Akka Coordinated Shutdown

Akka Coordinated Shutdown

• Akka의 Actor System 종료 과정에서 특정 Actor 또는 서비스에 대해 작동을 요청을 할수가 있다.
• 밀린 메시지 처리 및 Connect 정리 등에 활용 될수 있는 기능이다.

Configuration

# CoordinatedShutdown is enabled by default and will run the tasks that

# are added to these phases by individual Akka modules and user logic.

#

# The phases are ordered as a DAG by defining the dependencies between the phases

# to make sure shutdown tasks are run in the right order.

#

# In general user tasks belong in the first few phases, but there may be use

# cases where you would want to hook in new phases or register tasks later in

# the DAG.

#

# Each phase is defined as a named config section with the

# following optional properties:

# - timeout=15s: Override the default-phase-timeout for this phase.

# - recover=off: If the phase fails the shutdown is aborted

#                and depending phases will not be executed.

# - enabled=off: Skip all tasks registered in this phase. DO NOT use

#                this to disable phases unless you are absolutely sure what the

#                consequences are. Many of the built in tasks depend on other tasks

#                having been executed in earlier phases and may break if those are disabled.

# depends-on=[]: Run the phase after the given phases

phases {

  # The first pre-defined phase that applications can add tasks to.

  # Note that more phases can be added in the application's

  # configuration by overriding this phase with an additional

  # depends-on.

  before-service-unbind {

  }

  # Stop accepting new incoming connections.

  # This is where you can register tasks that makes a server stop accepting new connections. Already

  # established connections should be allowed to continue and complete if possible.

  service-unbind {

    depends-on = [before-service-unbind]

  }

  # Wait for requests that are in progress to be completed.

  # This is where you register tasks that will wait for already established connections to complete, potentially

  # also first telling them that it is time to close down.

  service-requests-done {

    depends-on = [service-unbind]

  }

  # Final shutdown of service endpoints.

  # This is where you would add tasks that forcefully kill connections that are still around.

  service-stop {

    depends-on = [service-requests-done]

  }

  # Phase for custom application tasks that are to be run

  # after service shutdown and before cluster shutdown.

  before-cluster-shutdown {

    depends-on = [service-stop]

  }

  # Graceful shutdown of the Cluster Sharding regions.

  # This phase is not meant for users to add tasks to.

  cluster-sharding-shutdown-region {

    timeout = 10 s

    depends-on = [before-cluster-shutdown]

  }

  # Emit the leave command for the node that is shutting down.

  # This phase is not meant for users to add tasks to.

  cluster-leave {

    depends-on = [cluster-sharding-shutdown-region]

  }

  # Shutdown cluster singletons

  # This is done as late as possible to allow the shard region shutdown triggered in

  # the "cluster-sharding-shutdown-region" phase to complete before the shard coordinator is shut down.

  # This phase is not meant for users to add tasks to.

  cluster-exiting {

    timeout = 10 s

    depends-on = [cluster-leave]

  }

  # Wait until exiting has been completed

  # This phase is not meant for users to add tasks to.

  cluster-exiting-done {

    depends-on = [cluster-exiting]

  }

  # Shutdown the cluster extension

  # This phase is not meant for users to add tasks to.

  cluster-shutdown {

    depends-on = [cluster-exiting-done]

  }

  # Phase for custom application tasks that are to be run

  # after cluster shutdown and before ActorSystem termination.

  before-actor-system-terminate {

    depends-on = [cluster-shutdown]

  }

  # Last phase. See terminate-actor-system and exit-jvm above.

  # Don't add phases that depends on this phase because the

  # dispatcher and scheduler of the ActorSystem have been shutdown.

  # This phase is not meant for users to add tasks to.

  actor-system-terminate {

    timeout = 10 s

    depends-on = [before-actor-system-terminate]

  }

}

• `akka.coordinated-shutdown.phases` 에 정의가 되어있다.

Shutdown Phase에 Task 추가하기

CoordinatedShutdown.get(system).addTask(

  CoordinatedShutdown.PhaseBeforeServiceUnbind(), "someTaskName",

  () -> {

    return akka.pattern.Patterns.ask(someActor, "stop", Duration.ofSeconds(5))

      .thenApply(reply -> Done.getInstance());

});

• Code를 이용하여 Phase에 필요한 작업을 추가를 할수가 있다.
• 리턴 값으로는 `CompletionStage<Done>`이 반드시 리턴이 되어야 한다.
• taskName은 debug용, logging용으로 사용되는 이름.

Etc

•  Coordinated Shutdown는 Jvm의 process exit에 의해 실행이 된다. jvm에 의해 종료되는 것이 아닌 custom이 필요한 경우에는 설정 등을 통해 할수있다. 자세한 내용은 문서를 참고

Reference

https://doc.akka.io/docs/akka/current/actors.html#coordinated-shutdown