Merge trunk to HDFS-4685.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/HDFS-4685@1566100 13f79535-47bb-0310-9956-ffa450edef68

dev-support/test-patch.sh

@@ -300,6 +300,17 @@ prebuildWithoutPatch () {
     {color:red}-1 patch{color}.  Trunk compilation may be broken."
     return 1
   fi
+
+  echo "$MVN clean test javadoc:javadoc -DskipTests -Pdocs -D${PROJECT_NAME}PatchProcess > $PATCH_DIR/trunkJavadocWarnings.txt 2>&1"
+  $MVN clean test javadoc:javadoc -DskipTests -Pdocs -D${PROJECT_NAME}PatchProcess > $PATCH_DIR/trunkJavadocWarnings.txt 2>&1
+  if [[ $? != 0 ]] ; then
+    echo "Trunk javadoc compilation is broken?"
+    JIRA_COMMENT="$JIRA_COMMENT
+
+    {color:red}-1 patch{color}.  Trunk compilation may be broken."
+    return 1
+  fi
+
   return 0
 }
 
@@ -401,6 +412,11 @@ applyPatch () {
 }
 
 ###############################################################################
+calculateJavadocWarnings() {
+    WARNING_FILE="$1"
+    RET=$(egrep "^[0-9]+ warnings$" "$WARNING_FILE" | awk '{sum+=$1} END {print sum}')
+}
+
 ### Check there are no javadoc warnings
 checkJavadocWarnings () {
   echo ""
@@ -420,24 +436,29 @@ checkJavadocWarnings () {
     (cd hadoop-common-project/hadoop-annotations; $MVN install > /dev/null 2>&1)
   fi
   $MVN clean test javadoc:javadoc -DskipTests -Pdocs -D${PROJECT_NAME}PatchProcess > $PATCH_DIR/patchJavadocWarnings.txt 2>&1
-  javadocWarnings=`$GREP '\[WARNING\]' $PATCH_DIR/patchJavadocWarnings.txt | $AWK '/Javadoc Warnings/,EOF' | $GREP warning | $AWK 'BEGIN {total = 0} {total += 1} END {print total}'`
-  echo ""
-  echo ""
-  echo "There appear to be $javadocWarnings javadoc warnings generated by the patched build."
+  calculateJavadocWarnings "$PATCH_DIR/trunkJavadocWarnings.txt"
+  numTrunkJavadocWarnings=$RET
+  calculateJavadocWarnings "$PATCH_DIR/patchJavadocWarnings.txt"
+  numPatchJavadocWarnings=$RET
+  grep -i warning "$PATCH_DIR/trunkJavadocWarnings.txt" > "$PATCH_DIR/trunkJavadocWarningsFiltered.txt"
+  grep -i warning "$PATCH_DIR/patchJavadocWarnings.txt" > "$PATCH_DIR/patchJavadocWarningsFiltered.txt"
+  diff -u "$PATCH_DIR/trunkJavadocWarningsFiltered.txt" \
+          "$PATCH_DIR/patchJavadocWarningsFiltered.txt" > \
+          "$PATCH_DIR/diffJavadocWarnings.txt"
+  rm -f "$PATCH_DIR/trunkJavadocWarningsFiltered.txt" "$PATCH_DIR/patchJavadocWarningsFiltered.txt"
+  echo "There appear to be $numTrunkJavadocWarnings javadoc warnings before the patch and $numPatchJavadocWarnings javadoc warnings after applying the patch."
+  if [[ $numTrunkJavadocWarnings != "" && $numPatchJavadocWarnings != "" ]] ; then
+    if [[ $numPatchJavadocWarnings -gt $numTrunkJavadocWarnings ]] ; then
+      JIRA_COMMENT="$JIRA_COMMENT
 
-  #There are 12 warnings that are caused by things that are caused by using sun internal APIs.
-  #There are 2 warnings that are caused by the Apache DS Dn class used in MiniKdc.
-  OK_JAVADOC_WARNINGS=14;
-  ### if current warnings greater than OK_JAVADOC_WARNINGS
-  if [[ $javadocWarnings -ne $OK_JAVADOC_WARNINGS ]] ; then
-    JIRA_COMMENT="$JIRA_COMMENT
-
-    {color:red}-1 javadoc{color}.  The javadoc tool appears to have generated `expr $(($javadocWarnings-$OK_JAVADOC_WARNINGS))` warning messages."
-    return 1
+    {color:red}-1 javadoc{color}.  The javadoc tool appears to have generated `expr $(($numPatchJavadocWarnings-$numTrunkJavadocWarnings))` warning messages.
+        See $BUILD_URL/artifact/trunk/patchprocess/diffJavadocWarnings.txt for details."
+        return 1
+    fi
   fi
   JIRA_COMMENT="$JIRA_COMMENT
 
-    {color:green}+1 javadoc{color}.  The javadoc tool did not generate any warning messages."
+    {color:green}+1 javadoc{color}.  There were no new javadoc warning messages."
   return 0
 }
 

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

@@ -113,6 +113,11 @@ Trunk (Unreleased)
 
     HADOOP-10177. Create CLI tools for managing keys. (Larry McCay via omalley)
 
+    HADOOP-10244. TestKeyShell improperly tests the results of delete (Larry
+    McCay via omalley)
+
+    HADOOP-10325. Improve jenkins javadoc warnings from test-patch.sh (cmccabe)
+
   BUG FIXES
 
     HADOOP-9451. Fault single-layer config if node group topology is enabled.
@@ -313,6 +318,15 @@ Release 2.4.0 - UNRELEASED
 
     HADOOP-10320. Javadoc in InterfaceStability.java lacks final </ul>.
     (René Nyffenegger via cnauroth)
+    
+    HADOOP-10085. CompositeService should allow adding services while being 
+    inited. (Steve Loughran via kasha)
+
+    HADOOP-10327. Trunk windows build broken after HDFS-5746.
+    (Vinay via cnauroth)
+
+    HADOOP-10330. TestFrameDecoder fails if it cannot bind port 12345.
+    (Arpit Agarwal)
 
 Release 2.3.0 - UNRELEASED
 
@@ -685,6 +699,8 @@ Release 2.3.0 - UNRELEASED
 
     HADOOP-10311. Cleanup vendor names from the code base. (tucu)
 
+    HADOOP-10273. Fix 'mvn site'. (Arpit Agarwal)
+
 Release 2.2.0 - 2013-10-13
 
   INCOMPATIBLE CHANGES

hadoop-common-project/hadoop-common/pom.xml

@@ -543,6 +543,7 @@
                     <javahClassName>org.apache.hadoop.io.compress.bzip2.Bzip2Decompressor</javahClassName>
                     <javahClassName>org.apache.hadoop.security.JniBasedUnixGroupsMapping</javahClassName>
                     <javahClassName>org.apache.hadoop.io.nativeio.NativeIO</javahClassName>
+                    <javahClassName>org.apache.hadoop.io.nativeio.SharedFileDescriptorFactory</javahClassName>
                     <javahClassName>org.apache.hadoop.security.JniBasedUnixGroupsNetgroupMapping</javahClassName>
                     <javahClassName>org.apache.hadoop.io.compress.snappy.SnappyCompressor</javahClassName>
                     <javahClassName>org.apache.hadoop.io.compress.snappy.SnappyDecompressor</javahClassName>
@@ -550,6 +551,7 @@
                     <javahClassName>org.apache.hadoop.io.compress.lz4.Lz4Decompressor</javahClassName>
                     <javahClassName>org.apache.hadoop.util.NativeCrc32</javahClassName>
                     <javahClassName>org.apache.hadoop.net.unix.DomainSocket</javahClassName>
+                    <javahClassName>org.apache.hadoop.net.unix.DomainSocketWatcher</javahClassName>
                   </javahClassNames>
                   <javahOutputDirectory>${project.build.directory}/native/javah</javahOutputDirectory>
                 </configuration>

hadoop-common-project/hadoop-common/src/CMakeLists.txt

@@ -178,7 +178,9 @@ add_dual_library(hadoop
     ${D}/io/nativeio/NativeIO.c
     ${D}/io/nativeio/errno_enum.c
     ${D}/io/nativeio/file_descriptor.c
+    ${D}/io/nativeio/SharedFileDescriptorFactory.c
     ${D}/net/unix/DomainSocket.c
+    ${D}/net/unix/DomainSocketWatcher.c
     ${D}/security/JniBasedUnixGroupsMapping.c
     ${D}/security/JniBasedUnixGroupsNetgroupMapping.c
     ${D}/security/hadoop_group_info.c

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/http/HttpConfig.java

@@ -34,13 +34,14 @@ public enum Policy {
     HTTPS_ONLY,
     HTTP_AND_HTTPS;
 
+    private static final Policy[] VALUES = values();
     public static Policy fromString(String value) {
-      if (HTTPS_ONLY.name().equalsIgnoreCase(value)) {
-        return HTTPS_ONLY;
-      } else if (HTTP_AND_HTTPS.name().equalsIgnoreCase(value)) {
-        return HTTP_AND_HTTPS;
+      for (Policy p : VALUES) {
+        if (p.name().equalsIgnoreCase(value)) {
+          return p;
+        }
       }
-      return HTTP_ONLY;
+      return null;
     }
 
     public boolean isHttpEnabled() {

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

@@ -487,6 +487,16 @@ public CachedName(String name, long timestamp) {
       new ConcurrentHashMap<Integer, CachedName>();
 
     private enum IdCache { USER, GROUP }
+
+    public final static int MMAP_PROT_READ = 0x1; 
+    public final static int MMAP_PROT_WRITE = 0x2; 
+    public final static int MMAP_PROT_EXEC = 0x4; 
+
+    public static native long mmap(FileDescriptor fd, int prot,
+        boolean shared, long length) throws IOException;
+
+    public static native void munmap(long addr, long length)
+        throws IOException;
   }
 
   private static boolean workaroundNonThreadSafePasswdCalls = false;

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

@@ -0,0 +1,90 @@
+/**
+ * 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.io.nativeio;
+
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.io.FileDescriptor;
+
+import org.apache.commons.lang.SystemUtils;
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceStability;
+
+import com.google.common.base.Preconditions;
+
+/**
+ * A factory for creating shared file descriptors inside a given directory.
+ * Typically, the directory will be /dev/shm or /tmp.
+ *
+ * We will hand out file descriptors that correspond to unlinked files residing
+ * in that directory.  These file descriptors are suitable for sharing across
+ * multiple processes and are both readable and writable.
+ *
+ * Because we unlink the temporary files right after creating them, a JVM crash
+ * usually does not leave behind any temporary files in the directory.  However,
+ * it may happen that we crash right after creating the file and before
+ * unlinking it.  In the constructor, we attempt to clean up after any such
+ * remnants by trying to unlink any temporary files created by previous
+ * SharedFileDescriptorFactory instances that also used our prefix.
+ */
+@InterfaceAudience.Private
+@InterfaceStability.Unstable
+public class SharedFileDescriptorFactory {
+  private final String prefix;
+  private final String path;
+
+  /**
+   * Create a SharedFileDescriptorFactory.
+   *
+   * @param prefix    Prefix to add to all file names we use.
+   * @param path      Path to use.
+   */
+  public SharedFileDescriptorFactory(String prefix, String path)
+      throws IOException {
+    Preconditions.checkArgument(NativeIO.isAvailable());
+    Preconditions.checkArgument(SystemUtils.IS_OS_UNIX);
+    this.prefix = prefix;
+    this.path = path;
+    deleteStaleTemporaryFiles0(prefix, path);
+  }
+
+  /**
+   * Create a shared file descriptor which will be both readable and writable.
+   *
+   * @param length         The starting file length.
+   *
+   * @return               The file descriptor, wrapped in a FileInputStream.
+   * @throws IOException   If there was an I/O or configuration error creating
+   *                       the descriptor.
+   */
+  public FileInputStream createDescriptor(int length) throws IOException {
+    return new FileInputStream(createDescriptor0(prefix, path, length));
+  }
+
+  /**
+   * Delete temporary files in the directory, NOT following symlinks.
+   */
+  private static native void deleteStaleTemporaryFiles0(String prefix,
+      String path) throws IOException;
+
+  /**
+   * Create a file with O_EXCL, and then resize it to the desired size.
+   */
+  private static native FileDescriptor createDescriptor0(String prefix,
+      String path, int length) throws IOException;
+}

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

@@ -151,6 +151,13 @@ public static final RetryPolicy failoverOnNetworkException(
         delayMillis, maxDelayBase);
   }
   
+  public static final RetryPolicy failoverOnNetworkException(
+      RetryPolicy fallbackPolicy, int maxFailovers, int maxRetries,
+      long delayMillis, long maxDelayBase) {
+    return new FailoverOnNetworkExceptionRetry(fallbackPolicy, maxFailovers,
+        maxRetries, delayMillis, maxDelayBase);
+  }
+  
   static class TryOnceThenFail implements RetryPolicy {
     @Override
     public RetryAction shouldRetry(Exception e, int retries, int failovers,
@@ -516,18 +523,25 @@ static class FailoverOnNetworkExceptionRetry implements RetryPolicy {
     
     private RetryPolicy fallbackPolicy;
     private int maxFailovers;
+    private int maxRetries;
     private long delayMillis;
     private long maxDelayBase;
     
     public FailoverOnNetworkExceptionRetry(RetryPolicy fallbackPolicy,
         int maxFailovers) {
-      this(fallbackPolicy, maxFailovers, 0, 0);
+      this(fallbackPolicy, maxFailovers, 0, 0, 0);
     }
     
     public FailoverOnNetworkExceptionRetry(RetryPolicy fallbackPolicy,
         int maxFailovers, long delayMillis, long maxDelayBase) {
+      this(fallbackPolicy, maxFailovers, 0, delayMillis, maxDelayBase);
+    }
+    
+    public FailoverOnNetworkExceptionRetry(RetryPolicy fallbackPolicy,
+        int maxFailovers, int maxRetries, long delayMillis, long maxDelayBase) {
       this.fallbackPolicy = fallbackPolicy;
       this.maxFailovers = maxFailovers;
+      this.maxRetries = maxRetries;
       this.delayMillis = delayMillis;
       this.maxDelayBase = maxDelayBase;
     }
@@ -549,6 +563,10 @@ public RetryAction shouldRetry(Exception e, int retries,
             "failovers (" + failovers + ") exceeded maximum allowed ("
             + maxFailovers + ")");
       }
+      if (retries - failovers > maxRetries) {
+        return new RetryAction(RetryAction.RetryDecision.FAIL, 0, "retries ("
+            + retries + ") exceeded maximum allowed (" + maxRetries + ")");
+      }
       
       if (e instanceof ConnectException ||
           e instanceof NoRouteToHostException ||

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

@@ -450,6 +450,14 @@ public void refreshServiceAcl(Configuration conf, PolicyProvider provider) {
     serviceAuthorizationManager.refresh(conf, provider);
   }
 
+  /**
+   * Refresh the service authorization ACL for the service handled by this server
+   * using the specified Configuration.
+   */
+  public void refreshServiceAclWithConfigration(Configuration conf,
+      PolicyProvider provider) {
+    serviceAuthorizationManager.refreshWithConfiguration(conf, provider);
+  }
   /**
    * Returns a handle to the serviceAuthorizationManager (required in tests)
    * @return instance of ServiceAuthorizationManager for this server

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

@@ -24,17 +24,15 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
-import java.net.SocketException;
-import java.nio.channels.AsynchronousCloseException;
 import java.nio.channels.ClosedChannelException;
 import java.nio.channels.ReadableByteChannel;
 import java.nio.ByteBuffer;
-import java.util.concurrent.atomic.AtomicInteger;
 
 import org.apache.commons.lang.SystemUtils;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.util.NativeCodeLoader;
+import org.apache.hadoop.util.CloseableReferenceCount;
 
 import com.google.common.annotations.VisibleForTesting;
 
@@ -132,104 +130,14 @@ public static String getEffectivePath(String path, int port) {
   }
 
   /**
-   * Tracks the reference count of the file descriptor, and also whether it is
-   * open or closed.
+   * The socket reference count and closed bit.
    */
-  private static class Status {
-    /**
-     * Bit mask representing a closed domain socket. 
-     */
-    private static final int STATUS_CLOSED_MASK = 1 << 30;
-    
-    /**
-     * Status bits
-     * 
-     * Bit 30: 0 = DomainSocket open, 1 = DomainSocket closed
-     * Bits 29 to 0: the reference count.
-     */
-    private final AtomicInteger bits = new AtomicInteger(0);
-
-    Status() { }
-
-    /**
-     * Increment the reference count of the underlying file descriptor.
-     *
-     * @throws ClosedChannelException      If the file descriptor is closed.
-     */
-    void reference() throws ClosedChannelException {
-      int curBits = bits.incrementAndGet();
-      if ((curBits & STATUS_CLOSED_MASK) != 0) {
-        bits.decrementAndGet();
-        throw new ClosedChannelException();
-      }
-    }
-
-    /**
-     * Decrement the reference count of the underlying file descriptor.
-     *
-     * @param checkClosed        Whether to throw an exception if the file
-     *                           descriptor is closed.
-     *
-     * @throws AsynchronousCloseException  If the file descriptor is closed and
-     *                                     checkClosed is set.
-     */
-    void unreference(boolean checkClosed) throws AsynchronousCloseException {
-      int newCount = bits.decrementAndGet();
-      assert (newCount & ~STATUS_CLOSED_MASK) >= 0;
-      if (checkClosed && ((newCount & STATUS_CLOSED_MASK) != 0)) {
-        throw new AsynchronousCloseException();
-      }
-    }
-
-    /**
-     * Return true if the file descriptor is currently open.
-     * 
-     * @return                 True if the file descriptor is currently open.
-     */
-    boolean isOpen() {
-      return ((bits.get() & STATUS_CLOSED_MASK) == 0);
-    }
-
-    /**
-     * Mark the file descriptor as closed.
-     *
-     * Once the file descriptor is closed, it cannot be reopened.
-     *
-     * @return                         The current reference count.
-     * @throws ClosedChannelException  If someone else closes the file 
-     *                                 descriptor before we do.
-     */
-    int setClosed() throws ClosedChannelException {
-      while (true) {
-        int curBits = bits.get();
-        if ((curBits & STATUS_CLOSED_MASK) != 0) {
-          throw new ClosedChannelException();
-        }
-        if (bits.compareAndSet(curBits, curBits | STATUS_CLOSED_MASK)) {
-          return curBits & (~STATUS_CLOSED_MASK);
-        }
-      }
-    }
-
-    /**
-     * Get the current reference count.
-     *
-     * @return                 The current reference count.
-     */
-    int getReferenceCount() {
-      return bits.get() & (~STATUS_CLOSED_MASK);
-    }
-  }
-
-  /**
-   * The socket status.
-   */
-  private final Status status;
+  final CloseableReferenceCount refCount;
 
   /**
    * The file descriptor associated with this UNIX domain socket.
    */
-  private final int fd;
+  final int fd;
 
   /**
    * The path associated with this UNIX domain socket.
@@ -252,13 +160,21 @@ int getReferenceCount() {
   private final DomainChannel channel = new DomainChannel();
 
   private DomainSocket(String path, int fd) {
-    this.status = new Status();
+    this.refCount = new CloseableReferenceCount();
     this.fd = fd;
     this.path = path;
   }
 
   private static native int bind0(String path) throws IOException;
 
+  private void unreference(boolean checkClosed) throws ClosedChannelException {
+    if (checkClosed) {
+      refCount.unreferenceCheckClosed();
+    } else {
+      refCount.unreference();
+    }
+  }
+
   /**
    * Create a new DomainSocket listening on the given path.
    *
@@ -308,14 +224,14 @@ public static DomainSocket[] socketpair() throws IOException {
    * @throws SocketTimeoutException       If the accept timed out.
    */
   public DomainSocket accept() throws IOException {
-    status.reference();
+    refCount.reference();
     boolean exc = true;
     try {
       DomainSocket ret = new DomainSocket(path, accept0(fd));
       exc = false;
       return ret;
     } finally {
-      status.unreference(exc);
+      unreference(exc);
     }
   }
 
@@ -335,14 +251,14 @@ public static DomainSocket connect(String path) throws IOException {
     return new DomainSocket(path, fd);
   }
 
- /**
-  * Return true if the file descriptor is currently open.
-  *
-  * @return                 True if the file descriptor is currently open.
-  */
- public boolean isOpen() {
-   return status.isOpen();
- }
+  /**
+   * Return true if the file descriptor is currently open.
+   *
+   * @return                 True if the file descriptor is currently open.
+   */
+  public boolean isOpen() {
+    return refCount.isOpen();
+  }
 
   /**
    * @return                 The socket path.
@@ -381,20 +297,20 @@ private static native void setAttribute0(int fd, int type, int val)
       throws IOException;
 
   public void setAttribute(int type, int size) throws IOException {
-    status.reference();
+    refCount.reference();
     boolean exc = true;
     try {
       setAttribute0(fd, type, size);
       exc = false;
     } finally {
-      status.unreference(exc);
+      unreference(exc);
     }
   }
 
   private native int getAttribute0(int fd, int type) throws IOException;
 
   public int getAttribute(int type) throws IOException {
-    status.reference();
+    refCount.reference();
     int attribute;
     boolean exc = true;
     try {
@@ -402,7 +318,7 @@ public int getAttribute(int type) throws IOException {
       exc = false;
       return attribute;
     } finally {
-      status.unreference(exc);
+      unreference(exc);
     }
   }
 
@@ -419,9 +335,9 @@ private static native void closeFileDescriptor0(FileDescriptor fd)
   @Override
   public void close() throws IOException {
     // Set the closed bit on this DomainSocket
-    int refCount;
+    int count;
     try {
-      refCount = status.setClosed();
+      count = refCount.setClosed();
     } catch (ClosedChannelException e) {
       // Someone else already closed the DomainSocket.
       return;
@@ -429,7 +345,7 @@ public void close() throws IOException {
     // Wait for all references to go away
     boolean didShutdown = false;
     boolean interrupted = false;
-    while (refCount > 0) {
+    while (count > 0) {
       if (!didShutdown) {
         try {
           // Calling shutdown on the socket will interrupt blocking system
@@ -446,7 +362,7 @@ public void close() throws IOException {
       } catch (InterruptedException e) {
         interrupted = true;
       }
-      refCount = status.getReferenceCount();
+      count = refCount.getReferenceCount();
     }
 
     // At this point, nobody has a reference to the file descriptor, 
@@ -478,13 +394,13 @@ private native static void sendFileDescriptors0(int fd,
    */
   public void sendFileDescriptors(FileDescriptor descriptors[],
       byte jbuf[], int offset, int length) throws IOException {
-    status.reference();
+    refCount.reference();
     boolean exc = true;
     try {
       sendFileDescriptors0(fd, descriptors, jbuf, offset, length);
       exc = false;
     } finally {
-      status.unreference(exc);
+      unreference(exc);
     }
   }
 
@@ -515,14 +431,14 @@ private static native int receiveFileDescriptors0(int fd,
    */
   public int receiveFileDescriptors(FileDescriptor[] descriptors,
       byte jbuf[], int offset, int length) throws IOException {
-    status.reference();
+    refCount.reference();
     boolean exc = true;
     try {
       int nBytes = receiveFileDescriptors0(fd, descriptors, jbuf, offset, length);
       exc = false;
       return nBytes;
     } finally {
-      status.unreference(exc);
+      unreference(exc);
     }
   }
 
@@ -539,7 +455,7 @@ public int recvFileInputStreams(FileInputStream[] streams, byte buf[],
     for (int i = 0; i < streams.length; i++) {
       streams[i] = null;
     }
-    status.reference();
+    refCount.reference();
     try {
       int ret = receiveFileDescriptors0(fd, descriptors, buf, offset, length);
       for (int i = 0, j = 0; i < descriptors.length; i++) {
@@ -569,7 +485,7 @@ public int recvFileInputStreams(FileInputStream[] streams, byte buf[],
           }
         }
       }
-      status.unreference(!success);
+      unreference(!success);
     }
   }
 
@@ -593,7 +509,7 @@ private native static int readByteBufferDirect0(int fd, ByteBuffer dst,
   public class DomainInputStream extends InputStream {
     @Override
     public int read() throws IOException {
-      status.reference();
+      refCount.reference();
       boolean exc = true;
       try {
         byte b[] = new byte[1];
@@ -601,33 +517,33 @@ public int read() throws IOException {
         exc = false;
         return (ret >= 0) ? b[0] : -1;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
     
     @Override
     public int read(byte b[], int off, int len) throws IOException {
-      status.reference();
+      refCount.reference();
       boolean exc = true;
       try {
         int nRead = DomainSocket.readArray0(DomainSocket.this.fd, b, off, len);
         exc = false;
         return nRead;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
 
     @Override
     public int available() throws IOException {
-      status.reference();
+      refCount.reference();
       boolean exc = true;
       try {
         int nAvailable = DomainSocket.available0(DomainSocket.this.fd);
         exc = false;
         return nAvailable;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
 
@@ -649,7 +565,7 @@ public void close() throws IOException {
 
     @Override
     public void write(int val) throws IOException {
-      status.reference();
+      refCount.reference();
       boolean exc = true;
       try {
         byte b[] = new byte[1];
@@ -657,19 +573,19 @@ public void write(int val) throws IOException {
         DomainSocket.writeArray0(DomainSocket.this.fd, b, 0, 1);
         exc = false;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
 
     @Override
     public void write(byte[] b, int off, int len) throws IOException {
-      status.reference();
-        boolean exc = true;
+      refCount.reference();
+      boolean exc = true;
       try {
         DomainSocket.writeArray0(DomainSocket.this.fd, b, off, len);
         exc = false;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
   }
@@ -688,7 +604,7 @@ public void close() throws IOException {
 
     @Override
     public int read(ByteBuffer dst) throws IOException {
-      status.reference();
+      refCount.reference();
       boolean exc = true;
       try {
         int nread = 0;
@@ -710,7 +626,7 @@ public int read(ByteBuffer dst) throws IOException {
         exc = false;
         return nread;
       } finally {
-        status.unreference(exc);
+        unreference(exc);
       }
     }
   }

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

@@ -0,0 +1,478 @@
+/**
+ * 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.unix;
+
+import java.io.Closeable;
+import java.io.EOFException;
+
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.io.IOUtils;
+
+import java.io.IOException;
+import java.nio.channels.ClosedChannelException;
+import java.util.Iterator;
+import java.util.LinkedList;
+import java.util.TreeMap;
+import java.util.Map;
+import java.util.concurrent.locks.Condition;
+import java.util.concurrent.locks.ReentrantLock;
+
+import org.apache.commons.lang.SystemUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.util.NativeCodeLoader;
+
+import com.google.common.base.Preconditions;
+import com.google.common.util.concurrent.Uninterruptibles;
+
+/**
+ * The DomainSocketWatcher watches a set of domain sockets to see when they
+ * become readable, or closed.  When one of those events happens, it makes a
+ * callback.
+ *
+ * See {@link DomainSocket} for more information about UNIX domain sockets.
+ */
+@InterfaceAudience.LimitedPrivate("HDFS")
+public final class DomainSocketWatcher extends Thread implements Closeable {
+  static {
+    if (SystemUtils.IS_OS_WINDOWS) {
+      loadingFailureReason = "UNIX Domain sockets are not available on Windows.";
+    } else if (!NativeCodeLoader.isNativeCodeLoaded()) {
+      loadingFailureReason = "libhadoop cannot be loaded.";
+    } else {
+      String problem;
+      try {
+        anchorNative();
+        problem = null;
+      } catch (Throwable t) {
+        problem = "DomainSocketWatcher#anchorNative got error: " +
+          t.getMessage();
+      }
+      loadingFailureReason = problem;
+    }
+  }
+
+  static Log LOG = LogFactory.getLog(DomainSocketWatcher.class);
+
+  /**
+   * The reason why DomainSocketWatcher is not available, or null if it is
+   * available.
+   */
+  private final static String loadingFailureReason;
+
+  /**
+   * Initializes the native library code.
+   */
+  private static native void anchorNative();
+
+  interface Handler {
+    /**
+     * Handles an event on a socket.  An event may be the socket becoming
+     * readable, or the remote end being closed.
+     *
+     * @param sock    The socket that the event occurred on.
+     * @return        Whether we should close the socket.
+     */
+    boolean handle(DomainSocket sock);
+  }
+
+  /**
+   * Handler for {DomainSocketWatcher#notificationSockets[1]}
+   */
+  private class NotificationHandler implements Handler {
+    public boolean handle(DomainSocket sock) {
+      try {
+        if (LOG.isTraceEnabled()) {
+          LOG.trace(this + ": NotificationHandler: doing a read on " +
+            sock.fd);
+        }
+        if (sock.getInputStream().read() == -1) {
+          if (LOG.isTraceEnabled()) {
+            LOG.trace(this + ": NotificationHandler: got EOF on " + sock.fd);
+          }
+          throw new EOFException();
+        }
+        if (LOG.isTraceEnabled()) {
+          LOG.trace(this + ": NotificationHandler: read succeeded on " +
+            sock.fd);
+        }
+        return false;
+      } catch (IOException e) {
+        if (LOG.isTraceEnabled()) {
+          LOG.trace(this + ": NotificationHandler: setting closed to " +
+              "true for " + sock.fd);
+        }
+        closed = true;
+        return true;
+      }
+    }
+  }
+
+  private static class Entry {
+    final DomainSocket socket;
+    final Handler handler;
+
+    Entry(DomainSocket socket, Handler handler) {
+      this.socket = socket;
+      this.handler = handler;
+    }
+
+    DomainSocket getDomainSocket() {
+      return socket;
+    }
+
+    Handler getHandler() {
+      return handler;
+    }
+  }
+
+  /**
+   * The FdSet is a set of file descriptors that gets passed to poll(2).
+   * It contains a native memory segment, so that we don't have to copy
+   * in the poll0 function.
+   */
+  private static class FdSet {
+    private long data;
+
+    private native static long alloc0();
+
+    FdSet() {
+      data = alloc0();
+    }
+
+    /**
+     * Add a file descriptor to the set.
+     *
+     * @param fd   The file descriptor to add.
+     */
+    native void add(int fd);
+
+    /**
+     * Remove a file descriptor from the set.
+     *
+     * @param fd   The file descriptor to remove.
+     */
+    native void remove(int fd);
+
+    /**
+     * Get an array containing all the FDs marked as readable.
+     * Also clear the state of all FDs.
+     *
+     * @return     An array containing all of the currently readable file
+     *             descriptors.
+     */
+    native int[] getAndClearReadableFds();
+
+    /**
+     * Close the object and de-allocate the memory used.
+     */
+    native void close();
+  }
+
+  /**
+   * Lock which protects toAdd, toRemove, and closed.
+   */
+  private final ReentrantLock lock = new ReentrantLock();
+
+  /**
+   * Condition variable which indicates that toAdd and toRemove have been
+   * processed.
+   */
+  private final Condition processedCond = lock.newCondition();
+
+  /**
+   * Entries to add.
+   */
+  private final LinkedList<Entry> toAdd =
+      new LinkedList<Entry>();
+
+  /**
+   * Entries to remove.
+   */
+  private final TreeMap<Integer, DomainSocket> toRemove =
+      new TreeMap<Integer, DomainSocket>();
+
+  /**
+   * Maximum length of time to go between checking whether the interrupted
+   * bit has been set for this thread.
+   */
+  private final int interruptCheckPeriodMs;
+
+  /**
+   * A pair of sockets used to wake up the thread after it has called poll(2).
+   */
+  private final DomainSocket notificationSockets[];
+
+  /**
+   * Whether or not this DomainSocketWatcher is closed.
+   */
+  private boolean closed = false;
+
+  public DomainSocketWatcher(int interruptCheckPeriodMs) throws IOException {
+    if (loadingFailureReason != null) {
+      throw new UnsupportedOperationException(loadingFailureReason);
+    }
+    notificationSockets = DomainSocket.socketpair();
+    this.interruptCheckPeriodMs = interruptCheckPeriodMs;
+    Preconditions.checkArgument(interruptCheckPeriodMs > 0);
+    watcherThread.start();
+  }
+
+  /**
+   * Close the DomainSocketWatcher and wait for its thread to terminate.
+   *
+   * If there is more than one close, all but the first will be ignored.
+   */
+  @Override
+  public void close() throws IOException {
+    try {
+      lock.lock();
+      if (closed) return;
+      LOG.info(this + ": closing");
+      closed = true;
+    } finally {
+      lock.unlock();
+    }
+    // Close notificationSockets[0], so that notificationSockets[1] gets an EOF
+    // event.  This will wake up the thread immediately if it is blocked inside
+    // the select() system call.
+    notificationSockets[0].close();
+    // Wait for the select thread to terminate.
+    Uninterruptibles.joinUninterruptibly(watcherThread);
+  }
+
+  /**
+   * Add a socket.
+   *
+   * @param sock     The socket to add.  It is an error to re-add a socket that
+   *                   we are already watching.
+   * @param handler  The handler to associate with this socket.  This may be
+   *                   called any time after this function is called.
+   */
+  public void add(DomainSocket sock, Handler handler) {
+    try {
+      lock.lock();
+      checkNotClosed();
+      Entry entry = new Entry(sock, handler);
+      try {
+        sock.refCount.reference();
+      } catch (ClosedChannelException e) {
+        Preconditions.checkArgument(false,
+            "tried to add a closed DomainSocket to " + this);
+      }
+      toAdd.add(entry);
+      kick();
+      while (true) {
+        try {
+          processedCond.await();
+        } catch (InterruptedException e) {
+          this.interrupt();
+        }
+        if (!toAdd.contains(entry)) {
+          break;
+        }
+        checkNotClosed();
+      }
+    } finally {
+      lock.unlock();
+    }
+  }
+
+  /**
+   * Remove a socket.  Its handler will be called.
+   *
+   * @param sock     The socket to remove.
+   */
+  public void remove(DomainSocket sock) {
+    try {
+      lock.lock();
+      checkNotClosed();
+      toRemove.put(sock.fd, sock);
+      kick();
+      while (true) {
+        try {
+          processedCond.await();
+        } catch (InterruptedException e) {
+          this.interrupt();
+        }
+        if (!toRemove.containsKey(sock.fd)) {
+          break;
+        }
+        checkNotClosed();
+      }
+    } finally {
+      lock.unlock();
+    }
+  }
+
+  /**
+   * Wake up the DomainSocketWatcher thread.
+   */
+  private void kick() {
+    try {
+      notificationSockets[0].getOutputStream().write(0);
+    } catch (IOException e) {
+      LOG.error(this + ": error writing to notificationSockets[0]", e);
+    }
+  }
+
+  /**
+   * Check that the DomainSocketWatcher is not closed.
+   * Must be called while holding the lock.
+   */
+  private void checkNotClosed() {
+    Preconditions.checkState(lock.isHeldByCurrentThread());
+    if (closed) {
+      throw new RuntimeException("DomainSocketWatcher is closed.");
+    }
+  }
+
+  private void sendCallback(String caller, TreeMap<Integer, Entry> entries,
+      FdSet fdSet, int fd) {
+    if (LOG.isTraceEnabled()) {
+      LOG.trace(this + ": " + caller + " starting sendCallback for fd " + fd);
+    }
+    Entry entry = entries.get(fd);
+    Preconditions.checkNotNull(entry,
+        this + ": fdSet contained " + fd + ", which we were " +
+        "not tracking.");
+    DomainSocket sock = entry.getDomainSocket();
+    if (entry.getHandler().handle(sock)) {
+      if (LOG.isTraceEnabled()) {
+        LOG.trace(this + ": " + caller + ": closing fd " + fd +
+            " at the request of the handler.");
+      }
+      if (toRemove.remove(fd) != null) {
+        if (LOG.isTraceEnabled()) {
+          LOG.trace(this + ": " + caller + " : sendCallback processed fd " +
+            fd  + " in toRemove.");
+        }
+      }
+      try {
+        sock.refCount.unreferenceCheckClosed();
+      } catch (IOException e) {
+        Preconditions.checkArgument(false,
+            this + ": file descriptor " + sock.fd + " was closed while " +
+            "still in the poll(2) loop.");
+      }
+      IOUtils.cleanup(LOG, sock);
+      entries.remove(fd);
+      fdSet.remove(fd);
+    } else {
+      if (LOG.isTraceEnabled()) {
+        LOG.trace(this + ": " + caller + ": sendCallback not " +
+            "closing fd " + fd);
+      }
+    }
+  }
+
+  private final Thread watcherThread = new Thread(new Runnable() {
+    @Override
+    public void run() {
+      LOG.info(this + ": starting with interruptCheckPeriodMs = " +
+          interruptCheckPeriodMs);
+      final TreeMap<Integer, Entry> entries = new TreeMap<Integer, Entry>();
+      FdSet fdSet = new FdSet();
+      addNotificationSocket(entries, fdSet);
+      try {
+        while (true) {
+          lock.lock();
+          try {
+            for (int fd : fdSet.getAndClearReadableFds()) {
+              sendCallback("getAndClearReadableFds", entries, fdSet, fd);
+            }
+            if (!(toAdd.isEmpty() && toRemove.isEmpty())) {
+              // Handle pending additions (before pending removes).
+              for (Iterator<Entry> iter = toAdd.iterator(); iter.hasNext(); ) {
+                Entry entry = iter.next();
+                DomainSocket sock = entry.getDomainSocket();
+                Entry prevEntry = entries.put(sock.fd, entry);
+                Preconditions.checkState(prevEntry == null,
+                    this + ": tried to watch a file descriptor that we " +
+                    "were already watching: " + sock);
+                if (LOG.isTraceEnabled()) {
+                  LOG.trace(this + ": adding fd " + sock.fd);
+                }
+                fdSet.add(sock.fd);
+                iter.remove();
+              }
+              // Handle pending removals
+              while (true) {
+                Map.Entry<Integer, DomainSocket> entry = toRemove.firstEntry();
+                if (entry == null) break;
+                sendCallback("handlePendingRemovals",
+                    entries, fdSet, entry.getValue().fd);
+              }
+              processedCond.signalAll();
+            }
+            // Check if the thread should terminate.  Doing this check now is
+            // easier than at the beginning of the loop, since we know toAdd and
+            // toRemove are now empty and processedCond has been notified if it
+            // needed to be.
+            if (closed) {
+              LOG.info(toString() + " thread terminating.");
+              return;
+            }
+            // Check if someone sent our thread an InterruptedException while we
+            // were waiting in poll().
+            if (Thread.interrupted()) {
+              throw new InterruptedException();
+            }
+          } finally {
+            lock.unlock();
+          }
+          doPoll0(interruptCheckPeriodMs, fdSet);
+        }
+      } catch (InterruptedException e) {
+        LOG.info(toString() + " terminating on InterruptedException");
+      } catch (IOException e) {
+        LOG.error(toString() + " terminating on IOException", e);
+      } finally {
+        for (Entry entry : entries.values()) {
+          sendCallback("close", entries, fdSet, entry.getDomainSocket().fd);
+        }
+        entries.clear();
+        fdSet.close();
+      }
+    }
+  });
+
+  private void addNotificationSocket(final TreeMap<Integer, Entry> entries,
+      FdSet fdSet) {
+    entries.put(notificationSockets[1].fd, 
+        new Entry(notificationSockets[1], new NotificationHandler()));
+    try {
+      notificationSockets[1].refCount.reference();
+    } catch (IOException e) {
+      throw new RuntimeException(e);
+    }
+    fdSet.add(notificationSockets[1].fd);
+    if (LOG.isTraceEnabled()) {
+      LOG.trace(this + ": adding notificationSocket " +
+          notificationSockets[1].fd + ", connected to " +
+          notificationSockets[0].fd);
+    }
+  }
+
+  public String toString() {
+    return "DomainSocketWatcher(" + System.identityHashCode(this) + ")"; 
+  }
+
+  private static native int doPoll0(int maxWaitMs, FdSet readFds)
+      throws IOException;
+}

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

@@ -30,6 +30,8 @@
 import org.apache.hadoop.security.UserGroupInformation;
 import org.apache.hadoop.util.StringUtils;
 
+import com.google.common.annotations.VisibleForTesting;
+
 @InterfaceAudience.Private
 public class ProxyUsers {
 
@@ -177,4 +179,13 @@ private static boolean isWildcardList(Collection<String> list) {
       (list.contains("*"));
   }
 
+  @VisibleForTesting
+  public static Map<String, Collection<String>> getProxyGroups() {
+    return proxyGroups;
+  }
+
+  @VisibleForTesting
+  public static Map<String, Collection<String>> getProxyHosts() {
+    return proxyHosts;
+  }
 }

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

@@ -33,6 +33,8 @@
 import org.apache.hadoop.security.SecurityUtil;
 import org.apache.hadoop.security.UserGroupInformation;
 
+import com.google.common.annotations.VisibleForTesting;
+
 /**
  * An authorization manager which handles service-level authorization
  * for incoming service requests.
@@ -120,19 +122,23 @@ public synchronized void refresh(Configuration conf,
     // Make a copy of the original config, and load the policy file
     Configuration policyConf = new Configuration(conf);
     policyConf.addResource(policyFile);
-    
+    refreshWithConfiguration(policyConf, provider);
+  }
+
+  public synchronized void refreshWithConfiguration(Configuration conf,
+      PolicyProvider provider) {
     final Map<Class<?>, AccessControlList> newAcls =
-      new IdentityHashMap<Class<?>, AccessControlList>();
+        new IdentityHashMap<Class<?>, AccessControlList>();
 
     // Parse the config file
     Service[] services = provider.getServices();
     if (services != null) {
       for (Service service : services) {
-        AccessControlList acl = 
-          new AccessControlList(
-              policyConf.get(service.getServiceKey(), 
-                             AccessControlList.WILDCARD_ACL_VALUE)
-              );
+        AccessControlList acl =
+            new AccessControlList(
+                conf.get(service.getServiceKey(),
+                    AccessControlList.WILDCARD_ACL_VALUE)
+            );
         newAcls.put(service.getProtocol(), acl);
       }
     }
@@ -141,8 +147,13 @@ public synchronized void refresh(Configuration conf,
     protocolToAcl = newAcls;
   }
 
-  // Package-protected for use in tests.
-  Set<Class<?>> getProtocolsWithAcls() {
+  @VisibleForTesting
+  public Set<Class<?>> getProtocolsWithAcls() {
     return protocolToAcl.keySet();
   }
+
+  @VisibleForTesting
+  public AccessControlList getProtocolsAcls(Class<?> className) {
+    return protocolToAcl.get(className);
+  }
 }

hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/service/CompositeService.java

@@ -19,7 +19,6 @@
 package org.apache.hadoop.service;
 
 import java.util.ArrayList;
-import java.util.Collections;
 import java.util.List;
 
 import org.apache.commons.logging.Log;
@@ -54,13 +53,13 @@ public CompositeService(String name) {
   }
 
   /**
-   * Get an unmodifiable list of services
+   * Get a cloned list of services
    * @return a list of child services at the time of invocation -
    * added services will not be picked up.
    */
   public List<Service> getServices() {
     synchronized (serviceList) {
-      return Collections.unmodifiableList(serviceList);
+      return new ArrayList<Service>(serviceList);
     }
   }
 

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

@@ -0,0 +1,125 @@
+/**
+ * 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.util;
+
+import java.nio.channels.AsynchronousCloseException;
+import java.nio.channels.ClosedChannelException;
+import java.util.concurrent.atomic.AtomicInteger;
+
+import com.google.common.base.Preconditions;
+
+/**
+ * A closeable object that maintains a reference count.
+ *
+ * Once the object is closed, attempting to take a new reference will throw
+ * ClosedChannelException.
+ */
+public class CloseableReferenceCount {
+  /**
+   * Bit mask representing a closed domain socket.
+   */
+  private static final int STATUS_CLOSED_MASK = 1 << 30;
+
+  /**
+   * The status bits.
+   *
+   * Bit 30: 0 = open, 1 = closed.
+   * Bits 29 to 0: the reference count.
+   */
+  private final AtomicInteger status = new AtomicInteger(0);
+
+  public CloseableReferenceCount() { }
+
+  /**
+   * Increment the reference count.
+   *
+   * @throws ClosedChannelException      If the status is closed.
+   */
+  public void reference() throws ClosedChannelException {
+    int curBits = status.incrementAndGet();
+    if ((curBits & STATUS_CLOSED_MASK) != 0) {
+      status.decrementAndGet();
+      throw new ClosedChannelException();
+    }
+  }
+
+  /**
+   * Decrement the reference count.
+   *
+   * @return          True if the object is closed and has no outstanding
+   *                  references.
+   */
+  public boolean unreference() {
+    int newVal = status.decrementAndGet();
+    Preconditions.checkState(newVal != 0xffffffff,
+        "called unreference when the reference count was already at 0.");
+    return newVal == STATUS_CLOSED_MASK;
+  }
+
+  /**
+   * Decrement the reference count, checking to make sure that the
+   * CloseableReferenceCount is not closed.
+   *
+   * @throws AsynchronousCloseException  If the status is closed.
+   */
+  public void unreferenceCheckClosed() throws ClosedChannelException {
+    int newVal = status.decrementAndGet();
+    if ((newVal & STATUS_CLOSED_MASK) != 0) {
+      throw new AsynchronousCloseException();
+    }
+  }
+
+  /**
+   * Return true if the status is currently open.
+   *
+   * @return                 True if the status is currently open.
+   */
+  public boolean isOpen() {
+    return ((status.get() & STATUS_CLOSED_MASK) == 0);
+  }
+
+  /**
+   * Mark the status as closed.
+   *
+   * Once the status is closed, it cannot be reopened.
+   *
+   * @return                         The current reference count.
+   * @throws ClosedChannelException  If someone else closes the object
+   *                                 before we do.
+   */
+  public int setClosed() throws ClosedChannelException {
+    while (true) {
+      int curBits = status.get();
+      if ((curBits & STATUS_CLOSED_MASK) != 0) {
+        throw new ClosedChannelException();
+      }
+      if (status.compareAndSet(curBits, curBits | STATUS_CLOSED_MASK)) {
+        return curBits & (~STATUS_CLOSED_MASK);
+      }
+    }
+  }
+
+  /**
+   * Get the current reference count.
+   *
+   * @return                 The current reference count.
+   */
+  public int getReferenceCount() {
+    return status.get() & (~STATUS_CLOSED_MASK);
+  }
+}

hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/nativeio/NativeIO.c

@@ -18,6 +18,7 @@
 
 #include "org_apache_hadoop.h"
 #include "org_apache_hadoop_io_nativeio_NativeIO.h"
+#include "org_apache_hadoop_io_nativeio_NativeIO_POSIX.h"
 
 #ifdef UNIX
 #include <assert.h>
@@ -49,6 +50,10 @@
 #include "file_descriptor.h"
 #include "errno_enum.h"
 
+#define MMAP_PROT_READ org_apache_hadoop_io_nativeio_NativeIO_POSIX_MMAP_PROT_READ
+#define MMAP_PROT_WRITE org_apache_hadoop_io_nativeio_NativeIO_POSIX_MMAP_PROT_WRITE
+#define MMAP_PROT_EXEC org_apache_hadoop_io_nativeio_NativeIO_POSIX_MMAP_PROT_EXEC
+
 // the NativeIO$POSIX$Stat inner class and its constructor
 static jclass stat_clazz;
 static jmethodID stat_ctor;
@@ -661,6 +666,55 @@ cleanup:
 #endif
 }
 
+JNIEXPORT jlong JNICALL 
+Java_org_apache_hadoop_io_nativeio_NativeIO_00024POSIX_mmap(
+  JNIEnv *env, jclass clazz, jobject jfd, jint jprot,
+  jboolean jshared, jlong length)
+{
+#ifdef UNIX
+  void *addr = 0;
+  int prot, flags, fd;
+  
+  prot = ((jprot & MMAP_PROT_READ) ? PROT_READ : 0) |
+         ((jprot & MMAP_PROT_WRITE) ? PROT_WRITE : 0) |
+         ((jprot & MMAP_PROT_EXEC) ? PROT_EXEC : 0);
+  flags = (jshared == JNI_TRUE) ? MAP_SHARED : MAP_PRIVATE;
+  fd = fd_get(env, jfd);
+  addr = mmap(NULL, length, prot, flags, fd, 0);
+  if (addr == MAP_FAILED) {
+    throw_ioe(env, errno);
+  }
+  return (jlong)(intptr_t)addr;
+#endif  //   UNIX
+
+#ifdef WINDOWS
+  THROW(env, "java/io/IOException",
+    "The function POSIX.mmap() is not supported on Windows");
+  return NULL;
+#endif
+}
+
+JNIEXPORT void JNICALL 
+Java_org_apache_hadoop_io_nativeio_NativeIO_00024POSIX_munmap(
+  JNIEnv *env, jclass clazz, jlong jaddr, jlong length)
+{
+#ifdef UNIX
+  void *addr;
+
+  addr = (void*)(intptr_t)jaddr;
+  if (munmap(addr, length) < 0) {
+    throw_ioe(env, errno);
+  }
+#endif  //   UNIX
+
+#ifdef WINDOWS
+  THROW(env, "java/io/IOException",
+    "The function POSIX.munmap() is not supported on Windows");
+  return NULL;
+#endif
+}
+
+
 /*
  * static native String getGroupName(int gid);
  *
@@ -1012,4 +1066,3 @@ JNIEnv *env, jclass clazz)
 /**
  * vim: sw=2: ts=2: et:
  */
-

hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/io/nativeio/SharedFileDescriptorFactory.c

@@ -0,0 +1,162 @@
+/*
+ * 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.
+ */
+
+#include "org_apache_hadoop.h"
+
+#ifdef UNIX
+
+#include "exception.h"
+#include "file_descriptor.h"
+#include "org_apache_hadoop.h"
+#include "org_apache_hadoop_io_nativeio_SharedFileDescriptorFactory.h"
+
+#include <dirent.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <pthread.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+static pthread_mutex_t g_rand_lock = PTHREAD_MUTEX_INITIALIZER;
+
+JNIEXPORT void JNICALL
+Java_org_apache_hadoop_io_nativeio_SharedFileDescriptorFactory_deleteStaleTemporaryFiles0(
+  JNIEnv *env, jclass clazz, jstring jprefix, jstring jpath)
+{
+  const char *prefix = NULL, *path = NULL;
+  char target[PATH_MAX];
+  jthrowable jthr;
+  DIR *dp = NULL;
+  struct dirent *de;
+
+  prefix = (*env)->GetStringUTFChars(env, jprefix, NULL);
+  if (!prefix) goto done; // exception raised
+  path = (*env)->GetStringUTFChars(env, jpath, NULL);
+  if (!path) goto done; // exception raised
+
+  dp = opendir(path);
+  if (!dp) {
+    int ret = errno;
+    jthr = newIOException(env, "opendir(%s) error %d: %s",
+                          path, ret, terror(ret));
+    (*env)->Throw(env, jthr);
+    goto done;
+  }
+  while ((de = readdir(dp))) {
+    if (strncmp(prefix, de->d_name, strlen(prefix)) == 0) {
+      int ret = snprintf(target, PATH_MAX, "%s/%s", path, de->d_name);
+      if ((0 < ret) && (ret < PATH_MAX)) {
+        unlink(target);
+      }
+    }
+  }
+
+done:
+  if (dp) {
+    closedir(dp);
+  }
+  if (prefix) {
+    (*env)->ReleaseStringUTFChars(env, jprefix, prefix);
+  }
+  if (path) {
+    (*env)->ReleaseStringUTFChars(env, jpath, path);
+  }
+}
+
+JNIEXPORT jobject JNICALL
+Java_org_apache_hadoop_io_nativeio_SharedFileDescriptorFactory_createDescriptor0(
+  JNIEnv *env, jclass clazz, jstring jprefix, jstring jpath, jint length)
+{
+  const char *prefix = NULL, *path = NULL;
+  char target[PATH_MAX];
+  int ret, fd = -1, rnd;
+  jthrowable jthr;
+  jobject jret = NULL;
+
+  prefix = (*env)->GetStringUTFChars(env, jprefix, NULL);
+  if (!prefix) goto done; // exception raised
+  path = (*env)->GetStringUTFChars(env, jpath, NULL);
+  if (!path) goto done; // exception raised
+
+  pthread_mutex_lock(&g_rand_lock);
+  rnd = rand();
+  pthread_mutex_unlock(&g_rand_lock);
+  while (1) {
+    ret = snprintf(target, PATH_MAX, "%s/%s_%d",
+                   path, prefix, rnd);
+    if (ret < 0) {
+      jthr = newIOException(env, "snprintf error");
+      (*env)->Throw(env, jthr);
+      goto done;
+    } else if (ret >= PATH_MAX) {
+      jthr = newIOException(env, "computed path was too long.");
+      (*env)->Throw(env, jthr);
+      goto done;
+    }
+    fd = open(target, O_CREAT | O_EXCL | O_RDWR, 0700);
+    if (fd >= 0) break; // success
+    ret = errno;
+    if (ret == EEXIST) {
+      // Bad luck -- we got a very rare collision here between us and 
+      // another DataNode (or process).  Try again.
+      continue;
+    } else if (ret == EINTR) {
+      // Most of the time, this error is only possible when opening FIFOs.
+      // But let's be thorough.
+      continue;
+    }
+    jthr = newIOException(env, "open(%s, O_CREAT | O_EXCL | O_RDWR) "
+            "failed: error %d (%s)", target, ret, terror(ret));
+    (*env)->Throw(env, jthr);
+    goto done;
+  }
+  if (unlink(target) < 0) {
+    jthr = newIOException(env, "unlink(%s) failed: error %d (%s)",
+                          path, ret, terror(ret));
+    (*env)->Throw(env, jthr);
+    goto done;
+  }
+  if (ftruncate(fd, length) < 0) {
+    jthr = newIOException(env, "ftruncate(%s, %d) failed: error %d (%s)",
+                          path, length, ret, terror(ret));
+    (*env)->Throw(env, jthr);
+    goto done;
+  }
+  jret = fd_create(env, fd); // throws exception on error.
+
+done:
+  if (prefix) {
+    (*env)->ReleaseStringUTFChars(env, jprefix, prefix);
+  }
+  if (path) {
+    (*env)->ReleaseStringUTFChars(env, jpath, path);
+  }
+  if (!jret) {
+    if (fd >= 0) {
+      close(fd);
+    }
+  }
+  return jret;
+}
+
+#endif

hadoop-common-project/hadoop-common/src/main/native/src/org/apache/hadoop/net/unix/DomainSocketWatcher.c

@@ -0,0 +1,247 @@
+/*
+ * 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.
+ */
+
+#include "config.h"
+#include "exception.h"
+#include "org_apache_hadoop.h"
+#include "org_apache_hadoop_net_unix_DomainSocketWatcher.h"
+
+#include <errno.h>
+#include <fcntl.h>
+#include <jni.h>
+#include <poll.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/select.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+static jfieldID fd_set_data_fid;
+
+#define FD_SET_DATA_MIN_SIZE 2
+
+struct fd_set_data {
+  /**
+   * Number of fds we have allocated space for.
+   */
+  int alloc_size;
+
+  /**
+   * Number of fds actually in use.
+   */
+  int used_size;
+
+  /**
+   * Beginning of pollfd data.
+   */
+  struct pollfd pollfd[0];
+};
+
+JNIEXPORT void JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_anchorNative(
+JNIEnv *env, jclass clazz)
+{
+  jclass fd_set_class;
+
+  fd_set_class = (*env)->FindClass(env,
+          "org/apache/hadoop/net/unix/DomainSocketWatcher$FdSet");
+  if (!fd_set_class) return; // exception raised
+  fd_set_data_fid = (*env)->GetFieldID(env, fd_set_class, "data", "J");
+  if (!fd_set_data_fid) return; // exception raised
+}
+
+JNIEXPORT jlong JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_00024FdSet_alloc0(
+JNIEnv *env, jclass clazz)
+{
+  struct fd_set_data *sd;
+
+  sd = calloc(1, sizeof(struct fd_set_data) +
+              (sizeof(struct pollfd) * FD_SET_DATA_MIN_SIZE));
+  if (!sd) {
+    (*env)->Throw(env, newRuntimeException(env, "out of memory allocating "
+                                            "DomainSocketWatcher#FdSet"));
+    return 0L;
+  }
+  sd->alloc_size = FD_SET_DATA_MIN_SIZE;
+  sd->used_size = 0;
+  return (jlong)(intptr_t)sd;
+}
+
+JNIEXPORT void JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_00024FdSet_add(
+JNIEnv *env, jobject obj, jint fd)
+{
+  struct fd_set_data *sd, *nd;
+  struct pollfd *pollfd;
+
+  sd = (struct fd_set_data*)(intptr_t)(*env)->
+    GetLongField(env, obj, fd_set_data_fid);
+  if (sd->used_size + 1 > sd->alloc_size) {
+    nd = realloc(sd, sizeof(struct fd_set_data) +
+            (sizeof(struct pollfd) * sd->alloc_size * 2));
+    if (!nd) {
+      (*env)->Throw(env, newRuntimeException(env, "out of memory adding "
+            "another fd to DomainSocketWatcher#FdSet.  we have %d already",
+            sd->alloc_size));
+      return;
+    }
+    nd->alloc_size = nd->alloc_size * 2;
+    (*env)->SetLongField(env, obj, fd_set_data_fid, (jlong)(intptr_t)nd);
+    sd = nd;
+  }
+  pollfd = &sd->pollfd[sd->used_size];
+  sd->used_size++;
+  pollfd->fd = fd;
+  pollfd->events = POLLIN;
+  pollfd->revents = 0;
+}
+
+JNIEXPORT void JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_00024FdSet_remove(
+JNIEnv *env, jobject obj, jint fd)
+{
+  struct fd_set_data *sd;
+  struct pollfd *pollfd, *last_pollfd;
+  int used_size, i;
+
+  sd = (struct fd_set_data*)(intptr_t)(*env)->
+      GetLongField(env, obj, fd_set_data_fid);
+  used_size = sd->used_size;
+  for (i = 0; i < used_size; i++) {
+    pollfd = sd->pollfd + i;
+    if (pollfd->fd == fd) break;
+  }
+  if (i == used_size) {
+    (*env)->Throw(env, newRuntimeException(env, "failed to remove fd %d "
+          "from the FdSet because it was never present.", fd));
+    return;
+  }
+  last_pollfd = sd->pollfd + (used_size - 1);
+  if (used_size > 1) {
+    // Move last pollfd to the new empty slot if needed
+    pollfd->fd = last_pollfd->fd;
+    pollfd->events = last_pollfd->events;
+    pollfd->revents = last_pollfd->revents;
+  }
+  memset(last_pollfd, 0, sizeof(struct pollfd));
+  sd->used_size--;
+}
+
+JNIEXPORT jobject JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_00024FdSet_getAndClearReadableFds(
+JNIEnv *env, jobject obj)
+{
+  int *carr = NULL;
+  jobject jarr = NULL;
+  struct fd_set_data *sd;
+  int used_size, num_readable = 0, i, j;
+  jthrowable jthr = NULL;
+
+  sd = (struct fd_set_data*)(intptr_t)(*env)->
+      GetLongField(env, obj, fd_set_data_fid);
+  used_size = sd->used_size;
+  for (i = 0; i < used_size; i++) {
+    if (sd->pollfd[i].revents & POLLIN) {
+      num_readable++;
+    } else {
+      sd->pollfd[i].revents = 0;
+    }
+  }
+  if (num_readable > 0) {
+    carr = malloc(sizeof(int) * num_readable);
+    if (!carr) {
+      jthr = newRuntimeException(env, "failed to allocate a temporary array "
+            "of %d ints", num_readable);
+      goto done;
+    }
+    j = 0;
+    for (i = 0; ((i < used_size) && (j < num_readable)); i++) {
+      if (sd->pollfd[i].revents & POLLIN) {
+        carr[j] = sd->pollfd[i].fd;
+        j++;
+        sd->pollfd[i].revents = 0;
+      }
+    }
+    if (j != num_readable) {
+      jthr = newRuntimeException(env, "failed to fill entire carr "
+            "array of size %d: only filled %d elements", num_readable, j);
+      goto done;
+    }
+  }
+  jarr = (*env)->NewIntArray(env, num_readable);
+  if (!jarr) {
+    jthr = (*env)->ExceptionOccurred(env);
+    (*env)->ExceptionClear(env);
+    goto done;
+  }
+  if (num_readable > 0) {
+    (*env)->SetIntArrayRegion(env, jarr, 0, num_readable, carr);
+    jthr = (*env)->ExceptionOccurred(env);
+    if (jthr) {
+      (*env)->ExceptionClear(env);
+      goto done;
+    }
+  }
+
+done:
+  free(carr);
+  if (jthr) {
+    (*env)->DeleteLocalRef(env, jarr);
+    jarr = NULL;
+  }
+  return jarr;
+}
+
+JNIEXPORT void JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_00024FdSet_close(
+JNIEnv *env, jobject obj)
+{
+  struct fd_set_data *sd;
+
+  sd = (struct fd_set_data*)(intptr_t)(*env)->
+      GetLongField(env, obj, fd_set_data_fid);
+  if (sd) {
+    free(sd);
+    (*env)->SetLongField(env, obj, fd_set_data_fid, 0L);
+  }
+}
+
+JNIEXPORT jint JNICALL
+Java_org_apache_hadoop_net_unix_DomainSocketWatcher_doPoll0(
+JNIEnv *env, jclass clazz, jint checkMs, jobject fdSet)
+{
+  struct fd_set_data *sd;
+  int ret, err;
+
+  sd = (struct fd_set_data*)(intptr_t)(*env)->
+      GetLongField(env, fdSet, fd_set_data_fid);
+  ret = poll(sd->pollfd, sd->used_size, checkMs);
+  if (ret >= 0) {
+    return ret;
+  }
+  err = errno;
+  if (err != EINTR) { // treat EINTR as 0 fds ready
+    (*env)->Throw(env, newIOException(env,
+            "poll(2) failed with error code %d: %s", err, terror(err)));
+  }
+  return 0;
+}

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

@@ -352,7 +352,8 @@ Configuration for <<<conf/core-site.xml>>>
 | | | This value is deprecated. Use dfs.http.policy |
 *-------------------------+-------------------------+------------------------+
 | <<<dfs.http.policy>>> | <HTTP_ONLY> or <HTTPS_ONLY> or <HTTP_AND_HTTPS> | |
-| | | HTTPS_ONLY turns off http access |
+| | | HTTPS_ONLY turns off http access. This option takes precedence over |
+| | | the deprecated configuration dfs.https.enable and hadoop.ssl.enabled. |
 *-------------------------+-------------------------+------------------------+
 | <<<dfs.namenode.https-address>>> | <nn_host_fqdn:50470> | |
 *-------------------------+-------------------------+------------------------+

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/crypto/key/TestKeyShell.java

@@ -41,7 +41,7 @@ public void setup() throws Exception {
   
   @Test
   public void testKeySuccessfulKeyLifecycle() throws Exception {
-    outContent.flush();
+    outContent.reset();
     String[] args1 = {"create", "key1", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     int rc = 0;
@@ -52,14 +52,14 @@ public void testKeySuccessfulKeyLifecycle() throws Exception {
     assertTrue(outContent.toString().contains("key1 has been successfully " +
     		"created."));
 
-    outContent.flush();
+    outContent.reset();
     String[] args2 = {"list", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     rc = ks.run(args2);
     assertEquals(0, rc);
     assertTrue(outContent.toString().contains("key1"));
 
-    outContent.flush();
+    outContent.reset();
     String[] args3 = {"roll", "key1", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     rc = ks.run(args3);
@@ -67,7 +67,7 @@ public void testKeySuccessfulKeyLifecycle() throws Exception {
     assertTrue(outContent.toString().contains("key1 has been successfully " +
     		"rolled."));
 
-    outContent.flush();
+    outContent.reset();
     String[] args4 = {"delete", "key1", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     rc = ks.run(args4);
@@ -75,12 +75,12 @@ public void testKeySuccessfulKeyLifecycle() throws Exception {
     assertTrue(outContent.toString().contains("key1 has been successfully " +
     		"deleted."));
 
-    outContent.flush();
+    outContent.reset();
     String[] args5 = {"list", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     rc = ks.run(args5);
     assertEquals(0, rc);
-    assertTrue(outContent.toString().contains("key1"));
+    assertFalse(outContent.toString(), outContent.toString().contains("key1"));
   }
   
   @Test
@@ -165,7 +165,7 @@ public void testFullCipher() throws Exception {
     assertTrue(outContent.toString().contains("key1 has been successfully " +
     		"created."));
 
-    outContent.flush();
+    outContent.reset();
     String[] args2 = {"delete", "key1", "--provider", 
         "jceks://file" + tmpDir + "/keystore.jceks"};
     rc = ks.run(args2);

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

@@ -0,0 +1,82 @@
+/**
+ * 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.io.nativeio;
+
+import java.io.File;
+import java.io.FileInputStream;
+import java.io.FileOutputStream;
+
+import org.junit.Assert;
+import org.junit.Assume;
+import org.junit.Test;
+import org.apache.commons.lang.SystemUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.fs.FileUtil;
+import org.apache.hadoop.fs.Path;
+
+public class TestSharedFileDescriptorFactory {
+  static final Log LOG = LogFactory.getLog(TestSharedFileDescriptorFactory.class);
+
+  private static final File TEST_BASE =
+      new File(System.getProperty("test.build.data", "/tmp"));
+
+  @Test(timeout=10000)
+  public void testReadAndWrite() throws Exception {
+    Assume.assumeTrue(NativeIO.isAvailable());
+    Assume.assumeTrue(SystemUtils.IS_OS_UNIX);
+    File path = new File(TEST_BASE, "testReadAndWrite");
+    path.mkdirs();
+    SharedFileDescriptorFactory factory =
+        new SharedFileDescriptorFactory("woot_", path.getAbsolutePath());
+    FileInputStream inStream = factory.createDescriptor(4096);
+    FileOutputStream outStream = new FileOutputStream(inStream.getFD());
+    outStream.write(101);
+    inStream.getChannel().position(0);
+    Assert.assertEquals(101, inStream.read());
+    inStream.close();
+    outStream.close();
+    FileUtil.fullyDelete(path);
+  }
+
+  static private void createTempFile(String path) throws Exception {
+    FileOutputStream fos = new FileOutputStream(path);
+    fos.write(101);
+    fos.close();
+  }
+  
+  @Test(timeout=10000)
+  public void testCleanupRemainders() throws Exception {
+    Assume.assumeTrue(NativeIO.isAvailable());
+    Assume.assumeTrue(SystemUtils.IS_OS_UNIX);
+    File path = new File(TEST_BASE, "testCleanupRemainders");
+    path.mkdirs();
+    String remainder1 = path.getAbsolutePath() + 
+        Path.SEPARATOR + "woot2_remainder1";
+    String remainder2 = path.getAbsolutePath() +
+        Path.SEPARATOR + "woot2_remainder2";
+    createTempFile(remainder1);
+    createTempFile(remainder2);
+    new SharedFileDescriptorFactory("woot2_", path.getAbsolutePath());
+    // creating the SharedFileDescriptorFactory should have removed 
+    // the remainders
+    Assert.assertFalse(new File(remainder1).exists());
+    Assert.assertFalse(new File(remainder2).exists());
+    FileUtil.fullyDelete(path);
+  }
+}

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/net/unix/TestDomainSocketWatcher.java

@@ -0,0 +1,150 @@
+/**
+ * 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.unix;
+
+import java.util.ArrayList;
+import java.util.Random;
+import java.util.concurrent.CountDownLatch;
+import java.util.concurrent.atomic.AtomicInteger;
+import java.util.concurrent.locks.ReentrantLock;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.junit.Assume;
+import org.junit.Before;
+import org.junit.Test;
+
+import com.google.common.util.concurrent.Uninterruptibles;
+
+public class TestDomainSocketWatcher {
+  static final Log LOG = LogFactory.getLog(TestDomainSocketWatcher.class);
+
+  @Before
+  public void before() {
+    Assume.assumeTrue(DomainSocket.getLoadingFailureReason() == null);
+  }
+
+  /**
+   * Test that we can create a DomainSocketWatcher and then shut it down.
+   */
+  @Test(timeout=60000)
+  public void testCreateShutdown() throws Exception {
+    DomainSocketWatcher watcher = new DomainSocketWatcher(10000000);
+    watcher.close();
+  }
+
+  /**
+   * Test that we can get notifications out a DomainSocketWatcher.
+   */
+  @Test(timeout=180000)
+  public void testDeliverNotifications() throws Exception {
+    DomainSocketWatcher watcher = new DomainSocketWatcher(10000000);
+    DomainSocket pair[] = DomainSocket.socketpair();
+    final CountDownLatch latch = new CountDownLatch(1);
+    watcher.add(pair[1], new DomainSocketWatcher.Handler() {
+      @Override
+      public boolean handle(DomainSocket sock) {
+        latch.countDown();
+        return true;
+      }
+    });
+    pair[0].close();
+    latch.await();
+    watcher.close();
+  }
+
+  /**
+   * Test that a java interruption can stop the watcher thread
+   */
+  @Test(timeout=60000)
+  public void testInterruption() throws Exception {
+    DomainSocketWatcher watcher = new DomainSocketWatcher(10);
+    watcher.interrupt();
+    Uninterruptibles.joinUninterruptibly(watcher);
+  }
+  
+  @Test(timeout=300000)
+  public void testStress() throws Exception {
+    final int SOCKET_NUM = 250;
+    final ReentrantLock lock = new ReentrantLock();
+    final DomainSocketWatcher watcher = new DomainSocketWatcher(10000000);
+    final ArrayList<DomainSocket[]> pairs = new ArrayList<DomainSocket[]>();
+    final AtomicInteger handled = new AtomicInteger(0);
+
+    final Thread adderThread = new Thread(new Runnable() {
+      @Override
+      public void run() {
+        try {
+          for (int i = 0; i < SOCKET_NUM; i++) {
+            DomainSocket pair[] = DomainSocket.socketpair();
+            watcher.add(pair[1], new DomainSocketWatcher.Handler() {
+              @Override
+              public boolean handle(DomainSocket sock) {
+                handled.incrementAndGet();
+                return true;
+              }
+            });
+            lock.lock();
+            try {
+              pairs.add(pair);
+            } finally {
+              lock.unlock();
+            }
+          }
+        } catch (Throwable e) {
+          LOG.error(e);
+          throw new RuntimeException(e);
+        }
+      }
+    });
+    
+    final Thread removerThread = new Thread(new Runnable() {
+      @Override
+      public void run() {
+        final Random random = new Random();
+        try {
+          while (handled.get() != SOCKET_NUM) {
+            lock.lock();
+            try {
+              if (!pairs.isEmpty()) {
+                int idx = random.nextInt(pairs.size());
+                DomainSocket pair[] = pairs.remove(idx);
+                if (random.nextBoolean()) {
+                  pair[0].close();
+                } else {
+                  watcher.remove(pair[1]);
+                }
+              }
+            } finally {
+              lock.unlock();
+            }
+          }
+        } catch (Throwable e) {
+          LOG.error(e);
+          throw new RuntimeException(e);
+        }
+      }
+    });
+
+    adderThread.start();
+    removerThread.start();
+    Uninterruptibles.joinUninterruptibly(adderThread);
+    Uninterruptibles.joinUninterruptibly(removerThread);
+    watcher.close();
+  }
+}

hadoop-common-project/hadoop-common/src/test/java/org/apache/hadoop/service/TestCompositeService.java

@@ -16,26 +16,20 @@
  * limitations under the License.
  */
 
-package org.apache.hadoop.yarn.util;
+package org.apache.hadoop.service;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.service.Service.STATE;
+import org.junit.Before;
+import org.junit.Test;
 
 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.assertFalse;
 import static org.junit.Assert.assertTrue;
 import static org.junit.Assert.fail;
 
-import org.apache.commons.logging.Log;
-import org.apache.commons.logging.LogFactory;
-import org.apache.hadoop.conf.Configuration;
-import org.apache.hadoop.service.AbstractService;
-import org.apache.hadoop.service.BreakableService;
-import org.apache.hadoop.service.CompositeService;
-import org.apache.hadoop.service.Service;
-import org.apache.hadoop.service.ServiceStateException;
-import org.apache.hadoop.service.Service.STATE;
-import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
-import org.junit.Before;
-import org.junit.Test;
-
 public class TestCompositeService {
 
   private static final int NUM_OF_SERVICES = 5;
@@ -156,7 +150,7 @@ public void testServiceStartup() {
     try {
       serviceManager.start();
       fail("Exception should have been thrown due to startup failure of last service");
-    } catch (YarnRuntimeException e) {
+    } catch (ServiceTestRuntimeException e) {
       for (int i = 0; i < NUM_OF_SERVICES - 1; i++) {
         if (i >= FAILED_SERVICE_SEQ_NUMBER && STOP_ONLY_STARTED_SERVICES) {
           // Failed service state should be INITED
@@ -197,7 +191,7 @@ public void testServiceStop() {
     // Stop the composite service
     try {
       serviceManager.stop();
-    } catch (YarnRuntimeException e) {
+    } catch (ServiceTestRuntimeException e) {
     }
     assertInState(STATE.STOPPED, services);
   }
@@ -335,7 +329,41 @@ public void serviceInit(Configuration conf) {
 
     testService.init(new Configuration());
     assertEquals("Incorrect number of services",
-        1, testService.getServices().size());
+                 1, testService.getServices().size());
+  }
+
+  @Test(timeout = 1000)
+  public void testAddInitedSiblingInInit() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    sibling.init(new Configuration());
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.INITED));
+    parent.init(new Configuration());
+    parent.start();
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
+  }
+
+  @Test(timeout = 1000)
+  public void testAddUninitedSiblingInInit() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.INITED));
+    parent.init(new Configuration());
+    try {
+      parent.start();
+      fail("Expected an exception, got " + parent);
+    } catch (ServiceStateException e) {
+      //expected
+    }
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
   }
 
   @Test
@@ -365,6 +393,105 @@ public void serviceInit(Configuration conf) {
         2, testService.getServices().size());
   }
 
+  @Test(timeout = 1000)
+  public void testAddStartedChildBeforeInit() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService child = new BreakableService();
+    child.init(new Configuration());
+    child.start();
+    AddSiblingService.addChildToService(parent, child);
+    try {
+      parent.init(new Configuration());
+      fail("Expected an exception, got " + parent);
+    } catch (ServiceStateException e) {
+      //expected
+    }
+    parent.stop();
+  }
+
+  @Test(timeout = 1000)
+  public void testAddStoppedChildBeforeInit() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService child = new BreakableService();
+    child.init(new Configuration());
+    child.start();
+    child.stop();
+    AddSiblingService.addChildToService(parent, child);
+    try {
+      parent.init(new Configuration());
+      fail("Expected an exception, got " + parent);
+    } catch (ServiceStateException e) {
+      //expected
+    }
+    parent.stop();
+  }
+
+  @Test(timeout = 1000)
+  public void testAddStartedSiblingInStart() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    sibling.init(new Configuration());
+    sibling.start();
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.STARTED));
+    parent.init(new Configuration());
+    parent.start();
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
+  }
+
+  @Test(timeout = 1000)
+  public void testAddUninitedSiblingInStart() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.STARTED));
+    parent.init(new Configuration());
+    assertInState(STATE.NOTINITED, sibling);
+    parent.start();
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
+  }
+
+  @Test(timeout = 1000)
+  public void testAddStartedSiblingInInit() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    sibling.init(new Configuration());
+    sibling.start();
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.INITED));
+    parent.init(new Configuration());
+    assertInState(STATE.STARTED, sibling);
+    parent.start();
+    assertInState(STATE.STARTED, sibling);
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
+    assertInState(STATE.STOPPED, sibling);
+  }
+
+  @Test(timeout = 1000)
+  public void testAddStartedSiblingInStop() throws Throwable {
+    CompositeService parent = new CompositeService("parent");
+    BreakableService sibling = new BreakableService();
+    sibling.init(new Configuration());
+    sibling.start();
+    parent.addService(new AddSiblingService(parent,
+                                            sibling,
+                                            STATE.STOPPED));
+    parent.init(new Configuration());
+    parent.start();
+    parent.stop();
+    assertEquals("Incorrect number of services",
+                 2, parent.getServices().size());
+  }
+
   public static class CompositeServiceAddingAChild extends CompositeService{
     Service child;
 
@@ -379,7 +506,18 @@ protected void serviceInit(Configuration conf) throws Exception {
       super.serviceInit(conf);
     }
   }
-  
+
+  public static class ServiceTestRuntimeException extends RuntimeException {
+    public ServiceTestRuntimeException(String message) {
+      super(message);
+    }
+  }
+
+  /**
+   * This is a composite service that keeps a count of the number of lifecycle
+   * events called, and can be set to throw a {@link ServiceTestRuntimeException }
+   * during service start or stop
+   */
   public static class CompositeServiceImpl extends CompositeService {
 
     public static boolean isPolicyToStopOnlyStartedServices() {
@@ -408,7 +546,7 @@ protected void serviceInit(Configuration conf) throws Exception {
     @Override
     protected void serviceStart() throws Exception {
       if (throwExceptionOnStart) {
-        throw new YarnRuntimeException("Fake service start exception");
+        throw new ServiceTestRuntimeException("Fake service start exception");
       }
       counter++;
       callSequenceNumber = counter;
@@ -420,7 +558,7 @@ protected void serviceStop() throws Exception {
       counter++;
       callSequenceNumber = counter;
       if (throwExceptionOnStop) {
-        throw new YarnRuntimeException("Fake service stop exception");
+        throw new ServiceTestRuntimeException("Fake service stop exception");
       }
       super.serviceStop();
     }
@@ -457,6 +595,9 @@ public String toString() {
 
   }
 
+  /**
+   * Composite service that makes the addService method public to all
+   */
   public static class ServiceManager extends CompositeService {
 
     public void addTestService(CompositeService service) {
@@ -468,4 +609,55 @@ public ServiceManager(String name) {
     }
   }
 
+  public static class AddSiblingService extends CompositeService {
+    private final CompositeService parent;
+    private final Service serviceToAdd;
+    private STATE triggerState;
+
+    public AddSiblingService(CompositeService parent,
+                             Service serviceToAdd,
+                             STATE triggerState) {
+      super("ParentStateManipulatorService");
+      this.parent = parent;
+      this.serviceToAdd = serviceToAdd;
+      this.triggerState = triggerState;
+    }
+
+    /**
+     * Add the serviceToAdd to the parent if this service
+     * is in the state requested
+     */
+    private void maybeAddSibling() {
+      if (getServiceState() == triggerState) {
+        parent.addService(serviceToAdd);
+      }
+    }
+
+    @Override
+    protected void serviceInit(Configuration conf) throws Exception {
+      maybeAddSibling();
+      super.serviceInit(conf);
+    }
+
+    @Override
+    protected void serviceStart() throws Exception {
+      maybeAddSibling();
+      super.serviceStart();
+    }
+
+    @Override
+    protected void serviceStop() throws Exception {
+      maybeAddSibling();
+      super.serviceStop();
+    }
+
+    /**
+     * Expose addService method
+     * @param parent parent service
+     * @param child child to add
+     */
+    public static void addChildToService(CompositeService parent, Service child) {
+      parent.addService(child);
+    }
+  }
 }

hadoop-common-project/hadoop-nfs/src/main/java/org/apache/hadoop/nfs/nfs3/IdUserGroup.java

@@ -50,14 +50,6 @@ public class IdUserGroup {
   private BiMap<Integer, String> gidNameMap = HashBiMap.create();
 
   private long lastUpdateTime = 0; // Last time maps were updated
-
-  static public class DuplicateNameOrIdException extends IOException {
-    private static final long serialVersionUID = 1L;
-
-    public DuplicateNameOrIdException(String msg) {
-      super(msg);
-    }
-  }
   
   public IdUserGroup() throws IOException {
     updateMaps();
@@ -80,7 +72,8 @@ private void checkAndUpdateMaps() {
     }
   }
 
-  private static final String DUPLICATE_NAME_ID_DEBUG_INFO = "NFS gateway can't start with duplicate name or id on the host system.\n"
+  private static final String DUPLICATE_NAME_ID_DEBUG_INFO =
+      "NFS gateway could have problem starting with duplicate name or id on the host system.\n"
       + "This is because HDFS (non-kerberos cluster) uses name as the only way to identify a user or group.\n"
       + "The host system with duplicated user/group name or id might work fine most of the time by itself.\n"
       + "However when NFS gateway talks to HDFS, HDFS accepts only user and group name.\n"
@@ -88,6 +81,16 @@ private void checkAndUpdateMaps() {
       + "<getent passwd | cut -d: -f1,3> and <getent group | cut -d: -f1,3> on Linux systms,\n"
       + "<dscl . -list /Users UniqueID> and <dscl . -list /Groups PrimaryGroupID> on MacOS.";
   
+  private static void reportDuplicateEntry(final String header,
+      final Integer key, final String value,
+      final Integer ekey, final String evalue) {    
+      LOG.warn("\n" + header + String.format(
+          "new entry (%d, %s), existing entry: (%d, %s).\n%s\n%s",
+          key, value, ekey, evalue,
+          "The new entry is to be ignored for the following reason.",
+          DUPLICATE_NAME_ID_DEBUG_INFO));
+  }
+      
   /**
    * Get the whole list of users and groups and save them in the maps.
    * @throws IOException 
@@ -108,22 +111,27 @@ public static void updateMapInternal(BiMap<Integer, String> map, String mapName,
         }
         LOG.debug("add to " + mapName + "map:" + nameId[0] + " id:" + nameId[1]);
         // HDFS can't differentiate duplicate names with simple authentication
-        Integer key = Integer.valueOf(nameId[1]);
-        String value = nameId[0];
+        final Integer key = Integer.valueOf(nameId[1]);
+        final String value = nameId[0];        
         if (map.containsKey(key)) {
-          LOG.error(String.format(
-              "Got duplicate id:(%d, %s), existing entry: (%d, %s).\n%s", key,
-              value, key, map.get(key), DUPLICATE_NAME_ID_DEBUG_INFO));
-          throw new DuplicateNameOrIdException("Got duplicate id.");
+          final String prevValue = map.get(key);
+          if (value.equals(prevValue)) {
+            // silently ignore equivalent entries
+            continue;
+          }
+          reportDuplicateEntry(
+              "Got multiple names associated with the same id: ",
+              key, value, key, prevValue);           
+          continue;
         }
-        if (map.containsValue(nameId[0])) {
-          LOG.error(String.format(
-              "Got duplicate name:(%d, %s), existing entry: (%d, %s) \n%s",
-              key, value, map.inverse().get(value), value,
-              DUPLICATE_NAME_ID_DEBUG_INFO));
-          throw new DuplicateNameOrIdException("Got duplicate name");
+        if (map.containsValue(value)) {
+          final Integer prevKey = map.inverse().get(value);
+          reportDuplicateEntry(
+              "Got multiple ids associated with the same name: ",
+              key, value, prevKey, value);
+          continue;
         }
-        map.put(Integer.valueOf(nameId[1]), nameId[0]);
+        map.put(key, value);
       }
       LOG.info("Updated " + mapName + " map size:" + map.size());
       

hadoop-common-project/hadoop-nfs/src/test/java/org/apache/hadoop/nfs/nfs3/TestIdUserGroup.java

@@ -17,11 +17,10 @@
  */
 package org.apache.hadoop.nfs.nfs3;
 
-import static org.junit.Assert.fail;
-
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertTrue;
 import java.io.IOException;
 
-import org.apache.hadoop.nfs.nfs3.IdUserGroup.DuplicateNameOrIdException;
 import org.junit.Test;
 
 import com.google.common.collect.BiMap;
@@ -33,24 +32,36 @@ public class TestIdUserGroup {
   public void testDuplicates() throws IOException {
     String GET_ALL_USERS_CMD = "echo \"root:x:0:0:root:/root:/bin/bash\n"
         + "hdfs:x:11501:10787:Grid Distributed File System:/home/hdfs:/bin/bash\n"
-        + "hdfs:x:11502:10788:Grid Distributed File System:/home/hdfs:/bin/bash\""
+        + "hdfs:x:11502:10788:Grid Distributed File System:/home/hdfs:/bin/bash\n"
+        + "hdfs1:x:11501:10787:Grid Distributed File System:/home/hdfs:/bin/bash\n"
+        + "hdfs2:x:11502:10787:Grid Distributed File System:/home/hdfs:/bin/bash\n"
+        + "bin:x:2:2:bin:/bin:/bin/sh\n"
+        + "bin:x:1:1:bin:/bin:/sbin/nologin\n"
+        + "daemon:x:1:1:daemon:/usr/sbin:/bin/sh\n"
+        + "daemon:x:2:2:daemon:/sbin:/sbin/nologin\""
         + " | cut -d: -f1,3";
     String GET_ALL_GROUPS_CMD = "echo \"hdfs:*:11501:hrt_hdfs\n"
-        + "mapred:x:497\n" + "mapred2:x:497\"" + " | cut -d: -f1,3";
+        + "mapred:x:497\n"
+        + "mapred2:x:497\n"
+        + "mapred:x:498\n" 
+        + "mapred3:x:498\"" 
+        + " | cut -d: -f1,3";
     // Maps for id to name map
     BiMap<Integer, String> uMap = HashBiMap.create();
     BiMap<Integer, String> gMap = HashBiMap.create();
 
-    try {
-      IdUserGroup.updateMapInternal(uMap, "user", GET_ALL_USERS_CMD, ":");
-      fail("didn't detect the duplicate name");
-    } catch (DuplicateNameOrIdException e) {
-    }
+    IdUserGroup.updateMapInternal(uMap, "user", GET_ALL_USERS_CMD, ":");
+    assertTrue(uMap.size() == 5);
+    assertEquals(uMap.get(0), "root");
+    assertEquals(uMap.get(11501), "hdfs");
+    assertEquals(uMap.get(11502), "hdfs2");
+    assertEquals(uMap.get(2), "bin");
+    assertEquals(uMap.get(1), "daemon");
     
-    try {
-      IdUserGroup.updateMapInternal(gMap, "group", GET_ALL_GROUPS_CMD, ":");
-      fail("didn't detect the duplicate id");
-    } catch (DuplicateNameOrIdException e) {
-    }
+    IdUserGroup.updateMapInternal(gMap, "group", GET_ALL_GROUPS_CMD, ":");
+    assertTrue(gMap.size() == 3);
+    assertEquals(gMap.get(11501), "hdfs");
+    assertEquals(gMap.get(497), "mapred");
+    assertEquals(gMap.get(498), "mapred3");    
   }
 }

hadoop-common-project/hadoop-nfs/src/test/java/org/apache/hadoop/oncrpc/TestFrameDecoder.java

@@ -23,6 +23,7 @@
 import static org.junit.Assert.assertTrue;
 
 import java.nio.ByteBuffer;
+import java.util.Random;
 
 import org.apache.hadoop.oncrpc.RpcUtil.RpcFrameDecoder;
 import org.apache.hadoop.oncrpc.security.CredentialsNone;
@@ -31,17 +32,17 @@
 import org.jboss.netty.buffer.ChannelBuffer;
 import org.jboss.netty.buffer.ChannelBuffers;
 import org.jboss.netty.channel.Channel;
+import org.jboss.netty.channel.ChannelException;
 import org.jboss.netty.channel.ChannelHandlerContext;
 import org.junit.Test;
 import org.mockito.Mockito;
 
 public class TestFrameDecoder {
 
-  private static int port = 12345; // some random server port
   private static int resultSize;
 
-  static void testRequest(XDR request) {
-    SimpleTcpClient tcpClient = new SimpleTcpClient("localhost", port, request,
+  static void testRequest(XDR request, int serverPort) {
+    SimpleTcpClient tcpClient = new SimpleTcpClient("localhost", serverPort, request,
         true);
     tcpClient.run();
   }
@@ -148,10 +149,25 @@ public void testMultipleFrames() {
   @Test
   public void testFrames() {
 
-    RpcProgram program = new TestFrameDecoder.TestRpcProgram("TestRpcProgram",
-        "localhost", port, 100000, 1, 2);
-    SimpleTcpServer tcpServer = new SimpleTcpServer(port, program, 1);
-    tcpServer.run();
+    Random rand = new Random();
+    int serverPort = 30000 + rand.nextInt(10000);
+    int retries = 10;    // A few retries in case initial choice is in use.
+
+    while (true) {
+      try {
+        RpcProgram program = new TestFrameDecoder.TestRpcProgram("TestRpcProgram",
+            "localhost", serverPort, 100000, 1, 2);
+        SimpleTcpServer tcpServer = new SimpleTcpServer(serverPort, program, 1);
+        tcpServer.run();
+        break;          // Successfully bound a port, break out.
+      } catch (ChannelException ce) {
+        if (retries-- > 0) {
+          serverPort += rand.nextInt(20); // Port in use? Try another.
+        } else {
+          throw ce;     // Out of retries.
+        }
+      }
+    }
 
     XDR xdrOut = createGetportMount();
     int headerSize = xdrOut.size();
@@ -161,7 +177,7 @@ public void testFrames() {
     int requestSize = xdrOut.size() - headerSize;
 
     // Send the request to the server
-    testRequest(xdrOut);
+    testRequest(xdrOut, serverPort);
 
     // Verify the server got the request with right size
     assertEquals(requestSize, resultSize);

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

@@ -303,6 +303,14 @@ Release 2.4.0 - UNRELEASED
     HDFS-5804. HDFS NFS Gateway fails to mount and proxy when using Kerberos.
     (Abin Shahab via jing9)
 
+    HDFS-5859. DataNode#checkBlockToken should check block tokens even if
+    security is not enabled. (cmccabe)
+
+    HDFS-5746.  Add ShortCircuitSharedMemorySegment (cmccabe)
+
+    HDFS-4911.  Reduce PeerCache timeout to be commensurate with
+    dfs.datanode.socket.reuse.keepalive (cmccabe)
+
   OPTIMIZATIONS
 
     HDFS-5790. LeaseManager.findPath is very slow when many leases need recovery
@@ -319,6 +327,32 @@ Release 2.4.0 - UNRELEASED
     HDFS-5856. DataNode.checkDiskError might throw NPE.
     (Josh Elser via suresh)
 
+    HDFS-5828. BlockPlacementPolicyWithNodeGroup can place multiple replicas on
+    the same node group when dfs.namenode.avoid.write.stale.datanode is true. 
+    (Buddy via junping_du)
+
+    HDFS-5767. NFS implementation assumes userName userId mapping to be unique,
+    which is not true sometimes (Yongjun Zhang via brandonli)
+
+    HDFS-5791. TestHttpsFileSystem should use a random port to avoid binding
+    error during testing (Haohui Mai via brandonli)
+
+    HDFS-5709. Improve NameNode upgrade with existing reserved paths and path
+    components. (Andrew Wang via atm)
+
+    HDFS-5881. Fix skip() of the short-circuit local reader(legacy). (kihwal)
+
+    HDFS-5895. HDFS cacheadmin -listPools has exit_code of 1 when the command
+    returns 0 result. (Tassapol Athiapinya via cnauroth)
+
+    HDFS-5807. TestBalancerWithNodeGroup.testBalancerWithNodeGroup fails
+    intermittently. (Chen He via kihwal)
+
+    HDFS-5882. TestAuditLogs is flaky (jxiang via cmccabe)
+
+    HDFS-5900. Cannot set cache pool limit of "unlimited" via CacheAdmin.
+    (wang)
+
 Release 2.3.0 - UNRELEASED
 
   INCOMPATIBLE CHANGES
@@ -848,6 +882,15 @@ Release 2.3.0 - UNRELEASED
     HDFS-5842. Cannot create hftp filesystem when using a proxy user ugi and a doAs 
     on a secure cluster. (jing9)
 
+    HDFS-5399. Revisit SafeModeException and corresponding retry policies.
+    (Jing Zhao via todd)
+
+    HDFS-5876. SecureDataNodeStarter does not pick up configuration in 
+    hdfs-site.xml. (Haohui Mai via jing9)
+
+    HDFS-5873. dfs.http.policy should have higher precedence over dfs.https.enable.
+    (Haohui Mai via jing9)
+
   BREAKDOWN OF HDFS-2832 SUBTASKS AND RELATED JIRAS
 
     HDFS-4985. Add storage type to the protocol and expose it in block report

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

@@ -629,7 +629,7 @@ public synchronized long skip(long n) throws IOException {
         skipBuf = new byte[bytesPerChecksum];
       }
       int ret = read(skipBuf, 0, (int)(n - remaining));
-      return ret;
+      return (remaining + ret);
     }
   
     // optimize for big gap: discard the current buffer, skip to
@@ -660,9 +660,9 @@ public synchronized long skip(long n) throws IOException {
     int ret = read(skipBuf, 0, myOffsetFromChunkBoundary);
 
     if (ret == -1) {  // EOS
-      return toskip;
+      return (toskip + remaining);
     } else {
-      return (toskip + ret);
+      return (toskip + remaining + ret);
     }
   }
 

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

@@ -36,6 +36,8 @@
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_MAX_DEFAULT;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_MAX_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_RETRY_MAX_ATTEMPTS_KEY;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_MAX_BLOCK_ACQUIRE_FAILURES_DEFAULT;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_MAX_BLOCK_ACQUIRE_FAILURES_KEY;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_READ_PREFETCH_SIZE_KEY;
@@ -261,6 +263,7 @@ synchronized void unref(ClientMmapManager mmapManager) {
   public static class Conf {
     final int hdfsTimeout;    // timeout value for a DFS operation.
     final int maxFailoverAttempts;
+    final int maxRetryAttempts;
     final int failoverSleepBaseMillis;
     final int failoverSleepMaxMillis;
     final int maxBlockAcquireFailures;
@@ -306,6 +309,9 @@ public Conf(Configuration conf) {
       maxFailoverAttempts = conf.getInt(
           DFS_CLIENT_FAILOVER_MAX_ATTEMPTS_KEY,
           DFS_CLIENT_FAILOVER_MAX_ATTEMPTS_DEFAULT);
+      maxRetryAttempts = conf.getInt(
+          DFS_CLIENT_RETRY_MAX_ATTEMPTS_KEY,
+          DFS_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT);
       failoverSleepBaseMillis = conf.getInt(
           DFS_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY,
           DFS_CLIENT_FAILOVER_SLEEPTIME_BASE_DEFAULT);

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

@@ -84,9 +84,11 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final int     DFS_CLIENT_FAILOVER_CONNECTION_RETRIES_DEFAULT = 0;
   public static final String  DFS_CLIENT_FAILOVER_CONNECTION_RETRIES_ON_SOCKET_TIMEOUTS_KEY = "dfs.client.failover.connection.retries.on.timeouts";
   public static final int     DFS_CLIENT_FAILOVER_CONNECTION_RETRIES_ON_SOCKET_TIMEOUTS_DEFAULT = 0;
+  public static final String  DFS_CLIENT_RETRY_MAX_ATTEMPTS_KEY = "dfs.client.retry.max.attempts";
+  public static final int     DFS_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT = 10;
   
   public static final String  DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_KEY = "dfs.client.socketcache.expiryMsec";
-  public static final long    DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_DEFAULT = 2 * 60 * 1000;
+  public static final long    DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_DEFAULT = 3000;
   public static final String  DFS_CLIENT_WRITE_EXCLUDE_NODES_CACHE_EXPIRY_INTERVAL = "dfs.client.write.exclude.nodes.cache.expiry.interval.millis";
   public static final long    DFS_CLIENT_WRITE_EXCLUDE_NODES_CACHE_EXPIRY_INTERVAL_DEFAULT = 10 * 60 * 1000; // 10 minutes, in ms
   public static final String  DFS_NAMENODE_BACKUP_ADDRESS_KEY = "dfs.namenode.backup.address";
@@ -217,7 +219,7 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final String  DFS_DATANODE_SYNCONCLOSE_KEY = "dfs.datanode.synconclose";
   public static final boolean DFS_DATANODE_SYNCONCLOSE_DEFAULT = false;
   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;
+  public static final int     DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT = 4000;
 
   public static final String DFS_NAMENODE_DATANODE_REGISTRATION_IP_HOSTNAME_CHECK_KEY = "dfs.namenode.datanode.registration.ip-hostname-check";
   public static final boolean DFS_NAMENODE_DATANODE_REGISTRATION_IP_HOSTNAME_CHECK_DEFAULT = true;
@@ -578,6 +580,8 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
   public static final String  DFS_HTTP_CLIENT_RETRY_POLICY_SPEC_DEFAULT = "10000,6,60000,10"; //t1,n1,t2,n2,...
   public static final String  DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_KEY = "dfs.http.client.failover.max.attempts";
   public static final int     DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_DEFAULT = 15;
+  public static final String  DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_KEY = "dfs.http.client.retry.max.attempts";
+  public static final int     DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT = 10;
   public static final String  DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY = "dfs.http.client.failover.sleep.base.millis";
   public static final int     DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_DEFAULT = 500;
   public static final String  DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_MAX_KEY = "dfs.http.client.failover.sleep.max.millis";

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

@@ -1345,6 +1345,14 @@ public synchronized void seek(long targetPos) throws IOException {
           pos += blockReader.skip(diff);
           if (pos == targetPos) {
             done = true;
+          } else {
+            // The range was already checked. If the block reader returns
+            // something unexpected instead of throwing an exception, it is
+            // most likely a bug. 
+            String errMsg = "BlockReader failed to seek to " + 
+                targetPos + ". Instead, it seeked to " + pos + ".";
+            DFSClient.LOG.warn(errMsg);
+            throw new IOException(errMsg);
           }
         } catch (IOException e) {//make following read to retry
           if(DFSClient.LOG.isDebugEnabled()) {

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

@@ -261,6 +261,47 @@ public static boolean isValidName(String src) {
     return true;
   }
 
+  /**
+   * Checks if a string is a valid path component. For instance, components
+   * cannot contain a ":" or "/", and cannot be equal to a reserved component
+   * like ".snapshot".
+   * <p>
+   * The primary use of this method is for validating paths when loading the
+   * FSImage. During normal NN operation, paths are sometimes allowed to
+   * contain reserved components.
+   * 
+   * @return If component is valid
+   */
+  public static boolean isValidNameForComponent(String component) {
+    if (component.equals(".") ||
+        component.equals("..") ||
+        component.indexOf(":") >= 0 ||
+        component.indexOf("/") >= 0) {
+      return false;
+    }
+    return !isReservedPathComponent(component);
+  }
+
+
+  /**
+   * Returns if the component is reserved.
+   * 
+   * <p>
+   * Note that some components are only reserved under certain directories, e.g.
+   * "/.reserved" is reserved, while "/hadoop/.reserved" is not.
+   * 
+   * @param component
+   * @return if the component is reserved
+   */
+  public static boolean isReservedPathComponent(String component) {
+    for (String reserved : HdfsConstants.RESERVED_PATH_COMPONENTS) {
+      if (component.equals(reserved)) {
+        return true;
+      }
+    }
+    return false;
+  }
+
   /**
    * Converts a byte array to a string using UTF8 encoding.
    */
@@ -312,7 +353,25 @@ public static String byteArray2PathString(byte[][] pathComponents) {
     }
     return result.toString();
   }
-  
+
+  /**
+   * Converts a list of path components into a path using Path.SEPARATOR.
+   * 
+   * @param components Path components
+   * @return Combined path as a UTF-8 string
+   */
+  public static String strings2PathString(String[] components) {
+    if (components.length == 0) {
+      return "";
+    }
+    if (components.length == 1) {
+      if (components[0] == null || components[0].isEmpty()) {
+        return Path.SEPARATOR;
+      }
+    }
+    return Joiner.on(Path.SEPARATOR).join(components);
+  }
+
   /**
    * Given a list of path components returns a byte array
    */
@@ -1508,31 +1567,34 @@ public static String getSpnegoKeytabKey(Configuration conf, String defaultKey) {
    * configuration settings.
    */
   public static HttpConfig.Policy getHttpPolicy(Configuration conf) {
-    String httpPolicy = conf.get(DFSConfigKeys.DFS_HTTP_POLICY_KEY,
-        DFSConfigKeys.DFS_HTTP_POLICY_DEFAULT);
-
-    HttpConfig.Policy policy = HttpConfig.Policy.fromString(httpPolicy);
-
-    if (policy == HttpConfig.Policy.HTTP_ONLY) {
-      boolean httpsEnabled = conf.getBoolean(
-          DFSConfigKeys.DFS_HTTPS_ENABLE_KEY,
+    String policyStr = conf.get(DFSConfigKeys.DFS_HTTP_POLICY_KEY);
+    if (policyStr == null) {
+      boolean https = conf.getBoolean(DFSConfigKeys.DFS_HTTPS_ENABLE_KEY,
           DFSConfigKeys.DFS_HTTPS_ENABLE_DEFAULT);
 
-      boolean hadoopSslEnabled = conf.getBoolean(
+      boolean hadoopSsl = conf.getBoolean(
           CommonConfigurationKeys.HADOOP_SSL_ENABLED_KEY,
           CommonConfigurationKeys.HADOOP_SSL_ENABLED_DEFAULT);
 
-      if (hadoopSslEnabled) {
+      if (hadoopSsl) {
         LOG.warn(CommonConfigurationKeys.HADOOP_SSL_ENABLED_KEY
-            + " is deprecated. Please use "
-            + DFSConfigKeys.DFS_HTTPS_ENABLE_KEY + ".");
-        policy = HttpConfig.Policy.HTTPS_ONLY;
-      } else if (httpsEnabled) {
-        LOG.warn(DFSConfigKeys.DFS_HTTPS_ENABLE_KEY
-            + " is deprecated. Please use "
-            + DFSConfigKeys.DFS_HTTPS_ENABLE_KEY + ".");
-        policy = HttpConfig.Policy.HTTP_AND_HTTPS;
+            + " is deprecated. Please use " + DFSConfigKeys.DFS_HTTP_POLICY_KEY
+            + ".");
       }
+      if (https) {
+        LOG.warn(DFSConfigKeys.DFS_HTTPS_ENABLE_KEY
+            + " is deprecated. Please use " + DFSConfigKeys.DFS_HTTP_POLICY_KEY
+            + ".");
+      }
+
+      return (hadoopSsl || https) ? HttpConfig.Policy.HTTP_AND_HTTPS
+          : HttpConfig.Policy.HTTP_ONLY;
+    }
+
+    HttpConfig.Policy policy = HttpConfig.Policy.fromString(policyStr);
+    if (policy == null) {
+      throw new HadoopIllegalArgumentException("Unregonized value '"
+          + policyStr + "' for " + DFSConfigKeys.DFS_HTTP_POLICY_KEY);
     }
 
     conf.set(DFSConfigKeys.DFS_HTTP_POLICY_KEY, policy.name());

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

@@ -24,6 +24,8 @@
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_MAX_DEFAULT;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_FAILOVER_SLEEPTIME_MAX_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_RETRY_MAX_ATTEMPTS_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT;
 
 import java.io.IOException;
 import java.lang.reflect.Constructor;
@@ -144,9 +146,10 @@ public static <T> ProxyAndInfo<T> createProxy(Configuration conf,
           .createFailoverProxyProvider(conf, failoverProxyProviderClass, xface,
               nameNodeUri);
       Conf config = new Conf(conf);
-      T proxy = (T) RetryProxy.create(xface, failoverProxyProvider, RetryPolicies
-          .failoverOnNetworkException(RetryPolicies.TRY_ONCE_THEN_FAIL,
-              config.maxFailoverAttempts, config.failoverSleepBaseMillis,
+      T proxy = (T) RetryProxy.create(xface, failoverProxyProvider,
+          RetryPolicies.failoverOnNetworkException(
+              RetryPolicies.TRY_ONCE_THEN_FAIL, config.maxFailoverAttempts,
+              config.maxRetryAttempts, config.failoverSleepBaseMillis,
               config.failoverSleepMaxMillis));
       
       Text dtService = HAUtil.buildTokenServiceForLogicalUri(nameNodeUri);
@@ -192,11 +195,14 @@ public static <T> ProxyAndInfo<T> createProxyWithLossyRetryHandler(
       int maxFailoverAttempts = config.getInt(
           DFS_CLIENT_FAILOVER_MAX_ATTEMPTS_KEY,
           DFS_CLIENT_FAILOVER_MAX_ATTEMPTS_DEFAULT);
+      int maxRetryAttempts = config.getInt(
+          DFS_CLIENT_RETRY_MAX_ATTEMPTS_KEY,
+          DFS_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT);
       InvocationHandler dummyHandler = new LossyRetryInvocationHandler<T>(
               numResponseToDrop, failoverProxyProvider,
               RetryPolicies.failoverOnNetworkException(
-                  RetryPolicies.TRY_ONCE_THEN_FAIL, 
-                  Math.max(numResponseToDrop + 1, maxFailoverAttempts), delay, 
+                  RetryPolicies.TRY_ONCE_THEN_FAIL, maxFailoverAttempts, 
+                  Math.max(numResponseToDrop + 1, maxRetryAttempts), delay, 
                   maxCap));
       
       T proxy = (T) Proxy.newProxyInstance(

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

@@ -18,6 +18,7 @@
 
 package org.apache.hadoop.hdfs;
 
+import java.io.IOException;
 import java.util.Iterator;
 import java.util.List;
 import java.util.Map.Entry;
@@ -25,6 +26,7 @@
 import com.google.common.annotations.VisibleForTesting;
 import com.google.common.base.Preconditions;
 import com.google.common.collect.LinkedListMultimap;
+
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.hdfs.protocol.DatanodeID;
@@ -118,6 +120,11 @@ public static synchronized PeerCache getInstance(int c, long e) {
     return instance;
   }
 
+  @VisibleForTesting
+  public static synchronized void setInstance(int c, long e) {
+    instance = new PeerCache(c, e);
+  }
+
   private boolean isDaemonStarted() {
     return (daemon == null)? false: true;
   }
@@ -171,8 +178,17 @@ public synchronized Peer get(DatanodeID dnId, boolean isDomain) {
     while (iter.hasNext()) {
       Value candidate = iter.next();
       iter.remove();
-      if (!candidate.getPeer().isClosed()) {
-        return candidate.getPeer();
+      long ageMs = Time.monotonicNow() - candidate.getTime();
+      Peer peer = candidate.getPeer();
+      if (ageMs >= expiryPeriod) {
+        try {
+          peer.close();
+        } catch (IOException e) {
+          LOG.warn("got IOException closing stale peer " + peer +
+                ", which is " + ageMs + " ms old");
+        }
+      } else if (!peer.isClosed()) {
+        return peer;
       }
     }
     return null;

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/client/ShortCircuitSharedMemorySegment.java

@@ -0,0 +1,302 @@
+/**
+ * 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.client;
+
+import java.io.Closeable;
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.lang.reflect.Field;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.io.IOUtils;
+import org.apache.hadoop.io.nativeio.NativeIO;
+import org.apache.hadoop.io.nativeio.NativeIO.POSIX;
+import org.apache.hadoop.util.CloseableReferenceCount;
+import org.apache.hadoop.util.Shell;
+
+import com.google.common.base.Preconditions;
+import com.google.common.primitives.Ints;
+
+import sun.misc.Unsafe;
+
+public class ShortCircuitSharedMemorySegment implements Closeable {
+  private static final Log LOG =
+    LogFactory.getLog(ShortCircuitSharedMemorySegment.class);
+
+  private static final int BYTES_PER_SLOT = 64;
+
+  private static final Unsafe unsafe;
+
+  static {
+    Unsafe theUnsafe = null;
+    try {
+      Field f = Unsafe.class.getDeclaredField("theUnsafe");
+      f.setAccessible(true);
+      theUnsafe = (Unsafe)f.get(null);
+    } catch (Throwable e) {
+      LOG.error("failed to load misc.Unsafe", e);
+    }
+    unsafe = theUnsafe;
+  }
+
+  /**
+   * A slot containing information about a replica.
+   *
+   * The format is:
+   * word 0
+   *   bit 0:32   Slot flags (see below).
+   *   bit 33:63  Anchor count.
+   * word 1:7
+   *   Reserved for future use, such as statistics.
+   *   Padding is also useful for avoiding false sharing.
+   *
+   * Little-endian versus big-endian is not relevant here since both the client
+   * and the server reside on the same computer and use the same orientation.
+   */
+  public class Slot implements Closeable {
+    /**
+     * Flag indicating that the slot is in use.
+     */
+    private static final long SLOT_IN_USE_FLAG =    1L<<63;
+
+    /**
+     * Flag indicating that the slot can be anchored.
+     */
+    private static final long ANCHORABLE_FLAG =     1L<<62;
+
+    private long slotAddress;
+
+    Slot(long slotAddress) {
+      this.slotAddress = slotAddress;
+    }
+
+    /**
+     * Make a given slot anchorable.
+     */
+    public void makeAnchorable() {
+      Preconditions.checkState(slotAddress != 0,
+          "Called makeAnchorable on a slot that was closed.");
+      long prev;
+      do {
+        prev = unsafe.getLongVolatile(null, this.slotAddress);
+        if ((prev & ANCHORABLE_FLAG) != 0) {
+          return;
+        }
+      } while (!unsafe.compareAndSwapLong(null, this.slotAddress,
+                  prev, prev | ANCHORABLE_FLAG));
+    }
+
+    /**
+     * Make a given slot unanchorable.
+     */
+    public void makeUnanchorable() {
+      Preconditions.checkState(slotAddress != 0,
+          "Called makeUnanchorable on a slot that was closed.");
+      long prev;
+      do {
+        prev = unsafe.getLongVolatile(null, this.slotAddress);
+        if ((prev & ANCHORABLE_FLAG) == 0) {
+          return;
+        }
+      } while (!unsafe.compareAndSwapLong(null, this.slotAddress,
+                  prev, prev & (~ANCHORABLE_FLAG)));
+    }
+
+    /**
+     * Try to add an anchor for a given slot.
+     *
+     * When a slot is anchored, we know that the block it refers to is resident
+     * in memory.
+     *
+     * @return          True if the slot is anchored.
+     */
+    public boolean addAnchor() {
+      long prev;
+      do {
+        prev = unsafe.getLongVolatile(null, this.slotAddress);
+        if ((prev & 0x7fffffff) == 0x7fffffff) {
+          // Too many other threads have anchored the slot (2 billion?)
+          return false;
+        }
+        if ((prev & ANCHORABLE_FLAG) == 0) {
+          // Slot can't be anchored right now.
+          return false;
+        }
+      } while (!unsafe.compareAndSwapLong(null, this.slotAddress,
+                  prev, prev + 1));
+      return true;
+    }
+
+    /**
+     * Remove an anchor for a given slot.
+     */
+    public void removeAnchor() {
+      long prev;
+      do {
+        prev = unsafe.getLongVolatile(null, this.slotAddress);
+        Preconditions.checkState((prev & 0x7fffffff) != 0,
+            "Tried to remove anchor for slot " + slotAddress +", which was " +
+            "not anchored.");
+      } while (!unsafe.compareAndSwapLong(null, this.slotAddress,
+                  prev, prev - 1));
+    }
+
+    /**
+     * @return      The index of this slot.
+     */
+    public int getIndex() {
+      Preconditions.checkState(slotAddress != 0);
+      return Ints.checkedCast(
+          (slotAddress - baseAddress) / BYTES_PER_SLOT);
+    }
+
+    @Override
+    public void close() throws IOException {
+      if (slotAddress == 0) return;
+      long prev;
+      do {
+        prev = unsafe.getLongVolatile(null, this.slotAddress);
+        Preconditions.checkState((prev & SLOT_IN_USE_FLAG) != 0,
+            "tried to close slot that wasn't open");
+      } while (!unsafe.compareAndSwapLong(null, this.slotAddress,
+                  prev, 0));
+      slotAddress = 0;
+      if (ShortCircuitSharedMemorySegment.this.refCount.unreference()) {
+        ShortCircuitSharedMemorySegment.this.free();
+      }
+    }
+  }
+
+  /**
+   * The stream that we're going to use to create this shared memory segment.
+   *
+   * Although this is a FileInputStream, we are going to assume that the
+   * underlying file descriptor is writable as well as readable.
+   * It would be more appropriate to use a RandomAccessFile here, but that class
+   * does not have any public accessor which returns a FileDescriptor, unlike
+   * FileInputStream.
+   */
+  private final FileInputStream stream;
+
+  /**
+   * Length of the shared memory segment.
+   */
+  private final int length;
+
+  /**
+   * The base address of the memory-mapped file.
+   */
+  private final long baseAddress;
+
+  /**
+   * Reference count and 'closed' status.
+   */
+  private final CloseableReferenceCount refCount = new CloseableReferenceCount();
+
+  public ShortCircuitSharedMemorySegment(FileInputStream stream)
+        throws IOException {
+    if (!NativeIO.isAvailable()) {
+      throw new UnsupportedOperationException("NativeIO is not available.");
+    }
+    if (Shell.WINDOWS) {
+      throw new UnsupportedOperationException(
+          "ShortCircuitSharedMemorySegment is not yet implemented " +
+          "for Windows.");
+    }
+    if (unsafe == null) {
+      throw new UnsupportedOperationException(
+          "can't use ShortCircuitSharedMemorySegment because we failed to " +
+          "load misc.Unsafe.");
+    }
+    this.refCount.reference();
+    this.stream = stream;
+    this.length = getEffectiveLength(stream);
+    this.baseAddress = POSIX.mmap(this.stream.getFD(), 
+      POSIX.MMAP_PROT_READ | POSIX.MMAP_PROT_WRITE, true, this.length);
+  }
+
+  /**
+   * Calculate the effective usable size of the shared memory segment.
+   * We round down to a multiple of the slot size and do some validation.
+   *
+   * @param stream The stream we're using.
+   * @return       The effective usable size of the shared memory segment.
+   */
+  private static int getEffectiveLength(FileInputStream stream)
+      throws IOException {
+    int intSize = Ints.checkedCast(stream.getChannel().size());
+    int slots = intSize / BYTES_PER_SLOT;
+    Preconditions.checkState(slots > 0, "size of shared memory segment was " +
+        intSize + ", but that is not enough to hold even one slot.");
+    return slots * BYTES_PER_SLOT;
+  }
+
+  private boolean allocateSlot(long address) {
+    long prev;
+    do {
+      prev = unsafe.getLongVolatile(null, address);
+      if ((prev & Slot.SLOT_IN_USE_FLAG) != 0) {
+        return false;
+      }
+    } while (!unsafe.compareAndSwapLong(null, address,
+                prev, prev | Slot.SLOT_IN_USE_FLAG));
+    return true;
+  }
+
+  /**
+   * Allocate a new Slot in this shared memory segment.
+   *
+   * @return        A newly allocated Slot, or null if there were no available
+   *                slots.
+   */
+  public Slot allocateNextSlot() throws IOException {
+    ShortCircuitSharedMemorySegment.this.refCount.reference();
+    Slot slot = null;
+    try {
+      final int numSlots = length / BYTES_PER_SLOT;
+      for (int i = 0; i < numSlots; i++) {
+        long address = this.baseAddress + (i * BYTES_PER_SLOT);
+        if (allocateSlot(address)) {
+          slot = new Slot(address);
+          break;
+        }
+      }
+    } finally {
+      if (slot == null) {
+        if (refCount.unreference()) {
+          free();
+        }
+      }
+    }
+    return slot;
+  }
+
+  @Override
+  public void close() throws IOException {
+    refCount.setClosed();
+    if (refCount.unreference()) {
+      free();
+    }
+  }
+
+  void free() throws IOException {
+    IOUtils.cleanup(LOG, stream);
+    POSIX.munmap(baseAddress, length);
+  }
+}

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

@@ -22,6 +22,7 @@
 import org.apache.hadoop.hdfs.DFSConfigKeys;
 import org.apache.hadoop.hdfs.DFSUtil;
 import org.apache.hadoop.hdfs.HdfsConfiguration;
+import org.apache.hadoop.hdfs.server.namenode.FSDirectory;
 
 /************************************
  * Some handy constants
@@ -108,7 +109,17 @@ public static enum DatanodeReportType {
    */
   public static final int LAYOUT_VERSION = LayoutVersion
       .getCurrentLayoutVersion();
-  
+
+  /**
+   * Path components that are reserved in HDFS.
+   * <p>
+   * .reserved is only reserved under root ("/").
+   */
+  public static final String[] RESERVED_PATH_COMPONENTS = new String[] {
+    HdfsConstants.DOT_SNAPSHOT_DIR,
+    FSDirectory.DOT_RESERVED_STRING
+  };
+
   /**
    * A special path component contained in the path for a snapshot file/dir
    */

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

@@ -317,7 +317,7 @@ private Node chooseTarget(int numOfReplicas,
         // We need to additionally exclude the nodes that were added to the 
         // result list in the successful calls to choose*() above.
         for (DatanodeStorageInfo resultStorage : results) {
-          oldExcludedNodes.add(resultStorage.getDatanodeDescriptor());
+          addToExcludedNodes(resultStorage.getDatanodeDescriptor(), oldExcludedNodes);
         }
         // Set numOfReplicas, since it can get out of sync with the result list
         // if the NotEnoughReplicasException was thrown in chooseRandom().

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

@@ -59,7 +59,8 @@ static public enum StartupOption{
     INITIALIZESHAREDEDITS("-initializeSharedEdits"),
     RECOVER  ("-recover"),
     FORCE("-force"),
-    NONINTERACTIVE("-nonInteractive");
+    NONINTERACTIVE("-nonInteractive"),
+    RENAMERESERVED("-renameReserved");
     
     private final String name;
     

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java

@@ -362,13 +362,13 @@ private void startInfoServer(Configuration conf) throws IOException {
         .setConf(conf).setACL(new AccessControlList(conf.get(DFS_ADMIN, " ")));
 
     HttpConfig.Policy policy = DFSUtil.getHttpPolicy(conf);
-    InetSocketAddress infoSocAddr = DataNode.getInfoAddr(conf);
-    String infoHost = infoSocAddr.getHostName();
 
     if (policy.isHttpEnabled()) {
       if (secureResources == null) {
+        InetSocketAddress infoSocAddr = DataNode.getInfoAddr(conf);
         int port = infoSocAddr.getPort();
-        builder.addEndpoint(URI.create("http://" + infoHost + ":" + port));
+        builder.addEndpoint(URI.create("http://"
+            + NetUtils.getHostPortString(infoSocAddr)));
         if (port == 0) {
           builder.setFindPort(true);
         }
@@ -381,7 +381,7 @@ private void startInfoServer(Configuration conf) throws IOException {
 
     if (policy.isHttpsEnabled()) {
       InetSocketAddress secInfoSocAddr = NetUtils.createSocketAddr(conf.get(
-          DFS_DATANODE_HTTPS_ADDRESS_KEY, infoHost + ":" + 0));
+          DFS_DATANODE_HTTPS_ADDRESS_KEY, DFS_DATANODE_HTTPS_ADDRESS_DEFAULT));
 
       Configuration sslConf = DFSUtil.loadSslConfiguration(conf);
       DFSUtil.loadSslConfToHttpServerBuilder(builder, sslConf);
@@ -390,7 +390,8 @@ private void startInfoServer(Configuration conf) throws IOException {
       if (port == 0) {
         builder.setFindPort(true);
       }
-      builder.addEndpoint(URI.create("https://" + infoHost + ":" + port));
+      builder.addEndpoint(URI.create("https://"
+          + NetUtils.getHostPortString(secInfoSocAddr)));
     }
 
     this.infoServer = builder.build();
@@ -1194,7 +1195,7 @@ public HdfsBlocksMetadata getHdfsBlocksMetadata(List<ExtendedBlock> blocks,
   
   private void checkBlockToken(ExtendedBlock block, Token<BlockTokenIdentifier> token,
       AccessMode accessMode) throws IOException {
-    if (isBlockTokenEnabled && UserGroupInformation.isSecurityEnabled()) {
+    if (isBlockTokenEnabled) {
       BlockTokenIdentifier id = new BlockTokenIdentifier();
       ByteArrayInputStream buf = new ByteArrayInputStream(token.getIdentifier());
       DataInputStream in = new DataInputStream(buf);

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/SecureDataNodeStarter.java

@@ -25,6 +25,7 @@
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.hdfs.DFSConfigKeys;
 import org.apache.hadoop.hdfs.DFSUtil;
+import org.apache.hadoop.hdfs.HdfsConfiguration;
 import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
 import org.apache.hadoop.http.HttpConfig;
 import org.apache.hadoop.http.HttpServer2;
@@ -62,7 +63,9 @@ public SecureResources(ServerSocket streamingSocket,
   @Override
   public void init(DaemonContext context) throws Exception {
     System.err.println("Initializing secure datanode resources");
-    Configuration conf = new Configuration();
+    // Create a new HdfsConfiguration object to ensure that the configuration in
+    // hdfs-site.xml is picked up.
+    Configuration conf = new HdfsConfiguration();
     
     // Stash command-line arguments for regular datanode
     args = context.getArguments();

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

@@ -17,6 +17,7 @@
  */
 package org.apache.hadoop.hdfs.server.namenode;
 
+import static org.apache.hadoop.hdfs.server.namenode.FSImageFormat.renameReservedPathsOnUpgrade;
 import static org.apache.hadoop.util.Time.now;
 
 import java.io.FilterInputStream;
@@ -296,8 +297,10 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
       if (addCloseOp.aclEntries != null) {
         fsNamesys.getAclConfigFlag().checkForEditLog();
       }
+      final String path =
+          renameReservedPathsOnUpgrade(addCloseOp.path, logVersion);
       if (FSNamesystem.LOG.isDebugEnabled()) {
-        FSNamesystem.LOG.debug(op.opCode + ": " + addCloseOp.path +
+        FSNamesystem.LOG.debug(op.opCode + ": " + path +
             " numblocks : " + addCloseOp.blocks.length +
             " clientHolder " + addCloseOp.clientName +
             " clientMachine " + addCloseOp.clientMachine);
@@ -308,9 +311,9 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
       // 3. OP_ADD to open file for append
 
       // See if the file already exists (persistBlocks call)
-      final INodesInPath iip = fsDir.getLastINodeInPath(addCloseOp.path);
+      final INodesInPath iip = fsDir.getLastINodeInPath(path);
       final INodeFile oldFile = INodeFile.valueOf(
-          iip.getINode(0), addCloseOp.path, true);
+          iip.getINode(0), path, true);
       INodeFile newFile = oldFile;
       if (oldFile == null) { // this is OP_ADD on a new file (case 1)
         // versions > 0 support per file replication
@@ -323,11 +326,11 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
         inodeId = getAndUpdateLastInodeId(addCloseOp.inodeId, logVersion,
             lastInodeId);
         newFile = fsDir.unprotectedAddFile(inodeId,
-            addCloseOp.path, addCloseOp.permissions, addCloseOp.aclEntries,
+            path, addCloseOp.permissions, addCloseOp.aclEntries,
             replication, addCloseOp.mtime, addCloseOp.atime,
             addCloseOp.blockSize, true, addCloseOp.clientName,
             addCloseOp.clientMachine);
-        fsNamesys.leaseManager.addLease(addCloseOp.clientName, addCloseOp.path);
+        fsNamesys.leaseManager.addLease(addCloseOp.clientName, path);
 
         // add the op into retry cache if necessary
         if (toAddRetryCache) {
@@ -343,11 +346,11 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
             FSNamesystem.LOG.debug("Reopening an already-closed file " +
                 "for append");
           }
-          LocatedBlock lb = fsNamesys.prepareFileForWrite(addCloseOp.path,
+          LocatedBlock lb = fsNamesys.prepareFileForWrite(path,
               oldFile, addCloseOp.clientName, addCloseOp.clientMachine, null,
               false, iip.getLatestSnapshotId(), false);
-          newFile = INodeFile.valueOf(fsDir.getINode(addCloseOp.path),
-              addCloseOp.path, true);
+          newFile = INodeFile.valueOf(fsDir.getINode(path),
+              path, true);
           
           // add the op into retry cache is necessary
           if (toAddRetryCache) {
@@ -368,16 +371,17 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_CLOSE: {
       AddCloseOp addCloseOp = (AddCloseOp)op;
-      
+      final String path =
+          renameReservedPathsOnUpgrade(addCloseOp.path, logVersion);
       if (FSNamesystem.LOG.isDebugEnabled()) {
-        FSNamesystem.LOG.debug(op.opCode + ": " + addCloseOp.path +
+        FSNamesystem.LOG.debug(op.opCode + ": " + path +
             " numblocks : " + addCloseOp.blocks.length +
             " clientHolder " + addCloseOp.clientName +
             " clientMachine " + addCloseOp.clientMachine);
       }
 
-      final INodesInPath iip = fsDir.getLastINodeInPath(addCloseOp.path);
-      final INodeFile file = INodeFile.valueOf(iip.getINode(0), addCloseOp.path);
+      final INodesInPath iip = fsDir.getLastINodeInPath(path);
+      final INodeFile file = INodeFile.valueOf(iip.getINode(0), path);
 
       // Update the salient file attributes.
       file.setAccessTime(addCloseOp.atime, Snapshot.CURRENT_STATE_ID);
@@ -391,24 +395,26 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
         // could show up twice in a row. But after that version, this
         // should be fixed, so we should treat it as an error.
         throw new IOException(
-            "File is not under construction: " + addCloseOp.path);
+            "File is not under construction: " + path);
       }
       // One might expect that you could use removeLease(holder, path) here,
       // but OP_CLOSE doesn't serialize the holder. So, remove by path.
       if (file.isUnderConstruction()) {
-        fsNamesys.leaseManager.removeLeaseWithPrefixPath(addCloseOp.path);
+        fsNamesys.leaseManager.removeLeaseWithPrefixPath(path);
         file.toCompleteFile(file.getModificationTime());
       }
       break;
     }
     case OP_UPDATE_BLOCKS: {
       UpdateBlocksOp updateOp = (UpdateBlocksOp)op;
+      final String path =
+          renameReservedPathsOnUpgrade(updateOp.path, logVersion);
       if (FSNamesystem.LOG.isDebugEnabled()) {
-        FSNamesystem.LOG.debug(op.opCode + ": " + updateOp.path +
+        FSNamesystem.LOG.debug(op.opCode + ": " + path +
             " numblocks : " + updateOp.blocks.length);
       }
-      INodeFile oldFile = INodeFile.valueOf(fsDir.getINode(updateOp.path),
-          updateOp.path);
+      INodeFile oldFile = INodeFile.valueOf(fsDir.getINode(path),
+          path);
       // Update in-memory data structures
       updateBlocks(fsDir, updateOp, oldFile);
       
@@ -419,7 +425,7 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_ADD_BLOCK: {
       AddBlockOp addBlockOp = (AddBlockOp) op;
-      String path = addBlockOp.getPath();
+      String path = renameReservedPathsOnUpgrade(addBlockOp.getPath(), logVersion);
       if (FSNamesystem.LOG.isDebugEnabled()) {
         FSNamesystem.LOG.debug(op.opCode + ": " + path +
             " new block id : " + addBlockOp.getLastBlock().getBlockId());
@@ -433,14 +439,20 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
       SetReplicationOp setReplicationOp = (SetReplicationOp)op;
       short replication = fsNamesys.getBlockManager().adjustReplication(
           setReplicationOp.replication);
-      fsDir.unprotectedSetReplication(setReplicationOp.path,
+      fsDir.unprotectedSetReplication(
+          renameReservedPathsOnUpgrade(setReplicationOp.path, logVersion),
                                       replication, null);
       break;
     }
     case OP_CONCAT_DELETE: {
       ConcatDeleteOp concatDeleteOp = (ConcatDeleteOp)op;
-      fsDir.unprotectedConcat(concatDeleteOp.trg, concatDeleteOp.srcs,
-          concatDeleteOp.timestamp);
+      String trg = renameReservedPathsOnUpgrade(concatDeleteOp.trg, logVersion);
+      String[] srcs = new String[concatDeleteOp.srcs.length];
+      for (int i=0; i<srcs.length; i++) {
+        srcs[i] =
+            renameReservedPathsOnUpgrade(concatDeleteOp.srcs[i], logVersion);
+      }
+      fsDir.unprotectedConcat(trg, srcs, concatDeleteOp.timestamp);
       
       if (toAddRetryCache) {
         fsNamesys.addCacheEntry(concatDeleteOp.rpcClientId,
@@ -450,7 +462,9 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_RENAME_OLD: {
       RenameOldOp renameOp = (RenameOldOp)op;
-      fsDir.unprotectedRenameTo(renameOp.src, renameOp.dst,
+      final String src = renameReservedPathsOnUpgrade(renameOp.src, logVersion);
+      final String dst = renameReservedPathsOnUpgrade(renameOp.dst, logVersion);
+      fsDir.unprotectedRenameTo(src, dst,
                                 renameOp.timestamp);
       
       if (toAddRetryCache) {
@@ -460,7 +474,9 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_DELETE: {
       DeleteOp deleteOp = (DeleteOp)op;
-      fsDir.unprotectedDelete(deleteOp.path, deleteOp.timestamp);
+      fsDir.unprotectedDelete(
+          renameReservedPathsOnUpgrade(deleteOp.path, logVersion),
+          deleteOp.timestamp);
       
       if (toAddRetryCache) {
         fsNamesys.addCacheEntry(deleteOp.rpcClientId, deleteOp.rpcCallId);
@@ -474,8 +490,9 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
       }
       inodeId = getAndUpdateLastInodeId(mkdirOp.inodeId, logVersion,
           lastInodeId);
-      fsDir.unprotectedMkdir(inodeId, mkdirOp.path, mkdirOp.permissions,
-                             mkdirOp.aclEntries, mkdirOp.timestamp);
+      fsDir.unprotectedMkdir(inodeId,
+          renameReservedPathsOnUpgrade(mkdirOp.path, logVersion),
+          mkdirOp.permissions, mkdirOp.aclEntries, mkdirOp.timestamp);
       break;
     }
     case OP_SET_GENSTAMP_V1: {
@@ -485,53 +502,56 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_SET_PERMISSIONS: {
       SetPermissionsOp setPermissionsOp = (SetPermissionsOp)op;
-      fsDir.unprotectedSetPermission(setPermissionsOp.src,
-                                     setPermissionsOp.permissions);
+      fsDir.unprotectedSetPermission(
+          renameReservedPathsOnUpgrade(setPermissionsOp.src, logVersion),
+          setPermissionsOp.permissions);
       break;
     }
     case OP_SET_OWNER: {
       SetOwnerOp setOwnerOp = (SetOwnerOp)op;
-      fsDir.unprotectedSetOwner(setOwnerOp.src, setOwnerOp.username,
-                                setOwnerOp.groupname);
+      fsDir.unprotectedSetOwner(
+          renameReservedPathsOnUpgrade(setOwnerOp.src, logVersion),
+          setOwnerOp.username, setOwnerOp.groupname);
       break;
     }
     case OP_SET_NS_QUOTA: {
       SetNSQuotaOp setNSQuotaOp = (SetNSQuotaOp)op;
-      fsDir.unprotectedSetQuota(setNSQuotaOp.src,
-                                setNSQuotaOp.nsQuota,
-                                HdfsConstants.QUOTA_DONT_SET);
+      fsDir.unprotectedSetQuota(
+          renameReservedPathsOnUpgrade(setNSQuotaOp.src, logVersion),
+          setNSQuotaOp.nsQuota, HdfsConstants.QUOTA_DONT_SET);
       break;
     }
     case OP_CLEAR_NS_QUOTA: {
       ClearNSQuotaOp clearNSQuotaOp = (ClearNSQuotaOp)op;
-      fsDir.unprotectedSetQuota(clearNSQuotaOp.src,
-                                HdfsConstants.QUOTA_RESET,
-                                HdfsConstants.QUOTA_DONT_SET);
+      fsDir.unprotectedSetQuota(
+          renameReservedPathsOnUpgrade(clearNSQuotaOp.src, logVersion),
+          HdfsConstants.QUOTA_RESET, HdfsConstants.QUOTA_DONT_SET);
       break;
     }
 
     case OP_SET_QUOTA:
       SetQuotaOp setQuotaOp = (SetQuotaOp)op;
-      fsDir.unprotectedSetQuota(setQuotaOp.src,
-                                setQuotaOp.nsQuota,
-                                setQuotaOp.dsQuota);
+      fsDir.unprotectedSetQuota(
+          renameReservedPathsOnUpgrade(setQuotaOp.src, logVersion),
+          setQuotaOp.nsQuota, setQuotaOp.dsQuota);
       break;
 
     case OP_TIMES: {
       TimesOp timesOp = (TimesOp)op;
 
-      fsDir.unprotectedSetTimes(timesOp.path,
-                                timesOp.mtime,
-                                timesOp.atime, true);
+      fsDir.unprotectedSetTimes(
+          renameReservedPathsOnUpgrade(timesOp.path, logVersion),
+          timesOp.mtime, timesOp.atime, true);
       break;
     }
     case OP_SYMLINK: {
       SymlinkOp symlinkOp = (SymlinkOp)op;
       inodeId = getAndUpdateLastInodeId(symlinkOp.inodeId, logVersion,
           lastInodeId);
-      fsDir.unprotectedAddSymlink(inodeId, symlinkOp.path,
-                                  symlinkOp.value, symlinkOp.mtime, 
-                                  symlinkOp.atime, symlinkOp.permissionStatus);
+      fsDir.unprotectedAddSymlink(inodeId,
+          renameReservedPathsOnUpgrade(symlinkOp.path, logVersion),
+          symlinkOp.value, symlinkOp.mtime, symlinkOp.atime,
+          symlinkOp.permissionStatus);
       
       if (toAddRetryCache) {
         fsNamesys.addCacheEntry(symlinkOp.rpcClientId, symlinkOp.rpcCallId);
@@ -540,8 +560,10 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_RENAME: {
       RenameOp renameOp = (RenameOp)op;
-      fsDir.unprotectedRenameTo(renameOp.src, renameOp.dst,
-                                renameOp.timestamp, renameOp.options);
+      fsDir.unprotectedRenameTo(
+          renameReservedPathsOnUpgrade(renameOp.src, logVersion),
+          renameReservedPathsOnUpgrade(renameOp.dst, logVersion),
+          renameOp.timestamp, renameOp.options);
       
       if (toAddRetryCache) {
         fsNamesys.addCacheEntry(renameOp.rpcClientId, renameOp.rpcCallId);
@@ -584,10 +606,12 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
 
       Lease lease = fsNamesys.leaseManager.getLease(
           reassignLeaseOp.leaseHolder);
-      INodeFile pendingFile = fsDir.getINode(reassignLeaseOp.path).asFile();
+      final String path =
+          renameReservedPathsOnUpgrade(reassignLeaseOp.path, logVersion);
+      INodeFile pendingFile = fsDir.getINode(path).asFile();
       Preconditions.checkState(pendingFile.isUnderConstruction());
       fsNamesys.reassignLeaseInternal(lease,
-          reassignLeaseOp.path, reassignLeaseOp.newHolder, pendingFile);
+          path, reassignLeaseOp.newHolder, pendingFile);
       break;
     }
     case OP_START_LOG_SEGMENT:
@@ -597,8 +621,11 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_CREATE_SNAPSHOT: {
       CreateSnapshotOp createSnapshotOp = (CreateSnapshotOp) op;
+      final String snapshotRoot =
+          renameReservedPathsOnUpgrade(createSnapshotOp.snapshotRoot,
+              logVersion);
       String path = fsNamesys.getSnapshotManager().createSnapshot(
-          createSnapshotOp.snapshotRoot, createSnapshotOp.snapshotName);
+          snapshotRoot, createSnapshotOp.snapshotName);
       if (toAddRetryCache) {
         fsNamesys.addCacheEntryWithPayload(createSnapshotOp.rpcClientId,
             createSnapshotOp.rpcCallId, path);
@@ -609,8 +636,11 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
       DeleteSnapshotOp deleteSnapshotOp = (DeleteSnapshotOp) op;
       BlocksMapUpdateInfo collectedBlocks = new BlocksMapUpdateInfo();
       List<INode> removedINodes = new ChunkedArrayList<INode>();
+      final String snapshotRoot =
+          renameReservedPathsOnUpgrade(deleteSnapshotOp.snapshotRoot,
+              logVersion);
       fsNamesys.getSnapshotManager().deleteSnapshot(
-          deleteSnapshotOp.snapshotRoot, deleteSnapshotOp.snapshotName,
+          snapshotRoot, deleteSnapshotOp.snapshotName,
           collectedBlocks, removedINodes);
       fsNamesys.removeBlocksAndUpdateSafemodeTotal(collectedBlocks);
       collectedBlocks.clear();
@@ -625,8 +655,11 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_RENAME_SNAPSHOT: {
       RenameSnapshotOp renameSnapshotOp = (RenameSnapshotOp) op;
+      final String snapshotRoot =
+          renameReservedPathsOnUpgrade(renameSnapshotOp.snapshotRoot,
+              logVersion);
       fsNamesys.getSnapshotManager().renameSnapshot(
-          renameSnapshotOp.snapshotRoot, renameSnapshotOp.snapshotOldName,
+          snapshotRoot, renameSnapshotOp.snapshotOldName,
           renameSnapshotOp.snapshotNewName);
       
       if (toAddRetryCache) {
@@ -637,14 +670,19 @@ private long applyEditLogOp(FSEditLogOp op, FSDirectory fsDir,
     }
     case OP_ALLOW_SNAPSHOT: {
       AllowSnapshotOp allowSnapshotOp = (AllowSnapshotOp) op;
+      final String snapshotRoot =
+          renameReservedPathsOnUpgrade(allowSnapshotOp.snapshotRoot, logVersion);
       fsNamesys.getSnapshotManager().setSnapshottable(
-          allowSnapshotOp.snapshotRoot, false);
+          snapshotRoot, false);
       break;
     }
     case OP_DISALLOW_SNAPSHOT: {
       DisallowSnapshotOp disallowSnapshotOp = (DisallowSnapshotOp) op;
+      final String snapshotRoot =
+          renameReservedPathsOnUpgrade(disallowSnapshotOp.snapshotRoot,
+              logVersion);
       fsNamesys.getSnapshotManager().resetSnapshottable(
-          disallowSnapshotOp.snapshotRoot);
+          snapshotRoot);
       break;
     }
     case OP_SET_GENSTAMP_V2: {

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

@@ -32,12 +32,13 @@
 import java.security.MessageDigest;
 import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.Collection;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
+import java.util.TreeMap;
 
 import org.apache.commons.logging.Log;
-import org.apache.hadoop.HadoopIllegalArgumentException;
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.classification.InterfaceStability;
 import org.apache.hadoop.conf.Configuration;
@@ -45,7 +46,9 @@
 import org.apache.hadoop.fs.PathIsNotDirectoryException;
 import org.apache.hadoop.fs.UnresolvedLinkException;
 import org.apache.hadoop.fs.permission.PermissionStatus;
+import org.apache.hadoop.hdfs.DFSUtil;
 import org.apache.hadoop.hdfs.protocol.HdfsConstants;
+import org.apache.hadoop.hdfs.protocol.LayoutFlags;
 import org.apache.hadoop.hdfs.protocol.LayoutVersion;
 import org.apache.hadoop.hdfs.protocol.LayoutVersion.Feature;
 import org.apache.hadoop.hdfs.protocol.proto.AclProtos.AclFsImageProto;
@@ -54,6 +57,7 @@
 import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
 import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoUnderConstruction;
 import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
+import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.StartupOption;
 import org.apache.hadoop.hdfs.server.common.InconsistentFSStateException;
 import org.apache.hadoop.hdfs.server.namenode.snapshot.DirectoryWithSnapshotFeature;
 import org.apache.hadoop.hdfs.server.namenode.snapshot.FileDiffList;
@@ -69,6 +73,10 @@
 import org.apache.hadoop.hdfs.util.ReadOnlyList;
 import org.apache.hadoop.io.MD5Hash;
 import org.apache.hadoop.io.Text;
+import org.apache.hadoop.util.StringUtils;
+
+import com.google.common.annotations.VisibleForTesting;
+import com.google.common.base.Preconditions;
 
 /**
  * Contains inner classes for reading or writing the on-disk format for
@@ -415,7 +423,8 @@ private void loadLocalNameINodesWithSnapshot(long numFiles, DataInput in,
     }
     
   /** 
-   * load fsimage files assuming only local names are stored
+   * load fsimage files assuming only local names are stored. Used when
+   * snapshots are not supported by the layout version.
    *   
    * @param numFiles number of files expected to be read
    * @param in image input stream
@@ -531,6 +540,8 @@ private void loadDirectoryWithSnapshot(DataInput in, Counter counter)
     */
    private int loadDirectory(DataInput in, Counter counter) throws IOException {
      String parentPath = FSImageSerialization.readString(in);
+     // Rename .snapshot paths if we're doing an upgrade
+     parentPath = renameReservedPathsOnUpgrade(parentPath, getLayoutVersion());
      final INodeDirectory parent = INodeDirectory.valueOf(
          namesystem.dir.rootDir.getNode(parentPath, true), parentPath);
      return loadChildren(parent, in, counter);
@@ -590,11 +601,9 @@ private INodeDirectory getParentINodeDirectory(byte[][] pathComponents
    */
   private void addToParent(INodeDirectory parent, INode child) {
     FSDirectory fsDir = namesystem.dir;
-    if (parent == fsDir.rootDir && FSDirectory.isReservedName(child)) {
-        throw new HadoopIllegalArgumentException("File name \""
-            + child.getLocalName() + "\" is reserved. Please "
-            + " change the name of the existing file or directory to another "
-            + "name before upgrading to this release.");
+    if (parent == fsDir.rootDir) {
+        child.setLocalName(renameReservedRootComponentOnUpgrade(
+            child.getLocalNameBytes(), getLayoutVersion()));
     }
     // NOTE: This does not update space counts for parents
     if (!parent.addChild(child)) {
@@ -631,7 +640,9 @@ public INode loadINodeWithLocalName(boolean isSnapshotINode, DataInput in,
     public INode loadINodeWithLocalName(boolean isSnapshotINode,
         DataInput in, boolean updateINodeMap, Counter counter)
         throws IOException {
-      final byte[] localName = FSImageSerialization.readLocalName(in);
+      byte[] localName = FSImageSerialization.readLocalName(in);
+      localName =
+          renameReservedComponentOnUpgrade(localName, getLayoutVersion());
       INode inode = loadINode(localName, isSnapshotINode, in, counter);
       if (updateINodeMap
           && LayoutVersion.supports(Feature.ADD_INODE_ID, getLayoutVersion())) {
@@ -968,7 +979,156 @@ public Snapshot getSnapshot(DataInput in) throws IOException {
       return snapshotMap.get(in.readInt());
     }
   }
-  
+
+  @VisibleForTesting
+  public static TreeMap<String, String> renameReservedMap =
+      new TreeMap<String, String>();
+
+  /**
+   * Use the default key-value pairs that will be used to determine how to
+   * rename reserved paths on upgrade.
+   */
+  @VisibleForTesting
+  public static void useDefaultRenameReservedPairs() {
+    renameReservedMap.clear();
+    for (String key: HdfsConstants.RESERVED_PATH_COMPONENTS) {
+      renameReservedMap.put(
+          key,
+          key + "." + LayoutVersion.getCurrentLayoutVersion() + "."
+              + "UPGRADE_RENAMED");
+    }
+  }
+
+  /**
+   * Set the key-value pairs that will be used to determine how to rename
+   * reserved paths on upgrade.
+   */
+  @VisibleForTesting
+  public static void setRenameReservedPairs(String renameReserved) {
+    // Clear and set the default values
+    useDefaultRenameReservedPairs();
+    // Overwrite with provided values
+    setRenameReservedMapInternal(renameReserved);
+  }
+
+  private static void setRenameReservedMapInternal(String renameReserved) {
+    Collection<String> pairs =
+        StringUtils.getTrimmedStringCollection(renameReserved);
+    for (String p : pairs) {
+      String[] pair = StringUtils.split(p, '/', '=');
+      Preconditions.checkArgument(pair.length == 2,
+          "Could not parse key-value pair " + p);
+      String key = pair[0];
+      String value = pair[1];
+      Preconditions.checkArgument(DFSUtil.isReservedPathComponent(key),
+          "Unknown reserved path " + key);
+      Preconditions.checkArgument(DFSUtil.isValidNameForComponent(value),
+          "Invalid rename path for " + key + ": " + value);
+      LOG.info("Will rename reserved path " + key + " to " + value);
+      renameReservedMap.put(key, value);
+    }
+  }
+
+  /**
+   * When upgrading from an old version, the filesystem could contain paths
+   * that are now reserved in the new version (e.g. .snapshot). This renames
+   * these new reserved paths to a user-specified value to avoid collisions
+   * with the reserved name.
+   * 
+   * @param path Old path potentially containing a reserved path
+   * @return New path with reserved path components renamed to user value
+   */
+  static String renameReservedPathsOnUpgrade(String path,
+      final int layoutVersion) {
+    final String oldPath = path;
+    // If any known LVs aren't supported, we're doing an upgrade
+    if (!LayoutVersion.supports(Feature.ADD_INODE_ID, layoutVersion)) {
+      String[] components = INode.getPathNames(path);
+      // Only need to worry about the root directory
+      if (components.length > 1) {
+        components[1] = DFSUtil.bytes2String(
+            renameReservedRootComponentOnUpgrade(
+                DFSUtil.string2Bytes(components[1]),
+                layoutVersion));
+        path = DFSUtil.strings2PathString(components);
+      }
+    }
+    if (!LayoutVersion.supports(Feature.SNAPSHOT, layoutVersion)) {
+      String[] components = INode.getPathNames(path);
+      // Special case the root path
+      if (components.length == 0) {
+        return path;
+      }
+      for (int i=0; i<components.length; i++) {
+        components[i] = DFSUtil.bytes2String(
+            renameReservedComponentOnUpgrade(
+                DFSUtil.string2Bytes(components[i]),
+                layoutVersion));
+      }
+      path = DFSUtil.strings2PathString(components);
+    }
+
+    if (!path.equals(oldPath)) {
+      LOG.info("Upgrade process renamed reserved path " + oldPath + " to "
+          + path);
+    }
+    return path;
+  }
+
+  private final static String RESERVED_ERROR_MSG = 
+      FSDirectory.DOT_RESERVED_PATH_PREFIX + " is a reserved path and "
+      + HdfsConstants.DOT_SNAPSHOT_DIR + " is a reserved path component in"
+      + " this version of HDFS. Please rollback and delete or rename"
+      + " this path, or upgrade with the "
+      + StartupOption.RENAMERESERVED.getName()
+      + " [key-value pairs]"
+      + " option to automatically rename these paths during upgrade.";
+
+  /**
+   * Same as {@link #renameReservedPathsOnUpgrade(String)}, but for a single
+   * byte array path component.
+   */
+  private static byte[] renameReservedComponentOnUpgrade(byte[] component,
+      final int layoutVersion) {
+    // If the LV doesn't support snapshots, we're doing an upgrade
+    if (!LayoutVersion.supports(Feature.SNAPSHOT, layoutVersion)) {
+      if (Arrays.equals(component, HdfsConstants.DOT_SNAPSHOT_DIR_BYTES)) {
+        Preconditions.checkArgument(
+            renameReservedMap != null &&
+            renameReservedMap.containsKey(HdfsConstants.DOT_SNAPSHOT_DIR),
+            RESERVED_ERROR_MSG);
+        component =
+            DFSUtil.string2Bytes(renameReservedMap
+                .get(HdfsConstants.DOT_SNAPSHOT_DIR));
+      }
+    }
+    return component;
+  }
+
+  /**
+   * Same as {@link #renameReservedPathsOnUpgrade(String)}, but for a single
+   * byte array path component.
+   */
+  private static byte[] renameReservedRootComponentOnUpgrade(byte[] component,
+      final int layoutVersion) {
+    // If the LV doesn't support inode IDs, we're doing an upgrade
+    if (!LayoutVersion.supports(Feature.ADD_INODE_ID, layoutVersion)) {
+      if (Arrays.equals(component, FSDirectory.DOT_RESERVED)) {
+        Preconditions.checkArgument(
+            renameReservedMap != null &&
+            renameReservedMap.containsKey(FSDirectory.DOT_RESERVED_STRING),
+            RESERVED_ERROR_MSG);
+        final String renameString = renameReservedMap
+            .get(FSDirectory.DOT_RESERVED_STRING);
+        component =
+            DFSUtil.string2Bytes(renameString);
+        LOG.info("Renamed root path " + FSDirectory.DOT_RESERVED_STRING
+            + " to " + renameString);
+      }
+    }
+    return component;
+  }
+
   /**
    * A one-shot class responsible for writing an image file.
    * The write() function should be called once, after which the getter

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

@@ -1166,7 +1166,8 @@ private void checkNameNodeSafeMode(String errorMsg)
     if (isInSafeMode()) {
       SafeModeException se = new SafeModeException(errorMsg, safeMode);
       if (haEnabled && haContext != null
-          && haContext.getState().getServiceState() == HAServiceState.ACTIVE) {
+          && haContext.getState().getServiceState() == HAServiceState.ACTIVE
+          && shouldRetrySafeMode(this.safeMode)) {
         throw new RetriableException(se);
       } else {
         throw se;
@@ -1174,6 +1175,18 @@ private void checkNameNodeSafeMode(String errorMsg)
     }
   }
   
+  /**
+   * We already know that the safemode is on. We will throw a RetriableException
+   * if the safemode is not manual or caused by low resource.
+   */
+  private boolean shouldRetrySafeMode(SafeModeInfo safeMode) {
+    if (safeMode == null) {
+      return false;
+    } else {
+      return !safeMode.isManual() && !safeMode.areResourcesLow();
+    }
+  }
+  
   public static Collection<URI> getNamespaceDirs(Configuration conf) {
     return getStorageDirs(conf, DFS_NAMENODE_NAME_DIR_KEY);
   }

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

@@ -212,7 +212,9 @@ public static enum OperationCategory {
       + StartupOption.CLUSTERID.getName() + " cid ] ["
       + StartupOption.FORCE.getName() + "] ["
       + StartupOption.NONINTERACTIVE.getName() + "] ] | ["
-      + StartupOption.UPGRADE.getName() + "] | ["
+      + StartupOption.UPGRADE.getName() + 
+        " [" + StartupOption.CLUSTERID.getName() + " cid]" +
+        " [" + StartupOption.RENAMERESERVED.getName() + "<k-v pairs>] ] | ["
       + StartupOption.ROLLBACK.getName() + "] | ["
       + StartupOption.FINALIZE.getName() + "] | ["
       + StartupOption.IMPORT.getName() + "] | ["
@@ -1056,7 +1058,8 @@ private static void printUsage(PrintStream out) {
     out.println(USAGE + "\n");
   }
 
-  private static StartupOption parseArguments(String args[]) {
+  @VisibleForTesting
+  static StartupOption parseArguments(String args[]) {
     int argsLen = (args == null) ? 0 : args.length;
     StartupOption startOpt = StartupOption.REGULAR;
     for(int i=0; i < argsLen; i++) {
@@ -1103,11 +1106,33 @@ private static StartupOption parseArguments(String args[]) {
         startOpt = StartupOption.CHECKPOINT;
       } else if (StartupOption.UPGRADE.getName().equalsIgnoreCase(cmd)) {
         startOpt = StartupOption.UPGRADE;
-        // might be followed by two args
-        if (i + 2 < argsLen
-            && args[i + 1].equalsIgnoreCase(StartupOption.CLUSTERID.getName())) {
-          i += 2;
-          startOpt.setClusterId(args[i]);
+        /* Can be followed by CLUSTERID with a required parameter or
+         * RENAMERESERVED with an optional parameter
+         */
+        while (i + 1 < argsLen) {
+          String flag = args[i + 1];
+          if (flag.equalsIgnoreCase(StartupOption.CLUSTERID.getName())) {
+            if (i + 2 < argsLen) {
+              i += 2;
+              startOpt.setClusterId(args[i]);
+            } else {
+              LOG.fatal("Must specify a valid cluster ID after the "
+                  + StartupOption.CLUSTERID.getName() + " flag");
+              return null;
+            }
+          } else if (flag.equalsIgnoreCase(StartupOption.RENAMERESERVED
+              .getName())) {
+            if (i + 2 < argsLen) {
+              FSImageFormat.setRenameReservedPairs(args[i + 2]);
+              i += 2;
+            } else {
+              FSImageFormat.useDefaultRenameReservedPairs();
+              i += 1;
+            }
+          } else {
+            LOG.fatal("Unknown upgrade flag " + flag);
+            return null;
+          }
         }
       } else if (StartupOption.ROLLBACK.getName().equalsIgnoreCase(cmd)) {
         startOpt = StartupOption.ROLLBACK;

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

@@ -140,6 +140,18 @@ private static Long parseTtlString(String maxTtlString) throws IOException {
     return maxTtl;
   }
 
+  private static Long parseLimitString(String limitString) {
+    Long limit = null;
+    if (limitString != null) {
+      if (limitString.equalsIgnoreCase("unlimited")) {
+        limit = CachePoolInfo.LIMIT_UNLIMITED;
+      } else {
+        limit = Long.parseLong(limitString);
+      }
+    }
+    return limit;
+  }
+
   private static Expiration parseExpirationString(String ttlString)
       throws IOException {
     Expiration ex = null;
@@ -650,8 +662,8 @@ public int run(Configuration conf, List<String> args) throws IOException {
         info.setMode(new FsPermission(mode));
       }
       String limitString = StringUtils.popOptionWithArgument("-limit", args);
-      if (limitString != null) {
-        long limit = Long.parseLong(limitString);
+      Long limit = parseLimitString(limitString);
+      if (limit != null) {
         info.setLimit(limit);
       }
       String maxTtlString = StringUtils.popOptionWithArgument("-maxTtl", args);
@@ -726,8 +738,7 @@ public int run(Configuration conf, List<String> args) throws IOException {
       Integer mode = (modeString == null) ?
           null : Integer.parseInt(modeString, 8);
       String limitString = StringUtils.popOptionWithArgument("-limit", args);
-      Long limit = (limitString == null) ?
-          null : Long.parseLong(limitString);
+      Long limit = parseLimitString(limitString);
       String maxTtlString = StringUtils.popOptionWithArgument("-maxTtl", args);
       Long maxTtl = null;
       try {
@@ -962,9 +973,8 @@ public int run(Configuration conf, List<String> args) throws IOException {
       if (numResults > 0) { 
         System.out.print(listing);
       }
-      // If there are no results, we return 1 (failure exit code);
-      // otherwise we return 0 (success exit code).
-      return (numResults == 0) ? 1 : 0;
+      // If list pools succeed, we return 0 (success exit code)
+      return 0;
     }
   }
 

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

@@ -191,6 +191,9 @@ public synchronized void initialize(URI uri, Configuration conf
       int maxFailoverAttempts = conf.getInt(
           DFSConfigKeys.DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_KEY,
           DFSConfigKeys.DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_DEFAULT);
+      int maxRetryAttempts = conf.getInt(
+          DFSConfigKeys.DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_KEY,
+          DFSConfigKeys.DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT);
       int failoverSleepBaseMillis = conf.getInt(
           DFSConfigKeys.DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY,
           DFSConfigKeys.DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_DEFAULT);
@@ -200,7 +203,7 @@ public synchronized void initialize(URI uri, Configuration conf
 
       this.retryPolicy = RetryPolicies
           .failoverOnNetworkException(RetryPolicies.TRY_ONCE_THEN_FAIL,
-              maxFailoverAttempts, failoverSleepBaseMillis,
+              maxFailoverAttempts, maxRetryAttempts, failoverSleepBaseMillis,
               failoverSleepMaxMillis);
     }
 

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

@@ -435,7 +435,7 @@ HDFS Users Guide
    state it was in before the upgrade. HDFS upgrade is described in more
    detail in {{{http://wiki.apache.org/hadoop/Hadoop_Upgrade}Hadoop Upgrade}}
    Wiki page. HDFS can have one such backup at a time. Before upgrading,
-   administrators need to remove existing backupusing bin/hadoop dfsadmin
+   administrators need to remove existing backup using bin/hadoop dfsadmin
    <<<-finalizeUpgrade>>> command. The following briefly describes the
    typical upgrade procedure:
 
@@ -459,6 +459,33 @@ HDFS Users Guide
 
           * start the cluster with rollback option. (<<<bin/start-dfs.sh -rollback>>>).
 
+    When upgrading to a new version of HDFS, it is necessary to rename or
+    delete any paths that are reserved in the new version of HDFS. If the
+    NameNode encounters a reserved path during upgrade, it will print an
+    error like the following:
+
+    <<< /.reserved is a reserved path and .snapshot is a
+    reserved path component in this version of HDFS. Please rollback and delete
+    or rename this path, or upgrade with the -renameReserved [key-value pairs]
+    option to automatically rename these paths during upgrade.>>>
+
+    Specifying <<<-upgrade -renameReserved [optional key-value pairs]>>> causes
+    the NameNode to automatically rename any reserved paths found during
+    startup. For example, to rename all paths named <<<.snapshot>>> to
+    <<<.my-snapshot>>> and <<<.reserved>>> to <<<.my-reserved>>>, a user would
+    specify <<<-upgrade -renameReserved
+    .snapshot=.my-snapshot,.reserved=.my-reserved>>>.
+
+    If no key-value pairs are specified with <<<-renameReserved>>>, the
+    NameNode will then suffix reserved paths with
+    <<<.<LAYOUT-VERSION>.UPGRADE_RENAMED>>>, e.g.
+    <<<.snapshot.-51.UPGRADE_RENAMED>>>.
+
+    There are some caveats to this renaming process. It's recommended,
+    if possible, to first <<<hdfs dfsadmin -saveNamespace>>> before upgrading.
+    This is because data inconsistency can result if an edit log operation
+    refers to the destination of an automatically renamed file.
+
 * File Permissions and Security
 
    The file permissions are designed to be similar to file permissions on

hadoop-hdfs-project/hadoop-hdfs/src/site/xdoc/HdfsSnapshots.xml

@@ -20,7 +20,7 @@
   xsi:schemaLocation="http://maven.apache.org/XDOC/2.0 http://maven.apache.org/xsd/xdoc-2.0.xsd">
 
   <properties>
-    <title>HFDS Snapshots</title>
+    <title>HDFS Snapshots</title>
   </properties>
 
   <body>
@@ -99,15 +99,22 @@
     <li>Copying a file from snapshot <code>s0</code>:
       <source>hdfs dfs -cp /foo/.snapshot/s0/bar /tmp</source></li>
   </ul>
-  <p>
-    <b>Note</b> that the name ".snapshot" is now a reserved file name in HDFS
-    so that users cannot create a file/directory with ".snapshot" as the name.
-    If ".snapshot" is used in a previous version of HDFS, it must be renamed before upgrade;
-    otherwise, upgrade will fail. 
-  </p>
   </subsection>
   </section>
 
+  <section name="Upgrading to a version of HDFS with snapshots" id="Upgrade">
+
+  <p>
+    The HDFS snapshot feature introduces a new reserved path name used to
+    interact with snapshots: <tt>.snapshot</tt>. When upgrading from an
+    older version of HDFS, existing paths named <tt>.snapshot</tt> need
+    to first be renamed or deleted to avoid conflicting with the reserved path.
+    See the upgrade section in
+    <a href="HdfsUserGuide.html#Upgrade_and_Rollback">the HDFS user guide</a>
+    for more information.  </p>
+
+  </section>
+
   <section name="Snapshot Operations" id="SnapshotOperations">
   <subsection name="Administrator Operations" id="AdministratorOperations">
   <p>

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

@@ -27,6 +27,7 @@
 import java.io.FileOutputStream;
 import java.io.FileReader;
 import java.io.IOException;
+import java.util.ArrayList;
 import java.util.Iterator;
 import java.util.LinkedList;
 import java.util.TreeMap;
@@ -43,7 +44,9 @@
 import org.apache.hadoop.hdfs.protocol.HdfsConstants;
 import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
 import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.StartupOption;
+import org.apache.hadoop.hdfs.server.namenode.FSImageFormat;
 import org.apache.hadoop.hdfs.server.namenode.FSImageTestUtil;
+import org.apache.hadoop.test.GenericTestUtils;
 import org.apache.hadoop.util.StringUtils;
 import org.apache.log4j.Logger;
 import org.junit.Test;
@@ -67,6 +70,7 @@ public class TestDFSUpgradeFromImage {
   private static final String HADOOP_DFS_DIR_TXT = "hadoop-dfs-dir.txt";
   private static final String HADOOP22_IMAGE = "hadoop-22-dfs-dir.tgz";
   private static final String HADOOP1_BBW_IMAGE = "hadoop1-bbw.tgz";
+  private static final String HADOOP2_RESERVED_IMAGE = "hadoop-2-reserved.tgz";
 
   private static class ReferenceFileInfo {
     String path;
@@ -320,6 +324,87 @@ public void testUpgradeFromCorruptRel22Image() throws IOException {
       assertEquals("Upgrade did not fail with bad MD5", 1, md5failures);
     }
   }
+
+  /**
+   * Test upgrade from 2.0 image with a variety of .snapshot and .reserved
+   * paths to test renaming on upgrade
+   */
+  @Test
+  public void testUpgradeFromRel2ReservedImage() throws IOException {
+    unpackStorage(HADOOP2_RESERVED_IMAGE);
+    MiniDFSCluster cluster = null;
+    // Try it once without setting the upgrade flag to ensure it fails
+    try {
+      cluster =
+          new MiniDFSCluster.Builder(new Configuration())
+              .format(false)
+              .startupOption(StartupOption.UPGRADE)
+              .numDataNodes(0).build();
+    } catch (IllegalArgumentException e) {
+      GenericTestUtils.assertExceptionContains(
+          "reserved path component in this version",
+          e);
+    } finally {
+      if (cluster != null) {
+        cluster.shutdown();
+      }
+    }
+    // Try it again with a custom rename string
+    try {
+      FSImageFormat.setRenameReservedPairs(
+          ".snapshot=.user-snapshot," +
+          ".reserved=.my-reserved");
+      cluster =
+          new MiniDFSCluster.Builder(new Configuration())
+              .format(false)
+              .startupOption(StartupOption.UPGRADE)
+              .numDataNodes(0).build();
+      // Make sure the paths were renamed as expected
+      DistributedFileSystem dfs = cluster.getFileSystem();
+      ArrayList<Path> toList = new ArrayList<Path>();
+      ArrayList<String> found = new ArrayList<String>();
+      toList.add(new Path("/"));
+      while (!toList.isEmpty()) {
+        Path p = toList.remove(0);
+        FileStatus[] statuses = dfs.listStatus(p);
+        for (FileStatus status: statuses) {
+          final String path = status.getPath().toUri().getPath();
+          System.out.println("Found path " + path);
+          found.add(path);
+          if (status.isDirectory()) {
+            toList.add(status.getPath());
+          }
+        }
+      }
+      String[] expected = new String[] {
+          "/edits",
+          "/edits/.reserved",
+          "/edits/.user-snapshot",
+          "/edits/.user-snapshot/editsdir",
+          "/edits/.user-snapshot/editsdir/editscontents",
+          "/edits/.user-snapshot/editsdir/editsdir2",
+          "/image",
+          "/image/.reserved",
+          "/image/.user-snapshot",
+          "/image/.user-snapshot/imagedir",
+          "/image/.user-snapshot/imagedir/imagecontents",
+          "/image/.user-snapshot/imagedir/imagedir2",
+          "/.my-reserved",
+          "/.my-reserved/edits-touch",
+          "/.my-reserved/image-touch"
+      };
+
+      for (String s: expected) {
+        assertTrue("Did not find expected path " + s, found.contains(s));
+      }
+      assertEquals("Found an unexpected path while listing filesystem",
+          found.size(), expected.length);
+    } finally {
+      if (cluster != null) {
+        cluster.shutdown();
+      }
+    }
+  }
     
   static void recoverAllLeases(DFSClient dfs, 
       Path path) throws IOException {

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

@@ -19,16 +19,19 @@
 
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_MAX_BLOCK_ACQUIRE_FAILURES_KEY;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT;
 import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_SOCKET_WRITE_TIMEOUT_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT;
 import static org.junit.Assert.assertEquals;
 import static org.junit.Assert.assertNotNull;
 import static org.junit.Assert.assertTrue;
 import static org.junit.Assert.fail;
 
+import com.google.common.base.Supplier;
+
 import java.io.InputStream;
 import java.io.PrintWriter;
-import java.net.InetSocketAddress;
-import java.net.Socket;
 
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.FSDataInputStream;
@@ -37,10 +40,8 @@
 import org.apache.hadoop.hdfs.MiniDFSCluster.DataNodeProperties;
 import org.apache.hadoop.hdfs.net.Peer;
 import org.apache.hadoop.hdfs.server.datanode.DataNode;
-import org.apache.hadoop.hdfs.server.datanode.DataNodeTestUtils;
-import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
 import org.apache.hadoop.io.IOUtils;
-import org.apache.hadoop.net.NetUtils;
+import org.apache.hadoop.test.GenericTestUtils;
 import org.apache.hadoop.util.ReflectionUtils;
 import org.junit.After;
 import org.junit.Before;
@@ -51,10 +52,7 @@
 public class TestDataTransferKeepalive {
   Configuration conf = new HdfsConfiguration();
   private MiniDFSCluster cluster;
-  private FileSystem fs;
-  private InetSocketAddress dnAddr;
   private DataNode dn;
-  private DFSClient dfsClient;
   private static Path TEST_FILE = new Path("/test");
   
   private static final int KEEPALIVE_TIMEOUT = 1000;
@@ -69,15 +67,7 @@ public void setup() throws Exception {
     
     cluster = new MiniDFSCluster.Builder(conf)
       .numDataNodes(1).build();
-    fs = cluster.getFileSystem();
-    dfsClient = ((DistributedFileSystem)fs).dfs;
-    dfsClient.peerCache.clear();
-
-    String poolId = cluster.getNamesystem().getBlockPoolId();
     dn = cluster.getDataNodes().get(0);
-    DatanodeRegistration dnReg = DataNodeTestUtils.getDNRegistrationForBP(
-        dn, poolId);
-    dnAddr = NetUtils.createSocketAddr(dnReg.getXferAddr());
   }
   
   @After
@@ -90,34 +80,86 @@ public void teardown() {
    * its configured keepalive timeout.
    */
   @Test(timeout=30000)
-  public void testKeepaliveTimeouts() throws Exception {
+  public void testDatanodeRespectsKeepAliveTimeout() throws Exception {
+    Configuration clientConf = new Configuration(conf);
+    // Set a client socket cache expiry time much longer than 
+    // the datanode-side expiration time.
+    final long CLIENT_EXPIRY_MS = 60000L;
+    clientConf.setLong(DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_KEY, CLIENT_EXPIRY_MS);
+    PeerCache.setInstance(DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT, CLIENT_EXPIRY_MS);
+    DistributedFileSystem fs =
+        (DistributedFileSystem)FileSystem.get(cluster.getURI(),
+            clientConf);
+
     DFSTestUtil.createFile(fs, TEST_FILE, 1L, (short)1, 0L);
 
     // Clients that write aren't currently re-used.
-    assertEquals(0, dfsClient.peerCache.size());
+    assertEquals(0, fs.dfs.peerCache.size());
     assertXceiverCount(0);
 
     // Reads the file, so we should get a
     // cached socket, and should have an xceiver on the other side.
     DFSTestUtil.readFile(fs, TEST_FILE);
-    assertEquals(1, dfsClient.peerCache.size());
+    assertEquals(1, fs.dfs.peerCache.size());
     assertXceiverCount(1);
 
     // Sleep for a bit longer than the keepalive timeout
     // and make sure the xceiver died.
-    Thread.sleep(KEEPALIVE_TIMEOUT * 2);
+    Thread.sleep(DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT + 1);
     assertXceiverCount(0);
     
     // The socket is still in the cache, because we don't
     // notice that it's closed until we try to read
     // from it again.
-    assertEquals(1, dfsClient.peerCache.size());
+    assertEquals(1, fs.dfs.peerCache.size());
     
     // Take it out of the cache - reading should
     // give an EOF.
-    Peer peer = dfsClient.peerCache.get(dn.getDatanodeId(), false);
+    Peer peer = fs.dfs.peerCache.get(dn.getDatanodeId(), false);
     assertNotNull(peer);
     assertEquals(-1, peer.getInputStream().read());
+    PeerCache.setInstance(DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT,
+        DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT);
+  }
+
+  /**
+   * Test that the client respects its keepalive timeout.
+   */
+  @Test(timeout=30000)
+  public void testClientResponsesKeepAliveTimeout() throws Exception {
+    Configuration clientConf = new Configuration(conf);
+    // Set a client socket cache expiry time much shorter than 
+    // the datanode-side expiration time.
+    final long CLIENT_EXPIRY_MS = 10L;
+    clientConf.setLong(DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_KEY, CLIENT_EXPIRY_MS);
+    PeerCache.setInstance(DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT, CLIENT_EXPIRY_MS);
+    DistributedFileSystem fs =
+        (DistributedFileSystem)FileSystem.get(cluster.getURI(),
+            clientConf);
+
+    DFSTestUtil.createFile(fs, TEST_FILE, 1L, (short)1, 0L);
+
+    // Clients that write aren't currently re-used.
+    assertEquals(0, fs.dfs.peerCache.size());
+    assertXceiverCount(0);
+
+    // Reads the file, so we should get a
+    // cached socket, and should have an xceiver on the other side.
+    DFSTestUtil.readFile(fs, TEST_FILE);
+    assertEquals(1, fs.dfs.peerCache.size());
+    assertXceiverCount(1);
+
+    // Sleep for a bit longer than the client keepalive timeout.
+    Thread.sleep(CLIENT_EXPIRY_MS + 1);
+    
+    // Taking out a peer which is expired should give a null.
+    Peer peer = fs.dfs.peerCache.get(dn.getDatanodeId(), false);
+    assertTrue(peer == null);
+
+    // The socket cache is now empty.
+    assertEquals(0, fs.dfs.peerCache.size());
+    PeerCache.setInstance(DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT,
+        DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_DEFAULT);
   }
 
   /**
@@ -125,8 +167,17 @@ public void testKeepaliveTimeouts() throws Exception {
    * read bytes off the stream quickly. The datanode should time out sending the
    * chunks and the transceiver should die, even if it has a long keepalive.
    */
-  @Test(timeout=30000)
+  @Test(timeout=300000)
   public void testSlowReader() throws Exception {
+    // Set a client socket cache expiry time much longer than 
+    // the datanode-side expiration time.
+    final long CLIENT_EXPIRY_MS = 600000L;
+    Configuration clientConf = new Configuration(conf);
+    clientConf.setLong(DFS_CLIENT_SOCKET_CACHE_EXPIRY_MSEC_KEY, CLIENT_EXPIRY_MS);
+    PeerCache.setInstance(DFS_CLIENT_SOCKET_CACHE_CAPACITY_DEFAULT, CLIENT_EXPIRY_MS);
+    DistributedFileSystem fs =
+        (DistributedFileSystem)FileSystem.get(cluster.getURI(),
+            clientConf);
     // Restart the DN with a shorter write timeout.
     DataNodeProperties props = cluster.stopDataNode(0);
     props.conf.setInt(DFS_DATANODE_SOCKET_WRITE_TIMEOUT_KEY,
@@ -134,38 +185,31 @@ public void testSlowReader() throws Exception {
     props.conf.setInt(DFS_DATANODE_SOCKET_REUSE_KEEPALIVE_KEY,
         120000);
     assertTrue(cluster.restartDataNode(props, true));
+    dn = cluster.getDataNodes().get(0);
     // Wait for heartbeats to avoid a startup race where we
     // try to write the block while the DN is still starting.
     cluster.triggerHeartbeats();
     
-    dn = cluster.getDataNodes().get(0);
-    
     DFSTestUtil.createFile(fs, TEST_FILE, 1024*1024*8L, (short)1, 0L);
     FSDataInputStream stm = fs.open(TEST_FILE);
-    try {
-      stm.read();
-      assertXceiverCount(1);
+    stm.read();
+    assertXceiverCount(1);
 
-      // Poll for 0 running xceivers.  Allow up to 5 seconds for some slack.
-      long totalSleepTime = 0;
-      long sleepTime = WRITE_TIMEOUT + 100;
-      while (getXceiverCountWithoutServer() > 0 && totalSleepTime < 5000) {
-        Thread.sleep(sleepTime);
-        totalSleepTime += sleepTime;
-        sleepTime = 100;
+    GenericTestUtils.waitFor(new Supplier<Boolean>() {
+      public Boolean get() {
+        // DN should time out in sendChunks, and this should force
+        // the xceiver to exit.
+        return getXceiverCountWithoutServer() == 0;
       }
+    }, 500, 50000);
 
-      // DN should time out in sendChunks, and this should force
-      // the xceiver to exit.
-      assertXceiverCount(0);
-    } finally {
-      IOUtils.closeStream(stm);
-    }
+    IOUtils.closeStream(stm);
   }
   
   @Test(timeout=30000)
   public void testManyClosedSocketsInCache() throws Exception {
     // Make a small file
+    DistributedFileSystem fs = cluster.getFileSystem();
     DFSTestUtil.createFile(fs, TEST_FILE, 1L, (short)1, 0L);
 
     // Insert a bunch of dead sockets in the cache, by opening

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

@@ -0,0 +1,54 @@
+/**
+ * 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;
+
+import static org.apache.hadoop.http.HttpConfig.Policy.HTTP_AND_HTTPS;
+import static org.apache.hadoop.http.HttpConfig.Policy.HTTP_ONLY;
+
+import org.apache.hadoop.HadoopIllegalArgumentException;
+import org.apache.hadoop.conf.Configuration;
+import org.junit.Assert;
+import org.junit.Test;
+
+public final class TestHttpPolicy {
+
+  @Test(expected = HadoopIllegalArgumentException.class)
+  public void testInvalidPolicyValue() {
+    Configuration conf = new Configuration();
+    conf.set(DFSConfigKeys.DFS_HTTP_POLICY_KEY, "invalid");
+    DFSUtil.getHttpPolicy(conf);
+  }
+
+  @Test
+  public void testDeprecatedConfiguration() {
+    Configuration conf = new Configuration(false);
+    Assert.assertSame(HTTP_ONLY, DFSUtil.getHttpPolicy(conf));
+
+    conf.setBoolean(DFSConfigKeys.DFS_HTTPS_ENABLE_KEY, true);
+    Assert.assertSame(HTTP_AND_HTTPS, DFSUtil.getHttpPolicy(conf));
+
+    conf = new Configuration(false);
+    conf.setBoolean(DFSConfigKeys.HADOOP_SSL_ENABLED_KEY, true);
+    Assert.assertSame(HTTP_AND_HTTPS, DFSUtil.getHttpPolicy(conf));
+
+    conf = new Configuration(false);
+    conf.set(DFSConfigKeys.DFS_HTTP_POLICY_KEY, HTTP_ONLY.name());
+    conf.setBoolean(DFSConfigKeys.DFS_HTTPS_ENABLE_KEY, true);
+    Assert.assertSame(HTTP_ONLY, DFSUtil.getHttpPolicy(conf));
+  }
+}

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/client/TestShortCircuitSharedMemorySegment.java

@@ -0,0 +1,104 @@
+/**
+ * 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.client;
+
+import java.io.File;
+import java.io.FileInputStream;
+import java.util.ArrayList;
+
+import org.apache.commons.lang.SystemUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.fs.FileUtil;
+import org.apache.hadoop.io.nativeio.NativeIO;
+import org.apache.hadoop.io.nativeio.SharedFileDescriptorFactory;
+import org.apache.hadoop.hdfs.client.ShortCircuitSharedMemorySegment.Slot;
+import org.junit.Assume;
+import org.junit.Before;
+import org.junit.Test;
+import org.junit.Assert;
+
+public class TestShortCircuitSharedMemorySegment {
+  public static final Log LOG =
+      LogFactory.getLog(TestShortCircuitSharedMemorySegment.class);
+  
+  private static final File TEST_BASE =
+      new File(System.getProperty("test.build.data", "/tmp"));
+
+  @Before
+  public void before() {
+    Assume.assumeTrue(NativeIO.isAvailable());
+    Assume.assumeTrue(SystemUtils.IS_OS_UNIX);
+  }
+
+  @Test(timeout=60000)
+  public void testStartupShutdown() throws Exception {
+    File path = new File(TEST_BASE, "testStartupShutdown");
+    path.mkdirs();
+    SharedFileDescriptorFactory factory =
+        new SharedFileDescriptorFactory("shm_", path.getAbsolutePath());
+    FileInputStream stream = factory.createDescriptor(4096);
+    ShortCircuitSharedMemorySegment shm = 
+        new ShortCircuitSharedMemorySegment(stream);
+    shm.close();
+    stream.close();
+    FileUtil.fullyDelete(path);
+  }
+
+  @Test(timeout=60000)
+  public void testAllocateSlots() throws Exception {
+    File path = new File(TEST_BASE, "testAllocateSlots");
+    path.mkdirs();
+    SharedFileDescriptorFactory factory =
+        new SharedFileDescriptorFactory("shm_", path.getAbsolutePath());
+    FileInputStream stream = factory.createDescriptor(4096);
+    ShortCircuitSharedMemorySegment shm = 
+        new ShortCircuitSharedMemorySegment(stream);
+    int numSlots = 0;
+    ArrayList<Slot> slots = new ArrayList<Slot>();
+    while (true) {
+      Slot slot = shm.allocateNextSlot();
+      if (slot == null) {
+        LOG.info("allocated " + numSlots + " slots before running out.");
+        break;
+      }
+      slots.add(slot);
+      numSlots++;
+    }
+    int slotIdx = 0;
+    for (Slot slot : slots) {
+      Assert.assertFalse(slot.addAnchor());
+      Assert.assertEquals(slotIdx++, slot.getIndex());
+    }
+    for (Slot slot : slots) {
+      slot.makeAnchorable();
+    }
+    for (Slot slot : slots) {
+      Assert.assertTrue(slot.addAnchor());
+    }
+    for (Slot slot : slots) {
+      slot.removeAnchor();
+    }
+    shm.close();
+    for (Slot slot : slots) {
+      slot.close();
+    }
+    stream.close();
+    FileUtil.fullyDelete(path);
+  }
+}

hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/balancer/TestBalancerWithNodeGroup.java

@@ -65,7 +65,7 @@ public class TestBalancerWithNodeGroup {
 
   ClientProtocol client;
 
-  static final long TIMEOUT = 20000L; //msec
+  static final long TIMEOUT = 40000L; //msec
   static final double CAPACITY_ALLOWED_VARIANCE = 0.005;  // 0.5%
   static final double BALANCE_ALLOWED_VARIANCE = 0.11;    // 10%+delta
   static final int DEFAULT_BLOCK_SIZE = 10;

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

@@ -124,6 +124,8 @@ public void setUp() throws Exception {
     CONF.set(CommonConfigurationKeysPublic.NET_TOPOLOGY_IMPL_KEY, 
         NetworkTopologyWithNodeGroup.class.getName());
     
+    CONF.setBoolean(DFSConfigKeys.DFS_NAMENODE_AVOID_STALE_DATANODE_FOR_WRITE_KEY, true);
+    
     File baseDir = PathUtils.getTestDir(TestReplicationPolicyWithNodeGroup.class);
     
     CONF.set(DFSConfigKeys.DFS_NAMENODE_NAME_DIR_KEY,

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

@@ -28,6 +28,7 @@
 import java.util.ArrayList;
 import java.util.Collection;
 import java.util.Collections;
+import java.util.Enumeration;
 import java.util.List;
 import java.util.regex.Pattern;
 
@@ -301,11 +302,18 @@ private void verifyAuditLogsRepeat(boolean expectSuccess, int ndupe)
     // Turn off the logs
     Logger logger = ((Log4JLogger) FSNamesystem.auditLog).getLogger();
     logger.setLevel(Level.OFF);
-    
+
+    // Close the appenders and force all logs to be flushed
+    Enumeration<?> appenders = logger.getAllAppenders();
+    while (appenders.hasMoreElements()) {
+      Appender appender = (Appender)appenders.nextElement();
+      appender.close();
+    }
+
     BufferedReader reader = new BufferedReader(new FileReader(auditLogFile));
     String line = null;
     boolean ret = true;
-   
+
     try {
       for (int i = 0; i < ndupe; i++) {
         line = reader.readLine();

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

@@ -85,6 +85,7 @@ public static void tearDown() throws Exception {
   @Test
   public void testHttpPolicy() throws Exception {
     conf.set(DFSConfigKeys.DFS_HTTP_POLICY_KEY, policy.name());
+    conf.set(DFSConfigKeys.DFS_NAMENODE_HTTPS_ADDRESS_KEY, "localhost:0");
 
     InetSocketAddress addr = InetSocketAddress.createUnresolved("localhost", 0);
     NameNodeHttpServer server = null;
@@ -103,7 +104,9 @@ public void testHttpPolicy() throws Exception {
           server.getHttpsAddress() == null));
 
     } finally {
-      server.stop();
+      if (server != null) {
+        server.stop();
+      }
     }
   }
 

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

@@ -0,0 +1,105 @@
+/**
+ * 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.apache.hadoop.test.GenericTestUtils.assertExceptionContains;
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertNull;
+import static org.junit.Assert.assertTrue;
+
+import org.apache.hadoop.hdfs.protocol.LayoutVersion;
+import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.StartupOption;
+import org.junit.Test;
+
+public class TestNameNodeOptionParsing {
+
+  @Test(timeout = 10000)
+  public void testUpgrade() {
+    StartupOption opt = null;
+    // UPGRADE is set, but nothing else
+    opt = NameNode.parseArguments(new String[] {"-upgrade"});
+    assertEquals(opt, StartupOption.UPGRADE);
+    assertNull(opt.getClusterId());
+    assertTrue(FSImageFormat.renameReservedMap.isEmpty());
+    // cluster ID is set
+    opt = NameNode.parseArguments(new String[] { "-upgrade", "-clusterid",
+        "mycid" });
+    assertEquals(StartupOption.UPGRADE, opt);
+    assertEquals("mycid", opt.getClusterId());
+    assertTrue(FSImageFormat.renameReservedMap.isEmpty());
+    // Everything is set
+    opt = NameNode.parseArguments(new String[] { "-upgrade", "-clusterid",
+        "mycid", "-renameReserved",
+        ".snapshot=.my-snapshot,.reserved=.my-reserved" });
+    assertEquals(StartupOption.UPGRADE, opt);
+    assertEquals("mycid", opt.getClusterId());
+    assertEquals(".my-snapshot",
+        FSImageFormat.renameReservedMap.get(".snapshot"));
+    assertEquals(".my-reserved",
+        FSImageFormat.renameReservedMap.get(".reserved"));
+    // Reset the map
+    FSImageFormat.renameReservedMap.clear();
+    // Everything is set, but in a different order
+    opt = NameNode.parseArguments(new String[] { "-upgrade", "-renameReserved",
+        ".reserved=.my-reserved,.snapshot=.my-snapshot", "-clusterid",
+        "mycid"});
+    assertEquals(StartupOption.UPGRADE, opt);
+    assertEquals("mycid", opt.getClusterId());
+    assertEquals(".my-snapshot",
+        FSImageFormat.renameReservedMap.get(".snapshot"));
+    assertEquals(".my-reserved",
+        FSImageFormat.renameReservedMap.get(".reserved"));
+    // Try the default renameReserved
+    opt = NameNode.parseArguments(new String[] { "-upgrade", "-renameReserved"});
+    assertEquals(StartupOption.UPGRADE, opt);
+    assertEquals(
+        ".snapshot." + LayoutVersion.getCurrentLayoutVersion()
+            + ".UPGRADE_RENAMED",
+        FSImageFormat.renameReservedMap.get(".snapshot"));
+    assertEquals(
+        ".reserved." + LayoutVersion.getCurrentLayoutVersion()
+            + ".UPGRADE_RENAMED",
+        FSImageFormat.renameReservedMap.get(".reserved"));
+
+    // Try some error conditions
+    try {
+      opt =
+          NameNode.parseArguments(new String[] { "-upgrade", "-renameReserved",
+              ".reserved=.my-reserved,.not-reserved=.my-not-reserved" });
+    } catch (IllegalArgumentException e) {
+      assertExceptionContains("Unknown reserved path", e);
+    }
+    try {
+      opt =
+          NameNode.parseArguments(new String[] { "-upgrade", "-renameReserved",
+              ".reserved=.my-reserved,.snapshot=.snapshot" });
+    } catch (IllegalArgumentException e) {
+      assertExceptionContains("Invalid rename path", e);
+    }
+    try {
+      opt =
+          NameNode.parseArguments(new String[] { "-upgrade", "-renameReserved",
+              ".snapshot=.reserved" });
+    } catch (IllegalArgumentException e) {
+      assertExceptionContains("Invalid rename path", e);
+    }
+    opt = NameNode.parseArguments(new String[] { "-upgrade", "-cid"});
+    assertNull(opt);
+  }
+
+}

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

@@ -55,6 +55,7 @@
 import org.apache.hadoop.hdfs.server.blockmanagement.BlockManagerTestUtil;
 import org.apache.hadoop.hdfs.server.namenode.FSImage;
 import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
+import org.apache.hadoop.hdfs.server.namenode.FSNamesystem.SafeModeInfo;
 import org.apache.hadoop.hdfs.server.namenode.NameNode;
 import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
 import org.apache.hadoop.io.IOUtils;
@@ -65,6 +66,7 @@
 import org.junit.After;
 import org.junit.Before;
 import org.junit.Test;
+import org.mockito.internal.util.reflection.Whitebox;
 
 import com.google.common.base.Supplier;
 import com.google.common.collect.Lists;
@@ -124,6 +126,9 @@ public void testClientRetrySafeMode() throws Exception {
     final Path test = new Path("/test");
     // let nn0 enter safemode
     NameNodeAdapter.enterSafeMode(nn0, false);
+    SafeModeInfo safeMode = (SafeModeInfo) Whitebox.getInternalState(
+        nn0.getNamesystem(), "safeMode");
+    Whitebox.setInternalState(safeMode, "extension", Integer.valueOf(30000));
     LOG.info("enter safemode");
     new Thread() {
       @Override

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

@@ -52,6 +52,7 @@ public static void setUp() throws Exception {
     conf.setBoolean(DFSConfigKeys.DFS_WEBHDFS_ENABLED_KEY, true);
     conf.set(DFSConfigKeys.DFS_HTTP_POLICY_KEY, HttpConfig.Policy.HTTPS_ONLY.name());
     conf.set(DFSConfigKeys.DFS_NAMENODE_HTTPS_ADDRESS_KEY, "localhost:0");
+    conf.set(DFSConfigKeys.DFS_DATANODE_HTTPS_ADDRESS_KEY, "localhost:0");
 
     File base = new File(BASEDIR);
     FileUtil.fullyDelete(base);

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

@@ -469,6 +469,8 @@
       </test-commands>
       <cleanup-commands>
         <cache-admin-command>-removePool pool1</cache-admin-command>
+        <cache-admin-command>-removePool pool2</cache-admin-command>
+        <cache-admin-command>-removePool pool3</cache-admin-command>
       </cleanup-commands>
       <comparators>
         <comparator>
@@ -489,5 +491,33 @@
         </comparator>
       </comparators>
     </test>
+
+    <test> <!--Tested -->
+      <description>Testing setting pool unlimited limits</description>
+      <test-commands>
+        <cache-admin-command>-addPool pool1 -limit unlimited -owner andrew -group andrew</cache-admin-command>
+        <cache-admin-command>-addPool pool2 -limit 10 -owner andrew -group andrew</cache-admin-command>
+        <cache-admin-command>-modifyPool pool2 -limit unlimited</cache-admin-command>
+        <cache-admin-command>-listPools</cache-admin-command>
+      </test-commands>
+      <cleanup-commands>
+        <cache-admin-command>-removePool pool1</cache-admin-command>
+        <cache-admin-command>-removePool pool2</cache-admin-command>
+      </cleanup-commands>
+      <comparators>
+        <comparator>
+          <type>SubstringComparator</type>
+          <expected-output>Found 2 results</expected-output>
+        </comparator>
+        <comparator>
+          <type>SubstringComparator</type>
+          <expected-output>pool1  andrew  andrew  rwxr-xr-x   unlimited   never</expected-output>
+        </comparator>
+        <comparator>
+          <type>SubstringComparator</type>
+          <expected-output>pool2  andrew  andrew  rwxr-xr-x   unlimited   never</expected-output>
+        </comparator>
+      </comparators>
+    </test>
   </tests>
 </configuration>

hadoop-mapreduce-project/CHANGES.txt

@@ -153,6 +153,8 @@ Release 2.4.0 - UNRELEASED
     MAPREDUCE-5732. Report proper queue when job has been automatically placed
     (Sandy Ryza)
 
+    MAPREDUCE-5699. Allow setting tags on MR jobs (kasha)
+
   OPTIMIZATIONS
 
   BUG FIXES
@@ -228,6 +230,10 @@ Release 2.3.0 - UNRELEASED
     MAPREDUCE-5725. Make explicit that TestNetworkedJob relies on the Capacity
     Scheduler (Sandy Ryza)
 
+    MAPREDUCE-5744. Job hangs because 
+    RMContainerAllocator$AssignedRequests.preemptReduce() violates the 
+    comparator contract (Gera Shegalov via kasha)
+
   OPTIMIZATIONS
 
     MAPREDUCE-4680. Job history cleaner should only check timestamps of files in
@@ -347,6 +353,9 @@ Release 2.3.0 - UNRELEASED
     MAPREDUCE-5723. MR AM container log can be truncated or empty.
     (Mohammad Kamrul Islam via kasha)
 
+    MAPREDUCE-5743. Fixed the test failure in TestRMContainerAllocator.
+    (Ted Yu and Vinod Kumar Vavilapalli via zjshen)
+
 Release 2.2.0 - 2013-10-13
 
   INCOMPATIBLE CHANGES

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

@@ -1143,9 +1143,9 @@ void preemptReduce(int toPreempt) {
           new Comparator<TaskAttemptId>() {
         @Override
         public int compare(TaskAttemptId o1, TaskAttemptId o2) {
-          float p = getJob().getTask(o1.getTaskId()).getAttempt(o1).getProgress() -
-              getJob().getTask(o2.getTaskId()).getAttempt(o2).getProgress();
-          return p >= 0 ? 1 : -1;
+          return Float.compare(
+              getJob().getTask(o1.getTaskId()).getAttempt(o1).getProgress(),
+              getJob().getTask(o2.getTaskId()).getAttempt(o2).getProgress());
         }
       });
       

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

@@ -1652,8 +1652,16 @@ public void testCompletedTasksRecalculateSchedule() throws Exception {
     RMApp app = rm.submitApp(1024);
     dispatcher.await();
 
+    // Make a node to register so as to launch the AM.
+    MockNM amNodeManager = rm.registerNode("amNM:1234", 2048);
+    amNodeManager.nodeHeartbeat(true);
+    dispatcher.await();
+
     ApplicationAttemptId appAttemptId = app.getCurrentAppAttempt()
         .getAppAttemptId();
+    rm.sendAMLaunched(appAttemptId);
+    dispatcher.await();
+
     JobId jobId = MRBuilderUtils.newJobId(appAttemptId.getApplicationId(), 0);
     Job job = mock(Job.class);
     when(job.getReport()).thenReturn(

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

@@ -60,6 +60,8 @@ public interface MRJobConfig {
 
   public static final String QUEUE_NAME = "mapreduce.job.queuename";
 
+  public static final String JOB_TAGS = "mapreduce.job.tags";
+
   public static final String JVM_NUMTASKS_TORUN = "mapreduce.job.jvm.numtasks";
 
   public static final String SPLIT_FILE = "mapreduce.job.splitfile";

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/resources/mapred-default.xml

@@ -1,4 +1,5 @@
 <?xml version="1.0"?>
+<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
 <!--
    Licensed to the Apache Software Foundation (ASF) under one or more
    contributor license agreements.  See the NOTICE file distributed with
@@ -15,7 +16,6 @@
    See the License for the specific language governing permissions and
    limitations under the License.
 -->
-<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
 
 <!-- Do not modify this file directly.  Instead, copy entries that you -->
 <!-- wish to modify from this file into mapred-site.xml and change them -->
@@ -727,6 +727,14 @@
   </description>
 </property>
 
+  <property>
+    <name>mapreduce.job.tags</name>
+    <value></value>
+    <description> Tags for the job that will be passed to YARN at submission 
+      time. Queries to YARN for applications can filter on these tags.
+    </description>
+  </property>
+
 <property>
   <name>mapreduce.cluster.local.dir</name>
   <value>${hadoop.tmp.dir}/mapred/local</value>

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/main/java/org/apache/hadoop/mapred/ResourceMgrDelegate.java

@@ -400,4 +400,10 @@ public List<ContainerReport> getContainers(
       IOException {
     return client.getContainers(applicationAttemptId);
   }
+
+  @Override
+  public void moveApplicationAcrossQueues(ApplicationId appId, String queue)
+      throws YarnException, IOException {
+    client.moveApplicationAcrossQueues(appId, queue);
+  }
 }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/main/java/org/apache/hadoop/mapred/YARNRunner.java

@@ -21,7 +21,9 @@
 import java.io.IOException;
 import java.nio.ByteBuffer;
 import java.util.ArrayList;
+import java.util.Collection;
 import java.util.HashMap;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Map;
 import java.util.Vector;
@@ -467,6 +469,8 @@ public ApplicationSubmissionContext createApplicationSubmissionContext(
         ContainerLaunchContext.newInstance(localResources, environment,
           vargsFinal, null, securityTokens, acls);
 
+    Collection<String> tagsFromConf =
+        jobConf.getTrimmedStringCollection(MRJobConfig.JOB_TAGS);
 
     // Set up the ApplicationSubmissionContext
     ApplicationSubmissionContext appContext =
@@ -486,6 +490,9 @@ public ApplicationSubmissionContext createApplicationSubmissionContext(
             MRJobConfig.DEFAULT_MR_AM_MAX_ATTEMPTS));
     appContext.setResource(capability);
     appContext.setApplicationType(MRJobConfig.MR_APPLICATION_TYPE);
+    if (tagsFromConf != null && !tagsFromConf.isEmpty()) {
+      appContext.setApplicationTags(new HashSet<String>(tagsFromConf));
+    }
     return appContext;
   }
 

hadoop-project/pom.xml

@@ -784,6 +784,12 @@
         <artifactId>grizzly-http-servlet</artifactId>
         <version>2.1.2</version>
       </dependency>
+
+      <dependency>
+        <groupId>org.fusesource.leveldbjni</groupId>
+        <artifactId>leveldbjni-all</artifactId>
+        <version>1.8</version>
+      </dependency>
     </dependencies>
   </dependencyManagement>
 

hadoop-tools/hadoop-sls/src/main/java/org/apache/hadoop/yarn/sls/scheduler/ResourceSchedulerWrapper.java

@@ -54,6 +54,7 @@
 import org.apache.hadoop.yarn.api.records.QueueUserACLInfo;
 import org.apache.hadoop.yarn.api.records.Resource;
 import org.apache.hadoop.yarn.api.records.ResourceRequest;
+import org.apache.hadoop.yarn.exceptions.YarnException;
 import org.apache.hadoop.yarn.server.resourcemanager.RMContext;
 import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore;
 import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
@@ -865,4 +866,10 @@ public List<ApplicationAttemptId> getAppsInQueue(String queue) {
   public RMContainer getRMContainer(ContainerId containerId) {
     return null;
   }
+
+  @Override
+  public String moveApplication(ApplicationId appId, String newQueue)
+      throws YarnException {
+    return scheduler.moveApplication(appId, newQueue);
+  }
 }

hadoop-yarn-project/CHANGES.txt

@@ -12,6 +12,14 @@ Trunk - Unreleased
     YARN-1498. Common scheduler changes for moving apps between queues (Sandy
     Ryza)
 
+    YARN-1504. RM changes for moving apps between queues (Sandy Ryza)
+
+    YARN-1499. Fair Scheduler changes for moving apps between queues (Sandy
+    Ryza)
+
+    YARN-1497. Command line additions for moving apps between queues (Sandy
+    Ryza)
+
   IMPROVEMENTS
 
   OPTIMIZATIONS
@@ -77,6 +85,40 @@ Release 2.4.0 - UNRELEASED
     YARN-1413. Implemented serving of aggregated-logs in the ApplicationHistory
     server. (Mayank Bansal via vinodkv)
 
+    YARN-1633. Defined user-facing entity, entity-info and event objects related
+    to Application Timeline feature. (Zhijie Shen via vinodkv)
+
+    YARN-1611. Introduced the concept of a configuration provider which can be
+    used by ResourceManager to read configuration locally or from remote systems
+    so as to help RM failover. (Xuan Gong via vinodkv)
+
+    YARN-1659. Defined the ApplicationTimelineStore store as an abstraction for
+    implementing different storage impls for storing timeline information.
+    (Billie Rinaldi via vinodkv)
+
+    YARN-1634. Added a testable in-memory implementation of
+    ApplicationTimelineStore. (Zhijie Shen via vinodkv)
+
+    YARN-1461. Added tags for YARN applications and changed RM to handle them.
+    (Karthik Kambatla via zjshen)
+
+    YARN-1636. Augmented Application-history server's web-services to also expose
+    new APIs for retrieving and storing timeline information. (Zhijie Shen via
+    vinodkv)
+
+    YARN-1490. Introduced the ability to make ResourceManager optionally not kill
+    all containers when an ApplicationMaster exits. (Jian He via vinodkv)
+
+    YARN-1041. Added the ApplicationMasterProtocol API for applications to use the
+    ability in ResourceManager to optionally not kill containers when the
+    ApplicationMaster exits. (Jian He via vinodkv)
+
+    YARN-1566. Changed Distributed Shell to retain containers across application
+    attempts. (Jian He via vinodkv)
+
+    YARN-1635. Implemented a Leveldb based ApplicationTimelineStore. (Billie
+    Rinaldi via zjshen)
+
   IMPROVEMENTS
 
     YARN-1007. Enhance History Reader interface for Containers. (Mayank Bansal via
@@ -94,6 +136,36 @@ Release 2.4.0 - UNRELEASED
     YARN-1617. Remove ancient comment and surround LOG.debug in
     AppSchedulingInfo.allocate (Sandy Ryza)
 
+    YARN-1639. Modified RM HA configuration handling to have a way of not
+    requiring separate configuration files for each RM. (Xuan Gong via vinodkv)
+
+    YARN-1668. Modified RM HA handling of admin-acls to be available across RM
+    failover by making using of a remote configuration-provider. (Xuan Gong via
+    vinodkv)
+
+    YARN-1667. Modified RM HA handling of super users (with proxying ability) to
+    be available across RM failover by making using of a remote
+    configuration-provider. (Xuan Gong via vinodkv)
+
+    YARN-1285. Changed the default value of yarn.acl.enable in yarn-default.xml
+    to be consistent with what exists (false) in the code and documentation.
+    (Kenji Kikushima via vinodkv)
+
+    YARN-1669. Modified RM HA handling of protocol level service-ACLS to
+    be available across RM failover by making using of a remote
+    configuration-provider. (Xuan Gong via vinodkv)
+
+    YARN-1665. Simplify the configuration of RM HA by having better default
+    values. (Xuan Gong via vinodkv)
+
+    YARN-1660. Simplified the RM HA configuration to accept and be able to simply
+    depend just on configuration properties of the form
+    yarn.resourcemanager.hostname.RMID and use the default ports for all service
+    addresses. (Xuan Gong via vinodkv)
+
+    YARN-1493. Changed ResourceManager and Scheduler interfacing to recognize
+    app-attempts separately from apps. (Jian He via vinodkv)
+
   OPTIMIZATIONS
 
   BUG FIXES
@@ -143,6 +215,24 @@ Release 2.4.0 - UNRELEASED
     YARN-1632. TestApplicationMasterServices should be under
     org.apache.hadoop.yarn.server.resourcemanager package (Chen He via jeagles)
 
+    YARN-1673. Fix option parsing in YARN's application CLI after it is broken
+    by YARN-967. (Mayank Bansal via vinodkv)
+
+    YARN-1684. Fixed history server heap size in yarn script. (Billie Rinaldi
+    via zjshen)
+
+    YARN-1166. Fixed app-specific and attempt-specific QueueMetrics to be
+    triggered by accordingly app event and attempt event. 
+
+    YARN-1689. Made RMAppAttempt get killed when RMApp is at ACCEPTED. (Vinod
+    Kumar Vavilapalli via zjshen)
+
+    YARN-1661. Fixed DS ApplicationMaster to write the correct exit log. (Vinod
+    Kumar Vavilapalli via zjshen)
+
+    YARN-1672. YarnConfiguration is missing a default for 
+    yarn.nodemanager.log.retain-seconds (Naren Koneru via kasha)
+
 Release 2.3.0 - UNRELEASED
 
   INCOMPATIBLE CHANGES
@@ -186,18 +276,8 @@ Release 2.3.0 - UNRELEASED
     YARN-1029. Added embedded leader election in the ResourceManager. (Karthik
     Kambatla via vinodkv)
 
-    YARN-1490. Introduced the ability to make ResourceManager optionally not kill
-    all containers when an ApplicationMaster exits. (Jian He via vinodkv)
-
     YARN-1033. Expose RM active/standby state to Web UI and REST API (kasha)
 
-    YARN-1041. Added the ApplicationMasterProtocol API for applications to use the
-    ability in ResourceManager to optionally not kill containers when the
-    ApplicationMaster exits. (Jian He via vinodkv)
-
-    YARN-1566. Changed Distributed Shell to retain containers across application
-    attempts. (Jian He via vinodkv)
-
   IMPROVEMENTS
 
     YARN-305. Fair scheduler logs too many "Node offered to app" messages.
@@ -361,9 +441,6 @@ Release 2.3.0 - UNRELEASED
     YARN-1541. Changed ResourceManager to invalidate ApplicationMaster host/port
     information once an AM crashes. (Jian He via vinodkv)
 
-    YARN-1493. Changed ResourceManager and Scheduler interfacing to recognize
-    app-attempts separately from apps. (Jian He via vinodkv)
-
     YARN-1482. Modified WebApplicationProxy to make it work across ResourceManager
     fail-over. (Xuan Gong via vinodkv)
 
@@ -549,9 +626,6 @@ Release 2.3.0 - UNRELEASED
     YARN-1574. RMDispatcher should be reset on transition to standby. (Xuan Gong
     via kasha)
 
-    YARN-1166. Fixed app-specific and attempt-specific QueueMetrics to be
-    triggered by accordingly app event and attempt event. 
-
     YARN-1598. HA-related rmadmin commands don't work on a secure cluster (kasha)
 
     YARN-1603. Remove two *.orig files which were unexpectedly committed. 
@@ -576,6 +650,9 @@ Release 2.3.0 - UNRELEASED
 
     YARN-1629. IndexOutOfBoundsException in MaxRunningAppsEnforcer (Sandy Ryza)
 
+    YARN-1628. Fixed the test failure in TestContainerManagerSecurity. (Vinod
+    Kumar Vavilapalli via zjshen)
+
 Release 2.2.0 - 2013-10-13
 
   INCOMPATIBLE CHANGES

hadoop-yarn-project/hadoop-yarn/bin/yarn

@@ -204,7 +204,7 @@ elif [ "$COMMAND" = "historyserver" ] ; then
   CLASSPATH=${CLASSPATH}:$YARN_CONF_DIR/ahs-config/log4j.properties
   CLASS='org.apache.hadoop.yarn.server.applicationhistoryservice.ApplicationHistoryServer'
   YARN_OPTS="$YARN_OPTS $YARN_HISTORYSERVER_OPTS"
-  if [ "$YARN_RESOURCEMANAGER_HEAPSIZE" != "" ]; then
+  if [ "$YARN_HISTORYSERVER_HEAPSIZE" != "" ]; then
     JAVA_HEAP_MAX="-Xmx""$YARN_HISTORYSERVER_HEAPSIZE""m"
   fi  
 elif [ "$COMMAND" = "nodemanager" ] ; then

hadoop-yarn-project/hadoop-yarn/bin/yarn.cmd

@@ -207,7 +207,7 @@ goto :eof
   set CLASSPATH=%CLASSPATH%;%YARN_CONF_DIR%\ahs-config\log4j.properties
   set CLASS=org.apache.hadoop.yarn.server.applicationhistoryservice.ApplicationHistoryServer
   set YARN_OPTS=%YARN_OPTS% %HADOOP_HISTORYSERVER_OPTS%
-  if defined YARN_RESOURCEMANAGER_HEAPSIZE (
+  if defined YARN_HISTORYSERVER_HEAPSIZE (
     set JAVA_HEAP_MAX=-Xmx%YARN_HISTORYSERVER_HEAPSIZE%m
   )
   goto :eof

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/protocolrecords/ApplicationsRequestScope.java

@@ -0,0 +1,38 @@
+/**
+ * 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.yarn.api.protocolrecords;
+
+import org.apache.hadoop.classification.InterfaceAudience;
+import org.apache.hadoop.classification.InterfaceStability;
+
+/**
+ * Enumeration that controls the scope of applications fetched
+ */
+@InterfaceAudience.Public
+@InterfaceStability.Unstable
+public enum ApplicationsRequestScope {
+  /** All jobs */
+  ALL,
+
+  /** Jobs viewable by current user */
+  VIEWABLE,
+
+  /** Jobs owned by current user */
+  OWN
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/protocolrecords/GetApplicationsRequest.java

@@ -21,7 +21,6 @@
 import java.util.EnumSet;
 import java.util.Set;
 
-import org.apache.commons.collections.buffer.UnboundedFifoBuffer;
 import org.apache.commons.lang.math.LongRange;
 import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.classification.InterfaceAudience.Public;
@@ -49,6 +48,86 @@ public static GetApplicationsRequest newInstance() {
     return request;
   }
 
+  /**
+   * <p>
+   * The request from clients to get a report of Applications matching the
+   * giving application types in the cluster from the
+   * <code>ResourceManager</code>.
+   * </p>
+   *
+   * @see ApplicationClientProtocol#getApplications(GetApplicationsRequest)
+   *
+   * <p>Setting any of the parameters to null, would just disable that
+   * filter</p>
+   *
+   * @param scope {@link ApplicationsRequestScope} to filter by
+   * @param users list of users to filter by
+   * @param queues list of scheduler queues to filter by
+   * @param applicationTypes types of applications
+   * @param applicationTags application tags to filter by
+   * @param applicationStates application states to filter by
+   * @param startRange range of application start times to filter by
+   * @param finishRange range of application finish times to filter by
+   * @param limit number of applications to limit to
+   * @return {@link GetApplicationsRequest} to be used with
+   * {@link ApplicationClientProtocol#getApplications(GetApplicationsRequest)}
+   */
+  @Public
+  @Stable
+  public static GetApplicationsRequest newInstance(
+      ApplicationsRequestScope scope,
+      Set<String> users,
+      Set<String> queues,
+      Set<String> applicationTypes,
+      Set<String> applicationTags,
+      EnumSet<YarnApplicationState> applicationStates,
+      LongRange startRange,
+      LongRange finishRange,
+      Long limit) {
+    GetApplicationsRequest request =
+        Records.newRecord(GetApplicationsRequest.class);
+    if (scope != null) {
+      request.setScope(scope);
+    }
+    request.setUsers(users);
+    request.setQueues(queues);
+    request.setApplicationTypes(applicationTypes);
+    request.setApplicationTags(applicationTags);
+    request.setApplicationStates(applicationStates);
+    if (startRange != null) {
+      request.setStartRange(
+          startRange.getMinimumLong(), startRange.getMaximumLong());
+    }
+    if (finishRange != null) {
+      request.setFinishRange(
+          finishRange.getMinimumLong(), finishRange.getMaximumLong());
+    }
+    if (limit != null) {
+      request.setLimit(limit);
+    }
+    return request;
+  }
+
+  /**
+   * <p>
+   * The request from clients to get a report of Applications matching the
+   * giving application types in the cluster from the
+   * <code>ResourceManager</code>.
+   * </p>
+   *
+   * @param scope {@link ApplicationsRequestScope} to filter by
+   * @see ApplicationClientProtocol#getApplications(GetApplicationsRequest)
+   */
+  @Public
+  @Stable
+  public static GetApplicationsRequest newInstance(
+      ApplicationsRequestScope scope) {
+    GetApplicationsRequest request =
+        Records.newRecord(GetApplicationsRequest.class);
+    request.setScope(scope);
+    return request;
+  }
+
   /**
    * <p>
    * The request from clients to get a report of Applications matching the
@@ -257,4 +336,40 @@ public abstract void setStartRange(long begin, long end)
   @Private
   @Unstable
   public abstract void setFinishRange(long begin, long end);
+
+  /**
+   * Get the tags to filter applications on
+   *
+   * @return list of tags to filter on
+   */
+  @Private
+  @Unstable
+  public abstract Set<String> getApplicationTags();
+
+  /**
+   * Set the list of tags to filter applications on
+   *
+   * @param tags list of tags to filter on
+   */
+  @Private
+  @Unstable
+  public abstract void setApplicationTags(Set<String> tags);
+
+  /**
+   * Get the {@link ApplicationsRequestScope} of applications to be filtered.
+   *
+   * @return {@link ApplicationsRequestScope} of applications to return.
+   */
+  @Private
+  @Unstable
+  public abstract ApplicationsRequestScope getScope();
+
+  /**
+   * Set the {@link ApplicationsRequestScope} of applications to filter.
+   *
+   * @param scope scope to use for filtering applications
+   */
+  @Private
+  @Unstable
+  public abstract void setScope(ApplicationsRequestScope scope);
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/ApplicationReport.java

@@ -25,6 +25,8 @@
 import org.apache.hadoop.yarn.api.ApplicationClientProtocol;
 import org.apache.hadoop.yarn.util.Records;
 
+import java.util.Set;
+
 /**
  * <p><code>ApplicationReport</code> is a report of an application.</p>
  *
@@ -321,6 +323,18 @@ public static ApplicationReport newInstance(ApplicationId applicationId,
   @Unstable
   public abstract void setApplicationType(String applicationType);
 
+  /**
+   * Get all tags corresponding to the application
+   * @return Application's tags
+   */
+  @Public
+  @Stable
+  public abstract Set<String> getApplicationTags();
+
+  @Private
+  @Unstable
+  public abstract void setApplicationTags(Set<String> tags);
+
   @Private
   @Stable
   public abstract void setAMRMToken(Token amRmToken);

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/ApplicationSubmissionContext.java

@@ -25,8 +25,11 @@
 import org.apache.hadoop.classification.InterfaceStability.Unstable;
 import org.apache.hadoop.yarn.api.ApplicationClientProtocol;
 import org.apache.hadoop.yarn.api.ApplicationMasterProtocol;
+import org.apache.hadoop.yarn.conf.YarnConfiguration;
 import org.apache.hadoop.yarn.util.Records;
 
+import java.util.Set;
+
 /**
  * <p><code>ApplicationSubmissionContext</code> represents all of the
  * information needed by the <code>ResourceManager</code> to launch 
@@ -284,7 +287,6 @@ public static ApplicationSubmissionContext newInstance(
   @Stable
   public abstract void setApplicationType(String applicationType);
 
-
   /**
    * Get the flag which indicates whether to keep containers across application
    * attempts or not.
@@ -314,4 +316,26 @@ public static ApplicationSubmissionContext newInstance(
   @Stable
   public abstract void setKeepContainersAcrossApplicationAttempts(
       boolean keepContainers);
+
+  /**
+   * Get tags for the application
+   *
+   * @return the application tags
+   */
+  @Public
+  @Stable
+  public abstract Set<String> getApplicationTags();
+
+  /**
+   * Set tags for the application. A maximum of
+   * {@link YarnConfiguration#APPLICATION_MAX_TAGS} are allowed
+   * per application. Each tag can be at most
+   * {@link YarnConfiguration#APPLICATION_MAX_TAG_LENGTH}
+   * characters, and can contain only ASCII characters.
+   *
+   * @param tags tags to set
+   */
+  @Public
+  @Stable
+  public abstract void setApplicationTags(Set<String> tags);
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/ATSEntities.java

@@ -0,0 +1,88 @@
+/**
+ * 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.yarn.api.records.apptimeline;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlElement;
+import javax.xml.bind.annotation.XmlRootElement;
+
+import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+
+/**
+ * The class that hosts a list of application timeline entities.
+ */
+@XmlRootElement(name = "entities")
+@XmlAccessorType(XmlAccessType.NONE)
+@Public
+@Unstable
+public class ATSEntities {
+
+  private List<ATSEntity> entities =
+      new ArrayList<ATSEntity>();
+
+  public ATSEntities() {
+
+  }
+
+  /**
+   * Get a list of entities
+   * 
+   * @return a list of entities
+   */
+  @XmlElement(name = "entities")
+  public List<ATSEntity> getEntities() {
+    return entities;
+  }
+
+  /**
+   * Add a single entity into the existing entity list
+   * 
+   * @param entity
+   *          a single entity
+   */
+  public void addEntity(ATSEntity entity) {
+    entities.add(entity);
+  }
+
+  /**
+   * All a list of entities into the existing entity list
+   * 
+   * @param entities
+   *          a list of entities
+   */
+  public void addEntities(List<ATSEntity> entities) {
+    this.entities.addAll(entities);
+  }
+
+  /**
+   * Set the entity list to the given list of entities
+   * 
+   * @param entities
+   *          a list of entities
+   */
+  public void setEntities(List<ATSEntity> entities) {
+    this.entities = entities;
+  }
+
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/ATSEntity.java

@@ -0,0 +1,401 @@
+/**
+ * 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.yarn.api.records.apptimeline;
+
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Map.Entry;
+
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlElement;
+import javax.xml.bind.annotation.XmlRootElement;
+
+import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+
+/**
+ * <p>
+ * The class that contains the the meta information of some conceptual entity of
+ * an application and its related events. The entity can be an application, an
+ * application attempt, a container or whatever the user-defined object.
+ * </p>
+ * 
+ * <p>
+ * Primary filters will be used to index the entities in
+ * <code>ApplicationTimelineStore</code>, such that users should carefully
+ * choose the information they want to store as the primary filters. The
+ * remaining can be stored as other information.
+ * </p>
+ */
+@XmlRootElement(name = "entity")
+@XmlAccessorType(XmlAccessType.NONE)
+@Public
+@Unstable
+public class ATSEntity implements Comparable<ATSEntity> {
+
+  private String entityType;
+  private String entityId;
+  private Long startTime;
+  private List<ATSEvent> events = new ArrayList<ATSEvent>();
+  private Map<String, List<String>> relatedEntities =
+      new HashMap<String, List<String>>();
+  private Map<String, Object> primaryFilters =
+      new HashMap<String, Object>();
+  private Map<String, Object> otherInfo =
+      new HashMap<String, Object>();
+
+  public ATSEntity() {
+
+  }
+
+  /**
+   * Get the entity type
+   * 
+   * @return the entity type
+   */
+  @XmlElement(name = "entitytype")
+  public String getEntityType() {
+    return entityType;
+  }
+
+  /**
+   * Set the entity type
+   * 
+   * @param entityType
+   *          the entity type
+   */
+  public void setEntityType(String entityType) {
+    this.entityType = entityType;
+  }
+
+  /**
+   * Get the entity Id
+   * 
+   * @return the entity Id
+   */
+  @XmlElement(name = "entity")
+  public String getEntityId() {
+    return entityId;
+  }
+
+  /**
+   * Set the entity Id
+   * 
+   * @param entityId
+   *          the entity Id
+   */
+  public void setEntityId(String entityId) {
+    this.entityId = entityId;
+  }
+
+  /**
+   * Get the start time of the entity
+   * 
+   * @return the start time of the entity
+   */
+  @XmlElement(name = "starttime")
+  public Long getStartTime() {
+    return startTime;
+  }
+
+  /**
+   * Set the start time of the entity
+   * 
+   * @param startTime
+   *          the start time of the entity
+   */
+  public void setStartTime(Long startTime) {
+    this.startTime = startTime;
+  }
+
+  /**
+   * Get a list of events related to the entity
+   * 
+   * @return a list of events related to the entity
+   */
+  @XmlElement(name = "events")
+  public List<ATSEvent> getEvents() {
+    return events;
+  }
+
+  /**
+   * Add a single event related to the entity to the existing event list
+   * 
+   * @param event
+   *          a single event related to the entity
+   */
+  public void addEvent(ATSEvent event) {
+    events.add(event);
+  }
+
+  /**
+   * Add a list of events related to the entity to the existing event list
+   * 
+   * @param events
+   *          a list of events related to the entity
+   */
+  public void addEvents(List<ATSEvent> events) {
+    this.events.addAll(events);
+  }
+
+  /**
+   * Set the event list to the given list of events related to the entity
+   * 
+   * @param events
+   *          events a list of events related to the entity
+   */
+  public void setEvents(List<ATSEvent> events) {
+    this.events = events;
+  }
+
+  /**
+   * Get the related entities
+   * 
+   * @return the related entities
+   */
+  @XmlElement(name = "relatedentities")
+  public Map<String, List<String>> getRelatedEntities() {
+    return relatedEntities;
+  }
+
+  /**
+   * Add an entity to the existing related entity map
+   * 
+   * @param entityType
+   *          the entity type
+   * @param entityId
+   *          the entity Id
+   */
+  public void addRelatedEntity(String entityType, String entityId) {
+    List<String> thisRelatedEntity = relatedEntities.get(entityType);
+    if (thisRelatedEntity == null) {
+      thisRelatedEntity = new ArrayList<String>();
+      relatedEntities.put(entityType, thisRelatedEntity);
+    }
+    thisRelatedEntity.add(entityId);
+  }
+
+  /**
+   * Add a map of related entities to the existing related entity map
+   * 
+   * @param relatedEntities
+   *          a map of related entities
+   */
+  public void addRelatedEntities(Map<String, List<String>> relatedEntities) {
+    for (Entry<String, List<String>> relatedEntity :
+        relatedEntities.entrySet()) {
+      List<String> thisRelatedEntity =
+          this.relatedEntities.get(relatedEntity.getKey());
+      if (thisRelatedEntity == null) {
+        this.relatedEntities.put(
+            relatedEntity.getKey(), relatedEntity.getValue());
+      } else {
+        thisRelatedEntity.addAll(relatedEntity.getValue());
+      }
+    }
+  }
+
+  /**
+   * Set the related entity map to the given map of related entities
+   * 
+   * @param relatedEntities
+   *          a map of related entities
+   */
+  public void setRelatedEntities(
+      Map<String, List<String>> relatedEntities) {
+    this.relatedEntities = relatedEntities;
+  }
+
+  /**
+   * Get the primary filters
+   * 
+   * @return the primary filters
+   */
+  @XmlElement(name = "primaryfilters")
+  public Map<String, Object> getPrimaryFilters() {
+    return primaryFilters;
+  }
+
+  /**
+   * Add a single piece of primary filter to the existing primary filter map
+   * 
+   * @param key
+   *          the primary filter key
+   * @param value
+   *          the primary filter value
+   */
+  public void addPrimaryFilter(String key, Object value) {
+    primaryFilters.put(key, value);
+  }
+
+  /**
+   * Add a map of primary filters to the existing primary filter map
+   * 
+   * @param primaryFilters
+   *          a map of primary filters
+   */
+  public void addPrimaryFilters(Map<String, Object> primaryFilters) {
+    this.primaryFilters.putAll(primaryFilters);
+  }
+
+  /**
+   * Set the primary filter map to the given map of primary filters
+   * 
+   * @param primaryFilters
+   *          a map of primary filters
+   */
+  public void setPrimaryFilters(Map<String, Object> primaryFilters) {
+    this.primaryFilters = primaryFilters;
+  }
+
+  /**
+   * Get the other information of the entity
+   * 
+   * @return the other information of the entity
+   */
+  @XmlElement(name = "otherinfo")
+  public Map<String, Object> getOtherInfo() {
+    return otherInfo;
+  }
+
+  /**
+   * Add one piece of other information of the entity to the existing other info
+   * map
+   * 
+   * @param key
+   *          the other information key
+   * @param value
+   *          the other information value
+   */
+  public void addOtherInfo(String key, Object value) {
+    this.otherInfo.put(key, value);
+  }
+
+  /**
+   * Add a map of other information of the entity to the existing other info map
+   * 
+   * @param otherInfo
+   *          a map of other information
+   */
+  public void addOtherInfo(Map<String, Object> otherInfo) {
+    this.otherInfo.putAll(otherInfo);
+  }
+
+  /**
+   * Set the other info map to the given map of other information
+   * 
+   * @param otherInfo
+   *          a map of other information
+   */
+  public void setOtherInfo(Map<String, Object> otherInfo) {
+    this.otherInfo = otherInfo;
+  }
+
+  @Override
+  public int hashCode() {
+    // generated by eclipse
+    final int prime = 31;
+    int result = 1;
+    result = prime * result + ((entityId == null) ? 0 : entityId.hashCode());
+    result =
+        prime * result + ((entityType == null) ? 0 : entityType.hashCode());
+    result = prime * result + ((events == null) ? 0 : events.hashCode());
+    result = prime * result + ((otherInfo == null) ? 0 : otherInfo.hashCode());
+    result =
+        prime * result
+            + ((primaryFilters == null) ? 0 : primaryFilters.hashCode());
+    result =
+        prime * result
+            + ((relatedEntities == null) ? 0 : relatedEntities.hashCode());
+    result = prime * result + ((startTime == null) ? 0 : startTime.hashCode());
+    return result;
+  }
+
+  @Override
+  public boolean equals(Object obj) {
+    // generated by eclipse
+    if (this == obj)
+      return true;
+    if (obj == null)
+      return false;
+    if (getClass() != obj.getClass())
+      return false;
+    ATSEntity other = (ATSEntity) obj;
+    if (entityId == null) {
+      if (other.entityId != null)
+        return false;
+    } else if (!entityId.equals(other.entityId))
+      return false;
+    if (entityType == null) {
+      if (other.entityType != null)
+        return false;
+    } else if (!entityType.equals(other.entityType))
+      return false;
+    if (events == null) {
+      if (other.events != null)
+        return false;
+    } else if (!events.equals(other.events))
+      return false;
+    if (otherInfo == null) {
+      if (other.otherInfo != null)
+        return false;
+    } else if (!otherInfo.equals(other.otherInfo))
+      return false;
+    if (primaryFilters == null) {
+      if (other.primaryFilters != null)
+        return false;
+    } else if (!primaryFilters.equals(other.primaryFilters))
+      return false;
+    if (relatedEntities == null) {
+      if (other.relatedEntities != null)
+        return false;
+    } else if (!relatedEntities.equals(other.relatedEntities))
+      return false;
+    if (startTime == null) {
+      if (other.startTime != null)
+        return false;
+    } else if (!startTime.equals(other.startTime))
+      return false;
+    return true;
+  }
+
+  @Override
+  public int compareTo(ATSEntity other) {
+    int comparison = entityType.compareTo(other.entityType);
+    if (comparison == 0) {
+      long thisStartTime =
+          startTime == null ? Long.MIN_VALUE : startTime;
+      long otherStartTime =
+          other.startTime == null ? Long.MIN_VALUE : other.startTime;
+      if (thisStartTime > otherStartTime) {
+        return -1;
+      } else if (thisStartTime < otherStartTime) {
+        return 1;
+      } else {
+        return entityId.compareTo(other.entityId);
+      }
+    } else {
+      return comparison;
+    }
+  }
+
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/ATSEvent.java

@@ -0,0 +1,172 @@
+/**
+ * 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.yarn.api.records.apptimeline;
+
+import java.util.HashMap;
+import java.util.Map;
+
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlElement;
+import javax.xml.bind.annotation.XmlRootElement;
+
+import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+
+/**
+ * The class that contains the information of an event that is related to some
+ * conceptual entity of an application. Users are free to define what the event
+ * means, such as starting an application, getting allocated a container and
+ * etc.
+ */
+@XmlRootElement(name = "event")
+@XmlAccessorType(XmlAccessType.NONE)
+@Public
+@Unstable
+public class ATSEvent implements Comparable<ATSEvent> {
+
+  private long timestamp;
+  private String eventType;
+  private Map<String, Object> eventInfo = new HashMap<String, Object>();
+
+  public ATSEvent() {
+  }
+
+  /**
+   * Get the timestamp of the event
+   * 
+   * @return the timestamp of the event
+   */
+  @XmlElement(name = "timestamp")
+  public long getTimestamp() {
+    return timestamp;
+  }
+
+  /**
+   * Set the timestamp of the event
+   * 
+   * @param timestamp
+   *          the timestamp of the event
+   */
+  public void setTimestamp(long timestamp) {
+    this.timestamp = timestamp;
+  }
+
+  /**
+   * Get the event type
+   * 
+   * @return the event type
+   */
+  @XmlElement(name = "eventtype")
+  public String getEventType() {
+    return eventType;
+  }
+
+  /**
+   * Set the event type
+   * 
+   * @param eventType
+   *          the event type
+   */
+  public void setEventType(String eventType) {
+    this.eventType = eventType;
+  }
+
+  /**
+   * Set the information of the event
+   * 
+   * @return the information of the event
+   */
+  @XmlElement(name = "eventinfo")
+  public Map<String, Object> getEventInfo() {
+    return eventInfo;
+  }
+
+  /**
+   * Add one piece of the information of the event to the existing information
+   * map
+   * 
+   * @param key
+   *          the information key
+   * @param value
+   *          the information value
+   */
+  public void addEventInfo(String key, Object value) {
+    this.eventInfo.put(key, value);
+  }
+
+  /**
+   * Add a map of the information of the event to the existing information map
+   * 
+   * @param eventInfo
+   *          a map of of the information of the event
+   */
+  public void addEventInfo(Map<String, Object> eventInfo) {
+    this.eventInfo.putAll(eventInfo);
+  }
+
+  /**
+   * Set the information map to the given map of the information of the event
+   * 
+   * @param eventInfo
+   *          a map of of the information of the event
+   */
+  public void setEventInfo(Map<String, Object> eventInfo) {
+    this.eventInfo = eventInfo;
+  }
+
+  @Override
+  public int compareTo(ATSEvent other) {
+    if (timestamp > other.timestamp) {
+      return -1;
+    } else if (timestamp < other.timestamp) {
+      return 1;
+    } else {
+      return eventType.compareTo(other.eventType);
+    }
+  }
+
+  @Override
+  public boolean equals(Object o) {
+    if (this == o)
+      return true;
+    if (o == null || getClass() != o.getClass())
+      return false;
+
+    ATSEvent atsEvent = (ATSEvent) o;
+
+    if (timestamp != atsEvent.timestamp)
+      return false;
+    if (!eventType.equals(atsEvent.eventType))
+      return false;
+    if (eventInfo != null ? !eventInfo.equals(atsEvent.eventInfo) :
+        atsEvent.eventInfo != null)
+      return false;
+
+    return true;
+  }
+
+  @Override
+  public int hashCode() {
+    int result = (int) (timestamp ^ (timestamp >>> 32));
+    result = 31 * result + eventType.hashCode();
+    result = 31 * result + (eventInfo != null ? eventInfo.hashCode() : 0);
+    return result;
+  }
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/ATSEvents.java

@@ -0,0 +1,189 @@
+/**
+ * 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.yarn.api.records.apptimeline;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlElement;
+import javax.xml.bind.annotation.XmlRootElement;
+
+import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+
+/**
+ * The class that hosts a list of events, which are categorized according to
+ * their related entities.
+ */
+@XmlRootElement(name = "events")
+@XmlAccessorType(XmlAccessType.NONE)
+@Public
+@Unstable
+public class ATSEvents {
+
+  private List<ATSEventsOfOneEntity> allEvents =
+      new ArrayList<ATSEventsOfOneEntity>();
+
+  public ATSEvents() {
+
+  }
+
+  /**
+   * Get a list of {@link ATSEventsOfOneEntity} instances
+   * 
+   * @return a list of {@link ATSEventsOfOneEntity} instances
+   */
+  @XmlElement(name = "events")
+  public List<ATSEventsOfOneEntity> getAllEvents() {
+    return allEvents;
+  }
+
+  /**
+   * Add a single {@link ATSEventsOfOneEntity} instance into the existing list
+   * 
+   * @param eventsOfOneEntity
+   *          a single {@link ATSEventsOfOneEntity} instance
+   */
+  public void addEvent(ATSEventsOfOneEntity eventsOfOneEntity) {
+    allEvents.add(eventsOfOneEntity);
+  }
+
+  /**
+   * Add a list of {@link ATSEventsOfOneEntity} instances into the existing list
+   * 
+   * @param allEvents
+   *          a list of {@link ATSEventsOfOneEntity} instances
+   */
+  public void addEvents(List<ATSEventsOfOneEntity> allEvents) {
+    this.allEvents.addAll(allEvents);
+  }
+
+  /**
+   * Set the list to the given list of {@link ATSEventsOfOneEntity} instances
+   * 
+   * @param allEvents
+   *          a list of {@link ATSEventsOfOneEntity} instances
+   */
+  public void setEvents(List<ATSEventsOfOneEntity> allEvents) {
+    this.allEvents.clear();
+    this.allEvents.addAll(allEvents);
+  }
+
+  /**
+   * The class that hosts a list of events that are only related to one entity.
+   */
+  @XmlRootElement(name = "events")
+  @XmlAccessorType(XmlAccessType.NONE)
+  @Public
+  @Unstable
+  public static class ATSEventsOfOneEntity {
+
+    private String entityId;
+    private String entityType;
+    private List<ATSEvent> events = new ArrayList<ATSEvent>();
+
+    public ATSEventsOfOneEntity() {
+
+    }
+
+    /**
+     * Get the entity Id
+     * 
+     * @return the entity Id
+     */
+    @XmlElement(name = "entity")
+    public String getEntityId() {
+      return entityId;
+    }
+
+    /**
+     * Set the entity Id
+     * 
+     * @param entityId
+     *          the entity Id
+     */
+    public void setEntityId(String entityId) {
+      this.entityId = entityId;
+    }
+
+    /**
+     * Get the entity type
+     * 
+     * @return the entity type
+     */
+    @XmlElement(name = "entitytype")
+    public String getEntityType() {
+      return entityType;
+    }
+
+    /**
+     * Set the entity type
+     * 
+     * @param entityType
+     *          the entity type
+     */
+    public void setEntityType(String entityType) {
+      this.entityType = entityType;
+    }
+
+    /**
+     * Get a list of events
+     * 
+     * @return a list of events
+     */
+    @XmlElement(name = "events")
+    public List<ATSEvent> getEvents() {
+      return events;
+    }
+
+    /**
+     * Add a single event to the existing event list
+     * 
+     * @param event
+     *          a single event
+     */
+    public void addEvent(ATSEvent event) {
+      events.add(event);
+    }
+
+    /**
+     * Add a list of event to the existing event list
+     * 
+     * @param events
+     *          a list of events
+     */
+    public void addEvents(List<ATSEvent> events) {
+      this.events.addAll(events);
+    }
+
+    /**
+     * Set the event list to the given list of events
+     * 
+     * @param events
+     *          a list of events
+     */
+    public void setEvents(List<ATSEvent> events) {
+      this.events = events;
+    }
+
+  }
+
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/ATSPutErrors.java

@@ -0,0 +1,175 @@
+/**
+ * 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.yarn.api.records.apptimeline;
+
+import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+
+import javax.xml.bind.annotation.XmlAccessType;
+import javax.xml.bind.annotation.XmlAccessorType;
+import javax.xml.bind.annotation.XmlElement;
+import javax.xml.bind.annotation.XmlRootElement;
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * A class that holds a list of put errors.  This is the response returned
+ * when a list of {@link ATSEntity} objects is added to the application
+ * timeline.  If there are errors in storing individual entity objects,
+ * they will be indicated in the list of errors.
+ */
+@XmlRootElement(name = "errors")
+@XmlAccessorType(XmlAccessType.NONE)
+@Public
+@Unstable
+public class ATSPutErrors {
+
+  private List<ATSPutError> errors = new ArrayList<ATSPutError>();
+
+  public ATSPutErrors() {
+
+  }
+
+  /**
+   * Get a list of {@link ATSPutError} instances
+   *
+   * @return a list of {@link ATSPutError} instances
+   */
+  @XmlElement(name = "errors")
+  public List<ATSPutError> getErrors() {
+    return errors;
+  }
+
+  /**
+   * Add a single {@link ATSPutError} instance into the existing list
+   *
+   * @param error
+   *          a single {@link ATSPutError} instance
+   */
+  public void addError(ATSPutError error) {
+    errors.add(error);
+  }
+
+  /**
+   * Add a list of {@link ATSPutError} instances into the existing list
+   *
+   * @param errors
+   *          a list of {@link ATSPutError} instances
+   */
+  public void addErrors(List<ATSPutError> errors) {
+    this.errors.addAll(errors);
+  }
+
+  /**
+   * Set the list to the given list of {@link ATSPutError} instances
+   *
+   * @param errors
+   *          a list of {@link ATSPutError} instances
+   */
+  public void setErrors(List<ATSPutError> errors) {
+    this.errors.clear();
+    this.errors.addAll(errors);
+  }
+
+  /**
+   * A class that holds the error code for one entity.
+   */
+  @XmlRootElement(name = "error")
+  @XmlAccessorType(XmlAccessType.NONE)
+  @Public
+  @Unstable
+  public static class ATSPutError {
+    /**
+     * Error code returned when no start time can be found when putting an
+     * entity. This occurs when the entity does not already exist in the
+     * store and it is put with no start time or events specified.
+     */
+    public static final int NO_START_TIME = 1;
+    /**
+     * Error code returned if an IOException is encountered when putting an
+     * entity.
+     */
+    public static final int IO_EXCEPTION = 2;
+
+    private String entityId;
+    private String entityType;
+    private int errorCode;
+
+    /**
+     * Get the entity Id
+     *
+     * @return the entity Id
+     */
+    @XmlElement(name = "entity")
+    public String getEntityId() {
+      return entityId;
+    }
+
+    /**
+     * Set the entity Id
+     *
+     * @param entityId
+     *          the entity Id
+     */
+    public void setEntityId(String entityId) {
+      this.entityId = entityId;
+    }
+
+    /**
+     * Get the entity type
+     *
+     * @return the entity type
+     */
+    @XmlElement(name = "entitytype")
+    public String getEntityType() {
+      return entityType;
+    }
+
+    /**
+     * Set the entity type
+     *
+     * @param entityType
+     *          the entity type
+     */
+    public void setEntityType(String entityType) {
+      this.entityType = entityType;
+    }
+
+    /**
+     * Get the error code
+     *
+     * @return an error code
+     */
+    @XmlElement(name = "errorcode")
+    public int getErrorCode() {
+      return errorCode;
+    }
+
+    /**
+     * Set the error code to the given error code
+     *
+     * @param errorCode
+     *          an error code
+     */
+    public void setErrorCode(int errorCode) {
+      this.errorCode = errorCode;
+    }
+
+  }
+
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/api/records/apptimeline/package-info.java

@@ -0,0 +1,21 @@
+/**
+ * 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.
+ */
+@InterfaceAudience.Public
+package org.apache.hadoop.yarn.api.records.apptimeline;
+import org.apache.hadoop.classification.InterfaceAudience;
+

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/ConfigurationProvider.java

@@ -0,0 +1,64 @@
+/**
+ * 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.yarn.conf;
+
+import java.io.IOException;
+
+import org.apache.hadoop.classification.InterfaceAudience.Private;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.yarn.exceptions.YarnException;
+
+@Private
+@Unstable
+/**
+ * Base class to implement ConfigurationProvider.
+ * Real ConfigurationProvider implementations need to derive from it and
+ * implement load methods to actually load the configuration.
+ */
+public abstract class ConfigurationProvider {
+
+  public void init(Configuration conf) throws Exception {
+    initInternal(conf);
+  }
+
+  public void close() throws Exception {
+    closeInternal();
+  }
+
+  /**
+   * Get the configuration.
+   * @param name The configuration file name
+   * @return configuration
+   * @throws YarnException
+   * @throws IOException
+   */
+  public abstract Configuration getConfiguration(String name)
+      throws YarnException, IOException;
+
+  /**
+   * Derived classes initialize themselves using this method.
+   */
+  public abstract void initInternal(Configuration conf) throws Exception;
+
+  /**
+   * Derived classes close themselves using this method.
+   */
+  public abstract void closeInternal() throws Exception;
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/ConfigurationProviderFactory.java

@@ -0,0 +1,57 @@
+/**
+ * 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.yarn.conf;
+
+import org.apache.hadoop.classification.InterfaceAudience.Private;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.util.ReflectionUtils;
+import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
+
+@Private
+@Unstable
+/**
+ * Factory for {@link ConfigurationProvider} implementations.
+ */
+public class ConfigurationProviderFactory {
+  /**
+   * Creates an instance of {@link ConfigurationProvider} using given
+   * configuration.
+   * @param conf
+   * @return configurationProvider
+   */
+  @SuppressWarnings("unchecked")
+  public static ConfigurationProvider
+      getConfigurationProvider(Configuration conf) {
+    Class<? extends ConfigurationProvider> defaultProviderClass;
+    try {
+      defaultProviderClass = (Class<? extends ConfigurationProvider>)
+          Class.forName(
+              YarnConfiguration.DEFAULT_RM_CONFIGURATION_PROVIDER_CLASS);
+    } catch (Exception e) {
+      throw new YarnRuntimeException(
+          "Invalid default configuration provider class"
+              + YarnConfiguration.DEFAULT_RM_CONFIGURATION_PROVIDER_CLASS, e);
+    }
+    ConfigurationProvider configurationProvider = ReflectionUtils.newInstance(
+        conf.getClass(YarnConfiguration.RM_CONFIGURATION_PROVIDER_CLASS,
+            defaultProviderClass, ConfigurationProvider.class), conf);
+    return configurationProvider;
+  }
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/HAUtil.java

@@ -21,10 +21,13 @@
 import com.google.common.annotations.VisibleForTesting;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.HadoopIllegalArgumentException;
 import org.apache.hadoop.classification.InterfaceAudience;
 import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.net.NetUtils;
 import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
 
+import java.net.InetSocketAddress;
 import java.util.Collection;
 
 @InterfaceAudience.Private
@@ -98,21 +101,7 @@ private static void verifyAndSetRMHAIdsList(Configuration conf) {
     for (String id: ids) {
       // verify the RM service addresses configurations for every RMIds
       for (String prefix : YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS) {
-        String confKey = null;
-        try {
-          confKey = addSuffix(prefix, id);
-          if (conf.getTrimmed(confKey) == null) {
-            throwBadConfigurationException(getNeedToSetValueMessage(confKey));
-          }
-        } catch (IllegalArgumentException iae) {
-          String errmsg = iae.getMessage();
-          if (confKey == null) {
-            // Error at addSuffix
-            errmsg = getInvalidValueMessage(YarnConfiguration.RM_HA_ID,
-              getRMHAId(conf));
-          }
-          throwBadConfigurationException(errmsg);
-        }
+        checkAndSetRMRPCAddress(prefix, id, conf);
       }
       setValue.append(id);
       setValue.append(",");
@@ -122,10 +111,18 @@ private static void verifyAndSetRMHAIdsList(Configuration conf) {
   }
 
   private static void verifyAndSetCurrentRMHAId(Configuration conf) {
-    String rmId = conf.getTrimmed(YarnConfiguration.RM_HA_ID);
+    String rmId = getRMHAId(conf);
     if (rmId == null) {
-      throwBadConfigurationException(
-        getNeedToSetValueMessage(YarnConfiguration.RM_HA_ID));
+      StringBuilder msg = new StringBuilder();
+      msg.append("Can not find valid RM_HA_ID. None of ");
+      for (String id : conf
+          .getTrimmedStringCollection(YarnConfiguration.RM_HA_IDS)) {
+        msg.append(addSuffix(YarnConfiguration.RM_ADDRESS, id) + " ");
+      }
+      msg.append(" are matching" +
+          " the local address OR " + YarnConfiguration.RM_HA_ID + " is not" +
+          " specified in HA Configuration");
+      throwBadConfigurationException(msg.toString());
     } else {
       Collection<String> ids = getRMHAIds(conf);
       if (!ids.contains(rmId)) {
@@ -179,7 +176,34 @@ public static Collection<String> getRMHAIds(Configuration conf) {
    * @return RM Id on success
    */
   public static String getRMHAId(Configuration conf) {
-    return conf.get(YarnConfiguration.RM_HA_ID);
+    int found = 0;
+    String currentRMId = conf.getTrimmed(YarnConfiguration.RM_HA_ID);
+    if(currentRMId == null) {
+      for(String rmId : getRMHAIds(conf)) {
+        String key = addSuffix(YarnConfiguration.RM_ADDRESS, rmId);
+        String addr = conf.get(key);
+        if (addr == null) {
+          continue;
+        }
+        InetSocketAddress s;
+        try {
+          s = NetUtils.createSocketAddr(addr);
+        } catch (Exception e) {
+          LOG.warn("Exception in creating socket address " + addr, e);
+          continue;
+        }
+        if (!s.isUnresolved() && NetUtils.isLocalAddress(s.getAddress())) {
+          currentRMId = rmId.trim();
+          found++;
+        }
+      }
+    }
+    if (found > 1) { // Only one address must match the local address
+      String msg = "The HA Configuration has multiple addresses that match "
+          + "local node's address.";
+      throw new HadoopIllegalArgumentException(msg);
+    }
+    return currentRMId;
   }
 
   @VisibleForTesting
@@ -212,9 +236,13 @@ static String getRMHAIdsWarningMessage(String ids) {
   @InterfaceAudience.Private
   @VisibleForTesting
   static String getConfKeyForRMInstance(String prefix, Configuration conf) {
-    return YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS.contains(prefix)
-        ? addSuffix(prefix, getRMHAId(conf))
-        : prefix;
+    if (!YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS.contains(prefix)) {
+      return prefix;
+    } else {
+      String RMId = getRMHAId(conf);
+      checkAndSetRMRPCAddress(prefix, RMId, conf);
+      return addSuffix(prefix, RMId);
+    }
   }
 
   public static String getConfValueForRMInstance(String prefix,
@@ -247,4 +275,30 @@ public static String addSuffix(String key, String suffix) {
     }
     return key + "." + suffix;
   }
+
+  private static void checkAndSetRMRPCAddress(String prefix, String RMId,
+      Configuration conf) {
+    String rpcAddressConfKey = null;
+    try {
+      rpcAddressConfKey = addSuffix(prefix, RMId);
+      if (conf.getTrimmed(rpcAddressConfKey) == null) {
+        String hostNameConfKey = addSuffix(YarnConfiguration.RM_HOSTNAME, RMId);
+        String confVal = conf.getTrimmed(hostNameConfKey);
+        if (confVal == null) {
+          throwBadConfigurationException(getNeedToSetValueMessage(
+              hostNameConfKey + " or " + addSuffix(prefix, RMId)));
+        } else {
+          conf.set(addSuffix(prefix, RMId), confVal + ":"
+              + YarnConfiguration.getRMDefaultPortNumber(prefix));
+        }
+      }
+    } catch (IllegalArgumentException iae) {
+      String errmsg = iae.getMessage();
+      if (rpcAddressConfKey == null) {
+        // Error at addSuffix
+        errmsg = getInvalidValueMessage(YarnConfiguration.RM_HA_ID, RMId);
+      }
+      throwBadConfigurationException(errmsg);
+    }
+  }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java

@@ -23,8 +23,10 @@
 import java.util.Collections;
 import java.util.List;
 
+import org.apache.hadoop.HadoopIllegalArgumentException;
 import org.apache.hadoop.classification.InterfaceAudience.Private;
 import org.apache.hadoop.classification.InterfaceAudience.Public;
+import org.apache.hadoop.classification.InterfaceStability;
 import org.apache.hadoop.classification.InterfaceStability.Evolving;
 import org.apache.hadoop.conf.Configuration;
 import org.apache.hadoop.fs.CommonConfigurationKeysPublic;
@@ -37,8 +39,26 @@
 @Evolving
 public class YarnConfiguration extends Configuration {
 
+  @Private
+  public static final String CS_CONFIGURATION_FILE= "capacity-scheduler.xml";
+
+  @Private
+  public static final String HADOOP_POLICY_CONFIGURATION_FILE =
+      "hadoop-policy.xml";
+
+  @Private
+  public static final String YARN_SITE_XML_FILE = "yarn-site.xml";
+
+  @Private
+  public static final String CORE_SITE_CONFIGURATION_FILE = "core-site.xml";
+
+  @Evolving
+  public static final int APPLICATION_MAX_TAGS = 10;
+
+  @Evolving
+  public static final int APPLICATION_MAX_TAG_LENGTH = 100;
+
   private static final String YARN_DEFAULT_XML_FILE = "yarn-default.xml";
-  private static final String YARN_SITE_XML_FILE = "yarn-site.xml";
 
   static {
     Configuration.addDefaultResource(YARN_DEFAULT_XML_FILE);
@@ -89,6 +109,8 @@ public class YarnConfiguration extends Configuration {
 
   public static final String RM_CLUSTER_ID = RM_PREFIX + "cluster-id";
 
+  public static final String RM_HOSTNAME = RM_PREFIX + "hostname";
+
   /** The address of the applications manager interface in the RM.*/
   public static final String RM_ADDRESS = 
     RM_PREFIX + "address";
@@ -329,6 +351,16 @@ public class YarnConfiguration extends Configuration {
   public static final String RM_HA_IDS = RM_HA_PREFIX + "rm-ids";
   public static final String RM_HA_ID = RM_HA_PREFIX + "id";
 
+  /** Store the related configuration files in File System */
+  public static final String FS_BASED_RM_CONF_STORE = RM_PREFIX
+      + "configuration.file-system-based-store";
+  public static final String DEFAULT_FS_BASED_RM_CONF_STORE = "/yarn/conf";
+
+  public static final String RM_CONFIGURATION_PROVIDER_CLASS = RM_PREFIX
+      + "configuration.provider-class";
+  public static final String DEFAULT_RM_CONFIGURATION_PROVIDER_CLASS =
+      "org.apache.hadoop.yarn.LocalConfigurationProvider";
+
   @Private
   public static final List<String> RM_SERVICES_ADDRESS_CONF_KEYS =
       Collections.unmodifiableList(Arrays.asList(
@@ -344,11 +376,11 @@ public class YarnConfiguration extends Configuration {
 
   public static final String AUTO_FAILOVER_ENABLED =
       AUTO_FAILOVER_PREFIX + "enabled";
-  public static final boolean DEFAULT_AUTO_FAILOVER_ENABLED = false;
+  public static final boolean DEFAULT_AUTO_FAILOVER_ENABLED = true;
 
   public static final String AUTO_FAILOVER_EMBEDDED =
       AUTO_FAILOVER_PREFIX + "embedded";
-  public static final boolean DEFAULT_AUTO_FAILOVER_EMBEDDED = false;
+  public static final boolean DEFAULT_AUTO_FAILOVER_EMBEDDED = true;
 
   public static final String AUTO_FAILOVER_ZK_BASE_PATH =
       AUTO_FAILOVER_PREFIX + "zk-base-path";
@@ -597,6 +629,7 @@ public class YarnConfiguration extends Configuration {
    */
   public static final String NM_LOG_RETAIN_SECONDS = NM_PREFIX
       + "log.retain-seconds";
+  public static final long DEFAULT_NM_LOG_RETAIN_SECONDS = 3 * 60 * 60;
 
   /**
    * Number of threads used in log cleanup. Only applicable if Log aggregation
@@ -999,6 +1032,19 @@ public class YarnConfiguration extends Configuration {
   public static final String AHS_WEBAPP_SPNEGO_KEYTAB_FILE_KEY =
       AHS_PREFIX + "webapp.spnego-keytab-file";
 
+  ////////////////////////////////
+  // ATS Configs
+  ////////////////////////////////
+
+  public static final String ATS_PREFIX = YARN_PREFIX + "ats.";
+
+  /** ATS store class */
+  public static final String ATS_STORE = ATS_PREFIX + "store.class";
+
+  /** ATS leveldb path */
+  public static final String ATS_LEVELDB_PATH_PROPERTY =
+      ATS_PREFIX + "leveldb-apptimeline-store.path";
+
   ////////////////////////////////
   // Other Configs
   ////////////////////////////////
@@ -1101,4 +1147,27 @@ public InetSocketAddress updateConnectAddr(String name,
     }
     return super.updateConnectAddr(prefix, addr);
   }
+
+  @Private
+  public static int getRMDefaultPortNumber(String addressPrefix) {
+    if (addressPrefix.equals(YarnConfiguration.RM_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_PORT;
+    } else if (addressPrefix.equals(YarnConfiguration.RM_SCHEDULER_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_SCHEDULER_PORT;
+    } else if (addressPrefix.equals(YarnConfiguration.RM_WEBAPP_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_WEBAPP_PORT;
+    } else if (addressPrefix.equals(YarnConfiguration.RM_WEBAPP_HTTPS_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_WEBAPP_HTTPS_PORT;
+    } else if (addressPrefix
+        .equals(YarnConfiguration.RM_RESOURCE_TRACKER_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_RESOURCE_TRACKER_PORT;
+    } else if (addressPrefix.equals(YarnConfiguration.RM_ADMIN_ADDRESS)) {
+      return YarnConfiguration.DEFAULT_RM_ADMIN_PORT;
+    } else {
+      throw new HadoopIllegalArgumentException(
+          "Invalid RM RPC address Prefix: " + addressPrefix
+              + ". The valid value should be one of "
+              + YarnConfiguration.RM_SERVICES_ADDRESS_CONF_KEYS);
+    }
+  }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/proto/yarn_protos.proto

@@ -190,6 +190,7 @@ message ApplicationReportProto {
   optional float progress = 17;
   optional string applicationType = 18;
   optional hadoop.common.TokenProto am_rm_token = 19;
+  repeated string applicationTags = 20;
 }
 
 message ApplicationAttemptReportProto {
@@ -287,6 +288,7 @@ message ApplicationSubmissionContextProto {
   optional ResourceProto resource = 9;
   optional string applicationType = 10 [default = "YARN"];
   optional bool keep_containers_across_application_attempts = 11 [default = false];
+  repeated string applicationTags = 12;
 }
 
 enum ApplicationAccessTypeProto {

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/proto/yarn_service_protos.proto

@@ -136,6 +136,12 @@ message MoveApplicationAcrossQueuesRequestProto {
 message MoveApplicationAcrossQueuesResponseProto {
 }
 
+enum ApplicationsRequestScopeProto {
+  ALL = 0;
+  VIEWABLE = 1;
+  OWN = 2;
+}
+
 message GetApplicationsRequestProto {
   repeated string application_types = 1;
   repeated YarnApplicationStateProto application_states = 2;
@@ -146,6 +152,8 @@ message GetApplicationsRequestProto {
   optional int64 start_end = 7;
   optional int64 finish_begin = 8;
   optional int64 finish_end = 9;
+  repeated string applicationTags = 10;
+  optional ApplicationsRequestScopeProto scope = 11 [default = ALL];
 }
 
 message GetApplicationsResponseProto {

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-distributedshell/src/main/java/org/apache/hadoop/yarn/applications/distributedshell/ApplicationMaster.java

@@ -232,7 +232,6 @@ public class ApplicationMaster {
   private static final String shellArgsPath = "shellArgs";
 
   private volatile boolean done;
-  private volatile boolean success;
 
   private ByteBuffer allTokens;
 
@@ -254,8 +253,8 @@ public static void main(String[] args) {
       if (!doRun) {
         System.exit(0);
       }
-      result = appMaster.run();
-      appMaster.finish();
+      appMaster.run();
+      result = appMaster.finish();
     } catch (Throwable t) {
       LOG.fatal("Error running ApplicationMaster", t);
       System.exit(1);
@@ -480,7 +479,7 @@ private void printUsage(Options opts) {
    * @throws IOException
    */
   @SuppressWarnings({ "unchecked" })
-  public boolean run() throws YarnException, IOException {
+  public void run() throws YarnException, IOException {
     LOG.info("Starting ApplicationMaster");
 
     Credentials credentials =
@@ -561,7 +560,6 @@ public boolean run() throws YarnException, IOException {
       amRMClient.addContainerRequest(containerAsk);
     }
     numRequestedContainers.set(numTotalContainersToRequest);
-    return success;
   }
 
   @VisibleForTesting
@@ -569,7 +567,8 @@ NMCallbackHandler createNMCallbackHandler() {
     return new NMCallbackHandler(this);
   }
 
-  protected void finish() {
+  @VisibleForTesting
+  protected boolean finish() {
     // wait for completion.
     while (!done
         && (numCompletedContainers.get() != numTotalContainers)) {
@@ -600,7 +599,7 @@ protected void finish() {
 
     FinalApplicationStatus appStatus;
     String appMessage = null;
-    success = true;
+    boolean success = true;
     if (numFailedContainers.get() == 0 && 
         numCompletedContainers.get() == numTotalContainers) {
       appStatus = FinalApplicationStatus.SUCCEEDED;
@@ -621,6 +620,8 @@ protected void finish() {
     }
     
     amRMClient.stop();
+
+    return success;
   }
   
   private class RMCallbackHandler implements AMRMClientAsync.CallbackHandler {

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-distributedshell/src/test/java/org/apache/hadoop/yarn/applications/distributedshell/ContainerLaunchFailAppMaster.java

@@ -18,13 +18,13 @@
 
 package org.apache.hadoop.yarn.applications.distributedshell;
 
+import java.nio.ByteBuffer;
+import java.util.Map;
+
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.hadoop.yarn.api.records.ContainerId;
 
-import java.nio.ByteBuffer;
-import java.util.Map;
-
 public class ContainerLaunchFailAppMaster extends ApplicationMaster {
 
   private static final Log LOG =
@@ -66,8 +66,8 @@ public static void main(String[] args) {
       if (!doRun) {
         System.exit(0);
       }
-      result = appMaster.run();
-      appMaster.finish();
+      appMaster.run();
+      result = appMaster.finish();
     } catch (Throwable t) {
       LOG.fatal("Error running ApplicationMaster", t);
       System.exit(1);

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-distributedshell/src/test/java/org/apache/hadoop/yarn/applications/distributedshell/TestDSFailedAppMaster.java

@@ -29,8 +29,8 @@ public class TestDSFailedAppMaster extends ApplicationMaster {
   private static final Log LOG = LogFactory.getLog(TestDSFailedAppMaster.class);
 
   @Override
-  public boolean run() throws YarnException, IOException {
-    boolean res = super.run();
+  public void run() throws YarnException, IOException {
+    super.run();
 
     // for the 2nd attempt.
     if (appAttemptID.getAttemptId() == 2) {
@@ -39,11 +39,12 @@ public boolean run() throws YarnException, IOException {
       // numRequestedContainers should be set to 0.
       if (numAllocatedContainers.get() != 1
           || numRequestedContainers.get() != 0) {
-        LOG.info("Application Master failed. exiting");
+        LOG.info("NumAllocatedContainers is " + numAllocatedContainers.get()
+            + " and NumRequestedContainers is " + numAllocatedContainers.get()
+            + ".Application Master failed. exiting");
         System.exit(200);
       }
     }
-    return res;
   }
 
   public static void main(String[] args) {
@@ -54,7 +55,7 @@ public static void main(String[] args) {
       if (!doRun) {
         System.exit(0);
       }
-      result = appMaster.run();
+      appMaster.run();
       if (appMaster.appAttemptID.getAttemptId() == 1) {
         try {
           // sleep some time, wait for the AM to launch a container.
@@ -63,7 +64,7 @@ public static void main(String[] args) {
         // fail the first am.
         System.exit(100);
       }
-      appMaster.finish();
+      result = appMaster.finish();
     } catch (Throwable t) {
       System.exit(1);
     }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/main/java/org/apache/hadoop/yarn/client/api/YarnClient.java

@@ -436,4 +436,19 @@ public abstract ContainerReport getContainerReport(ContainerId containerId)
   public abstract List<ContainerReport> getContainers(
       ApplicationAttemptId applicationAttemptId) throws YarnException,
       IOException;
+  
+  /**
+   * <p>
+   * Attempts to move the given application to the given queue.
+   * </p>
+   * 
+   * @param appId
+   *    Application to move.
+   * @param queue
+   *    Queue to place it in to.
+   * @throws YarnException
+   * @throws IOException
+   */
+  public abstract void moveApplicationAcrossQueues(ApplicationId appId,
+      String queue) throws YarnException, IOException;
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/main/java/org/apache/hadoop/yarn/client/api/impl/YarnClientImpl.java

@@ -48,6 +48,7 @@
 import org.apache.hadoop.yarn.api.protocolrecords.GetQueueUserAclsInfoRequest;
 import org.apache.hadoop.yarn.api.protocolrecords.KillApplicationRequest;
 import org.apache.hadoop.yarn.api.protocolrecords.KillApplicationResponse;
+import org.apache.hadoop.yarn.api.protocolrecords.MoveApplicationAcrossQueuesRequest;
 import org.apache.hadoop.yarn.api.protocolrecords.SubmitApplicationRequest;
 import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
 import org.apache.hadoop.yarn.api.records.ApplicationAttemptReport;
@@ -478,4 +479,12 @@ public List<ContainerReport> getContainers(
     }
     throw new YarnException("History service is not enabled.");
   }
+  
+  @Override
+  public void moveApplicationAcrossQueues(ApplicationId appId,
+      String queue) throws YarnException, IOException {
+    MoveApplicationAcrossQueuesRequest request =
+        MoveApplicationAcrossQueuesRequest.newInstance(appId, queue);
+    rmClient.moveApplicationAcrossQueues(request);
+  }
 }

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/main/java/org/apache/hadoop/yarn/client/cli/ApplicationCLI.java

@@ -61,6 +61,7 @@ public class ApplicationCLI extends YarnCLI {
   private static final String APP_TYPE_CMD = "appTypes";
   private static final String APP_STATE_CMD = "appStates";
   private static final String ALLSTATES_OPTION = "ALL";
+  private static final String QUEUE_CMD = "queue";
   public static final String APPLICATION = "application";
   public static final String APPLICATION_ATTEMPT = "applicationattempt";
   public static final String CONTAINER = "container";
@@ -96,6 +97,10 @@ public int run(String[] args) throws Exception {
           + "and -appStates to filter applications based on application state");
     }
     opts.addOption(KILL_CMD, true, "Kills the application.");
+    opts.addOption(MOVE_TO_QUEUE_CMD, true, "Moves the application to a "
+        + "different queue.");
+    opts.addOption(QUEUE_CMD, true, "Works with the movetoqueue command to"
+        + " specify which queue to move an application to.");
     opts.addOption(HELP_CMD, false, "Displays help for all commands.");
     Option appTypeOpt = new Option(APP_TYPE_CMD, true, "Works with -list to "
         + "filter applications based on "
@@ -112,6 +117,8 @@ public int run(String[] args) throws Exception {
     appStateOpt.setArgName("States");
     opts.addOption(appStateOpt);
     opts.getOption(KILL_CMD).setArgName("Application ID");
+    opts.getOption(MOVE_TO_QUEUE_CMD).setArgName("Application ID");
+    opts.getOption(QUEUE_CMD).setArgName("Queue Name");
     opts.getOption(STATUS_CMD).setArgName("Application ID");
 
     int exitCode = -1;
@@ -197,11 +204,18 @@ public int run(String[] args) throws Exception {
         listApplications(appTypes, appStates);
       }
     } else if (cliParser.hasOption(KILL_CMD)) {
-      if (args.length != 2) {
+      if (args.length != 3) {
         printUsage(opts);
         return exitCode;
       }
       killApplication(cliParser.getOptionValue(KILL_CMD));
+    } else if (cliParser.hasOption(MOVE_TO_QUEUE_CMD)) {
+      if (!cliParser.hasOption(QUEUE_CMD)) {
+        printUsage(opts);
+        return exitCode;
+      }
+      moveApplicationAcrossQueues(cliParser.getOptionValue(MOVE_TO_QUEUE_CMD),
+          cliParser.getOptionValue(QUEUE_CMD));
     } else if (cliParser.hasOption(HELP_CMD)) {
       printUsage(opts);
       return 0;
@@ -366,6 +380,28 @@ private void killApplication(String applicationId) throws YarnException,
       client.killApplication(appId);
     }
   }
+  
+  /**
+   * Kills the application with the application id as appId
+   * 
+   * @param applicationId
+   * @throws YarnException
+   * @throws IOException
+   */
+  private void moveApplicationAcrossQueues(String applicationId, String queue)
+      throws YarnException, IOException {
+    ApplicationId appId = ConverterUtils.toApplicationId(applicationId);
+    ApplicationReport appReport = client.getApplicationReport(appId);
+    if (appReport.getYarnApplicationState() == YarnApplicationState.FINISHED
+        || appReport.getYarnApplicationState() == YarnApplicationState.KILLED
+        || appReport.getYarnApplicationState() == YarnApplicationState.FAILED) {
+      sysout.println("Application " + applicationId + " has already finished ");
+    } else {
+      sysout.println("Moving application " + applicationId + " to queue " + queue);
+      client.moveApplicationAcrossQueues(appId, queue);
+      sysout.println("Successfully completed move.");
+    }
+  }
 
   /**
    * Prints the application report for an application id.

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/main/java/org/apache/hadoop/yarn/client/cli/YarnCLI.java

@@ -33,6 +33,7 @@ public abstract class YarnCLI extends Configured implements Tool {
   public static final String STATUS_CMD = "status";
   public static final String LIST_CMD = "list";
   public static final String KILL_CMD = "kill";
+  public static final String MOVE_TO_QUEUE_CMD = "movetoqueue";
   public static final String HELP_CMD = "help";
   protected PrintStream sysout;
   protected PrintStream syserr;

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/test/java/org/apache/hadoop/yarn/client/TestRMFailover.java

@@ -172,8 +172,6 @@ public void testExplicitFailover()
   @Test
   public void testAutomaticFailover()
       throws YarnException, InterruptedException, IOException {
-    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, true);
-    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_EMBEDDED, true);
     conf.set(YarnConfiguration.RM_CLUSTER_ID, "yarn-test-cluster");
     conf.set(YarnConfiguration.RM_ZK_ADDRESS, hostPort);
     conf.setInt(YarnConfiguration.RM_ZK_TIMEOUT_MS, 2000);
@@ -193,6 +191,7 @@ public void testAutomaticFailover()
   @Test
   public void testWebAppProxyInStandAloneMode() throws YarnException,
       InterruptedException, IOException {
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, false);
     WebAppProxyServer webAppProxyServer = new WebAppProxyServer();
     try {
       conf.set(YarnConfiguration.PROXY_ADDRESS, "0.0.0.0:9099");
@@ -227,6 +226,7 @@ public void testWebAppProxyInStandAloneMode() throws YarnException,
   @Test
   public void testEmbeddedWebAppProxy() throws YarnException,
       InterruptedException, IOException {
+    conf.setBoolean(YarnConfiguration.AUTO_FAILOVER_ENABLED, false);
     cluster.init(conf);
     cluster.start();
     getAdminService(0).transitionToActive(req);

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-client/src/test/java/org/apache/hadoop/yarn/client/cli/TestYarnCLI.java

@@ -675,13 +675,14 @@ public void testAppsHelpCommand() throws Exception {
     int result = spyCli.run(new String[] { "-help" });
     Assert.assertTrue(result == 0);
     verify(spyCli).printUsage(any(Options.class));
+    System.err.println(sysOutStream.toString()); //todo sandyt remove this hejfkdsl
     Assert.assertEquals(createApplicationCLIHelpMessage(),
         sysOutStream.toString());
 
     sysOutStream.reset();
     ApplicationId applicationId = ApplicationId.newInstance(1234, 5);
     result =
-        cli.run(new String[] { "-kill", applicationId.toString(), "args" });
+        cli.run(new String[] {"application", "-kill", applicationId.toString(), "args" });
     verify(spyCli).printUsage(any(Options.class));
     Assert.assertEquals(createApplicationCLIHelpMessage(),
         sysOutStream.toString());
@@ -717,7 +718,7 @@ public void testKillApplication() throws Exception {
         FinalApplicationStatus.SUCCEEDED, null, "N/A", 0.53789f, "YARN", null);
     when(client.getApplicationReport(any(ApplicationId.class))).thenReturn(
         newApplicationReport2);
-    int result = cli.run(new String[] { "-kill", applicationId.toString() });
+    int result = cli.run(new String[] { "application","-kill", applicationId.toString() });
     assertEquals(0, result);
     verify(client, times(0)).killApplication(any(ApplicationId.class));
     verify(sysOut).println(
@@ -730,7 +731,7 @@ public void testKillApplication() throws Exception {
         FinalApplicationStatus.SUCCEEDED, null, "N/A", 0.53789f, "YARN", null);
     when(client.getApplicationReport(any(ApplicationId.class))).thenReturn(
         newApplicationReport);
-    result = cli.run(new String[] { "-kill", applicationId.toString() });
+    result = cli.run(new String[] { "application","-kill", applicationId.toString() });
     assertEquals(0, result);
     verify(client).killApplication(any(ApplicationId.class));
     verify(sysOut).println("Killing application application_1234_0005");
@@ -740,7 +741,57 @@ public void testKillApplication() throws Exception {
         .getApplicationReport(applicationId);
     cli = createAndGetAppCLI();
     try {
-      cli.run(new String[] { "-kill", applicationId.toString() });
+      cli.run(new String[] { "application","-kill", applicationId.toString() });
+      Assert.fail();
+    } catch (Exception ex) {
+      Assert.assertTrue(ex instanceof ApplicationNotFoundException);
+      Assert.assertEquals("Application with id '" + applicationId +
+          "' doesn't exist in RM.", ex.getMessage());
+    }
+  }
+  
+  @Test
+  public void testMoveApplicationAcrossQueues() throws Exception {
+    ApplicationCLI cli = createAndGetAppCLI();
+    ApplicationId applicationId = ApplicationId.newInstance(1234, 5);
+
+    ApplicationReport newApplicationReport2 = ApplicationReport.newInstance(
+        applicationId, ApplicationAttemptId.newInstance(applicationId, 1),
+        "user", "queue", "appname", "host", 124, null,
+        YarnApplicationState.FINISHED, "diagnostics", "url", 0, 0,
+        FinalApplicationStatus.SUCCEEDED, null, "N/A", 0.53789f, "YARN", null);
+    when(client.getApplicationReport(any(ApplicationId.class))).thenReturn(
+        newApplicationReport2);
+    int result = cli.run(new String[] { "-movetoqueue", applicationId.toString(),
+        "-queue", "targetqueue"});
+    assertEquals(0, result);
+    verify(client, times(0)).moveApplicationAcrossQueues(
+        any(ApplicationId.class), any(String.class));
+    verify(sysOut).println(
+        "Application " + applicationId + " has already finished ");
+
+    ApplicationReport newApplicationReport = ApplicationReport.newInstance(
+        applicationId, ApplicationAttemptId.newInstance(applicationId, 1),
+        "user", "queue", "appname", "host", 124, null,
+        YarnApplicationState.RUNNING, "diagnostics", "url", 0, 0,
+        FinalApplicationStatus.SUCCEEDED, null, "N/A", 0.53789f, "YARN", null);
+    when(client.getApplicationReport(any(ApplicationId.class))).thenReturn(
+        newApplicationReport);
+    result = cli.run(new String[] { "-movetoqueue", applicationId.toString(),
+        "-queue", "targetqueue"});
+    assertEquals(0, result);
+    verify(client).moveApplicationAcrossQueues(any(ApplicationId.class),
+        any(String.class));
+    verify(sysOut).println("Moving application application_1234_0005 to queue targetqueue");
+    verify(sysOut).println("Successfully completed move.");
+
+    doThrow(new ApplicationNotFoundException("Application with id '"
+        + applicationId + "' doesn't exist in RM.")).when(client)
+        .moveApplicationAcrossQueues(applicationId, "targetqueue");
+    cli = createAndGetAppCLI();
+    try {
+      result = cli.run(new String[] { "-movetoqueue", applicationId.toString(),
+          "-queue", "targetqueue"});
       Assert.fail();
     } catch (Exception ex) {
       Assert.assertTrue(ex instanceof ApplicationNotFoundException);
@@ -1087,23 +1138,28 @@ private String createApplicationCLIHelpMessage() throws IOException {
     ByteArrayOutputStream baos = new ByteArrayOutputStream();
     PrintWriter pw = new PrintWriter(baos);
     pw.println("usage: application");
-    pw.println(" -appStates <States>        Works with -list to filter applications based");
-    pw.println("                            on input comma-separated list of application");
-    pw.println("                            states. The valid application state can be one");
-    pw.println("                            of the following:");
-    pw.println("                            ALL,NEW,NEW_SAVING,SUBMITTED,ACCEPTED,RUNNING,");
-    pw.println("                            FINISHED,FAILED,KILLED");
-    pw.println(" -appTypes <Types>          Works with -list to filter applications based");
-    pw.println("                            on input comma-separated list of application");
-    pw.println("                            types.");
-    pw.println(" -help                      Displays help for all commands.");
-    pw.println(" -kill <Application ID>     Kills the application.");
-    pw.println(" -list                      List applications from the RM. Supports");
-    pw.println("                            optional use of -appTypes to filter");
-    pw.println("                            applications based on application type, and");
-    pw.println("                            -appStates to filter applications based on");
-    pw.println("                            application state");
-    pw.println(" -status <Application ID>   Prints the status of the application.");
+    pw.println(" -appStates <States>             Works with -list to filter applications");
+    pw.println("                                 based on input comma-separated list of");
+    pw.println("                                 application states. The valid application");
+    pw.println("                                 state can be one of the following:");
+    pw.println("                                 ALL,NEW,NEW_SAVING,SUBMITTED,ACCEPTED,RUN");
+    pw.println("                                 NING,FINISHED,FAILED,KILLED");
+    pw.println(" -appTypes <Types>               Works with -list to filter applications");
+    pw.println("                                 based on input comma-separated list of");
+    pw.println("                                 application types.");
+    pw.println(" -help                           Displays help for all commands.");
+    pw.println(" -kill <Application ID>          Kills the application.");
+    pw.println(" -list                           List applications from the RM. Supports");
+    pw.println("                                 optional use of -appTypes to filter");
+    pw.println("                                 applications based on application type,");
+    pw.println("                                 and -appStates to filter applications");
+    pw.println("                                 based on application state");
+    pw.println(" -movetoqueue <Application ID>   Moves the application to a different");
+    pw.println("                                 queue.");
+    pw.println(" -queue <Queue Name>             Works with the movetoqueue command to");
+    pw.println("                                 specify which queue to move an");
+    pw.println("                                 application to.");
+    pw.println(" -status <Application ID>        Prints the status of the application.");
     pw.close();
     String appsHelpStr = baos.toString("UTF-8");
     return appsHelpStr;

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/FileSystemBasedConfigurationProvider.java

@@ -0,0 +1,72 @@
+/**
+ * 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.yarn;
+
+import java.io.IOException;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.classification.InterfaceAudience.Private;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.yarn.conf.ConfigurationProvider;
+import org.apache.hadoop.yarn.conf.YarnConfiguration;
+import org.apache.hadoop.yarn.exceptions.YarnException;
+
+@Private
+@Unstable
+public class FileSystemBasedConfigurationProvider
+    extends ConfigurationProvider {
+
+  private static final Log LOG = LogFactory
+      .getLog(FileSystemBasedConfigurationProvider.class);
+  private FileSystem fs;
+  private Path configDir;
+
+  @Override
+  public synchronized Configuration getConfiguration(String name)
+      throws IOException, YarnException {
+    Path configPath = new Path(this.configDir, name);
+    if (!fs.exists(configPath)) {
+      throw new YarnException("Can not find Configuration: " + name + " in "
+          + configDir);
+    }
+    Configuration conf = new Configuration(false);
+    conf.addResource(fs.open(configPath));
+    return conf;
+  }
+
+  @Override
+  public synchronized void initInternal(Configuration conf) throws Exception {
+    configDir =
+        new Path(conf.get(YarnConfiguration.FS_BASED_RM_CONF_STORE,
+            YarnConfiguration.DEFAULT_FS_BASED_RM_CONF_STORE));
+    fs = configDir.getFileSystem(conf);
+    if (!fs.exists(configDir)) {
+      fs.mkdirs(configDir);
+    }
+  }
+
+  @Override
+  public synchronized void closeInternal() throws Exception {
+    fs.close();
+  }
+}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/LocalConfigurationProvider.java

@@ -0,0 +1,48 @@
+/**
+ * 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.yarn;
+
+import java.io.IOException;
+
+import org.apache.hadoop.classification.InterfaceAudience.Private;
+import org.apache.hadoop.classification.InterfaceStability.Unstable;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.yarn.conf.ConfigurationProvider;
+import org.apache.hadoop.yarn.exceptions.YarnException;
+
+@Private
+@Unstable
+public class LocalConfigurationProvider extends ConfigurationProvider {
+
+  @Override
+  public Configuration getConfiguration(String name)
+      throws IOException, YarnException {
+    return new Configuration();
+  }
+
+  @Override
+  public void initInternal(Configuration conf) throws Exception {
+    // Do nothing
+  }
+
+  @Override
+  public void closeInternal() throws Exception {
+    // Do nothing
+  }
+}