Merged trunk to branch-trunk-win

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-trunk-win@1443757 13f79535-47bb-0310-9956-ffa450edef68

hadoop-assemblies/src/main/resources/assemblies/hadoop-dist.xml

@@ -99,13 +99,19 @@
       <includes>
         <include>${project.artifactId}-${project.version}.jar</include>
         <include>${project.artifactId}-${project.version}-tests.jar</include>
-        <include>${project.artifactId}-${project.version}-sources.jar</include>
-        <include>${project.artifactId}-${project.version}-test-sources.jar</include>
       </includes>
       <excludes>
         <exclude>hadoop-tools-dist-*.jar</exclude>
       </excludes>
     </fileSet>
+    <fileSet>
+      <directory>${project.build.directory}</directory>
+      <outputDirectory>/share/hadoop/${hadoop.component}/sources</outputDirectory>
+      <includes>
+        <include>${project.artifactId}-${project.version}-sources.jar</include>
+        <include>${project.artifactId}-${project.version}-test-sources.jar</include>
+      </includes>
+    </fileSet>
     <fileSet>
       <directory>${basedir}/dev-support/jdiff</directory>
       <outputDirectory>/share/hadoop/${hadoop.component}/jdiff</outputDirectory>

hadoop-common-project/hadoop-common/CHANGES.txt

@@ -149,6 +149,8 @@ Trunk (Unreleased)
     HADOOP-8924. Add maven plugin alternative to shell script to save
     package-info.java. (Chris Nauroth via suresh)
 
+    HADOOP-9277. Improve javadoc for FileContext. (Andrew Wang via suresh)
+
   BUG FIXES
 
     HADOOP-8419. Fixed GzipCode NPE reset for IBM JDK. (Yu Li via eyang)
@@ -339,7 +341,23 @@ Trunk (Unreleased)
 
     HADOOP-9190. packaging docs is broken. (Andy Isaacson via atm)
 
-Release 2.0.3-alpha - Unreleased 
+Release 2.0.4-beta - UNRELEASED
+
+  INCOMPATIBLE CHANGES
+
+  NEW FEATURES
+
+    HADOOP-9283. Add support for running the Hadoop client on AIX. (atm)
+
+  IMPROVEMENTS
+    HADOOP-9253. Capture ulimit info in the logs at service start time.
+    (Arpit Gupta via suresh)
+
+  OPTIMIZATIONS
+
+  BUG FIXES
+
+Release 2.0.3-alpha - 2013-02-06 
 
   INCOMPATIBLE CHANGES
 
@@ -467,6 +485,9 @@ Release 2.0.3-alpha - Unreleased
     HADOOP-9231. Parametrize staging URL for the uniformity of
     distributionManagement. (Konstantin Boudnik via suresh)
 
+    HADOOP-9276. Allow BoundedByteArrayOutputStream to be resettable.
+    (Arun Murthy via hitesh)
+
   OPTIMIZATIONS
 
     HADOOP-8866. SampleQuantiles#query is O(N^2) instead of O(N). (Andrew Wang
@@ -592,10 +613,23 @@ Release 2.0.3-alpha - Unreleased
     HADOOP-9221. Convert remaining xdocs to APT. (Andy Isaacson via atm)
 
     HADOOP-8981. TestMetricsSystemImpl fails on Windows. (Xuan Gong via suresh)
+    
+    HADOOP-9124. SortedMapWritable violates contract of Map interface for
+    equals() and hashCode(). (Surenkumar Nihalani via tomwhite)
 
     HADOOP-9278. Fix the file handle leak in HarMetaData.parseMetaData() in
     HarFileSystem. (Chris Nauroth via szetszwo)
 
+    HADOOP-9252. In StringUtils, humanReadableInt(..) has a race condition and
+    the synchronization of limitDecimalTo2(double) can be avoided.  (szetszwo)
+
+    HADOOP-9260. Hadoop version may be not correct when starting name node or
+    data node. (Chris Nauroth via jlowe)
+
+    HADOOP-9289. FsShell rm -f fails for non-matching globs. (Daryn Sharp via
+    suresh)
+
+
 Release 2.0.2-alpha - 2012-09-07 
 
   INCOMPATIBLE CHANGES
@@ -1297,6 +1331,9 @@ Release 0.23.7 - UNRELEASED
     HADOOP-8849. FileUtil#fullyDelete should grant the target directories +rwx
     permissions (Ivan A. Veselovsky via bobby)
 
+    HADOOP-9067. provide test for LocalFileSystem.reportChecksumFailure
+    (Ivan A. Veselovsky via bobby) 
+
   OPTIMIZATIONS
 
   BUG FIXES

hadoop-common-project/hadoop-common/src/main/bin/hadoop-daemon.sh

@@ -83,7 +83,8 @@ fi
 if [ "$command" == "datanode" ] && [ "$EUID" -eq 0 ] && [ -n "$HADOOP_SECURE_DN_USER" ]; then
   export HADOOP_PID_DIR=$HADOOP_SECURE_DN_PID_DIR
   export HADOOP_LOG_DIR=$HADOOP_SECURE_DN_LOG_DIR
-  export HADOOP_IDENT_STRING=$HADOOP_SECURE_DN_USER   
+  export HADOOP_IDENT_STRING=$HADOOP_SECURE_DN_USER
+  starting_secure_dn="true"
 fi
 
 if [ "$HADOOP_IDENT_STRING" = "" ]; then
@@ -154,7 +155,17 @@ case $startStop in
       ;;
     esac
     echo $! > $pid
-    sleep 1; head "$log"
+    sleep 1
+    # capture the ulimit output
+    if [ "true" = "$starting_secure_dn" ]; then
+      echo "ulimit -a for secure datanode user $HADOOP_SECURE_DN_USER" >> $log
+      # capture the ulimit info for the appropriate user
+      su --shell=/bin/bash $HADOOP_SECURE_DN_USER -c 'ulimit -a' >> $log 2>&1
+    else
+      echo "ulimit -a for user $USER" >> $log
+      ulimit -a >> $log 2>&1
+    fi
+    head -30 "$log"
     sleep 3;
     if ! ps -p $! > /dev/null ; then
       exit 1

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/FileContext.java

@@ -57,70 +57,60 @@ import org.apache.hadoop.security.token.Token;
 import org.apache.hadoop.util.ShutdownHookManager;
 
 /**
- * The FileContext class provides an interface to the application writer for
- * using the Hadoop file system.
- * It provides a set of methods for the usual operation: create, open, 
- * list, etc 
+ * The FileContext class provides an interface for users of the Hadoop
+ * file system. It exposes a number of file system operations, e.g. create,
+ * open, list.
  * 
- * <p>
- * <b> *** Path Names *** </b>
- * <p>
+ * <h2>Path Names</h2>
  * 
- * The Hadoop file system supports a URI name space and URI names.
- * It offers a forest of file systems that can be referenced using fully
- * qualified URIs.
- * Two common Hadoop file systems implementations are
+ * The Hadoop file system supports a URI namespace and URI names. This enables
+ * multiple types of file systems to be referenced using fully-qualified URIs.
+ * Two common Hadoop file system implementations are
  * <ul>
- * <li> the local file system: file:///path
- * <li> the hdfs file system hdfs://nnAddress:nnPort/path
+ * <li>the local file system: file:///path
+ * <li>the HDFS file system: hdfs://nnAddress:nnPort/path
  * </ul>
  * 
- * While URI names are very flexible, it requires knowing the name or address
- * of the server. For convenience one often wants to access the default system
- * in one's environment without knowing its name/address. This has an
- * additional benefit that it allows one to change one's default fs
- *  (e.g. admin moves application from cluster1 to cluster2).
+ * The Hadoop file system also supports additional naming schemes besides URIs.
+ * Hadoop has the concept of a <i>default file system</i>, which implies a
+ * default URI scheme and authority. This enables <i>slash-relative names</i>
+ * relative to the default FS, which are more convenient for users and
+ * application writers. The default FS is typically set by the user's
+ * environment, though it can also be manually specified.
  * <p>
  * 
- * To facilitate this, Hadoop supports a notion of a default file system.
- * The user can set his default file system, although this is
- * typically set up for you in your environment via your default config.
- * A default file system implies a default scheme and authority; slash-relative
- * names (such as /for/bar) are resolved relative to that default FS.
- * Similarly a user can also have working-directory-relative names (i.e. names
- * not starting with a slash). While the working directory is generally in the
- * same default FS, the wd can be in a different FS.
+ * Hadoop also supports <i>working-directory-relative</i> names, which are paths
+ * relative to the current working directory (similar to Unix). The working
+ * directory can be in a different file system than the default FS.
  * <p>
- *  Hence Hadoop path names can be one of:
- *  <ul>
- *  <li> fully qualified URI: scheme://authority/path
- *  <li> slash relative names: /path relative to the default file system
- *  <li> wd-relative names: path  relative to the working dir
- *  </ul>   
+ * Thus, Hadoop path names can be specified as one of the following:
+ * <ul>
+ * <li>a fully-qualified URI: scheme://authority/path (e.g.
+ * hdfs://nnAddress:nnPort/foo/bar)
+ * <li>a slash-relative name: path relative to the default file system (e.g.
+ * /foo/bar)
+ * <li>a working-directory-relative name: path relative to the working dir (e.g.
+ * foo/bar)
+ * </ul>
  *  Relative paths with scheme (scheme:foo/bar) are illegal.
  *  
- *  <p>
- *  <b>****The Role of the FileContext and configuration defaults****</b>
- *  <p>
- *  The FileContext provides file namespace context for resolving file names;
- *  it also contains the umask for permissions, In that sense it is like the
- *  per-process file-related state in Unix system.
- *  These two properties
- *  <ul> 
- *  <li> default file system i.e your slash)
- *  <li> umask
- *  </ul>
- *  in general, are obtained from the default configuration file
- *  in your environment,  (@see {@link Configuration}).
- *  
- *  No other configuration parameters are obtained from the default config as 
- *  far as the file context layer is concerned. All file system instances
- *  (i.e. deployments of file systems) have default properties; we call these
- *  server side (SS) defaults. Operation like create allow one to select many 
- *  properties: either pass them in as explicit parameters or use
- *  the SS properties.
- *  <p>
- *  The file system related SS defaults are
+ * <h2>Role of FileContext and Configuration Defaults</h2>
+ *
+ * The FileContext is the analogue of per-process file-related state in Unix. It
+ * contains two properties:
+ * 
+ * <ul>
+ * <li>the default file system (for resolving slash-relative names)
+ * <li>the umask (for file permissions)
+ * </ul>
+ * In general, these properties are obtained from the default configuration file
+ * in the user's environment (see {@link Configuration}).
+ * 
+ * Further file system properties are specified on the server-side. File system
+ * operations default to using these server-side defaults unless otherwise
+ * specified.
+ * <p>
+ * The file system related server-side defaults are:
  *  <ul>
  *  <li> the home directory (default is "/user/userName")
  *  <li> the initial wd (only for local fs)
@@ -131,34 +121,34 @@ import org.apache.hadoop.util.ShutdownHookManager;
  *  <li> checksum option. (checksumType and  bytesPerChecksum)
  *  </ul>
  *
- * <p>
- * <b> *** Usage Model for the FileContext class *** </b>
- * <p>
+ * <h2>Example Usage</h2>
+ *
  * Example 1: use the default config read from the $HADOOP_CONFIG/core.xml.
  *   Unspecified values come from core-defaults.xml in the release jar.
  *  <ul>  
  *  <li> myFContext = FileContext.getFileContext(); // uses the default config
  *                                                // which has your default FS 
  *  <li>  myFContext.create(path, ...);
- *  <li>  myFContext.setWorkingDir(path)
+ *  <li>  myFContext.setWorkingDir(path);
  *  <li>  myFContext.open (path, ...);  
+ *  <li>...
  *  </ul>  
  * Example 2: Get a FileContext with a specific URI as the default FS
  *  <ul>  
- *  <li> myFContext = FileContext.getFileContext(URI)
+ *  <li> myFContext = FileContext.getFileContext(URI);
  *  <li> myFContext.create(path, ...);
- *   ...
- * </ul> 
+ *  <li>...
+ * </ul>
  * Example 3: FileContext with local file system as the default
  *  <ul> 
- *  <li> myFContext = FileContext.getLocalFSFileContext()
+ *  <li> myFContext = FileContext.getLocalFSFileContext();
  *  <li> myFContext.create(path, ...);
  *  <li> ...
  *  </ul> 
  * Example 4: Use a specific config, ignoring $HADOOP_CONFIG
  *  Generally you should not need use a config unless you are doing
  *   <ul> 
- *   <li> configX = someConfigSomeOnePassedToYou.
+ *   <li> configX = someConfigSomeOnePassedToYou;
  *   <li> myFContext = getFileContext(configX); // configX is not changed,
  *                                              // is passed down 
  *   <li> myFContext.create(path, ...);

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/shell/Delete.java

@@ -21,6 +21,7 @@ package org.apache.hadoop.fs.shell;
 import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.util.LinkedList;
+import java.util.List;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
@@ -28,6 +29,7 @@ import org.apache.hadoop.fs.PathIOException;
 import org.apache.hadoop.fs.PathIsDirectoryException;
 import org.apache.hadoop.fs.PathIsNotDirectoryException;
 import org.apache.hadoop.fs.PathIsNotEmptyDirectoryException;
+import org.apache.hadoop.fs.PathNotFoundException;
 import org.apache.hadoop.fs.Trash;
 
 /**
@@ -71,6 +73,19 @@ class Delete {
       skipTrash = cf.getOpt("skipTrash");
     }
 
+    @Override
+    protected List<PathData> expandArgument(String arg) throws IOException {
+      try {
+        return super.expandArgument(arg);
+      } catch (PathNotFoundException e) {
+        if (!ignoreFNF) {
+          throw e;
+        }
+        // prevent -f on a non-existent glob from failing
+        return new LinkedList<PathData>();
+      }
+    }
+
     @Override
     protected void processNonexistentPath(PathData item) throws IOException {
       if (!ignoreFNF) super.processNonexistentPath(item);

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/shell/FsUsage.java

@@ -48,7 +48,7 @@ class FsUsage extends FsCommand {
   
   protected String formatSize(long size) {
     return humanReadable
-        ? StringUtils.humanReadableInt(size)
+        ? StringUtils.TraditionalBinaryPrefix.long2String(size, "", 1)
         : String.valueOf(size);
   }
 

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/shell/Ls.java

@@ -67,7 +67,7 @@ class Ls extends FsCommand {
   protected boolean humanReadable = false;
   protected String formatSize(long size) {
     return humanReadable
-      ? StringUtils.humanReadableInt(size)
+      ? StringUtils.TraditionalBinaryPrefix.long2String(size, "", 1)
       : String.valueOf(size);
   }
 

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/BoundedByteArrayOutputStream.java

@@ -32,9 +32,10 @@ import org.apache.hadoop.classification.InterfaceStability;
 @InterfaceAudience.LimitedPrivate({"HDFS", "MapReduce"})
 @InterfaceStability.Unstable
 public class BoundedByteArrayOutputStream extends OutputStream {
-  private final byte[] buffer;
+  private byte[] buffer;
+  private int startOffset;
   private int limit;
-  private int count;
+  private int currentPointer;
 
   /**
    * Create a BoundedByteArrayOutputStream with the specified
@@ -52,20 +53,30 @@ public class BoundedByteArrayOutputStream extends OutputStream {
    * @param limit The maximum limit upto which data can be written
    */
   public BoundedByteArrayOutputStream(int capacity, int limit) {
+    this(new byte[capacity], 0, limit);
+  }
+
+  protected BoundedByteArrayOutputStream(byte[] buf, int offset, int limit) {
+    resetBuffer(buf, offset, limit);
+  }
+  
+  protected void resetBuffer(byte[] buf, int offset, int limit) {
+    int capacity = buf.length - offset;
     if ((capacity < limit) || (capacity | limit) < 0) {
       throw new IllegalArgumentException("Invalid capacity/limit");
     }
-    this.buffer = new byte[capacity];
-    this.limit = limit;
-    this.count = 0;
+    this.buffer = buf;
+    this.startOffset = offset;
+    this.currentPointer = offset;
+    this.limit = offset + limit;
   }
-
+  
   @Override
   public void write(int b) throws IOException {
-    if (count >= limit) {
+    if (currentPointer >= limit) {
       throw new EOFException("Reaching the limit of the buffer.");
     }
-    buffer[count++] = (byte) b;
+    buffer[currentPointer++] = (byte) b;
   }
 
   @Override
@@ -77,12 +88,12 @@ public class BoundedByteArrayOutputStream extends OutputStream {
       return;
     }
 
-    if (count + len > limit) {
+    if (currentPointer + len > limit) {
       throw new EOFException("Reach the limit of the buffer");
     }
 
-    System.arraycopy(b, off, buffer, count, len);
-    count += len;
+    System.arraycopy(b, off, buffer, currentPointer, len);
+    currentPointer += len;
   }
 
   /**
@@ -90,17 +101,17 @@ public class BoundedByteArrayOutputStream extends OutputStream {
    * @param newlim New Limit
    */
   public void reset(int newlim) {
-    if (newlim > buffer.length) {
+    if (newlim > (buffer.length - startOffset)) {
       throw new IndexOutOfBoundsException("Limit exceeds buffer size");
     }
     this.limit = newlim;
-    this.count = 0;
+    this.currentPointer = startOffset;
   }
 
   /** Reset the buffer */
   public void reset() {
-    this.limit = buffer.length;
-    this.count = 0;
+    this.limit = buffer.length - startOffset;
+    this.currentPointer = startOffset;
   }
 
   /** Return the current limit */
@@ -119,6 +130,10 @@ public class BoundedByteArrayOutputStream extends OutputStream {
    * currently in the buffer.
    */
   public int size() {
-    return count;
+    return currentPointer - startOffset;
+  }
+  
+  public int available() {
+    return limit - currentPointer;
   }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/SortedMapWritable.java

@@ -203,4 +203,27 @@ public class SortedMapWritable extends AbstractMapWritable
       e.getValue().write(out);
     }
   }
+
+  @Override
+  public boolean equals(Object obj) {
+    if (this == obj) {
+      return true;
+    }
+
+    if (obj instanceof SortedMapWritable) {
+      Map map = (Map) obj;
+      if (size() != map.size()) {
+        return false;
+      }
+
+      return entrySet().equals(map.entrySet());
+    }
+
+    return false;
+  }
+
+  @Override
+  public int hashCode() {
+    return instance.hashCode();
+  }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/retry/RetryPolicies.java

@@ -35,6 +35,7 @@ import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.ipc.RemoteException;
 import org.apache.hadoop.ipc.StandbyException;
+import org.apache.hadoop.net.ConnectTimeoutException;
 
 /**
  * <p>
@@ -543,6 +544,7 @@ public class RetryPolicies {
           e instanceof NoRouteToHostException ||
           e instanceof UnknownHostException ||
           e instanceof StandbyException ||
+          e instanceof ConnectTimeoutException ||
           isWrappedStandbyException(e)) {
         return new RetryAction(
             RetryAction.RetryDecision.FAILOVER_AND_RETRY,

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/ipc/Client.java

@@ -67,6 +67,7 @@ import org.apache.hadoop.ipc.protobuf.RpcHeaderProtos.RpcRequestHeaderProto;
 import org.apache.hadoop.ipc.protobuf.RpcHeaderProtos.RpcRequestHeaderProto.OperationProto;
 import org.apache.hadoop.ipc.protobuf.RpcHeaderProtos.RpcResponseHeaderProto;
 import org.apache.hadoop.ipc.protobuf.RpcHeaderProtos.RpcResponseHeaderProto.RpcStatusProto;
+import org.apache.hadoop.net.ConnectTimeoutException;
 import org.apache.hadoop.net.NetUtils;
 import org.apache.hadoop.security.KerberosInfo;
 import org.apache.hadoop.security.SaslRpcClient;
@@ -511,14 +512,14 @@ public class Client {
           }
           this.socket.setSoTimeout(pingInterval);
           return;
-        } catch (SocketTimeoutException toe) {
+        } catch (ConnectTimeoutException toe) {
           /* Check for an address change and update the local reference.
            * Reset the failure counter if the address was changed
            */
           if (updateAddress()) {
             timeoutFailures = ioFailures = 0;
           }
-          handleConnectionFailure(timeoutFailures++,
+          handleConnectionTimeout(timeoutFailures++,
               maxRetriesOnSocketTimeouts, toe);
         } catch (IOException ie) {
           if (updateAddress()) {
@@ -680,7 +681,7 @@ public class Client {
       socket = null;
     }
 
-    /* Handle connection failures
+    /* Handle connection failures due to timeout on connect
      *
      * If the current number of retries is equal to the max number of retries,
      * stop retrying and throw the exception; Otherwise backoff 1 second and
@@ -694,7 +695,7 @@ public class Client {
      * @param ioe failure reason
      * @throws IOException if max number of retries is reached
      */
-    private void handleConnectionFailure(
+    private void handleConnectionTimeout(
         int curRetries, int maxRetries, IOException ioe) throws IOException {
 
       closeConnection();

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/net/ConnectTimeoutException.java

@@ -0,0 +1,37 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.net;
+
+import java.net.SocketTimeoutException;
+
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceStability;
+
+/**
+ * Thrown by {@link NetUtils#connect(java.net.Socket, java.net.SocketAddress, int)}
+ * if it times out while connecting to the remote host.
+ */
+@InterfaceAudience.Public
+@InterfaceStability.Stable
+public class ConnectTimeoutException extends SocketTimeoutException {
+  private static final long serialVersionUID = 1L;
+
+  public ConnectTimeoutException(String msg) {
+    super(msg);
+  }
+}

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/net/NetUtils.java

@@ -20,6 +20,7 @@ package org.apache.hadoop.net;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
+import java.lang.reflect.Constructor;
 import java.net.BindException;
 import java.net.InetAddress;
 import java.net.InetSocketAddress;
@@ -517,11 +518,15 @@ public class NetUtils {
       socket.bind(localAddr);
     }
 
-    if (ch == null) {
-      // let the default implementation handle it.
-      socket.connect(endpoint, timeout);
-    } else {
-      SocketIOWithTimeout.connect(ch, endpoint, timeout);
+    try {
+      if (ch == null) {
+        // let the default implementation handle it.
+        socket.connect(endpoint, timeout);
+      } else {
+        SocketIOWithTimeout.connect(ch, endpoint, timeout);
+      }
+    } catch (SocketTimeoutException ste) {
+      throw new ConnectTimeoutException(ste.getMessage());
     }
 
     // There is a very rare case allowed by the TCP specification, such that
@@ -719,7 +724,7 @@ public class NetUtils {
               + see("BindException"));
     } else if (exception instanceof ConnectException) {
       // connection refused; include the host:port in the error
-      return (ConnectException) new ConnectException(
+      return wrapWithMessage(exception, 
           "Call From "
               + localHost
               + " to "
@@ -729,32 +734,28 @@ public class NetUtils {
               + " failed on connection exception: "
               + exception
               + ";"
-              + see("ConnectionRefused"))
-          .initCause(exception);
+              + see("ConnectionRefused"));
     } else if (exception instanceof UnknownHostException) {
-      return (UnknownHostException) new UnknownHostException(
+      return wrapWithMessage(exception,
           "Invalid host name: "
               + getHostDetailsAsString(destHost, destPort, localHost)
               + exception
               + ";"
-              + see("UnknownHost"))
-          .initCause(exception);
+              + see("UnknownHost"));
     } else if (exception instanceof SocketTimeoutException) {
-      return (SocketTimeoutException) new SocketTimeoutException(
+      return wrapWithMessage(exception,
           "Call From "
               + localHost + " to " + destHost + ":" + destPort
               + " failed on socket timeout exception: " + exception
               + ";"
-              + see("SocketTimeout"))
-          .initCause(exception);
+              + see("SocketTimeout"));
     } else if (exception instanceof NoRouteToHostException) {
-      return (NoRouteToHostException) new NoRouteToHostException(
+      return wrapWithMessage(exception,
           "No Route to Host from  "
               + localHost + " to " + destHost + ":" + destPort
               + " failed on socket timeout exception: " + exception
               + ";"
-              + see("NoRouteToHost"))
-          .initCause(exception);
+              + see("NoRouteToHost"));
     }
     else {
       return (IOException) new IOException("Failed on local exception: "
@@ -769,6 +770,21 @@ public class NetUtils {
   private static String see(final String entry) {
     return FOR_MORE_DETAILS_SEE + HADOOP_WIKI + entry;
   }
+  
+  @SuppressWarnings("unchecked")
+  private static <T extends IOException> T wrapWithMessage(
+      T exception, String msg) {
+    Class<? extends Throwable> clazz = exception.getClass();
+    try {
+      Constructor<? extends Throwable> ctor = clazz.getConstructor(String.class);
+      Throwable t = ctor.newInstance(msg);
+      return (T)(t.initCause(exception));
+    } catch (Throwable e) {
+      LOG.warn("Unable to wrap exception of type " +
+          clazz + ": it has no (String) constructor", e);
+      return exception;
+    }
+  }
 
   /**
    * Get the host details as a string

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/security/UserGroupInformation.java

@@ -301,17 +301,25 @@ public class UserGroupInformation {
   
   private static String OS_LOGIN_MODULE_NAME;
   private static Class<? extends Principal> OS_PRINCIPAL_CLASS;
+  
   private static final boolean windows =
       System.getProperty("os.name").startsWith("Windows");
   private static final boolean is64Bit =
       System.getProperty("os.arch").contains("64");
+  private static final boolean ibmJava = System.getProperty("java.vendor").contains("IBM");
+  private static final boolean aix = System.getProperty("os.name").equals("AIX");
+
   /* Return the OS login module class name */
   private static String getOSLoginModuleName() {
-    if (System.getProperty("java.vendor").contains("IBM")) {
-      return windows ? (is64Bit
-          ? "com.ibm.security.auth.module.Win64LoginModule"
-          : "com.ibm.security.auth.module.NTLoginModule")
-        : "com.ibm.security.auth.module.LinuxLoginModule";
+    if (ibmJava) {
+      if (windows) {
+        return is64Bit ? "com.ibm.security.auth.module.Win64LoginModule"
+            : "com.ibm.security.auth.module.NTLoginModule";
+      } else if (aix) {
+        return "com.ibm.security.auth.module.AIXLoginModule";
+      } else {
+        return "com.ibm.security.auth.module.LinuxLoginModule";
+      }
     } else {
       return windows ? "com.sun.security.auth.module.NTLoginModule"
         : "com.sun.security.auth.module.UnixLoginModule";
@@ -323,11 +331,14 @@ public class UserGroupInformation {
   private static Class<? extends Principal> getOsPrincipalClass() {
     ClassLoader cl = ClassLoader.getSystemClassLoader();
     try {
-      if (System.getProperty("java.vendor").contains("IBM")) {
+      if (ibmJava) {
         if (windows) {
           return (Class<? extends Principal>) (is64Bit
             ? cl.loadClass("com.ibm.security.auth.UsernamePrincipal")
             : cl.loadClass("com.ibm.security.auth.NTUserPrincipal"));
+        } else if (aix) {
+          return (Class<? extends Principal>)
+             cl.loadClass("com.ibm.security.auth.AIXPrincipal");
         } else {
           return (Class<? extends Principal>) (is64Bit
             ? cl.loadClass("com.ibm.security.auth.UsernamePrincipal")
@@ -418,12 +429,21 @@ public class UserGroupInformation {
     private static final Map<String,String> USER_KERBEROS_OPTIONS = 
       new HashMap<String,String>();
     static {
-      USER_KERBEROS_OPTIONS.put("doNotPrompt", "true");
-      USER_KERBEROS_OPTIONS.put("useTicketCache", "true");
-      USER_KERBEROS_OPTIONS.put("renewTGT", "true");
+      if (ibmJava) {
+        USER_KERBEROS_OPTIONS.put("useDefaultCcache", "true");
+      } else {
+        USER_KERBEROS_OPTIONS.put("doNotPrompt", "true");
+        USER_KERBEROS_OPTIONS.put("useTicketCache", "true");
+        USER_KERBEROS_OPTIONS.put("renewTGT", "true");
+      }
       String ticketCache = System.getenv("KRB5CCNAME");
       if (ticketCache != null) {
-        USER_KERBEROS_OPTIONS.put("ticketCache", ticketCache);
+        if (ibmJava) {
+          // The first value searched when "useDefaultCcache" is used.
+          System.setProperty("KRB5CCNAME", ticketCache);
+        } else {
+          USER_KERBEROS_OPTIONS.put("ticketCache", ticketCache);
+        }
       }
       USER_KERBEROS_OPTIONS.putAll(BASIC_JAAS_OPTIONS);
     }
@@ -434,10 +454,14 @@ public class UserGroupInformation {
     private static final Map<String,String> KEYTAB_KERBEROS_OPTIONS = 
       new HashMap<String,String>();
     static {
-      KEYTAB_KERBEROS_OPTIONS.put("doNotPrompt", "true");
-      KEYTAB_KERBEROS_OPTIONS.put("useKeyTab", "true");
-      KEYTAB_KERBEROS_OPTIONS.put("storeKey", "true");
-      KEYTAB_KERBEROS_OPTIONS.put("refreshKrb5Config", "true");
+      if (ibmJava) {
+        KEYTAB_KERBEROS_OPTIONS.put("credsType", "both");
+      } else {
+        KEYTAB_KERBEROS_OPTIONS.put("doNotPrompt", "true");
+        KEYTAB_KERBEROS_OPTIONS.put("useKeyTab", "true");
+        KEYTAB_KERBEROS_OPTIONS.put("storeKey", "true");
+        KEYTAB_KERBEROS_OPTIONS.put("refreshKrb5Config", "true");
+      }
       KEYTAB_KERBEROS_OPTIONS.putAll(BASIC_JAAS_OPTIONS);      
     }
     private static final AppConfigurationEntry KEYTAB_KERBEROS_LOGIN =
@@ -462,7 +486,12 @@ public class UserGroupInformation {
       } else if (USER_KERBEROS_CONFIG_NAME.equals(appName)) {
         return USER_KERBEROS_CONF;
       } else if (KEYTAB_KERBEROS_CONFIG_NAME.equals(appName)) {
-        KEYTAB_KERBEROS_OPTIONS.put("keyTab", keytabFile);
+        if (ibmJava) {
+          KEYTAB_KERBEROS_OPTIONS.put("useKeytab",
+              prependFileAuthority(keytabFile));
+        } else {
+          KEYTAB_KERBEROS_OPTIONS.put("keyTab", keytabFile);
+        }
         KEYTAB_KERBEROS_OPTIONS.put("principal", keytabPrincipal);
         return KEYTAB_KERBEROS_CONF;
       }
@@ -470,6 +499,11 @@ public class UserGroupInformation {
     }
   }
 
+  private static String prependFileAuthority(String keytabPath) {
+    return keytabPath.startsWith("file://") ? keytabPath
+        : "file://" + keytabPath;
+  }
+
   /**
    * Represents a javax.security configuration that is created at runtime.
    */
@@ -666,6 +700,7 @@ public class UserGroupInformation {
         }
         loginUser.spawnAutoRenewalThreadForUserCreds();
       } catch (LoginException le) {
+        LOG.debug("failure to login", le);
         throw new IOException("failure to login", le);
       }
       if (LOG.isDebugEnabled()) {

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/util/StringUtils.java

@@ -23,8 +23,6 @@ import java.io.StringWriter;
 import java.net.URI;
 import java.net.URISyntaxException;
 import java.text.DateFormat;
-import java.text.DecimalFormat;
-import java.text.NumberFormat;
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Collection;
@@ -37,13 +35,14 @@ import java.util.StringTokenizer;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 
-import com.google.common.net.InetAddresses;
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
 import org.apache.hadoop.fs.Path;
 import org.apache.hadoop.net.NetUtils;
 import org.apache.hadoop.util.Shell;
 
+import com.google.common.net.InetAddresses;
+
 /**
  * General string utils
  */
@@ -77,13 +76,6 @@ public class StringUtils {
   public static final Pattern ENV_VAR_PATTERN = Shell.WINDOWS ?
     WIN_ENV_VAR_PATTERN : SHELL_ENV_VAR_PATTERN;
 
-  private static final DecimalFormat decimalFormat;
-  static {
-          NumberFormat numberFormat = NumberFormat.getNumberInstance(Locale.ENGLISH);
-          decimalFormat = (DecimalFormat) numberFormat;
-          decimalFormat.applyPattern("#.##");
-  }
-
   /**
    * Make a string representation of the exception.
    * @param e The exception to stringify
@@ -112,50 +104,33 @@ public class StringUtils {
     }
     return fullHostname;
   }
-
-  private static DecimalFormat oneDecimal = new DecimalFormat("0.0");
   
   /**
    * Given an integer, return a string that is in an approximate, but human 
    * readable format. 
-   * It uses the bases 'k', 'm', and 'g' for 1024, 1024**2, and 1024**3.
    * @param number the number to format
    * @return a human readable form of the integer
+   *
+   * @deprecated use {@link TraditionalBinaryPrefix#long2String(long, String, int)}.
    */
+  @Deprecated
   public static String humanReadableInt(long number) {
-    long absNumber = Math.abs(number);
-    double result = number;
-    String suffix = "";
-    if (absNumber < 1024) {
-      // since no division has occurred, don't format with a decimal point
-      return String.valueOf(number);
-    } else if (absNumber < 1024 * 1024) {
-      result = number / 1024.0;
-      suffix = "k";
-    } else if (absNumber < 1024 * 1024 * 1024) {
-      result = number / (1024.0 * 1024);
-      suffix = "m";
-    } else {
-      result = number / (1024.0 * 1024 * 1024);
-      suffix = "g";
-    }
-    return oneDecimal.format(result) + suffix;
+    return TraditionalBinaryPrefix.long2String(number, "", 1);
   }
-  
+
+  /** The same as String.format(Locale.ENGLISH, format, objects). */
+  public static String format(final String format, final Object... objects) {
+    return String.format(Locale.ENGLISH, format, objects);
+  }
+
   /**
    * Format a percentage for presentation to the user.
-   * @param done the percentage to format (0.0 to 1.0)
-   * @param digits the number of digits past the decimal point
+   * @param fraction the percentage as a fraction, e.g. 0.1 = 10%
+   * @param decimalPlaces the number of decimal places
    * @return a string representation of the percentage
    */
-  public static String formatPercent(double done, int digits) {
-    DecimalFormat percentFormat = new DecimalFormat("0.00%");
-    double scale = Math.pow(10.0, digits+2);
-    double rounded = Math.floor(done * scale);
-    percentFormat.setDecimalSeparatorAlwaysShown(false);
-    percentFormat.setMinimumFractionDigits(digits);
-    percentFormat.setMaximumFractionDigits(digits);
-    return percentFormat.format(rounded / scale);
+  public static String formatPercent(double fraction, int decimalPlaces) {
+    return format("%." + decimalPlaces + "f%%", fraction*100);
   }
   
   /**
@@ -190,7 +165,7 @@ public class StringUtils {
     }
     StringBuilder s = new StringBuilder(); 
     for(int i = start; i < end; i++) {
-      s.append(String.format("%02x", bytes[i]));
+      s.append(format("%02x", bytes[i]));
     }
     return s.toString();
   }
@@ -655,18 +630,22 @@ public class StringUtils {
    * TraditionalBinaryPrefix symbol are case insensitive. 
    */
   public static enum TraditionalBinaryPrefix {
-    KILO(1024),
-    MEGA(KILO.value << 10),
-    GIGA(MEGA.value << 10),
-    TERA(GIGA.value << 10),
-    PETA(TERA.value << 10),
-    EXA(PETA.value << 10);
+    KILO(10),
+    MEGA(KILO.bitShift + 10),
+    GIGA(MEGA.bitShift + 10),
+    TERA(GIGA.bitShift + 10),
+    PETA(TERA.bitShift + 10),
+    EXA (PETA.bitShift + 10);
 
     public final long value;
     public final char symbol;
+    public final int bitShift;
+    public final long bitMask;
 
-    TraditionalBinaryPrefix(long value) {
-      this.value = value;
+    private TraditionalBinaryPrefix(int bitShift) {
+      this.bitShift = bitShift;
+      this.value = 1L << bitShift;
+      this.bitMask = this.value - 1L;
       this.symbol = toString().charAt(0);
     }
 
@@ -717,8 +696,58 @@ public class StringUtils {
         return num * prefix;
       }
     }
+
+    /**
+     * Convert a long integer to a string with traditional binary prefix.
+     * 
+     * @param n the value to be converted
+     * @param unit The unit, e.g. "B" for bytes.
+     * @param decimalPlaces The number of decimal places.
+     * @return a string with traditional binary prefix.
+     */
+    public static String long2String(long n, String unit, int decimalPlaces) {
+      if (unit == null) {
+        unit = "";
+      }
+      //take care a special case
+      if (n == Long.MIN_VALUE) {
+        return "-8 " + EXA.symbol + unit;
+      }
+
+      final StringBuilder b = new StringBuilder();
+      //take care negative numbers
+      if (n < 0) {
+        b.append('-');
+        n = -n;
+      }
+      if (n < KILO.value) {
+        //no prefix
+        b.append(n);
+        return (unit.isEmpty()? b: b.append(" ").append(unit)).toString();
+      } else {
+        //find traditional binary prefix
+        int i = 0;
+        for(; i < values().length && n >= values()[i].value; i++);
+        TraditionalBinaryPrefix prefix = values()[i - 1];
+
+        if ((n & prefix.bitMask) == 0) {
+          //exact division
+          b.append(n >> prefix.bitShift);
+        } else {
+          final String  format = "%." + decimalPlaces + "f";
+          String s = format(format, n/(double)prefix.value);
+          //check a special rounding up case
+          if (s.startsWith("1024")) {
+            prefix = values()[i];
+            s = format(format, n/(double)prefix.value);
+          }
+          b.append(s);
+        }
+        return b.append(' ').append(prefix.symbol).append(unit).toString();
+      }
+    }
   }
-  
+
     /**
      * Escapes HTML Special characters present in the string.
      * @param string
@@ -756,32 +785,16 @@ public class StringUtils {
     }
 
   /**
-   * Return an abbreviated English-language desc of the byte length
+   * @return a byte description of the given long interger value.
    */
   public static String byteDesc(long len) {
-    double val = 0.0;
-    String ending = "";
-    if (len < 1024 * 1024) {
-      val = (1.0 * len) / 1024;
-      ending = " KB";
-    } else if (len < 1024 * 1024 * 1024) {
-      val = (1.0 * len) / (1024 * 1024);
-      ending = " MB";
-    } else if (len < 1024L * 1024 * 1024 * 1024) {
-      val = (1.0 * len) / (1024 * 1024 * 1024);
-      ending = " GB";
-    } else if (len < 1024L * 1024 * 1024 * 1024 * 1024) {
-      val = (1.0 * len) / (1024L * 1024 * 1024 * 1024);
-      ending = " TB";
-    } else {
-      val = (1.0 * len) / (1024L * 1024 * 1024 * 1024 * 1024);
-      ending = " PB";
-    }
-    return limitDecimalTo2(val) + ending;
-  }
-
-  public static synchronized String limitDecimalTo2(double d) {
-    return decimalFormat.format(d);
+    return TraditionalBinaryPrefix.long2String(len, "B", 2);
+  }
+
+  /** @deprecated use StringUtils.format("%.2f", d). */
+  @Deprecated
+  public static String limitDecimalTo2(double d) {
+    return format("%.2f", d);
   }
   
   /**

hadoop-common-project/hadoop-common/src/site/apt/ClusterSetup.apt.vm

@@ -325,6 +325,13 @@ Hadoop MapReduce Next Generation - Cluster Setup
 | | | How long to keep aggregation logs before deleting them. -1 disables. |
 | | | Be careful, set this too small and you will spam the name node. |
 *-------------------------+-------------------------+------------------------+
+| <<<yarn.log-aggregation.retain-check-interval-seconds>>> | | |
+| | <-1> | |
+| | | Time between checks for aggregated log retention. If set to 0 or a |
+| | | negative value then the value is computed as one-tenth of the |
+| | | aggregated log retention time. |
+| | | Be careful, set this too small and you will spam the name node. |
+*-------------------------+-------------------------+------------------------+
 
 
 

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/TestFsShellReturnCode.java

@@ -310,6 +310,46 @@ public class TestFsShellReturnCode {
     }
   }
   
+  @Test
+  public void testRmWithNonexistentGlob() throws Exception {
+    Configuration conf = new Configuration();
+    FsShell shell = new FsShell();
+    shell.setConf(conf);
+    final ByteArrayOutputStream bytes = new ByteArrayOutputStream();
+    final PrintStream err = new PrintStream(bytes);
+    final PrintStream oldErr = System.err;
+    System.setErr(err);
+    final String results;
+    try {
+      int exit = shell.run(new String[]{"-rm", "nomatch*"});
+      assertEquals(1, exit);
+      results = bytes.toString();
+      assertTrue(results.contains("rm: `nomatch*': No such file or directory"));
+    } finally {
+      IOUtils.closeStream(err);
+      System.setErr(oldErr);
+    }
+  }
+
+  @Test
+  public void testRmForceWithNonexistentGlob() throws Exception {
+    Configuration conf = new Configuration();
+    FsShell shell = new FsShell();
+    shell.setConf(conf);
+    final ByteArrayOutputStream bytes = new ByteArrayOutputStream();
+    final PrintStream err = new PrintStream(bytes);
+    final PrintStream oldErr = System.err;
+    System.setErr(err);
+    try {
+      int exit = shell.run(new String[]{"-rm", "-f", "nomatch*"});
+      assertEquals(0, exit);
+      assertTrue(bytes.toString().isEmpty());
+    } finally {
+      IOUtils.closeStream(err);
+      System.setErr(oldErr);
+    }
+  }
+
   @Test
   public void testInvalidDefaultFS() throws Exception {
     // if default fs doesn't exist or is invalid, but the path provided in 

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/TestListFiles.java

@@ -45,19 +45,39 @@ public class TestListFiles {
 
   final protected static Configuration conf = new Configuration();
   protected static FileSystem fs;
-  final protected static Path TEST_DIR = getTestDir();
+  protected static Path TEST_DIR;
   final private static int FILE_LEN = 10;
-  final private static Path FILE1 = new Path(TEST_DIR, "file1");
-  final private static Path DIR1 = new Path(TEST_DIR, "dir1");
-  final private static Path FILE2 = new Path(DIR1, "file2");
-  final private static Path FILE3 = new Path(DIR1, "file3");
+  private static Path FILE1;
+  private static Path DIR1;
+  private static Path FILE2;
+  private static Path FILE3;
+
+  static {
+    setTestPaths(new Path(
+      System.getProperty("test.build.data", "build/test/data/work-dir/localfs"),
+      "main_"));
+  }
 
   protected static Path getTestDir() {
-    return new Path(
-      System.getProperty("test.build.data","build/test/data/work-dir/localfs"),
-      "main_");
+    return TEST_DIR;
   }
-  
+
+  /**
+   * Sets the root testing directory and reinitializes any additional test paths
+   * that are under the root.  This method is intended to be called from a
+   * subclass's @BeforeClass method if there is a need to override the testing
+   * directory.
+   * 
+   * @param testDir Path root testing directory
+   */
+  protected static void setTestPaths(Path testDir) {
+    TEST_DIR = testDir;
+    FILE1 = new Path(TEST_DIR, "file1");
+    DIR1 = new Path(TEST_DIR, "dir1");
+    FILE2 = new Path(DIR1, "file2");
+    FILE3 = new Path(DIR1, "file3");
+  }
+
   @BeforeClass
   public static void testSetUp() throws Exception {
     fs = FileSystem.getLocal(conf);

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/fs/TestLocalFileSystem.java

@@ -28,6 +28,7 @@ import java.io.*;
 import static org.junit.Assert.*;
 import static org.junit.Assume.assumeTrue;
 
+import org.junit.After;
 import org.junit.Before;
 import org.junit.Test;
 
@@ -38,8 +39,9 @@ public class TestLocalFileSystem {
   private static final String TEST_ROOT_DIR
     = System.getProperty("test.build.data","build/test/data") + "/work-dir/localfs";
 
+  private final File base = new File(TEST_ROOT_DIR);
   private Configuration conf;
-  private FileSystem fileSys;
+  private LocalFileSystem fileSys;
 
   private void cleanupFile(FileSystem fs, Path name) throws IOException {
     assertTrue(fs.exists(name));
@@ -53,6 +55,13 @@ public class TestLocalFileSystem {
     fileSys = FileSystem.getLocal(conf);
     fileSys.delete(new Path(TEST_ROOT_DIR), true);
   }
+  
+  @After
+  public void after() throws IOException {
+    base.setWritable(true);
+    FileUtil.fullyDelete(base);
+    assertTrue(!base.exists());
+  }
 
   /**
    * Test the capability of setting the working directory.
@@ -269,10 +278,83 @@ public class TestLocalFileSystem {
     LocalFileSystem fs = FileSystem.getLocal(conf);
     File colonFile = new File(TEST_ROOT_DIR, "foo:bar");
     colonFile.mkdirs();
-    colonFile.createNewFile();
     FileStatus[] stats = fs.listStatus(new Path(TEST_ROOT_DIR));
     assertEquals("Unexpected number of stats", 1, stats.length);
     assertEquals("Bad path from stat", colonFile.getAbsolutePath(),
         stats[0].getPath().toUri().getPath());
   }
+  
+  @Test
+  public void testReportChecksumFailure() throws IOException {
+    base.mkdirs();
+    assertTrue(base.exists() && base.isDirectory());
+    
+    final File dir1 = new File(base, "dir1");
+    final File dir2 = new File(dir1, "dir2");
+    dir2.mkdirs();
+    assertTrue(dir2.exists() && dir2.canWrite());
+    
+    final String dataFileName = "corruptedData";
+    final Path dataPath = new Path(new File(dir2, dataFileName).toURI());
+    final Path checksumPath = fileSys.getChecksumFile(dataPath);
+    final FSDataOutputStream fsdos = fileSys.create(dataPath);
+    try {
+      fsdos.writeUTF("foo");
+    } finally {
+      fsdos.close();
+    }
+    assertTrue(fileSys.pathToFile(dataPath).exists());
+    final long dataFileLength = fileSys.getFileStatus(dataPath).getLen();
+    assertTrue(dataFileLength > 0);
+    
+    // check the the checksum file is created and not empty:
+    assertTrue(fileSys.pathToFile(checksumPath).exists());
+    final long checksumFileLength = fileSys.getFileStatus(checksumPath).getLen();
+    assertTrue(checksumFileLength > 0);
+    
+    // this is a hack to force the #reportChecksumFailure() method to stop
+    // climbing up at the 'base' directory and use 'dir1/bad_files' as the 
+    // corrupted files storage:
+    base.setWritable(false);
+    
+    FSDataInputStream dataFsdis = fileSys.open(dataPath);
+    FSDataInputStream checksumFsdis = fileSys.open(checksumPath);
+    
+    boolean retryIsNecessary = fileSys.reportChecksumFailure(dataPath, dataFsdis, 0, checksumFsdis, 0);
+    assertTrue(!retryIsNecessary);
+    
+    // the data file should be moved:
+    assertTrue(!fileSys.pathToFile(dataPath).exists());
+    // the checksum file should be moved:
+    assertTrue(!fileSys.pathToFile(checksumPath).exists());
+    
+    // check that the files exist in the new location where they were moved:
+    File[] dir1files = dir1.listFiles(new FileFilter() {
+      @Override
+      public boolean accept(File pathname) {
+        return pathname != null && !pathname.getName().equals("dir2");
+      }
+    });
+    assertTrue(dir1files != null);
+    assertTrue(dir1files.length == 1);
+    File badFilesDir = dir1files[0];
+    
+    File[] badFiles = badFilesDir.listFiles();
+    assertTrue(badFiles != null);
+    assertTrue(badFiles.length == 2);
+    boolean dataFileFound = false;
+    boolean checksumFileFound = false;
+    for (File badFile: badFiles) {
+      if (badFile.getName().startsWith(dataFileName)) {
+        assertTrue(dataFileLength == badFile.length());
+        dataFileFound = true;
+      } else if (badFile.getName().contains(dataFileName + ".crc")) {
+        assertTrue(checksumFileLength == badFile.length());
+        checksumFileFound = true;
+      }
+    }
+    assertTrue(dataFileFound);
+    assertTrue(checksumFileFound);
+  }
+  
 }

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/io/TestBoundedByteArrayOutputStream.java

@@ -88,4 +88,61 @@ public class TestBoundedByteArrayOutputStream extends TestCase {
     assertTrue("Writing beyond limit did not throw an exception",
         caughtException);
   }
+  
+  
+  static class ResettableBoundedByteArrayOutputStream 
+  extends BoundedByteArrayOutputStream {
+
+    public ResettableBoundedByteArrayOutputStream(int capacity) {
+      super(capacity);
+    }
+
+    public void resetBuffer(byte[] buf, int offset, int length) {
+      super.resetBuffer(buf, offset, length);
+    }
+    
+  }
+  
+  public void testResetBuffer() throws IOException {
+    
+    ResettableBoundedByteArrayOutputStream stream = 
+      new ResettableBoundedByteArrayOutputStream(SIZE);
+
+    // Write to the stream, get the data back and check for contents
+    stream.write(INPUT, 0, SIZE);
+    assertTrue("Array Contents Mismatch",
+        Arrays.equals(INPUT, stream.getBuffer()));
+    
+    // Try writing beyond end of buffer. Should throw an exception
+    boolean caughtException = false;
+    
+    try {
+      stream.write(INPUT[0]);
+    } catch (Exception e) {
+      caughtException = true;
+    }
+    
+    assertTrue("Writing beyond limit did not throw an exception",
+        caughtException);
+    
+    //Reset the stream and try, should succeed
+    byte[] newBuf = new byte[SIZE];
+    stream.resetBuffer(newBuf, 0, newBuf.length);
+    assertTrue("Limit did not get reset correctly", 
+        (stream.getLimit() == SIZE));
+    stream.write(INPUT, 0, SIZE);
+    assertTrue("Array Contents Mismatch",
+        Arrays.equals(INPUT, stream.getBuffer()));
+  
+    // Try writing one more byte, should fail
+    caughtException = false;
+    try {
+      stream.write(INPUT[0]);
+    } catch (Exception e) {
+      caughtException = true;
+    }
+    assertTrue("Writing beyond limit did not throw an exception",
+        caughtException);
+  }
+
 }

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/io/TestSortedMapWritable.java

@@ -17,15 +17,20 @@
  */
 package org.apache.hadoop.io;
 
-import java.util.Map;
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertFalse;
+import static org.junit.Assert.assertNotNull;
+import static org.junit.Assert.assertTrue;
 
-import junit.framework.TestCase;
+import java.util.Map;
+import org.junit.Test;
 
 /**
  * Tests SortedMapWritable
  */
-public class TestSortedMapWritable extends TestCase {
+public class TestSortedMapWritable {
   /** the test */
+  @Test
   @SuppressWarnings("unchecked")
   public void testSortedMapWritable() {
     Text[] keys = {
@@ -90,6 +95,7 @@ public class TestSortedMapWritable extends TestCase {
   /**
    * Test that number of "unknown" classes is propagated across multiple copies.
    */
+  @Test
   @SuppressWarnings("deprecation")
   public void testForeignClass() {
     SortedMapWritable inMap = new SortedMapWritable();
@@ -99,4 +105,63 @@ public class TestSortedMapWritable extends TestCase {
     SortedMapWritable copyOfCopy = new SortedMapWritable(outMap);
     assertEquals(1, copyOfCopy.getNewClasses());
   }
+  
+  /**
+   * Tests if equal and hashCode method still hold the contract.
+   */
+  @Test
+  public void testEqualsAndHashCode() {
+    String failureReason;
+    SortedMapWritable mapA = new SortedMapWritable();
+    SortedMapWritable mapB = new SortedMapWritable();
+    
+    // Sanity checks
+    failureReason = "SortedMapWritable couldn't be initialized. Got null reference";
+    assertNotNull(failureReason, mapA);
+    assertNotNull(failureReason, mapB);
+    
+    // Basic null check
+    assertFalse("equals method returns true when passed null", mapA.equals(null));
+    
+    // When entry set is empty, they should be equal
+    assertTrue("Two empty SortedMapWritables are no longer equal", mapA.equals(mapB));
+    
+    // Setup
+    Text[] keys = {
+        new Text("key1"),
+        new Text("key2")
+    };
+    
+    BytesWritable[] values = {
+        new BytesWritable("value1".getBytes()),
+        new BytesWritable("value2".getBytes())
+    };
+    
+    mapA.put(keys[0], values[0]);
+    mapB.put(keys[1], values[1]);
+    
+    // entrySets are different
+    failureReason = "Two SortedMapWritables with different data are now equal";
+    assertTrue(failureReason, mapA.hashCode() != mapB.hashCode());
+    assertTrue(failureReason, !mapA.equals(mapB));
+    assertTrue(failureReason, !mapB.equals(mapA));
+    
+    mapA.put(keys[1], values[1]);
+    mapB.put(keys[0], values[0]);
+    
+    // entrySets are now same
+    failureReason = "Two SortedMapWritables with same entry sets formed in different order are now different";
+    assertEquals(failureReason, mapA.hashCode(), mapB.hashCode());
+    assertTrue(failureReason, mapA.equals(mapB));
+    assertTrue(failureReason, mapB.equals(mapA));
+    
+    // Let's check if entry sets of same keys but different values
+    mapA.put(keys[0], values[1]);
+    mapA.put(keys[1], values[0]);
+    
+    failureReason = "Two SortedMapWritables with different content are now equal";
+    assertTrue(failureReason, mapA.hashCode() != mapB.hashCode());
+    assertTrue(failureReason, !mapA.equals(mapB));
+    assertTrue(failureReason, !mapB.equals(mapA));
+  }
 }

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/ipc/TestIPC.java

@@ -26,6 +26,7 @@ import org.apache.hadoop.io.IntWritable;
 import org.apache.hadoop.io.Writable;
 import org.apache.hadoop.io.LongWritable;
 import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.net.ConnectTimeoutException;
 import org.apache.hadoop.net.NetUtils;
 
 import java.util.Random;
@@ -586,7 +587,7 @@ public class TestIPC {
   private void assertRetriesOnSocketTimeouts(Configuration conf,
       int maxTimeoutRetries) throws IOException, InterruptedException {
     SocketFactory mockFactory = Mockito.mock(SocketFactory.class);
-    doThrow(new SocketTimeoutException()).when(mockFactory).createSocket();
+    doThrow(new ConnectTimeoutException("fake")).when(mockFactory).createSocket();
     Client client = new Client(IntWritable.class, conf, mockFactory);
     InetSocketAddress address = new InetSocketAddress("127.0.0.1", 9090);
     try {

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/util/TestStringUtils.java

@@ -18,6 +18,8 @@
 
 package org.apache.hadoop.util;
 
+import static org.apache.hadoop.util.StringUtils.TraditionalBinaryPrefix.long2String;
+import static org.apache.hadoop.util.StringUtils.TraditionalBinaryPrefix.string2long;
 import static org.junit.Assert.assertArrayEquals;
 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.fail;
@@ -29,6 +31,7 @@ import java.util.Map;
 import java.util.regex.Pattern;
 
 import org.apache.hadoop.test.UnitTestcaseTimeLimit;
+import org.apache.hadoop.util.StringUtils.TraditionalBinaryPrefix;
 import org.junit.Test;
 
 public class TestStringUtils extends UnitTestcaseTimeLimit {
@@ -137,45 +140,34 @@ public class TestStringUtils extends UnitTestcaseTimeLimit {
   
   @Test
   public void testTraditionalBinaryPrefix() throws Exception {
+    //test string2long(..)
     String[] symbol = {"k", "m", "g", "t", "p", "e"};
     long m = 1024;
     for(String s : symbol) {
-      assertEquals(0, StringUtils.TraditionalBinaryPrefix.string2long(0 + s));
-      assertEquals(m, StringUtils.TraditionalBinaryPrefix.string2long(1 + s));
+      assertEquals(0, string2long(0 + s));
+      assertEquals(m, string2long(1 + s));
       m *= 1024;
     }
     
-    assertEquals(0L, StringUtils.TraditionalBinaryPrefix.string2long("0"));
-    assertEquals(1024L, StringUtils.TraditionalBinaryPrefix.string2long("1k"));
-    assertEquals(-1024L, StringUtils.TraditionalBinaryPrefix.string2long("-1k"));
-    assertEquals(1259520L,
-        StringUtils.TraditionalBinaryPrefix.string2long("1230K"));
-    assertEquals(-1259520L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-1230K"));
-    assertEquals(104857600L,
-        StringUtils.TraditionalBinaryPrefix.string2long("100m"));
-    assertEquals(-104857600L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-100M"));
-    assertEquals(956703965184L,
-        StringUtils.TraditionalBinaryPrefix.string2long("891g"));
-    assertEquals(-956703965184L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-891G"));
-    assertEquals(501377302265856L,
-        StringUtils.TraditionalBinaryPrefix.string2long("456t"));
-    assertEquals(-501377302265856L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-456T"));
-    assertEquals(11258999068426240L,
-        StringUtils.TraditionalBinaryPrefix.string2long("10p"));
-    assertEquals(-11258999068426240L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-10P"));
-    assertEquals(1152921504606846976L,
-        StringUtils.TraditionalBinaryPrefix.string2long("1e"));
-    assertEquals(-1152921504606846976L,
-        StringUtils.TraditionalBinaryPrefix.string2long("-1E"));
+    assertEquals(0L, string2long("0"));
+    assertEquals(1024L, string2long("1k"));
+    assertEquals(-1024L, string2long("-1k"));
+    assertEquals(1259520L, string2long("1230K"));
+    assertEquals(-1259520L, string2long("-1230K"));
+    assertEquals(104857600L, string2long("100m"));
+    assertEquals(-104857600L, string2long("-100M"));
+    assertEquals(956703965184L, string2long("891g"));
+    assertEquals(-956703965184L, string2long("-891G"));
+    assertEquals(501377302265856L, string2long("456t"));
+    assertEquals(-501377302265856L, string2long("-456T"));
+    assertEquals(11258999068426240L, string2long("10p"));
+    assertEquals(-11258999068426240L, string2long("-10P"));
+    assertEquals(1152921504606846976L, string2long("1e"));
+    assertEquals(-1152921504606846976L, string2long("-1E"));
 
     String tooLargeNumStr = "10e";
     try {
-      StringUtils.TraditionalBinaryPrefix.string2long(tooLargeNumStr);
+      string2long(tooLargeNumStr);
       fail("Test passed for a number " + tooLargeNumStr + " too large");
     } catch (IllegalArgumentException e) {
       assertEquals(tooLargeNumStr + " does not fit in a Long", e.getMessage());
@@ -183,7 +175,7 @@ public class TestStringUtils extends UnitTestcaseTimeLimit {
 
     String tooSmallNumStr = "-10e";
     try {
-      StringUtils.TraditionalBinaryPrefix.string2long(tooSmallNumStr);
+      string2long(tooSmallNumStr);
       fail("Test passed for a number " + tooSmallNumStr + " too small");
     } catch (IllegalArgumentException e) {
       assertEquals(tooSmallNumStr + " does not fit in a Long", e.getMessage());
@@ -192,7 +184,7 @@ public class TestStringUtils extends UnitTestcaseTimeLimit {
     String invalidFormatNumStr = "10kb";
     char invalidPrefix = 'b';
     try {
-      StringUtils.TraditionalBinaryPrefix.string2long(invalidFormatNumStr);
+      string2long(invalidFormatNumStr);
       fail("Test passed for a number " + invalidFormatNumStr
           + " has invalid format");
     } catch (IllegalArgumentException e) {
@@ -202,6 +194,74 @@ public class TestStringUtils extends UnitTestcaseTimeLimit {
           e.getMessage());
     }
 
+    //test long2string(..)
+    assertEquals("0", long2String(0, null, 2));
+    for(int decimalPlace = 0; decimalPlace < 2; decimalPlace++) {
+      for(int n = 1; n < TraditionalBinaryPrefix.KILO.value; n++) {
+        assertEquals(n + "", long2String(n, null, decimalPlace));
+        assertEquals(-n + "", long2String(-n, null, decimalPlace));
+      }
+      assertEquals("1 K", long2String(1L << 10, null, decimalPlace));
+      assertEquals("-1 K", long2String(-1L << 10, null, decimalPlace));
+    }
+
+    assertEquals("8.00 E", long2String(Long.MAX_VALUE, null, 2));
+    assertEquals("8.00 E", long2String(Long.MAX_VALUE - 1, null, 2));
+    assertEquals("-8 E", long2String(Long.MIN_VALUE, null, 2));
+    assertEquals("-8.00 E", long2String(Long.MIN_VALUE + 1, null, 2));
+
+    final String[] zeros = {" ", ".0 ", ".00 "};
+    for(int decimalPlace = 0; decimalPlace < zeros.length; decimalPlace++) {
+      final String trailingZeros = zeros[decimalPlace]; 
+
+      for(int e = 11; e < Long.SIZE - 1; e++) {
+        final TraditionalBinaryPrefix p
+            = TraditionalBinaryPrefix.values()[e/10 - 1]; 
+  
+        { // n = 2^e
+          final long n = 1L << e;
+          final String expected = (n/p.value) + " " + p.symbol;
+          assertEquals("n=" + n, expected, long2String(n, null, 2));
+        }
+  
+        { // n = 2^e + 1
+          final long n = (1L << e) + 1;
+          final String expected = (n/p.value) + trailingZeros + p.symbol;
+          assertEquals("n=" + n, expected, long2String(n, null, decimalPlace));
+        }
+  
+        { // n = 2^e - 1
+          final long n = (1L << e) - 1;
+          final String expected = ((n+1)/p.value) + trailingZeros + p.symbol;
+          assertEquals("n=" + n, expected, long2String(n, null, decimalPlace));
+        }
+      }
+    }
+
+    assertEquals("1.50 K", long2String(3L << 9, null, 2));
+    assertEquals("1.5 K", long2String(3L << 9, null, 1));
+    assertEquals("1.50 M", long2String(3L << 19, null, 2));
+    assertEquals("2 M", long2String(3L << 19, null, 0));
+    assertEquals("3 G", long2String(3L << 30, null, 2));
+
+    // test byteDesc(..)
+    assertEquals("0 B", StringUtils.byteDesc(0));
+    assertEquals("-100 B", StringUtils.byteDesc(-100));
+    assertEquals("1 KB", StringUtils.byteDesc(1024));
+    assertEquals("1.50 KB", StringUtils.byteDesc(3L << 9));
+    assertEquals("1.50 MB", StringUtils.byteDesc(3L << 19));
+    assertEquals("3 GB", StringUtils.byteDesc(3L << 30));
+    
+    // test formatPercent(..)
+    assertEquals("10%", StringUtils.formatPercent(0.1, 0));
+    assertEquals("10.0%", StringUtils.formatPercent(0.1, 1));
+    assertEquals("10.00%", StringUtils.formatPercent(0.1, 2));
+
+    assertEquals("1%", StringUtils.formatPercent(0.00543, 0));
+    assertEquals("0.5%", StringUtils.formatPercent(0.00543, 1));
+    assertEquals("0.54%", StringUtils.formatPercent(0.00543, 2));
+    assertEquals("0.543%", StringUtils.formatPercent(0.00543, 3));
+    assertEquals("0.5430%", StringUtils.formatPercent(0.00543, 4));
   }
 
   @Test
@@ -360,10 +420,9 @@ public class TestStringUtils extends UnitTestcaseTimeLimit {
         }
         long et = System.nanoTime();
         if (outer > 3) {
-          System.out.println(
-            (useOurs ? "StringUtils impl" : "Java impl") +
-            " #" + outer + ":" +
-            (et - st)/1000000 + "ms");
+          System.out.println( (useOurs ? "StringUtils impl" : "Java impl")
+              + " #" + outer + ":" + (et - st)/1000000 + "ms, components="
+              + components  );
         }
       }
     }

hadoop-hdfs-project/hadoop-hdfs-httpfs/src/main/java/org/apache/hadoop/fs/http/client/HttpFSFileSystem.java

@@ -17,6 +17,8 @@
  */
 package org.apache.hadoop.fs.http.client;
 
+import java.util.ArrayList;
+import java.util.List;
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.ContentSummary;
@@ -86,6 +88,7 @@ public class HttpFSFileSystem extends FileSystem
   public static final String PERMISSION_PARAM = "permission";
   public static final String DESTINATION_PARAM = "destination";
   public static final String RECURSIVE_PARAM = "recursive";
+  public static final String SOURCES_PARAM = "sources";
   public static final String OWNER_PARAM = "owner";
   public static final String GROUP_PARAM = "group";
   public static final String MODIFICATION_TIME_PARAM = "modificationtime";
@@ -167,7 +170,7 @@ public class HttpFSFileSystem extends FileSystem
     GETHOMEDIRECTORY(HTTP_GET), GETCONTENTSUMMARY(HTTP_GET),
     GETFILECHECKSUM(HTTP_GET),  GETFILEBLOCKLOCATIONS(HTTP_GET),
     INSTRUMENTATION(HTTP_GET),
-    APPEND(HTTP_POST),
+    APPEND(HTTP_POST), CONCAT(HTTP_POST),
     CREATE(HTTP_PUT), MKDIRS(HTTP_PUT), RENAME(HTTP_PUT), SETOWNER(HTTP_PUT),
     SETPERMISSION(HTTP_PUT), SETREPLICATION(HTTP_PUT), SETTIMES(HTTP_PUT),
     DELETE(HTTP_DELETE);
@@ -528,6 +531,29 @@ public class HttpFSFileSystem extends FileSystem
                       HttpURLConnection.HTTP_OK);
   }
 
+  /**
+   * Concat existing files together.
+   * @param f the path to the target destination.
+   * @param psrcs the paths to the sources to use for the concatenation.
+   *
+   * @throws IOException
+   */
+  @Override
+  public void concat(Path f, Path[] psrcs) throws IOException {
+    List<String> strPaths = new ArrayList<String>(psrcs.length);
+    for(Path psrc : psrcs) {
+      strPaths.add(psrc.toUri().getPath());
+    }
+    String srcs = StringUtils.join(",", strPaths);
+
+    Map<String, String> params = new HashMap<String, String>();
+    params.put(OP_PARAM, Operation.CONCAT.toString());
+    params.put(SOURCES_PARAM, srcs);
+    HttpURLConnection conn = getConnection(Operation.CONCAT.getMethod(),
+        params, f, true);
+    HttpFSUtils.validateResponse(conn, HttpURLConnection.HTTP_OK);
+  }
+
   /**
    * Renames Path src to Path dst.  Can take place on local fs
    * or remote DFS.

hadoop-hdfs-project/hadoop-hdfs-httpfs/src/main/java/org/apache/hadoop/fs/http/server/FSOperations.java

@@ -198,6 +198,47 @@ public class FSOperations {
 
   }
 
+  /**
+   * Executor that performs an append FileSystemAccess files system operation.
+   */
+  @InterfaceAudience.Private
+  public static class FSConcat implements FileSystemAccess.FileSystemExecutor<Void> {
+    private Path path;
+    private Path[] sources;
+
+    /**
+     * Creates a Concat executor.
+     *
+     * @param path target path to concat to.
+     * @param sources comma seperated absolute paths to use as sources.
+     */
+    public FSConcat(String path, String[] sources) {
+      this.sources = new Path[sources.length];
+
+      for(int i = 0; i < sources.length; i++) {
+        this.sources[i] = new Path(sources[i]);
+      }
+
+      this.path = new Path(path);
+    }
+
+    /**
+     * Executes the filesystem operation.
+     *
+     * @param fs filesystem instance to use.
+     *
+     * @return void.
+     *
+     * @throws IOException thrown if an IO error occured.
+     */
+    @Override
+    public Void execute(FileSystem fs) throws IOException {
+      fs.concat(path, sources);
+      return null;
+    }
+
+  }
+
   /**
    * Executor that performs a content-summary FileSystemAccess files system operation.
    */

hadoop-hdfs-project/hadoop-hdfs-httpfs/src/main/java/org/apache/hadoop/fs/http/server/HttpFSParametersProvider.java

@@ -58,6 +58,7 @@ public class HttpFSParametersProvider extends ParametersProvider {
     PARAMS_DEF.put(Operation.INSTRUMENTATION, new Class[]{DoAsParam.class});
     PARAMS_DEF.put(Operation.APPEND,
       new Class[]{DoAsParam.class, DataParam.class});
+    PARAMS_DEF.put(Operation.CONCAT, new Class[]{SourcesParam.class});
     PARAMS_DEF.put(Operation.CREATE,
       new Class[]{DoAsParam.class, PermissionParam.class, OverwriteParam.class,
                   ReplicationParam.class, BlockSizeParam.class, DataParam.class});
@@ -388,6 +389,25 @@ public class HttpFSParametersProvider extends ParametersProvider {
     }
   }
 
+  /**
+   * Class for concat sources parameter.
+   */
+  @InterfaceAudience.Private
+  public static class SourcesParam extends StringParam {
+
+    /**
+     * Parameter name.
+     */
+    public static final String NAME = HttpFSFileSystem.SOURCES_PARAM;
+
+    /**
+     * Constructor.
+     */
+    public SourcesParam() {
+      super(NAME, null);
+    }
+  }
+
   /**
    * Class for to-path parameter.
    */

hadoop-hdfs-project/hadoop-hdfs-httpfs/src/main/java/org/apache/hadoop/fs/http/server/HttpFSServer.java

@@ -22,22 +22,23 @@ import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.http.client.HttpFSFileSystem;
-import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.OperationParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.AccessTimeParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.BlockSizeParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.DataParam;
-import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.RecursiveParam;
+import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.DestinationParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.DoAsParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.FilterParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.GroupParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.LenParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.ModifiedTimeParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.OffsetParam;
+import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.OperationParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.OverwriteParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.OwnerParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.PermissionParam;
+import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.RecursiveParam;
 import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.ReplicationParam;
-import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.DestinationParam;
+import org.apache.hadoop.fs.http.server.HttpFSParametersProvider.SourcesParam;
 import org.apache.hadoop.lib.service.FileSystemAccess;
 import org.apache.hadoop.lib.service.FileSystemAccessException;
 import org.apache.hadoop.lib.service.Groups;
@@ -403,9 +404,9 @@ public class HttpFSServer {
     Response response;
     path = makeAbsolute(path);
     MDC.put(HttpFSFileSystem.OP_PARAM, op.value().name());
-    String doAs = params.get(DoAsParam.NAME, DoAsParam.class);
     switch (op.value()) {
       case APPEND: {
+        String doAs = params.get(DoAsParam.NAME, DoAsParam.class);
         Boolean hasData = params.get(DataParam.NAME, DataParam.class);
         if (!hasData) {
           response = Response.temporaryRedirect(
@@ -420,6 +421,18 @@ public class HttpFSServer {
         }
         break;
       }
+      case CONCAT: {
+        System.out.println("HTTPFS SERVER CONCAT");
+        String sources = params.get(SourcesParam.NAME, SourcesParam.class);
+
+        FSOperations.FSConcat command =
+            new FSOperations.FSConcat(path, sources.split(","));
+        fsExecute(user, null, command);
+        AUDIT_LOG.info("[{}]", path);
+        System.out.println("SENT RESPONSE");
+        response = Response.ok().build();
+        break;
+      }
       default: {
         throw new IOException(
           MessageFormat.format("Invalid HTTP POST operation [{0}]",

hadoop-hdfs-project/hadoop-hdfs-httpfs/src/test/java/org/apache/hadoop/fs/http/client/BaseTestHttpFSWith.java

@@ -28,6 +28,8 @@ import org.apache.hadoop.fs.Path;
 import org.apache.hadoop.fs.http.server.HttpFSServerWebApp;
 import org.apache.hadoop.fs.permission.FsAction;
 import org.apache.hadoop.fs.permission.FsPermission;
+import org.apache.hadoop.hdfs.DFSConfigKeys;
+import org.apache.hadoop.hdfs.DFSTestUtil;
 import org.apache.hadoop.security.UserGroupInformation;
 import org.apache.hadoop.test.HFSTestCase;
 import org.apache.hadoop.test.HadoopUsersConfTestHelper;
@@ -206,6 +208,30 @@ public abstract class BaseTestHttpFSWith extends HFSTestCase {
     }
   }
 
+  private void testConcat() throws Exception {
+    Configuration config = getProxiedFSConf();
+    config.setLong(DFSConfigKeys.DFS_BLOCK_SIZE_KEY, 1024);
+    if (!isLocalFS()) {
+      FileSystem fs = FileSystem.get(config);
+      fs.mkdirs(getProxiedFSTestDir());
+      Path path1 = new Path("/test/foo.txt");
+      Path path2 = new Path("/test/bar.txt");
+      Path path3 = new Path("/test/derp.txt");
+      DFSTestUtil.createFile(fs, path1, 1024, (short) 3, 0);
+      DFSTestUtil.createFile(fs, path2, 1024, (short) 3, 0);
+      DFSTestUtil.createFile(fs, path3, 1024, (short) 3, 0);
+      fs.close();
+      fs = getHttpFSFileSystem();
+      fs.concat(path1, new Path[]{path2, path3});
+      fs.close();
+      fs = FileSystem.get(config);
+      Assert.assertTrue(fs.exists(path1));
+      Assert.assertFalse(fs.exists(path2));
+      Assert.assertFalse(fs.exists(path3));
+      fs.close();
+    }
+  }
+
   private void testRename() throws Exception {
     FileSystem fs = FileSystem.get(getProxiedFSConf());
     Path path = new Path(getProxiedFSTestDir(), "foo");
@@ -450,7 +476,7 @@ public abstract class BaseTestHttpFSWith extends HFSTestCase {
   }
 
   protected enum Operation {
-    GET, OPEN, CREATE, APPEND, RENAME, DELETE, LIST_STATUS, WORKING_DIRECTORY, MKDIRS,
+    GET, OPEN, CREATE, APPEND, CONCAT, RENAME, DELETE, LIST_STATUS, WORKING_DIRECTORY, MKDIRS,
     SET_TIMES, SET_PERMISSION, SET_OWNER, SET_REPLICATION, CHECKSUM, CONTENT_SUMMARY
   }
 
@@ -468,6 +494,8 @@ public abstract class BaseTestHttpFSWith extends HFSTestCase {
       case APPEND:
         testAppend();
         break;
+      case CONCAT:
+        testConcat();
       case RENAME:
         testRename();
         break;

hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt

@@ -296,7 +296,25 @@ Trunk (Unreleased)
 
     HDFS-4382. Fix typo MAX_NOT_CHANGED_INTERATIONS. (Ted Yu via suresh)
 
-Release 2.0.3-alpha - Unreleased 
+    HDFS-4340. Update addBlock() to inculde inode id as additional argument.
+    (Brandon Li via suresh)
+
+Release 2.0.4-beta - UNRELEASED
+
+  INCOMPATIBLE CHANGES
+
+  NEW FEATURES
+
+  IMPROVEMENTS
+
+  OPTIMIZATIONS
+
+  BUG FIXES
+    
+    HDFS-4470. Several HDFS tests attempt file operations on invalid HDFS
+    paths when running on Windows. (Chris Nauroth via suresh)
+
+Release 2.0.3-alpha - 2013-02-06
 
   INCOMPATIBLE CHANGES
 
@@ -317,6 +335,9 @@ Release 2.0.3-alpha - Unreleased
     HDFS-4451. hdfs balancer command returns exit code 1 on success instead
     of 0. (Joshua Blatt via suresh)
 
+    HDFS-4350. Make enabling of stale marking on read and write paths
+    independent. (Andrew Wang via suresh)
+
 
   NEW FEATURES
 
@@ -505,6 +526,8 @@ Release 2.0.3-alpha - Unreleased
 
     HDFS-3598. WebHDFS support for file concat. (Plamen Jeliazkov via shv)
 
+    HDFS-4456. Add concat to HttpFS and WebHDFS REST API docs. (plamenj2003 via tucu)
+
   OPTIMIZATIONS
 
     HDFS-3429. DataNode reads checksums even if client does not need them (todd)
@@ -742,6 +765,28 @@ Release 2.0.3-alpha - Unreleased
     HDFS-4428. FsDatasetImpl should disclose what the error is when a rename
     fails. (Colin Patrick McCabe via atm)
 
+    HDFS-4452. getAdditionalBlock() can create multiple blocks if the client
+    times out and retries. (shv)
+
+    HDFS-4445. All BKJM ledgers are not checked while tailing, So failover will fail.
+    (Vinay via umamahesh)
+
+    HDFS-4462. 2NN will fail to checkpoint after an HDFS upgrade from a
+    pre-federation version of HDFS. (atm)
+
+    HDFS-4404. Create file failure when the machine of first attempted NameNode
+    is down. (Todd Lipcon via atm)
+
+    HDFS-4344. dfshealth.jsp throws NumberFormatException when
+    dfs.hosts/dfs.hosts.exclude includes port number. (Andy Isaacson via atm)
+
+    HDFS-4468.  Use the new StringUtils methods added by HADOOP-9252 and fix
+    TestHDFSCLI and TestQuota. (szetszwo)
+
+    HDFS-4458. In DFSUtil.getNameServiceUris(..), convert default fs URI using
+    NetUtils.createSocketAddr(..) for being consistent with other addresses.
+    (Binglin Chang via szetszwo)
+
   BREAKDOWN OF HDFS-3077 SUBTASKS
 
     HDFS-3077. Quorum-based protocol for reading and writing edit logs.

hadoop-hdfs-project/hadoop-hdfs/src/contrib/bkjournal/src/main/java/org/apache/hadoop/contrib/bkjournal/BookKeeperJournalManager.java

@@ -503,7 +503,8 @@ public class BookKeeperJournalManager implements JournalManager {
   @Override
   public void selectInputStreams(Collection<EditLogInputStream> streams,
       long fromTxId, boolean inProgressOk) throws IOException {
-    List<EditLogLedgerMetadata> currentLedgerList = getLedgerList(inProgressOk);
+    List<EditLogLedgerMetadata> currentLedgerList = getLedgerList(fromTxId,
+        inProgressOk);
     try {
       BookKeeperEditLogInputStream elis = null;
       for (EditLogLedgerMetadata l : currentLedgerList) {
@@ -511,6 +512,8 @@ public class BookKeeperJournalManager implements JournalManager {
         if (l.isInProgress()) {
           lastTxId = recoverLastTxId(l, false);
         }
+        // Check once again, required in case of InProgress and is case of any
+        // gap.
         if (fromTxId >= l.getFirstTxId() && fromTxId <= lastTxId) {
           LedgerHandle h;
           if (l.isInProgress()) { // we don't want to fence the current journal
@@ -523,6 +526,8 @@ public class BookKeeperJournalManager implements JournalManager {
           elis = new BookKeeperEditLogInputStream(h, l);
           elis.skipTo(fromTxId);
         } else {
+          // If mismatches then there might be some gap, so we should not check
+          // further.
           return;
         }
         streams.add(elis);
@@ -732,6 +737,11 @@ public class BookKeeperJournalManager implements JournalManager {
    */
   List<EditLogLedgerMetadata> getLedgerList(boolean inProgressOk)
       throws IOException {
+    return getLedgerList(-1, inProgressOk);
+  }
+
+  private List<EditLogLedgerMetadata> getLedgerList(long fromTxId,
+      boolean inProgressOk) throws IOException {
     List<EditLogLedgerMetadata> ledgers
       = new ArrayList<EditLogLedgerMetadata>();
     try {
@@ -744,6 +754,12 @@ public class BookKeeperJournalManager implements JournalManager {
         try {
           EditLogLedgerMetadata editLogLedgerMetadata = EditLogLedgerMetadata
               .read(zkc, legderMetadataPath);
+          if (editLogLedgerMetadata.getLastTxId() != HdfsConstants.INVALID_TXID
+              && editLogLedgerMetadata.getLastTxId() < fromTxId) {
+            // exclude already read closed edits, but include inprogress edits
+            // as this will be handled in caller
+            continue;
+          }
           ledgers.add(editLogLedgerMetadata);
         } catch (KeeperException.NoNodeException e) {
           LOG.warn("ZNode: " + legderMetadataPath

hadoop-hdfs-project/hadoop-hdfs/src/contrib/bkjournal/src/test/java/org/apache/hadoop/contrib/bkjournal/TestBookKeeperAsHASharedDir.java

@@ -21,7 +21,6 @@ import static org.junit.Assert.*;
 
 import org.junit.Test;
 import org.junit.Before;
-import org.junit.After;
 import org.junit.BeforeClass;
 import org.junit.AfterClass;
 
@@ -34,11 +33,9 @@ import org.apache.hadoop.hdfs.DFSConfigKeys;
 import org.apache.hadoop.hdfs.HAUtil;
 import org.apache.hadoop.hdfs.MiniDFSCluster;
 import org.apache.hadoop.hdfs.MiniDFSNNTopology;
-import org.apache.hadoop.hdfs.DFSTestUtil;
 
 import org.apache.hadoop.hdfs.server.namenode.ha.HATestUtil;
 import org.apache.hadoop.hdfs.server.namenode.NameNode;
-import org.apache.hadoop.hdfs.server.namenode.FSEditLogTestUtil;
 import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
 
 import org.apache.hadoop.ipc.RemoteException;
@@ -352,4 +349,42 @@ public class TestBookKeeperAsHASharedDir {
       }
     }
   }
+
+  /**
+   * NameNode should load the edits correctly if the applicable edits are
+   * present in the BKJM.
+   */
+  @Test
+  public void testNameNodeMultipleSwitchesUsingBKJM() throws Exception {
+    MiniDFSCluster cluster = null;
+    try {
+      Configuration conf = new Configuration();
+      conf.setInt(DFSConfigKeys.DFS_HA_TAILEDITS_PERIOD_KEY, 1);
+      conf.set(DFSConfigKeys.DFS_NAMENODE_SHARED_EDITS_DIR_KEY, BKJMUtil
+          .createJournalURI("/correctEditLogSelection").toString());
+      BKJMUtil.addJournalManagerDefinition(conf);
+
+      cluster = new MiniDFSCluster.Builder(conf)
+          .nnTopology(MiniDFSNNTopology.simpleHATopology()).numDataNodes(0)
+          .manageNameDfsSharedDirs(false).build();
+      NameNode nn1 = cluster.getNameNode(0);
+      NameNode nn2 = cluster.getNameNode(1);
+      cluster.waitActive();
+      cluster.transitionToActive(0);
+      nn1.getRpcServer().rollEditLog(); // Roll Edits from current Active.
+      // Transition to standby current active gracefully.
+      cluster.transitionToStandby(0);
+      // Make the other Active and Roll edits multiple times
+      cluster.transitionToActive(1);
+      nn2.getRpcServer().rollEditLog();
+      nn2.getRpcServer().rollEditLog();
+      // Now One more failover. So NN1 should be able to failover successfully.
+      cluster.transitionToStandby(1);
+      cluster.transitionToActive(0);
+    } finally {
+      if (cluster != null) {
+        cluster.shutdown();
+      }
+    }
+  }
 }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java

@@ -181,10 +181,10 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final String  DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_KEY = "dfs.datanode.socket.reuse.keepalive";
   public static final int     DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT = 1000;
   
-  // Whether to enable datanode's stale state detection and usage
-  public static final String DFS_NAMENODE_CHECK_STALE_DATANODE_KEY = "dfs.namenode.check.stale.datanode";
-  public static final boolean DFS_NAMENODE_CHECK_STALE_DATANODE_DEFAULT = false;
-  // Whether to enable datanode's stale state detection and usage
+  // Whether to enable datanode's stale state detection and usage for reads
+  public static final String DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_KEY = "dfs.namenode.avoid.read.stale.datanode";
+  public static final boolean DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_DEFAULT = false;
+  // Whether to enable datanode's stale state detection and usage for writes
   public static final String DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY = "dfs.namenode.avoid.write.stale.datanode";
   public static final boolean DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_DEFAULT = false;
   // The default value of the time interval for marking datanodes as stale
@@ -195,8 +195,8 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final String DFS_NAMENODE_STALE_DATANODE_MINIMUM_INTERVAL_KEY = "dfs.namenode.stale.datanode.minimum.interval";
   public static final int DFS_NAMENODE_STALE_DATANODE_MINIMUM_INTERVAL_DEFAULT = 3; // i.e. min_interval is 3 * heartbeat_interval = 9s
   
-  // When the number stale datanodes marked as stale reached this certian ratio, 
-  // stop avoiding writing to stale nodes so as to prevent causing hotspots.
+  // When the percentage of stale datanodes reaches this ratio,
+  // allow writing to stale nodes to prevent hotspots.
   public static final String DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_KEY = "dfs.namenode.write.stale.datanode.ratio";
   public static final float DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_DEFAULT = 0.5f;
 

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSOutputStream.java

@@ -115,6 +115,7 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
   private volatile boolean closed = false;
 
   private String src;
+  private final long fileId;
   private final long blockSize;
   private final DataChecksum checksum;
   // both dataQueue and ackQueue are protected by dataQueue lock
@@ -1148,7 +1149,8 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
         long localstart = Time.now();
         while (true) {
           try {
-            return dfsClient.namenode.addBlock(src, dfsClient.clientName, block, excludedNodes);
+            return dfsClient.namenode.addBlock(src, dfsClient.clientName,
+                block, excludedNodes, fileId);
           } catch (RemoteException e) {
             IOException ue = 
               e.unwrapRemoteException(FileNotFoundException.class,
@@ -1261,20 +1263,21 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
     return value;
   }
 
-  private DFSOutputStream(DFSClient dfsClient, String src, long blockSize, Progressable progress,
-      DataChecksum checksum, short replication) throws IOException {
+  private DFSOutputStream(DFSClient dfsClient, String src, Progressable progress,
+      HdfsFileStatus stat, DataChecksum checksum) throws IOException {
     super(checksum, checksum.getBytesPerChecksum(), checksum.getChecksumSize());
-    int bytesPerChecksum = checksum.getBytesPerChecksum();
     this.dfsClient = dfsClient;
     this.src = src;
-    this.blockSize = blockSize;
-    this.blockReplication = replication;
+    this.fileId = stat.getFileId();
+    this.blockSize = stat.getBlockSize();
+    this.blockReplication = stat.getReplication();
     this.progress = progress;
     if ((progress != null) && DFSClient.LOG.isDebugEnabled()) {
       DFSClient.LOG.debug(
           "Set non-null progress callback on DFSOutputStream " + src);
     }
     
+    final int bytesPerChecksum = checksum.getBytesPerChecksum();
     if ( bytesPerChecksum < 1 || blockSize % bytesPerChecksum != 0) {
       throw new IOException("io.bytes.per.checksum(" + bytesPerChecksum +
                             ") and blockSize(" + blockSize + 
@@ -1286,19 +1289,27 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
   }
 
   /** Construct a new output stream for creating a file. */
-  private DFSOutputStream(DFSClient dfsClient, String src, FsPermission masked,
-      EnumSet<CreateFlag> flag, boolean createParent, short replication,
-      long blockSize, Progressable progress, int buffersize,
+  private DFSOutputStream(DFSClient dfsClient, String src, HdfsFileStatus stat,
+      EnumSet<CreateFlag> flag, Progressable progress,
       DataChecksum checksum) throws IOException {
-    this(dfsClient, src, blockSize, progress, checksum, replication);
+    this(dfsClient, src, progress, stat, checksum);
     this.shouldSyncBlock = flag.contains(CreateFlag.SYNC_BLOCK);
 
     computePacketChunkSize(dfsClient.getConf().writePacketSize,
         checksum.getBytesPerChecksum());
 
+    streamer = new DataStreamer();
+  }
+
+  static DFSOutputStream newStreamForCreate(DFSClient dfsClient, String src,
+      FsPermission masked, EnumSet<CreateFlag> flag, boolean createParent,
+      short replication, long blockSize, Progressable progress, int buffersize,
+      DataChecksum checksum) throws IOException {
+    final HdfsFileStatus stat;
     try {
-      dfsClient.namenode.create(
-          src, masked, dfsClient.clientName, new EnumSetWritable<CreateFlag>(flag), createParent, replication, blockSize);
+      stat = dfsClient.namenode.create(src, masked, dfsClient.clientName,
+          new EnumSetWritable<CreateFlag>(flag), createParent, replication,
+          blockSize);
     } catch(RemoteException re) {
       throw re.unwrapRemoteException(AccessControlException.class,
                                      DSQuotaExceededException.class,
@@ -1309,30 +1320,20 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
                                      SafeModeException.class,
                                      UnresolvedPathException.class);
     }
-    streamer = new DataStreamer();
-  }
-
-  static DFSOutputStream newStreamForCreate(DFSClient dfsClient, String src,
-      FsPermission masked, EnumSet<CreateFlag> flag, boolean createParent,
-      short replication, long blockSize, Progressable progress, int buffersize,
-      DataChecksum checksum) throws IOException {
-    final DFSOutputStream out = new DFSOutputStream(dfsClient, src, masked,
-        flag, createParent, replication, blockSize, progress, buffersize,
-        checksum);
-    out.streamer.start();
+    final DFSOutputStream out = new DFSOutputStream(dfsClient, src, stat,
+        flag, progress, checksum);
+    out.start();
     return out;
   }
 
   /** Construct a new output stream for append. */
-  private DFSOutputStream(DFSClient dfsClient, String src, int buffersize, Progressable progress,
-      LocatedBlock lastBlock, HdfsFileStatus stat,
+  private DFSOutputStream(DFSClient dfsClient, String src,
+      Progressable progress, LocatedBlock lastBlock, HdfsFileStatus stat,
       DataChecksum checksum) throws IOException {
-    this(dfsClient, src, stat.getBlockSize(), progress, checksum, stat.getReplication());
+    this(dfsClient, src, progress, stat, checksum);
     initialFileSize = stat.getLen(); // length of file when opened
 
-    //
     // The last partial block of the file has to be filled.
-    //
     if (lastBlock != null) {
       // indicate that we are appending to an existing block
       bytesCurBlock = lastBlock.getBlockSize();
@@ -1347,9 +1348,9 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
   static DFSOutputStream newStreamForAppend(DFSClient dfsClient, String src,
       int buffersize, Progressable progress, LocatedBlock lastBlock,
       HdfsFileStatus stat, DataChecksum checksum) throws IOException {
-    final DFSOutputStream out = new DFSOutputStream(dfsClient, src, buffersize,
+    final DFSOutputStream out = new DFSOutputStream(dfsClient, src,
         progress, lastBlock, stat, checksum);
-    out.streamer.start();
+    out.start();
     return out;
   }
 
@@ -1716,6 +1717,10 @@ public class DFSOutputStream extends FSOutputSummer implements Syncable {
     isClosed();
   }
 
+  private synchronized void start() {
+    streamer.start();
+  }
+  
   /**
    * Aborts this output stream and releases any system 
    * resources associated with this stream.

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSUtil.java

@@ -134,7 +134,7 @@ public class DFSUtil {
   /**
    * Comparator for sorting DataNodeInfo[] based on decommissioned/stale states.
    * Decommissioned/stale nodes are moved to the end of the array on sorting
-   * with this compartor.
+   * with this comparator.
    */ 
   @InterfaceAudience.Private 
   public static class DecomStaleComparator implements Comparator<DatanodeInfo> {
@@ -144,7 +144,7 @@ public class DFSUtil {
      * Constructor of DecomStaleComparator
      * 
      * @param interval
-     *          The time invertal for marking datanodes as stale is passed from
+     *          The time interval for marking datanodes as stale is passed from
      *          outside, since the interval may be changed dynamically
      */
     public DecomStaleComparator(long interval) {
@@ -766,6 +766,13 @@ public class DFSUtil {
     
     // Add the default URI if it is an HDFS URI.
     URI defaultUri = FileSystem.getDefaultUri(conf);
+    // checks if defaultUri is ip:port format
+    // and convert it to hostname:port format
+    if (defaultUri != null && (defaultUri.getPort() != -1)) {
+      defaultUri = createUri(defaultUri.getScheme(),
+          NetUtils.createSocketAddr(defaultUri.getHost(), 
+              defaultUri.getPort()));
+    }
     if (defaultUri != null &&
         HdfsConstants.HDFS_URI_SCHEME.equals(defaultUri.getScheme()) &&
         !nonPreferredUris.contains(defaultUri)) {
@@ -925,6 +932,11 @@ public class DFSUtil {
     return capacity <= 0 ? 0 : (remaining * 100.0f)/capacity; 
   }
 
+  /** Convert percentage to a string. */
+  public static String percent2String(double percentage) {
+    return StringUtils.format("%.2f%%", percentage);
+  }
+
   /**
    * Round bytes to GiB (gibibyte)
    * @param bytes number of bytes

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocol/ClientProtocol.java

@@ -150,6 +150,8 @@ public interface ClientProtocol {
    * @param replication block replication factor.
    * @param blockSize maximum block size.
    * 
+   * @return the status of the created file, it could be null if the server
+   *           doesn't support returning the file status
    * @throws AccessControlException If access is denied
    * @throws AlreadyBeingCreatedException if the path does not exist.
    * @throws DSQuotaExceededException If file creation violates disk space 
@@ -168,13 +170,14 @@ public interface ClientProtocol {
    * RuntimeExceptions:
    * @throws InvalidPathException Path <code>src</code> is invalid
    */
-  public void create(String src, FsPermission masked, String clientName,
-      EnumSetWritable<CreateFlag> flag, boolean createParent,
-      short replication, long blockSize) throws AccessControlException,
-      AlreadyBeingCreatedException, DSQuotaExceededException,
-      FileAlreadyExistsException, FileNotFoundException,
-      NSQuotaExceededException, ParentNotDirectoryException, SafeModeException,
-      UnresolvedLinkException, IOException;
+  public HdfsFileStatus create(String src, FsPermission masked,
+      String clientName, EnumSetWritable<CreateFlag> flag,
+      boolean createParent, short replication, long blockSize)
+      throws AccessControlException, AlreadyBeingCreatedException,
+      DSQuotaExceededException, FileAlreadyExistsException,
+      FileNotFoundException, NSQuotaExceededException,
+      ParentNotDirectoryException, SafeModeException, UnresolvedLinkException,
+      IOException;
 
   /**
    * Append to the end of the file. 
@@ -296,6 +299,7 @@ public interface ClientProtocol {
    * @param previous  previous block
    * @param excludeNodes a list of nodes that should not be
    * allocated for the current block
+   * @param fileId the id uniquely identifying a file
    *
    * @return LocatedBlock allocated block information.
    *
@@ -310,7 +314,7 @@ public interface ClientProtocol {
    */
   @Idempotent
   public LocatedBlock addBlock(String src, String clientName,
-      ExtendedBlock previous, DatanodeInfo[] excludeNodes)
+      ExtendedBlock previous, DatanodeInfo[] excludeNodes, long fileId)
       throws AccessControlException, FileNotFoundException,
       NotReplicatedYetException, SafeModeException, UnresolvedLinkException,
       IOException;

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocol/DSQuotaExceededException.java

@@ -20,7 +20,7 @@ package org.apache.hadoop.hdfs.protocol;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
-import org.apache.hadoop.util.StringUtils;
+import static org.apache.hadoop.util.StringUtils.TraditionalBinaryPrefix.long2String;
 
 @InterfaceAudience.Private
 @InterfaceStability.Evolving
@@ -41,9 +41,9 @@ public class DSQuotaExceededException extends QuotaExceededException {
   public String getMessage() {
     String msg = super.getMessage();
     if (msg == null) {
-      return "The DiskSpace quota" + (pathName==null?"":(" of " + pathName)) + 
-          " is exceeded: quota=" + StringUtils.humanReadableInt(quota) + 
-          " diskspace consumed=" + StringUtils.humanReadableInt(count);
+      return "The DiskSpace quota" + (pathName==null?"": " of " + pathName)
+          + " is exceeded: quota = " + quota + " B = " + long2String(quota, "B", 2)
+          + " but diskspace consumed = " + count + " B = " + long2String(count, "B", 2);
     } else {
       return msg;
     }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocol/DatanodeInfo.java

@@ -17,10 +17,13 @@
  */
 package org.apache.hadoop.hdfs.protocol;
 
+import static org.apache.hadoop.hdfs.DFSUtil.percent2String;
+
 import java.util.Date;
 
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
+import org.apache.hadoop.hdfs.DFSConfigKeys;
 import org.apache.hadoop.hdfs.DFSUtil;
 import org.apache.hadoop.net.NetUtils;
 import org.apache.hadoop.net.NetworkTopology;
@@ -244,8 +247,8 @@ public class DatanodeInfo extends DatanodeID implements Node {
     buffer.append("DFS Used: "+u+" ("+StringUtils.byteDesc(u)+")"+"\n");
     buffer.append("Non DFS Used: "+nonDFSUsed+" ("+StringUtils.byteDesc(nonDFSUsed)+")"+"\n");
     buffer.append("DFS Remaining: " +r+ " ("+StringUtils.byteDesc(r)+")"+"\n");
-    buffer.append("DFS Used%: "+StringUtils.limitDecimalTo2(usedPercent)+"%\n");
-    buffer.append("DFS Remaining%: "+StringUtils.limitDecimalTo2(remainingPercent)+"%\n");
+    buffer.append("DFS Used%: "+percent2String(usedPercent) + "\n");
+    buffer.append("DFS Remaining%: "+percent2String(remainingPercent) + "\n");
     buffer.append("Last contact: "+new Date(lastUpdate)+"\n");
     return buffer.toString();
   }
@@ -269,7 +272,7 @@ public class DatanodeInfo extends DatanodeID implements Node {
     }
     buffer.append(" " + c + "(" + StringUtils.byteDesc(c)+")");
     buffer.append(" " + u + "(" + StringUtils.byteDesc(u)+")");
-    buffer.append(" " + StringUtils.limitDecimalTo2(((1.0*u)/c)*100)+"%");
+    buffer.append(" " + percent2String(u/(double)c));
     buffer.append(" " + r + "(" + StringUtils.byteDesc(r)+")");
     buffer.append(" " + new Date(lastUpdate));
     return buffer.toString();

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocol/HdfsFileStatus.java

@@ -40,6 +40,7 @@ public class HdfsFileStatus {
   private FsPermission permission;
   private String owner;
   private String group;
+  private long fileId;
   
   public static final byte[] EMPTY_NAME = new byte[0];
 
@@ -55,11 +56,12 @@ public class HdfsFileStatus {
    * @param owner the owner of the path
    * @param group the group of the path
    * @param path the local name in java UTF8 encoding the same as that in-memory
+   * @param fileId the file id
    */
   public HdfsFileStatus(long length, boolean isdir, int block_replication,
                     long blocksize, long modification_time, long access_time,
                     FsPermission permission, String owner, String group, 
-                    byte[] symlink, byte[] path) {
+                    byte[] symlink, byte[] path, long fileId) {
     this.length = length;
     this.isdir = isdir;
     this.block_replication = (short)block_replication;
@@ -75,6 +77,7 @@ public class HdfsFileStatus {
     this.group = (group == null) ? "" : group;
     this.symlink = symlink;
     this.path = path;
+    this.fileId = fileId;
   }
 
   /**
@@ -223,4 +226,8 @@ public class HdfsFileStatus {
   final public byte[] getSymlinkInBytes() {
     return symlink;
   }
+  
+  final public long getFileId() {
+    return fileId;
+  }
 }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocol/HdfsLocatedFileStatus.java

@@ -44,19 +44,19 @@ public class HdfsLocatedFileStatus extends HdfsFileStatus {
    * @param group group
    * @param symlink symbolic link
    * @param path local path name in java UTF8 format 
+   * @param fileId the file id
    * @param locations block locations
    */
   public HdfsLocatedFileStatus(long length, boolean isdir,
-      int block_replication,
-	    long blocksize, long modification_time, long access_time,
-	    FsPermission permission, String owner, String group, 
-	    byte[] symlink, byte[] path, LocatedBlocks locations) {
-	  super(length, isdir, block_replication, blocksize, modification_time,
-		  access_time, permission, owner, group, symlink, path);
+      int block_replication, long blocksize, long modification_time,
+      long access_time, FsPermission permission, String owner, String group,
+      byte[] symlink, byte[] path, long fileId, LocatedBlocks locations) {
+    super(length, isdir, block_replication, blocksize, modification_time,
+        access_time, permission, owner, group, symlink, path, fileId);
     this.locations = locations;
-	}
+  }
 	
-	public LocatedBlocks getBlockLocations() {
-		return locations;
-	}
+  public LocatedBlocks getBlockLocations() {
+    return locations;
+  }
 }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocolPB/ClientNamenodeProtocolServerSideTranslatorPB.java

@@ -268,14 +268,19 @@ public class ClientNamenodeProtocolServerSideTranslatorPB implements
   public CreateResponseProto create(RpcController controller,
       CreateRequestProto req) throws ServiceException {
     try {
-      server.create(req.getSrc(), PBHelper.convert(req.getMasked()),
-          req.getClientName(), PBHelper.convert(req.getCreateFlag()),
-          req.getCreateParent(), (short) req.getReplication(),
-          req.getBlockSize());
+      HdfsFileStatus result = server.create(req.getSrc(),
+          PBHelper.convert(req.getMasked()), req.getClientName(),
+          PBHelper.convert(req.getCreateFlag()), req.getCreateParent(),
+          (short) req.getReplication(), req.getBlockSize());
+
+      if (result != null) {
+        return CreateResponseProto.newBuilder().setFs(PBHelper.convert(result))
+            .build();
+      }
+      return VOID_CREATE_RESPONSE;
     } catch (IOException e) {
       throw new ServiceException(e);
     }
-    return VOID_CREATE_RESPONSE;
   }
   
   @Override
@@ -348,13 +353,14 @@ public class ClientNamenodeProtocolServerSideTranslatorPB implements
     
     try {
       List<DatanodeInfoProto> excl = req.getExcludeNodesList();
-      LocatedBlock result = server.addBlock(req.getSrc(), req.getClientName(),
+      LocatedBlock result = server.addBlock(
+          req.getSrc(),
+          req.getClientName(),
           req.hasPrevious() ? PBHelper.convert(req.getPrevious()) : null,
-          (excl == null || 
-           excl.size() == 0) ? null : 
-            PBHelper.convert(excl.toArray(new DatanodeInfoProto[excl.size()])));
-      return AddBlockResponseProto.newBuilder().setBlock(
-          PBHelper.convert(result)).build();
+          (excl == null || excl.size() == 0) ? null : PBHelper.convert(excl
+              .toArray(new DatanodeInfoProto[excl.size()])), req.getFileId());
+      return AddBlockResponseProto.newBuilder()
+          .setBlock(PBHelper.convert(result)).build();
     } catch (IOException e) {
       throw new ServiceException(e);
     }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocolPB/ClientNamenodeProtocolTranslatorPB.java

@@ -54,6 +54,7 @@ import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.Append
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.CompleteRequestProto;
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.ConcatRequestProto;
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.CreateRequestProto;
+import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.CreateResponseProto;
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.CreateSymlinkRequestProto;
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.DeleteRequestProto;
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.FinalizeUpgradeRequestProto;
@@ -100,6 +101,7 @@ import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.Update
 import org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos.UpdatePipelineRequestProto;
 import org.apache.hadoop.hdfs.security.token.block.DataEncryptionKey;
 import org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenIdentifier;
+import org.apache.hadoop.hdfs.server.namenode.INodeId;
 import org.apache.hadoop.hdfs.server.namenode.NotReplicatedYetException;
 import org.apache.hadoop.hdfs.server.namenode.SafeModeException;
 import org.apache.hadoop.io.EnumSetWritable;
@@ -193,13 +195,14 @@ public class ClientNamenodeProtocolTranslatorPB implements
   }
 
   @Override
-  public void create(String src, FsPermission masked, String clientName,
-      EnumSetWritable<CreateFlag> flag, boolean createParent,
-      short replication, long blockSize) throws AccessControlException,
-      AlreadyBeingCreatedException, DSQuotaExceededException,
-      FileAlreadyExistsException, FileNotFoundException,
-      NSQuotaExceededException, ParentNotDirectoryException, SafeModeException,
-      UnresolvedLinkException, IOException {
+  public HdfsFileStatus create(String src, FsPermission masked,
+      String clientName, EnumSetWritable<CreateFlag> flag,
+      boolean createParent, short replication, long blockSize)
+      throws AccessControlException, AlreadyBeingCreatedException,
+      DSQuotaExceededException, FileAlreadyExistsException,
+      FileNotFoundException, NSQuotaExceededException,
+      ParentNotDirectoryException, SafeModeException, UnresolvedLinkException,
+      IOException {
     CreateRequestProto req = CreateRequestProto.newBuilder()
         .setSrc(src)
         .setMasked(PBHelper.convert(masked))
@@ -210,7 +213,8 @@ public class ClientNamenodeProtocolTranslatorPB implements
         .setBlockSize(blockSize)
         .build();
     try {
-      rpcProxy.create(null, req);
+      CreateResponseProto res = rpcProxy.create(null, req);
+      return res.hasFs() ? PBHelper.convert(res.getFs()) : null;
     } catch (ServiceException e) {
       throw ProtobufHelper.getRemoteException(e);
     }
@@ -294,15 +298,15 @@ public class ClientNamenodeProtocolTranslatorPB implements
       throw ProtobufHelper.getRemoteException(e);
     }
   }
-
+  
   @Override
   public LocatedBlock addBlock(String src, String clientName,
-      ExtendedBlock previous, DatanodeInfo[] excludeNodes)
+      ExtendedBlock previous, DatanodeInfo[] excludeNodes, long fileId)
       throws AccessControlException, FileNotFoundException,
       NotReplicatedYetException, SafeModeException, UnresolvedLinkException,
       IOException {
-    AddBlockRequestProto.Builder req = AddBlockRequestProto.newBuilder().setSrc(src)
-        .setClientName(clientName);
+    AddBlockRequestProto.Builder req = AddBlockRequestProto.newBuilder()
+        .setSrc(src).setClientName(clientName).setFileId(fileId);
     if (previous != null) 
       req.setPrevious(PBHelper.convert(previous)); 
     if (excludeNodes != null) 

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/protocolPB/PBHelper.java

@@ -106,6 +106,7 @@ import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.NamenodeRole;
 import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.ReplicaState;
 import org.apache.hadoop.hdfs.server.common.StorageInfo;
 import org.apache.hadoop.hdfs.server.namenode.CheckpointSignature;
+import org.apache.hadoop.hdfs.server.namenode.INodeId;
 import org.apache.hadoop.hdfs.server.protocol.BalancerBandwidthCommand;
 import org.apache.hadoop.hdfs.server.protocol.BlockCommand;
 import org.apache.hadoop.hdfs.server.protocol.BlockRecoveryCommand;
@@ -1045,6 +1046,7 @@ public class PBHelper {
         fs.getFileType().equals(FileType.IS_SYMLINK) ? 
             fs.getSymlink().toByteArray() : null,
         fs.getPath().toByteArray(),
+        fs.hasFileId()? fs.getFileId(): INodeId.GRANDFATHER_INODE_ID,
         fs.hasLocations() ? PBHelper.convert(fs.getLocations()) : null);
   }
 
@@ -1069,6 +1071,7 @@ public class PBHelper {
       setPermission(PBHelper.convert(fs.getPermission())).
       setOwner(fs.getOwner()).
       setGroup(fs.getGroup()).
+      setFileId(fs.getFileId()).
       setPath(ByteString.copyFrom(fs.getLocalNameInBytes()));
     if (fs.isSymlink())  {
       builder.setSymlink(ByteString.copyFrom(fs.getSymlinkInBytes()));

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeManager.java

@@ -126,15 +126,26 @@ public class DatanodeManager {
   private final long heartbeatExpireInterval;
   /** Ask Datanode only up to this many blocks to delete. */
   final int blockInvalidateLimit;
-  
-  /** Whether or not to check stale DataNodes for read/write */
-  private final boolean checkForStaleDataNodes;
 
   /** The interval for judging stale DataNodes for read/write */
   private final long staleInterval;
   
-  /** Whether or not to avoid using stale DataNodes for writing */
-  private volatile boolean avoidStaleDataNodesForWrite;
+  /** Whether or not to avoid using stale DataNodes for reading */
+  private final boolean avoidStaleDataNodesForRead;
+
+  /**
+   * Whether or not to avoid using stale DataNodes for writing.
+   * Note that, even if this is configured, the policy may be
+   * temporarily disabled when a high percentage of the nodes
+   * are marked as stale.
+   */
+  private final boolean avoidStaleDataNodesForWrite;
+
+  /**
+   * When the ratio of stale datanodes reaches this number, stop avoiding 
+   * writing to stale datanodes, i.e., continue using stale nodes for writing.
+   */
+  private final float ratioUseStaleDataNodesForWrite;
   
   /** The number of stale DataNodes */
   private volatile int numStaleNodes;
@@ -183,14 +194,23 @@ public class DatanodeManager {
         DFSConfigKeys.DFS_BLOCK_INVALIDATE_LIMIT_KEY, blockInvalidateLimit);
     LOG.info(DFSConfigKeys.DFS_BLOCK_INVALIDATE_LIMIT_KEY
         + "=" + this.blockInvalidateLimit);
-    
-    checkForStaleDataNodes = conf.getBoolean(
-        DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY,
-        DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_DEFAULT);
 
-    staleInterval = getStaleIntervalFromConf(conf, heartbeatExpireInterval);
-    avoidStaleDataNodesForWrite = getAvoidStaleForWriteFromConf(conf,
-        checkForStaleDataNodes);
+    this.avoidStaleDataNodesForRead = conf.getBoolean(
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_KEY,
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_DEFAULT);
+    this.avoidStaleDataNodesForWrite = conf.getBoolean(
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY,
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_DEFAULT);
+    this.staleInterval = getStaleIntervalFromConf(conf, heartbeatExpireInterval);
+    this.ratioUseStaleDataNodesForWrite = conf.getFloat(
+        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_KEY,
+        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_DEFAULT);
+    Preconditions.checkArgument(
+        (ratioUseStaleDataNodesForWrite > 0 && 
+            ratioUseStaleDataNodesForWrite <= 1.0f),
+        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_KEY +
+        " = '" + ratioUseStaleDataNodesForWrite + "' is invalid. " +
+        "It should be a positive non-zero float value, not greater than 1.0f.");
   }
   
   private static long getStaleIntervalFromConf(Configuration conf,
@@ -230,22 +250,6 @@ public class DatanodeManager {
     return staleInterval;
   }
   
-  static boolean getAvoidStaleForWriteFromConf(Configuration conf,
-      boolean checkForStale) {
-    boolean avoid = conf.getBoolean(
-        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY,
-        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_DEFAULT);
-    boolean avoidStaleDataNodesForWrite = checkForStale && avoid;
-    if (!checkForStale && avoid) {
-      LOG.warn("Cannot set "
-          + DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY
-          + " as false while setting "
-          + DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY
-          + " as true.");
-    }
-    return avoidStaleDataNodesForWrite;
-  }
-  
   void activate(final Configuration conf) {
     final DecommissionManager dm = new DecommissionManager(namesystem, blockManager);
     this.decommissionthread = new Daemon(dm.new Monitor(
@@ -299,7 +303,7 @@ public class DatanodeManager {
         client = new NodeBase(rName + NodeBase.PATH_SEPARATOR_STR + targethost);
     }
     
-    Comparator<DatanodeInfo> comparator = checkForStaleDataNodes ? 
+    Comparator<DatanodeInfo> comparator = avoidStaleDataNodesForRead ?
         new DFSUtil.DecomStaleComparator(staleInterval) : 
         DFSUtil.DECOM_COMPARATOR;
         
@@ -825,32 +829,20 @@ public class DatanodeManager {
   }
   
   /* Getter and Setter for stale DataNodes related attributes */
-  
-  /**
-   * @return whether or not to avoid writing to stale datanodes
-   */
-  public boolean isAvoidingStaleDataNodesForWrite() {
-    return avoidStaleDataNodesForWrite;
-  }
 
   /**
-   * Set the value of {@link DatanodeManager#avoidStaleDataNodesForWrite}. 
-   * The HeartbeatManager disable avoidStaleDataNodesForWrite when more than
-   * half of the DataNodes are marked as stale.
+   * Whether stale datanodes should be avoided as targets on the write path.
+   * The result of this function may change if the number of stale datanodes
+   * eclipses a configurable threshold.
    * 
-   * @param avoidStaleDataNodesForWrite
-   *          The value to set to
-   *          {@link DatanodeManager#avoidStaleDataNodesForWrite}
-   */
-  void setAvoidStaleDataNodesForWrite(boolean avoidStaleDataNodesForWrite) {
-    this.avoidStaleDataNodesForWrite = avoidStaleDataNodesForWrite;
-  }
-
-  /**
-   * @return Whether or not to check stale DataNodes for R/W
+   * @return whether stale datanodes should be avoided on the write path
    */
-  boolean isCheckingForStaleDataNodes() {
-    return checkForStaleDataNodes;
+  public boolean shouldAvoidStaleDataNodesForWrite() {
+    // If # stale exceeds maximum staleness ratio, disable stale
+    // datanode avoidance on the write path
+    return avoidStaleDataNodesForWrite &&
+        (numStaleNodes <= heartbeatManager.getLiveDatanodeCount()
+            * ratioUseStaleDataNodesForWrite);
   }
   
   /**
@@ -967,7 +959,7 @@ public class DatanodeManager {
       port = DFSConfigKeys.DFS_DATANODE_DEFAULT_PORT;
     } else {
       hostStr = hostLine.substring(0, idx);
-      port = Integer.valueOf(hostLine.substring(idx));
+      port = Integer.valueOf(hostLine.substring(idx+1));
     }
 
     if (InetAddresses.isInetAddress(hostStr)) {

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java

@@ -30,8 +30,6 @@ import org.apache.hadoop.hdfs.server.namenode.Namesystem;
 import org.apache.hadoop.util.Daemon;
 import org.apache.hadoop.util.Time;
 
-import com.google.common.base.Preconditions;
-
 /**
  * Manage the heartbeats received from datanodes.
  * The datanode list and statistics are synchronized
@@ -56,16 +54,7 @@ class HeartbeatManager implements DatanodeStatistics {
   private final long heartbeatRecheckInterval;
   /** Heartbeat monitor thread */
   private final Daemon heartbeatThread = new Daemon(new Monitor());
-  /**
-   * The initial setting of end user which indicates whether or not to avoid
-   * writing to stale datanodes.
-   */
-  private final boolean initialAvoidWriteStaleNodes;
-  /**
-   * When the ratio of stale datanodes reaches this number, stop avoiding 
-   * writing to stale datanodes, i.e., continue using stale nodes for writing.
-   */
-  private final float ratioUseStaleDataNodesForWrite;
+
     
   final Namesystem namesystem;
   final BlockManager blockManager;
@@ -74,30 +63,25 @@ class HeartbeatManager implements DatanodeStatistics {
       final BlockManager blockManager, final Configuration conf) {
     this.namesystem = namesystem;
     this.blockManager = blockManager;
-    boolean checkStaleNodes = conf.getBoolean(
-        DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY,
-        DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_DEFAULT);
+    boolean avoidStaleDataNodesForWrite = conf.getBoolean(
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY,
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_DEFAULT);
     long recheckInterval = conf.getInt(
-        DFSConfigKeys.DFS_NAMENODE_HEARTBEAT_RECHECK_INTERVAL_KEY, 
+        DFSConfigKeys.DFS_NAMENODE_HEARTBEAT_RECHECK_INTERVAL_KEY,
         DFSConfigKeys.DFS_NAMENODE_HEARTBEAT_RECHECK_INTERVAL_DEFAULT); // 5 min
     long staleInterval = conf.getLong(
-        DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_KEY, 
+        DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_KEY,
         DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_DEFAULT);// 30s
-    this.initialAvoidWriteStaleNodes = DatanodeManager
-        .getAvoidStaleForWriteFromConf(conf, checkStaleNodes);
-    this.ratioUseStaleDataNodesForWrite = conf.getFloat(
-        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_KEY,
-        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_DEFAULT);
-    Preconditions.checkArgument(
-        (ratioUseStaleDataNodesForWrite > 0 && 
-            ratioUseStaleDataNodesForWrite <= 1.0f),
-        DFSConfigKeys.DFS_NAMENODE_USE_STALE_DATANODE_FOR_WRITE_RATIO_KEY +
-        " = '" + ratioUseStaleDataNodesForWrite + "' is invalid. " +
-        "It should be a positive non-zero float value, not greater than 1.0f.");
-    
-    this.heartbeatRecheckInterval = (checkStaleNodes 
-        && initialAvoidWriteStaleNodes 
-        && staleInterval < recheckInterval) ? staleInterval : recheckInterval;
+
+    if (avoidStaleDataNodesForWrite && staleInterval < recheckInterval) {
+      this.heartbeatRecheckInterval = staleInterval;
+      LOG.info("Setting heartbeat recheck interval to " + staleInterval
+          + " since " + DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_KEY
+          + " is less than "
+          + DFSConfigKeys.DFS_NAMENODE_HEARTBEAT_RECHECK_INTERVAL_KEY);
+    } else {
+      this.heartbeatRecheckInterval = recheckInterval;
+    }
   }
 
   void activate(Configuration conf) {
@@ -242,7 +226,6 @@ class HeartbeatManager implements DatanodeStatistics {
     if (namesystem.isInSafeMode()) {
       return;
     }
-    boolean checkStaleNodes = dm.isCheckingForStaleDataNodes();
     boolean allAlive = false;
     while (!allAlive) {
       // locate the first dead node.
@@ -254,29 +237,14 @@ class HeartbeatManager implements DatanodeStatistics {
           if (dead == null && dm.isDatanodeDead(d)) {
             stats.incrExpiredHeartbeats();
             dead = d;
-            if (!checkStaleNodes) {
-              break;
-            }
           }
-          if (checkStaleNodes && 
-              d.isStale(dm.getStaleInterval())) {
+          if (d.isStale(dm.getStaleInterval())) {
             numOfStaleNodes++;
           }
         }
         
-        // Change whether to avoid using stale datanodes for writing
-        // based on proportion of stale datanodes
-        if (checkStaleNodes) {
-          dm.setNumStaleNodes(numOfStaleNodes);
-          if (numOfStaleNodes > 
-                datanodes.size() * ratioUseStaleDataNodesForWrite) {
-            dm.setAvoidStaleDataNodesForWrite(false);
-          } else {
-            if (this.initialAvoidWriteStaleNodes) {
-              dm.setAvoidStaleDataNodesForWrite(true);
-            }
-          }
-        }
+        // Set the number of stale nodes in the DatanodeManager
+        dm.setNumStaleNodes(numOfStaleNodes);
       }
 
       allAlive = dead == null;

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/common/Storage.java

@@ -905,7 +905,7 @@ public abstract class Storage extends StorageInfo {
     props.setProperty("storageType", storageType.toString());
     props.setProperty("namespaceID", String.valueOf(namespaceID));
     // Set clusterID in version with federation support
-    if (LayoutVersion.supports(Feature.FEDERATION, layoutVersion)) {
+    if (versionSupportsFederation()) {
       props.setProperty("clusterID", clusterID);
     }
     props.setProperty("cTime", String.valueOf(cTime));

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/common/StorageInfo.java

@@ -18,6 +18,8 @@
 package org.apache.hadoop.hdfs.server.common;
 
 import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.hdfs.protocol.LayoutVersion;
+import org.apache.hadoop.hdfs.protocol.LayoutVersion.Feature;
 
 import com.google.common.base.Joiner;
 
@@ -77,6 +79,10 @@ public class StorageInfo {
     namespaceID = from.namespaceID;
     cTime = from.cTime;
   }
+
+  public boolean versionSupportsFederation() {
+    return LayoutVersion.supports(Feature.FEDERATION, layoutVersion);
+  }
   
   @Override
   public String toString() {

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/CheckpointSignature.java

@@ -123,6 +123,10 @@ public class CheckpointSignature extends StorageInfo
       blockpoolID.equals(si.getBlockPoolID());
   }
 
+  boolean namespaceIdMatches(FSImage si) {
+    return namespaceID == si.getStorage().namespaceID;
+  }
+
   void validateStorageInfo(FSImage si) throws IOException {
     if (!isSameCluster(si)
         || !storageVersionMatches(si.getStorage())) {

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/ClusterJspHelper.java

@@ -569,12 +569,10 @@ class ClusterJspHelper {
       toXmlItemBlock(doc, "DFS Remaining", StringUtils.byteDesc(free));
     
       // dfsUsedPercent
-      toXmlItemBlock(doc, "DFS Used%", 
-          StringUtils.limitDecimalTo2(dfsUsedPercent)+ "%");
+      toXmlItemBlock(doc, "DFS Used%", DFSUtil.percent2String(dfsUsedPercent));
     
       // dfsRemainingPercent
-      toXmlItemBlock(doc, "DFS Remaining%",
-          StringUtils.limitDecimalTo2(dfsRemainingPercent) + "%");
+      toXmlItemBlock(doc, "DFS Remaining%", DFSUtil.percent2String(dfsRemainingPercent));
     
       doc.endTag(); // storage
     

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSDirectory.java

@@ -1985,7 +1985,8 @@ public class FSDirectory implements Closeable {
         node.getUserName(),
         node.getGroupName(),
         node.isSymlink() ? ((INodeSymlink)node).getSymlink() : null,
-        path);
+        path,
+        node.getId());
   }
 
    /**
@@ -2022,6 +2023,7 @@ public class FSDirectory implements Closeable {
           node.getGroupName(),
           node.isSymlink() ? ((INodeSymlink)node).getSymlink() : null,
           path,
+          node.getId(),
           loc);
       }
 

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSNamesystem.java

@@ -1772,16 +1772,18 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
    * Create a new file entry in the namespace.
    * 
    * For description of parameters and exceptions thrown see 
-   * {@link ClientProtocol#create()}
+   * {@link ClientProtocol#create()}, except it returns valid  file status
+   * upon success
    */
-  void startFile(String src, PermissionStatus permissions, String holder,
-      String clientMachine, EnumSet<CreateFlag> flag, boolean createParent,
-      short replication, long blockSize) throws AccessControlException,
-      SafeModeException, FileAlreadyExistsException, UnresolvedLinkException,
+  HdfsFileStatus startFile(String src, PermissionStatus permissions,
+      String holder, String clientMachine, EnumSet<CreateFlag> flag,
+      boolean createParent, short replication, long blockSize)
+      throws AccessControlException, SafeModeException,
+      FileAlreadyExistsException, UnresolvedLinkException,
       FileNotFoundException, ParentNotDirectoryException, IOException {
     try {
-      startFileInt(src, permissions, holder, clientMachine, flag, createParent,
-                   replication, blockSize);
+      return startFileInt(src, permissions, holder, clientMachine, flag,
+          createParent, replication, blockSize);
     } catch (AccessControlException e) {
       if (isAuditEnabled() && isExternalInvocation()) {
         logAuditEvent(false, UserGroupInformation.getCurrentUser(),
@@ -1792,18 +1794,21 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
     }
   }
 
-  private void startFileInt(String src, PermissionStatus permissions, String holder,
-      String clientMachine, EnumSet<CreateFlag> flag, boolean createParent,
-      short replication, long blockSize) throws AccessControlException,
-      SafeModeException, FileAlreadyExistsException, UnresolvedLinkException,
+  private HdfsFileStatus startFileInt(String src, PermissionStatus permissions,
+      String holder, String clientMachine, EnumSet<CreateFlag> flag,
+      boolean createParent, short replication, long blockSize)
+      throws AccessControlException, SafeModeException,
+      FileAlreadyExistsException, UnresolvedLinkException,
       FileNotFoundException, ParentNotDirectoryException, IOException {
     boolean skipSync = false;
+    final HdfsFileStatus stat;
     writeLock();
     try {
       checkOperation(OperationCategory.WRITE);
 
       startFileInternal(src, permissions, holder, clientMachine, flag,
           createParent, replication, blockSize);
+      stat = dir.getFileInfo(src, false);
     } catch (StandbyException se) {
       skipSync = true;
       throw se;
@@ -1817,11 +1822,11 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
     } 
 
     if (isAuditEnabled() && isExternalInvocation()) {
-      final HdfsFileStatus stat = dir.getFileInfo(src, false);
       logAuditEvent(UserGroupInformation.getCurrentUser(),
                     getRemoteIp(),
                     "create", src, null, stat);
     }
+    return stat;
   }
 
   /**
@@ -2192,20 +2197,14 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
    * are replicated.  Will return an empty 2-elt array if we want the
    * client to "try again later".
    */
-  LocatedBlock getAdditionalBlock(String src,
-                                         String clientName,
-                                         ExtendedBlock previous,
-                                         HashMap<Node, Node> excludedNodes
-                                         ) 
+  LocatedBlock getAdditionalBlock(String src, long fileId, String clientName,
+      ExtendedBlock previous, HashMap<Node, Node> excludedNodes)
       throws LeaseExpiredException, NotReplicatedYetException,
       QuotaExceededException, SafeModeException, UnresolvedLinkException,
       IOException {
-    checkBlock(previous);
-    Block previousBlock = ExtendedBlock.getLocalBlock(previous);
-    long fileLength, blockSize;
+    long blockSize;
     int replication;
     DatanodeDescriptor clientNode = null;
-    Block newBlock = null;
 
     if(NameNode.stateChangeLog.isDebugEnabled()) {
       NameNode.stateChangeLog.debug(
@@ -2213,119 +2212,61 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
           +src+" for "+clientName);
     }
 
-    writeLock();
+    // Part I. Analyze the state of the file with respect to the input data.
+    readLock();
     try {
-      checkOperation(OperationCategory.WRITE);
+      LocatedBlock[] onRetryBlock = new LocatedBlock[1];
+      final INode[] inodes = analyzeFileState(
+          src, fileId, clientName, previous, onRetryBlock).getINodes();
+      final INodeFileUnderConstruction pendingFile =
+          (INodeFileUnderConstruction) inodes[inodes.length - 1];
 
-      if (isInSafeMode()) {
-        throw new SafeModeException("Cannot add block to " + src, safeMode);
-      }
-
-      // have we exceeded the configured limit of fs objects.
-      checkFsObjectLimit();
-
-      INodeFileUnderConstruction pendingFile = checkLease(src, clientName);
-      BlockInfo lastBlockInFile = pendingFile.getLastBlock();
-      if (!Block.matchingIdAndGenStamp(previousBlock, lastBlockInFile)) {
-        // The block that the client claims is the current last block
-        // doesn't match up with what we think is the last block. There are
-        // three possibilities:
-        // 1) This is the first block allocation of an append() pipeline
-        //    which started appending exactly at a block boundary.
-        //    In this case, the client isn't passed the previous block,
-        //    so it makes the allocateBlock() call with previous=null.
-        //    We can distinguish this since the last block of the file
-        //    will be exactly a full block.
-        // 2) This is a retry from a client that missed the response of a
-        //    prior getAdditionalBlock() call, perhaps because of a network
-        //    timeout, or because of an HA failover. In that case, we know
-        //    by the fact that the client is re-issuing the RPC that it
-        //    never began to write to the old block. Hence it is safe to
-        //    abandon it and allocate a new one.
-        // 3) This is an entirely bogus request/bug -- we should error out
-        //    rather than potentially appending a new block with an empty
-        //    one in the middle, etc
-
-        BlockInfo penultimateBlock = pendingFile.getPenultimateBlock();
-        if (previous == null &&
-            lastBlockInFile != null &&
-            lastBlockInFile.getNumBytes() == pendingFile.getPreferredBlockSize() &&
-            lastBlockInFile.isComplete()) {
-          // Case 1
-          if (NameNode.stateChangeLog.isDebugEnabled()) {
-             NameNode.stateChangeLog.debug(
-                 "BLOCK* NameSystem.allocateBlock: handling block allocation" +
-                 " writing to a file with a complete previous block: src=" +
-                 src + " lastBlock=" + lastBlockInFile);
-          }
-        } else if (Block.matchingIdAndGenStamp(penultimateBlock, previousBlock)) {
-          // Case 2
-          if (lastBlockInFile.getNumBytes() != 0) {
-            throw new IOException(
-                "Request looked like a retry to allocate block " +
-                lastBlockInFile + " but it already contains " +
-                lastBlockInFile.getNumBytes() + " bytes");
-          }
-
-          // The retry case ("b" above) -- abandon the old block.
-          NameNode.stateChangeLog.info("BLOCK* allocateBlock: " +
-              "caught retry for allocation of a new block in " +
-              src + ". Abandoning old block " + lastBlockInFile);
-          dir.removeBlock(src, pendingFile, lastBlockInFile);
-          dir.persistBlocks(src, pendingFile);
-        } else {
-          
-          throw new IOException("Cannot allocate block in " + src + ": " +
-              "passed 'previous' block " + previous + " does not match actual " +
-              "last block in file " + lastBlockInFile);
-        }
+      if(onRetryBlock[0] != null) {
+        // This is a retry. Just return the last block.
+        return onRetryBlock[0];
       }
 
-      // commit the last block and complete it if it has minimum replicas
-      commitOrCompleteLastBlock(pendingFile, previousBlock);
-
-      //
-      // If we fail this, bad things happen!
-      //
-      if (!checkFileProgress(pendingFile, false)) {
-        throw new NotReplicatedYetException("Not replicated yet:" + src);
-      }
-      fileLength = pendingFile.computeContentSummary().getLength();
       blockSize = pendingFile.getPreferredBlockSize();
       clientNode = pendingFile.getClientNode();
       replication = pendingFile.getBlockReplication();
     } finally {
-      writeUnlock();
+      readUnlock();
     }
 
     // choose targets for the new block to be allocated.
-    final DatanodeDescriptor targets[] = blockManager.chooseTarget(
+    final DatanodeDescriptor targets[] = getBlockManager().chooseTarget(
         src, replication, clientNode, excludedNodes, blockSize);
 
-    // Allocate a new block and record it in the INode. 
+    // Part II.
+    // Allocate a new block, add it to the INode and the BlocksMap. 
+    Block newBlock = null;
+    long offset;
     writeLock();
     try {
-      checkOperation(OperationCategory.WRITE);
-      if (isInSafeMode()) {
-        throw new SafeModeException("Cannot add block to " + src, safeMode);
-      }
+      // Run the full analysis again, since things could have changed
+      // while chooseTarget() was executing.
+      LocatedBlock[] onRetryBlock = new LocatedBlock[1];
+      INodesInPath inodesInPath =
+          analyzeFileState(src, fileId, clientName, previous, onRetryBlock);
+      INode[] inodes = inodesInPath.getINodes();
+      final INodeFileUnderConstruction pendingFile =
+          (INodeFileUnderConstruction) inodes[inodes.length - 1];
 
-      final INodesInPath inodesInPath = dir.rootDir.getExistingPathINodes(src, true);
-      final INode[] inodes = inodesInPath.getINodes();
-      final INodeFileUnderConstruction pendingFile
-          = checkLease(src, clientName, inodes[inodes.length - 1]);
-                                                           
-      if (!checkFileProgress(pendingFile, false)) {
-        throw new NotReplicatedYetException("Not replicated yet:" + src);
+      if(onRetryBlock[0] != null) {
+        // This is a retry. Just return the last block.
+        return onRetryBlock[0];
       }
 
-      // allocate new block record block locations in INode.
-      newBlock = allocateBlock(src, inodesInPath, targets);
-      
-      for (DatanodeDescriptor dn : targets) {
-        dn.incBlocksScheduled();
-      }
+      // commit the last block and complete it if it has minimum replicas
+      commitOrCompleteLastBlock(pendingFile,
+                                ExtendedBlock.getLocalBlock(previous));
+
+      // allocate new block, record block locations in INode.
+      newBlock = createNewBlock();
+      saveAllocatedBlock(src, inodesInPath, newBlock, targets);
+
       dir.persistBlocks(src, pendingFile);
+      offset = pendingFile.computeFileSize(true);
     } finally {
       writeUnlock();
     }
@@ -2333,10 +2274,115 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
       getEditLog().logSync();
     }
 
-    // Create next block
-    LocatedBlock b = new LocatedBlock(getExtendedBlock(newBlock), targets, fileLength);
-    blockManager.setBlockToken(b, BlockTokenSecretManager.AccessMode.WRITE);
-    return b;
+    // Return located block
+    return makeLocatedBlock(newBlock, targets, offset);
+  }
+
+  INodesInPath analyzeFileState(String src,
+                                long fileId,
+                                String clientName,
+                                ExtendedBlock previous,
+                                LocatedBlock[] onRetryBlock)
+          throws IOException  {
+    assert hasReadOrWriteLock();
+
+    checkBlock(previous);
+    onRetryBlock[0] = null;
+    checkOperation(OperationCategory.WRITE);
+    if (isInSafeMode()) {
+      throw new SafeModeException("Cannot add block to " + src, safeMode);
+    }
+
+    // have we exceeded the configured limit of fs objects.
+    checkFsObjectLimit();
+
+    Block previousBlock = ExtendedBlock.getLocalBlock(previous);
+    final INodesInPath inodesInPath =
+        dir.rootDir.getExistingPathINodes(src, true);
+    final INode[] inodes = inodesInPath.getINodes();
+    final INodeFileUnderConstruction pendingFile
+        = checkLease(src, fileId, clientName, inodes[inodes.length - 1]);
+    BlockInfo lastBlockInFile = pendingFile.getLastBlock();
+    if (!Block.matchingIdAndGenStamp(previousBlock, lastBlockInFile)) {
+      // The block that the client claims is the current last block
+      // doesn't match up with what we think is the last block. There are
+      // four possibilities:
+      // 1) This is the first block allocation of an append() pipeline
+      //    which started appending exactly at a block boundary.
+      //    In this case, the client isn't passed the previous block,
+      //    so it makes the allocateBlock() call with previous=null.
+      //    We can distinguish this since the last block of the file
+      //    will be exactly a full block.
+      // 2) This is a retry from a client that missed the response of a
+      //    prior getAdditionalBlock() call, perhaps because of a network
+      //    timeout, or because of an HA failover. In that case, we know
+      //    by the fact that the client is re-issuing the RPC that it
+      //    never began to write to the old block. Hence it is safe to
+      //    to return the existing block.
+      // 3) This is an entirely bogus request/bug -- we should error out
+      //    rather than potentially appending a new block with an empty
+      //    one in the middle, etc
+      // 4) This is a retry from a client that timed out while
+      //    the prior getAdditionalBlock() is still being processed,
+      //    currently working on chooseTarget(). 
+      //    There are no means to distinguish between the first and 
+      //    the second attempts in Part I, because the first one hasn't
+      //    changed the namesystem state yet.
+      //    We run this analysis again in Part II where case 4 is impossible.
+
+      BlockInfo penultimateBlock = pendingFile.getPenultimateBlock();
+      if (previous == null &&
+          lastBlockInFile != null &&
+          lastBlockInFile.getNumBytes() == pendingFile.getPreferredBlockSize() &&
+          lastBlockInFile.isComplete()) {
+        // Case 1
+        if (NameNode.stateChangeLog.isDebugEnabled()) {
+           NameNode.stateChangeLog.debug(
+               "BLOCK* NameSystem.allocateBlock: handling block allocation" +
+               " writing to a file with a complete previous block: src=" +
+               src + " lastBlock=" + lastBlockInFile);
+        }
+      } else if (Block.matchingIdAndGenStamp(penultimateBlock, previousBlock)) {
+        if (lastBlockInFile.getNumBytes() != 0) {
+          throw new IOException(
+              "Request looked like a retry to allocate block " +
+              lastBlockInFile + " but it already contains " +
+              lastBlockInFile.getNumBytes() + " bytes");
+        }
+
+        // Case 2
+        // Return the last block.
+        NameNode.stateChangeLog.info("BLOCK* allocateBlock: " +
+            "caught retry for allocation of a new block in " +
+            src + ". Returning previously allocated block " + lastBlockInFile);
+        long offset = pendingFile.computeFileSize(true);
+        onRetryBlock[0] = makeLocatedBlock(lastBlockInFile,
+            ((BlockInfoUnderConstruction)lastBlockInFile).getExpectedLocations(),
+            offset);
+        return inodesInPath;
+      } else {
+        // Case 3
+        throw new IOException("Cannot allocate block in " + src + ": " +
+            "passed 'previous' block " + previous + " does not match actual " +
+            "last block in file " + lastBlockInFile);
+      }
+    }
+
+    // Check if the penultimate block is minimally replicated
+    if (!checkFileProgress(pendingFile, false)) {
+      throw new NotReplicatedYetException("Not replicated yet: " + src);
+    }
+    return inodesInPath;
+  }
+
+  LocatedBlock makeLocatedBlock(Block blk,
+                                        DatanodeInfo[] locs,
+                                        long offset) throws IOException {
+    LocatedBlock lBlk = new LocatedBlock(
+        getExtendedBlock(blk), locs, offset);
+    getBlockManager().setBlockToken(
+        lBlk, BlockTokenSecretManager.AccessMode.WRITE);
+    return lBlk;
   }
 
   /** @see NameNode#getAdditionalDatanode(String, ExtendedBlock, DatanodeInfo[], DatanodeInfo[], int, String) */
@@ -2424,14 +2470,17 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
   }
   
   // make sure that we still have the lease on this file.
-  private INodeFileUnderConstruction checkLease(String src, String holder) 
-      throws LeaseExpiredException, UnresolvedLinkException {
+  private INodeFileUnderConstruction checkLease(String src, String holder)
+      throws LeaseExpiredException, UnresolvedLinkException,
+      FileNotFoundException {
     assert hasReadOrWriteLock();
-    return checkLease(src, holder, dir.getINode(src));
+    return checkLease(src, INodeId.GRANDFATHER_INODE_ID, holder,
+        dir.getINode(src));
   }
-
-  private INodeFileUnderConstruction checkLease(String src, String holder,
-      INode file) throws LeaseExpiredException {
+  
+  private INodeFileUnderConstruction checkLease(String src, long fileId,
+      String holder, INode file) throws LeaseExpiredException,
+      FileNotFoundException {
     assert hasReadOrWriteLock();
     if (file == null || !(file instanceof INodeFile)) {
       Lease lease = leaseManager.getLease(holder);
@@ -2452,6 +2501,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
       throw new LeaseExpiredException("Lease mismatch on " + src + " owned by "
           + pendingFile.getClientName() + " but is accessed by " + holder);
     }
+    INodeId.checkId(fileId, pendingFile);
     return pendingFile;
   }
  
@@ -2528,22 +2578,33 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
   }
 
   /**
-   * Allocate a block at the given pending filename
+   * Save allocated block at the given pending filename
    * 
    * @param src path to the file
    * @param inodesInPath representing each of the components of src. 
    *                     The last INode is the INode for the file.
    * @throws QuotaExceededException If addition of block exceeds space quota
    */
-  private Block allocateBlock(String src, INodesInPath inodesInPath,
-      DatanodeDescriptor targets[]) throws IOException {
+  BlockInfo saveAllocatedBlock(String src, INodesInPath inodesInPath,
+      Block newBlock, DatanodeDescriptor targets[]) throws IOException {
+    assert hasWriteLock();
+    BlockInfo b = dir.addBlock(src, inodesInPath, newBlock, targets);
+    NameNode.stateChangeLog.info("BLOCK* allocateBlock: " + src + ". "
+        + getBlockPoolId() + " " + b);
+    for (DatanodeDescriptor dn : targets) {
+      dn.incBlocksScheduled();
+    }
+    return b;
+  }
+
+  /**
+   * Create new block with a unique block id and a new generation stamp.
+   */
+  Block createNewBlock() throws IOException {
     assert hasWriteLock();
     Block b = new Block(getFSImage().getUniqueBlockId(), 0, 0); 
     // Increment the generation stamp for every new block.
     b.setGenerationStamp(nextGenerationStamp());
-    b = dir.addBlock(src, inodesInPath, b, targets);
-    NameNode.stateChangeLog.info("BLOCK* allocateBlock: " + src + ". "
-        + blockPoolId + " " + b);
     return b;
   }
 
@@ -5582,7 +5643,7 @@ public class FSNamesystem implements Namesystem, FSClusterStats,
   @Override
   public boolean isAvoidingStaleDataNodesForWrite() {
     return this.blockManager.getDatanodeManager()
-        .isAvoidingStaleDataNodesForWrite();
+        .shouldAvoidStaleDataNodesForWrite();
   }
 
   /**

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/INodeId.java

@@ -17,18 +17,21 @@
  */
 package org.apache.hadoop.hdfs.server.namenode;
 
+import java.io.FileNotFoundException;
+
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.util.SequentialNumber;
 
 /**
- * An id which uniquely identifies an inode
+ * An id which uniquely identifies an inode. Id 1 to 1000 are reserved for
+ * potential future usage. The id won't be recycled and is not expected to wrap
+ * around in a very long time. Root inode id is always 1001. Id 0 is used for
+ * backward compatibility support.
  */
 @InterfaceAudience.Private
-class INodeId extends SequentialNumber {
+public class INodeId extends SequentialNumber {
   /**
-   * The last reserved inode id. Reserve id 1 to 1000 for potential future
-   * usage. The id won't be recycled and is not expected to wrap around in a
-   * very long time. Root inode id will be 1001.
+   * The last reserved inode id. 
    */
   public static final long LAST_RESERVED_ID = 1000L;
 
@@ -38,6 +41,19 @@ class INodeId extends SequentialNumber {
    */
   public static final long GRANDFATHER_INODE_ID = 0;
 
+  /**
+   * To check if the request id is the same as saved id. Don't check fileId
+   * with GRANDFATHER_INODE_ID for backward compatibility.
+   */
+  public static void checkId(long requestId, INode inode)
+      throws FileNotFoundException {
+    if (requestId != GRANDFATHER_INODE_ID && requestId != inode.getId()) {
+      throw new FileNotFoundException(
+          "ID mismatch. Request id and saved id: " + requestId + " , "
+              + inode.getId());
+    }
+  }
+  
   INodeId() {
     super(LAST_RESERVED_ID);
   }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NNStorage.java

@@ -587,7 +587,7 @@ public class NNStorage extends Storage implements Closeable,
     }
 
     // Set Block pool ID in version with federation support
-    if (LayoutVersion.supports(Feature.FEDERATION, layoutVersion)) {
+    if (versionSupportsFederation()) {
       String sbpid = props.getProperty("blockpoolID");
       setBlockPoolID(sd.getRoot(), sbpid);
     }
@@ -634,7 +634,7 @@ public class NNStorage extends Storage implements Closeable,
                            ) throws IOException {
     super.setPropertiesFromFields(props, sd);
     // Set blockpoolID in version with federation support
-    if (LayoutVersion.supports(Feature.FEDERATION, layoutVersion)) {
+    if (versionSupportsFederation()) {
       props.setProperty("blockpoolID", blockpoolID);
     }
   }

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NameNodeRpcServer.java

@@ -422,13 +422,10 @@ class NameNodeRpcServer implements NamenodeProtocols {
   }
 
   @Override // ClientProtocol
-  public void create(String src, 
-                     FsPermission masked,
-                     String clientName, 
-                     EnumSetWritable<CreateFlag> flag,
-                     boolean createParent,
-                     short replication,
-                     long blockSize) throws IOException {
+  public HdfsFileStatus create(String src, FsPermission masked,
+      String clientName, EnumSetWritable<CreateFlag> flag,
+      boolean createParent, short replication, long blockSize)
+      throws IOException {
     String clientMachine = getClientMachine();
     if (stateChangeLog.isDebugEnabled()) {
       stateChangeLog.debug("*DIR* NameNode.create: file "
@@ -438,12 +435,13 @@ class NameNodeRpcServer implements NamenodeProtocols {
       throw new IOException("create: Pathname too long.  Limit "
           + MAX_PATH_LENGTH + " characters, " + MAX_PATH_DEPTH + " levels.");
     }
-    namesystem.startFile(src,
-        new PermissionStatus(UserGroupInformation.getCurrentUser().getShortUserName(),
-            null, masked),
-        clientName, clientMachine, flag.get(), createParent, replication, blockSize);
+    HdfsFileStatus fileStatus = namesystem.startFile(src, new PermissionStatus(
+        UserGroupInformation.getCurrentUser().getShortUserName(), null, masked),
+        clientName, clientMachine, flag.get(), createParent, replication,
+        blockSize);
     metrics.incrFilesCreated();
     metrics.incrCreateFileOps();
+    return fileStatus;
   }
 
   @Override // ClientProtocol
@@ -482,26 +480,24 @@ class NameNodeRpcServer implements NamenodeProtocols {
       throws IOException {
     namesystem.setOwner(src, username, groupname);
   }
-
-  @Override // ClientProtocol
-  public LocatedBlock addBlock(String src,
-                               String clientName,
-                               ExtendedBlock previous,
-                               DatanodeInfo[] excludedNodes)
+  
+  @Override
+  public LocatedBlock addBlock(String src, String clientName,
+      ExtendedBlock previous, DatanodeInfo[] excludedNodes, long fileId)
       throws IOException {
-    if(stateChangeLog.isDebugEnabled()) {
-      stateChangeLog.debug("*BLOCK* NameNode.addBlock: file "
-          +src+" for "+clientName);
+    if (stateChangeLog.isDebugEnabled()) {
+      stateChangeLog.debug("*BLOCK* NameNode.addBlock: file " + src
+          + " fileId=" + fileId + " for " + clientName);
     }
     HashMap<Node, Node> excludedNodesSet = null;
     if (excludedNodes != null) {
       excludedNodesSet = new HashMap<Node, Node>(excludedNodes.length);
-      for (Node node:excludedNodes) {
+      for (Node node : excludedNodes) {
         excludedNodesSet.put(node, node);
       }
     }
-    LocatedBlock locatedBlock = 
-      namesystem.getAdditionalBlock(src, clientName, previous, excludedNodesSet);
+    LocatedBlock locatedBlock = namesystem.getAdditionalBlock(src, fileId,
+        clientName, previous, excludedNodesSet);
     if (locatedBlock != null)
       metrics.incrAddBlockOps();
     return locatedBlock;

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NamenodeJspHelper.java

@@ -17,6 +17,8 @@
  */
 package org.apache.hadoop.hdfs.server.namenode;
 
+import static org.apache.hadoop.hdfs.DFSUtil.percent2String;
+
 import java.io.IOException;
 import java.lang.management.ManagementFactory;
 import java.lang.management.MemoryMXBean;
@@ -64,6 +66,14 @@ import org.znerd.xmlenc.XMLOutputter;
 import com.google.common.base.Preconditions;
 
 class NamenodeJspHelper {
+  static String fraction2String(double value) {
+    return StringUtils.format("%.2f", value);
+  }
+
+  static String fraction2String(long numerator, long denominator) {
+    return fraction2String(numerator/(double)denominator);
+  }
+
   static String getSafeModeText(FSNamesystem fsn) {
     if (!fsn.isInSafeMode())
       return "";
@@ -361,20 +371,20 @@ class NamenodeJspHelper {
           + "DFS Remaining" + colTxt() + ":" + colTxt()
           + StringUtils.byteDesc(remaining) + rowTxt() + colTxt() + "DFS Used%"
           + colTxt() + ":" + colTxt()
-          + StringUtils.limitDecimalTo2(percentUsed) + " %" + rowTxt()
+          + percent2String(percentUsed) + rowTxt()
           + colTxt() + "DFS Remaining%" + colTxt() + ":" + colTxt()
-          + StringUtils.limitDecimalTo2(percentRemaining) + " %"
+          + percent2String(percentRemaining)
           + rowTxt() + colTxt() + "Block Pool Used" + colTxt() + ":" + colTxt()
           + StringUtils.byteDesc(bpUsed) + rowTxt()
           + colTxt() + "Block Pool Used%"+ colTxt() + ":" + colTxt()
-          + StringUtils.limitDecimalTo2(percentBpUsed) + " %" 
+          + percent2String(percentBpUsed) 
           + rowTxt() + colTxt() + "DataNodes usages" + colTxt() + ":" + colTxt()
           + "Min %" + colTxt() + "Median %" + colTxt() + "Max %" + colTxt()
           + "stdev %" + rowTxt() + colTxt() + colTxt() + colTxt()
-          + StringUtils.limitDecimalTo2(min) + " %"
-          + colTxt() + StringUtils.limitDecimalTo2(median) + " %"
-          + colTxt() + StringUtils.limitDecimalTo2(max) + " %"
-          + colTxt() + StringUtils.limitDecimalTo2(dev) + " %"
+          + percent2String(min)
+          + colTxt() + percent2String(median)
+          + colTxt() + percent2String(max)
+          + colTxt() + percent2String(dev)
           + rowTxt() + colTxt()
           + "<a href=\"dfsnodelist.jsp?whatNodes=LIVE\">Live Nodes</a> "
           + colTxt() + ":" + colTxt() + live.size()
@@ -562,9 +572,9 @@ class NamenodeJspHelper {
       long u = d.getDfsUsed();
       long nu = d.getNonDfsUsed();
       long r = d.getRemaining();
-      String percentUsed = StringUtils.limitDecimalTo2(d.getDfsUsedPercent());
-      String percentRemaining = StringUtils.limitDecimalTo2(d
-          .getRemainingPercent());
+      final double percentUsedValue = d.getDfsUsedPercent();
+      String percentUsed = fraction2String(percentUsedValue);
+      String percentRemaining = fraction2String(d.getRemainingPercent());
 
       String adminState = d.getAdminState().toString();
 
@@ -572,32 +582,30 @@ class NamenodeJspHelper {
       long currentTime = Time.now();
       
       long bpUsed = d.getBlockPoolUsed();
-      String percentBpUsed = StringUtils.limitDecimalTo2(d
-          .getBlockPoolUsedPercent());
+      String percentBpUsed = fraction2String(d.getBlockPoolUsedPercent());
 
       out.print("<td class=\"lastcontact\"> "
           + ((currentTime - timestamp) / 1000)
           + "<td class=\"adminstate\">"
           + adminState
           + "<td align=\"right\" class=\"capacity\">"
-          + StringUtils.limitDecimalTo2(c * 1.0 / diskBytes)
+          + fraction2String(c, diskBytes)
           + "<td align=\"right\" class=\"used\">"
-          + StringUtils.limitDecimalTo2(u * 1.0 / diskBytes)
+          + fraction2String(u, diskBytes)
           + "<td align=\"right\" class=\"nondfsused\">"
-          + StringUtils.limitDecimalTo2(nu * 1.0 / diskBytes)
+          + fraction2String(nu, diskBytes)
           + "<td align=\"right\" class=\"remaining\">"
-          + StringUtils.limitDecimalTo2(r * 1.0 / diskBytes)
+          + fraction2String(r, diskBytes)
           + "<td align=\"right\" class=\"pcused\">"
           + percentUsed
           + "<td class=\"pcused\">"
-          + ServletUtil.percentageGraph((int) Double.parseDouble(percentUsed),
-              100) 
+          + ServletUtil.percentageGraph((int)percentUsedValue, 100) 
           + "<td align=\"right\" class=\"pcremaining\">"
           + percentRemaining 
           + "<td title=" + "\"blocks scheduled : "
           + d.getBlocksScheduled() + "\" class=\"blocks\">" + d.numBlocks()+"\n"
           + "<td align=\"right\" class=\"bpused\">"
-          + StringUtils.limitDecimalTo2(bpUsed * 1.0 / diskBytes)
+          + fraction2String(bpUsed, diskBytes)
           + "<td align=\"right\" class=\"pcbpused\">"
           + percentBpUsed
           + "<td align=\"right\" class=\"volfails\">"

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/SecondaryNameNode.java

@@ -475,14 +475,20 @@ public class SecondaryNameNode implements Runnable {
     // Returns a token that would be used to upload the merged image.
     CheckpointSignature sig = namenode.rollEditLog();
     
-    if ((checkpointImage.getNamespaceID() == 0) ||
-        (sig.isSameCluster(checkpointImage) &&
+    boolean loadImage = false;
+    boolean isFreshCheckpointer = (checkpointImage.getNamespaceID() == 0);
+    boolean isSameCluster =
+        (dstStorage.versionSupportsFederation() && sig.isSameCluster(checkpointImage)) ||
+        (!dstStorage.versionSupportsFederation() && sig.namespaceIdMatches(checkpointImage));
+    if (isFreshCheckpointer ||
+        (isSameCluster &&
          !sig.storageVersionMatches(checkpointImage.getStorage()))) {
       // if we're a fresh 2NN, or if we're on the same cluster and our storage
       // needs an upgrade, just take the storage info from the server.
       dstStorage.setStorageInfo(sig);
       dstStorage.setClusterID(sig.getClusterID());
       dstStorage.setBlockPoolID(sig.getBlockpoolID());
+      loadImage = true;
     }
     sig.validateStorageInfo(checkpointImage);
 
@@ -492,7 +498,7 @@ public class SecondaryNameNode implements Runnable {
     RemoteEditLogManifest manifest =
       namenode.getEditLogManifest(sig.mostRecentCheckpointTxId + 1);
 
-    boolean loadImage = downloadCheckpointFiles(
+    loadImage |= downloadCheckpointFiles(
         fsName, checkpointImage, sig, manifest);   // Fetch fsimage and edits
     doMerge(sig, manifest, loadImage, checkpointImage, namesystem);
     

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/tools/DFSAdmin.java

@@ -316,8 +316,7 @@ public class DFSAdmin extends FsShell {
       System.out.println("DFS Used: " + used
                          + " (" + StringUtils.byteDesc(used) + ")");
       System.out.println("DFS Used%: "
-                         + StringUtils.limitDecimalTo2(((1.0 * used) / presentCapacity) * 100)
-                         + "%");
+          + StringUtils.formatPercent(used/(double)presentCapacity, 2));
       
       /* These counts are not always upto date. They are updated after  
        * iteration of an internal list. Should be updated in a few seconds to 

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/web/JsonUtil.java

@@ -219,6 +219,7 @@ public class JsonUtil {
     m.put("modificationTime", status.getModificationTime());
     m.put("blockSize", status.getBlockSize());
     m.put("replication", status.getReplication());
+    m.put("fileId", status.getFileId());
     return includeType ? toJsonString(FileStatus.class, m): JSON.toString(m);
   }
 
@@ -243,9 +244,10 @@ public class JsonUtil {
     final long mTime = (Long) m.get("modificationTime");
     final long blockSize = (Long) m.get("blockSize");
     final short replication = (short) (long) (Long) m.get("replication");
+    final long fileId = (Long) m.get("fileId");
     return new HdfsFileStatus(len, type == PathType.DIRECTORY, replication,
         blockSize, mTime, aTime, permission, owner, group,
-        symlink, DFSUtil.string2Bytes(localName));
+        symlink, DFSUtil.string2Bytes(localName), fileId);
   }
 
   /** Convert an ExtendedBlock to a Json map. */

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/web/resources/ConcatSourcesParam.java

@@ -21,7 +21,7 @@ package org.apache.hadoop.hdfs.web.resources;
 /** The concat source paths parameter. */
 public class ConcatSourcesParam extends StringParam {
   /** Parameter name. */
-  public static final String NAME = "srcs";
+  public static final String NAME = "sources";
 
   public static final String DEFAULT = NULL;
 

hadoop-hdfs-project/hadoop-hdfs/src/main/proto/ClientNamenodeProtocol.proto

@@ -67,7 +67,8 @@ message CreateRequestProto {
   required uint64 blockSize = 7;
 }
 
-message CreateResponseProto { // void response
+message CreateResponseProto {
+  optional HdfsFileStatusProto fs = 1;
 }
 
 message AppendRequestProto {
@@ -119,6 +120,7 @@ message AddBlockRequestProto {
   required string clientName = 2;
   optional ExtendedBlockProto previous = 3;
   repeated DatanodeInfoProto excludeNodes = 4;
+  optional uint64 fileId = 5 [default = 0];  // default as a bogus id
 }
 
 message AddBlockResponseProto {

hadoop-hdfs-project/hadoop-hdfs/src/main/proto/hdfs.proto

@@ -170,6 +170,9 @@ message HdfsFileStatusProto {
   optional uint32 block_replication = 10 [default = 0]; // only 16bits used
   optional uint64 blocksize = 11 [default = 0];
   optional LocatedBlocksProto locations = 12;  // suppled only if asked by client
+
+  // Optional field for fileId
+  optional uint64 fileId = 13 [default = 0]; // default as an invalid id
 } 
 
 /**

hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml

@@ -999,17 +999,14 @@
 </property>
 
 <property>
-  <name>dfs.namenode.check.stale.datanode</name>
+  <name>dfs.namenode.avoid.read.stale.datanode</name>
   <value>false</value>
   <description>
-    Indicate whether or not to check "stale" datanodes whose 
+    Indicate whether or not to avoid reading from &quot;stale&quot; datanodes whose
     heartbeat messages have not been received by the namenode 
-    for more than a specified time interval. If this configuration 
-    parameter is set as true, the system will keep track 
-    of the number of stale datanodes. The stale datanodes will be 
+    for more than a specified time interval. Stale datanodes will be
     moved to the end of the node list returned for reading. See
-    dfs.namenode.avoid.write.stale.datanode for details on how this 
-    affects writes. 
+    dfs.namenode.avoid.write.stale.datanode for a similar setting for writes.
   </description>
 </property>
 
@@ -1017,13 +1014,13 @@
   <name>dfs.namenode.avoid.write.stale.datanode</name>
   <value>false</value>
   <description>
-    Indicate whether or not to avoid writing to "stale" datanodes whose 
+    Indicate whether or not to avoid writing to &quot;stale&quot; datanodes whose 
     heartbeat messages have not been received by the namenode 
-    for more than a specified time interval. If this configuration 
-    parameter and dfs.namenode.check.stale.datanode are both set as true, 
-    the writing will avoid using stale datanodes unless a high number 
-    of datanodes are marked as stale. See 
-    dfs.namenode.write.stale.datanode.ratio for details.
+    for more than a specified time interval. Writes will avoid using 
+    stale datanodes unless more than a configured ratio 
+    (dfs.namenode.write.stale.datanode.ratio) of datanodes are marked as 
+    stale. See dfs.namenode.avoid.read.stale.datanode for a similar setting
+    for reads.
   </description>
 </property>
 

hadoop-hdfs-project/hadoop-hdfs/src/site/apt/WebHDFS.apt.vm

@@ -109,6 +109,9 @@ WebHDFS REST API
     * {{{Append to a File}<<<APPEND>>>}}
         (see  {{{../../api/org/apache/hadoop/fs/FileSystem.html}FileSystem}}.append)
 
+    * {{{Concat File(s)}<<<CONCAT>>>}}
+        (see  {{{../../api/org/apache/hadoop/fs/FileSystem.html}FileSystem}}.concat)
+
   * HTTP DELETE
 
     * {{{Delete a File/Directory}<<<DELETE>>>}}
@@ -299,6 +302,32 @@ Content-Length: 0
    {{{../../api/org/apache/hadoop/fs/FileSystem.html}FileSystem}}.append
 
 
+** {Concat File(s)}
+
+  * Submit a HTTP POST request.
+
++---------------------------------
+curl -i -X POST "http://<HOST>:<PORT>/webhdfs/v1/<PATH>?op=CONCAT&sources=<SOURCES>"
++---------------------------------
+
+  The client receives a response with zero content length:
+
++---------------------------------
+HTTP/1.1 200 OK
+Content-Length: 0
++---------------------------------
+
+  []
+
+  This REST API call is available as of Hadoop version 2.0.3.
+  Please note that <SOURCES> is a comma seperated list of absolute paths.
+  (Example: sources=/test/file1,/test/file2,/test/file3)
+
+  See also:
+  {{{Sources}<<<sources>>>}},
+   {{{../../api/org/apache/hadoop/fs/FileSystem.html}FileSystem}}.concat
+
+
 ** {Open and Read a File}
 
   * Submit a HTTP GET request with automatically following redirects.
@@ -1727,6 +1756,29 @@ var tokenProperties =
   {{{Set Replication Factor}<<<SETREPLICATION>>>}}
 
 
+** {Sources}
+
+*----------------+-------------------------------------------------------------------+
+|| Name          | <<<sources>>> |
+*----------------+-------------------------------------------------------------------+
+|| Description   | The comma seperated absolute paths used for concatenation. |
+*----------------+-------------------------------------------------------------------+
+|| Type          | String |
+*----------------+-------------------------------------------------------------------+
+|| Default Value | \<empty\> |
+*----------------+-------------------------------------------------------------------+
+|| Valid Values  | A list of comma seperated absolute FileSystem paths without scheme and authority. |
+*----------------+-------------------------------------------------------------------+
+|| Syntax        | See the note in {{Delegation}}. |
+*----------------+-------------------------------------------------------------------+
+
+  <<Note>> that sources are absolute FileSystem paths.
+
+
+  See also:
+  {{{Concat File(s)}<<<CONCAT>>>}}
+
+
 ** {Token}
 
 *----------------+-------------------------------------------------------------------+

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/DFSTestUtil.java

@@ -33,6 +33,7 @@ import java.io.FileInputStream;
 import java.io.FileReader;
 import java.io.IOException;
 import java.io.InputStream;
+import java.net.HttpURLConnection;
 import java.net.InetSocketAddress;
 import java.net.Socket;
 import java.net.URL;
@@ -620,6 +621,9 @@ public class DFSTestUtil {
    */
   public static byte[] urlGetBytes(URL url) throws IOException {
     URLConnection conn = url.openConnection();
+    HttpURLConnection hc = (HttpURLConnection)conn;
+    
+    assertEquals(HttpURLConnection.HTTP_OK, hc.getResponseCode());
     ByteArrayOutputStream out = new ByteArrayOutputStream();
     IOUtils.copyBytes(conn.getInputStream(), out, 4096, true);
     return out.toByteArray();

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientFailover.java

@@ -23,22 +23,34 @@ import static org.junit.Assert.assertTrue;
 import static org.junit.Assert.fail;
 
 import java.io.IOException;
+import java.net.InetSocketAddress;
+import java.net.Socket;
+import java.net.SocketAddress;
 import java.net.URI;
 import java.net.URISyntaxException;
 
+import javax.net.SocketFactory;
+
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.CommonConfigurationKeysPublic;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.Path;
 import org.apache.hadoop.hdfs.server.namenode.NameNode;
 import org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider;
 import org.apache.hadoop.hdfs.server.namenode.ha.HATestUtil;
+import org.apache.hadoop.io.IOUtils;
+import org.apache.hadoop.net.ConnectTimeoutException;
+import org.apache.hadoop.net.StandardSocketFactory;
 import org.apache.hadoop.test.GenericTestUtils;
 import org.apache.hadoop.util.StringUtils;
+import org.hamcrest.BaseMatcher;
+import org.hamcrest.Description;
 import org.junit.After;
 import org.junit.Before;
 import org.junit.Test;
+import org.mockito.Mockito;
 
 public class TestDFSClientFailover {
   
@@ -91,6 +103,63 @@ public class TestDFSClientFailover {
     fs.close();
   }
   
+  /**
+   * Test that even a non-idempotent method will properly fail-over if the
+   * first IPC attempt times out trying to connect. Regression test for
+   * HDFS-4404. 
+   */
+  @Test
+  public void testFailoverOnConnectTimeout() throws Exception {
+    conf.setClass(CommonConfigurationKeysPublic.HADOOP_RPC_SOCKET_FACTORY_CLASS_DEFAULT_KEY,
+        InjectingSocketFactory.class, SocketFactory.class);
+    // Set up the InjectingSocketFactory to throw a ConnectTimeoutException
+    // when connecting to the first NN.
+    InjectingSocketFactory.portToInjectOn = cluster.getNameNodePort(0);
+
+    FileSystem fs = HATestUtil.configureFailoverFs(cluster, conf);
+    
+    // Make the second NN the active one.
+    cluster.shutdownNameNode(0);
+    cluster.transitionToActive(1);
+    
+    // Call a non-idempotent method, and ensure the failover of the call proceeds
+    // successfully.
+    IOUtils.closeStream(fs.create(TEST_FILE));
+  }
+  
+  private static class InjectingSocketFactory extends StandardSocketFactory {
+
+    static SocketFactory defaultFactory = SocketFactory.getDefault();
+
+    static int portToInjectOn;
+    
+    @Override
+    public Socket createSocket() throws IOException {
+      Socket spy = Mockito.spy(defaultFactory.createSocket());
+      // Simplify our spying job by not having to also spy on the channel
+      Mockito.doReturn(null).when(spy).getChannel();
+      // Throw a ConnectTimeoutException when connecting to our target "bad"
+      // host.
+      Mockito.doThrow(new ConnectTimeoutException("injected"))
+        .when(spy).connect(
+            Mockito.argThat(new MatchesPort()),
+            Mockito.anyInt());
+      return spy;
+    }
+
+    private class MatchesPort extends BaseMatcher<SocketAddress> {
+      @Override
+      public boolean matches(Object arg0) {
+        return ((InetSocketAddress)arg0).getPort() == portToInjectOn;
+      }
+
+      @Override
+      public void describeTo(Description desc) {
+        desc.appendText("matches port " + portToInjectOn);
+      }
+    }
+  }
+  
   /**
    * Regression test for HDFS-2683.
    */

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSClientRetries.java

@@ -23,7 +23,10 @@ import static org.junit.Assert.assertFalse;
 import static org.junit.Assert.assertTrue;
 import static org.junit.Assert.fail;
 import static org.mockito.Matchers.any;
+import static org.mockito.Matchers.anyBoolean;
 import static org.mockito.Matchers.anyLong;
+import static org.mockito.Matchers.anyObject;
+import static org.mockito.Matchers.anyShort;
 import static org.mockito.Matchers.anyString;
 import static org.mockito.Mockito.doAnswer;
 import static org.mockito.Mockito.mock;
@@ -49,13 +52,13 @@ import org.apache.commons.logging.LogFactory;
 import org.apache.commons.logging.impl.Log4JLogger;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.CommonConfigurationKeys;
+import org.apache.hadoop.fs.CreateFlag;
 import org.apache.hadoop.fs.FSDataInputStream;
 import org.apache.hadoop.fs.FSDataOutputStream;
 import org.apache.hadoop.fs.FileChecksum;
 import org.apache.hadoop.fs.FileStatus;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.Path;
-import org.apache.hadoop.fs.UnresolvedLinkException;
 import org.apache.hadoop.fs.permission.FsPermission;
 import org.apache.hadoop.hdfs.client.HdfsUtils;
 import org.apache.hadoop.hdfs.protocol.Block;
@@ -64,12 +67,14 @@ import org.apache.hadoop.hdfs.protocol.DatanodeID;
 import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
 import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
 import org.apache.hadoop.hdfs.protocol.HdfsConstants.SafeModeAction;
+import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
 import org.apache.hadoop.hdfs.protocol.LocatedBlock;
 import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
 import org.apache.hadoop.hdfs.server.namenode.NameNode;
 import org.apache.hadoop.hdfs.server.namenode.NotReplicatedYetException;
 import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
 import org.apache.hadoop.hdfs.web.WebHdfsTestUtil;
+import org.apache.hadoop.io.EnumSetWritable;
 import org.apache.hadoop.io.IOUtils;
 import org.apache.hadoop.io.LongWritable;
 import org.apache.hadoop.io.Writable;
@@ -208,7 +213,7 @@ public class TestDFSClientRetries {
    * Verify that client will correctly give up after the specified number
    * of times trying to add a block
    */
-  @SuppressWarnings("serial")
+  @SuppressWarnings({ "serial", "unchecked" })
   @Test
   public void testNotYetReplicatedErrors() throws IOException
   { 
@@ -235,7 +240,22 @@ public class TestDFSClientRetries {
     when(mockNN.addBlock(anyString(), 
                          anyString(),
                          any(ExtendedBlock.class),
-                         any(DatanodeInfo[].class))).thenAnswer(answer);
+                         any(DatanodeInfo[].class),
+                         anyLong())).thenAnswer(answer);
+    
+    Mockito.doReturn(
+            new HdfsFileStatus(0, false, 1, 1024, 0, 0, new FsPermission(
+                (short) 777), "owner", "group", new byte[0], new byte[0],
+                1010)).when(mockNN).getFileInfo(anyString());
+    
+    Mockito.doReturn(
+            new HdfsFileStatus(0, false, 1, 1024, 0, 0, new FsPermission(
+                (short) 777), "owner", "group", new byte[0], new byte[0],
+                1010))
+        .when(mockNN)
+        .create(anyString(), (FsPermission) anyObject(), anyString(),
+            (EnumSetWritable<CreateFlag>) anyObject(), anyBoolean(),
+            anyShort(), anyLong());
 
     final DFSClient client = new DFSClient(null, mockNN, conf, null);
     OutputStream os = client.create("testfile", true);
@@ -369,7 +389,8 @@ public class TestDFSClientRetries {
           return ret2;
         }
       }).when(spyNN).addBlock(Mockito.anyString(), Mockito.anyString(),
-          Mockito.<ExtendedBlock>any(), Mockito.<DatanodeInfo[]>any());
+          Mockito.<ExtendedBlock> any(), Mockito.<DatanodeInfo[]> any(),
+          Mockito.anyLong());
 
       doAnswer(new Answer<Boolean>() {
 
@@ -410,7 +431,8 @@ public class TestDFSClientRetries {
       // Make sure the mock was actually properly injected.
       Mockito.verify(spyNN, Mockito.atLeastOnce()).addBlock(
           Mockito.anyString(), Mockito.anyString(),
-          Mockito.<ExtendedBlock>any(), Mockito.<DatanodeInfo[]>any());
+          Mockito.<ExtendedBlock> any(), Mockito.<DatanodeInfo[]> any(),
+          Mockito.anyLong());
       Mockito.verify(spyNN, Mockito.atLeastOnce()).complete(
           Mockito.anyString(), Mockito.anyString(),
           Mockito.<ExtendedBlock>any());

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestDFSUtil.java

@@ -619,6 +619,16 @@ public class TestDFSUtil {
     
     assertEquals(1, uris.size());
     assertTrue(uris.contains(new URI("hdfs://" + NN1_SRVC_ADDR)));
+
+    // Make sure when config FS_DEFAULT_NAME_KEY using IP address,
+    // it will automatically convert it to hostname
+    conf = new HdfsConfiguration();
+    conf.set(CommonConfigurationKeys.FS_DEFAULT_NAME_KEY, "hdfs://127.0.0.1:8020");
+    uris = DFSUtil.getNameServiceUris(conf);
+    assertEquals(1, uris.size());
+    for (URI uri : uris) {
+      assertFalse(uri.getHost().equals("127.0.0.1"));
+    }
   }
   
   @Test

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestFileCreation.java

@@ -71,6 +71,7 @@ import org.apache.hadoop.hdfs.server.datanode.DataNodeTestUtils;
 import org.apache.hadoop.hdfs.server.datanode.SimulatedFSDataset;
 import org.apache.hadoop.hdfs.server.datanode.fsdataset.FsDatasetSpi;
 import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
+import org.apache.hadoop.hdfs.server.namenode.INodeId;
 import org.apache.hadoop.hdfs.server.namenode.LeaseManager;
 import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
 import org.apache.hadoop.io.EnumSetWritable;
@@ -517,8 +518,8 @@ public class TestFileCreation {
           + "The file has " + locations.locatedBlockCount() + " blocks.");
 
       // add one block to the file
-      LocatedBlock location = client.getNamenode().addBlock(file1.toString(), 
-          client.clientName, null, null);
+      LocatedBlock location = client.getNamenode().addBlock(file1.toString(),
+          client.clientName, null, null, INodeId.GRANDFATHER_INODE_ID);
       System.out.println("testFileCreationError2: "
           + "Added block " + location.getBlock());
 
@@ -568,8 +569,8 @@ public class TestFileCreation {
       final Path f = new Path("/foo.txt");
       createFile(dfs, f, 3);
       try {
-        cluster.getNameNodeRpc().addBlock(f.toString(), 
-            client.clientName, null, null);
+        cluster.getNameNodeRpc().addBlock(f.toString(), client.clientName,
+            null, null, INodeId.GRANDFATHER_INODE_ID);
         fail();
       } catch(IOException ioe) {
         FileSystem.LOG.info("GOOD!", ioe);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestFileLengthOnClusterRestart.java

@@ -43,7 +43,7 @@ public class TestFileLengthOnClusterRestart {
         .numDataNodes(2).build();
     HdfsDataInputStream in = null;
     try {
-      Path path = new Path(MiniDFSCluster.getBaseDirectory(), "test");
+      Path path = new Path("/tmp/TestFileLengthOnClusterRestart", "test");
       DistributedFileSystem dfs = (DistributedFileSystem) cluster
           .getFileSystem();
       FSDataOutputStream out = dfs.create(path);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestGetBlocks.java

@@ -88,7 +88,7 @@ public class TestGetBlocks {
   @Test
   public void testReadSelectNonStaleDatanode() throws Exception {
     HdfsConfiguration conf = new HdfsConfiguration();
-    conf.setBoolean(DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY, true);
+    conf.setBoolean(DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_KEY, true);
     long staleInterval = 30 * 1000 * 60;
     conf.setLong(DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_KEY,
         staleInterval);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestLargeBlock.java

@@ -183,8 +183,7 @@ public class TestLargeBlock {
     try {
 
       // create a new file in test data directory
-      Path file1 = new Path(System.getProperty("test.build.data") + "/" +
-          Long.toString(blockSize) + ".dat");
+      Path file1 = new Path("/tmp/TestLargeBlock", blockSize + ".dat");
       FSDataOutputStream stm = createFile(fs, file1, 1, blockSize);
       LOG.info("File " + file1 + " created with file size " +
           fileSize +

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestLease.java

@@ -18,6 +18,10 @@
 package org.apache.hadoop.hdfs;
 
 import static org.mockito.Matchers.anyString;
+import static org.mockito.Matchers.anyShort;
+import static org.mockito.Matchers.anyLong;
+import static org.mockito.Matchers.anyBoolean;
+import static org.mockito.Matchers.anyObject;
 import static org.mockito.Mockito.doNothing;
 import static org.mockito.Mockito.doThrow;
 import static org.mockito.Mockito.spy;
@@ -29,14 +33,18 @@ import java.security.PrivilegedExceptionAction;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.CreateFlag;
 import org.apache.hadoop.fs.FSDataOutputStream;
 import org.apache.hadoop.fs.FileStatus;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.Options;
 import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.fs.permission.FsPermission;
 import org.apache.hadoop.hdfs.protocol.ClientProtocol;
+import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
 import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
 import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
+import org.apache.hadoop.io.EnumSetWritable;
 import org.apache.hadoop.ipc.RemoteException;
 import org.apache.hadoop.security.UserGroupInformation;
 import org.apache.hadoop.security.token.SecretManager.InvalidToken;
@@ -256,6 +264,7 @@ public class TestLease {
     }
   }
 
+  @SuppressWarnings("unchecked")
   @Test
   public void testFactory() throws Exception {
     final String[] groups = new String[]{"supergroup"};
@@ -264,6 +273,20 @@ public class TestLease {
       ugi[i] = UserGroupInformation.createUserForTesting("user" + i, groups);
     }
 
+    Mockito.doReturn(
+        new HdfsFileStatus(0, false, 1, 1024, 0, 0, new FsPermission(
+            (short) 777), "owner", "group", new byte[0], new byte[0],
+            1010)).when(mcp).getFileInfo(anyString());
+    Mockito
+        .doReturn(
+            new HdfsFileStatus(0, false, 1, 1024, 0, 0, new FsPermission(
+                (short) 777), "owner", "group", new byte[0], new byte[0],
+                1010))
+        .when(mcp)
+        .create(anyString(), (FsPermission) anyObject(), anyString(),
+            (EnumSetWritable<CreateFlag>) anyObject(), anyBoolean(),
+            anyShort(), anyLong());
+
     final Configuration conf = new Configuration();
     final DFSClient c1 = createDFSClientAs(ugi[0], conf);
     FSDataOutputStream out1 = createFsOut(c1, "/out1");

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestListFilesInDFS.java

@@ -38,6 +38,7 @@ public class TestListFilesInDFS extends TestListFiles {
 
   @BeforeClass
   public static void testSetUp() throws Exception {
+    setTestPaths(new Path("/tmp/TestListFilesInDFS"));
     cluster = new MiniDFSCluster.Builder(conf).build();
     fs = cluster.getFileSystem();
     fs.delete(TEST_DIR, true);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/TestQuota.java

@@ -70,8 +70,8 @@ public class TestQuota {
       throw new DSQuotaExceededException(bytes, bytes);
     } catch(DSQuotaExceededException e) {
       
-      assertEquals("The DiskSpace quota is exceeded: quota=1.0k " +
-          "diskspace consumed=1.0k", e.getMessage());
+      assertEquals("The DiskSpace quota is exceeded: quota = 1024 B = 1 KB"
+          + " but diskspace consumed = 1024 B = 1 KB", e.getMessage());
     }
   }
   

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/blockmanagement/TestRBWBlockInvalidation.java

@@ -67,7 +67,7 @@ public class TestRBWBlockInvalidation {
     try {
       final FSNamesystem namesystem = cluster.getNamesystem();
       FileSystem fs = cluster.getFileSystem();
-      Path testPath = new Path(MiniDFSCluster.getBaseDirectory(), "foo1");
+      Path testPath = new Path("/tmp/TestRBWBlockInvalidation", "foo1");
       out = fs.create(testPath, (short) 2);
       out.writeBytes("HDFS-3157: " + testPath);
       out.hsync();

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/blockmanagement/TestReplicationPolicy.java

@@ -88,9 +88,11 @@ public class TestReplicationPolicy {
         "test.build.data", "build/test/data"), "dfs/");
     conf.set(DFSConfigKeys.DFS_NAMENODE_NAME_DIR_KEY,
         new File(baseDir, "name").getPath());
-    // Enable the checking for stale datanodes in the beginning
-    conf.setBoolean(DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY, true);
 
+    conf.setBoolean(
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_KEY, true);
+    conf.setBoolean(
+        DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY, true);
     DFSTestUtil.formatNameNode(conf);
     namenode = new NameNode(conf);
 
@@ -100,6 +102,8 @@ public class TestReplicationPolicy {
     // construct network topology
     for (int i=0; i < NUM_OF_DATANODES; i++) {
       cluster.add(dataNodes[i]);
+      bm.getDatanodeManager().getHeartbeatManager().addDatanode(
+          dataNodes[i]);
     }
     for (int i=0; i < NUM_OF_DATANODES; i++) {
       dataNodes[i].updateHeartbeat(
@@ -393,11 +397,11 @@ public class TestReplicationPolicy {
       throws Exception {
     try {
       namenode.getNamesystem().getBlockManager().getDatanodeManager()
-        .setAvoidStaleDataNodesForWrite(true);
+        .setNumStaleNodes(NUM_OF_DATANODES);
       testChooseTargetWithMoreThanAvailableNodes();
     } finally {
       namenode.getNamesystem().getBlockManager().getDatanodeManager()
-      .setAvoidStaleDataNodesForWrite(false);
+        .setNumStaleNodes(0);
     }
   }
   
@@ -479,12 +483,12 @@ public class TestReplicationPolicy {
   
   @Test
   public void testChooseTargetWithStaleNodes() throws Exception {
-    // Enable avoidng writing to stale datanodes
-    namenode.getNamesystem().getBlockManager().getDatanodeManager()
-        .setAvoidStaleDataNodesForWrite(true);
     // Set dataNodes[0] as stale
     dataNodes[0].setLastUpdate(Time.now() - staleInterval - 1);
-
+    namenode.getNamesystem().getBlockManager()
+      .getDatanodeManager().getHeartbeatManager().heartbeatCheck();
+    assertTrue(namenode.getNamesystem().getBlockManager()
+        .getDatanodeManager().shouldAvoidStaleDataNodesForWrite());
     DatanodeDescriptor[] targets;
     // We set the datanode[0] as stale, thus should choose datanode[1] since
     // datanode[1] is on the same rack with datanode[0] (writer)
@@ -503,9 +507,9 @@ public class TestReplicationPolicy {
     assertFalse(cluster.isOnSameRack(targets[0], dataNodes[0]));
     
     // reset
-    namenode.getNamesystem().getBlockManager().getDatanodeManager()
-        .setAvoidStaleDataNodesForWrite(false);
     dataNodes[0].setLastUpdate(Time.now());
+    namenode.getNamesystem().getBlockManager()
+      .getDatanodeManager().getHeartbeatManager().heartbeatCheck();
   }
 
   /**
@@ -518,20 +522,20 @@ public class TestReplicationPolicy {
    */
   @Test
   public void testChooseTargetWithHalfStaleNodes() throws Exception {
-    // Enable stale datanodes checking
-    namenode.getNamesystem().getBlockManager().getDatanodeManager()
-        .setAvoidStaleDataNodesForWrite(true);
     // Set dataNodes[0], dataNodes[1], and dataNodes[2] as stale
     for (int i = 0; i < 3; i++) {
       dataNodes[i].setLastUpdate(Time.now() - staleInterval - 1);
     }
+    namenode.getNamesystem().getBlockManager()
+      .getDatanodeManager().getHeartbeatManager().heartbeatCheck();
 
     DatanodeDescriptor[] targets;
     targets = replicator.chooseTarget(filename, 0, dataNodes[0],
         new ArrayList<DatanodeDescriptor>(), BLOCK_SIZE);
     assertEquals(targets.length, 0);
 
-    // We set the datanode[0] as stale, thus should choose datanode[1]
+    // Since we have 6 datanodes total, stale nodes should
+    // not be returned until we ask for more than 3 targets
     targets = replicator.chooseTarget(filename, 1, dataNodes[0],
         new ArrayList<DatanodeDescriptor>(), BLOCK_SIZE);
     assertEquals(targets.length, 1);
@@ -557,18 +561,16 @@ public class TestReplicationPolicy {
     assertTrue(containsWithinRange(dataNodes[4], targets, 0, 3));
     assertTrue(containsWithinRange(dataNodes[5], targets, 0, 3));
 
-    // reset
-    namenode.getNamesystem().getBlockManager().getDatanodeManager()
-        .setAvoidStaleDataNodesForWrite(false);
     for (int i = 0; i < dataNodes.length; i++) {
       dataNodes[i].setLastUpdate(Time.now());
     }
+    namenode.getNamesystem().getBlockManager()
+      .getDatanodeManager().getHeartbeatManager().heartbeatCheck();
   }
 
   @Test
   public void testChooseTargetWithMoreThanHalfStaleNodes() throws Exception {
     HdfsConfiguration conf = new HdfsConfiguration();
-    conf.setBoolean(DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY, true);
     conf.setBoolean(
         DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY, true);
     String[] hosts = new String[]{"host1", "host2", "host3", 
@@ -598,7 +600,7 @@ public class TestReplicationPolicy {
           .getBlockManager().getDatanodeManager().getNumStaleNodes();
       assertEquals(numStaleNodes, 2);
       assertTrue(miniCluster.getNameNode().getNamesystem().getBlockManager()
-          .getDatanodeManager().isAvoidingStaleDataNodesForWrite());
+          .getDatanodeManager().shouldAvoidStaleDataNodesForWrite());
       // Call chooseTarget
       DatanodeDescriptor staleNodeInfo = miniCluster.getNameNode()
           .getNamesystem().getBlockManager().getDatanodeManager()
@@ -627,7 +629,7 @@ public class TestReplicationPolicy {
       // According to our strategy, stale datanodes will be included for writing
       // to avoid hotspots
       assertFalse(miniCluster.getNameNode().getNamesystem().getBlockManager()
-          .getDatanodeManager().isAvoidingStaleDataNodesForWrite());     
+          .getDatanodeManager().shouldAvoidStaleDataNodesForWrite());
       // Call chooseTarget
       targets = replicator.chooseTarget(filename, 3,
           staleNodeInfo, new ArrayList<DatanodeDescriptor>(), BLOCK_SIZE);
@@ -650,7 +652,7 @@ public class TestReplicationPolicy {
           .getBlockManager().getDatanodeManager().getNumStaleNodes();
       assertEquals(numStaleNodes, 2);
       assertTrue(miniCluster.getNameNode().getNamesystem().getBlockManager()
-          .getDatanodeManager().isAvoidingStaleDataNodesForWrite());
+          .getDatanodeManager().shouldAvoidStaleDataNodesForWrite());
       // Call chooseTarget
       targets = replicator.chooseTarget(filename, 3,
           staleNodeInfo, new ArrayList<DatanodeDescriptor>(), BLOCK_SIZE);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/FSImageTestUtil.java

@@ -506,7 +506,11 @@ public abstract class FSImageTestUtil {
       props.load(fis);
       IOUtils.closeStream(fis);
   
-      props.setProperty(key, value);
+      if (value == null || value.isEmpty()) {
+        props.remove(key);
+      } else {
+        props.setProperty(key, value);
+      }
       
       out = new FileOutputStream(versionFile);
       props.store(out, null);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/NNThroughputBenchmark.java

@@ -1058,7 +1058,8 @@ public class NNThroughputBenchmark {
     throws IOException {
       ExtendedBlock prevBlock = null;
       for(int jdx = 0; jdx < blocksPerFile; jdx++) {
-        LocatedBlock loc = nameNodeProto.addBlock(fileName, clientName, prevBlock, null);
+        LocatedBlock loc = nameNodeProto.addBlock(fileName, clientName,
+            prevBlock, null, INodeId.GRANDFATHER_INODE_ID);
         prevBlock = loc.getBlock();
         for(DatanodeInfo dnInfo : loc.getLocations()) {
           int dnIdx = Arrays.binarySearch(datanodes, dnInfo.getXferAddr());

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestAddBlockRetry.java

@@ -0,0 +1,141 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.hdfs.server.namenode;
+
+
+import static org.junit.Assert.assertEquals;
+import static org.mockito.Mockito.doAnswer;
+import static org.mockito.Mockito.spy;
+
+import java.lang.reflect.Field;
+import java.util.EnumSet;
+import java.util.HashMap;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.CreateFlag;
+import org.apache.hadoop.fs.permission.FsPermission;
+import org.apache.hadoop.hdfs.MiniDFSCluster;
+import org.apache.hadoop.hdfs.protocol.LocatedBlock;
+import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
+import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
+import org.apache.hadoop.hdfs.server.blockmanagement.DatanodeDescriptor;
+import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
+import org.apache.hadoop.io.EnumSetWritable;
+import org.apache.hadoop.net.Node;
+import org.junit.After;
+import org.junit.Before;
+import org.junit.Test;
+import org.mockito.Mockito;
+import org.mockito.invocation.InvocationOnMock;
+import org.mockito.stubbing.Answer;
+
+/**
+ * Race between two threads simultaneously calling
+ * FSNamesystem.getAdditionalBlock().
+ */
+public class TestAddBlockRetry {
+  public static final Log LOG = LogFactory.getLog(TestAddBlockRetry.class);
+
+  private static final short REPLICATION = 3;
+
+  private Configuration conf;
+  private MiniDFSCluster cluster;
+
+  private int count = 0;
+  private LocatedBlock lb1;
+  private LocatedBlock lb2;
+
+  @Before
+  public void setUp() throws Exception {
+    conf = new Configuration();
+    cluster = new MiniDFSCluster.Builder(conf)
+      .numDataNodes(REPLICATION)
+      .build();
+    cluster.waitActive();
+  }
+
+  @After
+  public void tearDown() throws Exception {
+    if (cluster != null) {
+      cluster.shutdown();
+    }
+  }
+
+  /**
+   * Retry addBlock() while another thread is in chooseTarget().
+   * See HDFS-4452.
+   */
+  @Test
+  public void testRetryAddBlockWhileInChooseTarget() throws Exception {
+    final String src = "/testRetryAddBlockWhileInChooseTarget";
+
+    FSNamesystem ns = cluster.getNamesystem();
+    BlockManager spyBM = spy(ns.getBlockManager());
+    final NamenodeProtocols nn = cluster.getNameNodeRpc();
+
+    // substitute mocked BlockManager into FSNamesystem
+    Class<? extends FSNamesystem> nsClass = ns.getClass();
+    Field bmField = nsClass.getDeclaredField("blockManager");
+    bmField.setAccessible(true);
+    bmField.set(ns, spyBM);
+
+    doAnswer(new Answer<DatanodeDescriptor[]>() {
+      @Override
+      public DatanodeDescriptor[] answer(InvocationOnMock invocation)
+          throws Throwable {
+        LOG.info("chooseTarget for " + src);
+        DatanodeDescriptor[] ret =
+            (DatanodeDescriptor[]) invocation.callRealMethod();
+        count++;
+        if(count == 1) { // run second addBlock()
+          LOG.info("Starting second addBlock for " + src);
+          nn.addBlock(src, "clientName", null, null,
+              INodeId.GRANDFATHER_INODE_ID);
+          LocatedBlocks lbs = nn.getBlockLocations(src, 0, Long.MAX_VALUE);
+          assertEquals("Must be one block", 1, lbs.getLocatedBlocks().size());
+          lb2 = lbs.get(0);
+          assertEquals("Wrong replication",
+              REPLICATION, lb2.getLocations().length);
+        }
+        return ret;
+      }
+    }).when(spyBM).chooseTarget(Mockito.anyString(), Mockito.anyInt(),
+        Mockito.<DatanodeDescriptor>any(), Mockito.<HashMap<Node, Node>>any(),
+        Mockito.anyLong());
+
+    // create file
+    nn.create(src, FsPermission.getFileDefault(),
+        "clientName",
+        new EnumSetWritable<CreateFlag>(EnumSet.of(CreateFlag.CREATE)),
+        true, (short)3, 1024);
+
+    // start first addBlock()
+    LOG.info("Starting first addBlock for " + src);
+    nn.addBlock(src, "clientName", null, null, INodeId.GRANDFATHER_INODE_ID);
+
+    // check locations
+    LocatedBlocks lbs = nn.getBlockLocations(src, 0, Long.MAX_VALUE);
+    assertEquals("Must be one block", 1, lbs.getLocatedBlocks().size());
+    lb1 = lbs.get(0);
+    assertEquals("Wrong replication", REPLICATION, lb1.getLocations().length);
+    assertEquals("Blocks are not equal", lb1.getBlock(), lb2.getBlock());
+  }
+}

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestHostsFiles.java

@@ -0,0 +1,134 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.hdfs.server.namenode;
+
+import static org.junit.Assert.assertTrue;
+
+import java.net.InetSocketAddress;
+import java.net.URL;
+
+import org.apache.commons.lang.StringEscapeUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.BlockLocation;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.hdfs.DFSConfigKeys;
+import org.apache.hadoop.hdfs.DFSTestUtil;
+import org.apache.hadoop.hdfs.HdfsConfiguration;
+import org.apache.hadoop.hdfs.MiniDFSCluster;
+import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
+import org.junit.Test;
+
+/**
+ * DFS_HOSTS and DFS_HOSTS_EXCLUDE tests
+ * 
+ */
+public class TestHostsFiles {
+  private static final Log LOG =
+    LogFactory.getLog(TestHostsFiles.class.getName());
+
+  /*
+   * Return a configuration object with low timeouts for testing and 
+   * a topology script set (which enables rack awareness).  
+   */
+  private Configuration getConf() {
+    Configuration conf = new HdfsConfiguration();
+
+    // Lower the heart beat interval so the NN quickly learns of dead
+    // or decommissioned DNs and the NN issues replication and invalidation
+    // commands quickly (as replies to heartbeats)
+    conf.setLong(DFSConfigKeys.DFS_HEARTBEAT_INTERVAL_KEY, 1L);
+
+    // Have the NN ReplicationMonitor compute the replication and
+    // invalidation commands to send DNs every second.
+    conf.setInt(DFSConfigKeys.DFS_NAMENODE_REPLICATION_INTERVAL_KEY, 1);
+
+    // Have the NN check for pending replications every second so it
+    // quickly schedules additional replicas as they are identified.
+    conf.setInt(DFSConfigKeys.DFS_NAMENODE_REPLICATION_PENDING_TIMEOUT_SEC_KEY, 1);
+
+    // The DNs report blocks every second.
+    conf.setLong(DFSConfigKeys.DFS_BLOCKREPORT_INTERVAL_MSEC_KEY, 1000L);
+
+    // Indicates we have multiple racks
+    conf.set(DFSConfigKeys.NET_TOPOLOGY_SCRIPT_FILE_NAME_KEY, "xyz");
+    return conf;
+  }
+
+  @Test
+  public void testHostsExcludeDfshealthJsp() throws Exception {
+    Configuration conf = getConf();
+    short REPLICATION_FACTOR = 2;
+    final Path filePath = new Path("/testFile");
+
+    // Configure an excludes file
+    FileSystem localFileSys = FileSystem.getLocal(conf);
+    Path workingDir = localFileSys.getWorkingDirectory();
+    Path dir = new Path(workingDir, "build/test/data/temp/decommission");
+    Path excludeFile = new Path(dir, "exclude");
+    Path includeFile = new Path(dir, "include");
+    assertTrue(localFileSys.mkdirs(dir));
+    DFSTestUtil.writeFile(localFileSys, excludeFile, "");
+    DFSTestUtil.writeFile(localFileSys, includeFile, "");
+    conf.set(DFSConfigKeys.DFS_HOSTS_EXCLUDE, excludeFile.toUri().getPath());
+    conf.set(DFSConfigKeys.DFS_HOSTS, includeFile.toUri().getPath());
+
+    // Two blocks and four racks
+    String racks[] = {"/rack1", "/rack1", "/rack2", "/rack2"};
+    MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf)
+      .numDataNodes(racks.length).racks(racks).build();
+    final FSNamesystem ns = cluster.getNameNode().getNamesystem();
+
+    try {
+      // Create a file with one block
+      final FileSystem fs = cluster.getFileSystem();
+      DFSTestUtil.createFile(fs, filePath, 1L, REPLICATION_FACTOR, 1L);
+      ExtendedBlock b = DFSTestUtil.getFirstBlock(fs, filePath);
+      DFSTestUtil.waitForReplication(cluster, b, 2, REPLICATION_FACTOR, 0);
+
+      // Decommission one of the hosts with the block, this should cause 
+      // the block to get replicated to another host on the same rack,
+      // otherwise the rack policy is violated.
+      BlockLocation locs[] = fs.getFileBlockLocations(
+          fs.getFileStatus(filePath), 0, Long.MAX_VALUE);
+      String name = locs[0].getNames()[0];
+      String names = name + "\n" + "localhost:42\n";
+      LOG.info("adding '" + names + "' to exclude file " + excludeFile.toUri().getPath());
+      DFSTestUtil.writeFile(localFileSys, excludeFile, name);
+      ns.getBlockManager().getDatanodeManager().refreshNodes(conf);
+      DFSTestUtil.waitForDecommission(fs, name);
+
+      // Check the block still has sufficient # replicas across racks
+      DFSTestUtil.waitForReplication(cluster, b, 2, REPLICATION_FACTOR, 0);
+      
+      InetSocketAddress nnHttpAddress = cluster.getNameNode().getHttpAddress();
+      LOG.info("nnaddr = '" + nnHttpAddress + "'");
+      URL nnjsp = new URL("http://" + nnHttpAddress.getHostName() + ":" + nnHttpAddress.getPort() + "/dfshealth.jsp");
+      LOG.info("fetching " + nnjsp);
+      String dfshealthPage = StringEscapeUtils.unescapeHtml(DFSTestUtil.urlGet(nnjsp));
+      LOG.info("got " + dfshealthPage);
+      assertTrue("dfshealth should contain localhost, got:" + dfshealthPage,
+          dfshealthPage.contains("localhost"));
+
+    } finally {
+      cluster.shutdown();
+    }
+  }
+}

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestINodeFile.java

@@ -24,8 +24,10 @@ import static org.junit.Assert.fail;
 
 import java.io.FileNotFoundException;
 import java.io.IOException;
+import java.util.EnumSet;
 
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.CreateFlag;
 import org.apache.hadoop.fs.FSDataOutputStream;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.Options;
@@ -39,6 +41,8 @@ import org.apache.hadoop.hdfs.DistributedFileSystem;
 import org.apache.hadoop.hdfs.MiniDFSCluster;
 import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
 import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
+import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
+import org.apache.hadoop.io.EnumSetWritable;
 import org.junit.Test;
 
 public class TestINodeFile {
@@ -376,7 +380,7 @@ public class TestINodeFile {
    * @throws IOException
    */
   @Test
-  public void TestInodeId() throws IOException {
+  public void testInodeId() throws IOException {
 
     Configuration conf = new Configuration();
     conf.setInt(DFSConfigKeys.DFS_BLOCK_SIZE_KEY,
@@ -396,9 +400,14 @@ public class TestINodeFile {
     assertTrue(fs.mkdirs(path));
     assertTrue(fsn.getLastInodeId() == 1002);
 
-    Path filePath = new Path("/test1/file");
-    fs.create(filePath);
+    // Use namenode rpc to create a file
+    NamenodeProtocols nnrpc = cluster.getNameNodeRpc();
+    HdfsFileStatus fileStatus = nnrpc.create("/test1/file", new FsPermission(
+        (short) 0755), "client",
+        new EnumSetWritable<CreateFlag>(EnumSet.of(CreateFlag.CREATE)), true,
+        (short) 1, 128 * 1024 * 1024L);
     assertTrue(fsn.getLastInodeId() == 1003);
+    assertTrue(fileStatus.getFileId() == 1003);
 
     // Rename doesn't increase inode id
     Path renamedPath = new Path("/test2");
@@ -412,4 +421,44 @@ public class TestINodeFile {
     cluster.waitActive();
     assertTrue(fsn.getLastInodeId() == 1003);
   }
+
+  @Test
+  public void testWriteToRenamedFile() throws IOException {
+
+    Configuration conf = new Configuration();
+    MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf).numDataNodes(1)
+        .build();
+    cluster.waitActive();
+    FileSystem fs = cluster.getFileSystem();
+
+    Path path = new Path("/test1");
+    assertTrue(fs.mkdirs(path));
+
+    int size = conf.getInt(DFSConfigKeys.DFS_BYTES_PER_CHECKSUM_KEY, 512);
+    byte[] data = new byte[size];
+
+    // Create one file
+    Path filePath = new Path("/test1/file");
+    FSDataOutputStream fos = fs.create(filePath);
+
+    // Rename /test1 to test2, and recreate /test1/file
+    Path renamedPath = new Path("/test2");
+    fs.rename(path, renamedPath);
+    fs.create(filePath, (short) 1);
+
+    // Add new block should fail since /test1/file has a different fileId
+    try {
+      fos.write(data, 0, data.length);
+      // make sure addBlock() request gets to NN immediately
+      fos.hflush();
+
+      fail("Write should fail after rename");
+    } catch (Exception e) {
+      /* Ignore */
+    } finally {
+      if (cluster != null) {
+        cluster.shutdown();
+      }
+    }
+  }
 }

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestSecondaryNameNodeUpgrade.java

@@ -17,9 +17,12 @@
  */
 package org.apache.hadoop.hdfs.server.namenode;
 
+import com.google.common.collect.ImmutableMap;
+
 import java.io.File;
 import java.io.IOException;
 import java.util.List;
+import java.util.Map;
 
 import org.junit.Test;
 import org.junit.Before;
@@ -51,7 +54,7 @@ public class TestSecondaryNameNodeUpgrade {
     }
   }
 
-  private void doIt(String param, String val) throws IOException {
+  private void doIt(Map<String, String> paramsToCorrupt) throws IOException {
     MiniDFSCluster cluster = null;
     FileSystem fs = null;
     SecondaryNameNode snn = null;
@@ -76,8 +79,12 @@ public class TestSecondaryNameNodeUpgrade {
       snn.shutdown();
 
       for (File versionFile : versionFiles) {
-        System.out.println("Changing '" + param + "' to '" + val + "' in " + versionFile);
-        FSImageTestUtil.corruptVersionFile(versionFile, param, val);
+        for (Map.Entry<String, String> paramToCorrupt : paramsToCorrupt.entrySet()) {
+          String param = paramToCorrupt.getKey();
+          String val = paramToCorrupt.getValue();
+          System.out.println("Changing '" + param + "' to '" + val + "' in " + versionFile);
+          FSImageTestUtil.corruptVersionFile(versionFile, param, val);
+        }
       }
 
       snn = new SecondaryNameNode(conf);
@@ -94,13 +101,19 @@ public class TestSecondaryNameNodeUpgrade {
 
   @Test
   public void testUpgradeLayoutVersionSucceeds() throws IOException {
-    doIt("layoutVersion", "-39");
+    doIt(ImmutableMap.of("layoutVersion", "-39"));
+  }
+
+  @Test
+  public void testUpgradePreFedSucceeds() throws IOException {
+    doIt(ImmutableMap.of("layoutVersion", "-19", "clusterID", "",
+          "blockpoolID", ""));
   }
 
   @Test
   public void testChangeNsIDFails() throws IOException {
     try {
-      doIt("namespaceID", "2");
+      doIt(ImmutableMap.of("namespaceID", "2"));
       Assert.fail("Should throw InconsistentFSStateException");
     } catch(IOException e) {
       GenericTestUtils.assertExceptionContains("Inconsistent checkpoint fields", e);

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/metrics/TestNameNodeMetrics.java

@@ -82,7 +82,7 @@ public class TestNameNodeMetrics {
     CONF.set(DFSConfigKeys.DFS_METRICS_PERCENTILES_INTERVALS_KEY, 
         "" + PERCENTILES_INTERVAL);
     // Enable stale DataNodes checking
-    CONF.setBoolean(DFSConfigKeys.DFS_NAMENODE_CHECK_STALE_DATANODE_KEY, true);
+    CONF.setBoolean(DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_READ_KEY, true);
     ((Log4JLogger)LogFactory.getLog(MetricsAsserts.class))
       .getLogger().setLevel(Level.DEBUG);
   }

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/web/TestJsonUtil.java

@@ -24,6 +24,7 @@ import org.apache.hadoop.fs.Path;
 import org.apache.hadoop.fs.permission.FsPermission;
 import org.apache.hadoop.hdfs.DFSUtil;
 import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
+import org.apache.hadoop.hdfs.server.namenode.INodeId;
 import org.apache.hadoop.util.Time;
 import org.junit.Assert;
 import org.junit.Test;
@@ -42,9 +43,10 @@ public class TestJsonUtil {
   public void testHdfsFileStatus() {
     final long now = Time.now();
     final String parent = "/dir";
-    final HdfsFileStatus status = new HdfsFileStatus(1001L, false, 3, 1L<<26,
-        now, now + 10, new FsPermission((short)0644), "user", "group",
-        DFSUtil.string2Bytes("bar"), DFSUtil.string2Bytes("foo"));
+    final HdfsFileStatus status = new HdfsFileStatus(1001L, false, 3, 1L << 26,
+        now, now + 10, new FsPermission((short) 0644), "user", "group",
+        DFSUtil.string2Bytes("bar"), DFSUtil.string2Bytes("foo"),
+        INodeId.GRANDFATHER_INODE_ID);
     final FileStatus fstatus = toFileStatus(status, parent);
     System.out.println("status  = " + status);
     System.out.println("fstatus = " + fstatus);

hadoop-hdfs-project/hadoop-hdfs/src/test/resources/testHDFSConf.xml

@@ -1182,7 +1182,7 @@
         </comparator>
         <comparator>
           <type>RegexpComparator</type>
-          <expected-output>^1\.0k\s+hdfs:///dir0/data1k</expected-output>
+          <expected-output>^1\.0 K\s+hdfs:///dir0/data1k</expected-output>
         </comparator>
       </comparators>
     </test>
@@ -15590,7 +15590,7 @@
       <comparators>
         <comparator>
           <type>RegexpComparator</type>
-          <expected-output>put: The DiskSpace quota of /dir1 is exceeded: quota=1.0k diskspace consumed=[0-9.]+[kmg]*</expected-output>
+          <expected-output>put: The DiskSpace quota of /dir1 is exceeded: quota = 1024 B = 1 KB but diskspace consumed = [0-9]+ B = [0-9.]+ [KMG]B*</expected-output>
         </comparator>
       </comparators>
     </test>

hadoop-mapreduce-project/CHANGES.txt

@@ -152,7 +152,22 @@ Trunk (Unreleased)
     MAPREDUCE-4884. Streaming tests fail to start MiniMRCluster due to missing
     queue configuration. (Chris Nauroth via suresh)
 
-Release 2.0.3-alpha - Unreleased 
+Release 2.0.4-beta - UNRELEASED
+
+  INCOMPATIBLE CHANGES
+
+  NEW FEATURES
+
+  IMPROVEMENTS
+
+  OPTIMIZATIONS
+
+  BUG FIXES
+
+    MAPREDUCE-4671. AM does not tell the RM about container requests which are
+    no longer needed. (Bikas Saha via sseth)
+
+Release 2.0.3-alpha - 2013-02-06 
 
   INCOMPATIBLE CHANGES
 
@@ -215,6 +230,12 @@ Release 2.0.3-alpha - Unreleased
     MAPREDUCE-4838. Add additional fields like Locality, Avataar to the
     JobHistory logs. (Zhijie Shen via sseth)
 
+    MAPREDUCE-4971. Minor extensibility enhancements to Counters & 
+    FileOutputFormat. (Arun C Murthy via sseth)
+
+    MAPREDUCE-4977. Documentation for pluggable shuffle and pluggable sort. 
+    (tucu)
+
   OPTIMIZATIONS
 
     MAPREDUCE-4893. Fixed MR ApplicationMaster to do optimal assignment of
@@ -284,6 +305,8 @@ Release 2.0.3-alpha - Unreleased
     MAPREDUCE-4969. TestKeyValueTextInputFormat test fails with Open JDK 7.
     (Arpit Agarwal via suresh)
 
+    MAPREDUCE-4953. HadoopPipes misuses fprintf. (Andy Isaacson via atm)
+
 Release 2.0.2-alpha - 2012-09-07 
 
   INCOMPATIBLE CHANGES
@@ -668,11 +691,17 @@ Release 0.23.7 - UNRELEASED
 
   IMPROVEMENTS
 
+    MAPREDUCE-4905. test org.apache.hadoop.mapred.pipes 
+    (Aleksey Gorshkov via bobby)
+
   OPTIMIZATIONS
 
     MAPREDUCE-4946. Fix a performance problem for large jobs by reducing the
     number of map completion event type conversions. (Jason Lowe via sseth)
 
+    MAPREDUCE-4822. Unnecessary conversions in History Events. (Chu Tong via
+    jlowe)
+
   BUG FIXES
 
     MAPREDUCE-4458. Warn if java.library.path is used for AM or Task

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/main/java/org/apache/hadoop/mapreduce/v2/app/rm/RMContainerRequestor.java

@@ -72,7 +72,10 @@ public abstract class RMContainerRequestor extends RMCommunicator {
   remoteRequestsTable =
       new TreeMap<Priority, Map<String, Map<Resource, ResourceRequest>>>();
 
-  private final Set<ResourceRequest> ask = new TreeSet<ResourceRequest>();
+  // use custom comparator to make sure ResourceRequest objects differing only in 
+  // numContainers dont end up as duplicates
+  private final Set<ResourceRequest> ask = new TreeSet<ResourceRequest>(
+      new org.apache.hadoop.yarn.util.BuilderUtils.ResourceRequestComparator());
   private final Set<ContainerId> release = new TreeSet<ContainerId>(); 
 
   private boolean nodeBlacklistingEnabled;
@@ -235,7 +238,7 @@ public abstract class RMContainerRequestor extends RMCommunicator {
               ResourceRequest zeroedRequest = BuilderUtils.newResourceRequest(req);
               zeroedRequest.setNumContainers(0);
               // to be sent to RM on next heartbeat
-              ask.add(zeroedRequest);
+              addResourceRequestToAsk(zeroedRequest);
             }
           }
           // if all requests were still in ask queue
@@ -320,7 +323,7 @@ public abstract class RMContainerRequestor extends RMCommunicator {
     remoteRequest.setNumContainers(remoteRequest.getNumContainers() + 1);
 
     // Note this down for next interaction with ResourceManager
-    ask.add(remoteRequest);
+    addResourceRequestToAsk(remoteRequest);
     if (LOG.isDebugEnabled()) {
       LOG.debug("addResourceRequest:" + " applicationId="
           + applicationId.getId() + " priority=" + priority.getPriority()
@@ -353,7 +356,12 @@ public abstract class RMContainerRequestor extends RMCommunicator {
           + remoteRequest.getNumContainers() + " #asks=" + ask.size());
     }
 
-    remoteRequest.setNumContainers(remoteRequest.getNumContainers() -1);
+    if(remoteRequest.getNumContainers() > 0) {
+      // based on blacklisting comments above we can end up decrementing more 
+      // than requested. so guard for that.
+      remoteRequest.setNumContainers(remoteRequest.getNumContainers() -1);
+    }
+    
     if (remoteRequest.getNumContainers() == 0) {
       reqMap.remove(capability);
       if (reqMap.size() == 0) {
@@ -362,13 +370,12 @@ public abstract class RMContainerRequestor extends RMCommunicator {
       if (remoteRequests.size() == 0) {
         remoteRequestsTable.remove(priority);
       }
-      //remove from ask if it may have
-      ask.remove(remoteRequest);
-    } else {
-      ask.add(remoteRequest);//this will override the request if ask doesn't
-      //already have it.
     }
 
+    // send the updated resource request to RM
+    // send 0 container count requests also to cancel previous requests
+    addResourceRequestToAsk(remoteRequest);
+
     if (LOG.isDebugEnabled()) {
       LOG.info("AFTER decResourceRequest:" + " applicationId="
           + applicationId.getId() + " priority=" + priority.getPriority()
@@ -376,6 +383,16 @@ public abstract class RMContainerRequestor extends RMCommunicator {
           + remoteRequest.getNumContainers() + " #asks=" + ask.size());
     }
   }
+  
+  private void addResourceRequestToAsk(ResourceRequest remoteRequest) {
+    // because objects inside the resource map can be deleted ask can end up 
+    // containing an object that matches new resource object but with different
+    // numContainers. So exisintg values must be replaced explicitly
+    if(ask.contains(remoteRequest)) {
+      ask.remove(remoteRequest);
+    }
+    ask.add(remoteRequest);    
+  }
 
   protected void release(ContainerId containerId) {
     release.add(containerId);

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapreduce/v2/app/TestRMContainerAllocator.java

@@ -167,6 +167,7 @@ public class TestRMContainerAllocator {
     List<TaskAttemptContainerAssignedEvent> assigned = allocator.schedule();
     dispatcher.await();
     Assert.assertEquals("No of assignments must be 0", 0, assigned.size());
+    Assert.assertEquals(4, rm.getMyFifoScheduler().lastAsk.size());
 
     // send another request with different resource and priority
     ContainerRequestEvent event3 = createReq(jobId, 3, 1024,
@@ -178,7 +179,8 @@ public class TestRMContainerAllocator {
     assigned = allocator.schedule();
     dispatcher.await();
     Assert.assertEquals("No of assignments must be 0", 0, assigned.size());
-
+    Assert.assertEquals(3, rm.getMyFifoScheduler().lastAsk.size());
+    
     // update resources in scheduler
     nodeManager1.nodeHeartbeat(true); // Node heartbeat
     nodeManager2.nodeHeartbeat(true); // Node heartbeat
@@ -187,8 +189,14 @@ public class TestRMContainerAllocator {
 
     assigned = allocator.schedule();
     dispatcher.await();
+    Assert.assertEquals(0, rm.getMyFifoScheduler().lastAsk.size());
     checkAssignments(new ContainerRequestEvent[] { event1, event2, event3 },
         assigned, false);
+    
+    // check that the assigned container requests are cancelled
+    assigned = allocator.schedule();
+    dispatcher.await();
+    Assert.assertEquals(5, rm.getMyFifoScheduler().lastAsk.size());    
   }
   
   @Test 
@@ -422,7 +430,7 @@ public class TestRMContainerAllocator {
   }
 
   private static class MyResourceManager extends MockRM {
-
+    
     public MyResourceManager(Configuration conf) {
       super(conf);
     }
@@ -446,6 +454,10 @@ public class TestRMContainerAllocator {
     protected ResourceScheduler createScheduler() {
       return new MyFifoScheduler(this.getRMContext());
     }
+    
+    MyFifoScheduler getMyFifoScheduler() {
+      return (MyFifoScheduler) scheduler;
+    }
   }
 
   @Test
@@ -1194,7 +1206,9 @@ public class TestRMContainerAllocator {
         assert (false);
       }
     }
-
+    
+    List<ResourceRequest> lastAsk = null;
+    
     // override this to copy the objects otherwise FifoScheduler updates the
     // numContainers in same objects as kept by RMContainerAllocator
     @Override
@@ -1208,6 +1222,7 @@ public class TestRMContainerAllocator {
             .getNumContainers());
         askCopy.add(reqCopy);
       }
+      lastAsk = ask;
       return super.allocate(applicationAttemptId, askCopy, release);
     }
   }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-app/src/test/java/org/apache/hadoop/mapreduce/v2/app/job/impl/TestTaskImpl.java

@@ -60,7 +60,7 @@ import org.apache.hadoop.yarn.Clock;
 import org.apache.hadoop.yarn.SystemClock;
 import org.apache.hadoop.yarn.api.records.ApplicationId;
 import org.apache.hadoop.yarn.event.EventHandler;
-import org.apache.hadoop.yarn.server.resourcemanager.resourcetracker.InlineDispatcher;
+import org.apache.hadoop.yarn.event.InlineDispatcher;
 import org.apache.hadoop.yarn.util.Records;
 import org.junit.After;
 import org.junit.Before;

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/Counters.java

@@ -230,6 +230,10 @@ public class Counters
   public static class Group implements CounterGroupBase<Counter> {
     private CounterGroupBase<Counter> realGroup;
     
+    protected Group() {
+      realGroup = null;
+    }
+    
     Group(GenericGroup group) {
       this.realGroup = group;
     }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/FileOutputCommitter.java

@@ -92,7 +92,7 @@ public class FileOutputCommitter extends OutputCommitter {
   }
 
   @Private
-  Path getTaskAttemptPath(TaskAttemptContext context) throws IOException {
+  public Path getTaskAttemptPath(TaskAttemptContext context) throws IOException {
     Path out = getOutputPath(context);
     return out == null ? null : getTaskAttemptPath(context, out);
   }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/FileOutputFormat.java

@@ -22,6 +22,7 @@ import java.io.IOException;
 import java.text.NumberFormat;
 
 import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.classification.InterfaceStability;
 import org.apache.hadoop.fs.FileSystem;
 import org.apache.hadoop.fs.Path;
@@ -152,8 +153,8 @@ public abstract class FileOutputFormat<K, V> implements OutputFormat<K, V> {
    * @param outputDir the {@link Path} of the output directory 
    * for the map-reduce job.
    */
-  
-  static void setWorkOutputPath(JobConf conf, Path outputDir) {
+  @Private
+  public static void setWorkOutputPath(JobConf conf, Path outputDir) {
     outputDir = new Path(conf.getWorkingDirectory(), outputDir);
     conf.set(JobContext.TASK_OUTPUT_DIR, outputDir.toString());
   }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/JobConf.java

@@ -28,6 +28,7 @@ import java.util.regex.Pattern;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.classification.InterfaceStability;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.FileStatus;
@@ -419,7 +420,8 @@ public class JobConf extends Configuration {
     return credentials;
   }
   
-  void setCredentials(Credentials credentials) {
+  @Private
+  public void setCredentials(Credentials credentials) {
     this.credentials = credentials;
   }
   

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/pipes/Submitter.java

@@ -58,6 +58,7 @@ import org.apache.hadoop.mapred.lib.LazyOutputFormat;
 import org.apache.hadoop.mapred.lib.NullOutputFormat;
 import org.apache.hadoop.mapreduce.MRJobConfig;
 import org.apache.hadoop.mapreduce.filecache.DistributedCache;
+import org.apache.hadoop.util.ExitUtil;
 import org.apache.hadoop.util.GenericOptionsParser;
 import org.apache.hadoop.util.Tool;
 
@@ -515,7 +516,7 @@ public class Submitter extends Configured implements Tool {
    */
   public static void main(String[] args) throws Exception {
     int exitCode =  new Submitter().run(args);
-    System.exit(exitCode);
+    ExitUtil.terminate(exitCode);
   }
 
 }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/counters/FileSystemCounterGroup.java

@@ -35,6 +35,7 @@ import com.google.common.collect.Iterators;
 import com.google.common.collect.Maps;
 
 import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.io.WritableUtils;
 import org.apache.hadoop.mapreduce.Counter;
 import org.apache.hadoop.mapreduce.FileSystemCounter;
@@ -72,6 +73,16 @@ public abstract class FileSystemCounterGroup<C extends Counter>
       this.scheme = scheme;
       key = ref;
     }
+    
+    @Private
+    public String getScheme() {
+      return scheme;
+    }
+    
+    @Private
+    public FileSystemCounter getFileSystemCounter() {
+      return key;
+    }
 
     @Override
     public String getName() {

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/counters/FrameworkCounterGroup.java

@@ -29,6 +29,7 @@ import java.util.Iterator;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.io.WritableUtils;
 import org.apache.hadoop.mapreduce.Counter;
 import org.apache.hadoop.mapreduce.util.ResourceBundles;
@@ -66,7 +67,17 @@ public abstract class FrameworkCounterGroup<T extends Enum<T>,
       key = ref;
       this.groupName = groupName;
     }
+    
+    @Private
+    public T getKey() {
+      return key;
+    }
 
+    @Private
+    public String getGroupName() {
+      return groupName;
+    }
+    
     @Override
     public String getName() {
       return key.name();

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/jobhistory/MapAttemptFinishedEvent.java

@@ -178,7 +178,7 @@ public class MapAttemptFinishedEvent  implements HistoryEvent {
 
   /** Get the task type */
   public TaskType getTaskType() {
-    return TaskType.valueOf(taskType.toString());
+    return taskType;
   }
   /** Get the task status */
   public String getTaskStatus() { return taskStatus.toString(); }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/jobhistory/ReduceAttemptFinishedEvent.java

@@ -176,11 +176,11 @@ public class ReduceAttemptFinishedEvent  implements HistoryEvent {
   public TaskID getTaskId() { return attemptId.getTaskID(); }
   /** Get the attempt id */
   public TaskAttemptID getAttemptId() {
-    return TaskAttemptID.forName(attemptId.toString());
+    return attemptId;
   }
   /** Get the task type */
   public TaskType getTaskType() {
-    return TaskType.valueOf(taskType.toString());
+    return taskType;
   }
   /** Get the task status */
   public String getTaskStatus() { return taskStatus.toString(); }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/jobhistory/TaskAttemptFinishedEvent.java

@@ -105,11 +105,11 @@ public class TaskAttemptFinishedEvent  implements HistoryEvent {
   public TaskID getTaskId() { return attemptId.getTaskID(); }
   /** Get the task attempt id */
   public TaskAttemptID getAttemptId() {
-    return TaskAttemptID.forName(attemptId.toString());
+    return attemptId;
   }
   /** Get the task type */
   public TaskType getTaskType() {
-    return TaskType.valueOf(taskType.toString());
+    return taskType;
   }
   /** Get the task status */
   public String getTaskStatus() { return taskStatus.toString(); }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/jobhistory/TaskFinishedEvent.java

@@ -95,14 +95,10 @@ public class TaskFinishedEvent implements HistoryEvent {
   }
 
   /** Get task id */
-  public TaskID getTaskId() { return TaskID.forName(taskid.toString()); }
+  public TaskID getTaskId() { return taskid; }
   /** Get successful task attempt id */
   public TaskAttemptID getSuccessfulTaskAttemptId() {
-    if(successfulAttemptId != null)
-    {
-      return TaskAttemptID.forName(successfulAttemptId.toString());
-    }
-    return null;
+    return successfulAttemptId;
   }
   /** Get the task finish time */
   public long getFinishTime() { return finishTime; }
@@ -110,7 +106,7 @@ public class TaskFinishedEvent implements HistoryEvent {
   public Counters getCounters() { return counters; }
   /** Get task type */
   public TaskType getTaskType() {
-    return TaskType.valueOf(taskType.toString());
+    return taskType;
   }
   /** Get task status */
   public String getTaskStatus() { return status.toString(); }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/site/apt/PluggableShuffleAndPluggableSort.apt.vm

@@ -0,0 +1,96 @@
+~~ Licensed under the Apache License, Version 2.0 (the "License");
+~~ you may not use this file except in compliance with the License.
+~~ You may obtain a copy of the License at
+~~
+~~   http://www.apache.org/licenses/LICENSE-2.0
+~~
+~~ Unless required by applicable law or agreed to in writing, software
+~~ distributed under the License is distributed on an "AS IS" BASIS,
+~~ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+~~ See the License for the specific language governing permissions and
+~~ limitations under the License. See accompanying LICENSE file.
+
+  ---
+  Hadoop Map Reduce Next Generation-${project.version} - Pluggable Shuffle and Pluggable Sort
+  ---
+  ---
+  ${maven.build.timestamp}
+
+Hadoop MapReduce Next Generation - Pluggable Shuffle and Pluggable Sort
+
+  \[ {{{./index.html}Go Back}} \]
+
+* Introduction
+
+  The pluggable shuffle and pluggable sort capabilities allow replacing the 
+  built in shuffle and sort logic with alternate implementations. Example use 
+  cases for this are: using a different application protocol other than HTTP 
+  such as RDMA for shuffling data from the Map nodes to the Reducer nodes; or
+  replacing the sort logic with custom algorithms that enable Hash aggregation 
+  and Limit-N query.
+
+  <<IMPORTANT:>> The pluggable shuffle and pluggable sort capabilities are 
+  experimental and unstable. This means the provided APIs may change and break 
+  compatibility in future versions of Hadoop.
+
+* Implementing a Custom Shuffle and a Custom Sort 
+
+  A custom shuffle implementation requires a
+  <<<org.apache.hadoop.yarn.server.nodemanager.containermanager.AuxServices.AuxiliaryService>>> 
+  implementation class running in the NodeManagers and a 
+  <<<org.apache.hadoop.mapred.ShuffleConsumerPlugin>>> implementation class
+  running in the Reducer tasks.
+
+  The default implementations provided by Hadoop can be used as references:
+
+    * <<<org.apache.hadoop.mapred.ShuffleHandler>>>
+    
+    * <<<org.apache.hadoop.mapreduce.task.reduce.Shuffle>>>
+
+  A custom sort implementation requires a <<<org.apache.hadoop.mapred.MapOutputCollector>>>
+  implementation class running in the Mapper tasks and (optionally, depending
+  on the sort implementation) a <<<org.apache.hadoop.mapred.ShuffleConsumerPlugin>>> 
+  implementation class running in the Reducer tasks.
+
+  The default implementations provided by Hadoop can be used as references:
+
+  * <<<org.apache.hadoop.mapred.MapTask$MapOutputBuffer>>>
+  
+  * <<<org.apache.hadoop.mapreduce.task.reduce.Shuffle>>>
+
+* Configuration
+
+  Except for the auxiliary service running in the NodeManagers serving the 
+  shuffle (by default the <<<ShuffleHandler>>>), all the pluggable components 
+  run in the job tasks. This means, they can be configured on per job basis. 
+  The auxiliary service servicing the Shuffle must be configured in the 
+  NodeManagers configuration.
+
+** Job Configuration Properties (on per job basis):
+
+*--------------------------------------+---------------------+-----------------+
+| <<Property>>                         | <<Default Value>>   | <<Explanation>> |
+*--------------------------------------+---------------------+-----------------+
+| <<<mapreduce.job.reduce.shuffle.consumer.plugin.class>>> | <<<org.apache.hadoop.mapreduce.task.reduce.Shuffle>>>         | The <<<ShuffleConsumerPlugin>>> implementation to use |
+*--------------------------------------+---------------------+-----------------+
+| <<<mapreduce.job.map.output.collector.class>>>   | <<<org.apache.hadoop.mapred.MapTask$MapOutputBuffer>>> | The <<<MapOutputCollector>>> implementation to use |
+*--------------------------------------+---------------------+-----------------+
+
+  These properties can also be set in the <<<mapred-site.xml>>> to change the default values for all jobs.
+
+** NodeManager Configuration properties, <<<yarn-site.xml>>> in all nodes:
+
+*--------------------------------------+---------------------+-----------------+
+| <<Property>>                         | <<Default Value>>   | <<Explanation>> |
+*--------------------------------------+---------------------+-----------------+
+| <<<yarn.nodemanager.aux-services>>> | <<<...,mapreduce.shuffle>>>  | The auxiliary service name |
+*--------------------------------------+---------------------+-----------------+
+| <<<yarn.nodemanager.aux-services.mapreduce.shuffle.class>>>   | <<<org.apache.hadoop.mapred.ShuffleHandler>>> | The auxiliary service class to use |
+*--------------------------------------+---------------------+-----------------+
+
+  <<IMPORTANT:>> If setting an auxiliary service in addition the default 
+  <<<mapreduce.shuffle>>> service, then a new service key should be added to the
+  <<<yarn.nodemanager.aux-services>>> property, for example <<<mapred.shufflex>>>.
+  Then the property defining the corresponding class must be
+  <<<yarn.nodemanager.aux-services.mapreduce.shufflex.class>>>.
+