Repository: tez
Updated Branches:
refs/heads/branch-0.6 6751a67f2 -> 3a8bf0e0a
TEZ-2781. Fallback to send only TaskAttemptFailedEvent if taskFailed heartbeat
fails (zjffdu)
(cherry picked from commit f9d15c8695de7975817631b051450336bc5eadee)
Conflicts:
CHANGES.txt
tez-dag/src/main/java/org/apache/tez/dag/app/TaskCommunicatorManager.java
(cherry picked from commit 8d49fd5285016fb64ebccdc9cf31c408c79ebaaf)
Conflicts:
CHANGES.txt
tez-dag/src/main/java/org/apache/tez/dag/app/TaskAttemptListenerImpTezDag.java
tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
Project: http://git-wip-us.apache.org/repos/asf/tez/repo
Commit: http://git-wip-us.apache.org/repos/asf/tez/commit/3a8bf0e0
Tree: http://git-wip-us.apache.org/repos/asf/tez/tree/3a8bf0e0
Diff: http://git-wip-us.apache.org/repos/asf/tez/diff/3a8bf0e0
Branch: refs/heads/branch-0.6
Commit: 3a8bf0e0afb0a30d5c4cbf82d1e2ce9e0dd22b4a
Parents: 6751a67
Author: Jeff Zhang <[email protected]>
Authored: Fri Oct 9 15:07:17 2015 +0800
Committer: Jeff Zhang <[email protected]>
Committed: Fri Oct 9 15:27:59 2015 +0800
----------------------------------------------------------------------
CHANGES.txt | 2 ++
.../org/apache/tez/runtime/task/TaskReporter.java | 14 +++++++++++---
.../org/apache/tez/test/TestExceptionPropagation.java | 12 +++++++++++-
3 files changed, 24 insertions(+), 4 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/tez/blob/3a8bf0e0/CHANGES.txt
----------------------------------------------------------------------
diff --git a/CHANGES.txt b/CHANGES.txt
index 4fdcb66..62a4929 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -6,6 +6,7 @@ Release 0.6.3: Unreleased
INCOMPATIBLE CHANGES
ALL CHANGES:
+ TEZ-2781. Fallback to send only TaskAttemptFailedEvent if taskFailed
heartbeat fails
TEZ-2855. Fix a potential NPE while routing VertexManager events.
TEZ-2716. DefaultSorter.isRleNeeded not thread safe.
TEZ-2758. Remove append API in RecoveryService after TEZ-1909.
@@ -249,6 +250,7 @@ INCOMPATIBLE CHANGES
TEZ-2552. CRC errors can cause job to run for very long time in large jobs.
ALL CHANGES:
+ TEZ-2781. Fallback to send only TaskAttemptFailedEvent if taskFailed
heartbeat fails
TEZ-2398. Flaky test: TestFaultTolerance
TEZ-2834. Make Tez preemption resilient to incorrect free resource reported
by YARN
http://git-wip-us.apache.org/repos/asf/tez/blob/3a8bf0e0/tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
----------------------------------------------------------------------
diff --git
a/tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
b/tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
index defc6bd..4def43f 100644
--- a/tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
+++ b/tez-dag/src/main/java/org/apache/tez/runtime/task/TaskReporter.java
@@ -332,8 +332,15 @@ public class TaskReporter {
*/
private boolean taskFailed(TezTaskAttemptID taskAttemptID, Throwable t,
String diagnostics,
EventMetaData srcMeta) throws IOException, TezException {
- TezEvent statusUpdateEvent = new TezEvent(new
TaskStatusUpdateEvent(task.getCounters(),
- task.getProgress()), updateEventMetadata);
+ List<TezEvent> tezEvents = new ArrayList<TezEvent>();
+ try {
+ TezEvent statusUpdateEvent = new TezEvent(new
TaskStatusUpdateEvent(task.getCounters(),
+ task.getProgress()), updateEventMetadata);
+ tezEvents.add(statusUpdateEvent);
+ } catch (Exception e) {
+ // Counter may exceed limitation
+ LOG.warn("Error when get constructing TaskStatusUpdateEvent");
+ }
if (diagnostics == null) {
diagnostics = ExceptionUtils.getStackTrace(t);
} else {
@@ -341,7 +348,8 @@ public class TaskReporter {
}
TezEvent taskAttemptFailedEvent = new TezEvent(new
TaskAttemptFailedEvent(diagnostics),
srcMeta == null ? updateEventMetadata : srcMeta);
- return !heartbeat(Lists.newArrayList(statusUpdateEvent,
taskAttemptFailedEvent)).shouldDie;
+ tezEvents.add(taskAttemptFailedEvent);
+ return !heartbeat(tezEvents).shouldDie;
}
private void addEvents(TezTaskAttemptID taskAttemptID,
Collection<TezEvent> events) {
http://git-wip-us.apache.org/repos/asf/tez/blob/3a8bf0e0/tez-tests/src/test/java/org/apache/tez/test/TestExceptionPropagation.java
----------------------------------------------------------------------
diff --git
a/tez-tests/src/test/java/org/apache/tez/test/TestExceptionPropagation.java
b/tez-tests/src/test/java/org/apache/tez/test/TestExceptionPropagation.java
index bb9888a..d30d73f 100644
--- a/tez-tests/src/test/java/org/apache/tez/test/TestExceptionPropagation.java
+++ b/tez-tests/src/test/java/org/apache/tez/test/TestExceptionPropagation.java
@@ -223,7 +223,11 @@ public class TestExceptionPropagation {
DAGStatus dagStatus = dagClient.waitForCompletion();
String diagnostics = StringUtils.join(dagStatus.getDiagnostics(), ",");
LOG.info("Diagnostics:" + diagnostics);
- assertTrue(diagnostics.contains(exLocation.name()));
+ if (exLocation == ExceptionLocation.PROCESSOR_COUNTER_EXCEEDED) {
+ assertTrue(diagnostics.contains("Too many counters"));
+ } else {
+ assertTrue(diagnostics.contains(exLocation.name()));
+ }
}
} finally {
stopSessionClient();
@@ -300,6 +304,7 @@ public class TestExceptionPropagation {
// PROCESSOR_HANDLE_EVENTS
PROCESSOR_RUN_ERROR, PROCESSOR_CLOSE_ERROR, PROCESSOR_INITIALIZE_ERROR,
PROCESSOR_RUN_EXCEPTION, PROCESSOR_CLOSE_EXCEPTION,
PROCESSOR_INITIALIZE_EXCEPTION,
+ PROCESSOR_COUNTER_EXCEEDED,
// VM
VM_INITIALIZE, VM_ON_ROOTVERTEX_INITIALIZE,VM_ON_SOURCETASK_COMPLETED,
VM_ON_VERTEX_STARTED,
@@ -624,6 +629,11 @@ public class TestExceptionPropagation {
throw new Error(this.exLocation.name());
} else if (this.exLocation == ExceptionLocation.PROCESSOR_RUN_EXCEPTION)
{
throw new Exception(this.exLocation.name());
+ } else if (this.exLocation ==
ExceptionLocation.PROCESSOR_COUNTER_EXCEEDED) {
+ // simulate the counter limitation exceeded
+ for (int i=0;i< TezConfiguration.TEZ_COUNTERS_MAX_DEFAULT+1; ++i) {
+ getContext().getCounters().findCounter("mycounter",
"counter_"+i).increment(1);
+ }
}
}