HBASE-21521 Expose master startup status via web UI

Author: sunhelly

hbase-server/src/main/jamon/org/apache/hadoop/hbase/tmpl/master/MasterStatusTmpl.jamon

@@ -177,6 +177,7 @@ AssignmentManager assignmentManager = master.getAssignmentManager();
                 <%if HBaseConfiguration.isShowConfInServlet()%>
                 <li><a href="/conf">HBase Configuration</a></li>
                 </%if>
+                <li><a href="/startupProgress.jsp">Startup Progress</a></li>
                </ul>
             </div><!--/.nav-collapse -->
         </div>

hbase-server/src/main/java/org/apache/hadoop/hbase/master/ActiveMasterManager.java

@@ -27,6 +27,7 @@
 import org.apache.hadoop.hbase.ZNodeClearer;
 import org.apache.hadoop.hbase.exceptions.DeserializationException;
 import org.apache.hadoop.hbase.monitoring.MonitoredTask;
+import org.apache.hadoop.hbase.monitoring.TaskGroup;
 import org.apache.hadoop.hbase.zookeeper.MasterAddressTracker;
 import org.apache.hadoop.hbase.zookeeper.ZKListener;
 import org.apache.hadoop.hbase.zookeeper.ZKUtil;
@@ -218,16 +219,18 @@ private void handleMasterNodeChange() {
    * Block until becoming the active master. Method blocks until there is not another active master
    * and our attempt to become the new active master is successful. This also makes sure that we are
    * watching the master znode so will be notified if another master dies.
-   * @param checkInterval the interval to check if the master is stopped
-   * @param startupStatus the monitor status to track the progress
+   * @param checkInterval    the interval to check if the master is stopped
+   * @param startupTaskGroup the task group for master startup to track the progress
    * @return True if no issue becoming active master else false if another master was running or if
    *         some other problem (zookeeper, stop flag has been set on this Master)
    */
-  boolean blockUntilBecomingActiveMaster(int checkInterval, MonitoredTask startupStatus) {
+  boolean blockUntilBecomingActiveMaster(int checkInterval, TaskGroup startupTaskGroup) {
+    MonitoredTask blockUntilActive =
+      startupTaskGroup.addTask("Blocking until becoming active master");
     String backupZNode = ZNodePaths
       .joinZNode(this.watcher.getZNodePaths().backupMasterAddressesZNode, this.sn.toString());
     while (!(master.isAborted() || master.isStopped())) {
-      startupStatus.setStatus("Trying to register in ZK as active master");
+      blockUntilActive.setStatus("Trying to register in ZK as active master");
       // Try to become the active master, watch if there is another master.
       // Write out our ServerName as versioned bytes.
       try {
@@ -246,7 +249,7 @@ boolean blockUntilBecomingActiveMaster(int checkInterval, MonitoredTask startupS
           ZNodeClearer.writeMyEphemeralNodeOnDisk(this.sn.toString());
 
           // We are the master, return
-          startupStatus.setStatus("Successfully registered as active master.");
+          blockUntilActive.setStatus("Successfully registered as active master.");
           this.clusterHasActiveMaster.set(true);
           activeMasterServerName = sn;
           LOG.info("Registered as active master=" + this.sn);
@@ -291,7 +294,7 @@ boolean blockUntilBecomingActiveMaster(int checkInterval, MonitoredTask startupS
           }
         }
         LOG.info(msg);
-        startupStatus.setStatus(msg);
+        blockUntilActive.setStatus(msg);
       } catch (KeeperException ke) {
         master.abort("Received an unexpected KeeperException, aborting", ke);
         return false;

hbase-server/src/main/java/org/apache/hadoop/hbase/master/HMaster.java

@@ -186,6 +186,7 @@
 import org.apache.hadoop.hbase.mob.MobFileCompactionChore;
 import org.apache.hadoop.hbase.monitoring.MemoryBoundedLogMessageBuffer;
 import org.apache.hadoop.hbase.monitoring.MonitoredTask;
+import org.apache.hadoop.hbase.monitoring.TaskGroup;
 import org.apache.hadoop.hbase.monitoring.TaskMonitor;
 import org.apache.hadoop.hbase.namequeues.NamedQueueRecorder;
 import org.apache.hadoop.hbase.procedure.MasterProcedureManagerHost;
@@ -465,6 +466,8 @@ public class HMaster extends HBaseServerBase<MasterRpcServices> implements Maste
   public static final String WARMUP_BEFORE_MOVE = "hbase.master.warmup.before.move";
   private static final boolean DEFAULT_WARMUP_BEFORE_MOVE = true;
 
+  private TaskGroup startupTaskGroup;
+
   /**
    * Initializes the HMaster. The steps are as follows:
    * <p>
@@ -473,9 +476,8 @@ public class HMaster extends HBaseServerBase<MasterRpcServices> implements Maste
    * <li>Start the ActiveMasterManager.
    * </ol>
    * <p>
-   * Remaining steps of initialization occur in
-   * {@link #finishActiveMasterInitialization(MonitoredTask)} after the master becomes the active
-   * one.
+   * Remaining steps of initialization occur in {@link #finishActiveMasterInitialization(TaskGroup)}
+   * after the master becomes the active one.
    */
   public HMaster(final Configuration conf) throws IOException {
     super(conf, "Master");
@@ -908,12 +910,12 @@ private void tryMigrateMetaLocationsFromZooKeeper() throws IOException, KeeperEx
    * Notice that now we will not schedule a special procedure to make meta online(unless the first
    * time where meta has not been created yet), we will rely on SCP to bring meta online.
    */
-  private void finishActiveMasterInitialization(MonitoredTask status) throws IOException,
+  private void finishActiveMasterInitialization(TaskGroup startupTaskGro) throws IOException,
     InterruptedException, KeeperException, ReplicationException, DeserializationException {
     /*
      * We are active master now... go initialize components we need to run.
      */
-    status.setStatus("Initializing Master file system");
+    startupTaskGroup.addTask("Initializing Master file system");
 
     this.masterActiveTime = EnvironmentEdgeManager.currentTime();
     // TODO: Do this using Dependency Injection, using PicoContainer, Guice or Spring.
@@ -926,15 +928,15 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
 
     // warm-up HTDs cache on master initialization
     if (preLoadTableDescriptors) {
-      status.setStatus("Pre-loading table descriptors");
+      startupTaskGroup.addTask("Pre-loading table descriptors");
       this.tableDescriptors.getAll();
     }
 
     // Publish cluster ID; set it in Master too. The superclass RegionServer does this later but
     // only after it has checked in with the Master. At least a few tests ask Master for clusterId
     // before it has called its run method and before RegionServer has done the reportForDuty.
     ClusterId clusterId = fileSystemManager.getClusterId();
-    status.setStatus("Publishing Cluster ID " + clusterId + " in ZooKeeper");
+    startupTaskGroup.addTask("Publishing Cluster ID " + clusterId + " in ZooKeeper");
     ZKClusterId.setClusterId(this.zooKeeper, fileSystemManager.getClusterId());
     this.clusterId = clusterId.toString();
 
@@ -953,7 +955,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
       }
     }
 
-    status.setStatus("Initialize ServerManager and schedule SCP for crash servers");
+    startupTaskGroup.addTask("Initialize ServerManager and schedule SCP for crash servers");
     // The below two managers must be created before loading procedures, as they will be used during
     // loading.
     // initialize master local region
@@ -1000,9 +1002,9 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     // This manager must be accessed AFTER hbase:meta is confirmed on line..
     this.tableStateManager = new TableStateManager(this);
 
-    status.setStatus("Initializing ZK system trackers");
+    startupTaskGroup.addTask("Initializing ZK system trackers");
     initializeZKBasedSystemTrackers();
-    status.setStatus("Loading last flushed sequence id of regions");
+    startupTaskGroup.addTask("Loading last flushed sequence id of regions");
     try {
       this.serverManager.loadLastFlushedSequenceIds();
     } catch (IOException e) {
@@ -1018,7 +1020,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     zombieDetector.start();
 
     if (!maintenanceMode) {
-      status.setStatus("Initializing master coprocessors");
+      startupTaskGroup.addTask("Initializing master coprocessors");
       setQuotasObserver(conf);
       initializeCoprocessorHost(conf);
     } else {
@@ -1029,7 +1031,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     }
 
     // Checking if meta needs initializing.
-    status.setStatus("Initializing meta table if this is a new deploy");
+    startupTaskGroup.addTask("Initializing meta table if this is a new deploy");
     InitMetaProcedure initMetaProc = null;
     // Print out state of hbase:meta on startup; helps debugging.
     if (!this.assignmentManager.getRegionStates().hasTableRegionStates(TableName.META_TABLE_NAME)) {
@@ -1049,7 +1051,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     this.balancer.updateClusterMetrics(getClusterMetricsWithoutCoprocessor());
 
     // start up all service threads.
-    status.setStatus("Initializing master service threads");
+    startupTaskGroup.addTask("Initializing master service threads");
     startServiceThreads();
     // wait meta to be initialized after we start procedure executor
     if (initMetaProc != null) {
@@ -1062,16 +1064,16 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     // With this as part of master initialization, it precludes our being able to start a single
     // server that is both Master and RegionServer. Needs more thought. TODO.
     String statusStr = "Wait for region servers to report in";
-    status.setStatus(statusStr);
-    LOG.info(Objects.toString(status));
-    waitForRegionServers(status);
+    MonitoredTask waitRegionServer = startupTaskGroup.addTask(statusStr);
+    LOG.info(Objects.toString(waitRegionServer));
+    waitForRegionServers(waitRegionServer);
 
     // Check if master is shutting down because issue initializing regionservers or balancer.
     if (isStopped()) {
       return;
     }
 
-    status.setStatus("Starting assignment manager");
+    startupTaskGroup.addTask("Starting assignment manager");
     // FIRST HBASE:META READ!!!!
     // The below cannot make progress w/o hbase:meta being online.
     // This is the FIRST attempt at going to hbase:meta. Meta on-lining is going on in background
@@ -1136,7 +1138,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     this.balancer.updateClusterMetrics(getClusterMetricsWithoutCoprocessor());
 
     // Start balancer and meta catalog janitor after meta and regions have been assigned.
-    status.setStatus("Starting balancer and catalog janitor");
+    startupTaskGroup.addTask("Starting balancer and catalog janitor");
     this.clusterStatusChore = new ClusterStatusChore(this, balancer);
     getChoreService().scheduleChore(clusterStatusChore);
     this.balancerChore = new BalancerChore(this);
@@ -1156,7 +1158,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     if (!waitForNamespaceOnline()) {
       return;
     }
-    status.setStatus("Starting cluster schema service");
+    startupTaskGroup.addTask("Starting cluster schema service");
     try {
       initClusterSchemaService();
     } catch (IllegalStateException e) {
@@ -1179,7 +1181,6 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
       }
     }
 
-    status.markComplete("Initialization successful");
     LOG.info(String.format("Master has completed initialization %.3fsec",
       (EnvironmentEdgeManager.currentTime() - masterActiveTime) / 1000.0f));
     this.masterFinishedInitializationTime = EnvironmentEdgeManager.currentTime();
@@ -1198,6 +1199,9 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
     }
     // Set master as 'initialized'.
     setInitialized(true);
+    startupTaskGroup.markComplete("Initialization successful");
+    MonitoredTask status =
+      TaskMonitor.get().createStatus("Progress after master initialized", false, true);
 
     if (tableFamilyDesc == null && replBarrierFamilyDesc == null) {
       // create missing CFs in meta table after master is set to 'initialized'.
@@ -1286,6 +1290,7 @@ private void finishActiveMasterInitialization(MonitoredTask status) throws IOExc
 
     this.rollingUpgradeChore = new RollingUpgradeChore(this);
     getChoreService().scheduleChore(rollingUpgradeChore);
+    status.markComplete("Progress after master initialized complete");
   }
 
   private void createMissingCFsInMetaDuringUpgrade(TableDescriptor metaDescriptor)
@@ -2399,14 +2404,19 @@ private void startActiveMasterManager(int infoPort) throws KeeperException {
         Threads.sleep(timeout);
       }
     }
-    MonitoredTask status = TaskMonitor.get().createStatus("Master startup");
-    status.setDescription("Master startup");
+
+    // Here for the master startup process, we use TaskGroup to monitor the whole progress.
+    // The UI is similar to how Hadoop designed the startup page for the NameNode.
+    // See HBASE-21521 for more details.
+    // We do not cleanup the startupTaskGroup, let the startup progress information
+    // be permanent in the MEM.
+    startupTaskGroup = TaskGroup.createTaskGroup(false, "Master startup");
     try {
-      if (activeMasterManager.blockUntilBecomingActiveMaster(timeout, status)) {
-        finishActiveMasterInitialization(status);
+      if (activeMasterManager.blockUntilBecomingActiveMaster(timeout, startupTaskGroup)) {
+        finishActiveMasterInitialization(startupTaskGroup);
       }
     } catch (Throwable t) {
-      status.setStatus("Failed to become active: " + t.getMessage());
+      startupTaskGroup.abort("Failed to become active master due to:" + t.getMessage());
       LOG.error(HBaseMarkers.FATAL, "Failed to become active master", t);
       // HBASE-5680: Likely hadoop23 vs hadoop 20.x/1.x incompatibility
       if (
@@ -2420,8 +2430,6 @@ private void startActiveMasterManager(int infoPort) throws KeeperException {
       } else {
         abort("Unhandled exception. Starting shutdown.", t);
       }
-    } finally {
-      status.cleanup();
     }
   }
 
@@ -3097,6 +3105,10 @@ public MemoryBoundedLogMessageBuffer getRegionServerFatalLogBuffer() {
     return rsFatals;
   }
 
+  public TaskGroup getStartupProgress() {
+    return startupTaskGroup;
+  }
+
   /**
    * Shutdown the cluster. Master runs a coordinated stop of all RegionServers and then itself.
    */

hbase-server/src/main/java/org/apache/hadoop/hbase/monitoring/TaskGroup.java

@@ -0,0 +1,116 @@
+/*
+ * 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.hbase.monitoring;
+
+import java.util.Collection;
+import java.util.Collections;
+import java.util.concurrent.ConcurrentLinkedDeque;
+import org.apache.yetus.audience.InterfaceAudience;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+/**
+ * The {@link TaskGroup} can be seen as a big {@link MonitoredTask}, which contains a list of sub
+ * monitored tasks. The monitored tasks in the group are still be managed by the
+ * {@link TaskMonitor}, but whether to clear/expire the monitored tasks in a task group is optional.
+ * Since the monitored task already has journals, which mark the phases in a task, we still also
+ * need a task group to monitor a big task/process because the journals in a task is serial but the
+ * tasks in the task group can be parallel, then we have more flexible ability to monitor the
+ * process. Grouping the tasks is not strictly necessary but it is cleaner for presentation to
+ * operators. We might want to display the tasks in a group in a list view where each task can be
+ * collapsed (probably by default) or expanded.
+ */
+@InterfaceAudience.Private
+public class TaskGroup extends MonitoredTaskImpl {
+  private static final Logger LOG = LoggerFactory.getLogger(TaskGroup.class);
+
+  private final ConcurrentLinkedDeque<MonitoredTask> tasks = new ConcurrentLinkedDeque<>();
+
+  /** Whether to be tracked(e.g. show/clear/expire) in the singleton {@link TaskMonitor} */
+  private final boolean trackSubTasksInTaskMonitor;
+  private final MonitoredTask delegate;
+
+  public TaskGroup(boolean trackSubTasksInTaskMonitor, String description) {
+    super(false, description);
+    this.trackSubTasksInTaskMonitor = trackSubTasksInTaskMonitor;
+    this.delegate = TaskMonitor.get().createStatus(description, true, true);
+  }
+
+  public static TaskGroup createTaskGroup(boolean trackSubTasksInTaskMonitor, String description) {
+    return new TaskGroup(trackSubTasksInTaskMonitor, description);
+  }
+
+  public synchronized MonitoredTask addTask(String description) {
+    return addTask(description, true);
+  }
+
+  /**
+   * add a new task to the group, and before that might complete the last task in the group
+   * @param description      the description of the new task
+   * @param withCompleteLast whether to complete the last task in the group
+   * @return the added new task
+   */
+  public synchronized MonitoredTask addTask(String description, boolean withCompleteLast) {
+    if (withCompleteLast) {
+      MonitoredTask previousTask = this.tasks.peekLast();
+      if (
+        previousTask != null && previousTask.getState() != State.COMPLETE
+          && previousTask.getState() != State.ABORTED
+      ) {
+        previousTask.markComplete("Completed");
+      }
+    }
+    MonitoredTask task =
+      TaskMonitor.get().createStatus(description, trackSubTasksInTaskMonitor, true);
+    this.setStatus(description);
+    this.tasks.addLast(task);
+    delegate.setStatus(description);
+    return task;
+  }
+
+  public synchronized Collection<MonitoredTask> getTasks() {
+    return Collections.unmodifiableCollection(this.tasks);
+  }
+
+  @Override
+  public synchronized void abort(String msg) {
+    setStatus(msg);
+    setState(State.ABORTED);
+    for (MonitoredTask task : tasks) {
+      if (task.getState() != State.COMPLETE && task.getState() != State.ABORTED) {
+        task.abort(msg);
+      }
+    }
+    delegate.abort(msg);
+  }
+
+  @Override
+  public synchronized void markComplete(String msg) {
+    setState(State.COMPLETE);
+    setStatus(msg);
+    if (tasks.getLast() != null) {
+      tasks.getLast().markComplete(msg);
+    }
+    delegate.markComplete(msg);
+  }
+
+  @Override
+  public synchronized void cleanup() {
+    this.tasks.clear();
+  }
+}