YARN-6695. Fixed NPE in publishing appFinished events to ATSv2.

           Contributed by Prabhu Joseph

(cherry picked from commit df76cdc8959c51b71704ab5c38335f745a6f35d8)

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/metrics/TimelineServiceV2Publisher.java

@@ -454,10 +454,15 @@ public class TimelineServiceV2Publisher extends AbstractSystemMetricsPublisher {
       }
       TimelineCollector timelineCollector =
           rmTimelineCollectorManager.get(appId);
-      TimelineEntities entities = new TimelineEntities();
-      entities.addEntity(entity);
-      timelineCollector.putEntities(entities,
-          UserGroupInformation.getCurrentUser());
+      if (timelineCollector != null) {
+        TimelineEntities entities = new TimelineEntities();
+        entities.addEntity(entity);
+        timelineCollector.putEntities(entities,
+                UserGroupInformation.getCurrentUser());
+      } else {
+        LOG.debug("Cannot find active collector while publishing entity "
+            + entity);
+      }
     } catch (IOException e) {
       LOG.error("Error when publishing entity " + entity);
       if (LOG.isDebugEnabled()) {

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/metrics/TestSystemMetricsPublisherForV2.java

@@ -28,12 +28,17 @@ import java.io.BufferedReader;
 import java.io.File;
 import java.io.FileReader;
 import java.io.IOException;
+import java.util.ArrayList;
 import java.util.Collections;
 import java.util.HashMap;
+import java.util.List;
 import java.util.Map;
 import java.util.concurrent.ConcurrentHashMap;
 import java.util.concurrent.ConcurrentMap;
 
+import org.apache.log4j.AppenderSkeleton;
+import org.apache.log4j.Logger;
+import org.apache.log4j.spi.LoggingEvent;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.FileContext;
 import org.apache.hadoop.fs.Path;
@@ -291,6 +296,48 @@ public class TestSystemMetricsPublisherForV2 {
         ContainerMetricsConstants.CREATED_IN_RM_EVENT_TYPE, 0, 0);
   }
 
+  @Test(timeout = 10000)
+  public void testPutEntityWhenNoCollector() throws Exception {
+    // Validating the logs as DrainDispatcher won't throw exception
+    class TestAppender extends AppenderSkeleton {
+      private final List<LoggingEvent> log = new ArrayList<>();
+
+      @Override
+      public boolean requiresLayout() {
+        return false;
+      }
+
+      @Override
+      protected void append(final LoggingEvent loggingEvent) {
+        log.add(loggingEvent);
+      }
+
+      @Override
+      public void close() {
+      }
+
+      public List<LoggingEvent> getLog() {
+        return new ArrayList<>(log);
+      }
+    }
+
+    TestAppender appender = new TestAppender();
+    final Logger logger = Logger.getRootLogger();
+    logger.addAppender(appender);
+
+    try {
+      RMApp app = createRMApp(ApplicationId.newInstance(0, 1));
+      metricsPublisher.appCreated(app, app.getStartTime());
+      dispatcher.await();
+      for (LoggingEvent event : appender.getLog()) {
+        assertFalse("Dispatcher Crashed",
+            event.getRenderedMessage().contains("Error in dispatcher thread"));
+      }
+    } finally {
+      logger.removeAppender(appender);
+    }
+  }
+
   private RMApp createAppAndRegister(ApplicationId appId) {
     RMApp app = createRMApp(appId);
 

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-site/src/site/markdown/TimelineServiceV2.md

@@ -295,14 +295,6 @@ Following are the basic configurations to start Timeline service v.2:
   <name>yarn.system-metrics-publisher.enabled</name>
   <value>true</value>
 </property>
-
-<property>
-  <description>The setting that controls whether yarn container events are
-  published to the timeline service or not by RM. This configuration setting
-  is for ATS V2.</description>
-  <name>yarn.rm.system-metrics-publisher.emit-container-events</name>
-  <value>true</value>
-</property>
 ```
 
 In addition, you may want to set the YARN cluster name to a reasonably unique value in case you