hdfs datanode 启动失败

hadoop-root-datanode-macmini.log中：

2015-03-12 23:52:33,671 FATAL org.apache.hadoop.hdfs.server.datanode.DataNode: Initialization failed for Block pool <registering> (Datanode Uuid unassigned) service to localhost/127.0.0.1:9000. Exiting.
java.io.IOException: Incompatible clusterIDs in /hdfs/name/dfs/data: namenode clusterID = CID-70d64aad-1dfe-4f87-af15-d53ff80db3dd; datanode clusterID = CID-388a9ec6-cb87-4b0d-97c4-3b4d5c787b76

原因是：

namenode与datanode的clusterID在重新格式化namenode以后已经不再匹配，datanode无法启动。

另外：

此错误会导致在hive导入数据时发生如下错误（由于metadata不存在hdfs中，故create table并无报错）：

hive> load data local inpath '/root/dbfile' overwrite into table employees PARTITION (country='US', state='IL');
Loading data to table default.employees partition (country=US, state=IL)
Failed with exception Unable to move source file:/root/dbfile to destination hdfs://localhost:9000/user/hive/warehouse/employees/country=US/state=IL/dbfile
FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.MoveTas

 

解决方法：

将hdfs存储数据的所在目录删掉，重新格式化hdfs（相关参数：dfs.name.dir  dfs.data.dir）：(具体目录看hadfs-site.xml文件）

hadoop namenode -format