You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@flink.apache.org by "Stephan Ewen (Jira)" <ji...@apache.org> on 2020/02/10 11:43:00 UTC

[jira] [Created] (FLINK-15966) Capture the call stack of RPC ask() calls.

Stephan Ewen created FLINK-15966:
------------------------------------

             Summary: Capture the call stack of RPC ask() calls.
                 Key: FLINK-15966
                 URL: https://issues.apache.org/jira/browse/FLINK-15966
             Project: Flink
          Issue Type: Improvement
          Components: Runtime / Coordination
            Reporter: Stephan Ewen
            Assignee: Stephan Ewen
             Fix For: 1.10.1, 1.11.0


Currently, when an RPC ask() call fails, we get a rather unhelpful exception with a stack trace from akka's internal scheduler.

Instead, we should capture the call stack during the invocation and use it to give a helpful error message when the RPC call failed. This is especially helpful in cases where the future (and future handlers) are passed for later asynchronous result handling (which is the common case in most components (JM / TM / RM).

The options should have a flag to turn it off, because when having a lot of concurrent ask calls (hundreds of thousands, during large deploy phases), it may be possible that the captured call.



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