#!/bin/bash # # # # # # # # # # # # # # # # # # # # # # # @file : adn_request_other_install.sh # @author: feng.liang # @date : 2018-05-29 # @desc : 将天数据合并到安装列表 # # # # # # # # # # # # # # # # # # # # # # source ../../dmp_env.sh LOG_TIME=$(date -d "$ScheduleTime 1 days ago" "+%Y%m%d") old_path=$(date -d "$ScheduleTime 2 days ago" "+%Y/%m/%d") date_path=$(date -d "$ScheduleTime 1 days ago" "+%Y/%m/%d") date=$(date +"%Y-%m-%d" -d "-1 day $ScheduleTime") INPUT_PATH="$ETL_ADN_REQUEST_OTHER_DAILY/${date_path}/" OLD_INPUT_PATH="${DM_INSTALL_LIST}/$old_path/adn_request_other" OUTPUT_PATH="$DM_INSTALL_LIST/$date_path/adn_request_other" check_await $INPUT_PATH/_SUCCESS hadoop fs -rm -r "$OUTPUT_PATH" spark-submit --class mobvista.dmp.datasource.adn_request_other.AdnRequestOtherInstall \ --conf spark.yarn.executor.memoryOverhead=2048 \ --conf spark.sql.shuffle.partitions=2000 \ --master yarn --deploy-mode cluster --executor-memory 6g --driver-memory 4g --executor-cores 2 --num-executors 200 \ ../../${JAR} \ -input "${INPUT_PATH}" -output $OUTPUT_PATH -date $date -oldInput $OLD_INPUT_PATH -parallelism 2000 -coalesce 2000 if [ $? -ne 0 ]; then exit 255 fi mount_partition "dm_install_list" "year='${LOG_TIME:0:4}', month='${LOG_TIME:4:2}', day='${LOG_TIME:6:2}', business='adn_request_other'" "$OUTPUT_PATH" || exit 1 echo "[Etl adn request other install end!]"