[jira] [Closed] (FLINK-8923) Set an uncaught exception handler for ForkJoinPool

["Nico Kruber (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/FLINK-8923?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Nico Kruber closed FLINK-8923.
------------------------------
       Resolution: Won't Do
    Fix Version/s:     (was: 1.6.0)
                       (was: 1.5.0)

Actually, after investigating further and trying to cause an uncaught exception in a {{ForkJoinPool}} worker thread/task, and also looking into the implementation, I think, we don't really need to set this property since both {{ForkJoinTask}} and {{ForkJoinWorkerThread}} take care of exceptions, store thrown ones and re-throw them when joining/getting the result.

> Set an uncaught exception handler for ForkJoinPool
> --------------------------------------------------
>
>                 Key: FLINK-8923
>                 URL: https://issues.apache.org/jira/browse/FLINK-8923
>             Project: Flink
>          Issue Type: Improvement
>          Components: Core, Startup Shell Scripts
>    Affects Versions: 1.4.0, 1.5.0, 1.4.1, 1.4.2, 1.6.0
>            Reporter: Nico Kruber
>            Assignee: Nico Kruber
>            Priority: Major
>
> We should also set an uncaught exception handler for the {{ForkJoinPool}} which may be used directly, indirectly via Java's parallel steam operations, or through any library we use. We'd like to make sure Flink doesn't linger around while waiting for something to happen which will never happen due to an exception in another thread.
> This should be as easy as setting the {{java.util.concurrent.ForkJoinPool.common.exceptionHandler}} property to {{org.apache.flink.runtime.util.FatalExitExceptionHandler}}.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)