#!/bin/bash source ../dmp_env.sh region=$1 today=${ScheduleTime:-$2} datetime=$(date +"%Y/%m/%d %H" -d "1 hour ago $today") date=${datetime:0:10} hour=${datetime:11:2} date_path="${date}/${region}/${hour}" sleep 10 host="ip-172-31-20-35.ec2.internal" cluster="cluster_1st" database="dwc" table="realtime_service_result" S3_OUTPUT="s3://mob-emr-test/dataplatform/DataWareHouse/data/dev/realtime_service_device/${date_path}" if [[ ${region} == "cn" ]]; then partition=10 cores=2 else partition=100 cores=4 fi spark-submit --class mobvista.dmp.clickhouse.realtime.ReadFromCKWriteCS \ --name "ReadFromCKWriteCS_${date}_${hour}_${region}" \ --conf spark.sql.shuffle.partitions=${partition} \ --conf spark.default.parallelism=${partition} \ --conf spark.sql.files.maxPartitionBytes=33554432 \ --conf spark.yarn.executor.memoryOverhead=2048 \ --conf spark.driver.extraJavaOptions="-XX:+UseG1GC" \ --conf spark.executor.extraJavaOptions="-XX:+UseG1GC" \ --master yarn --deploy-mode cluster --executor-memory 4g --driver-memory 2g --executor-cores 2 --num-executors ${cores} \ ../${JAR} \ -date ${date} -hour ${hour} -region ${region} -host ${host} -cluster ${cluster} -database ${database} -table ${table} --output ${S3_OUTPUT} -part ${partition} if [[ $? -ne 0 ]]; then exit 255 fi