@halorgium @mperham I have a strong suspicion this is strongly coupled to this issue: https://github.com/cowboyd/therubyracer/issues/206 … its the root cause
-
-
Replying to @samsaffron
@samsaffron@mperham good find. have you a short repro? This crash should not affect celluloid or sidekiq.1 reply 0 retweets 0 likes -
Replying to @halorgium
@halorgium@mperham checked a nasty functioning workaround into@discourse https://github.com/discourse/discourse/commit/b2b01fd26cfa0e1d46b5955aef5790aec1bf750c … (wish sidekiq had a non fiber mode)2 replies 0 retweets 1 like -
Replying to @samsaffron
@samsaffron@mperham this happens on Ubuntu with MRI 1.9.3 as well. Did you see my info on the bug report?1 reply 0 retweets 0 likes -
Replying to @halorgium
@halorgium@mperham interesting, was not able to repro it there with Discourse. Yeah makes total sense.1 reply 0 retweets 0 likes -
Replying to @samsaffron
@samsaffron@mperham sidekiq does actually use a separate Thread on 1.9.3 due to the stack limit. MRI 2.0.0 has a larger Fiber stack btw.2 replies 0 retweets 0 likes -
Replying to @halorgium
@halorgium@samsaffron For some reason, Fibers have proven to be more stable for Sidekiq users overall. No idea why.2 replies 0 retweets 0 likes -
Replying to @mperham
@mperham@halorgium I'm curious are people actually yielding these fibers in sidekiq, what is the benefit in execute?4 replies 0 retweets 0 likes -
Replying to @samsaffron
@samsaffron@mperham@halorgium we've tried running Sidekiq using "Erlang mode" and it deadlocks for indeterminate reasons1 reply 0 retweets 0 likes -
Replying to @bascule
@bascule@samsaffron@mperham could you elaborate? i wonder if we’ve gotten past this point.1 reply 0 retweets 0 likes
@halorgium @samsaffron @mperham we tried making Sidekiq::Workers exclusive...
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.