Spark版本2.1.1
一 Spark Submit本地解析
1.1 现象
提交命令：
spark-submit --master local[10] --driver-memory 30g --class app.package.AppClass app-1.0.jar
进程：
hadoop 225653 0.0 0.0 11256 364 ? S Aug24 0:00 bash /$spark-dir/bin/spark-class org.apache.spark.deploy.SparkSubmit --master local[10] --driver-memory 30g --class app.package.AppClass app-1.0.jar
hadoop 225654 0.0 0.0 34424 2860 ? Sl Aug24 0:00 /$jdk_dir/bin/java -Xmx128m -cp /spark-dir/jars/* org.apache.spark.launcher.Main org.apache.spark.deploy.SparkSubmit --master local[10] --driver-memory 30g --class app.package.AppClass app-1.0.jar
1.2 执行过程
1.2.1 脚本执行
-bash-4.1$ cat bin/spark-submit
#!/usr/bin/env bash
if [ -z "${SPARK_HOME}" ]; then
source "$(dirname "$0")"/find-spark-home
fi
# disable randomized hash for string in Python 3.3+
export PYTHONHASHSEED=0
exec "${SPARK_HOME}"/bin/spark-class org.apache.spark.deploy.SparkSubmit "$@"
注释：这里执行了另一个脚本spark-class，具体如下：
-bash-4.1$ cat bin/spark-class
...
build_command() {
"$RUNNER" -Xmx128m -cp "$LAUNCH_CLASSPATH" org.apache.spark.launcher.Main "$@"
printf "%d\0" $?
}
CMD=()
while IFS= read -d '' -r ARG; do
CMD+=("$ARG")
done < <(build_command "$@")
...
CMD=("${CMD[@]:0:$LAST}")
exec "${CMD[@]}"
注释：这里执行java class： org.apache.spark.launcher.Main，并传入参数，具体如下：
1.2.2 代码执行
org.apache.spark.launcher.Main
...
builder = new SparkSubmitCommandBuilder(help);
...
List<String> cmd = builder.buildCommand(env);
...
List<String> bashCmd = prepareBashCommand(cmd, env);
for (String c : bashCmd) {
System.out.print(c);
System.out.print('\0');
}
...
注释：其中会调用SparkSubmitCommandBuilder来生成Spark Submit命令，具体如下：
org.apache.spark.launcher.SparkSubmitCommandBuilder
...
private List<String> buildSparkSubmitCommand(Map<String, String> env)
...
addPermGenSizeOpt(cmd);
cmd.add("org.apache.spark.deploy.SparkSubmit");
cmd.addAll(buildSparkSubmitArgs());
return cmd;
...
注释：这里创建了本地命令，其中java class：org.apache.spark.deploy.SparkSubmit，具体如下：
org.apache.spark.deploy.SparkSubmit
def main(args: Array[String]): Unit = {
val appArgs = new SparkSubmitArguments(args) //parse command line parameter
if (appArgs.verbose) {
// scalastyle:off println
printStream.println(appArgs)
// scalastyle:on println
}
appArgs.action match {
case SparkSubmitAction.SUBMIT => submit(appArgs)
case SparkSubmitAction.KILL => kill(appArgs)
case SparkSubmitAction.REQUEST_STATUS => requestStatus(appArgs)
}
}
private def submit(args: SparkSubmitArguments): Unit = {
val (childArgs, childClasspath, sysProps, childMainClass) = prepareSubmitEnvironment(args) //merge all parameters from: command line, properties file, system property, etc...
def doRunMain(): Unit = {
...
runMain(childArgs, childClasspath, sysProps, childMainClass, args.verbose)
...
}
...
private[deploy] def prepareSubmitEnvironment(args: SparkSubmitArguments)
: (Seq[String], Seq[String], Map[String, String], String) = {
if (deployMode == CLIENT || isYarnCluster) {
childMainClass = args.mainClass
...
if (isYarnCluster) {
childMainClass = "org.apache.spark.deploy.yarn.Client"
...
private def runMain(
childArgs: Seq[String],
childClasspath: Seq[String],
sysProps: Map[String, String],
childMainClass: String,
verbose: Boolean): Unit = {
// scalastyle:off println
if (verbose) {
printStream.println(s"Main class:\n$childMainClass")
printStream.println(s"Arguments:\n${childArgs.mkString("\n")}")
printStream.println(s"System properties:\n${sysProps.mkString("\n")}")
printStream.println(s"Classpath elements:\n${childClasspath.mkString("\n")}")
printStream.println("\n")
}
// scalastyle:on println
val loader =
if (sysProps.getOrElse("spark.driver.userClassPathFirst", "false").toBoolean) {
new ChildFirstURLClassLoader(new Array[URL](0),
Thread.currentThread.getContextClassLoader)
} else {
new MutableURLClassLoader(new Array[URL](0),
Thread.currentThread.getContextClassLoader)
}
Thread.currentThread.setContextClassLoader(loader)
for (jar <- childClasspath) {
addJarToClasspath(jar, loader)
}
for ((key, value) <- sysProps) {
System.setProperty(key, value)
}
var mainClass: Class[_] = null
try {
mainClass = Utils.classForName(childMainClass)
} catch {
...
val mainMethod = mainClass.getMethod("main", new Array[String](0).getClass)
...
mainMethod.invoke(null, childArgs.toArray)
...
注释：这里首先会解析命令行参数，比如mainClass，准备运行环境包括System Property以及classpath等，然后使用一个新的classloader：ChildFirstURLClassLoader来加载用户的mainClass，然后反射调用mainClass的main方法，这样用户的app.package.AppClass的main方法就开始执行了。
org.apache.spark.SparkConf
class SparkConf(loadDefaults: Boolean) extends Cloneable with Logging with Serializable {
import SparkConf._
/** Create a SparkConf that loads defaults from system properties and the classpath */
def this() = this(true)
...
if (loadDefaults) {
loadFromSystemProperties(false)
}
private[spark] def loadFromSystemProperties(silent: Boolean): SparkConf = {
// Load any spark.* system properties
for ((key, value) <- Utils.getSystemProperties if key.startsWith("spark.")) {
set(key, value, silent)
}
this
}
注释：这里可以看到spark是怎样加载配置的
1.2.3 --verbose
spark-submit --master local[*] --class app.package.AppClass --jars /$other-dir/other.jar --driver-memory 1g --verbose app-1.0.jar
输出示例：
Main class:
app.package.AppClass
Arguments:
System properties:
spark.executor.logs.rolling.maxSize -> 1073741824
spark.driver.memory -> 1g
spark.driver.extraLibraryPath -> /$hadoop-dir/lib/native
spark.eventLog.enabled -> true
spark.eventLog.compress -> true
spark.executor.logs.rolling.time.interval -> daily
SPARK_SUBMIT -> true
spark.app.name -> app.package.AppClass
spark.driver.extraJavaOptions -> -XX:+PrintGCDetails -XX:+UseG1GC -XX:G1HeapRegionSize=32M -XX:+UseGCOverheadLimit -XX:+ExplicitGCInvokesConcurrent -XX:+HeapDumpOnOutOfMemoryError -XX:-UseCompressedClassPointers -XX:CompressedClassSpaceSize=3G -XX:+PrintGCTimeStamps -Xloggc:/export/Logs/hadoop/g1gc.log
spark.jars -> file:/$other-dir/other.jar
spark.sql.adaptive.enabled -> true
spark.submit.deployMode -> client
spark.executor.logs.rolling.maxRetainedFiles -> 10
spark.executor.extraClassPath -> /usr/lib/hadoop/lib/hadoop-lzo.jar
spark.eventLog.dir -> hdfs://myhdfs/spark/history
spark.master -> local[*]
spark.sql.crossJoin.enabled -> true
spark.driver.extraClassPath -> /usr/lib/hadoop/lib/hadoop-lzo.jar
Classpath elements:
file:/$other-dir/other.jar
file:/app-1.0.jar
启动时添加--verbose参数后，可以输出所有的运行时信息，有助于判断问题。