Thanks for reporting. I have created github issue: https://github.com/akka/akka/issues/17729
On Wed, Jun 10, 2015 at 2:53 PM, Matej Krchniak <[email protected]> wrote: > Hi. > > We hit serious PermGen leak when used akka in war/ear application. We used > this pattern for shutting down ActorSystem while destroying application: > > actorSystem.shutdown(); > actorSystem.awaitTermination(); > > > Redeploying application then caused java.lang.OutOfMemoryError: PermGen > space. Root cause was leaked thread ForkJoinPool-3-worker-3, which > prevented WebappClassLoader from being garbage collected. All dispatcher > threads have been destroyed. After hours of debugging and by elemination we > identified source of this leaked thread. Leak is caused by > Remoting#shutdown method which imports > scala.concurrent.ExecutionContext.Implicits.global. When shutting down > remoting, new ForkJoinPool (ForkJoinPool-3) is created and one thread > remains, which holds ClassLoader in field contextClassLoader. ForkJoinPool-1 > and ForkJoinPool-2 were created by default-dispatcher and > default-remote-dispatcher. > > Our current workaround is termination callback which is responsible for > shutting down global ExecutionContext. I'm not sure, if this is 100% > working solution (or rather call executorService.shutdown() after > actorSystem.awaitTermination()), but at least leaking thread is no longer > present after shutting down application. > > actorSystem.registerOnTermination(new Runnable() { > @Override > public void run() { > ExecutionContextImpl executorImpl = (ExecutionContextImpl) > ExecutionContext.Implicits$.MODULE$.global(); > AbstractExecutorService executorService = (AbstractExecutorService) > executorImpl.executor(); > > executorService.shutdown(); > try { > executorService.awaitTermination(10, TimeUnit.SECONDS); > LOGGER.info("Shutting down ExecutionContext.Implicits.global > completed."); > } catch (InterruptedException e) { > LOGGER.error("Shutting down ExecutionContext.Implicits.global > failed.", e); > } > } > }); > > > Is there some reason why is Remoting#shutdown using > scala.concurrent.ExecutionContext.Implicits.global instead of some thread > created by akka (e.g. dispatcher thread)? Who is then responsible for > shutting down global ExecutionContext thread pool? > > Thanks > > -- > >>>>>>>>>> Read the docs: http://akka.io/docs/ > >>>>>>>>>> Check the FAQ: > http://doc.akka.io/docs/akka/current/additional/faq.html > >>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user > --- > You received this message because you are subscribed to the Google Groups > "Akka User List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/akka-user. > For more options, visit https://groups.google.com/d/optout. > -- Patrik Nordwall Typesafe <http://typesafe.com/> - Reactive apps on the JVM Twitter: @patriknw -- >>>>>>>>>> Read the docs: http://akka.io/docs/ >>>>>>>>>> Check the FAQ: >>>>>>>>>> http://doc.akka.io/docs/akka/current/additional/faq.html >>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user --- You received this message because you are subscribed to the Google Groups "Akka User List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/akka-user. For more options, visit https://groups.google.com/d/optout.
