HDFS-6160. TestSafeMode occasionally fails. (Contributed by Arpit Agarwal)

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1586007 13f79535-47bb-0310-9956-ffa450edef68

Author: arp7

hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt

@@ -322,6 +322,8 @@ Release 2.5.0 - UNRELEASED
 
     HDFS-6169. Move the address in WebImageViewer. (Akira Ajisaka via wheat9)
 
+    HDFS-6160. TestSafeMode occasionally fails. (Arpit Agarwal)
+
 Release 2.4.1 - UNRELEASED
 
   INCOMPATIBLE CHANGES

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NameNodeRpcServer.java

@@ -1022,6 +1022,7 @@ public DatanodeCommand blockReport(DatanodeRegistration nodeReg,
     for(StorageBlockReport r : reports) {
       final BlockListAsLongs blocks = new BlockListAsLongs(r.getBlocks());
       hasStaleStorages = bm.processReport(nodeReg, r.getStorage(), poolId, blocks);
+      metrics.incrStorageBlockReportOps();
     }
 
     if (nn.getFSImage().isUpgradeFinalized() &&

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/metrics/NameNodeMetrics.java

@@ -73,6 +73,8 @@ public class NameNodeMetrics {
   MutableCounterLong snapshotDiffReportOps;
   @Metric("Number of blockReceivedAndDeleted calls")
   MutableCounterLong blockReceivedAndDeletedOps;
+  @Metric("Number of blockReports from individual storages")
+  MutableCounterLong storageBlockReportOps;
 
   @Metric("Journal transactions") MutableRate transactions;
   @Metric("Journal syncs") MutableRate syncs;
@@ -221,6 +223,10 @@ public void incrSnapshotDiffReportOps() {
   public void incrBlockReceivedAndDeletedOps() {
     blockReceivedAndDeletedOps.incr();
   }
+  
+  public void incrStorageBlockReportOps() {
+    storageBlockReportOps.incr();
+  }
 
   public void addTransaction(long latency) {
     transactions.add(latency);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestSafeMode.java

@@ -18,6 +18,8 @@
 
 package org.apache.hadoop.hdfs;
 
+import static org.apache.hadoop.test.MetricsAsserts.getLongCounter;
+import static org.apache.hadoop.test.MetricsAsserts.getMetrics;
 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.assertFalse;
 import static org.junit.Assert.assertTrue;
@@ -26,6 +28,8 @@
 import java.io.IOException;
 import java.util.List;
 
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.FSDataOutputStream;
 import org.apache.hadoop.fs.FileStatus;
@@ -55,12 +59,14 @@
  * Tests to verify safe mode correctness.
  */
 public class TestSafeMode {
+  public static final Log LOG = LogFactory.getLog(TestSafeMode.class);
   private static final Path TEST_PATH = new Path("/test");
   private static final int BLOCK_SIZE = 1024;
   Configuration conf; 
   MiniDFSCluster cluster;
   FileSystem fs;
   DistributedFileSystem dfs;
+  private static final String NN_METRICS = "NameNodeActivity";
 
   @Before
   public void startUp() throws IOException {
@@ -158,16 +164,19 @@ public void testNoExtensionIfNoBlocks() throws IOException {
    */
   @Test(timeout=45000)
   public void testInitializeReplQueuesEarly() throws Exception {
+    LOG.info("Starting testInitializeReplQueuesEarly");
     // Spray the blocks around the cluster when we add DNs instead of
     // concentrating all blocks on the first node.
     BlockManagerTestUtil.setWritingPrefersLocalNode(
         cluster.getNamesystem().getBlockManager(), false);
 
     cluster.startDataNodes(conf, 2, true, StartupOption.REGULAR, null);
     cluster.waitActive();
+
+    LOG.info("Creating files");
     DFSTestUtil.createFile(fs, TEST_PATH, 15*BLOCK_SIZE, (short)1, 1L);
 
-    
+    LOG.info("Stopping all DataNodes");
     List<DataNodeProperties> dnprops = Lists.newLinkedList();
     dnprops.add(cluster.stopDataNode(0));
     dnprops.add(cluster.stopDataNode(0));
@@ -176,6 +185,7 @@ public void testInitializeReplQueuesEarly() throws Exception {
     cluster.getConfiguration(0).setFloat(
         DFSConfigKeys.DFS_NAMENODE_REPL_QUEUE_THRESHOLD_PCT_KEY, 1f/15f);
 
+    LOG.info("Restarting NameNode");
     cluster.restartNameNode();
     final NameNode nn = cluster.getNameNode();
 
@@ -189,27 +199,37 @@ public void testInitializeReplQueuesEarly() throws Exception {
         "until threshold is crossed",
         NameNodeAdapter.safeModeInitializedReplQueues(nn));
 
+    LOG.info("Restarting one DataNode");
     cluster.restartDataNode(dnprops.remove(0));
 
-    // Wait for the block report from the restarted DN to come in.
+    // Wait for block reports from all attached storages of
+    // the restarted DN to come in.
     GenericTestUtils.waitFor(new Supplier<Boolean>() {
       @Override
       public Boolean get() {
-        return NameNodeAdapter.getSafeModeSafeBlocks(nn) > 0;
+        return getLongCounter("StorageBlockReportOps", getMetrics(NN_METRICS)) ==
+            MiniDFSCluster.DIRS_PER_DATANODE;
       }
     }, 10, 10000);
-    // SafeMode is fine-grain synchronized, so the processMisReplicatedBlocks
-    // call is still going on at this point - wait until it's done by grabbing
-    // the lock.
-    nn.getNamesystem().writeLock();
-    nn.getNamesystem().writeUnlock();
-    int safe = NameNodeAdapter.getSafeModeSafeBlocks(nn);
-    assertTrue("Expected first block report to make some but not all blocks " +
-        "safe. Got: " + safe, safe >= 1 && safe < 15);
-    BlockManagerTestUtil.updateState(nn.getNamesystem().getBlockManager());
-    
+
+    final int safe = NameNodeAdapter.getSafeModeSafeBlocks(nn);
+    assertTrue("Expected first block report to make some blocks safe.", safe > 0);
+    assertTrue("Did not expect first block report to make all blocks safe.", safe < 15);
+
     assertTrue(NameNodeAdapter.safeModeInitializedReplQueues(nn));
-    assertEquals(15 - safe, nn.getNamesystem().getUnderReplicatedBlocks());
+
+    // Ensure that UnderReplicatedBlocks goes up to 15 - safe. Misreplicated
+    // blocks are processed asynchronously so this may take a few seconds.
+    // Failure here will manifest as a test timeout.
+    BlockManagerTestUtil.updateState(nn.getNamesystem().getBlockManager());
+    long underReplicatedBlocks = nn.getNamesystem().getUnderReplicatedBlocks();
+    while (underReplicatedBlocks != (15 - safe)) {
+      LOG.info("UnderReplicatedBlocks expected=" + (15 - safe) +
+               ", actual=" + underReplicatedBlocks);
+      Thread.sleep(100);
+      BlockManagerTestUtil.updateState(nn.getNamesystem().getBlockManager());
+      underReplicatedBlocks = nn.getNamesystem().getUnderReplicatedBlocks();
+    }
 
     cluster.restartDataNodes();
   }