#!/bin/bash source ../dmp_env.sh part=$1 today=${ScheduleTime} date_time=$(date +"%Y-%m-%d.%H" -d "-1 hour $today") date_path=$(date +%Y/%m/%d/%H -d "-1 hour $today") part_num=$(hadoop fs -ls s3://mob-emr-test/dataplatform/rtdmp_pre/${date_path}/ | wc -l) if [[ ${part_num} -le 50 && ${part} -ne 0 ]]; then echo "This Dir No Data !!!" exit 0 else INPUT="s3://mob-emr-test/dataplatform/rtdmp_pre/${date_path}/part-000${part}*" OUTPUT="s3://mob-emr-test/dataplatform/rtdmp_deal/${date_path}/${part}" sleep 10 spark-submit --class mobvista.dmp.datasource.rtdmp.RTDmpMainDealV2 \ --name "RTDmpMainDealV2.${date_time}.${part}" \ --conf spark.sql.shuffle.partitions=2000 \ --conf spark.default.parallelism=2000 \ --conf spark.kryoserializer.buffer.max=256m \ --master yarn --deploy-mode cluster --executor-memory 10g --driver-memory 4g --executor-cores 5 --num-executors 20 \ ../${JAR} -time "${date_time}" -input ${INPUT} -output ${OUTPUT} -coalesce 100 -partition 2000 if [[ $? -ne 0 ]]; then exit 255 fi fi