概述

A集群HBase中有个表testTableCopy, 要将其复制到B集群的HBase中。

使用HBase表复制工具./hbase org.apache.hadoop.hbase.mapreduce.CopyTable

1root@host:/tstar/hbase/bin-->./hbase org.apache.hadoop.hbase.mapreduce.CopyTable
2Usage: CopyTable [general options] [--starttime=X] [--endtime=Y] [--new.name=NEW] [--peer.adr=ADR] <tablename>
3
4Options:
5 rs.class     hbase.regionserver.class of the peer cluster
6              specify if different from current cluster
7 rs.impl      hbase.regionserver.impl of the peer cluster
8 startrow     the start row
9 stoprow      the stop row
10 starttime    beginning of the time range (unixtime in millis)
11              without endtime means from starttime to forever
12 endtime      end of the time range.  Ignored if no starttime specified.
13 versions     number of cell versions to copy
14 new.name     new table's name
15 peer.adr     Address of the peer cluster given in the format
16              hbase.zookeeer.quorum:hbase.zookeeper.client.port:zookeeper.znode.parent
17 families     comma-separated list of families to copy
18              To copy from cf1 to cf2, give sourceCfName:destCfName. 
19              To keep the same name, just give "cfName"
20 all.cells    also copy delete markers and deleted cells
21 bulkload     Write input into HFiles and bulk load to the destination table
22
23Args:
24 tablename    Name of the table to copy
25
26Examples:
27 To copy 'TestTable' to a cluster that uses replication for a 1 hour window:
28 $ bin/hbase org.apache.hadoop.hbase.mapreduce.CopyTable --starttime=1265875194289 --endtime=1265878794289 --peer.adr=server1,server2,server3:2181:/hbase --families=myOldCf:myNewCf,cf2,cf3 TestTable 
29For performance consider the following general options:
30-Dhbase.client.scanner.caching=100
31-Dmapred.map.tasks.speculative.execution=false

准备

1. 工具必须在A集群HBase master节点上进行；
2. 在A集群HBase master节点上须要配置B集群HBase的相关主机名到hosts中，原则上是zookeeper及HBase相关节点主机名；
3. 在B集群HBase中预先建好复制的目标表，目标表的表名可以和原始的表名，目标表的列族名和来源表的列族名，可以不同；

执行复制

./hbase -Dhbase.root.logger=INFO,RFA -Dhbase.log.dir=/tmp -Dhbase.log.file=copytable.log org.apache.hadoop.hbase.mapreduce.CopyTable --peer.adr=test61,test62,test63:2181:/hbase --families=1 "tstar:testTableCopy"