diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala b/core/src/main/scala/org/apache/spark/SparkContext.scala
index 66c226e4913621e11fd11a63898d026d35fa2f5e..1647d904a22474094ed22ae6e02017120f16748e 100644
--- a/core/src/main/scala/org/apache/spark/SparkContext.scala
+++ b/core/src/main/scala/org/apache/spark/SparkContext.scala
@@ -677,10 +677,10 @@ class SparkContext(
         key = uri.getScheme match {
           // A JAR file which exists only on the driver node
           case null | "file" =>
-            if (SparkHadoopUtil.get.isYarnMode()) {
-              // In order for this to work on yarn the user must specify the --addjars option to
-              // the client to upload the file into the distributed cache to make it show up in the
-              // current working directory.
+            if (SparkHadoopUtil.get.isYarnMode() && master == "yarn-standalone") {
+              // In order for this to work in yarn standalone mode the user must specify the 
+              // --addjars option to the client to upload the file into the distributed cache 
+              // of the AM to make it show up in the current working directory.
               val fileName = new Path(uri.getPath).getName()
               try {
                 env.httpFileServer.addJar(new File(fileName))
diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md
index b20627010798a9fe8461a661c281f12847593eab..3bd62646bab060eba316406fa9a1c7a4983a31b1 100644
--- a/docs/running-on-yarn.md
+++ b/docs/running-on-yarn.md
@@ -101,7 +101,19 @@ With this mode, your application is actually run on the remote machine where the
 
 With yarn-client mode, the application will be launched locally. Just like running application or spark-shell on Local / Mesos / Standalone mode. The launch method is also the similar with them, just make sure that when you need to specify a master url, use "yarn-client" instead. And you also need to export the env value for SPARK_JAR and SPARK_YARN_APP_JAR
 
-In order to tune worker core/number/memory etc. You need to export SPARK_WORKER_CORES, SPARK_WORKER_MEMORY, SPARK_WORKER_INSTANCES e.g. by ./conf/spark-env.sh
+Configuration in yarn-client mode:
+
+In order to tune worker core/number/memory etc. You need to export environment variables or add them to the spark configuration file (./conf/spark_env.sh). The following are the list of options.
+
+* `SPARK_YARN_APP_JAR`, Path to your application's JAR file (required)
+* `SPARK_WORKER_INSTANCES`, Number of workers to start (Default: 2)
+* `SPARK_WORKER_CORES`, Number of cores for the workers (Default: 1).
+* `SPARK_WORKER_MEMORY`, Memory per Worker (e.g. 1000M, 2G) (Default: 1G)
+* `SPARK_MASTER_MEMORY`, Memory for Master (e.g. 1000M, 2G) (Default: 512 Mb)
+* `SPARK_YARN_APP_NAME`, The name of your application (Default: Spark)
+* `SPARK_YARN_QUEUE`, The hadoop queue to use for allocation requests (Default: 'default')
+* `SPARK_YARN_DIST_FILES`, Comma separated list of files to be distributed with the job.
+* `SPARK_YARN_DIST_ARCHIVES`, Comma separated list of archives to be distributed with the job.
 
 For example:
 
@@ -114,7 +126,6 @@ For example:
     SPARK_YARN_APP_JAR=examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \
     MASTER=yarn-client ./bin/spark-shell
 
-You can also send extra files to yarn cluster for worker to use by exporting SPARK_YARN_DIST_FILES=file1,file2... etc.
 
 # Building Spark for Hadoop/YARN 2.2.x
 
diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
index 2bb11e54c549af037fc75a2dd92300e3588828b4..2e46d750c4a3801f5d4cc381dcd2f8dd759561a3 100644
--- a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
+++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
@@ -127,14 +127,13 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration,
     // local dirs, so lets check both. We assume one of the 2 is set.
     // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X
     val localDirs = Option(System.getenv("YARN_LOCAL_DIRS"))
-      .getOrElse(Option(System.getenv("LOCAL_DIRS"))
-      .getOrElse(""))
-
-    if (localDirs.isEmpty()) {
-      throw new Exception("Yarn Local dirs can't be empty")
+      .orElse(Option(System.getenv("LOCAL_DIRS")))
+ 
+    localDirs match {
+      case None => throw new Exception("Yarn Local dirs can't be empty")
+      case Some(l) => l
     }
-    localDirs
-  }
+  } 
 
   private def getApplicationAttemptId(): ApplicationAttemptId = {
     val envs = System.getenv()
diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
index ddfec1a4ac6728e8dd0030a6b5fcf164a1dff42c..62b20b8fbaf19da6cc13d2eb531e4ec78aa6f199 100644
--- a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
+++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
@@ -76,6 +76,10 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
 
   def run() {
 
+    // Setup the directories so things go to yarn approved directories rather
+    // then user specified and /tmp.
+    System.setProperty("spark.local.dir", getLocalDirs())
+
     appAttemptId = getApplicationAttemptId()
     resourceManager = registerWithResourceManager()
     val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster()
@@ -103,10 +107,12 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
     // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse.
 
     val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000)
-    // must be <= timeoutInterval/ 2.
-    // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM.
-    // so atleast 1 minute or timeoutInterval / 10 - whichever is higher.
-    val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval/ 10, 60000L))
+    // we want to be reasonably responsive without causing too many requests to RM.
+    val schedulerInterval =
+      System.getProperty("spark.yarn.scheduler.heartbeat.interval-ms", "5000").toLong
+    // must be <= timeoutInterval / 2.
+    val interval = math.min(timeoutInterval / 2, schedulerInterval)
+
     reporterThread = launchReporterThread(interval)
 
     // Wait for the reporter thread to Finish.
@@ -119,6 +125,20 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
     System.exit(0)
   }
 
+  /** Get the Yarn approved local directories. */
+  private def getLocalDirs(): String = {
+    // Hadoop 0.23 and 2.x have different Environment variable names for the
+    // local dirs, so lets check both. We assume one of the 2 is set.
+    // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X
+    val localDirs = Option(System.getenv("YARN_LOCAL_DIRS"))
+      .orElse(Option(System.getenv("LOCAL_DIRS")))
+
+    localDirs match {
+      case None => throw new Exception("Yarn Local dirs can't be empty")
+      case Some(l) => l
+    }
+  }
+
   private def getApplicationAttemptId(): ApplicationAttemptId = {
     val envs = System.getenv()
     val containerIdString = envs.get(ApplicationConstants.AM_CONTAINER_ID_ENV)
diff --git a/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala b/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala
index 4b1b5da048df4d801dacb24f7df18245e98735ac..22e55e0c60647978d4543d14f78444fdfa0d2e8d 100644
--- a/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala
+++ b/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala
@@ -22,6 +22,8 @@ import org.apache.spark.{SparkException, Logging, SparkContext}
 import org.apache.spark.deploy.yarn.{Client, ClientArguments}
 import org.apache.spark.scheduler.TaskSchedulerImpl
 
+import scala.collection.mutable.ArrayBuffer
+
 private[spark] class YarnClientSchedulerBackend(
     scheduler: TaskSchedulerImpl,
     sc: SparkContext)
@@ -31,45 +33,47 @@ private[spark] class YarnClientSchedulerBackend(
   var client: Client = null
   var appId: ApplicationId = null
 
+  private[spark] def addArg(optionName: String, optionalParam: String, arrayBuf: ArrayBuffer[String]) {
+    Option(System.getenv(optionalParam)) foreach {
+      optParam => {
+        arrayBuf += (optionName, optParam)
+      }
+    }
+  }
+
   override def start() {
     super.start()
 
-    val defalutWorkerCores = "2"
-    val defalutWorkerMemory = "512m"
-    val defaultWorkerNumber = "1"
-
     val userJar = System.getenv("SPARK_YARN_APP_JAR")
-    val distFiles = System.getenv("SPARK_YARN_DIST_FILES")
-    var workerCores = System.getenv("SPARK_WORKER_CORES")
-    var workerMemory = System.getenv("SPARK_WORKER_MEMORY")
-    var workerNumber = System.getenv("SPARK_WORKER_INSTANCES")
-
     if (userJar == null)
       throw new SparkException("env SPARK_YARN_APP_JAR is not set")
 
-    if (workerCores == null)
-      workerCores = defalutWorkerCores
-    if (workerMemory == null)
-      workerMemory = defalutWorkerMemory
-    if (workerNumber == null)
-      workerNumber = defaultWorkerNumber
-
     val driverHost = conf.get("spark.driver.host")
     val driverPort = conf.get("spark.driver.port")
     val hostport = driverHost + ":" + driverPort
 
-    val argsArray = Array[String](
+    val argsArrayBuf = new ArrayBuffer[String]()
+    argsArrayBuf += (
       "--class", "notused",
       "--jar", userJar,
       "--args", hostport,
-      "--worker-memory", workerMemory,
-      "--worker-cores", workerCores,
-      "--num-workers", workerNumber,
-      "--master-class", "org.apache.spark.deploy.yarn.WorkerLauncher",
-      "--files", distFiles
+      "--master-class", "org.apache.spark.deploy.yarn.WorkerLauncher"
     )
 
-    val args = new ClientArguments(argsArray, conf)
+    // process any optional arguments, use the defaults already defined in ClientArguments 
+    // if things aren't specified
+    Map("--master-memory" -> "SPARK_MASTER_MEMORY",
+      "--num-workers" -> "SPARK_WORKER_INSTANCES",
+      "--worker-memory" -> "SPARK_WORKER_MEMORY",
+      "--worker-cores" -> "SPARK_WORKER_CORES",
+      "--queue" -> "SPARK_YARN_QUEUE",
+      "--name" -> "SPARK_YARN_APP_NAME",
+      "--files" -> "SPARK_YARN_DIST_FILES",
+      "--archives" -> "SPARK_YARN_DIST_ARCHIVES")
+    .foreach { case (optName, optParam) => addArg(optName, optParam, argsArrayBuf) }
+      
+    logDebug("ClientArguments called with: " + argsArrayBuf)
+    val args = new ClientArguments(argsArrayBuf.toArray, conf)
     client = new Client(args, conf)
     appId = client.runApp()
     waitForApp()
diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
index 69ae14ce8385cfb3b978481d90b2681d02a2ae80..4b777d5fa7a283e78744923dce484b13b1cc1431 100644
--- a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
+++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala
@@ -116,14 +116,13 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration,
     // local dirs, so lets check both. We assume one of the 2 is set.
     // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X
     val localDirs = Option(System.getenv("YARN_LOCAL_DIRS"))
-      .getOrElse(Option(System.getenv("LOCAL_DIRS"))
-      .getOrElse(""))
-
-    if (localDirs.isEmpty()) {
-      throw new Exception("Yarn Local dirs can't be empty")
+      .orElse(Option(System.getenv("LOCAL_DIRS")))
+ 
+    localDirs match {
+      case None => throw new Exception("Yarn Local dirs can't be empty")
+      case Some(l) => l
     }
-    localDirs
-  }
+  } 
 
   private def getApplicationAttemptId(): ApplicationAttemptId = {
     val envs = System.getenv()
diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
index be323d77835a8892eb3e481eba83f31f7dc3e8b9..952e963389c0aa7f1b9e3a3a4557ff4736920964 100644
--- a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
+++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
@@ -99,6 +99,7 @@ class Client(args: ClientArguments, conf: Configuration, sparkConf: SparkConf)
     appContext.setApplicationName(args.appName)
     appContext.setQueue(args.amQueue)
     appContext.setAMContainerSpec(amContainer)
+    appContext.setApplicationType("SPARK")
 
     // Memory for the ApplicationMaster.
     val memoryResource = Records.newRecord(classOf[Resource]).asInstanceOf[Resource]
diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
index 49248a8516b9cef6941524500965354d9ac30d08..78353224fa4b8b51aea9ba6d56ab04c2e1a479cd 100644
--- a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
+++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala
@@ -78,6 +78,10 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
 
   def run() {
 
+    // Setup the directories so things go to yarn approved directories rather
+    // then user specified and /tmp.
+    System.setProperty("spark.local.dir", getLocalDirs())
+
     amClient = AMRMClient.createAMRMClient()
     amClient.init(yarnConf)
     amClient.start()
@@ -94,10 +98,12 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
     // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse.
 
     val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000)
-    // must be <= timeoutInterval/ 2.
-    // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM.
-    // so atleast 1 minute or timeoutInterval / 10 - whichever is higher.
-    val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval / 10, 60000L))
+    // we want to be reasonably responsive without causing too many requests to RM.
+    val schedulerInterval =
+      System.getProperty("spark.yarn.scheduler.heartbeat.interval-ms", "5000").toLong
+    // must be <= timeoutInterval / 2.
+    val interval = math.min(timeoutInterval / 2, schedulerInterval)
+
     reporterThread = launchReporterThread(interval)
 
     // Wait for the reporter thread to Finish.
@@ -110,6 +116,20 @@ class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration, spar
     System.exit(0)
   }
 
+  /** Get the Yarn approved local directories. */
+  private def getLocalDirs(): String = {
+    // Hadoop 0.23 and 2.x have different Environment variable names for the
+    // local dirs, so lets check both. We assume one of the 2 is set.
+    // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X
+    val localDirs = Option(System.getenv("YARN_LOCAL_DIRS"))
+      .orElse(Option(System.getenv("LOCAL_DIRS")))
+ 
+    localDirs match {
+      case None => throw new Exception("Yarn Local dirs can't be empty")
+      case Some(l) => l
+    }
+  } 
+
   private def getApplicationAttemptId(): ApplicationAttemptId = {
     val envs = System.getenv()
     val containerIdString = envs.get(ApplicationConstants.Environment.CONTAINER_ID.name())