#!/bin/sh # # # # # # # # # # # # # # # # # # # # # # # @author : jiangfan # @date : 2021-01-24 12:06:00 # # # # # # # # # # # # # # # # # # # # # # #!/usr/bin/env bash source ../../dmp_env.sh source ././../../ga_rawdata_analysis/common/tools.sh dt_today=$(date -d "$ScheduleTime" +"%Y%m%d") dt_dash_today=$(date -d "$ScheduleTime" +"%Y-%m-%d") dt_slash_today=$(date -d "$ScheduleTime" +"%Y/%m/%d") #检查表分区是否已经存在,不存在等待十个小时退出脚本 partition_info="dt='$dt_dash_today'" check_await_hive_partition "uparpu_main.uparpu_device_active" $partition_info 36000 OUTPUT_PATH="${TO_DAILY_PATH}/${dt_slash_today}" spark-submit --class mobvista.dmp.datasource.TO.TODaily \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=3000 \ --conf spark.sql.shuffle.partitions=3000 \ --conf spark.sql.broadcastTimeout=1200 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --master yarn --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 5 --num-executors 20 \ ../../${JAR} \ -output ${OUTPUT_PATH} -coalesce 200 -dt_dash_today ${dt_dash_today} if [[ $? -ne 0 ]];then exit 255 fi mount_partition "etl_to_daily" "dt='${dt_today}'" "$OUTPUT_PATH" hadoop fs -touchz ${OUTPUT_PATH}/_SUCCESS