From 6d2aaf8833bfced99269c749881b428b2638b564 Mon Sep 17 00:00:00 2001
From: Michael Stack <stack@apache.org>
Date: Wed, 20 Oct 2010 05:12:19 +0000
Subject: [PATCH] HBASE-3128 On assign, if ConnectException, reassign another
 server

git-svn-id: https://svn.apache.org/repos/asf/hbase/trunk@1024524 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                   |  1 +
 .../hbase/master/AssignmentManager.java       | 13 ++------
 .../hadoop/hbase/master/ServerManager.java    | 33 ++++++++++---------
 3 files changed, 22 insertions(+), 25 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 62a470daf2a..2a6b7e4d3e3 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -1018,6 +1018,7 @@ Release 0.21.0 - Unreleased
    HBASE-3115  HBaseClient wastes 1 TCP packet per RPC
    HBASE-3076  Allow to disable automatic shipping of dependency jars
                for mapreduce jobs (Bruno Dumon)
+   HBASE-3128  On assign, if ConnectException, reassign another server
 
   NEW FEATURES
    HBASE-1961  HBase EC2 scripts
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 f594dfd8acc..05341aa079e 100644
--- a/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java
+++ b/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java
@@ -729,7 +729,7 @@ public class AssignmentManager extends ZooKeeperListener {
     } catch (Throwable t) {
       LOG.warn("Failed assignment of " +
         state.getRegion().getRegionNameAsString() + " to " +
-        plan.getDestination(), t);
+        plan.getDestination() + ", trying to assign elsewhere instead", t);
       // Clean out plan we failed execute and one that doesn't look like it'll
       // succeed anyways; we need a new plan!
       this.regionPlans.remove(state.getRegion().getEncodedName());
@@ -873,18 +873,11 @@ public class AssignmentManager extends ZooKeeperListener {
     try {
       serverManager.sendRegionClose(regions.get(region), state.getRegion());
     } catch (IOException e) {
-      if (e instanceof RemoteException) {
-        e = ((RemoteException)e).unwrapRemoteException();
-      }
-      if (e instanceof NotServingRegionException) {
-        LOG.warn("Attempted to close region " + region.getRegionNameAsString() +
-          " but got an NSRE", e);
-      }
-      // For now call abort if unexpected exception -- seeing it up in hudson.
+      // For now call abort if unexpected exception -- radical, but will get fellas attention.
       // St.Ack 20101012
       this.master.abort("Remote unexpected exception", e);
     } catch (Throwable t) {
-      // For now call abort if unexpected exception -- seeing it up in hudson.
+      // For now call abort if unexpected exception -- radical, but will get fellas attention.
       // St.Ack 20101012
       this.master.abort("Unexpected exception", t);
     }
diff --git a/src/main/java/org/apache/hadoop/hbase/master/ServerManager.java b/src/main/java/org/apache/hadoop/hbase/master/ServerManager.java
index 29d02529edf..200e1e47e6f 100644
--- a/src/main/java/org/apache/hadoop/hbase/master/ServerManager.java
+++ b/src/main/java/org/apache/hadoop/hbase/master/ServerManager.java
@@ -41,6 +41,7 @@ import org.apache.hadoop.hbase.PleaseHoldException;
 import org.apache.hadoop.hbase.Server;
 import org.apache.hadoop.hbase.Stoppable;
 import org.apache.hadoop.hbase.YouAreDeadException;
+import org.apache.hadoop.hbase.ZooKeeperConnectionException;
 import org.apache.hadoop.hbase.client.HConnection;
 import org.apache.hadoop.hbase.client.HConnectionManager;
 import org.apache.hadoop.hbase.ipc.HRegionInterface;
@@ -48,7 +49,6 @@ import org.apache.hadoop.hbase.master.handler.ServerShutdownHandler;
 import org.apache.hadoop.hbase.master.metrics.MasterMetrics;
 import org.apache.hadoop.hbase.regionserver.Leases.LeaseStillHeldException;
 import org.apache.hadoop.hbase.util.Threads;
-import org.apache.hadoop.ipc.RemoteException;
 import org.apache.hadoop.util.StringUtils;
 
 /**
@@ -558,21 +558,24 @@ public class ServerManager {
     hri.closeRegion(region);
   }
 
-  private HRegionInterface getServerConnection(HServerInfo info) {
-    try {
-      HConnection connection =
-        HConnectionManager.getConnection(this.master.getConfiguration());
-      HRegionInterface hri = serverConnections.get(info.getServerName());
-      if (hri == null) {
-        LOG.debug("New connection to " + info.getServerName());
-        hri = connection.getHRegionConnection(info.getServerAddress(), false);
-        serverConnections.put(info.getServerName(), hri);
-      }
-      return hri;
-    } catch (IOException e) {
-      LOG.error("Error connecting to region server", e);
-      throw new RuntimeException("Fatal error connection to RS", e);
+  /**
+   * @param info
+   * @return
+   * @throws IOException
+   * @throws RetriesExhaustedException wrapping a ConnectException if failed
+   * putting up proxy.
+   */
+  private HRegionInterface getServerConnection(HServerInfo info)
+  throws IOException {
+    HConnection connection =
+      HConnectionManager.getConnection(this.master.getConfiguration());
+    HRegionInterface hri = serverConnections.get(info.getServerName());
+    if (hri == null) {
+      LOG.debug("New connection to " + info.getServerName());
+      hri = connection.getHRegionConnection(info.getServerAddress(), false);
+      this.serverConnections.put(info.getServerName(), hri);
     }
+    return hri;
   }
 
   /**