going forward (particularly server-side), but you and others are already brainstorming in this thread, so I'll slink back to my quiet world of client-side scripting. :-) (And thanks again to Asher for depooling the lagged server, even as a temporary fix. I was able to do about 2500 talk page deliveries this morning without any issue.) MZMcBride _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
