1. there are a large amount of data in Hbase (row numer = 3849920), read all data from mongodb again and below is the log: [azureuser@h3crd-wlan31 ~]$ kubectl logs histclientsdataetl-20161028141055-doize |grep @ETL@ @ETL@ getLatestOnlineTime: 2016-10-27 23:59:59.815@ETL@ get uplineDate from hbase: 2016-10-27 23:59:59.815@ETL@ getCurrentPosFromDB: 2016-10-27 23:59:59.815@ETL@ Start finding at time: 2016-10-28 15:00:17.991, startFindDayIndex=17101@ETL@ Finding finished at time: 2016-10-28 15:00:21.551@ETL@ Wrting to Hbase has been finished !@ETL@readCount: 3849920,putCount: 0,time: 2016-10-28 15:11:48.784@ETL@ child file name: 1.1.1.1-histclientsdataetl@ETL@ Change logfile name has finished: 2016-10-28 15:11:48.792@ETL@ cmd: sh /home/mongo-hive-hbase/Hive_HistClientsInfoAnalysis.sh@ETL@: exec has finished!@ETL@ new Executing remote hive shell has finished: 2016-10-28 15:12:49.273@ETL@ result:
将代码改为批量获取后的效果(10695026条数据共4个G左右):
public MongoCursorfind(Bson filter) { FindIterable findIterable = this.collection.find(filter).batchSize(50000); MongoCursor mongoCursor = findIterable.iterator(); return mongoCursor; }
[azureuser@h3crd-wlan31 ~]$ kubectl logs histclientsdataetl-xubd1 |grep @ETL@@ETL@ getLatestOnlineTime: 2016-10-28 23:59:59.982@ETL@ get uplineDate from hbase: 2016-10-28 23:59:59.982@ETL@ getCurrentPosFromDB: 2016-10-28 23:59:59.982@ETL@ Start finding at time: 2016-10-31 00:05:02.858, startFindDayIndex=17102@ETL@ Finding finished at time: 2016-10-31 00:05:04.154@ETL@ Wrting to Hbase has been finished !@ETL@ upDateLatestOnlineTime: 2016-10-30 23:59:59.999@ETL@readCount: 10695026,putCount: 10695026,time: 2016-10-31 00:48:41.78
数据量是之前的2.7倍,拉取时间4倍,读取的时候是从Mongodb的备服务器上读取
继续修改代码加batchsize增加一倍后:
public MongoCursorfind(Bson filter) { FindIterable findIterable = this.collection.find(filter).batchSize(100000); MongoCursor mongoCursor = findIterable.iterator(); return mongoCursor; }
[azureuser@h3crd-wlan31 ~]$ kubectl logs histclientsdataetl-20161031095007-qv4fd |grep @ETL@@ETL@ getLatestOnlineTime: 2016-10-30 23:59:59.999@ETL@ get uplineDate from hbase: 2016-10-30 23:59:59.999@ETL@ getCurrentPosFromDB: 2016-10-30 23:59:59.999@ETL@ Start finding at time: 2016-10-31 10:26:55.071, startFindDayIndex=17104@ETL@ Finding finished at time: 2016-10-31 10:26:56.406@ETL@ Wrting to Hbase has been finished !@ETL@readCount: 7936993,putCount: 0,time: 2016-10-31 11:04:02.219
数据量2.06倍,时间3.08倍