我们在上一篇文章《HBase复制》中讲述了如何建立主/从集群，实现数据的实时备份。但是，HBase复制只对设置好复制以后的数据生效，也即，配置好复制之后插入HBase主集群的数据才能同步复制到HBase从集群中，而对之前的历史数据，采用HBase复制这种办法是无能为力的。本文介绍如何使用HBase的导入导出功能来实现历史数据的备份。
1）将HBase表数据导出到hdfs的一个指定目录中，具体命令如下：
$ cd $HBASE_HOME/ $ bin/hbase org.apache.hadoop.hbase.mapreduce.Export test_table /data/test_table
其中，$HBASE_HOME为HBase主目录，test_table为要导出的表名，/data/test_table为hdfs中的目录地址。
执行结果太长，这里截取最后一部分，如下所示：
2014-08-11 16:49:44,484 INFO [main] mapreduce.Job: Running job: job_1407491918245_0021 2014-08-11 16:49:51,658 INFO [main] mapreduce.Job: Job job_1407491918245_0021 running in uber mode : false 2014-08-11 16:49:51,659 INFO [main] mapreduce.Job: map 0% reduce 0% 2014-08-11 16:49:57,706 INFO [main] mapreduce.Job: map 100% reduce 0% 2014-08-11 16:49:57,715 INFO [main] mapreduce.Job: Job job_1407491918245_0021 completed successfully 2014-08-11 16:49:57,789 INFO [main] mapreduce.Job: Counters: 37 File System Counters FILE: Number of bytes read=0 FILE: Number of bytes written=118223 FILE: Number of read operations=0 FILE: Number of large read operations=0 FILE: Number of write operations=0 HDFS: Number of bytes read=84 HDFS: Number of bytes written=243 HDFS: Number of read operations=4 HDFS: Number of large read operations=0 HDFS: Number of write operations=2 Job Counters Launched map tasks=1 Rack-local map tasks=1 Total time spent by all maps in occupied slots (ms)=9152 Total time spent by all reduces in occupied slots (ms)=0 Map-Reduce Framework Map input records=3 Map output records=3 Input split bytes=84 Spilled Records=0 Failed Shuffles=0 Merged Map outputs=0 GC time elapsed (ms)=201 CPU time spent (ms)=5210 Physical memory (bytes) snapshot=377470976 Virtual memory (bytes) snapshot=1863364608 Total committed heap usage (bytes)=1029177344 HBase Counters BYTES_IN_REMOTE_RESULTS=87 BYTES_IN_RESULTS=87 MILLIS_BETWEEN_NEXTS=444 NOT_SERVING_REGION_EXCEPTION=0 NUM_SCANNER_RESTARTS=0 REGIONS_SCANNED=1 REMOTE_RPC_CALLS=3 REMOTE_RPC_RETRIES=0 RPC_CALLS=3 RPC_RETRIES=0 File Input Format Counters Bytes Read=0 File Output Format Counters Bytes Written=243
查看以下指定的导出目录，命令如下：
$ cd $HADOOP_HOME/ $ bin/hadoop fs -ls /data/test_table
其中$HADOOP_HOME为hadoop的主目录。结果如下：
Found 2 items -rw-r--r-- 3 hbase supergroup 0 2014-08-11 16:49 /data/test_table/_SUCCESS -rw-r--r-- 3 hbase supergroup 243 2014-08-11 16:49 /data/test_table/part-m-00000
执行以下hbase shell命令，查看以下test_table表中的数据：
$ cd $HBASE_HOME/
$ bin/hbase shell
2014-08-11 17:05:52,589 INFO [main] Configuration.deprecation: hadoop.native.lib is deprecated. Instead, use io.native.lib.available
HBase Shell; enter 'help<RETURN>' for list of supported commands.
Type "exit<RETURN>" to leave the HBase Shell
Version 0.98.2-hadoop2, r1591526, Wed Apr 30 20:17:33 PDT 2014
hbase(main):001:0> describe 'test_table'
DESCRIPTION ENABLED
'test_table', {NAME => 'cf', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '1', COMPRESSION => 'NONE', VERSIONS => true
'1', TTL => '2147483647', MIN_VERSIONS => '0', KEEP_DELETED_CELLS => 'false', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE =>
'true'}
1 row(s) in 1.3400 seconds
hbase(main):002:0> scan 'test_table'
ROW COLUMN+CELL
r1 column=cf:q1, timestamp=1406788229440, value=va1
r2 column=cf:q1, timestamp=1406788265646, value=va2
r3 column=cf:q1, timestamp=1406788474301, value=va3
3 row(s) in 0.0560 seconds
至此，HBase表数据导出结束。接下来开始导入工作。
2）将导出到hdfs中的数据导入到hbase创建好的表中。注意，该表可以和之前的表不同名，但模式一定要相同。我们领取一个名字，使用test_copy这个表名。创建表的命令如下：
$ cd $HBASE_HOME/ $ bin/hbase shell 2014-08-11 17:05:52,589 INFO [main] Configuration.deprecation: hadoop.native.lib is deprecated. Instead, use io.native.lib.available HBase Shell; enter 'help<RETURN>' for list of supported commands. Type "exit<RETURN>" to leave the HBase Shell Version 0.98.2-hadoop2, r1591526, Wed Apr 30 20:17:33 PDT 2014 hbase(main):001:0> create 'test_copy', 'cf' 0 row(s) in 1.1980 seconds => Hbase::Table - test_copy
接下来，执行导入命令。具体的命令如下：
$ cd $HBASE_HOME/ $ bin/hbase org.apache.hadoop.hbase.mapreduce.Import test_copy hdfs://l-master.data/data/test_table
其中，test_copy为我们想要导入的表名。而hdfs://l-master.data/data/test_table为master集群的hdfs中，我们之前将test_table表导出hdfs的全路径。
导入命令执行的结果如下，因为结果很长，所以取最后一部分：
2014-08-11 17:13:08,706 INFO [main] mapreduce.Job: map 100% reduce 0% 2014-08-11 17:13:08,710 INFO [main] mapreduce.Job: Job job_1407728839061_0014 completed successfully 2014-08-11 17:13:08,715 INFO [main] mapreduce.Job: Counters: 27 File System Counters FILE: Number of bytes read=0 FILE: Number of bytes written=117256 FILE: Number of read operations=0 FILE: Number of large read operations=0 FILE: Number of write operations=0 HDFS: Number of bytes read=356 HDFS: Number of bytes written=0 HDFS: Number of read operations=3 HDFS: Number of large read operations=0 HDFS: Number of write operations=0 Job Counters Launched map tasks=1 Rack-local map tasks=1 Total time spent by all maps in occupied slots (ms)=6510 Total time spent by all reduces in occupied slots (ms)=0 Map-Reduce Framework Map input records=3 Map output records=3 Input split bytes=113 Spilled Records=0 Failed Shuffles=0 Merged Map outputs=0 GC time elapsed (ms)=21 CPU time spent (ms)=1110 Physical memory (bytes) snapshot=379494400 Virtual memory (bytes) snapshot=1855762432 Total committed heap usage (bytes)=1029177344 File Input Format Counters Bytes Read=243 File Output Format Counters Bytes Written=0
接下来，我们看看从集群test_copy表中的数据是否和主集群test_table表的数据一致，执行hbase shell命令：
$ cd $HBASE_HOME/ $ bin/hbase shell 2014-08-11 17:15:52,117 INFO [main] Configuration.deprecation: hadoop.native.lib is deprecated. Instead, use io.native.lib.available HBase Shell; enter 'help<RETURN>' for list of supported commands. Type "exit<RETURN>" to leave the HBase Shell Version 0.98.2-hadoop2, r1591526, Wed Apr 30 20:17:33 PDT 2014 hbase(main):001:0> scan 'test_copy' ROW COLUMN+CELL r1 column=cf:q1, timestamp=1406788229440, value=va1 r2 column=cf:q1, timestamp=1406788265646, value=va2 r3 column=cf:q1, timestamp=1406788474301, value=va3 3 row(s) in 0.3640 seconds
对照后，就可以发现，两个表的数据是完全一致的。
转载请注明出处：http://blog.csdn.net/iAm333
HBase备份之导入导出
时间： 2024-11-05 22:07:31