On Thu, Sep 29, 2016 at 9:29 PM, Shannon Carey <sca...@expedia.com> wrote: > It looks like Flink is disabling the objects that the FlatMap collector > relies on before disabling the operator itself. Is that expected/normal? Is > there anything I should change in my FlatMap function or job code to account > for it?
Hey Shannon, Flink actually does cancel the tasks *before* cleaning up the network resources that throw the root Exception here. We actually don't log any Exceptions that are thrown during cancellation, because it is possible that the user code/operator use the closed resources concurrently with cancellation (your stack traces essentially), but it looks like in some places we don't respect this. Can you tell which classes actually log this? Would be good to fix this if possible as it is very confusing and looks quite bad. I don't expect it to be an actual problem though. – Ufuk
