rtdmp.sh 1.48 KB
#!/bin/bash

source ../dmp_env.sh

today=${ScheduleTime}

date_time=$(date +"%Y%m%d%H" -d "-1 hour $today")

date_path=$(date +%Y/%m/%d/%H -d "-1 hour $today")

INPUT="s3://mob-emr-test/dataplatform/rtdmp_pre/${date_path}"

old_date_time=$(date +"%Y%m%d%H" -d "-2 hour $today")

old_date_path=$(date +%Y/%m/%d/%H -d "-2 hour $today")

OLD_MERGE_INPUT="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/audience_merge/${old_date_path}"

check_await ${OLD_MERGE_INPUT}/_SUCCESS

sleep 120

OUTPUT="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/audience_merge/${date_path}"

spark-submit --class mobvista.dmp.datasource.rtdmp.RTDmpMain \
  --name "RTDmpMain.${date_time}" \
  --conf spark.sql.shuffle.partitions=2000 \
  --conf spark.default.parallelism=2000 \
  --conf spark.kryoserializer.buffer.max=512m \
  --conf spark.kryoserializer.buffer=64m \
  --master yarn --deploy-mode cluster \
  --executor-memory 8g --driver-memory 6g --executor-cores 4 --num-executors 100 \
  .././DMP.jar \
  -datetime ${date_time} -old_datetime ${old_date_time} -input ${INPUT} -output ${OUTPUT} -coalesce 400

if [[ $? -ne 0 ]]; then
  exit 255
fi

mount_partition "audience_merge" "dt='${date_time}'" "$OUTPUT"

expire_time=$(date +"%Y%m%d%H" -d "-24 hour $today")

expire_date_path=$(date +%Y/%m/%d/%H -d "-24 hour $today")

EXPIRE_OUTPUT_PATH="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/audience_merge/${expire_date_path}"

unmount_partition "audience_merge" "dt='${expire_time}'" "${EXPIRE_OUTPUT_PATH}"