config: #导出数据任务 user.to.proxy: foo day: $(new("org.joda.time.DateTime").minusDays(1).toString("yyyyMMdd")) jarPATH: ${working.dir}/jars/algorithm-framework-1.0-SNAPSHOT.jar mainClassName: com.iov.flow.task.StarterTask systype: idc nodes: - name: nsure type: command dependsOn: config: command: spark-submit --class ${mainClassName} \ --master yarn \ --deploy-mode client \ --driver-memory 4g \ --executor-memory 16g \ --executor-cores 1 \ --num-executors 100 \ --queue default \ --conf spark.default.parallelism=1000 \ --conf spark.task.maxFailures=10 \ --conf spark.storage.memoryFraction=0.4 \ --conf spark.shuffle.memoryFraction=0.4 \ --conf spark.shuffle.io.maxRetries=120 \ --conf spark.sql.shuffle.partitions=1000 \ --conf spark.sql.files.maxPartitionBytes=1073741824 \ --conf spark.network.timeout=36000 \ --conf dfs.client.socket-timeout=3600000 \ --conf spark.shuffle.spill.numElementsForceSpillThreshold=2000000 \ ${jarPATH} -DtaskType=spark -DtaskName=PARKINGPOINTCALCTASK -DtaskKind=PARKINGPOINTCALCTASK -DorgDataType=1 -DcalcDate=${day} -DtaskOutputPath=/data/tmp/ -DparamsPgEnv="" -Dtime=1800
标签:shuffle,--,flow,azkaban,conf,spark,2.0,day From: https://www.cnblogs.com/zyanrong/p/17851800.html