From cbb6aac2a7ee941bdaa87b7f9b141e875f4568f8 Mon Sep 17 00:00:00 2001
From: Michael Stack <stack@apache.org>
Date: Mon, 29 Aug 2011 17:51:44 +0000
Subject: [PATCH] HBASE-4124 ZK restarted while a region is being assigned, new
 active HM re-assigns it but the RS warns 'already online on this server'

git-svn-id: https://svn.apache.org/repos/asf/hbase/trunk@1162919 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                   |  3 +++
 .../hbase/master/AssignmentManager.java       | 27 ++++++++++++++-----
 2 files changed, 24 insertions(+), 6 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 211654e30c3..53f885ca864 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -496,6 +496,9 @@ Release 0.90.5 - Unreleased
    HBASE-4253  TestScannerTimeOut.test3686a and TestHTablePool.
                testReturnDifferentTable() failure because of using new
                HTable(tablename) (ramkrishna.s.vasudevan)
+   HBASE-4124  ZK restarted while a region is being assigned, new active HM
+               re-assigns it but the RS warns 'already online on this server'
+               (Gaojinchao)
 
   IMPROVEMENT
    HBASE-4205  Enhance HTable javadoc (Eric Charles)
diff --git a/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java b/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java
index 29858b2c13b..90e1a3c8b75 100644
--- a/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java
+++ b/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java
@@ -369,8 +369,11 @@ public class AssignmentManager extends ZooKeeperListener {
     synchronized (regionsInTransition) {
       switch (data.getEventType()) {
       case RS_ZK_REGION_CLOSING:
-        if (isOnDeadServer(regionInfo, deadServers)) {
-          // If was on dead server, its closed now.  Force to OFFLINE and this
+        // If zk node of the region was updated by a live server skip this
+        // region and just add it into RIT.
+        if (isOnDeadServer(regionInfo, deadServers) &&
+            (data.getOrigin() == null || !serverManager.isServerOnline(data.getOrigin()))) {
+          // If was on dead server, its closed now. Force to OFFLINE and this
           // will get it reassigned if appropriate
           forceOffline(regionInfo, data);
         } else {
@@ -416,10 +419,10 @@ public class AssignmentManager extends ZooKeeperListener {
           LOG.warn("Region in transition " + regionInfo.getEncodedName() +
             " references a null server; letting RIT timeout so will be " +
             "assigned elsewhere");
-          break;
-        }
-        if (isOnDeadServer(regionInfo, deadServers)) {
-          // If was on a dead server, then its not open any more; needs handling.
+        } else if (isOnDeadServer(regionInfo, deadServers) &&
+            !serverManager.isServerOnline(sn)) {
+          // If was on a dead server, then its not open any more; needs
+          // handling.
           forceOffline(regionInfo, data);
         } else {
           new OpenedRegionHandler(master, this, regionInfo, sn).process();
@@ -1957,6 +1960,18 @@ public class AssignmentManager extends ZooKeeperListener {
         Result result = region.getSecond();
         // If region was in transition (was in zk) force it offline for reassign
         try {
+          RegionTransitionData data = ZKAssign.getData(watcher,
+              regionInfo.getEncodedName());
+
+          // If zk node of this region has been updated by a live server,
+          // we consider that this region is being handled.
+          // So we should skip it and process it in processRegionsInTransition.
+          if (data != null && data.getOrigin() != null &&
+	      serverManager.isServerOnline(data.getOrigin())) {
+            LOG.info("The region " + regionInfo.getEncodedName()
+                + "is being handled on " + data.getOrigin());
+            continue;
+          }
           // Process with existing RS shutdown code
           boolean assign =
             ServerShutdownHandler.processDeadRegion(regionInfo, result, this,