You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Dmitri Chmelev (Jira)" <ji...@apache.org> on 2020/03/21 01:20:00 UTC

[jira] [Created] (FLINK-16703) AkkaRpcActor state machine does record transition to terminating state.

Dmitri Chmelev created FLINK-16703:
--------------------------------------

             Summary: AkkaRpcActor state machine does record transition to terminating state.
                 Key: FLINK-16703
                 URL: https://issues.apache.org/jira/browse/FLINK-16703
             Project: Flink
          Issue Type: Bug
          Components: Runtime / Coordination
    Affects Versions: 1.10.0, 1.9.0, 1.8.0, 1.11.0, 2.0.0
            Reporter: Dmitri Chmelev


As part of FLINK-11551, the state machine of AkkaRpcActor has been updated to include 'terminating' and 'terminated' states. However, when actor termination request is handled, the resulting 'terminating' state is not updated by the FSM.

[https://github.com/apache/flink/blame/master/flink-runtime/src/main/java/org/apache/flink/runtime/rpc/akka/AkkaRpcActor.java#L175]

As a side-effect, {{isRunning()}} predicate can return that the actor is still running after terminate was initiated and to still handle messages.

I believe the fix is trivial and the private fieldĀ {{state}} should be updated with the return value of the call to {{state.terminate()}}.

Feel free to adjust the priority accordingly.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)