Why spark on yarn applicationmaster cannot get a proper resourcemanager address from yarnconfiguration?
Hello, I tried to submit a spark job to yarn cluster, there is an error occured with those messages: [root@saturn00 bin]# ./spark-submit --class SparkHiveJoin --master yarn-cluster --num-executors 10 --executor-memory 12g --executor-cores 1 spark.jarSpark assembly has been built with Hive, including Datanucleus jars on classpathWarning: Ignoring non-spark config property: yarn.resourcemanager.address=10.150.20.22:8032Warning: Ignoring non-spark config property: yarn.resourcemanager.address=10.150.20.22:803214/09/04 11:01:27 INFO client.RMProxy: Connecting to ResourceManager at /10.150.20.22:803214/09/04 11:01:27 INFO yarn.Client: Got Cluster metric info from ApplicationsManager (ASM), number of NodeManagers: 1014/09/04 11:01:27 INFO yarn.Client: Queue info ... queueName: default, queueCurrentCapacity: 0.0, queueMaxCapacity: 1.0, queueApplicationCount = 0, queueChildQueueCount = 014/09/04 11:01:27 INFO yarn.Client: Max mem capabililty of a single resource in this cluster 1228814 /09/04 11:01:27 INFO yarn.Client: Preparing Local resources14/09/04 11:01:27 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable14/09/04 11:01:27 INFO yarn.Client: Uploading file:/opt/spark-1.0.2-bin-hadoop2/bin/spark.jar to hdfs://10.150.20.22:8020/user/root/.sparkStaging/application_1409759471992_0007/spark.jar14/09/04 11:01:28 INFO yarn.Client: Uploading file:/opt/spark-1.0.2-bin-hadoop2/lib/spark-assembly-1.0.2-hadoop2.4.0.jar to hdfs://10.150.20.22:8020/user/root/.sparkStaging/application_1409759471992_0007/spark-assembly-1.0.2-hadoop2.4.0.jar14/09/04 11:01:30 INFO yarn.Client: Setting up the launch environment14/09/04 11:01:30 INFO yarn.Client: Setting up container launch context, , /stderr)14/09/04 11:01:30 INFO yarn.Client: Submitting application to ASM14/09/04 11:01:30 INFO impl.YarnClientImpl: Submitted application application_1409759471992_000714/09/04 11:01:31 INFO yarn.Client: Application repo rt from ASM:application identifier: application_1409759471992_0007 appId: 7clientToAMToken: null appDiagnostics: appMasterHost: N/A appQueue: default appMasterRpcPort: -1appStartTime: 1409796090132 yarnAppState: ACCEPTED distributedFinalState: UNDEFINED appTrackingUrl: http://saturn00:8088/proxy/application_1409759471992_0007/ appUser: root14/09/04 11:01:32 INFO yarn.Client: Application report from ASM: application identifier: application_1409759471992_0007 appId: 7 clientToAMToken: null appDiagnostics: appMasterHost: N/A appQueue: default appMasterRpcPort: -1appStartTime: 1409796090132 yarnAppState: ACCEPTED distributedFinalState: UNDEFINEDappTrackingUrl: http://saturn00:8088/proxy/application_1409759471992_0007/ appUser: root So, I found the log of applicationmaster of spark on yarn job, here is the logs: 14/09/04 11:01:33 INFO ApplicationMaster: ApplicationAttemptId: appattempt_1409759471992_0007_01 14/09/04 11:01:33 INFO RMProxy: Connecting to ResourceManager at /0.0.0.0:8030 14/09/04 11:01:33 INFO SecurityManager: Changing view acls to: root 14/09/04 11:01:33 INFO SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(root) 14/09/04 11:01:33 INFO ApplicationMaster: Starting the user JAR in a separate Thread 14/09/04 11:01:33 INFO ApplicationMaster: Waiting for Spark context initialization 14/09/04 11:01:33 INFO ApplicationMaster: Waiting for Spark context initialization ... 0 14/09/04 11:01:33 WARN SparkConf: In Spark 1.0 and later spark.local.dir will be overridden by the value set by the cluster manager (via SPARK_LOCAL_DIRS in mesos/standalone and LOCAL_DIRS in YARN). 14/09/04 11:01:33 INFO SecurityManager: Changing view acls to: root 14/09/04 11:01:33 INFO SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(root) 14/09/04 11:01:33 INFO Slf4jLogger: Slf4jLogger started 14/09/04 11:01:33 INFO Remoting: Starting remoting 14/09/04 11:01:33 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://spark@saturn02:58444] 14/09/04 11:01:33 INFO Remoting: Remoting now listens on addresses: [akka.tcp://spark@saturn02:58444] 14/09/04 11:01:33 INFO SparkEnv: Registering MapOutputTracker 14/09/04 11:01:33 INFO SparkEnv: Registering BlockManagerMaster 14/09/04 11:01:33 INFO DiskBlockManager: Created local directory at /home/hadoop/logs/hadoop_tmp/usercache/root/appcache/application_1409759471992_0007/spark-local-20140904110133-a55e 14/09/04 11:01:33 INFO MemoryStore: MemoryStore started with capacity 273.1 MB. 14/09/04 11:01:33 INFO ConnectionManager: Bound socket to port 43903 with id = ConnectionManagerId(saturn02,43903) 14/09/04 11:01:33 INFO BlockManagerMaster: Trying to register BlockManager 14/09/04 11:01:33 INFO BlockManagerInfo: Registering block manager
Re: Why spark on yarn applicationmaster cannot get a proper resourcemanager address from yarnconfiguration?
Did you follow the exact step in this page https://spark.apache.org/docs/1.0.2/running-on-yarn.html ? Please be sure that HADOOP_CONF_DIR or YARN_CONF_DIR points to the directory which contains the (client side) configuration files for the Hadoop cluster. Guodong On Thu, Sep 4, 2014 at 10:15 AM, 남윤민 rony...@dgist.ac.kr wrote: Hello, I tried to submit a spark job to yarn cluster, there is an error occured with those messages: [root@saturn00 bin]# ./spark-submit --class SparkHiveJoin --master yarn-cluster --num-executors 10 --executor-memory 12g --executor-cores 1 spark.jar Spark assembly has been built with Hive, including Datanucleus jars on classpath Warning: Ignoring non-spark config property: yarn.resourcemanager.address= 10.150.20.22:8032 Warning: Ignoring non-spark config property: yarn.resourcemanager.address= 10.150.20.22:8032 14/09/04 11:01:27 INFO client.RMProxy: Connecting to ResourceManager at / 10.150.20.22:8032 14/09/04 11:01:27 INFO yarn.Client: Got Cluster metric info from ApplicationsManager (ASM), number of NodeManagers: 10 14/09/04 11:01:27 INFO yarn.Client: Queue info ... queueName: default, queueCurrentCapacity: 0.0, queueMaxCapacity: 1.0, queueApplicationCount = 0, queueChildQueueCount = 0 14/09/04 11:01:27 INFO yarn.Client: Max mem capabililty of a single resource in this cluster 12288 14/09/04 11:01:27 INFO yarn.Client: Preparing Local resources 14/09/04 11:01:27 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 14/09/04 11:01:27 INFO yarn.Client: Uploading file:/opt/spark-1.0.2-bin-hadoop2/bin/spark.jar to hdfs:// 10.150.20.22:8020/user/root/.sparkStaging/application_1409759471992_0007/spark.jar 14/09/04 11:01:28 INFO yarn.Client: Uploading file:/opt/spark-1.0.2-bin-hadoop2/lib/spark-assembly-1.0.2-hadoop2.4.0.jar to hdfs:// 10.150.20.22:8020/user/root/.sparkStaging/application_1409759471992_000 7/spark-assembly-1.0.2-hadoop2.4.0.jar 14/09/04 11:01:30 INFO yarn.Client: Setting up the launch environment 14/09/04 11:01:30 INFO yarn.Client: Setting up container launch context 14/09/04 11:01:30 INFO yarn.Client: Command for starting the Spark ApplicationMaster: List($JAVA_HOME/bin/java, -server, -Xmx512m, -Djava.io.tmpdir=$PWD/tmp, -Dspark.local.dir=\/hadoop/spark\, -Dspark.shuffle.consolidateFiles=\true\, -Dspark.executor.memory=\12g\, -Dspark.master=\spark://10.150.20.22:8081\, -Dspark.app.name=\SparkHiveJoin\, -Dspark.eventLog.enabled=\true\, -Dspark.spill=\true\, -Dspark.serializer=\org.apache.spark.serializer.KryoSerializer\, -Dspark.eventLog.dir=\hdfs://10.150.20.22:8020/sparkLog\ http://10.150.20.22:8020/sparkLog%5C, -Dlog4j.configuration=log4j-spark-container.properties, org.apache.spark.deploy.yarn.ApplicationMaster, --class, SparkHiveJoin, --jar , file:/opt/spark-1.0.2-bin-hadoop2/bin/spark.jar, , --executor-memory, 12288, --executor-cores, 1, --num-e xecutors , 10, 1, LOG_DIR/stdout, 2, LOG_DIR/stderr) 14/09/04 11:01:30 INFO yarn.Client: Submitting application to ASM 14/09/04 11:01:30 INFO impl.YarnClientImpl: Submitted application application_1409759471992_0007 14/09/04 11:01:31 INFO yarn.Client: Application report from ASM: application identifier: application_1409759471992_0007 appId: 7 clientToAMToken: null appDiagnostics: appMasterHost: N/A appQueue: default appMast erRpcPort: -1 appStartTime: 1409796090132 yarnAppState: ACCEPTED distributedFinalState: UNDEFINED appTrackingUrl: http://saturn00:8088/proxy/application_1409759471992_0007/ appUser: root 14/09/04 11:01:32 INFO yarn.Client: Application report from ASM: application identifier: application_1409759471992_0007 appId: 7 clientToAMToken: null appDiagnostics: appMasterHost: N/A appQueue: default appMasterRpcPort: -1 appStartTime: 1409796090132 yarnAppState: ACCEPTED distributedFinalState: UNDEFINED appTrackingUrl: http://saturn00:8088/proxy/application_1409759471992_0007/ appUser: root So, I found the log of applicationmaster of spark on yarn job, here is the logs: 14/09/04 11:01:33 INFO ApplicationMaster: ApplicationAttemptId: appattempt_1409759471992_0007_01 14/09/04 11:01:33 INFO RMProxy: Connecting to ResourceManager at /0.0.0.0:8030 14/09/04 11:01:33 INFO SecurityManager: Changing view acls to: root 14/09/04 11:01:33 INFO SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(root) 14/09/04 11:01:33 INFO ApplicationMaster: Starting the user JAR in a separate Thread 14/09/04 11:01:33 INFO ApplicationMaster: Waiting for Spark context initialization 14/09/04 11:01:33 INFO ApplicationMaster: Waiting for Spark context initialization ... 0 14/09/04 11:01:33 WARN SparkConf: In Spark 1.0 and later spark.local.dir will be overridden by the value set by the cluster manager (via SPARK_LOCAL_DIRS in mesos/standalone and LOCAL_DIRS in