From 08f061b3665935b8c9c368c21b3483e064968ab4 Mon Sep 17 00:00:00 2001
From: Eli Collins <eli@apache.org>
Date: Wed, 12 Sep 2012 04:12:49 +0000
Subject: [PATCH] HDFS-3664. BlockManager race when stopping active services.
 Contributed by Colin Patrick McCabe

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-2@1383756 13f79535-47bb-0310-9956-ffa450edef68
---
 hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt               | 3 +++
 .../hadoop/hdfs/server/blockmanagement/BlockManager.java  | 7 +++----
 .../hdfs/server/blockmanagement/DatanodeManager.java      | 8 +++++++-
 .../hdfs/server/blockmanagement/HeartbeatManager.java     | 5 +++++
 4 files changed, 18 insertions(+), 5 deletions(-)

diff --git a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
index c60c67daaf6..bdaf7a7880a 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
+++ b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
@@ -590,6 +590,9 @@ Release 2.0.2-alpha - 2012-09-07
 
     HDFS-2757. Cannot read a local block that's being written to when
     using the local read short circuit. (Jean-Daniel Cryans via eli)
+
+    HDFS-3664. BlockManager race when stopping active services.
+    (Colin Patrick McCabe via eli)
     
   BREAKDOWN OF HDFS-3042 SUBTASKS
 
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 45612186d98..e851279ba07 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
@@ -364,11 +364,10 @@ public class BlockManager {
         replicationThread.join(3000);
       }
     } catch (InterruptedException ie) {
-    } finally {
-      if (pendingReplications != null) pendingReplications.stop();
-      blocksMap.close();
-      datanodeManager.close();
     }
+    datanodeManager.close();
+    pendingReplications.stop();
+    blocksMap.close();
   }
 
   /** @return the datanodeManager */
diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeManager.java b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeManager.java
index 3ccbc8462a3..5f28f99c7df 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeManager.java
+++ b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeManager.java
@@ -183,7 +183,13 @@ public class DatanodeManager {
   }
 
   void close() {
-    if (decommissionthread != null) decommissionthread.interrupt();
+    if (decommissionthread != null) {
+      decommissionthread.interrupt();
+      try {
+        decommissionthread.join(3000);
+      } catch (InterruptedException e) {
+      }
+    }
     heartbeatManager.close();
   }
 
diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java
index 449619ced72..30300077077 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java
+++ b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java
@@ -74,6 +74,11 @@ class HeartbeatManager implements DatanodeStatistics {
 
   void close() {
     heartbeatThread.interrupt();
+    try {
+      // This will no effect if the thread hasn't yet been started.
+      heartbeatThread.join(3000);
+    } catch (InterruptedException e) {
+    }
   }
   
   synchronized int getLiveDatanodeCount() {