diff options
Diffstat (limited to 'core/src/main/scala/org/apache/spark/deploy/master/Master.scala')
| -rw-r--r-- | core/src/main/scala/org/apache/spark/deploy/master/Master.scala | 41 |
1 files changed, 23 insertions, 18 deletions
diff --git a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala index eebd0794b8..7b696cfcca 100644 --- a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala +++ b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala @@ -29,7 +29,7 @@ import akka.pattern.ask import akka.remote.{DisassociatedEvent, RemotingLifecycleEvent} import akka.serialization.SerializationExtension -import org.apache.spark.{Logging, SparkException} +import org.apache.spark.{SparkConf, SparkContext, Logging, SparkException} import org.apache.spark.deploy.{ApplicationDescription, ExecutorState} import org.apache.spark.deploy.DeployMessages._ import org.apache.spark.deploy.master.MasterMessages._ @@ -38,14 +38,16 @@ import org.apache.spark.metrics.MetricsSystem import org.apache.spark.util.{AkkaUtils, Utils} private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Actor with Logging { - import context.dispatcher + import context.dispatcher // to use Akka's scheduler.schedule() + + val conf = new SparkConf val DATE_FORMAT = new SimpleDateFormat("yyyyMMddHHmmss") // For application IDs - val WORKER_TIMEOUT = System.getProperty("spark.worker.timeout", "60").toLong * 1000 - val RETAINED_APPLICATIONS = System.getProperty("spark.deploy.retainedApplications", "200").toInt - val REAPER_ITERATIONS = System.getProperty("spark.dead.worker.persistence", "15").toInt - val RECOVERY_DIR = System.getProperty("spark.deploy.recoveryDirectory", "") - val RECOVERY_MODE = System.getProperty("spark.deploy.recoveryMode", "NONE") + val WORKER_TIMEOUT = conf.get("spark.worker.timeout", "60").toLong * 1000 + val RETAINED_APPLICATIONS = conf.get("spark.deploy.retainedApplications", "200").toInt + val REAPER_ITERATIONS = conf.get("spark.dead.worker.persistence", "15").toInt + val RECOVERY_DIR = conf.get("spark.deploy.recoveryDirectory", "") + val RECOVERY_MODE = conf.get("spark.deploy.recoveryMode", "NONE") var nextAppNumber = 0 val workers = new HashSet[WorkerInfo] @@ -63,8 +65,8 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act Utils.checkHost(host, "Expected hostname") - val masterMetricsSystem = MetricsSystem.createMetricsSystem("master") - val applicationMetricsSystem = MetricsSystem.createMetricsSystem("applications") + val masterMetricsSystem = MetricsSystem.createMetricsSystem("master", conf) + val applicationMetricsSystem = MetricsSystem.createMetricsSystem("applications", conf) val masterSource = new MasterSource(this) val webUi = new MasterWebUI(this, webUiPort) @@ -86,7 +88,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act // As a temporary workaround before better ways of configuring memory, we allow users to set // a flag that will perform round-robin scheduling across the nodes (spreading out each app // among all the nodes) instead of trying to consolidate each app onto a small # of nodes. - val spreadOutApps = System.getProperty("spark.deploy.spreadOut", "true").toBoolean + val spreadOutApps = conf.get("spark.deploy.spreadOut", "true").toBoolean override def preStart() { logInfo("Starting Spark master at " + masterUrl) @@ -103,7 +105,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act persistenceEngine = RECOVERY_MODE match { case "ZOOKEEPER" => logInfo("Persisting recovery state to ZooKeeper") - new ZooKeeperPersistenceEngine(SerializationExtension(context.system)) + new ZooKeeperPersistenceEngine(SerializationExtension(context.system), conf) case "FILESYSTEM" => logInfo("Persisting recovery state to directory: " + RECOVERY_DIR) new FileSystemPersistenceEngine(RECOVERY_DIR, SerializationExtension(context.system)) @@ -113,7 +115,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act leaderElectionAgent = RECOVERY_MODE match { case "ZOOKEEPER" => - context.actorOf(Props(classOf[ZooKeeperLeaderElectionAgent], self, masterUrl)) + context.actorOf(Props(classOf[ZooKeeperLeaderElectionAgent], self, masterUrl, conf)) case _ => context.actorOf(Props(classOf[MonarchyLeaderAgent], self)) } @@ -495,7 +497,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act removeWorker(worker) } else { if (worker.lastHeartbeat < currentTime - ((REAPER_ITERATIONS + 1) * WORKER_TIMEOUT)) - workers -= worker // we've seen this DEAD worker in the UI, etc. for long enough; cull it + workers -= worker // we've seen this DEAD worker in the UI, etc. for long enough; cull it } } } @@ -507,8 +509,9 @@ private[spark] object Master { val sparkUrlRegex = "spark://([^:]+):([0-9]+)".r def main(argStrings: Array[String]) { - val args = new MasterArguments(argStrings) - val (actorSystem, _, _) = startSystemAndActor(args.host, args.port, args.webUiPort) + val conf = new SparkConf + val args = new MasterArguments(argStrings, conf) + val (actorSystem, _, _) = startSystemAndActor(args.host, args.port, args.webUiPort, conf) actorSystem.awaitTermination() } @@ -522,10 +525,12 @@ private[spark] object Master { } } - def startSystemAndActor(host: String, port: Int, webUiPort: Int): (ActorSystem, Int, Int) = { - val (actorSystem, boundPort) = AkkaUtils.createActorSystem(systemName, host, port) + def startSystemAndActor(host: String, port: Int, webUiPort: Int, conf: SparkConf) + : (ActorSystem, Int, Int) = + { + val (actorSystem, boundPort) = AkkaUtils.createActorSystem(systemName, host, port, conf = conf) val actor = actorSystem.actorOf(Props(classOf[Master], host, boundPort, webUiPort), actorName) - val timeout = AkkaUtils.askTimeout + val timeout = AkkaUtils.askTimeout(conf) val respFuture = actor.ask(RequestWebUIPort)(timeout) val resp = Await.result(respFuture, timeout).asInstanceOf[WebUIPortResponse] (actorSystem, boundPort, resp.webUIBoundPort) |