static {
specialClasses.put("org.apache.spark.repl.Main", "spark-shell");
specialClasses.put(
"org.apache.spark.sql.hive.thriftserver.SparkSQLCLIDriver", "spark-internal");
specialClasses.put(
"org.apache.spark.sql.hive.thriftserver.HiveThriftServer2", "spark-internal");
}
private boolean isClientMode(Map<String, String> userProps) {
String userMaster = firstNonEmpty(master, userProps.get(SparkLauncher.SPARK_MASTER));
String userDeployMode = firstNonEmpty(deployMode, userProps.get(SparkLauncher.DEPLOY_MODE));
// Default master is "local[*]", so assume client mode in that case
return userMaster == null
|| "client".equals(userDeployMode)
|| (!userMaster.equals("yarn-cluster") && userDeployMode == null);
}
private List<String> buildSparkSubmitCommand(Map<String, String> env) throws IOException {
// Load the properties file and check whether spark-submit will be running the app's driver
// or just launching a cluster app. When running the driver, the JVM's argument will be
// modified to cover the driver's configuration.
Map<String, String> config = getEffectiveConfig();
boolean isClientMode = isClientMode(config);
String extraClassPath = isClientMode ? config.get(SparkLauncher.DRIVER_EXTRA_CLASSPATH) : null;
List<String> cmd = buildJavaCommand(extraClassPath);
// Take Thrift Server as daemon
if (isThriftServer(mainClass)) {
addOptionString(cmd, System.getenv("SPARK_DAEMON_JAVA_OPTS"));
}
addOptionString(cmd, System.getenv("SPARK_SUBMIT_OPTS"));
addOptionString(cmd, System.getenv("SPARK_JAVA_OPTS"));
if (isClientMode) {
// Figuring out where the memory value come from is a little tricky due to precedence.
// Precedence is observed in the following order:
// - explicit configuration (setConf()), which also covers --driver-memory cli argument.
// - properties file.
// - SPARK_DRIVER_MEMORY env variable
// - SPARK_MEM env variable
// - default value (1g)
// Take Thrift Server as daemon
String tsMemory = isThriftServer(mainClass) ? System.getenv("SPARK_DAEMON_MEMORY") : null;
String memory =
firstNonEmpty(
tsMemory,
config.get(SparkLauncher.DRIVER_MEMORY),
System.getenv("SPARK_DRIVER_MEMORY"),
System.getenv("SPARK_MEM"),
DEFAULT_MEM);
cmd.add("-Xms" + memory);
cmd.add("-Xmx" + memory);
addOptionString(cmd, config.get(SparkLauncher.DRIVER_EXTRA_JAVA_OPTIONS));
mergeEnvPathList(
env, getLibPathEnvName(), config.get(SparkLauncher.DRIVER_EXTRA_LIBRARY_PATH));
}
addPermGenSizeOpt(cmd);
cmd.add("org.apache.spark.deploy.SparkSubmit");
cmd.addAll(buildSparkSubmitArgs());
return cmd;
}
private void constructEnvVarArgs(Map<String, String> env, String submitArgsEnvVariable)
throws IOException {
mergeEnvPathList(
env,
getLibPathEnvName(),
getEffectiveConfig().get(SparkLauncher.DRIVER_EXTRA_LIBRARY_PATH));
StringBuilder submitArgs = new StringBuilder();
for (String arg : buildSparkSubmitArgs()) {
if (submitArgs.length() > 0) {
submitArgs.append(" ");
}
submitArgs.append(quoteForCommandString(arg));
}
env.put(submitArgsEnvVariable, submitArgs.toString());
}
private List<String> buildSparkRCommand(Map<String, String> env) throws IOException {
if (!appArgs.isEmpty() && appArgs.get(0).endsWith(".R")) {
System.err.println(
"Running R applications through 'sparkR' is not supported as of Spark 2.0.\n"
+ "Use ./bin/spark-submit <R file>");
System.exit(-1);
}
// When launching the SparkR shell, store the spark-submit arguments in the SPARKR_SUBMIT_ARGS
// env variable.
constructEnvVarArgs(env, "SPARKR_SUBMIT_ARGS");
// Set shell.R as R_PROFILE_USER to load the SparkR package when the shell comes up.
String sparkHome = System.getenv("SPARK_HOME");
env.put(
"R_PROFILE_USER",
join(File.separator, sparkHome, "R", "lib", "SparkR", "profile", "shell.R"));
List<String> args = new ArrayList<>();
args.add(firstNonEmpty(System.getenv("SPARKR_DRIVER_R"), "R"));
return args;
}