#!/usr/bin/env bash source ../../dmp_env.sh source ././../../ga_rawdata_analysis/common/tools.sh echo "job begin!!!" dt_today=$(date -d "$ScheduleTime" +"%Y%m%d") dt_begin_days=$(date -d "$ScheduleTime" +"%Y%m%d") dt_slash_today=$(date -d "$ScheduleTime" +"%Y/%m/%d") update=$(date -d "$ScheduleTime" +"%Y-%m-%d") INPUT_PATH="${YOUKU_LAXIN_DAILY_TMP_PATH}/${dt_slash_today}" OUTPUT_PATH="${YOUKU_LAXIN_TMP_DAILY_TO_S3}/${dt_slash_today}/youku_acquisition" check_await "${INPUT_PATH}/imeimd5/_SUCCESS" check_await "${INPUT_PATH}/oaidmd5/_SUCCESS" hadoop fs -rm -r "${OUTPUT_PATH}" spark-submit --class mobvista.dmp.datasource.taobao.YoukuTmpDataToDmp \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=2000 \ --conf spark.sql.shuffle.partitions=2000 \ --conf spark.sql.broadcastTimeout=1200 \ --conf spark.yarn.executor.memoryOverhead=4096 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --files ${HIVE_SITE_PATH} \ --master yarn --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 4 --num-executors 40 \ ../../${JAR} -Input "${INPUT_PATH}/*/*" -Output ${OUTPUT_PATH} \ -update ${update} if [ $? -ne 0 ];then exit 255 fi DMP_INSTALL_LIST_ACQUISITION_PATH="${DM_INSTALL_LIST}_v2/${dt_slash_today}/youku_acquisition" mount_partition "dm_install_list_v2" "dt='${dt_today}', business='youku_acquisition'" "${DMP_INSTALL_LIST_ACQUISITION_PATH}" mount_partition "youku_laxin_tmp_daily_to_s3" "dt='${dt_today}', business='youku_acquisition'" "${OUTPUT_PATH}" if [ $? -ne 0 ];then exit 255 fi ACQUISITIONOUTPUT="${YOUKU_LAXIN_TMP_DAILY_TO_S3}/${dt_slash_today}/youku_acquisition_polling" spark-submit --class mobvista.dmp.datasource.taobao.YoukuLaXinPollingDataDeduplication \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=2000 \ --conf spark.sql.shuffle.partitions=2000 \ --conf spark.sql.broadcastTimeout=1200 \ --conf spark.yarn.executor.memoryOverhead=4096 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --files ${HIVE_SITE_PATH} \ --master yarn --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 4 --num-executors 40 \ ../../${JAR} -dt_today ${dt_today} -dt_begin_days ${dt_begin_days} \ -AcquisitionOutput ${ACQUISITIONOUTPUT} if [ $? -ne 0 ];then exit 255 fi mount_partition "youku_laxin_tmp_daily_to_s3" "dt='${dt_today}', business='youku_acquisition_polling'" "${ACQUISITIONOUTPUT}" if [ $? -ne 0 ];then exit 255 fi HIVE_CMD=$(hive_func) $HIVE_CMD -v -hivevar dt_today ${dt_today} -f youku_laxin_data_to_dmp.sql if [ $? -ne 0 ];then exit 255 fi hadoop fs -touchz ${DMP_INSTALL_LIST_ACQUISITION_PATH}/_SUCCESS expire_date_path=$(date -d "$ScheduleTime 61 days ago" +"%Y/%m/%d") EXPIRE_PATH="${YOUKU_LAXIN_TMP_DAILY_TO_S3}/${expire_date_path}" hadoop fs -rm -r "${EXPIRE_PATH}" echo "good job"