重新build hadoop后进入hive客户端异常

在自己虚拟机上使用mysql作为hive元数据存储, 修改配置如下:

1. <property>
2. <name>javax.jdo.option.ConnectionURL</name>
3. <value>jdbc:mysql://localhost:3306/metastore</value>
4. <description>JDBCconnectstringforaJDBCmetastore</description>
5. </property>
7. <property>
8. <name>javax.jdo.option.ConnectionDriverName</name>
9. <value>com.mysql.jdbc.Driver</value>
10. <description>DriverclassnameforaJDBCmetastore</description>
11. </property>
13. <property>
14. <name>javax.jdo.option.ConnectionUserName</name>
15. <value>hive</value>
16. <description>usernametouseagainstmetastoredatabase</description>
17. </property>
19. <property>
20. <name>javax.jdo.option.ConnectionPassword</name>
21. <value>hive</value>
22. <description>passwordtouseagainstmetastoredatabase</description>
23. </property>

并将mysql jdbc驱动拷贝到HIVE_HOME/lib目录下.
接下去登陆hive客户端，执行show databases；命令, 异常:

1. [ruizhe@localhost~]$hive
2. Hivehistoryfile=/tmp/ruize/hive_job_log_ruize_201204091822_467986476.txt
3. hive>showdatabases;
4. FAILED:Errorinmetadata:javax.jdo.JDOFatalInternalException:Unexpectedexceptioncaught.
5. NestedThrowables:
6. java.lang.reflect.InvocationTargetException
7. FAILED:ExecutionError,returncode1fromorg.apache.hadoop.hive.ql.exec.DDLTask
8. hive>

被这个问题困扰好久, 同样的配置在自己笔记本环境上可以，但在公司机器上异常,
最后网上搜索了一把，发现解决方法如下:

1. delete$HADOOP_HOME/buildandeverythingshouldbefine

切换到 HADOOP_HOME目录:

1. [hadoop@localhostbuild]$pwd
2. /opt/app/hadoop-0.20.2-cdh3u3/build
3. [hadoop@localhostbuild]$ls
4. antc++classescontribexampleshadoop-core-0.20.2-cdh3u3.jarhadoop-tools-0.20.2-cdh3u3.jarivysrctesttoolswebapps
5. [hadoop@localhostbuild]$

果然在HADOOP_HOME/build目录下有新build的信息(我用ant重新build过hadoop)
直接删除build目录:

1. [hadoop@localhosthadoop-0.20.2-cdh3u3]$rm-rfbuild

重新进入hive客户端:

1. [ruize@localhost~]$hive
2. Hivehistoryfile=/tmp/ruize/hive_job_log_ruize_201204091826_1452110220.txt
3. hive>showdatabases;
4. OK
5. default
6. Timetaken:1.786seconds
7. hive>

这次OK了!

出处http://yaoyinjie.blog.51cto.com/3189782/830481