#!/bin/bash

source ../../dmp_env.sh

# region name
region=$1

# 等待时间
hour=$2

schedule_time=$(date +"%Y-%m-%d %H:%M:%S" -d "${hour} hour $ScheduleTime")

final_schedule_time=$(date +%s -d "$schedule_time")

currtime=`date '+%s'`

while [[ true ]];
do
    if [ ${currtime} -gt ${final_schedule_time} ];then
        break
    fi
    sleep 300
    currtime=`date '+%s'`
done

# 回流备份开始时间
backflow_start_time=$(date +%Y%m%d -d "-1 day $ScheduleTime")

# 回流备份开始时间戳 * 1000000，Cassandra 中 writeTime 为16位
writetime_start=`expr $(date +%s -d "$backflow_start_time") \* 1000000`

# 回流备份结束时间
backflow_end_time=$(date +%Y%m%d -d "$ScheduleTime")

# 回流备份结束时间戳 * 1000000
writetime_end=`expr $(date +%s -d "$backflow_end_time") \* 1000000`

# package name
system="mapping"

# keyspace name
keyspace="mapping"

# table name
table="mapping_server"

# 需指定一个 value column，用于计算 writeTime
value_column="devid_value"

date_path=$(date +"%Y/%m/%d" -d "-1 day ${ScheduleTime}")

# 备份 s3 地址
output="${BACKFLOW_OUTPUT}/${keyspace}/${table}/${date_path}/${region}/"

spark-submit --class mobvista.dmp.datasource.backflow.BackFlow \
  --name "BackFlow.${keyspace}.${table}.${region}" \
  --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 4g --driver-memory 4g --executor-cores 2 --num-executors 4 \
  ../.././DMP.jar \
  -keyspace ${keyspace} -table ${table} -region ${region} -output ${output} -system ${system} \
  -writetime_start ${writetime_start} -writetime_end ${writetime_end} -value_column ${value_column}

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