Merge branch 'master' into hist-server-single-log

Conflicts:
	core/src/main/scala/org/apache/spark/util/FileLogger.scala
	core/src/test/scala/org/apache/spark/scheduler/EventLoggingListenerSuite.scala
	core/src/test/scala/org/apache/spark/util/FileLoggerSuite.scala

Author: Marcelo Vanzin

bin/pyspark

@@ -50,22 +50,47 @@ fi
 
 . "$FWDIR"/bin/load-spark-env.sh
 
-# Figure out which Python executable to use
+# In Spark <= 1.1, setting IPYTHON=1 would cause the driver to be launched using the `ipython`
+# executable, while the worker would still be launched using PYSPARK_PYTHON.
+#
+# In Spark 1.2, we removed the documentation of the IPYTHON and IPYTHON_OPTS variables and added
+# PYSPARK_DRIVER_PYTHON and PYSPARK_DRIVER_PYTHON_OPTS to allow IPython to be used for the driver.
+# Now, users can simply set PYSPARK_DRIVER_PYTHON=ipython to use IPython and set
+# PYSPARK_DRIVER_PYTHON_OPTS to pass options when starting the Python driver
+# (e.g. PYSPARK_DRIVER_PYTHON_OPTS='notebook').  This supports full customization of the IPython
+# and executor Python executables.
+#
+# For backwards-compatibility, we retain the old IPYTHON and IPYTHON_OPTS variables.
+
+# Determine the Python executable to use if PYSPARK_PYTHON or PYSPARK_DRIVER_PYTHON isn't set:
+if hash python2.7 2>/dev/null; then
+  # Attempt to use Python 2.7, if installed:
+  DEFAULT_PYTHON="python2.7"
+else
+  DEFAULT_PYTHON="python"
+fi
+
+# Determine the Python executable to use for the driver:
+if [[ -n "$IPYTHON_OPTS" || "$IPYTHON" == "1" ]]; then
+  # If IPython options are specified, assume user wants to run IPython
+  # (for backwards-compatibility)
+  PYSPARK_DRIVER_PYTHON_OPTS="$PYSPARK_DRIVER_PYTHON_OPTS $IPYTHON_OPTS"
+  PYSPARK_DRIVER_PYTHON="ipython"
+elif [[ -z "$PYSPARK_DRIVER_PYTHON" ]]; then
+  PYSPARK_DRIVER_PYTHON="${PYSPARK_PYTHON:-"$DEFAULT_PYTHON"}"
+fi
+
+# Determine the Python executable to use for the executors:
 if [[ -z "$PYSPARK_PYTHON" ]]; then
-  if [[ "$IPYTHON" = "1" || -n "$IPYTHON_OPTS" ]]; then
-    # for backward compatibility
-    PYSPARK_PYTHON="ipython"
+  if [[ $PYSPARK_DRIVER_PYTHON == *ipython* && $DEFAULT_PYTHON != "python2.7" ]]; then
+    echo "IPython requires Python 2.7+; please install python2.7 or set PYSPARK_PYTHON" 1>&2
+    exit 1
   else
-    PYSPARK_PYTHON="python"
+    PYSPARK_PYTHON="$DEFAULT_PYTHON"
   fi
 fi
 export PYSPARK_PYTHON
 
-if [[ -z "$PYSPARK_PYTHON_OPTS" && -n "$IPYTHON_OPTS" ]]; then
-  # for backward compatibility
-  PYSPARK_PYTHON_OPTS="$IPYTHON_OPTS"
-fi
-
 # Add the PySpark classes to the Python path:
 export PYTHONPATH="$SPARK_HOME/python/:$PYTHONPATH"
 export PYTHONPATH="$SPARK_HOME/python/lib/py4j-0.8.2.1-src.zip:$PYTHONPATH"
@@ -93,9 +118,9 @@ if [[ -n "$SPARK_TESTING" ]]; then
   unset YARN_CONF_DIR
   unset HADOOP_CONF_DIR
   if [[ -n "$PYSPARK_DOC_TEST" ]]; then
-    exec "$PYSPARK_PYTHON" -m doctest $1
+    exec "$PYSPARK_DRIVER_PYTHON" -m doctest $1
   else
-    exec "$PYSPARK_PYTHON" $1
+    exec "$PYSPARK_DRIVER_PYTHON" $1
   fi
   exit
 fi
@@ -111,5 +136,5 @@ if [[ "$1" =~ \.py$ ]]; then
 else
   # PySpark shell requires special handling downstream
   export PYSPARK_SHELL=1
-  exec "$PYSPARK_PYTHON" $PYSPARK_PYTHON_OPTS
+  exec "$PYSPARK_DRIVER_PYTHON" $PYSPARK_DRIVER_PYTHON_OPTS
 fi

bin/spark-class

@@ -105,7 +105,7 @@ else
     exit 1
   fi
 fi
-JAVA_VERSION=$("$RUNNER" -version 2>&1 | sed 's/.* version "\(.*\)\.\(.*\)\..*"/\1\2/; 1q')
+JAVA_VERSION=$("$RUNNER" -version 2>&1 | grep 'version' | sed 's/.* version "\(.*\)\.\(.*\)\..*"/\1\2/; 1q')
 
 # Set JAVA_OPTS to be able to load native libraries and to set heap size
 if [ "$JAVA_VERSION" -ge 18 ]; then

bin/spark-shell.cmd

@@ -17,6 +17,7 @@ rem See the License for the specific language governing permissions and
 rem limitations under the License.
 rem
 
-set SPARK_HOME=%~dp0..
+rem This is the entry point for running Spark shell. To avoid polluting the
+rem environment, it just launches a new cmd to do the real work.
 
-cmd /V /E /C %SPARK_HOME%\bin\spark-submit.cmd --class org.apache.spark.repl.Main %* spark-shell
+cmd /V /E /C %~dp0spark-shell2.cmd %*

bin/spark-shell2.cmd

@@ -0,0 +1,22 @@
+@echo off
+
+rem
+rem Licensed to the Apache Software Foundation (ASF) under one or more
+rem contributor license agreements.  See the NOTICE file distributed with
+rem this work for additional information regarding copyright ownership.
+rem The ASF licenses this file to You under the Apache License, Version 2.0
+rem (the "License"); you may not use this file except in compliance with
+rem the License.  You may obtain a copy of the License at
+rem
+rem    http://www.apache.org/licenses/LICENSE-2.0
+rem
+rem Unless required by applicable law or agreed to in writing, software
+rem distributed under the License is distributed on an "AS IS" BASIS,
+rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+rem See the License for the specific language governing permissions and
+rem limitations under the License.
+rem
+
+set SPARK_HOME=%~dp0..
+
+cmd /V /E /C %SPARK_HOME%\bin\spark-submit.cmd --class org.apache.spark.repl.Main %* spark-shell

bin/spark-submit.cmd

@@ -17,52 +17,7 @@ rem See the License for the specific language governing permissions and
 rem limitations under the License.
 rem
 
-rem NOTE: Any changes in this file must be reflected in SparkSubmitDriverBootstrapper.scala!
+rem This is the entry point for running Spark submit. To avoid polluting the
+rem environment, it just launches a new cmd to do the real work.
 
-set SPARK_HOME=%~dp0..
-set ORIG_ARGS=%*
-
-rem Reset the values of all variables used
-set SPARK_SUBMIT_DEPLOY_MODE=client
-set SPARK_SUBMIT_PROPERTIES_FILE=%SPARK_HOME%\conf\spark-defaults.conf
-set SPARK_SUBMIT_DRIVER_MEMORY=
-set SPARK_SUBMIT_LIBRARY_PATH=
-set SPARK_SUBMIT_CLASSPATH=
-set SPARK_SUBMIT_OPTS=
-set SPARK_SUBMIT_BOOTSTRAP_DRIVER=
-
-:loop
-if [%1] == [] goto continue
-  if [%1] == [--deploy-mode] (
-    set SPARK_SUBMIT_DEPLOY_MODE=%2
-  ) else if [%1] == [--properties-file] (
-    set SPARK_SUBMIT_PROPERTIES_FILE=%2
-  ) else if [%1] == [--driver-memory] (
-    set SPARK_SUBMIT_DRIVER_MEMORY=%2
-  ) else if [%1] == [--driver-library-path] (
-    set SPARK_SUBMIT_LIBRARY_PATH=%2
-  ) else if [%1] == [--driver-class-path] (
-    set SPARK_SUBMIT_CLASSPATH=%2
-  ) else if [%1] == [--driver-java-options] (
-    set SPARK_SUBMIT_OPTS=%2
-  )
-  shift
-goto loop
-:continue
-
-rem For client mode, the driver will be launched in the same JVM that launches
-rem SparkSubmit, so we may need to read the properties file for any extra class
-rem paths, library paths, java options and memory early on. Otherwise, it will
-rem be too late by the time the driver JVM has started.
-
-if [%SPARK_SUBMIT_DEPLOY_MODE%] == [client] (
-  if exist %SPARK_SUBMIT_PROPERTIES_FILE% (
-    rem Parse the properties file only if the special configs exist
-    for /f %%i in ('findstr /r /c:"^[\t ]*spark.driver.memory" /c:"^[\t ]*spark.driver.extra" ^
-      %SPARK_SUBMIT_PROPERTIES_FILE%') do (
-      set SPARK_SUBMIT_BOOTSTRAP_DRIVER=1
-    )
-  )
-)
-
-cmd /V /E /C %SPARK_HOME%\bin\spark-class.cmd org.apache.spark.deploy.SparkSubmit %ORIG_ARGS%
+cmd /V /E /C %~dp0spark-submit2.cmd %*

bin/spark-submit2.cmd

@@ -0,0 +1,68 @@
+@echo off
+
+rem
+rem Licensed to the Apache Software Foundation (ASF) under one or more
+rem contributor license agreements.  See the NOTICE file distributed with
+rem this work for additional information regarding copyright ownership.
+rem The ASF licenses this file to You under the Apache License, Version 2.0
+rem (the "License"); you may not use this file except in compliance with
+rem the License.  You may obtain a copy of the License at
+rem
+rem    http://www.apache.org/licenses/LICENSE-2.0
+rem
+rem Unless required by applicable law or agreed to in writing, software
+rem distributed under the License is distributed on an "AS IS" BASIS,
+rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+rem See the License for the specific language governing permissions and
+rem limitations under the License.
+rem
+
+rem NOTE: Any changes in this file must be reflected in SparkSubmitDriverBootstrapper.scala!
+
+set SPARK_HOME=%~dp0..
+set ORIG_ARGS=%*
+
+rem Reset the values of all variables used
+set SPARK_SUBMIT_DEPLOY_MODE=client
+set SPARK_SUBMIT_PROPERTIES_FILE=%SPARK_HOME%\conf\spark-defaults.conf
+set SPARK_SUBMIT_DRIVER_MEMORY=
+set SPARK_SUBMIT_LIBRARY_PATH=
+set SPARK_SUBMIT_CLASSPATH=
+set SPARK_SUBMIT_OPTS=
+set SPARK_SUBMIT_BOOTSTRAP_DRIVER=
+
+:loop
+if [%1] == [] goto continue
+  if [%1] == [--deploy-mode] (
+    set SPARK_SUBMIT_DEPLOY_MODE=%2
+  ) else if [%1] == [--properties-file] (
+    set SPARK_SUBMIT_PROPERTIES_FILE=%2
+  ) else if [%1] == [--driver-memory] (
+    set SPARK_SUBMIT_DRIVER_MEMORY=%2
+  ) else if [%1] == [--driver-library-path] (
+    set SPARK_SUBMIT_LIBRARY_PATH=%2
+  ) else if [%1] == [--driver-class-path] (
+    set SPARK_SUBMIT_CLASSPATH=%2
+  ) else if [%1] == [--driver-java-options] (
+    set SPARK_SUBMIT_OPTS=%2
+  )
+  shift
+goto loop
+:continue
+
+rem For client mode, the driver will be launched in the same JVM that launches
+rem SparkSubmit, so we may need to read the properties file for any extra class
+rem paths, library paths, java options and memory early on. Otherwise, it will
+rem be too late by the time the driver JVM has started.
+
+if [%SPARK_SUBMIT_DEPLOY_MODE%] == [client] (
+  if exist %SPARK_SUBMIT_PROPERTIES_FILE% (
+    rem Parse the properties file only if the special configs exist
+    for /f %%i in ('findstr /r /c:"^[\t ]*spark.driver.memory" /c:"^[\t ]*spark.driver.extra" ^
+      %SPARK_SUBMIT_PROPERTIES_FILE%') do (
+      set SPARK_SUBMIT_BOOTSTRAP_DRIVER=1
+    )
+  )
+)
+
+cmd /V /E /C %SPARK_HOME%\bin\spark-class.cmd org.apache.spark.deploy.SparkSubmit %ORIG_ARGS%

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

@@ -21,6 +21,7 @@ import scala.language.implicitConversions
 
 import java.io._
 import java.net.URI
+import java.util.Arrays
 import java.util.concurrent.atomic.AtomicInteger
 import java.util.{Properties, UUID}
 import java.util.UUID.randomUUID
@@ -1429,7 +1430,10 @@ object SparkContext extends Logging {
     simpleWritableConverter[Boolean, BooleanWritable](_.get)
 
   implicit def bytesWritableConverter(): WritableConverter[Array[Byte]] = {
-    simpleWritableConverter[Array[Byte], BytesWritable](_.getBytes)
+    simpleWritableConverter[Array[Byte], BytesWritable](bw =>
+      // getBytes method returns array which is longer then data to be returned
+      Arrays.copyOfRange(bw.getBytes, 0, bw.getLength)
+    )
   }
 
   implicit def stringWritableConverter(): WritableConverter[String] =

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

@@ -26,6 +26,8 @@ import scala.collection.JavaConversions._
 import javax.tools.{JavaFileObject, SimpleJavaFileObject, ToolProvider}
 import com.google.common.io.Files
 
+import org.apache.spark.util.Utils
+
 /**
  * Utilities for tests. Included in main codebase since it's used by multiple
  * projects.
@@ -42,8 +44,7 @@ private[spark] object TestUtils {
    * in order to avoid interference between tests.
    */
   def createJarWithClasses(classNames: Seq[String], value: String = ""): URL = {
-    val tempDir = Files.createTempDir()
-    tempDir.deleteOnExit()
+    val tempDir = Utils.createTempDir()
     val files = for (name <- classNames) yield createCompiledClass(name, tempDir, value)
     val jarFile = new File(tempDir, "testJar-%s.jar".format(System.currentTimeMillis()))
     createJar(files, jarFile)

core/src/main/scala/org/apache/spark/api/python/PythonRDD.scala

@@ -25,8 +25,6 @@ import java.util.{List => JList, ArrayList => JArrayList, Map => JMap, Collectio
 import scala.collection.JavaConversions._
 import scala.collection.mutable
 import scala.language.existentials
-import scala.reflect.ClassTag
-import scala.util.{Try, Success, Failure}
 
 import net.razorvine.pickle.{Pickler, Unpickler}
 
@@ -42,7 +40,7 @@ import org.apache.spark.rdd.RDD
 import org.apache.spark.util.Utils
 
 private[spark] class PythonRDD(
-    parent: RDD[_],
+    @transient parent: RDD[_],
     command: Array[Byte],
     envVars: JMap[String, String],
     pythonIncludes: JList[String],
@@ -55,9 +53,9 @@ private[spark] class PythonRDD(
   val bufferSize = conf.getInt("spark.buffer.size", 65536)
   val reuse_worker = conf.getBoolean("spark.python.worker.reuse", true)
 
-  override def getPartitions = parent.partitions
+  override def getPartitions = firstParent.partitions
 
-  override val partitioner = if (preservePartitoning) parent.partitioner else None
+  override val partitioner = if (preservePartitoning) firstParent.partitioner else None
 
   override def compute(split: Partition, context: TaskContext): Iterator[Array[Byte]] = {
     val startTime = System.currentTimeMillis
@@ -234,7 +232,7 @@ private[spark] class PythonRDD(
         dataOut.writeInt(command.length)
         dataOut.write(command)
         // Data values
-        PythonRDD.writeIteratorToStream(parent.iterator(split, context), dataOut)
+        PythonRDD.writeIteratorToStream(firstParent.iterator(split, context), dataOut)
         dataOut.writeInt(SpecialLengths.END_OF_DATA_SECTION)
         dataOut.flush()
       } catch {

core/src/main/scala/org/apache/spark/api/python/PythonWorkerFactory.scala

@@ -108,10 +108,12 @@ private[spark] class PythonWorkerFactory(pythonExec: String, envVars: Map[String
       serverSocket = new ServerSocket(0, 1, InetAddress.getByAddress(Array(127, 0, 0, 1)))
 
       // Create and start the worker
-      val pb = new ProcessBuilder(Seq(pythonExec, "-u", "-m", "pyspark.worker"))
+      val pb = new ProcessBuilder(Seq(pythonExec, "-m", "pyspark.worker"))
       val workerEnv = pb.environment()
       workerEnv.putAll(envVars)
       workerEnv.put("PYTHONPATH", pythonPath)
+      // This is equivalent to setting the -u flag; we use it because ipython doesn't support -u:
+      workerEnv.put("PYTHONUNBUFFERED", "YES")
       val worker = pb.start()
 
       // Redirect worker stdout and stderr
@@ -149,10 +151,12 @@ private[spark] class PythonWorkerFactory(pythonExec: String, envVars: Map[String
 
       try {
         // Create and start the daemon
-        val pb = new ProcessBuilder(Seq(pythonExec, "-u", "-m", "pyspark.daemon"))
+        val pb = new ProcessBuilder(Seq(pythonExec, "-m", "pyspark.daemon"))
         val workerEnv = pb.environment()
         workerEnv.putAll(envVars)
         workerEnv.put("PYTHONPATH", pythonPath)
+        // This is equivalent to setting the -u flag; we use it because ipython doesn't support -u:
+        workerEnv.put("PYTHONUNBUFFERED", "YES")
         daemon = pb.start()
 
         val in = new DataInputStream(daemon.getInputStream)