#!/bin/sh # # # # # # # # # # # # # # # # # # # # # # # @author : wangjf # @date : 2019-02-27 11:24:06 # # # # # # # # # # # # # # # # # # # # # # source ../dmp_env.sh LOG_TIME=$(date -d "$ScheduleTime 1 days ago" +"%Y%m%d") dt=$(date -d "$ScheduleTime 1 days ago" +"%Y-%m-%d") date_path=$(date -d "$ScheduleTime 1 days ago" +"%Y/%m/%d") host="ip-172-31-20-35.ec2.internal" cluster="cluster_1st" database="dwh" table="joypac_daily" part='etl' INPUT_PATH="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/stage_joypac_upload_daily/$date_path/*/*" OUTPUT_PATH="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/joypac_result/${date_path}/${part}" spark-submit --class mobvista.dmp.datasource.joypac.JoypacResultEtl \ --conf spark.yarn.executor.memoryOverhead=2048 \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=20 \ --conf spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive=true \ --conf spark.hadoop.fs.s3.impl=org.apache.hadoop.fs.s3a.S3AFileSystem \ --conf spark.hadoop.fs.s3a.impl=org.apache.hadoop.fs.s3a.S3AFileSystem \ --master yarn --deploy-mode cluster --name JoypacResutlEtl \ --executor-memory 4g --driver-memory 4g --executor-cores 2 --num-executors 5 \ ../${JAR} -input ${INPUT_PATH} -output ${OUTPUT_PATH} -coalesce 10 -date ${LOG_TIME} -host ${host} -cluster ${cluster} -database ${database} -table ${table} if [[ $? -ne 0 ]];then exit 255 fi mount_partition "joypac_result" "dt='${LOG_TIME}',part='${part}'" "$OUTPUT_PATH"