Merge remote-tracking branch 'upstream/master' into mllib-stats-api-check

Author: jkbradley

bin/pyspark

@@ -23,12 +23,18 @@ FWDIR="$(cd `dirname $0`/..; pwd)"
 # Export this as SPARK_HOME
 export SPARK_HOME="$FWDIR"
 
+source $FWDIR/bin/utils.sh
+
 SCALA_VERSION=2.10
 
-if [[ "$@" = *--help ]] || [[ "$@" = *-h ]]; then
+function usage() {
   echo "Usage: ./bin/pyspark [options]" 1>&2
   $FWDIR/bin/spark-submit --help 2>&1 | grep -v Usage 1>&2
   exit 0
+}
+
+if [[ "$@" = *--help ]] || [[ "$@" = *-h ]]; then
+  usage
 fi
 
 # Exit if the user hasn't compiled Spark
@@ -66,10 +72,11 @@ fi
 # Build up arguments list manually to preserve quotes and backslashes.
 # We export Spark submit arguments as an environment variable because shell.py must run as a
 # PYTHONSTARTUP script, which does not take in arguments. This is required for IPython notebooks.
-
+SUBMIT_USAGE_FUNCTION=usage
+gatherSparkSubmitOpts "$@"
 PYSPARK_SUBMIT_ARGS=""
 whitespace="[[:space:]]"
-for i in "$@"; do
+for i in "${SUBMISSION_OPTS[@]}"; do
   if [[ $i =~ \" ]]; then i=$(echo $i | sed 's/\"/\\\"/g'); fi
   if [[ $i =~ $whitespace ]]; then i=\"$i\"; fi
   PYSPARK_SUBMIT_ARGS="$PYSPARK_SUBMIT_ARGS $i"
@@ -90,7 +97,10 @@ fi
 if [[ "$1" =~ \.py$ ]]; then
   echo -e "\nWARNING: Running python applications through ./bin/pyspark is deprecated as of Spark 1.0." 1>&2
   echo -e "Use ./bin/spark-submit <python file>\n" 1>&2
-  exec $FWDIR/bin/spark-submit "$@"
+  primary=$1
+  shift
+  gatherSparkSubmitOpts "$@"
+  exec $FWDIR/bin/spark-submit "${SUBMISSION_OPTS[@]}" $primary "${APPLICATION_OPTS[@]}"
 else
   # Only use ipython if no command line arguments were provided [SPARK-1134]
   if [[ "$IPYTHON" = "1" ]]; then

bin/spark-shell

@@ -31,13 +31,21 @@ set -o posix
 ## Global script variables
 FWDIR="$(cd `dirname $0`/..; pwd)"
 
+function usage() {
+    echo "Usage: ./bin/spark-shell [options]"
+    $FWDIR/bin/spark-submit --help 2>&1 | grep -v Usage 1>&2
+    exit 0
+}
+
 if [[ "$@" = *--help ]] || [[ "$@" = *-h ]]; then
-  echo "Usage: ./bin/spark-shell [options]"
-  $FWDIR/bin/spark-submit --help 2>&1 | grep -v Usage 1>&2
-  exit 0
+  usage
 fi
 
-function main(){
+source $FWDIR/bin/utils.sh
+SUBMIT_USAGE_FUNCTION=usage
+gatherSparkSubmitOpts "$@"
+
+function main() {
     if $cygwin; then
         # Workaround for issue involving JLine and Cygwin
         # (see http://sourceforge.net/p/jline/bugs/40/).
@@ -46,11 +54,11 @@ function main(){
         # (see https://github.com/sbt/sbt/issues/562).
         stty -icanon min 1 -echo > /dev/null 2>&1
         export SPARK_SUBMIT_OPTS="$SPARK_SUBMIT_OPTS -Djline.terminal=unix"
-        $FWDIR/bin/spark-submit --class org.apache.spark.repl.Main spark-shell "$@"
+        $FWDIR/bin/spark-submit --class org.apache.spark.repl.Main "${SUBMISSION_OPTS[@]}" spark-shell "${APPLICATION_OPTS[@]}"
         stty icanon echo > /dev/null 2>&1
     else
         export SPARK_SUBMIT_OPTS
-        $FWDIR/bin/spark-submit --class org.apache.spark.repl.Main spark-shell "$@"
+        $FWDIR/bin/spark-submit --class org.apache.spark.repl.Main "${SUBMISSION_OPTS[@]}" spark-shell "${APPLICATION_OPTS[@]}"
     fi
 }
 

bin/utils.sh

@@ -0,0 +1,59 @@
+#!/usr/bin/env bash
+
+#
+# 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.
+#
+
+# Gather all all spark-submit options into SUBMISSION_OPTS
+function gatherSparkSubmitOpts() {
+
+  if [ -z "$SUBMIT_USAGE_FUNCTION" ]; then
+    echo "Function for printing usage of $0 is not set." 1>&2
+    echo "Please set usage function to shell variable 'SUBMIT_USAGE_FUNCTION' in $0" 1>&2
+    exit 1
+  fi
+
+  # NOTE: If you add or remove spark-sumbmit options,
+  # modify NOT ONLY this script but also SparkSubmitArgument.scala
+  SUBMISSION_OPTS=()
+  APPLICATION_OPTS=()
+  while (($#)); do
+    case "$1" in
+      --master | --deploy-mode | --class | --name | --jars | --py-files | --files | \
+      --conf | --properties-file | --driver-memory | --driver-java-options | \
+      --driver-library-path | --driver-class-path | --executor-memory | --driver-cores | \
+      --total-executor-cores | --executor-cores | --queue | --num-executors | --archives)
+        if [[ $# -lt 2 ]]; then
+          "$SUBMIT_USAGE_FUNCTION"
+          exit 1;
+        fi
+        SUBMISSION_OPTS+=("$1"); shift
+        SUBMISSION_OPTS+=("$1"); shift
+        ;;
+
+      --verbose | -v | --supervise)
+        SUBMISSION_OPTS+=("$1"); shift
+        ;;
+
+      *)
+        APPLICATION_OPTS+=("$1"); shift
+        ;;
+    esac
+  done
+
+  export SUBMISSION_OPTS
+  export APPLICATION_OPTS
+}

core/src/main/scala/org/apache/spark/SparkEnv.scala

@@ -156,11 +156,9 @@ object SparkEnv extends Logging {
       conf.set("spark.driver.port", boundPort.toString)
     }
 
-    // Create an instance of the class named by the given Java system property, or by
-    // defaultClassName if the property is not set, and return it as a T
-    def instantiateClass[T](propertyName: String, defaultClassName: String): T = {
-      val name = conf.get(propertyName,  defaultClassName)
-      val cls = Class.forName(name, true, Utils.getContextOrSparkClassLoader)
+    // Create an instance of the class with the given name, possibly initializing it with our conf
+    def instantiateClass[T](className: String): T = {
+      val cls = Class.forName(className, true, Utils.getContextOrSparkClassLoader)
       // Look for a constructor taking a SparkConf and a boolean isDriver, then one taking just
       // SparkConf, then one taking no arguments
       try {
@@ -178,11 +176,17 @@ object SparkEnv extends Logging {
       }
     }
 
-    val serializer = instantiateClass[Serializer](
+    // Create an instance of the class named by the given SparkConf property, or defaultClassName
+    // if the property is not set, possibly initializing it with our conf
+    def instantiateClassFromConf[T](propertyName: String, defaultClassName: String): T = {
+      instantiateClass[T](conf.get(propertyName, defaultClassName))
+    }
+
+    val serializer = instantiateClassFromConf[Serializer](
       "spark.serializer", "org.apache.spark.serializer.JavaSerializer")
     logDebug(s"Using serializer: ${serializer.getClass}")
 
-    val closureSerializer = instantiateClass[Serializer](
+    val closureSerializer = instantiateClassFromConf[Serializer](
       "spark.closure.serializer", "org.apache.spark.serializer.JavaSerializer")
 
     def registerOrLookup(name: String, newActor: => Actor): ActorRef = {
@@ -246,8 +250,13 @@ object SparkEnv extends Logging {
       "."
     }
 
-    val shuffleManager = instantiateClass[ShuffleManager](
-      "spark.shuffle.manager", "org.apache.spark.shuffle.hash.HashShuffleManager")
+    // Let the user specify short names for shuffle managers
+    val shortShuffleMgrNames = Map(
+      "hash" -> "org.apache.spark.shuffle.hash.HashShuffleManager",
+      "sort" -> "org.apache.spark.shuffle.sort.SortShuffleManager")
+    val shuffleMgrName = conf.get("spark.shuffle.manager", "hash")
+    val shuffleMgrClass = shortShuffleMgrNames.getOrElse(shuffleMgrName.toLowerCase, shuffleMgrName)
+    val shuffleManager = instantiateClass[ShuffleManager](shuffleMgrClass)
 
     val shuffleMemoryManager = new ShuffleMemoryManager(conf)
 

core/src/main/scala/org/apache/spark/broadcast/TorrentBroadcast.scala

@@ -17,14 +17,15 @@
 
 package org.apache.spark.broadcast
 
-import java.io.{ByteArrayInputStream, ObjectInputStream, ObjectOutputStream}
+import java.io.{ByteArrayOutputStream, ByteArrayInputStream, InputStream,
+  ObjectInputStream, ObjectOutputStream, OutputStream}
 
 import scala.reflect.ClassTag
 import scala.util.Random
 
 import org.apache.spark.{Logging, SparkConf, SparkEnv, SparkException}
+import org.apache.spark.io.CompressionCodec
 import org.apache.spark.storage.{BroadcastBlockId, StorageLevel}
-import org.apache.spark.util.Utils
 
 /**
  *  A [[org.apache.spark.broadcast.Broadcast]] implementation that uses a BitTorrent-like
@@ -214,11 +215,15 @@ private[broadcast] object TorrentBroadcast extends Logging {
   private lazy val BLOCK_SIZE = conf.getInt("spark.broadcast.blockSize", 4096) * 1024
   private var initialized = false
   private var conf: SparkConf = null
+  private var compress: Boolean = false
+  private var compressionCodec: CompressionCodec = null
 
   def initialize(_isDriver: Boolean, conf: SparkConf) {
     TorrentBroadcast.conf = conf // TODO: we might have to fix it in tests
     synchronized {
       if (!initialized) {
+        compress = conf.getBoolean("spark.broadcast.compress", true)
+        compressionCodec = CompressionCodec.createCodec(conf)
         initialized = true
       }
     }
@@ -228,8 +233,13 @@ private[broadcast] object TorrentBroadcast extends Logging {
     initialized = false
   }
 
-  def blockifyObject[T](obj: T): TorrentInfo = {
-    val byteArray = Utils.serialize[T](obj)
+  def blockifyObject[T: ClassTag](obj: T): TorrentInfo = {
+    val bos = new ByteArrayOutputStream()
+    val out: OutputStream = if (compress) compressionCodec.compressedOutputStream(bos) else bos
+    val ser = SparkEnv.get.serializer.newInstance()
+    val serOut = ser.serializeStream(out)
+    serOut.writeObject[T](obj).close()
+    val byteArray = bos.toByteArray
     val bais = new ByteArrayInputStream(byteArray)
 
     var blockNum = byteArray.length / BLOCK_SIZE
@@ -255,7 +265,7 @@ private[broadcast] object TorrentBroadcast extends Logging {
     info
   }
 
-  def unBlockifyObject[T](
+  def unBlockifyObject[T: ClassTag](
       arrayOfBlocks: Array[TorrentBlock],
       totalBytes: Int,
       totalBlocks: Int): T = {
@@ -264,7 +274,16 @@ private[broadcast] object TorrentBroadcast extends Logging {
       System.arraycopy(arrayOfBlocks(i).byteArray, 0, retByteArray,
         i * BLOCK_SIZE, arrayOfBlocks(i).byteArray.length)
     }
-    Utils.deserialize[T](retByteArray, Thread.currentThread.getContextClassLoader)
+
+    val in: InputStream = {
+      val arrIn = new ByteArrayInputStream(retByteArray)
+      if (compress) compressionCodec.compressedInputStream(arrIn) else arrIn
+    }
+    val ser = SparkEnv.get.serializer.newInstance()
+    val serIn = ser.deserializeStream(in)
+    val obj = serIn.readObject[T]()
+    serIn.close()
+    obj
   }
 
   /**

core/src/main/scala/org/apache/spark/deploy/SparkSubmitArguments.scala

@@ -219,12 +219,15 @@ private[spark] class SparkSubmitArguments(args: Seq[String]) {
 
   /** Fill in values by parsing user options. */
   private def parseOpts(opts: Seq[String]): Unit = {
-    var inSparkOpts = true
     val EQ_SEPARATED_OPT="""(--[^=]+)=(.+)""".r
 
     // Delineates parsing of Spark options from parsing of user options.
     parse(opts)
 
+    /**
+     * NOTE: If you add or remove spark-submit options,
+     * modify NOT ONLY this file but also utils.sh
+     */
     def parse(opts: Seq[String]): Unit = opts match {
       case ("--name") :: value :: tail =>
         name = value

core/src/main/scala/org/apache/spark/deploy/worker/Worker.scala

@@ -136,7 +136,7 @@ private[spark] class Worker(
     logInfo("Spark home: " + sparkHome)
     createWorkDir()
     context.system.eventStream.subscribe(self, classOf[RemotingLifecycleEvent])
-    webUi = new WorkerWebUI(this, workDir, Some(webUiPort))
+    webUi = new WorkerWebUI(this, workDir, webUiPort)
     webUi.bind()
     registerWithMaster()
 
@@ -373,7 +373,8 @@ private[spark] class Worker(
 private[spark] object Worker extends Logging {
   def main(argStrings: Array[String]) {
     SignalLogger.register(log)
-    val args = new WorkerArguments(argStrings)
+    val conf = new SparkConf
+    val args = new WorkerArguments(argStrings, conf)
     val (actorSystem, _) = startSystemAndActor(args.host, args.port, args.webUiPort, args.cores,
       args.memory, args.masters, args.workDir)
     actorSystem.awaitTermination()

core/src/main/scala/org/apache/spark/deploy/worker/WorkerArguments.scala

@@ -20,11 +20,12 @@ package org.apache.spark.deploy.worker
 import java.lang.management.ManagementFactory
 
 import org.apache.spark.util.{IntParam, MemoryParam, Utils}
+import org.apache.spark.SparkConf
 
 /**
  * Command-line parser for the worker.
  */
-private[spark] class WorkerArguments(args: Array[String]) {
+private[spark] class WorkerArguments(args: Array[String], conf: SparkConf) {
   var host = Utils.localHostName()
   var port = 0
   var webUiPort = 8081
@@ -46,6 +47,9 @@ private[spark] class WorkerArguments(args: Array[String]) {
   if (System.getenv("SPARK_WORKER_WEBUI_PORT") != null) {
     webUiPort = System.getenv("SPARK_WORKER_WEBUI_PORT").toInt
   }
+  if (conf.contains("spark.worker.ui.port")) {
+    webUiPort = conf.get("spark.worker.ui.port").toInt
+  }
   if (System.getenv("SPARK_WORKER_DIR") != null) {
     workDir = System.getenv("SPARK_WORKER_DIR")
   }

core/src/main/scala/org/apache/spark/deploy/worker/ui/WorkerWebUI.scala

@@ -34,8 +34,8 @@ private[spark]
 class WorkerWebUI(
     val worker: Worker,
     val workDir: File,
-    port: Option[Int] = None)
-  extends WebUI(worker.securityMgr, getUIPort(port, worker.conf), worker.conf, name = "WorkerUI")
+    requestedPort: Int)
+  extends WebUI(worker.securityMgr, requestedPort, worker.conf, name = "WorkerUI")
   with Logging {
 
   val timeout = AkkaUtils.askTimeout(worker.conf)
@@ -55,10 +55,5 @@ class WorkerWebUI(
 }
 
 private[spark] object WorkerWebUI {
-  val DEFAULT_PORT = 8081
   val STATIC_RESOURCE_BASE = SparkUI.STATIC_RESOURCE_DIR
-
-  def getUIPort(requestedPort: Option[Int], conf: SparkConf): Int = {
-    requestedPort.getOrElse(conf.getInt("spark.worker.ui.port", WorkerWebUI.DEFAULT_PORT))
-  }
 }

core/src/main/scala/org/apache/spark/executor/Executor.scala

@@ -374,6 +374,7 @@ private[spark] class Executor(
           for (taskRunner <- runningTasks.values()) {
             if (!taskRunner.attemptedTask.isEmpty) {
               Option(taskRunner.task).flatMap(_.metrics).foreach { metrics =>
+                metrics.updateShuffleReadMetrics
                 tasksMetrics += ((taskRunner.taskId, metrics))
               }
             }