YARN-1029. Added embedded leader election in the ResourceManager. Contributed by Karthik Kambatla.

svn merge --ignore-ancestry -c 1556103 ../../trunk/


git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1556105 13f79535-47bb-0310-9956-ffa450edef68

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/ha/ActiveStandbyElector.java

@@ -766,8 +766,9 @@ public class ActiveStandbyElector implements StatCallback, StringCallback {
     zkClient = getNewZooKeeper();
     LOG.debug("Created new connection for " + this);
   }
-  
-  void terminateConnection() {
+
+  @InterfaceAudience.Private
+  public void terminateConnection() {
     if (zkClient == null) {
       return;
     }

hadoop-yarn-project/CHANGES.txt

@@ -37,6 +37,9 @@ Release 2.4.0 - UNRELEASED
     YARN-1028. Added FailoverProxyProvider capability to ResourceManager to help
     with RM failover. (Karthik Kambatla via vinodkv)
 
+    YARN-1029. Added embedded leader election in the ResourceManager. (Karthik
+    Kambatla via vinodkv)
+
   IMPROVEMENTS
 
     YARN-7. Support CPU resource for DistributedShell. (Junping Du via llu)

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/HAUtil.java

@@ -51,6 +51,22 @@ public class HAUtil {
         YarnConfiguration.DEFAULT_RM_HA_ENABLED);
   }
 
+  public static boolean isAutomaticFailoverEnabled(Configuration conf) {
+    return conf.getBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED,
+        YarnConfiguration.DEFAULT_AUTO_FAILOVER_ENABLED);
+  }
+
+  public static boolean isAutomaticFailoverEnabledAndEmbedded(
+      Configuration conf) {
+    return isAutomaticFailoverEnabled(conf) &&
+        isAutomaticFailoverEmbedded(conf);
+  }
+
+  public static boolean isAutomaticFailoverEmbedded(Configuration conf) {
+    return conf.getBoolean(YarnConfiguration.AUTO_FAILOVER_EMBEDDED,
+        YarnConfiguration.DEFAULT_AUTO_FAILOVER_EMBEDDED);
+  }
+
   /**
    * Verify configuration for Resource Manager HA.
    * @param conf Configuration
@@ -162,8 +178,7 @@ public class HAUtil {
    * @param conf Configuration. Please use verifyAndSetRMHAId to check.
    * @return RM Id on success
    */
-  @VisibleForTesting
-  static String getRMHAId(Configuration conf) {
+  public static String getRMHAId(Configuration conf) {
     return conf.get(YarnConfiguration.RM_HA_ID);
   }
 

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java

@@ -59,7 +59,7 @@ public class YarnConfiguration extends Configuration {
   public static final String IPC_PREFIX = YARN_PREFIX + "ipc.";
 
   /** Factory to create client IPC classes.*/
-  public static final String IPC_CLIENT_FACTORY_CLASS = 
+  public static final String IPC_CLIENT_FACTORY_CLASS =
     IPC_PREFIX + "client.factory.class";
   public static final String DEFAULT_IPC_CLIENT_FACTORY_CLASS = 
       "org.apache.hadoop.yarn.factories.impl.pb.RpcClientFactoryPBImpl";
@@ -87,6 +87,8 @@ public class YarnConfiguration extends Configuration {
   ////////////////////////////////
   public static final String RM_PREFIX = "yarn.resourcemanager.";
 
+  public static final String RM_CLUSTER_ID = RM_PREFIX + "cluster-id";
+
   /** The address of the applications manager interface in the RM.*/
   public static final String RM_ADDRESS = 
     RM_PREFIX + "address";
@@ -278,6 +280,36 @@ public class YarnConfiguration extends Configuration {
   public static final String RECOVERY_ENABLED = RM_PREFIX + "recovery.enabled";
   public static final boolean DEFAULT_RM_RECOVERY_ENABLED = false;
 
+  /** Zookeeper interaction configs */
+  public static final String RM_ZK_PREFIX = RM_PREFIX + "zk-";
+
+  public static final String RM_ZK_ADDRESS = RM_ZK_PREFIX + "address";
+
+  public static final String RM_ZK_NUM_RETRIES = RM_ZK_PREFIX + "num-retries";
+  public static final int DEFAULT_ZK_RM_NUM_RETRIES = 500;
+
+  public static final String RM_ZK_RETRY_INTERVAL_MS =
+      RM_ZK_PREFIX + "retry-interval-ms";
+  public static final long DEFAULT_RM_ZK_RETRY_INTERVAL_MS = 2000;
+
+  public static final String RM_ZK_TIMEOUT_MS = RM_ZK_PREFIX + "timeout-ms";
+  public static final int DEFAULT_RM_ZK_TIMEOUT_MS = 10000;
+
+  public static final String RM_ZK_ACL = RM_ZK_PREFIX + "acl";
+  public static final String DEFAULT_RM_ZK_ACL = "world:anyone:rwcda";
+
+  public static final String ZK_STATE_STORE_PREFIX =
+      RM_PREFIX + "zk-state-store.";
+
+  /** Parent znode path under which ZKRMStateStore will create znodes */
+  public static final String ZK_RM_STATE_STORE_PARENT_PATH =
+      ZK_STATE_STORE_PREFIX + "parent-path";
+  public static final String DEFAULT_ZK_RM_STATE_STORE_PARENT_PATH = "/rmstore";
+
+  /** Root node ACLs for fencing */
+  public static final String ZK_RM_STATE_STORE_ROOT_NODE_ACL =
+      ZK_STATE_STORE_PREFIX + "root-node.acl";
+
   /** HA related configs */
   public static final String RM_HA_PREFIX = RM_PREFIX + "ha.";
   public static final String RM_HA_ENABLED = RM_HA_PREFIX + "enabled";
@@ -296,6 +328,22 @@ public class YarnConfiguration extends Configuration {
           HttpConfig.isSecure() ? RM_WEBAPP_HTTPS_ADDRESS
               : RM_WEBAPP_ADDRESS));
 
+  public static final String AUTO_FAILOVER_PREFIX =
+      RM_HA_PREFIX + "automatic-failover.";
+
+  public static final String AUTO_FAILOVER_ENABLED =
+      AUTO_FAILOVER_PREFIX + "enabled";
+  public static final boolean DEFAULT_AUTO_FAILOVER_ENABLED = false;
+
+  public static final String AUTO_FAILOVER_EMBEDDED =
+      AUTO_FAILOVER_PREFIX + "embedded";
+  public static final boolean DEFAULT_AUTO_FAILOVER_EMBEDDED = false;
+
+  public static final String AUTO_FAILOVER_ZK_BASE_PATH =
+      AUTO_FAILOVER_PREFIX + "zk-base-path";
+  public static final String DEFAULT_AUTO_FAILOVER_ZK_BASE_PATH =
+      "/yarn-leader-election";
+
   public static final String CLIENT_FAILOVER_PREFIX =
       YARN_PREFIX + "client.failover-";
   public static final String CLIENT_FAILOVER_PROXY_PROVIDER =
@@ -334,36 +382,6 @@ public class YarnConfiguration extends Configuration {
       + "fs.state-store.retry-policy-spec";
   public static final String DEFAULT_FS_RM_STATE_STORE_RETRY_POLICY_SPEC =
       "2000, 500";
-  /**
-   * Comma separated host:port pairs, each corresponding to a ZK server for
-   * ZKRMStateStore
-   */
-  public static final String ZK_STATE_STORE_PREFIX =
-      RM_PREFIX + "zk-state-store.";
-  public static final String ZK_RM_STATE_STORE_NUM_RETRIES =
-      ZK_STATE_STORE_PREFIX + "num-retries";
-  public static final int DEFAULT_ZK_RM_STATE_STORE_NUM_RETRIES = 500;
-  /** retry interval when connecting to zookeeper*/
-  public static final String ZK_RM_STATE_STORE_RETRY_INTERVAL_MS =
-      ZK_STATE_STORE_PREFIX + "retry-interval-ms";
-  public static final long DEFAULT_ZK_RM_STATE_STORE_RETRY_INTERVAL_MS = 2000;
-  public static final String ZK_RM_STATE_STORE_ADDRESS =
-      ZK_STATE_STORE_PREFIX + "address";
-  /** Timeout in millisec for ZK server connection for ZKRMStateStore */
-  public static final String ZK_RM_STATE_STORE_TIMEOUT_MS =
-      ZK_STATE_STORE_PREFIX + "timeout-ms";
-  public static final int DEFAULT_ZK_RM_STATE_STORE_TIMEOUT_MS = 60000;
-  /** Parent znode path under which ZKRMStateStore will create znodes */
-  public static final String ZK_RM_STATE_STORE_PARENT_PATH =
-      ZK_STATE_STORE_PREFIX + "parent-path";
-  public static final String DEFAULT_ZK_RM_STATE_STORE_PARENT_PATH = "/rmstore";
-  /** ACL for znodes in ZKRMStateStore */
-  public static final String ZK_RM_STATE_STORE_ACL =
-      ZK_STATE_STORE_PREFIX + "acl";
-  public static final String DEFAULT_ZK_RM_STATE_STORE_ACL =
-      "world:anyone:rwcda";
-  public static final String ZK_RM_STATE_STORE_ROOT_NODE_ACL =
-      ZK_STATE_STORE_PREFIX + "root-node.acl";
 
   /** The maximum number of completed applications RM keeps. */ 
   public static final String RM_MAX_COMPLETED_APPLICATIONS =

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/proto/server/yarn_server_resourcemanager_service_protos.proto

@@ -133,3 +133,11 @@ message RMStateVersionProto {
   optional int32 major_version = 1;
   optional int32 minor_version = 2;
 }
+
+//////////////////////////////////////////////////////////////////
+///////////// RM Failover related records ////////////////////////
+//////////////////////////////////////////////////////////////////
+message ActiveRMInfoProto {
+  required string clusterid = 1;
+  required string rmId = 2;
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/pom.xml

@@ -30,7 +30,13 @@
   </properties>
 
   <dependencies>
-  	<dependency>
+    <dependency>
+      <groupId>org.apache.zookeeper</groupId>
+      <artifactId>zookeeper</artifactId>
+      <type>test-jar</type>
+      <scope>test</scope>
+    </dependency>
+    <dependency>
   		<groupId>org.apache.hadoop</groupId>
   		<artifactId>hadoop-yarn-api</artifactId>
   	</dependency>

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/test/java/org/apache/hadoop/yarn/client/TestRMFailover.java

@@ -28,37 +28,39 @@ import java.io.IOException;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.ha.ClientBaseWithFixes;
 import org.apache.hadoop.ha.HAServiceProtocol;
+import org.apache.hadoop.ha.proto.HAServiceProtocolProtos;
 import org.apache.hadoop.yarn.client.api.YarnClient;
 import org.apache.hadoop.yarn.conf.HAUtil;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
 import org.apache.hadoop.yarn.exceptions.YarnException;
 import org.apache.hadoop.yarn.server.MiniYARNCluster;
 import org.apache.hadoop.yarn.server.resourcemanager.AdminService;
-import org.junit.AfterClass;
-import org.junit.BeforeClass;
+import org.junit.After;
+import org.junit.Before;
 import org.junit.Test;
 
-public class TestRMFailover {
+public class TestRMFailover extends ClientBaseWithFixes {
   private static final Log LOG =
       LogFactory.getLog(TestRMFailover.class.getName());
+  private static final HAServiceProtocol.StateChangeRequestInfo req =
+      new HAServiceProtocol.StateChangeRequestInfo(
+          HAServiceProtocol.RequestSource.REQUEST_BY_USER);
 
   private static final String RM1_NODE_ID = "rm1";
   private static final int RM1_PORT_BASE = 10000;
   private static final String RM2_NODE_ID = "rm2";
   private static final int RM2_PORT_BASE = 20000;
-  private static final HAServiceProtocol.StateChangeRequestInfo req =
-      new HAServiceProtocol.StateChangeRequestInfo(
-          HAServiceProtocol.RequestSource.REQUEST_BY_USER_FORCED);
 
-  private static Configuration conf;
-  private static MiniYARNCluster cluster;
+  private Configuration conf;
+  private MiniYARNCluster cluster;
 
-  private static void setConfForRM(String rmId, String prefix, String value) {
+  private void setConfForRM(String rmId, String prefix, String value) {
     conf.set(HAUtil.addSuffix(prefix, rmId), value);
   }
 
-  private static void setRpcAddressForRM(String rmId, int base) {
+  private void setRpcAddressForRM(String rmId, int base) {
     setConfForRM(rmId, YarnConfiguration.RM_ADDRESS, "0.0.0.0:" +
         (base + YarnConfiguration.DEFAULT_RM_PORT));
     setConfForRM(rmId, YarnConfiguration.RM_SCHEDULER_ADDRESS, "0.0.0.0:" +
@@ -73,13 +75,8 @@ public class TestRMFailover {
         (base + YarnConfiguration.DEFAULT_RM_WEBAPP_HTTPS_PORT));
   }
 
-  private static AdminService getRMAdminService(int index) {
-    return
-        cluster.getResourceManager(index).getRMContext().getRMAdminService();
-  }
-
-  @BeforeClass
-  public static void setup() throws IOException {
+  @Before
+  public void setup() throws IOException {
     conf = new YarnConfiguration();
     conf.setBoolean(YarnConfiguration.RM_HA_ENABLED, true);
     conf.set(YarnConfiguration.RM_HA_IDS, RM1_NODE_ID + "," + RM2_NODE_ID);
@@ -87,27 +84,22 @@ public class TestRMFailover {
     setRpcAddressForRM(RM2_NODE_ID, RM2_PORT_BASE);
 
     conf.setLong(YarnConfiguration.CLIENT_FAILOVER_SLEEPTIME_BASE_MS, 100L);
+
     conf.setBoolean(YarnConfiguration.YARN_MINICLUSTER_FIXED_PORTS, true);
     conf.setBoolean(YarnConfiguration.YARN_MINICLUSTER_USE_RPC, true);
 
     cluster = new MiniYARNCluster(TestRMFailover.class.getName(), 2, 1, 1, 1);
-    cluster.init(conf);
-    cluster.start();
-
-    cluster.getResourceManager(0).getRMContext().getRMAdminService()
-        .transitionToActive(req);
-    assertFalse("RM never turned active", -1 == cluster.getActiveRMIndex());
   }
 
-  @AfterClass
-  public static void teardown() {
+  @After
+  public void teardown() {
     cluster.stop();
   }
 
   private void verifyClientConnection() {
     int numRetries = 3;
     while(numRetries-- > 0) {
-      Configuration conf = new YarnConfiguration(TestRMFailover.conf);
+      Configuration conf = new YarnConfiguration(this.conf);
       YarnClient client = YarnClient.createYarnClient();
       client.init(conf);
       client.start();
@@ -123,31 +115,68 @@ public class TestRMFailover {
     fail("Client couldn't connect to the Active RM");
   }
 
+  private void verifyConnections() throws InterruptedException, YarnException {
+    assertTrue("NMs failed to connect to the RM",
+        cluster.waitForNodeManagersToConnect(20000));
+    verifyClientConnection();
+  }
+
+  private AdminService getAdminService(int index) {
+    return cluster.getResourceManager(index).getRMContext().getRMAdminService();
+  }
+
+  private void explicitFailover() throws IOException {
+    int activeRMIndex = cluster.getActiveRMIndex();
+    int newActiveRMIndex = (activeRMIndex + 1) % 2;
+    getAdminService(activeRMIndex).transitionToStandby(req);
+    getAdminService(newActiveRMIndex).transitionToActive(req);
+    assertEquals("Failover failed", newActiveRMIndex, cluster.getActiveRMIndex());
+  }
+
+  private void failover()
+      throws IOException, InterruptedException, YarnException {
+    int activeRMIndex = cluster.getActiveRMIndex();
+    cluster.stopResourceManager(activeRMIndex);
+    assertEquals("Failover failed",
+        (activeRMIndex + 1) % 2, cluster.getActiveRMIndex());
+    cluster.restartResourceManager(activeRMIndex);
+  }
+
   @Test
   public void testExplicitFailover()
       throws YarnException, InterruptedException, IOException {
-    assertTrue("NMs failed to connect to the RM",
-        cluster.waitForNodeManagersToConnect(5000));
-    verifyClientConnection();
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, false);
+    cluster.init(conf);
+    cluster.start();
+    getAdminService(0).transitionToActive(req);
+    assertFalse("RM never turned active", -1 == cluster.getActiveRMIndex());
+    verifyConnections();
 
-    // Failover to the second RM
-    getRMAdminService(0).transitionToStandby(req);
-    getRMAdminService(1).transitionToActive(req);
-    assertEquals("Wrong ResourceManager is active",
-        HAServiceProtocol.HAServiceState.ACTIVE,
-        getRMAdminService(1).getServiceStatus().getState());
-    assertTrue("NMs failed to connect to the RM",
-        cluster.waitForNodeManagersToConnect(5000));
-    verifyClientConnection();
+    explicitFailover();
+    verifyConnections();
 
-    // Failover back to the first RM
-    getRMAdminService(1).transitionToStandby(req);
-    getRMAdminService(0).transitionToActive(req);
-    assertEquals("Wrong ResourceManager is active",
-        HAServiceProtocol.HAServiceState.ACTIVE,
-        getRMAdminService(0).getServiceStatus().getState());
-    assertTrue("NMs failed to connect to the RM",
-        cluster.waitForNodeManagersToConnect(5000));
-    verifyClientConnection();
+    explicitFailover();
+    verifyConnections();
+  }
+
+  @Test
+  public void testAutomaticFailover()
+      throws YarnException, InterruptedException, IOException {
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, true);
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_EMBEDDED, true);
+    conf.set(YarnConfiguration.RM_CLUSTER_ID, "yarn-test-cluster");
+    conf.set(YarnConfiguration.RM_ZK_ADDRESS, hostPort);
+    conf.setInt(YarnConfiguration.RM_ZK_TIMEOUT_MS, 2000);
+
+    cluster.init(conf);
+    cluster.start();
+    assertFalse("RM never turned active", -1 == cluster.getActiveRMIndex());
+    verifyConnections();
+
+    failover();
+    verifyConnections();
+
+    failover();
+    verifyConnections();
   }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/client/RMHAServiceTarget.java

@@ -21,16 +21,19 @@ package org.apache.hadoop.yarn.client;
 import org.apache.hadoop.ha.BadFencingConfigurationException;
 import org.apache.hadoop.ha.HAServiceTarget;
 import org.apache.hadoop.ha.NodeFencer;
+import org.apache.hadoop.yarn.conf.HAUtil;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
 
 import java.io.IOException;
 import java.net.InetSocketAddress;
 
 public class RMHAServiceTarget extends HAServiceTarget {
-  private InetSocketAddress haAdminServiceAddress;
+  private final boolean autoFailoverEnabled;
+  private final InetSocketAddress haAdminServiceAddress;
 
   public RMHAServiceTarget(YarnConfiguration conf)
       throws IOException {
+    autoFailoverEnabled = HAUtil.isAutomaticFailoverEnabled(conf);
     haAdminServiceAddress = conf.getSocketAddr(
         YarnConfiguration.RM_ADMIN_ADDRESS,
         YarnConfiguration.DEFAULT_RM_ADMIN_ADDRESS,
@@ -44,19 +47,23 @@ public class RMHAServiceTarget extends HAServiceTarget {
 
   @Override
   public InetSocketAddress getZKFCAddress() {
-    // TODO (YARN-1177): Hook up ZKFC information
-    return null;
+    // TODO (YARN-1177): ZKFC implementation
+    throw new UnsupportedOperationException("RMHAServiceTarget doesn't have " +
+        "a corresponding ZKFC address");
   }
 
   @Override
   public NodeFencer getFencer() {
-    // TODO (YARN-1026): Hook up fencing implementation
     return null;
   }
 
   @Override
-  public void checkFencingConfigured()
-      throws BadFencingConfigurationException {
-    // TODO (YARN-1026): Based on fencing implementation
+  public void checkFencingConfigured() throws BadFencingConfigurationException {
+    throw new BadFencingConfigurationException("Fencer not configured");
+  }
+
+  @Override
+  public boolean isAutoFailoverEnabled() {
+    return autoFailoverEnabled;
   }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml

@@ -264,13 +264,20 @@
 
   <property>
     <description>Enable RM to recover state after starting. If true, then 
-    yarn.resourcemanager.store.class must be specified</description>
+      yarn.resourcemanager.store.class must be specified. </description>
     <name>yarn.resourcemanager.recovery.enabled</name>
     <value>false</value>
   </property>
 
   <property>
-    <description>The class to use as the persistent store.</description>
+    <description>The class to use as the persistent store.
+
+      If org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
+      is used, the store is implicitly fenced; meaning a single ResourceManager
+      is able to use the store at any point in time. More details on this
+      implicit fencing, along with setting up appropriate ACLs is discussed
+      under yarn.resourcemanager.zk-state-store.root-node.acl.
+    </description>
     <name>yarn.resourcemanager.store.class</name>
     <value>org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore</value>
   </property>
@@ -291,31 +298,24 @@
   </property>
 
   <property>
-    <description>Host:Port of the ZooKeeper server where RM state will 
-    be stored. This must be supplied when using
-    org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
-    as the value for yarn.resourcemanager.store.class. ZKRMStateStore
-    is implicitly fenced, meaning a single ResourceManager is
-    able to use the store at any point in time. More details on this, along
-    with setting up appropriate ACLs is discussed under the description for
-    yarn.resourcemanager.zk-state-store.root-node.acl.</description>
-    <name>yarn.resourcemanager.zk-state-store.address</name>
+    <description>Host:Port of the ZooKeeper server to be used by the RM. This
+      must be supplied when using the ZooKeeper based implementation of the
+      RM state store and/or embedded automatic failover in a HA setting.
+    </description>
+    <name>yarn.resourcemanager.zk-address</name>
     <!--value>127.0.0.1:2181</value-->
   </property>
 
   <property>
-    <description>Number of times ZKRMStateStore tries to connect to
-    ZooKeeper. This may be supplied when using
-    org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
-    as the value for yarn.resourcemanager.store.class</description>
-    <name>yarn.resourcemanager.zk-state-store.num-retries</name>
+    <description>Number of times RM tries to connect to ZooKeeper.</description>
+    <name>yarn.resourcemanager.zk-num-retries</name>
     <value>500</value>
   </property>
 
   <property>
-    <description>Retry interval in milliseconds when ZKRMStateStore tries to
-    connect to ZooKeeper.</description>
-    <name>yarn.resourcemanager.zk-state-store.retry-interval-ms</name>
+    <description>Retry interval in milliseconds when connecting to ZooKeeper.
+    </description>
+    <name>yarn.resourcemanager.zk-retry-interval-ms</name>
     <value>2000</value>
   </property>
 
@@ -333,20 +333,14 @@
     is managed by the ZooKeeper cluster itself, not by the client. This value is
     used by the cluster to determine when the client's session expires.
     Expirations happens when the cluster does not hear from the client within
-    the specified session timeout period (i.e. no heartbeat).
-    This may be supplied when using
-    org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
-    as the value for yarn.resourcemanager.store.class</description>
-    <name>yarn.resourcemanager.zk-state-store.timeout-ms</name>
+    the specified session timeout period (i.e. no heartbeat).</description>
+    <name>yarn.resourcemanager.zk-timeout-ms</name>
     <value>60000</value>
   </property>
 
   <property>
-    <description>ACL's to be used for ZooKeeper znodes.
-    This may be supplied when using
-    org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
-    as the value for yarn.resourcemanager.store.class</description>
-    <name>yarn.resourcemanager.zk-state-store.acl</name>
+    <description>ACL's to be used for ZooKeeper znodes.</description>
+    <name>yarn.resourcemanager.zk-acl</name>
     <value>world:anyone:rwcda</value>
   </property>
 
@@ -362,7 +356,7 @@
       permissions.
 
       By default, when this property is not set, we use the ACLs from
-      yarn.resourcemanager.zk-state-store.acl for shared admin access and
+      yarn.resourcemanager.zk-acl for shared admin access and
       rm-address:cluster-timestamp for username-based exclusive create-delete
       access.
 
@@ -408,6 +402,36 @@
     <value>false</value>
   </property>
 
+  <property>
+    <description>Enable automatic failover.</description>
+    <name>yarn.resourcemanager.ha.automatic-failover.enabled</name>
+    <value>false</value>
+  </property>
+
+  <property>
+    <description>Enable embedded automatic failover. The embedded elector
+      relies on the RM state store to handle fencing, and is primarily intended
+      to be used in conjunction with ZKRMStateStore.</description>
+    <name>yarn.resourcemanager.ha.automatic-failover.embedded</name>
+    <value>false</value>
+  </property>
+
+  <property>
+    <description>The base znode path to use for storing leader information,
+      when using ZooKeeper based leader election.</description>
+    <name>yarn.resourcemanager.ha.automatic-failover.zk-base-path</name>
+    <value>/yarn-leader-election</value>
+  </property>
+
+  <property>
+    <description>Name of the cluster. In a HA setting,
+      this is used to ensure the RM participates in leader
+      election fo this cluster and ensures it does not affect
+      other clusters</description>
+    <name>yarn.resourcemanager.cluster-id</name>
+    <!--value>yarn-cluster</value-->
+  </property>
+
   <property>
     <description>The list of RM nodes in the cluster when HA is
       enabled. See description of yarn.resourcemanager.ha

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

@@ -20,11 +20,8 @@ package org.apache.hadoop.yarn.server.resourcemanager;
 
 import java.io.IOException;
 import java.net.InetSocketAddress;
-import java.util.ArrayList;
-import java.util.List;
 import java.util.Map;
 import java.util.Set;
-import java.util.concurrent.ConcurrentMap;
 
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
@@ -41,14 +38,16 @@ import org.apache.hadoop.ipc.ProtobufRpcEngine;
 import org.apache.hadoop.ipc.RPC;
 import org.apache.hadoop.ipc.RPC.Server;
 import org.apache.hadoop.ipc.StandbyException;
+import org.apache.hadoop.security.AccessControlException;
 import org.apache.hadoop.security.Groups;
 import org.apache.hadoop.security.UserGroupInformation;
 import org.apache.hadoop.security.authorize.AccessControlList;
 import org.apache.hadoop.security.authorize.PolicyProvider;
 import org.apache.hadoop.security.authorize.ProxyUsers;
-import org.apache.hadoop.service.AbstractService;
+import org.apache.hadoop.service.CompositeService;
 import org.apache.hadoop.yarn.api.records.NodeId;
 import org.apache.hadoop.yarn.api.records.ResourceOption;
+import org.apache.hadoop.yarn.conf.HAUtil;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
 import org.apache.hadoop.yarn.exceptions.YarnException;
 import org.apache.hadoop.yarn.factories.RecordFactory;
@@ -75,7 +74,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.security.authorize.RMPolicy
 
 import com.google.protobuf.BlockingService;
 
-public class AdminService extends AbstractService implements
+public class AdminService extends CompositeService implements
     HAServiceProtocol, ResourceManagerAdministrationProtocol {
 
   private static final Log LOG = LogFactory.getLog(AdminService.class);
@@ -84,6 +83,8 @@ public class AdminService extends AbstractService implements
   private final ResourceManager rm;
   private String rmId;
 
+  private boolean autoFailoverEnabled;
+
   private Server server;
   private InetSocketAddress masterServiceAddress;
   private AccessControlList adminAcl;
@@ -99,6 +100,15 @@ public class AdminService extends AbstractService implements
 
   @Override
   public synchronized void serviceInit(Configuration conf) throws Exception {
+    if (rmContext.isHAEnabled()) {
+      autoFailoverEnabled = HAUtil.isAutomaticFailoverEnabled(conf);
+      if (autoFailoverEnabled) {
+        if (HAUtil.isAutomaticFailoverEmbedded(conf)) {
+          addIfService(createEmbeddedElectorService());
+        }
+      }
+    }
+
     masterServiceAddress = conf.getSocketAddr(
         YarnConfiguration.RM_ADMIN_ADDRESS,
         YarnConfiguration.DEFAULT_RM_ADMIN_ADDRESS,
@@ -162,6 +172,10 @@ public class AdminService extends AbstractService implements
     }
   }
 
+  protected EmbeddedElectorService createEmbeddedElectorService() {
+    return new EmbeddedElectorService(rmContext);
+  }
+
   private UserGroupInformation checkAccess(String method) throws IOException {
     return RMServerUtils.verifyAccess(adminAcl, method, LOG);
   }
@@ -174,6 +188,43 @@ public class AdminService extends AbstractService implements
     }
   }
 
+  /**
+   * Check that a request to change this node's HA state is valid.
+   * In particular, verifies that, if auto failover is enabled, non-forced
+   * requests from the HAAdmin CLI are rejected, and vice versa.
+   *
+   * @param req the request to check
+   * @throws AccessControlException if the request is disallowed
+   */
+  private void checkHaStateChange(StateChangeRequestInfo req)
+      throws AccessControlException {
+    switch (req.getSource()) {
+      case REQUEST_BY_USER:
+        if (autoFailoverEnabled) {
+          throw new AccessControlException(
+              "Manual failover for this ResourceManager is disallowed, " +
+                  "because automatic failover is enabled.");
+        }
+        break;
+      case REQUEST_BY_USER_FORCED:
+        if (autoFailoverEnabled) {
+          LOG.warn("Allowing manual failover from " +
+              org.apache.hadoop.ipc.Server.getRemoteAddress() +
+              " even though automatic failover is enabled, because the user " +
+              "specified the force flag");
+        }
+        break;
+      case REQUEST_BY_ZKFC:
+        if (!autoFailoverEnabled) {
+          throw new AccessControlException(
+              "Request from ZK failover controller at " +
+                  org.apache.hadoop.ipc.Server.getRemoteAddress() + " denied " +
+                  "since automatic failover is not enabled");
+        }
+        break;
+    }
+  }
+
   private synchronized boolean isRMActive() {
     return HAServiceState.ACTIVE == rmContext.getHAServiceState();
   }
@@ -196,8 +247,7 @@ public class AdminService extends AbstractService implements
   public synchronized void transitionToActive(
       HAServiceProtocol.StateChangeRequestInfo reqInfo) throws IOException {
     UserGroupInformation user = checkAccess("transitionToActive");
-    // TODO (YARN-1177): When automatic failover is enabled,
-    // check if transition should be allowed for this request
+    checkHaStateChange(reqInfo);
     try {
       rm.transitionToActive();
       RMAuditLogger.logSuccess(user.getShortUserName(),
@@ -215,8 +265,7 @@ public class AdminService extends AbstractService implements
   public synchronized void transitionToStandby(
       HAServiceProtocol.StateChangeRequestInfo reqInfo) throws IOException {
     UserGroupInformation user = checkAccess("transitionToStandby");
-    // TODO (YARN-1177): When automatic failover is enabled,
-    // check if transition should be allowed for this request
+    checkHaStateChange(reqInfo);
     try {
       rm.transitionToStandby(true);
       RMAuditLogger.logSuccess(user.getShortUserName(),
@@ -394,7 +443,7 @@ public class AdminService extends AbstractService implements
   public String[] getGroupsForUser(String user) throws IOException {
     return UserGroupInformation.createRemoteUser(user).getGroupNames();
   }
-  
+
   @Override
   public UpdateNodeResourceResponse updateNodeResource(
       UpdateNodeResourceRequest request) throws YarnException, IOException {

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

@@ -0,0 +1,209 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.yarn.server.resourcemanager;
+
+import com.google.protobuf.InvalidProtocolBufferException;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceStability;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.ha.ActiveStandbyElector;
+import org.apache.hadoop.ha.HAServiceProtocol;
+import org.apache.hadoop.ha.ServiceFailedException;
+import org.apache.hadoop.service.AbstractService;
+import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.util.ZKUtil;
+import org.apache.hadoop.yarn.conf.HAUtil;
+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.proto.YarnServerResourceManagerServiceProtos;
+import org.apache.zookeeper.KeeperException;
+import org.apache.zookeeper.data.ACL;
+
+import java.io.IOException;
+import java.util.Collections;
+import java.util.List;
+
+@InterfaceAudience.Private
+@InterfaceStability.Unstable
+public class EmbeddedElectorService extends AbstractService
+    implements ActiveStandbyElector.ActiveStandbyElectorCallback {
+  private static final Log LOG =
+      LogFactory.getLog(EmbeddedElectorService.class.getName());
+  private static final HAServiceProtocol.StateChangeRequestInfo req =
+      new HAServiceProtocol.StateChangeRequestInfo(
+          HAServiceProtocol.RequestSource.REQUEST_BY_ZKFC);
+
+  private RMContext rmContext;
+
+  private byte[] localActiveNodeInfo;
+  private ActiveStandbyElector elector;
+
+  EmbeddedElectorService(RMContext rmContext) {
+    super(EmbeddedElectorService.class.getName());
+    this.rmContext = rmContext;
+  }
+
+  @Override
+  protected synchronized void serviceInit(Configuration conf)
+      throws Exception {
+    conf = conf instanceof YarnConfiguration ? conf : new YarnConfiguration(conf);
+
+    String zkQuorum = conf.get(YarnConfiguration.RM_ZK_ADDRESS);
+    if (zkQuorum == null) {
+     throw new YarnRuntimeException("Embedded automatic failover " +
+          "is enabled, but " + YarnConfiguration.RM_ZK_ADDRESS +
+          " is not set");
+    }
+
+    String rmId = HAUtil.getRMHAId(conf);
+    String clusterId = conf.get(YarnConfiguration.RM_CLUSTER_ID);
+    if (clusterId == null) {
+      throw new YarnRuntimeException(YarnConfiguration.RM_CLUSTER_ID +
+          " is not specified!");
+    }
+    localActiveNodeInfo = createActiveNodeInfo(clusterId, rmId);
+
+    String zkBasePath = conf.get(YarnConfiguration.AUTO_FAILOVER_ZK_BASE_PATH,
+        YarnConfiguration.DEFAULT_AUTO_FAILOVER_ZK_BASE_PATH);
+    String electionZNode = zkBasePath + "/" + clusterId;
+
+    long zkSessionTimeout = conf.getLong(YarnConfiguration.RM_ZK_TIMEOUT_MS,
+        YarnConfiguration.DEFAULT_RM_ZK_TIMEOUT_MS);
+
+    String zkAclConf = conf.get(YarnConfiguration.RM_ZK_ACL,
+        YarnConfiguration.DEFAULT_RM_ZK_ACL);
+    List<ACL> zkAcls;
+    try {
+      zkAcls = ZKUtil.parseACLs(ZKUtil.resolveConfIndirection(zkAclConf));
+    } catch (ZKUtil.BadAclFormatException bafe) {
+      throw new YarnRuntimeException(
+          YarnConfiguration.RM_ZK_ACL + "has ill-formatted ACLs");
+    }
+
+    // TODO (YARN-1528): ZKAuthInfo to be set for rm-store and elector
+    List<ZKUtil.ZKAuthInfo> zkAuths = Collections.emptyList();
+
+    elector = new ActiveStandbyElector(zkQuorum, (int) zkSessionTimeout,
+        electionZNode, zkAcls, zkAuths, this);
+
+    elector.ensureParentZNode();
+    if (!isParentZnodeSafe(clusterId)) {
+      notifyFatalError(electionZNode + " znode has invalid data! "+
+          "Might need formatting!");
+    }
+
+    super.serviceInit(conf);
+  }
+
+  @Override
+  protected synchronized void serviceStart() throws Exception {
+    elector.joinElection(localActiveNodeInfo);
+    super.serviceStart();
+  }
+
+  @Override
+  protected synchronized void serviceStop() throws Exception {
+    elector.quitElection(false);
+    elector.terminateConnection();
+    super.serviceStop();
+  }
+
+  @Override
+  public synchronized void becomeActive() throws ServiceFailedException {
+    try {
+      rmContext.getRMAdminService().transitionToActive(req);
+    } catch (Exception e) {
+      throw new ServiceFailedException("RM could not transition to Active", e);
+    }
+  }
+
+  @Override
+  public synchronized void becomeStandby() {
+    try {
+      rmContext.getRMAdminService().transitionToStandby(req);
+    } catch (Exception e) {
+      LOG.error("RM could not transition to Standby", e);
+    }
+  }
+
+  @Override
+  public void enterNeutralMode() {
+    /**
+     * Possibly due to transient connection issues. Do nothing.
+     * TODO: Might want to keep track of how long in this state and transition
+     * to standby.
+     */
+  }
+
+  @SuppressWarnings(value = "unchecked")
+  @Override
+  public synchronized void notifyFatalError(String errorMessage) {
+    rmContext.getDispatcher().getEventHandler().handle(
+        new RMFatalEvent(RMFatalEventType.EMBEDDED_ELECTOR_FAILED, errorMessage));
+  }
+
+  @Override
+  public synchronized void fenceOldActive(byte[] oldActiveData) {
+    if (LOG.isDebugEnabled()) {
+      LOG.debug("Request to fence old active being ignored, " +
+          "as embedded leader election doesn't support fencing");
+    }
+  }
+
+  private static byte[] createActiveNodeInfo(String clusterId, String rmId)
+      throws IOException {
+    return YarnServerResourceManagerServiceProtos.ActiveRMInfoProto
+        .newBuilder()
+        .setClusterid(clusterId)
+        .setRmId(rmId)
+        .build()
+        .toByteArray();
+  }
+
+  private synchronized boolean isParentZnodeSafe(String clusterId)
+      throws InterruptedException, IOException, KeeperException {
+    byte[] data;
+    try {
+      data = elector.getActiveData();
+    } catch (ActiveStandbyElector.ActiveNotFoundException e) {
+      // no active found, parent znode is safe
+      return true;
+    }
+
+    YarnServerResourceManagerServiceProtos.ActiveRMInfoProto proto;
+    try {
+      proto = YarnServerResourceManagerServiceProtos.ActiveRMInfoProto
+          .parseFrom(data);
+    } catch (InvalidProtocolBufferException e) {
+      LOG.error("Invalid data in ZK: " + StringUtils.byteToHexString(data));
+      return false;
+    }
+
+    // Check if the passed proto corresponds to an RM in the same cluster
+    if (!proto.getClusterid().equals(clusterId)) {
+      LOG.error("Mismatched cluster! The other RM seems " +
+          "to be from a different cluster. Current cluster = " + clusterId +
+          "Other RM's cluster = " + proto.getClusterid());
+      return false;
+    }
+    return true;
+  }
+}

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

@@ -0,0 +1,37 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.yarn.server.resourcemanager;
+
+import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.yarn.event.AbstractEvent;
+
+public class RMFatalEvent extends AbstractEvent<RMFatalEventType> {
+  private String cause;
+
+  public RMFatalEvent(RMFatalEventType rmFatalEventType, String cause) {
+    super(rmFatalEventType);
+    this.cause = cause;
+  }
+
+  public RMFatalEvent(RMFatalEventType rmFatalEventType, Exception cause) {
+    super(rmFatalEventType);
+    this.cause = StringUtils.stringifyException(cause);
+  }
+
+  public String getCause() {return this.cause;}
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/RMStateStoreOperationFailedEventType.java → hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMFatalEventType.java

@@ -15,9 +15,17 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
-package org.apache.hadoop.yarn.server.resourcemanager.recovery;
 
-public enum RMStateStoreOperationFailedEventType {
-  FENCED, // Store operation failed because it was fenced
-  FAILED // Store operation failed for no known reason
+package org.apache.hadoop.yarn.server.resourcemanager;
+
+import org.apache.hadoop.classification.InterfaceAudience;
+
+@InterfaceAudience.Private
+public enum RMFatalEventType {
+  // Source <- Store
+  STATE_STORE_FENCED,
+  STATE_STORE_OP_FAILED,
+
+  // Source <- Embedded Elector
+  EMBEDDED_ELECTOR_FAILED
 }

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

@@ -59,8 +59,6 @@ import org.apache.hadoop.yarn.server.resourcemanager.recovery.NullRMStateStore;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.RMState;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStoreFactory;
-import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStoreOperationFailedEvent;
-import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStoreOperationFailedEventType;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.Recoverable;
 import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp;
 import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMAppEvent;
@@ -121,6 +119,7 @@ public class ResourceManager extends CompositeService implements Recoverable {
    */
   @VisibleForTesting
   protected RMContextImpl rmContext;
+  private Dispatcher rmDispatcher;
   @VisibleForTesting
   protected AdminService adminService;
 
@@ -134,7 +133,6 @@ public class ResourceManager extends CompositeService implements Recoverable {
    */
   protected RMActiveServices activeServices;
   protected RMSecretManagerService rmSecretManagerService;
-  private Dispatcher rmDispatcher;
 
   protected ResourceScheduler scheduler;
   private ClientRMService clientRM;
@@ -179,6 +177,13 @@ public class ResourceManager extends CompositeService implements Recoverable {
     this.conf = conf;
     this.rmContext = new RMContextImpl();
 
+    rmDispatcher = createDispatcher();
+    addIfService(rmDispatcher);
+    rmContext.setDispatcher(rmDispatcher);
+
+    rmDispatcher.register(RMFatalEventType.class,
+        new ResourceManager.RMFatalEventDispatcher(this.rmContext, this));
+
     adminService = createAdminService();
     addService(adminService);
     rmContext.setRMAdminService(adminService);
@@ -207,11 +212,6 @@ public class ResourceManager extends CompositeService implements Recoverable {
     return new SchedulerEventDispatcher(this.scheduler);
   }
 
-  protected RMStateStoreOperationFailedEventDispatcher
-      createRMStateStoreOperationFailedEventDispatcher() {
-    return new RMStateStoreOperationFailedEventDispatcher(rmContext, this);
-  }
-
   protected Dispatcher createDispatcher() {
     return new AsyncDispatcher();
   }
@@ -297,10 +297,6 @@ public class ResourceManager extends CompositeService implements Recoverable {
     protected void serviceInit(Configuration configuration) throws Exception {
       conf.setBoolean(Dispatcher.DISPATCHER_EXIT_ON_ERROR_KEY, true);
 
-      rmDispatcher = createDispatcher();
-      addIfService(rmDispatcher);
-      rmContext.setDispatcher(rmDispatcher);
-
       rmSecretManagerService = createRMSecretManagerService();
       addService(rmSecretManagerService);
 
@@ -332,8 +328,6 @@ public class ResourceManager extends CompositeService implements Recoverable {
       try {
         rmStore.init(conf);
         rmStore.setRMDispatcher(rmDispatcher);
-        rmDispatcher.register(RMStateStoreOperationFailedEventType.class,
-            createRMStateStoreOperationFailedEventDispatcher());
       } catch (Exception e) {
         // the Exception from stateStore.init() needs to be handled for
         // HA and we need to give up master status if we got fenced
@@ -605,26 +599,23 @@ public class ResourceManager extends CompositeService implements Recoverable {
   }
 
   @Private
-  public static class RMStateStoreOperationFailedEventDispatcher implements
-      EventHandler<RMStateStoreOperationFailedEvent> {
-
+  public static class RMFatalEventDispatcher
+      implements EventHandler<RMFatalEvent> {
     private final RMContext rmContext;
     private final ResourceManager rm;
 
-    public RMStateStoreOperationFailedEventDispatcher(RMContext rmContext,
-        ResourceManager resourceManager) {
+    public RMFatalEventDispatcher(
+        RMContext rmContext, ResourceManager resourceManager) {
       this.rmContext = rmContext;
       this.rm = resourceManager;
     }
 
     @Override
-    public void handle(RMStateStoreOperationFailedEvent event) {
-      if (LOG.isDebugEnabled()) {
-        LOG.debug("Received a " +
-            RMStateStoreOperationFailedEvent.class.getName() + " of type " +
-            event.getType().name());
-      }
-      if (event.getType() == RMStateStoreOperationFailedEventType.FENCED) {
+    public void handle(RMFatalEvent event) {
+      LOG.fatal("Received a " + RMFatalEvent.class.getName() + " of type " +
+          event.getType().name());
+
+      if (event.getType() == RMFatalEventType.STATE_STORE_FENCED) {
         LOG.info("RMStateStore has been fenced");
         if (rmContext.isHAEnabled()) {
           try {
@@ -633,14 +624,11 @@ public class ResourceManager extends CompositeService implements Recoverable {
             rm.transitionToStandby(true);
             return;
           } catch (Exception e) {
-            LOG.error("Failed to transition RM to Standby mode.");
+            LOG.fatal("Failed to transition RM to Standby mode.");
           }
         }
       }
 
-      LOG.error("Shutting down RM on receiving a " +
-          RMStateStoreOperationFailedEvent.class.getName() + " of type " +
-          event.getType().name());
       ExitUtil.terminate(1, event.getCause());
     }
   }

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

@@ -48,6 +48,8 @@ import org.apache.hadoop.yarn.event.Dispatcher;
 import org.apache.hadoop.yarn.event.EventHandler;
 import org.apache.hadoop.yarn.security.AMRMTokenIdentifier;
 import org.apache.hadoop.yarn.security.client.RMDelegationTokenIdentifier;
+import org.apache.hadoop.yarn.server.resourcemanager.RMFatalEvent;
+import org.apache.hadoop.yarn.server.resourcemanager.RMFatalEventType;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.RMStateVersion;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.impl.pb.ApplicationAttemptStateDataPBImpl;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.impl.pb.ApplicationStateDataPBImpl;
@@ -679,15 +681,13 @@ public abstract class RMStateStore extends AbstractService {
    * @param failureCause the exception due to which the operation failed
    */
   private void notifyStoreOperationFailed(Exception failureCause) {
-    RMStateStoreOperationFailedEventType type;
+    RMFatalEventType type;
     if (failureCause instanceof StoreFencedException) {
-      type = RMStateStoreOperationFailedEventType.FENCED;
+      type = RMFatalEventType.STATE_STORE_FENCED;
     } else {
-      type = RMStateStoreOperationFailedEventType.FAILED;
+      type = RMFatalEventType.STATE_STORE_OP_FAILED;
     }
-
-    rmDispatcher.getEventHandler().handle(
-        new RMStateStoreOperationFailedEvent(type, failureCause));
+    rmDispatcher.getEventHandler().handle(new RMFatalEvent(type, failureCause));
   }
 
   @SuppressWarnings("unchecked")

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

@@ -1,36 +0,0 @@
-/**
-* Licensed to the Apache Software Foundation (ASF) under one
-* or more contributor license agreements.  See the NOTICE file
-* distributed with this work for additional information
-* regarding copyright ownership.  The ASF licenses this file
-* to you under the Apache License, Version 2.0 (the
-* "License"); you may not use this file except in compliance
-* with the License.  You may obtain a copy of the License at
-*
-*     http://www.apache.org/licenses/LICENSE-2.0
-*
-* Unless required by applicable law or agreed to in writing, software
-* distributed under the License is distributed on an "AS IS" BASIS,
-* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-* See the License for the specific language governing permissions and
-* limitations under the License.
-*/
-
-package org.apache.hadoop.yarn.server.resourcemanager.recovery;
-
-import org.apache.hadoop.yarn.event.AbstractEvent;
-
-public class RMStateStoreOperationFailedEvent
-    extends AbstractEvent<RMStateStoreOperationFailedEventType> {
-  private Exception cause;
-
-  RMStateStoreOperationFailedEvent(
-      RMStateStoreOperationFailedEventType type, Exception cause) {
-    super(type);
-    this.cause = cause;
-  }
-
-  public Exception getCause() {
-    return this.cause;
-  }
-}

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

@@ -182,34 +182,34 @@ public class ZKRMStateStore extends RMStateStore {
 
   @Override
   public synchronized void initInternal(Configuration conf) throws Exception {
-    zkHostPort = conf.get(YarnConfiguration.ZK_RM_STATE_STORE_ADDRESS);
+    zkHostPort = conf.get(YarnConfiguration.RM_ZK_ADDRESS);
     if (zkHostPort == null) {
       throw new YarnRuntimeException("No server address specified for " +
           "zookeeper state store for Resource Manager recovery. " +
-          YarnConfiguration.ZK_RM_STATE_STORE_ADDRESS + " is not configured.");
+          YarnConfiguration.RM_ZK_ADDRESS + " is not configured.");
     }
     numRetries =
-        conf.getInt(YarnConfiguration.ZK_RM_STATE_STORE_NUM_RETRIES,
-            YarnConfiguration.DEFAULT_ZK_RM_STATE_STORE_NUM_RETRIES);
+        conf.getInt(YarnConfiguration.RM_ZK_NUM_RETRIES,
+            YarnConfiguration.DEFAULT_ZK_RM_NUM_RETRIES);
     znodeWorkingPath =
         conf.get(YarnConfiguration.ZK_RM_STATE_STORE_PARENT_PATH,
             YarnConfiguration.DEFAULT_ZK_RM_STATE_STORE_PARENT_PATH);
     zkSessionTimeout =
-        conf.getInt(YarnConfiguration.ZK_RM_STATE_STORE_TIMEOUT_MS,
-            YarnConfiguration.DEFAULT_ZK_RM_STATE_STORE_TIMEOUT_MS);
+        conf.getInt(YarnConfiguration.RM_ZK_TIMEOUT_MS,
+            YarnConfiguration.DEFAULT_RM_ZK_TIMEOUT_MS);
     zkRetryInterval =
-        conf.getLong(YarnConfiguration.ZK_RM_STATE_STORE_RETRY_INTERVAL_MS,
-          YarnConfiguration.DEFAULT_ZK_RM_STATE_STORE_RETRY_INTERVAL_MS);
+        conf.getLong(YarnConfiguration.RM_ZK_RETRY_INTERVAL_MS,
+          YarnConfiguration.DEFAULT_RM_ZK_RETRY_INTERVAL_MS);
     // Parse authentication from configuration.
     String zkAclConf =
-        conf.get(YarnConfiguration.ZK_RM_STATE_STORE_ACL,
-            YarnConfiguration.DEFAULT_ZK_RM_STATE_STORE_ACL);
+        conf.get(YarnConfiguration.RM_ZK_ACL,
+            YarnConfiguration.DEFAULT_RM_ZK_ACL);
     zkAclConf = ZKUtil.resolveConfIndirection(zkAclConf);
 
     try {
       zkAcl = ZKUtil.parseACLs(zkAclConf);
     } catch (ZKUtil.BadAclFormatException bafe) {
-      LOG.error("Invalid format for " + YarnConfiguration.ZK_RM_STATE_STORE_ACL);
+      LOG.error("Invalid format for " + YarnConfiguration.RM_ZK_ACL);
       throw bafe;
     }
 

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

@@ -396,6 +396,11 @@ public class MockRM extends ResourceManager {
       protected void stopServer() {
         // don't do anything
       }
+
+      @Override
+      protected EmbeddedElectorService createEmbeddedElectorService() {
+        return null;
+      }
     };
   }
 

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

@@ -25,6 +25,7 @@ import org.apache.hadoop.ha.HAServiceProtocol;
 import org.apache.hadoop.ha.HAServiceProtocol.HAServiceState;
 import org.apache.hadoop.ha.HAServiceProtocol.StateChangeRequestInfo;
 import org.apache.hadoop.ha.HealthCheckFailedException;
+import org.apache.hadoop.security.AccessControlException;
 import org.apache.hadoop.yarn.conf.YarnConfiguration;
 import org.apache.hadoop.yarn.conf.HAUtil;
 import org.junit.Before;
@@ -39,6 +40,7 @@ import static org.junit.Assert.fail;
 
 public class TestRMHA {
   private Log LOG = LogFactory.getLog(TestRMHA.class);
+  private final Configuration configuration = new YarnConfiguration();
   private MockRM rm = null;
   private static final String STATE_ERR =
       "ResourceManager is in wrong HA state";
@@ -51,17 +53,13 @@ public class TestRMHA {
 
   @Before
   public void setUp() throws Exception {
-    Configuration conf = new YarnConfiguration();
-    conf.setBoolean(YarnConfiguration.RM_HA_ENABLED, true);
-    conf.set(YarnConfiguration.RM_HA_IDS, RM1_NODE_ID + "," + RM2_NODE_ID);
+    configuration.setBoolean(YarnConfiguration.RM_HA_ENABLED, true);
+    configuration.set(YarnConfiguration.RM_HA_IDS, RM1_NODE_ID + "," + RM2_NODE_ID);
     for (String confKey : YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS) {
-      conf.set(HAUtil.addSuffix(confKey, RM1_NODE_ID), RM1_ADDRESS);
-      conf.set(HAUtil.addSuffix(confKey, RM2_NODE_ID), RM2_ADDRESS);
+      configuration.set(HAUtil.addSuffix(confKey, RM1_NODE_ID), RM1_ADDRESS);
+      configuration.set(HAUtil.addSuffix(confKey, RM2_NODE_ID), RM2_ADDRESS);
     }
-    conf.set(YarnConfiguration.RM_HA_ID, RM1_NODE_ID);
-
-    rm = new MockRM(conf);
-    rm.init(conf);
+    configuration.set(YarnConfiguration.RM_HA_ID, RM1_NODE_ID);
   }
 
   private void checkMonitorHealth() throws IOException {
@@ -113,6 +111,9 @@ public class TestRMHA {
    */
   @Test (timeout = 30000)
   public void testStartAndTransitions() throws IOException {
+    Configuration conf = new YarnConfiguration(configuration);
+    rm = new MockRM(conf);
+    rm.init(conf);
     StateChangeRequestInfo requestInfo = new StateChangeRequestInfo(
         HAServiceProtocol.RequestSource.REQUEST_BY_USER);
 
@@ -162,4 +163,63 @@ public class TestRMHA {
         rm.areActiveServicesRunning());
     checkMonitorHealth();
   }
+
+  @Test
+  public void testTransitionsWhenAutomaticFailoverEnabled() throws IOException {
+    final String ERR_UNFORCED_REQUEST = "User request succeeded even when " +
+        "automatic failover is enabled";
+
+    Configuration conf = new YarnConfiguration(configuration);
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, true);
+
+    rm = new MockRM(conf);
+    rm.init(conf);
+    rm.start();
+    StateChangeRequestInfo requestInfo = new StateChangeRequestInfo(
+        HAServiceProtocol.RequestSource.REQUEST_BY_USER);
+
+    // Transition to standby
+    try {
+      rm.adminService.transitionToStandby(requestInfo);
+      fail(ERR_UNFORCED_REQUEST);
+    } catch (AccessControlException e) {
+      // expected
+    }
+    checkMonitorHealth();
+    checkStandbyRMFunctionality();
+
+    // Transition to active
+    try {
+      rm.adminService.transitionToActive(requestInfo);
+      fail(ERR_UNFORCED_REQUEST);
+    } catch (AccessControlException e) {
+      // expected
+    }
+    checkMonitorHealth();
+    checkStandbyRMFunctionality();
+
+
+    final String ERR_FORCED_REQUEST = "Forced request by user should work " +
+        "even if automatic failover is enabled";
+    requestInfo = new StateChangeRequestInfo(
+        HAServiceProtocol.RequestSource.REQUEST_BY_USER_FORCED);
+
+    // Transition to standby
+    try {
+      rm.adminService.transitionToStandby(requestInfo);
+    } catch (AccessControlException e) {
+      fail(ERR_FORCED_REQUEST);
+    }
+    checkMonitorHealth();
+    checkStandbyRMFunctionality();
+
+    // Transition to active
+    try {
+      rm.adminService.transitionToActive(requestInfo);
+    } catch (AccessControlException e) {
+      fail(ERR_FORCED_REQUEST);
+    }
+    checkMonitorHealth();
+    checkActiveRMFunctionality();
+  }
 }

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

@@ -92,7 +92,7 @@ public class TestZKRMStateStore extends RMStateStoreTestBase {
     public RMStateStore getRMStateStore() throws Exception {
       YarnConfiguration conf = new YarnConfiguration();
       workingZnode = "/Test";
-      conf.set(YarnConfiguration.ZK_RM_STATE_STORE_ADDRESS, hostPort);
+      conf.set(YarnConfiguration.RM_ZK_ADDRESS, hostPort);
       conf.set(YarnConfiguration.ZK_RM_STATE_STORE_PARENT_PATH, workingZnode);
       this.client = createClient();
       this.store = new TestZKRMStateStoreInternal(conf, workingZnode);
@@ -140,7 +140,7 @@ public class TestZKRMStateStore extends RMStateStoreTestBase {
     conf.set(YarnConfiguration.RM_HA_IDS, rmIds);
     conf.setBoolean(YarnConfiguration.RECOVERY_ENABLED, true);
     conf.set(YarnConfiguration.RM_STORE, ZKRMStateStore.class.getName());
-    conf.set(YarnConfiguration.ZK_RM_STATE_STORE_ADDRESS, hostPort);
+    conf.set(YarnConfiguration.RM_ZK_ADDRESS, hostPort);
     conf.set(YarnConfiguration.RM_HA_ID, rmId);
     for (String rpcAddress : YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS) {
       for (String id : HAUtil.getRMHAIds(conf)) {

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

@@ -107,7 +107,7 @@ public class TestZKRMStateStoreZKClientConnections extends
 
     public RMStateStore getRMStateStore(Configuration conf) throws Exception {
       String workingZnode = "/Test";
-      conf.set(YarnConfiguration.ZK_RM_STATE_STORE_ADDRESS, hostPort);
+      conf.set(YarnConfiguration.RM_ZK_ADDRESS, hostPort);
       conf.set(YarnConfiguration.ZK_RM_STATE_STORE_PARENT_PATH, workingZnode);
       watcher = new TestForwardingWatcher();
       this.store = new TestZKRMStateStore(conf, workingZnode);
@@ -120,8 +120,8 @@ public class TestZKRMStateStoreZKClientConnections extends
     TestZKClient zkClientTester = new TestZKClient();
     final String path = "/test";
     YarnConfiguration conf = new YarnConfiguration();
-    conf.setInt(YarnConfiguration.ZK_RM_STATE_STORE_TIMEOUT_MS, 1000);
-    conf.setLong(YarnConfiguration.ZK_RM_STATE_STORE_RETRY_INTERVAL_MS, 100);
+    conf.setInt(YarnConfiguration.RM_ZK_TIMEOUT_MS, 1000);
+    conf.setLong(YarnConfiguration.RM_ZK_RETRY_INTERVAL_MS, 100);
     final ZKRMStateStore store =
         (ZKRMStateStore) zkClientTester.getRMStateStore(conf);
     TestDispatcher dispatcher = new TestDispatcher();
@@ -153,7 +153,7 @@ public class TestZKRMStateStoreZKClientConnections extends
     TestZKClient zkClientTester = new TestZKClient();
     String path = "/test";
     YarnConfiguration conf = new YarnConfiguration();
-    conf.setInt(YarnConfiguration.ZK_RM_STATE_STORE_TIMEOUT_MS, 100);
+    conf.setInt(YarnConfiguration.RM_ZK_TIMEOUT_MS, 100);
     ZKRMStateStore store =
         (ZKRMStateStore) zkClientTester.getRMStateStore(conf);
     TestDispatcher dispatcher = new TestDispatcher();
@@ -195,7 +195,7 @@ public class TestZKRMStateStoreZKClientConnections extends
     TestZKClient zkClientTester = new TestZKClient();
     String path = "/test";
     YarnConfiguration conf = new YarnConfiguration();
-    conf.setInt(YarnConfiguration.ZK_RM_STATE_STORE_TIMEOUT_MS, 100);
+    conf.setInt(YarnConfiguration.RM_ZK_TIMEOUT_MS, 100);
     ZKRMStateStore store =
         (ZKRMStateStore) zkClientTester.getRMStateStore(conf);
     TestDispatcher dispatcher = new TestDispatcher();
@@ -227,7 +227,7 @@ public class TestZKRMStateStoreZKClientConnections extends
   public void testSetZKAcl() {
     TestZKClient zkClientTester = new TestZKClient();
     YarnConfiguration conf = new YarnConfiguration();
-    conf.set(YarnConfiguration.ZK_RM_STATE_STORE_ACL, "world:anyone:rwca");
+    conf.set(YarnConfiguration.RM_ZK_ACL, "world:anyone:rwca");
     try {
       zkClientTester.store.zkClient.delete(zkClientTester.store
           .znodeWorkingPath, -1);
@@ -240,7 +240,7 @@ public class TestZKRMStateStoreZKClientConnections extends
   public void testInvalidZKAclConfiguration() {
     TestZKClient zkClientTester = new TestZKClient();
     YarnConfiguration conf = new YarnConfiguration();
-    conf.set(YarnConfiguration.ZK_RM_STATE_STORE_ACL, "randomstring&*");
+    conf.set(YarnConfiguration.RM_ZK_ACL, "randomstring&*");
     try {
       zkClientTester.getRMStateStore(conf);
       fail("ZKRMStateStore created with bad ACL");

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-tests/src/test/java/org/apache/hadoop/yarn/server/MiniYARNCluster.java

@@ -58,6 +58,7 @@ import org.apache.hadoop.yarn.server.nodemanager.NodeHealthCheckerService;
 import org.apache.hadoop.yarn.server.nodemanager.NodeManager;
 import org.apache.hadoop.yarn.server.nodemanager.NodeStatusUpdater;
 import org.apache.hadoop.yarn.server.nodemanager.NodeStatusUpdaterImpl;
+import org.apache.hadoop.yarn.server.resourcemanager.RMFatalEvent;
 import org.apache.hadoop.yarn.server.resourcemanager.ResourceManager;
 import org.apache.hadoop.yarn.server.resourcemanager.ResourceTrackerService;
 import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.RMAppAttemptEvent;
@@ -98,6 +99,7 @@ public class MiniYARNCluster extends CompositeService {
 
   private boolean useFixedPorts;
   private boolean useRpc = false;
+  private int failoverTimeout;
 
   private ConcurrentMap<ApplicationAttemptId, Long> appMasters =
       new ConcurrentHashMap<ApplicationAttemptId, Long>(16, 0.75f, 2);
@@ -189,12 +191,15 @@ public class MiniYARNCluster extends CompositeService {
         YarnConfiguration.DEFAULT_YARN_MINICLUSTER_FIXED_PORTS);
     useRpc = conf.getBoolean(YarnConfiguration.YARN_MINICLUSTER_USE_RPC,
         YarnConfiguration.DEFAULT_YARN_MINICLUSTER_USE_RPC);
+    failoverTimeout = conf.getInt(YarnConfiguration.RM_ZK_TIMEOUT_MS,
+        YarnConfiguration.DEFAULT_RM_ZK_TIMEOUT_MS);
 
     if (useRpc && !useFixedPorts) {
       throw new YarnRuntimeException("Invalid configuration!" +
           " Minicluster can use rpc only when configured to use fixed ports");
     }
 
+    conf.setBoolean(YarnConfiguration.IS_MINI_YARN_CLUSTER, true);
     if (resourceManagers.length > 1) {
       conf.setBoolean(YarnConfiguration.RM_HA_ENABLED, true);
       if (conf.get(YarnConfiguration.RM_HA_IDS) == null) {
@@ -218,6 +223,13 @@ public class MiniYARNCluster extends CompositeService {
           // Don't try to login using keytab in the testcases.
         }
       };
+      if (!useFixedPorts) {
+        if (HAUtil.isHAEnabled(conf)) {
+          setHARMConfiguration(i, conf);
+        } else {
+          setNonHARMConfiguration(conf);
+        }
+      }
       addService(new ResourceManagerWrapper(i));
     }
     for(int index = 0; index < nodeManagers.length; index++) {
@@ -230,18 +242,103 @@ public class MiniYARNCluster extends CompositeService {
         conf instanceof YarnConfiguration ? conf : new YarnConfiguration(conf));
   }
 
+  private void setNonHARMConfiguration(Configuration conf) {
+    String hostname = MiniYARNCluster.getHostname();
+    conf.set(YarnConfiguration.RM_ADDRESS, hostname + ":0");
+    conf.set(YarnConfiguration.RM_ADMIN_ADDRESS, hostname + ":0");
+    conf.set(YarnConfiguration.RM_SCHEDULER_ADDRESS, hostname + ":0");
+    conf.set(YarnConfiguration.RM_RESOURCE_TRACKER_ADDRESS, hostname + ":0");
+    WebAppUtils.setRMWebAppHostnameAndPort(conf, hostname, 0);
+  }
+
+  private void setHARMConfiguration(final int index, Configuration conf) {
+    String hostname = MiniYARNCluster.getHostname();
+    for (String confKey : YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS) {
+      conf.set(HAUtil.addSuffix(confKey, rmIds[index]), hostname + ":0");
+    }
+  }
+
+  private synchronized void initResourceManager(int index, Configuration conf) {
+    if (HAUtil.isHAEnabled(conf)) {
+      conf.set(YarnConfiguration.RM_HA_ID, rmIds[index]);
+    }
+    resourceManagers[index].init(conf);
+    resourceManagers[index].getRMContext().getDispatcher().register(
+        RMAppAttemptEventType.class,
+        new EventHandler<RMAppAttemptEvent>() {
+          public void handle(RMAppAttemptEvent event) {
+            if (event instanceof RMAppAttemptRegistrationEvent) {
+              appMasters.put(event.getApplicationAttemptId(),
+                  event.getTimestamp());
+            } else if (event instanceof RMAppAttemptUnregistrationEvent) {
+              appMasters.remove(event.getApplicationAttemptId());
+            }
+          }
+        });
+  }
+
+  private synchronized void startResourceManager(final int index) {
+    try {
+      Thread rmThread = new Thread() {
+        public void run() {
+          resourceManagers[index].start();
+        }
+      };
+      rmThread.setName("RM-" + index);
+      rmThread.start();
+      int waitCount = 0;
+      while (resourceManagers[index].getServiceState() == STATE.INITED
+          && waitCount++ < 60) {
+        LOG.info("Waiting for RM to start...");
+        Thread.sleep(1500);
+      }
+      if (resourceManagers[index].getServiceState() != STATE.STARTED) {
+        // RM could have failed.
+        throw new IOException(
+            "ResourceManager failed to start. Final state is "
+                + resourceManagers[index].getServiceState());
+      }
+    } catch (Throwable t) {
+      throw new YarnRuntimeException(t);
+    }
+    LOG.info("MiniYARN ResourceManager address: " +
+        getConfig().get(YarnConfiguration.RM_ADDRESS));
+    LOG.info("MiniYARN ResourceManager web address: " +
+        WebAppUtils.getRMWebAppURLWithoutScheme(getConfig()));
+  }
+
+  @InterfaceAudience.Private
+  @VisibleForTesting
+  public synchronized void stopResourceManager(int index) {
+    if (resourceManagers[index] != null) {
+      resourceManagers[index].stop();
+      resourceManagers[index] = null;
+    }
+  }
+
+  @InterfaceAudience.Private
+  @VisibleForTesting
+  public synchronized void restartResourceManager(int index)
+      throws InterruptedException {
+    if (resourceManagers[index] != null) {
+      resourceManagers[index].stop();
+      resourceManagers[index] = null;
+    }
+    Configuration conf = getConfig();
+    resourceManagers[index] = new ResourceManager();
+    initResourceManager(index, getConfig());
+    startResourceManager(index);
+  }
+
   public File getTestWorkDir() {
     return testWorkDir;
   }
 
   /**
-   * In a HA cluster, go through all the RMs and find the Active RM. If none
-   * of them are active, wait upto 5 seconds for them to transition to Active.
+   * In a HA cluster, go through all the RMs and find the Active RM. In a
+   * non-HA cluster, return the index of the only RM.
    *
-   * In an non-HA cluster, return the index of the only RM.
-   *
-   * @return index of the active RM or -1 if none of them transition to
-   * active even after 5 seconds of waiting
+   * @return index of the active RM or -1 if none of them turn active
    */
   @InterfaceAudience.Private
   @VisibleForTesting
@@ -250,9 +347,12 @@ public class MiniYARNCluster extends CompositeService {
       return 0;
     }
 
-    int numRetriesForRMBecomingActive = 5;
+    int numRetriesForRMBecomingActive = failoverTimeout / 100;
     while (numRetriesForRMBecomingActive-- > 0) {
       for (int i = 0; i < resourceManagers.length; i++) {
+        if (resourceManagers[i] == null) {
+          continue;
+        }
         try {
           if (HAServiceProtocol.HAServiceState.ACTIVE ==
               resourceManagers[i].getRMContext().getRMAdminService()
@@ -265,7 +365,7 @@ public class MiniYARNCluster extends CompositeService {
         }
       }
       try {
-        Thread.sleep(1000);
+        Thread.sleep(100);
       } catch (InterruptedException e) {
         throw new YarnRuntimeException("Interrupted while waiting for one " +
             "of the ResourceManagers to become active");
@@ -282,7 +382,7 @@ public class MiniYARNCluster extends CompositeService {
     int activeRMIndex = getActiveRMIndex();
     return activeRMIndex == -1
         ? null
-        : this.resourceManagers[getActiveRMIndex()];
+        : this.resourceManagers[activeRMIndex];
   }
 
   public ResourceManager getResourceManager(int i) {
@@ -310,82 +410,21 @@ public class MiniYARNCluster extends CompositeService {
       index = i;
     }
 
-    private void setNonHARMConfiguration(Configuration conf) {
-      String hostname = MiniYARNCluster.getHostname();
-      conf.set(YarnConfiguration.RM_ADDRESS, hostname + ":0");
-      conf.set(YarnConfiguration.RM_ADMIN_ADDRESS, hostname + ":0");
-      conf.set(YarnConfiguration.RM_SCHEDULER_ADDRESS, hostname + ":0");
-      conf.set(YarnConfiguration.RM_RESOURCE_TRACKER_ADDRESS, hostname + ":0");
-      WebAppUtils.setRMWebAppHostnameAndPort(conf, hostname, 0);
-    }
-
-    private void setHARMConfiguration(Configuration conf) {
-      String hostname = MiniYARNCluster.getHostname();
-      for (String confKey : YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS) {
-        for (String id : HAUtil.getRMHAIds(conf)) {
-          conf.set(HAUtil.addSuffix(confKey, id), hostname + ":0");
-        }
-      }
-    }
-
     @Override
     protected synchronized void serviceInit(Configuration conf)
         throws Exception {
-      conf.setBoolean(YarnConfiguration.IS_MINI_YARN_CLUSTER, true);
-
-      if (!useFixedPorts) {
-        if (HAUtil.isHAEnabled(conf)) {
-          setHARMConfiguration(conf);
-        } else {
-          setNonHARMConfiguration(conf);
-        }
-      }
-      if (HAUtil.isHAEnabled(conf)) {
-        conf.set(YarnConfiguration.RM_HA_ID, rmIds[index]);
-      }
-      resourceManagers[index].init(conf);
-      resourceManagers[index].getRMContext().getDispatcher().register
-          (RMAppAttemptEventType.class,
-          new EventHandler<RMAppAttemptEvent>() {
-            public void handle(RMAppAttemptEvent event) {
-              if (event instanceof RMAppAttemptRegistrationEvent) {
-                appMasters.put(event.getApplicationAttemptId(), event.getTimestamp());
-              } else if (event instanceof RMAppAttemptUnregistrationEvent) {
-                appMasters.remove(event.getApplicationAttemptId());
-              }
-            }
-          });
+      initResourceManager(index, conf);
       super.serviceInit(conf);
     }
 
     @Override
     protected synchronized void serviceStart() throws Exception {
-      try {
-        new Thread() {
-          public void run() {
-            resourceManagers[index].start();
-          }
-        }.start();
-        int waitCount = 0;
-        while (resourceManagers[index].getServiceState() == STATE.INITED
-            && waitCount++ < 60) {
-          LOG.info("Waiting for RM to start...");
-          Thread.sleep(1500);
-        }
-        if (resourceManagers[index].getServiceState() != STATE.STARTED) {
-          // RM could have failed.
-          throw new IOException(
-              "ResourceManager failed to start. Final state is "
-                  + resourceManagers[index].getServiceState());
-        }
-        super.serviceStart();
-      } catch (Throwable t) {
-        throw new YarnRuntimeException(t);
-      }
+      startResourceManager(index);
       LOG.info("MiniYARN ResourceManager address: " +
                getConfig().get(YarnConfiguration.RM_ADDRESS));
       LOG.info("MiniYARN ResourceManager web address: " +
                WebAppUtils.getRMWebAppURLWithoutScheme(getConfig()));
+      super.serviceStart();
     }
 
     private void waitForAppMastersToFinish(long timeoutMillis) throws InterruptedException {
@@ -406,7 +445,6 @@ public class MiniYARNCluster extends CompositeService {
         waitForAppMastersToFinish(5000);
         resourceManagers[index].stop();
       }
-      super.serviceStop();
 
       if (Shell.WINDOWS) {
         // On Windows, clean up the short temporary symlink that was created to
@@ -420,6 +458,7 @@ public class MiniYARNCluster extends CompositeService {
             testWorkDir.getAbsolutePath());
         }
       }
+      super.serviceStop();
     }
   }