From 2a05c630044a18af10dd8c74db394f23581e35a9 Mon Sep 17 00:00:00 2001
From: Konstantin Shvachko <shv@apache.org>
Date: Tue, 11 Jun 2013 01:40:52 +0000
Subject: [PATCH] HDFS-4878. On Remove Block, block is not removed from
 neededReplications queue. Contributed by Tao Luo.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1491672 13f79535-47bb-0310-9956-ffa450edef68
---
 hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt   |  3 +
 .../server/blockmanagement/BlockManager.java  |  3 +-
 .../hdfs/server/namenode/TestMetaSave.java    | 75 ++++++++++++++++---
 3 files changed, 70 insertions(+), 11 deletions(-)

diff --git a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
index b9203345400..863a885caf1 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
+++ b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
@@ -354,6 +354,9 @@ Release 2.1.0-beta - UNRELEASED
     HDFS-4586. TestDataDirs.testGetDataDirsFromURIs fails with all directories
     in dfs.datanode.data.dir are invalid. (Ivan Mitic via atm)
 
+    HDFS-4878. On Remove Block, block is not removed from neededReplications
+    queue. (Tao Luo via shv)
+
   BREAKDOWN OF HDFS-347 SUBTASKS AND RELATED JIRAS
 
     HDFS-4353. Encapsulate connections to peers in Peer and PeerServer classes.
diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java
index 4e082b702bc..b904ce36f4a 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java
+++ b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java
@@ -2869,8 +2869,9 @@ assert storedBlock.findDatanode(dn) < 0 : "Block " + block
     addToInvalidates(block);
     corruptReplicas.removeFromCorruptReplicasMap(block);
     blocksMap.removeBlock(block);
-    // Remove the block from pendingReplications
+    // Remove the block from pendingReplications and neededReplications
     pendingReplications.remove(block);
+    neededReplications.remove(block, UnderReplicatedBlocks.LEVEL);
     if (postponedMisreplicatedBlocks.remove(block)) {
       postponedMisreplicatedBlocksCount.decrementAndGet();
     }
diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestMetaSave.java b/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestMetaSave.java
index 2f2b6888fca..729358562b2 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestMetaSave.java
+++ b/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestMetaSave.java
@@ -99,16 +99,71 @@ public class TestMetaSave {
         + "metasave.out.txt";
     FileInputStream fstream = new FileInputStream(logFile);
     DataInputStream in = new DataInputStream(fstream);
-    BufferedReader reader = new BufferedReader(new InputStreamReader(in));
-    String line = reader.readLine();
-    assertTrue(line.equals("3 files and directories, 2 blocks = 5 total"));
-    line = reader.readLine();
-    assertTrue(line.equals("Live Datanodes: 1"));
-    line = reader.readLine();
-    assertTrue(line.equals("Dead Datanodes: 1"));
-    line = reader.readLine();
-    line = reader.readLine();
-    assertTrue(line.matches("^/filestatus[01]:.*"));
+    BufferedReader reader = null;
+    try {
+      reader = new BufferedReader(new InputStreamReader(in));
+      String line = reader.readLine();
+      assertTrue(line.equals(
+          "3 files and directories, 2 blocks = 5 total"));
+      line = reader.readLine();
+      assertTrue(line.equals("Live Datanodes: 1"));
+      line = reader.readLine();
+      assertTrue(line.equals("Dead Datanodes: 1"));
+      line = reader.readLine();
+      line = reader.readLine();
+      assertTrue(line.matches("^/filestatus[01]:.*"));
+    } finally {
+      if (reader != null)
+        reader.close();
+    }
+  }
+
+  /**
+   * Tests metasave after delete, to make sure there are no orphaned blocks
+   */
+  @Test
+  public void testMetasaveAfterDelete()
+      throws IOException, InterruptedException {
+
+    final FSNamesystem namesystem = cluster.getNamesystem();
+
+    for (int i = 0; i < 2; i++) {
+      Path file = new Path("/filestatus" + i);
+      createFile(fileSys, file);
+    }
+
+    cluster.stopDataNode(1);
+    // wait for namenode to discover that a datanode is dead
+    Thread.sleep(15000);
+    namesystem.setReplication("/filestatus0", (short) 4);
+    namesystem.delete("/filestatus0", true);
+    namesystem.delete("/filestatus1", true);
+
+    namesystem.metaSave("metasaveAfterDelete.out.txt");
+
+    // Verification
+    String logFile = System.getProperty("hadoop.log.dir") + "/"
+        + "metasaveAfterDelete.out.txt";
+    BufferedReader reader = null;
+    try {
+      FileInputStream fstream = new FileInputStream(logFile);
+      DataInputStream in = new DataInputStream(fstream);
+      reader = new BufferedReader(new InputStreamReader(in));
+      reader.readLine();
+      String line = reader.readLine();
+      assertTrue(line.equals("Live Datanodes: 1"));
+      line = reader.readLine();
+      assertTrue(line.equals("Dead Datanodes: 1"));
+      line = reader.readLine();
+      assertTrue(line.equals("Metasave: Blocks waiting for replication: 0"));
+      line = reader.readLine();
+      assertTrue(line.equals("Mis-replicated blocks that have been postponed:"));
+      line = reader.readLine();
+      assertTrue(line.equals("Metasave: Blocks being replicated: 0"));
+    } finally {
+      if (reader != null)
+        reader.close();
+    }
   }
 
   @AfterClass