#!/usr/bin/env bash source ../../dmp_env.sh LOG_TIME=$(date -d "$ScheduleTime" +"%Y%m%d") date_path=$(date -d "$ScheduleTime" +"%Y/%m/%d") hour="03" IMEIMD5_INPUT="${ALIPAY_LAHUO_DAILY_PATH}/$date_path/imeimd5_request_data/${hour}" host="ip-172-31-20-35.ec2.internal" cluster="cluster_1st" database="dmp" table="alipay_lahuo_daily" spark-submit --class mobvista.dmp.datasource.taobao.AlipayImeiLaHuoCK \ --name "AlipayImeiLaHuoCK.${LOG_TIME}" \ --conf spark.sql.shuffle.partitions=100 \ --conf spark.default.parallelism=100 \ --conf spark.kryoserializer.buffer.max=256m \ --conf hive.exec.orc.split.strategy="BI" \ --conf spark.driver.extraJavaOptions="-XX:+UseG1GC" \ --conf spark.executor.extraJavaOptions="-XX:+UseG1GC" \ --conf hive.exec.orc.default.stripe.size=268435456 \ --conf spark.sql.files.maxPartitionBytes=268435456 \ --master yarn --deploy-mode cluster --executor-memory 4g --driver-memory 4g --executor-cores 2 --num-executors 10 \ ../../${JAR} -imeiInput ${IMEIMD5_INPUT} -date ${LOG_TIME} -host ${host} -cluster ${cluster} -database ${database} -table ${table} -hour ${hour} if [[ $? -ne 0 ]];then exit 255 fi