diff options
author | WangTaoTheTonic <barneystinson@aliyun.com> | 2015-01-13 09:43:48 -0800 |
---|---|---|
committer | Andrew Or <andrew@databricks.com> | 2015-01-13 09:50:14 -0800 |
commit | 9dea64e53ad8df8a3160c0f4010811af1e73dd6f (patch) | |
tree | 5a48e84a492575ac97fff04d9f29c433eb479d39 /yarn/src | |
parent | f7741a9a72fef23c46f0ad9e1bd16b150967d816 (diff) | |
download | spark-9dea64e53ad8df8a3160c0f4010811af1e73dd6f.tar.gz spark-9dea64e53ad8df8a3160c0f4010811af1e73dd6f.tar.bz2 spark-9dea64e53ad8df8a3160c0f4010811af1e73dd6f.zip |
[SPARK-4697][YARN]System properties should override environment variables
I found some arguments in yarn module take environment variables before system properties while the latter override the former in core module.
Author: WangTaoTheTonic <barneystinson@aliyun.com>
Author: WangTao <barneystinson@aliyun.com>
Closes #3557 from WangTaoTheTonic/SPARK4697 and squashes the following commits:
836b9ef [WangTaoTheTonic] fix type mismatch
e3e486a [WangTaoTheTonic] remove the comma
1262d57 [WangTaoTheTonic] handle spark.app.name and SPARK_YARN_APP_NAME in SparkSubmitArguments
bee9447 [WangTaoTheTonic] wrong brace
81833bb [WangTaoTheTonic] rebase
40934b4 [WangTaoTheTonic] just switch blocks
5f43f45 [WangTao] System property can override environment variable
Diffstat (limited to 'yarn/src')
-rw-r--r-- | yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientArguments.scala | 4 | ||||
-rw-r--r-- | yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala | 19 |
2 files changed, 13 insertions, 10 deletions
diff --git a/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientArguments.scala b/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientArguments.scala index fdbf9f8eed..461a9ccd3c 100644 --- a/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientArguments.scala +++ b/yarn/src/main/scala/org/apache/spark/deploy/yarn/ClientArguments.scala @@ -64,12 +64,12 @@ private[spark] class ClientArguments(args: Array[String], sparkConf: SparkConf) // For backward compatibility, SPARK_YARN_DIST_{ARCHIVES/FILES} should be resolved to hdfs://, // while spark.yarn.dist.{archives/files} should be resolved to file:// (SPARK-2051). files = Option(files) - .orElse(sys.env.get("SPARK_YARN_DIST_FILES")) .orElse(sparkConf.getOption("spark.yarn.dist.files").map(p => Utils.resolveURIs(p))) + .orElse(sys.env.get("SPARK_YARN_DIST_FILES")) .orNull archives = Option(archives) - .orElse(sys.env.get("SPARK_YARN_DIST_ARCHIVES")) .orElse(sparkConf.getOption("spark.yarn.dist.archives").map(p => Utils.resolveURIs(p))) + .orElse(sys.env.get("SPARK_YARN_DIST_ARCHIVES")) .orNull // If dynamic allocation is enabled, start at the max number of executors if (isDynamicAllocationEnabled) { diff --git a/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala b/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala index f99291553b..690f927e93 100644 --- a/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala +++ b/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala @@ -74,8 +74,7 @@ private[spark] class YarnClientSchedulerBackend( ("--executor-memory", "SPARK_EXECUTOR_MEMORY", "spark.executor.memory"), ("--executor-cores", "SPARK_WORKER_CORES", "spark.executor.cores"), ("--executor-cores", "SPARK_EXECUTOR_CORES", "spark.executor.cores"), - ("--queue", "SPARK_YARN_QUEUE", "spark.yarn.queue"), - ("--name", "SPARK_YARN_APP_NAME", "spark.app.name") + ("--queue", "SPARK_YARN_QUEUE", "spark.yarn.queue") ) // Warn against the following deprecated environment variables: env var -> suggestion val deprecatedEnvVars = Map( @@ -86,18 +85,22 @@ private[spark] class YarnClientSchedulerBackend( // Do the same for deprecated properties: property -> suggestion val deprecatedProps = Map("spark.master.memory" -> "--driver-memory through spark-submit") optionTuples.foreach { case (optionName, envVar, sparkProp) => - if (System.getenv(envVar) != null) { - extraArgs += (optionName, System.getenv(envVar)) - if (deprecatedEnvVars.contains(envVar)) { - logWarning(s"NOTE: $envVar is deprecated. Use ${deprecatedEnvVars(envVar)} instead.") - } - } else if (sc.getConf.contains(sparkProp)) { + if (sc.getConf.contains(sparkProp)) { extraArgs += (optionName, sc.getConf.get(sparkProp)) if (deprecatedProps.contains(sparkProp)) { logWarning(s"NOTE: $sparkProp is deprecated. Use ${deprecatedProps(sparkProp)} instead.") } + } else if (System.getenv(envVar) != null) { + extraArgs += (optionName, System.getenv(envVar)) + if (deprecatedEnvVars.contains(envVar)) { + logWarning(s"NOTE: $envVar is deprecated. Use ${deprecatedEnvVars(envVar)} instead.") + } } } + // The app name is a special case because "spark.app.name" is required of all applications. + // As a result, the corresponding "SPARK_YARN_APP_NAME" is already handled preemptively in + // SparkSubmitArguments if "spark.app.name" is not explicitly set by the user. (SPARK-5222) + sc.getConf.getOption("spark.app.name").foreach(v => extraArgs += ("--name", v)) extraArgs } |