[jira] [Created] (SPARK-21564) TaskDescription decoding failure should fail the task

["Andrew Ash (JIRA)" <ji...@apache.org>]

Andrew Ash created SPARK-21564:
----------------------------------

             Summary: TaskDescription decoding failure should fail the task
                 Key: SPARK-21564
                 URL: https://issues.apache.org/jira/browse/SPARK-21564
             Project: Spark
          Issue Type: Bug
          Components: Spark Core
    Affects Versions: 2.2.0
            Reporter: Andrew Ash


I was seeing an issue where Spark was throwing this exception:

{noformat}
16:16:28.294 [dispatcher-event-loop-14] ERROR org.apache.spark.rpc.netty.Inbox - Ignoring error
java.io.EOFException: null
    at java.io.DataInputStream.readFully(DataInputStream.java:197)
    at java.io.DataInputStream.readUTF(DataInputStream.java:609)
    at java.io.DataInputStream.readUTF(DataInputStream.java:564)
    at org.apache.spark.scheduler.TaskDescription$$anonfun$decode$1.apply(TaskDescription.scala:127)
    at org.apache.spark.scheduler.TaskDescription$$anonfun$decode$1.apply(TaskDescription.scala:126)
    at scala.collection.immutable.Range.foreach(Range.scala:160)
    at org.apache.spark.scheduler.TaskDescription$.decode(TaskDescription.scala:126)
    at org.apache.spark.executor.CoarseGrainedExecutorBackend$$anonfun$receive$1.applyOrElse(CoarseGrainedExecutorBackend.scala:95)
    at org.apache.spark.rpc.netty.Inbox$$anonfun$process$1.apply$mcV$sp(Inbox.scala:117)
    at org.apache.spark.rpc.netty.Inbox.safelyCall(Inbox.scala:205)
    at org.apache.spark.rpc.netty.Inbox.process(Inbox.scala:101)
    at org.apache.spark.rpc.netty.Dispatcher$MessageLoop.run(Dispatcher.scala:213)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:748)
{noformat}

For details on the cause of that exception, see SPARK-21563

We've since changed the application and have a proposed fix in Spark at the ticket above, but it was troubling that decoding the TaskDescription wasn't failing the tasks.  So the Spark job ended up hanging and making no progress, permanently stuck, because the driver thinks the task is running but the thread has died in the executor.

We should make a change around https://github.com/apache/spark/blob/v2.2.0/core/src/main/scala/org/apache/spark/executor/CoarseGrainedExecutorBackend.scala#L96 so that when that decode throws an exception, the task is marked as failed.

cc [~kayousterhout] [~irashid]



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org