MAPREDUCE-7249. Fix Invalid event TA_TOO_MANY_FETCH_FAILURE at SUCCESS_CONTAINER_CLEANUP causes job failure.

Contributed by Wilfred Spiegelenburg.

(cherry picked from commit a97f7776bd05f957a998b4883141855cad5afb31)

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/job/impl/TaskAttemptImpl.java

@@ -469,12 +469,16 @@ public abstract class TaskAttemptImpl implements
          TaskAttemptStateInternal.COMMIT_PENDING,
          TaskAttemptEventType.TA_COMMIT_PENDING)
 
-     // Transitions from SUCCESS_CONTAINER_CLEANUP state
-     // kill and cleanup the container
-     .addTransition(TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
-         TaskAttemptStateInternal.SUCCEEDED,
-         TaskAttemptEventType.TA_CONTAINER_CLEANED)
-     .addTransition(
+      // Transitions from SUCCESS_CONTAINER_CLEANUP state
+      // kill and cleanup the container
+      .addTransition(TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
+          TaskAttemptStateInternal.SUCCEEDED,
+          TaskAttemptEventType.TA_CONTAINER_CLEANED)
+      .addTransition(TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
+          TaskAttemptStateInternal.FAILED,
+          TaskAttemptEventType.TA_TOO_MANY_FETCH_FAILURE,
+          new TooManyFetchFailureTransition())
+      .addTransition(
           TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
           TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
           TaskAttemptEventType.TA_DIAGNOSTICS_UPDATE,

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapreduce/v2/app/job/impl/TestTaskAttempt.java

@@ -1773,6 +1773,38 @@ public class TestTaskAttempt{
     createReduceTaskAttemptImplForTest(eventHandler, clock, jobConf);
   }
 
+  @Test
+  public void testTooManyFetchFailureWhileContainerCleanup() {
+    MockEventHandler eventHandler = new MockEventHandler();
+    TaskAttemptImpl taImpl = createTaskAttemptImpl(eventHandler);
+    TaskId reducetaskId = MRBuilderUtils.newTaskId(taImpl.getID().getTaskId()
+        .getJobId(), 1, TaskType.REDUCE);
+    TaskAttemptId reduceTAId =
+        MRBuilderUtils.newTaskAttemptId(reducetaskId, 0);
+
+    // move in two steps to the desired state (cannot get there directly)
+    taImpl.handle(new TaskAttemptEvent(taImpl.getID(),
+        TaskAttemptEventType.TA_DONE));
+    assertEquals("Task attempt's internal state is not " +
+        "SUCCESS_FINISHING_CONTAINER",
+        TaskAttemptStateInternal.SUCCESS_FINISHING_CONTAINER,
+        taImpl.getInternalState());
+
+    taImpl.handle(new TaskAttemptEvent(taImpl.getID(),
+        TaskAttemptEventType.TA_TIMED_OUT));
+    assertEquals("Task attempt's internal state is not " +
+        "SUCCESS_CONTAINER_CLEANUP",
+        TaskAttemptStateInternal.SUCCESS_CONTAINER_CLEANUP,
+        taImpl.getInternalState());
+
+    taImpl.handle(new TaskAttemptTooManyFetchFailureEvent(taImpl.getID(),
+        reduceTAId, "Host"));
+    assertEquals("Task attempt is not in FAILED state",
+        TaskAttemptState.FAILED,
+        taImpl.getState());
+    assertFalse("InternalError occurred", eventHandler.internalError);
+  }
+
   private void initResourceTypes() {
     Configuration conf = new Configuration();
     conf.set(YarnConfiguration.RM_CONFIGURATION_PROVIDER_CLASS,