Additional improvements.

- Made driver & executor options consistent.
- Some doc improvements to YARN code.
- Handled special flags on YARN.

Author: pwendell

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

@@ -224,8 +224,8 @@ class SparkConf(loadDefaults: Boolean) extends Cloneable with Logging {
         throw new Exception(msg)
       }
       if (javaOpts.contains("-Xmx") || javaOpts.contains("-Xms")) {
-        val msg = s"$executorOptsKey is not allowed to alter memory settings (was '$javaOpts'). Please use " +
-          "spark.executor.memory."
+        val msg = s"$executorOptsKey is not allowed to alter memory settings (was '$javaOpts'). Use " +
+          "spark.executor.memory instead."
         throw new Exception(msg)
       }
     }

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

@@ -246,27 +246,28 @@ class SparkContext(config: SparkConf) extends Logging {
     .map(Utils.memoryStringToMb)
     .getOrElse(512)
 
-  // Environment variables to pass to our executors
-  private[spark] val executorEnvs = HashMap[String, String]()
+  // Environment variables to pass to our executors.
+  // NOTE: This should only be used for test related settings.
+  private[spark] val testExecutorEnvs = HashMap[String, String]()
 
   // Convert java options to env vars as a work around
   // since we can't set env vars directly in sbt.
-  for { (envKey, propKey) <- Seq(("SPARK_HOME", "spark.home"), ("SPARK_TESTING", "spark.testing"))
+  for { (envKey, propKey) <- Seq(("SPARK_TESTING", "spark.testing"))
     value <- Option(System.getenv(envKey)).orElse(Option(System.getProperty(propKey)))} {
-    executorEnvs(envKey) = value
+    testExecutorEnvs(envKey) = value
   }
   // The Mesos scheduler backend relies on this environment variable to set executor memory.
   // TODO: Set this only in the Mesos scheduler.
-  executorEnvs("SPARK_EXECUTOR_MEMORY") = executorMemory + "m"
-  executorEnvs ++= conf.getExecutorEnv
+  testExecutorEnvs("SPARK_EXECUTOR_MEMORY") = executorMemory + "m"
+  testExecutorEnvs ++= conf.getExecutorEnv
 
   // Set SPARK_USER for user who is running SparkContext.
   val sparkUser = Option {
     Option(System.getProperty("user.name")).getOrElse(System.getenv("SPARK_USER"))
   }.getOrElse {
     SparkContext.SPARK_UNKNOWN_USER
   }
-  executorEnvs("SPARK_USER") = sparkUser
+  testExecutorEnvs("SPARK_USER") = sparkUser
 
   // Create and start the scheduler
   private[spark] var taskScheduler = SparkContext.createTaskScheduler(this, master)

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

@@ -54,10 +54,10 @@ private class ClientActor(driverArgs: ClientArguments, conf: SparkConf) extends
         System.getenv().foreach{case (k, v) => env(k) = v}
 
         val mainClass = "org.apache.spark.deploy.worker.DriverWrapper"
-        val classPathEntries = sys.props.get("spark.driver.classPath").toSeq.flatMap { cp =>
+        val classPathEntries = sys.props.get("spark.driver.extraClassPath").toSeq.flatMap { cp =>
           cp.split(java.io.File.pathSeparator)
         }
-        val libraryPathEntries = sys.props.get("spark.driver.libraryPath").toSeq.flatMap { cp =>
+        val libraryPathEntries = sys.props.get("spark.driver.extraLibraryPath").toSeq.flatMap { cp =>
           cp.split(java.io.File.pathSeparator)
         }
         val javaOpts = sys.props.get("spark.driver.extraJavaOptions")

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

@@ -144,11 +144,11 @@ object SparkSubmit {
       new OptionAssigner(appArgs.master, ALL_CLUSTER_MGRS, false, sysProp = "spark.master"),
 
       new OptionAssigner(appArgs.driverExtraClassPath, STANDALONE | YARN, true,
-        sysProp = "spark.driver.classPath"),
+        sysProp = "spark.driver.extraClassPath"),
       new OptionAssigner(appArgs.driverExtraJavaOptions, STANDALONE | YARN, true,
-        sysProp = "spark.driver.javaOpts"),
+        sysProp = "spark.driver.extraJavaOptions"),
       new OptionAssigner(appArgs.driverExtraLibraryPath, STANDALONE | YARN, true,
-        sysProp = "spark.driver.libraryPath"),
+        sysProp = "spark.driver.extraLibraryPath"),
 
       new OptionAssigner(appArgs.driverMemory, YARN, true, clOption = "--driver-memory"),
       new OptionAssigner(appArgs.name, YARN, true, clOption = "--name"),

core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala

@@ -47,11 +47,16 @@ private[spark] class SparkDeploySchedulerBackend(
     val args = Seq(driverUrl, "{{EXECUTOR_ID}}", "{{HOSTNAME}}",
       "{{CORES}}", "{{WORKER_URL}}")
     val extraJavaOpts = sc.conf.getOption("spark.executor.extraJavaOptions")
+    val classPathEntries = sys.props.get("spark.executor.extraClassPath").toSeq.flatMap { cp =>
+      cp.split(java.io.File.pathSeparator)
+    }
+    val libraryPathEntries = sys.props.get("spark.executor.extraLibraryPath").toSeq.flatMap { cp =>
+      cp.split(java.io.File.pathSeparator)
+    }
 
-    // TODO (pwendell) LOOK AT THIS
     val command = Command(
-      "org.apache.spark.executor.CoarseGrainedExecutorBackend", args, sc.executorEnvs,
-      Seq(), Seq(), extraJavaOpts)
+      "org.apache.spark.executor.CoarseGrainedExecutorBackend", args, sc.testExecutorEnvs,
+      classPathEntries, libraryPathEntries, extraJavaOpts)
     val sparkHome = sc.getSparkHome()
     val appDesc = new ApplicationDescription(sc.appName, maxCores, sc.executorMemory, command,
       sparkHome, sc.ui.appUIAddress, sc.eventLogger.map(_.logDir))

core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala

@@ -111,7 +111,15 @@ private[spark] class CoarseMesosSchedulerBackend(
 
   def createCommand(offer: Offer, numCores: Int): CommandInfo = {
     val environment = Environment.newBuilder()
-    sc.executorEnvs.foreach { case (key, value) =>
+    val extraClassPath = conf.getOption("spark.executor.extraClassPath")
+    extraClassPath.foreach { cp =>
+      envrionment.addVariables(Environment.Variable.newBuilder().setName("SPARK_CLASSPATH").setValue(cp).build())
+    }
+    val extraJavaOpts = conf.getOption("spark.executor.extraJavaOptions", "")
+    val extraLibraryPath = conf.getOption("spark.executor.extraLibraryPath").map(p => s"-Djava.library.path=$p")
+    val extraOpts = Seq(extraJavaOpts, extraLibraryPath).flatten.mkString(" ")
+
+    sc.testExecutorEnvs.foreach { case (key, value) =>
       environment.addVariables(Environment.Variable.newBuilder()
         .setName(key)
         .setValue(value)
@@ -123,7 +131,7 @@ private[spark] class CoarseMesosSchedulerBackend(
       conf.get("spark.driver.host"),
       conf.get("spark.driver.port"),
       CoarseGrainedSchedulerBackend.ACTOR_NAME)
-    val extraOpts = conf.get("spark.executor.extraJavaOptions")
+
     val uri = conf.get("spark.executor.uri", null)
     if (uri == null) {
       val runScript = new File(sparkHome, "./bin/spark-class").getCanonicalPath

core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosSchedulerBackend.scala

@@ -90,7 +90,7 @@ private[spark] class MesosSchedulerBackend(
       "Spark home is not set; set it through the spark.home system " +
       "property, the SPARK_HOME environment variable or the SparkContext constructor"))
     val environment = Environment.newBuilder()
-    sc.executorEnvs.foreach { case (key, value) =>
+    sc.testExecutorEnvs.foreach { case (key, value) =>
       environment.addVariables(Environment.Variable.newBuilder()
         .setName(key)
         .setValue(value)

docs/configuration.md

@@ -655,6 +655,22 @@ Apart from these, the following properties are also available, and may be useful
     option.
   </td>
 </tr>
+<tr>
+  <td>spark.executor.extraClassPath</td>
+  <td>(none)</td>
+  <td>
+    Extra classpath entries to append to the classpath of executors. This exists primarily
+    for backwards-compatiblity with older versions of Spark. Users typically should not need
+    to set this option.
+  </td>
+</tr>
+<tr>
+  <td>spark.executor.extraLibraryPath</td>
+  <td>(none)</td>
+  <td>
+    Set a special library path to use when launching executor JVM's.
+  </td>
+</tr>
 
 </table>
 

yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/Client.scala

@@ -33,7 +33,9 @@ import org.apache.hadoop.yarn.util.{Apps, Records}
 
 import org.apache.spark.{Logging, SparkConf}
 
-
+/**
+ * Version of [[org.apache.spark.deploy.yarn.ClientBase]] tailored to YARN's alpha API.
+ */
 class Client(clientArgs: ClientArguments, hadoopConf: Configuration, spConf: SparkConf)
   extends YarnClientImpl with ClientBase with Logging {
 

yarn/common/src/main/scala/org/apache/spark/deploy/yarn/ClientBase.scala

@@ -39,8 +39,11 @@ import org.apache.spark.{Logging, SparkConf}
 
 /**
  * The entry point (starting in Client#main() and Client#run()) for launching Spark on YARN. The
- * Client submits an application to the global ResourceManager to launch Spark's ApplicationMaster,
- * which will launch a Spark master process and negotiate resources throughout its duration.
+ * Client submits an application to the YARN ResourceManager.
+ *
+ * Depending on the deployment mode this will launch one of two application master classes:
+ * 1. In standalone mode, it will launch an [[org.apache.spark.deploy.yarn.ApplicationMaster]] which embeds a driver.
+ * 2. In client mode, it will launch an [[org.apache.spark.deploy.yarn.ExecutorLauncher]].
  */
 trait ClientBase extends Logging {
   val args: ClientArguments
@@ -259,8 +262,8 @@ trait ClientBase extends Logging {
 
     val env = new HashMap[String, String]()
 
-    ClientBase.populateClasspath(yarnConf, sparkConf, localResources.contains(ClientBase.LOG4J_PROP),
-      env)
+    val extraCp = sparkConf.getOption("spark.driver.extraClassPath")
+    ClientBase.populateClasspath(yarnConf, sparkConf, localResources.contains(ClientBase.LOG4J_PROP), env, extraCp)
     env("SPARK_YARN_MODE") = "true"
     env("SPARK_YARN_STAGING_DIR") = stagingDir
     env("SPARK_USER") = UserGroupInformation.getCurrentUser().getShortUserName()
@@ -272,9 +275,6 @@ trait ClientBase extends Logging {
     // Allow users to specify some environment variables.
     Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV"))
 
-    // Add each SPARK_* key to the environment.
-    System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v }
-
     env
   }
 
@@ -427,29 +427,29 @@ object ClientBase {
     }
   }
 
-  def populateClasspath(conf: Configuration, sparkConf: SparkConf, addLog4j: Boolean, env: HashMap[String, String]) {
-    Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$())
+  def populateClasspath(conf: Configuration, sparkConf: SparkConf, addLog4j: Boolean, env: HashMap[String, String],
+      extraClassPath: Option[String] = None) {
+
+    /** Add entry to the classpath. */
+    def addClasspathEntry(entry: String) = pps.addToEnvironment(env, Environment.CLASSPATH.name, entry)
+    /** Add entry to the classpath. Interpreted as a path relative to the working directory. */
+    def addPwdClasspathEntry(path: String) = addClasspathEntry(Environment.PWD.$() + Path.SEPARATOR + entry)
+
+    extraClassPath.foreach(addClasspathEntry)
+
+    addClasspathEntry(Environment.PWD.$())
     // If log4j present, ensure ours overrides all others
-    if (addLog4j) {
-      Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() +
-        Path.SEPARATOR + LOG4J_PROP)
-    }
+    if (addLog4j) addPwdClasspathEntry(LOG4J_PROP)
     // Normally the users app.jar is last in case conflicts with spark jars
-    val userClasspathFirst = sparkConf.get("spark.yarn.user.classpath.first", "false")
-      .toBoolean
-    if (userClasspathFirst) {
-      Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() +
-        Path.SEPARATOR + APP_JAR)
-    }
-    Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() +
-      Path.SEPARATOR + SPARK_JAR)
-    ClientBase.populateHadoopClasspath(conf, env)
-
-    if (!userClasspathFirst) {
-      Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() +
-        Path.SEPARATOR + APP_JAR)
+    if (sparkConf.get("spark.yarn.user.classpath.first", "false").toBoolean) {
+      addPwdClasspathEntry(APP_JAR)
+      addPwdClasspathEntry(SPARK_JAR)
+      ClientBase.populateHadoopClasspath(conf, env)
+    } else {
+      addPwdClasspathEntry(SPARK_JAR)
+      ClientBase.populateHadoopClasspath(conf, env)
+      addPwdClasspathEntry(APP_JAR)
     }
-    Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() +
-      Path.SEPARATOR + "*")
+    addPwdClasspathEntry("*")
   }
 }