YARN-6202. Configuration item Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY is disregarded

(Contributed by Yufei Gu via Daniel Templeton)

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/MRAppMaster.java

@@ -280,8 +280,6 @@ public class MRAppMaster extends CompositeService {
     // create the job classloader if enabled
     createJobClassLoader(conf);
 
-    conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
-
     initJobCredentialsAndUGI(conf);
 
     dispatcher = createDispatcher();

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-hs/src/main/java/org/apache/hadoop/mapreduce/v2/hs/JobHistoryServer.java

@@ -43,7 +43,6 @@ import org.apache.hadoop.util.ShutdownHookManager;
 import org.apache.hadoop.util.StringUtils;
 import org.apache.hadoop.yarn.YarnUncaughtExceptionHandler;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
-import org.apache.hadoop.yarn.event.Dispatcher;
 import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
 import org.apache.hadoop.yarn.logaggregation.AggregatedLogDeletionService;
 
@@ -120,8 +119,6 @@ public class JobHistoryServer extends CompositeService {
   protected void serviceInit(Configuration conf) throws Exception {
     Configuration config = new YarnConfiguration(conf);
 
-    config.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
-
     // This is required for WebApps to use https if enabled.
     MRWebAppUtil.initialize(getConfig());
     try {

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/event/AsyncDispatcher.java

@@ -29,7 +29,6 @@ import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.classification.InterfaceAudience.Public;
 import org.apache.hadoop.classification.InterfaceStability.Evolving;
-import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.service.AbstractService;
 import org.apache.hadoop.util.ShutdownHookManager;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
@@ -72,7 +71,7 @@ public class AsyncDispatcher extends AbstractService implements Dispatcher {
 
   private Thread eventHandlingThread;
   protected final Map<Class<? extends Enum>, EventHandler> eventDispatchers;
-  private boolean exitOnDispatchException;
+  private boolean exitOnDispatchException = true;
 
   /**
    * The thread name for dispatcher.
@@ -131,12 +130,9 @@ public class AsyncDispatcher extends AbstractService implements Dispatcher {
     };
   }
 
-  @Override
-  protected void serviceInit(Configuration conf) throws Exception {
-    this.exitOnDispatchException =
-        conf.getBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY,
-          Dispatcher.DEFAULT_DISPATCHER_EXIT_ON_ERROR);
-    super.serviceInit(conf);
+  @VisibleForTesting
+  public void disableExitOnDispatchException() {
+    exitOnDispatchException = false;
   }
 
   @Override

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/event/Dispatcher.java

@@ -31,15 +31,6 @@ import org.apache.hadoop.classification.InterfaceStability.Evolving;
 @Evolving
 public interface Dispatcher {
 
-  // Configuration to make sure dispatcher crashes but doesn't do system-exit in
-  // case of errors. By default, it should be false, so that tests are not
-  // affected. For all daemons it should be explicitly set to true so that
-  // daemons can crash instead of hanging around.
-  public static final String DISPATCHER_EXIT_ON_ERROR_KEY =
-      "yarn.dispatcher.exit-on-error";
-
-  public static final boolean DEFAULT_DISPATCHER_EXIT_ON_ERROR = false;
-
   EventHandler<Event> getEventHandler();
 
   void register(Class<? extends Enum> eventType, EventHandler handler);

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/event/EventDispatcher.java

@@ -18,9 +18,9 @@
 
 package org.apache.hadoop.yarn.event;
 
+import com.google.common.annotations.VisibleForTesting;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
-import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.service.AbstractService;
 import org.apache.hadoop.util.ShutdownHookManager;
 import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
@@ -44,7 +44,7 @@ public class EventDispatcher<T extends Event> extends
       new LinkedBlockingDeque<>();
   private final Thread eventProcessor;
   private volatile boolean stopped = false;
-  private boolean shouldExitOnError = false;
+  private boolean shouldExitOnError = true;
 
   private static final Log LOG = LogFactory.getLog(EventDispatcher.class);
 
@@ -91,14 +91,6 @@ public class EventDispatcher<T extends Event> extends
     this.eventProcessor.setName(getName() + ":Event Processor");
   }
 
-  @Override
-  protected void serviceInit(Configuration conf) throws Exception {
-    this.shouldExitOnError =
-        conf.getBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY,
-            Dispatcher.DEFAULT_DISPATCHER_EXIT_ON_ERROR);
-    super.serviceInit(conf);
-  }
-
   @Override
   protected void serviceStart() throws Exception {
     this.eventProcessor.start();
@@ -134,4 +126,9 @@ public class EventDispatcher<T extends Event> extends
       LOG.info("Interrupted. Trying to exit gracefully.");
     }
   }
+
+  @VisibleForTesting
+  public void disableExitOnError() {
+    shouldExitOnError = false;
+  }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/test/java/org/apache/hadoop/yarn/event/DrainDispatcher.java

@@ -17,8 +17,6 @@
 */
 package org.apache.hadoop.yarn.event;
 
-import org.apache.hadoop.conf.Configuration;
-
 import java.util.concurrent.BlockingQueue;
 import java.util.concurrent.LinkedBlockingQueue;
 
@@ -36,13 +34,8 @@ public class DrainDispatcher extends AsyncDispatcher {
     super(eventQueue);
     this.queue = eventQueue;
     this.mutex = this;
-  }
-
-  @Override
-  public void serviceInit(Configuration conf)
-      throws Exception {
-    conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, false);
-    super.serviceInit(conf);
+    // Disable system exit since this class is only for unit tests.
+    disableExitOnDispatchException();
   }
 
   /**

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/NodeManager.java

@@ -299,9 +299,6 @@ public class NodeManager extends CompositeService
 
   @Override
   protected void serviceInit(Configuration conf) throws Exception {
-
-    conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
-
     rmWorkPreservingRestartEnabled = conf.getBoolean(YarnConfiguration
             .RM_WORK_PRESERVING_RECOVERY_ENABLED,
         YarnConfiguration.DEFAULT_RM_WORK_PRESERVING_RECOVERY_ENABLED);

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/DummyContainerManager.java

@@ -68,6 +68,7 @@ public class DummyContainerManager extends ContainerManagerImpl {
       NodeManagerMetrics metrics, LocalDirsHandlerService dirsHandler) {
     super(context, exec, deletionContext, nodeStatusUpdater, metrics,
         dirsHandler);
+    dispatcher.disableExitOnDispatchException();
   }
 
   @Override

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/TestContainerManagerRecovery.java

@@ -392,6 +392,7 @@ public class TestContainerManagerRecovery extends BaseContainerManagerTest {
     stateStore.start();
     Context context = createContext(conf, stateStore);
     ContainerManagerImpl cm = createContainerManager(context, delSrvc);
+    cm.dispatcher.disableExitOnDispatchException();
     cm.init(conf);
     cm.start();
     // add an application by starting a container
@@ -732,7 +733,7 @@ public class TestContainerManagerRecovery extends BaseContainerManagerTest {
           }
     };
 
-    return new ContainerManagerImpl(context,
+    ContainerManagerImpl containerManager = new ContainerManagerImpl(context,
         mock(ContainerExecutor.class), mock(DeletionService.class),
         mock(NodeStatusUpdater.class), metrics, null) {
           @Override
@@ -767,5 +768,7 @@ public class TestContainerManagerRecovery extends BaseContainerManagerTest {
             return null;
           }
     };
+    containerManager.dispatcher.disableExitOnDispatchException();
+    return containerManager;
   }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/localizer/TestResourceLocalizationService.java

@@ -100,7 +100,6 @@ import org.apache.hadoop.yarn.api.records.SerializedException;
 import org.apache.hadoop.yarn.api.records.URL;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
 import org.apache.hadoop.yarn.event.AsyncDispatcher;
-import org.apache.hadoop.yarn.event.Dispatcher;
 import org.apache.hadoop.yarn.event.DrainDispatcher;
 import org.apache.hadoop.yarn.event.EventHandler;
 import org.apache.hadoop.yarn.exceptions.YarnException;
@@ -1759,7 +1758,6 @@ public class TestResourceLocalizationService {
       sDirs[i] = localDirs.get(i).toString();
     }
     conf.setStrings(YarnConfiguration.NM_LOCAL_DIRS, sDirs);
-    conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
 
     DrainDispatcher dispatcher = new DrainDispatcher();
     EventHandler<ApplicationEvent> applicationBus = mock(EventHandler.class);

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

@@ -580,7 +580,6 @@ public class ResourceManager extends CompositeService implements Recoverable {
       activeServiceContext = new RMActiveServiceContext();
       rmContext.setActiveServiceContext(activeServiceContext);
 
-      conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
       rmSecretManagerService = createRMSecretManagerService();
       addService(rmSecretManagerService);
 

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

@@ -20,7 +20,6 @@ package org.apache.hadoop.yarn.server.resourcemanager;
 
 import java.io.IOException;
 import java.nio.ByteBuffer;
-import java.security.PrivilegedAction;
 import java.security.PrivilegedExceptionAction;
 import java.util.Arrays;
 import java.util.Collection;
@@ -67,6 +66,7 @@ import org.apache.hadoop.yarn.api.records.Resource;
 import org.apache.hadoop.yarn.api.records.ResourceRequest;
 import org.apache.hadoop.yarn.api.records.SignalContainerCommand;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
+import org.apache.hadoop.yarn.event.AsyncDispatcher;
 import org.apache.hadoop.yarn.event.Dispatcher;
 import org.apache.hadoop.yarn.event.DrainDispatcher;
 import org.apache.hadoop.yarn.event.EventHandler;
@@ -1273,4 +1273,12 @@ public class MockRM extends ResourceManager {
         false, false, null, 0, null, true, priority, null, applicationTimeouts,
         null);
   }
+
+  @Override
+  protected void serviceInit(Configuration conf) throws Exception {
+    super.serviceInit(conf);
+    if (getRmDispatcher() instanceof AsyncDispatcher) {
+      ((AsyncDispatcher) getRmDispatcher()).disableExitOnDispatchException();
+    }
+  }
 }

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

@@ -46,6 +46,7 @@ public class TestRMDispatcher {
     YarnConfiguration conf = new YarnConfiguration();
     EventDispatcher schedulerDispatcher =
         new EventDispatcher(sched, sched.getClass().getName());
+    schedulerDispatcher.disableExitOnError();
     rmDispatcher.register(SchedulerEventType.class, schedulerDispatcher);
     rmDispatcher.init(conf);
     rmDispatcher.start();

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/TestFSRMStateStore.java

@@ -71,6 +71,7 @@ public class TestFSRMStateStore extends RMStateStoreTestBase {
         init(conf);
         Assert.assertNull(fs);
         assertTrue(workingDirPathURI.equals(fsWorkingPath));
+        dispatcher.disableExitOnDispatchException();
         start();
         Assert.assertNotNull(fs);
       }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/TestLeveldbRMStateStore.java

@@ -159,6 +159,7 @@ public class TestLeveldbRMStateStore extends RMStateStoreTestBase {
       stateStore = new LeveldbRMStateStore();
       stateStore.init(conf);
       stateStore.start();
+      stateStore.dispatcher.disableExitOnDispatchException();
       return stateStore;
     }
 

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/TestZKRMStateStore.java

@@ -118,6 +118,7 @@ public class TestZKRMStateStore extends RMStateStoreTestBase {
           throws Exception {
         setResourceManager(new ResourceManager());
         init(conf);
+        dispatcher.disableExitOnDispatchException();
         start();
         assertTrue(znodeWorkingPath.equals(workingZnode));
       }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/TestAbstractYarnScheduler.java

@@ -807,8 +807,8 @@ public class TestAbstractYarnScheduler extends ParameterizedSchedulerTestBase {
     try {
       rm.start();
 
-      conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, false);
       DrainDispatcher privateDispatcher = new DrainDispatcher();
+      privateDispatcher.disableExitOnDispatchException();
       SleepHandler sleepHandler = new SleepHandler();
       ResourceTrackerService privateResourceTrackerService =
           getPrivateResourceTrackerService(privateDispatcher, rm, sleepHandler);