这篇文章记录一下如何在eclipse中调试运行hbase。
下载并编译源代码
修改配置文件
修改 conf/hbase-site.xml文件：
<property>
<name>hbase.defaults.for.version</name>
<value>0.94.6-cdh4.4.0</value>
</property>
<property>
<name>hbase.rootdir</name>
<value>file:///home/june/tmp/data</value>
</property>
把conf文件夹加到Classpath中
运行HMaster
新建一个Debug Configuration, main class 是org.apache.hadoop.hbase.master.HMaster, 参数填start
调试运行该类，运行成功之后日志如下：
13/10/29 14:38:29 WARN zookeeper.RecoverableZooKeeper: Node /hbase/table/.META. already deleted, and this is not a retry
13/10/29 14:38:29 INFO regionserver.HRegionServer: Received request to open region: .META.,,1.1028785192
13/10/29 14:38:29 INFO regionserver.HRegion: Setting up tabledescriptor config now ...
13/10/29 14:38:29 INFO regionserver.Store: time to purge deletes set to 0ms in store info
13/10/29 14:38:29 INFO regionserver.HRegion: Onlined .META.,,1.1028785192; next sequenceid=1
13/10/29 14:38:29 INFO regionserver.HRegionServer: Post open deploy tasks for region=.META.,,1.1028785192, daughter=false
13/10/29 14:38:29 INFO catalog.MetaEditor: Updated row .META.,,1.1028785192 with server=june-mint,47477,1383028701871
13/10/29 14:38:29 INFO regionserver.HRegionServer: Done with post open deploy task for region=.META.,,1.1028785192, daughter=false
13/10/29 14:38:29 INFO handler.OpenedRegionHandler: Handling OPENED event for .META.,,1.1028785192 from june-mint,47477,1383028701871; deleting unassigned node
13/10/29 14:38:29 INFO master.AssignmentManager: The master has opened the region .META.,,1.1028785192 that was online on june-mint,47477,1383028701871
13/10/29 14:38:29 INFO master.HMaster: .META. assigned=2, rit=false, location=june-mint,47477,1383028701871
13/10/29 14:38:29 INFO catalog.MetaMigrationRemovingHTD: Meta version=0; migrated=true
13/10/29 14:38:29 INFO catalog.MetaMigrationRemovingHTD: ROOT/Meta already up-to date with new HRI.
13/10/29 14:38:29 INFO master.AssignmentManager: Clean cluster startup. Assigning userregions
13/10/29 14:38:29 INFO master.HMaster: Registered HMaster MXBean
13/10/29 14:38:29 INFO master.HMaster: Master has completed initialization
如果想修改日志级别，请修改conf/log4j.properties中级别为INFO:
#Logging Threshold
log4j.threshold=INFO
运行HRegionServer
参考上面的方法，运行HRegionServer，这时候会出现如下日志：
13/11/04 11:50:47 INFO util.VersionInfo: HBase 0.94.6-cdh4.4.0
13/11/04 11:50:47 INFO util.VersionInfo: Subversion git://june-mint/chan/workspace/hadoop/hbase -r 979969e1d0d95ce3b8c1d14593f55148da8bc98f
13/11/04 11:50:47 INFO util.VersionInfo: Compiled by june on Tue Oct 29 15:11:51 CST 2013
13/11/04 11:50:47 WARN regionserver.HRegionServerCommandLine: Not starting a distinct region server because hbase.cluster.distributed is false
这是因为当hbase.cluster.distributed=false时，hbase为本地模式，master和regionserver在同一个jvm启动，并且会启动一个最小化的zookeeper集群。请参看：HMasterCommandLine.java的startMaster()方法。
如果你把该值设为true，则hbase集群为分布式模式，这时候默认会连接127.0.0.1：2181对应的zookeeper集群（该集群需要在master启动之前启动）。当然，你可以修改参数让hbase自己维护一个zookeeper集群。
调试hbase shell
新建一个Debug Configuration, main class 是org.jruby.Main，在程序参数中添加bin/hirb.rb,然后运行即可。
一些技巧
- 调试java代码的时候, byte[]的变量总是显示成数字,如果要显示对应的字符
Window->Preference->Java->Debug->Primitive Display Options->Check some of them
- hbase源码中默认依赖的是hadoop 1.0.x版本，所以mavne依赖中会引入hadoop-core-1.0.4.jar，你可以修改pom.xml文件，将默认的profile修改为你需要的hadoop版本，如2.0版本的hadoop。这样做之后，当你看HMaster的源代码时，你会很方便的关联并浏览ToolRunner类中的源代码。
默认的profile是hadoop-1.0，配置文件如下：
<!-- profile for building against Hadoop 1.0.x: This is the default. -->
<profile>
<id>hadoop-1.0</id>
<activation>
<property>
<name>!hadoop.profile</name>
</property>
<profile>
<id>hadoop-2.0</id>
<activation>
<property>
<name>hadoop.profile</name>
<value>2.0</value>
</property>
你可以将默认的profile改为hadoop-2.0,修改之后的配置文件如下：
<profile>
<id>hadoop-1.0</id>
<activation>
<property>
<name>hadoop.profile</name>
<value>1.0</value>
</property>
<profile>
<id>hadoop-2.0</id>
<activation>
<property>
<name>!hadoop.profile</name>
</property>