[SPARK-4461][YARN] pass extra java options to yarn application master

Currently, there is no way to pass yarn am specific java options. It cause some potential issues when reading classpath from hadoop configuration file. Hadoop configuration actually replace variables in its property with the system property passed in java options. How to specify the value depends on different hadoop distribution.

The new options are SPARK_YARN_JAVA_OPTS or spark.yarn.extraJavaOptions. I make it as spark global level, because typically we don't want user to specify this in their command line each time submitting spark job after it is setup in spark-defaults.conf.

In addition, with this new extra options enabled to be passed to AM, it provides more flexibility.

For example int the following valid mapred-site.xml file, we have the class path which specify values using system property. Hadoop can correctly handle it because it has java options passed in.

This is the example, currently spark will break due to hadoop.version is not passed in.
  <property>
    <name>mapreduce.application.classpath</name>
    <value>/etc/hadoop/${hadoop.version}/mapreduce/*</value>
  </property>

In the meantime, we cannot relies on  mapreduce.admin.map.child.java.opts in mapred-site.xml, because it has its own extra java options specified, which does not apply to Spark.

Author: Zhan Zhang <zhazhan@gmail.com>

Closes #3409 from zhzhan/Spark-4461 and squashes the following commits:

daec3d0 [Zhan Zhang] solve review comments
08f44a7 [Zhan Zhang] add warning in driver mode if spark.yarn.am.extraJavaOptions is configured
5a505d3 [Zhan Zhang] solve review comments
4ed43ad [Zhan Zhang] solve review comments
ad777ed [Zhan Zhang] Merge branch 'master' into Spark-4461
3e9e574 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
e3f9abe [Zhan Zhang] solve review comments
8963552 [Zhan Zhang] rebase
f8f6700 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
dea1692 [Zhan Zhang] change the option key name to client mode specific
90d5dff [Zhan Zhang] rebase
8ac9254 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
092a25f [Zhan Zhang] solve review comments
bc5a9ae [Zhan Zhang] solve review comments
782b014 [Zhan Zhang] add new configuration to docs/running-on-yarn.md and remove it from spark-defaults.conf.template
6faaa97 [Zhan Zhang] solve review comments
369863f [Zhan Zhang] clean up unnecessary var
733de9c [Zhan Zhang] Merge branch 'master' into Spark-4461
a68e7f0 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
864505a [Zhan Zhang] Add extra java options to be passed to Yarn application master
15830fc [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
685d911 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
03ebad3 [Zhan Zhang] Merge branch 'master' of https://github.com/zhzhan/spark
46d9e3d [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
ebb213a [Zhan Zhang] revert
b983ef3 [Zhan Zhang] test
c4efb9b [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
779d67b [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
4daae6d [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
12e1be5 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
ce0ca7b [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
93f3081 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
3764505 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
a9d372b [Zhan Zhang] Merge branch 'master' of https://github.com/zhzhan/spark
a00f60f [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
497b0f4 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
4a2e36d [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
a72c0d4 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
301eb4a [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
cedcc6f [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
adf4924 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
d10bf00 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
7e0cc36 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
68deb11 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
3ee3b2b [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
2b0d513 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
1ccd7cc [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
af9feb9 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
e4c1982 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
921e914 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
789ea21 [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
cb53a2c [Zhan Zhang] Merge branch 'master' of https://github.com/apache/spark
f6a8a40 [Zhan Zhang] revert
ba14f28 [Zhan Zhang] test

1 files changed, 19 insertions, 0 deletions

diff --git a/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientBase.scala b/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientBase.scala
index f95d723791..5f0c67f05c 100644
--- a/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientBase.scala
+++ b/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientBase.scala
@@ -352,12 +352,31 @@ private[spark] trait ClientBase extends Logging {
     if (isLaunchingDriver) {
       sparkConf.getOption("spark.driver.extraJavaOptions")
         .orElse(sys.env.get("SPARK_JAVA_OPTS"))
+        .map(Utils.splitCommandString).getOrElse(Seq.empty)
         .foreach(opts => javaOpts += opts)
       val libraryPaths = Seq(sys.props.get("spark.driver.extraLibraryPath"),
         sys.props.get("spark.driver.libraryPath")).flatten
       if (libraryPaths.nonEmpty) {
         prefixEnv = Some(Utils.libraryPathEnvPrefix(libraryPaths))
       }
+      if (sparkConf.getOption("spark.yarn.am.extraJavaOptions").isDefined) {
+        logWarning("spark.yarn.am.extraJavaOptions will not take effect in cluster mode")
+      }
+    } else {
+      // Validate and include yarn am specific java options in yarn-client mode.
+      val amOptsKey = "spark.yarn.am.extraJavaOptions"
+      val amOpts = sparkConf.getOption(amOptsKey)
+      amOpts.foreach { opts =>
+        if (opts.contains("-Dspark")) {
+          val msg = s"$amOptsKey is not allowed to set Spark options (was '$opts'). "
+          throw new SparkException(msg)
+        }
+        if (opts.contains("-Xmx") || opts.contains("-Xms")) {
+          val msg = s"$amOptsKey is not allowed to alter memory settings (was '$opts')."
+          throw new SparkException(msg)
+        }
+        javaOpts ++= Utils.splitCommandString(opts)
+      }
     }
 
     // For log4j configuration to reference