From 327a995561270bc01c024e38ddd1063c42ff2e3c Mon Sep 17 00:00:00 2001
From: Junping Du <junping_du@apache.org>
Date: Wed, 11 Dec 2013 08:51:12 +0000
Subject: [PATCH] HDFS-5580. Fix infinite loop in
 Balancer.waitForMoveCompletion. (Binglin Chang via junping_du)

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1550074 13f79535-47bb-0310-9956-ffa450edef68
---
 hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt   |  3 ++
 .../hadoop/hdfs/server/balancer/Balancer.java | 31 ++++++++++---------
 2 files changed, 19 insertions(+), 15 deletions(-)

diff --git a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
index de970ba7d6..6090ef78e4 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
+++ b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
@@ -639,6 +639,9 @@ Release 2.4.0 - UNRELEASED
     HDFS-5533. Symlink delete/create should be treated as DELETE/CREATE in snapshot diff 
     report. (Binglin Chang via jing9)
 
+    HDFS-5580. Fix infinite loop in Balancer.waitForMoveCompletion.
+    (Binglin Chang via junping_du)
+
 Release 2.3.0 - UNRELEASED
 
   INCOMPATIBLE CHANGES
diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Balancer.java b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Balancer.java
index befdd90a79..ff09ca54dd 100644
--- a/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Balancer.java
+++ b/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/balancer/Balancer.java
@@ -292,26 +292,27 @@ private boolean markMovedIfGoodBlock(BalancerBlock block) {
      */
     private boolean chooseProxySource() {
       final DatanodeInfo targetDN = target.getDatanode();
-      boolean find = false;
-      for (BalancerDatanode loc : block.getLocations()) {
-        // check if there is replica which is on the same rack with the target
-        if (cluster.isOnSameRack(loc.getDatanode(), targetDN) && addTo(loc)) {
-          find = true;
-          // if cluster is not nodegroup aware or the proxy is on the same 
-          // nodegroup with target, then we already find the nearest proxy
-          if (!cluster.isNodeGroupAware() 
-              || cluster.isOnSameNodeGroup(loc.getDatanode(), targetDN)) {
+      // if node group is supported, first try add nodes in the same node group
+      if (cluster.isNodeGroupAware()) {
+        for (BalancerDatanode loc : block.getLocations()) {
+          if (cluster.isOnSameNodeGroup(loc.getDatanode(), targetDN) && addTo(loc)) {
             return true;
           }
         }
-        
-        if (!find) {
-          // find out a non-busy replica out of rack of target
-          find = addTo(loc);
+      }
+      // check if there is replica which is on the same rack with the target
+      for (BalancerDatanode loc : block.getLocations()) {
+        if (cluster.isOnSameRack(loc.getDatanode(), targetDN) && addTo(loc)) {
+          return true;
         }
       }
-      
-      return find;
+      // find out a non-busy replica
+      for (BalancerDatanode loc : block.getLocations()) {
+        if (addTo(loc)) {
+          return true;
+        }
+      }
+      return false;
     }
     
     // add a BalancerDatanode as proxy source for specific block movement