最近将公司各个运营点的集群逐步转移入CDH。第一个遇到的问题就是Sqoop2的使用问题。
具体说来，分为下面几个方面。
Date类型字段的parse错误
测试sqoop2的时候就简单的测了下varchar和number类型的数据，发现没问题后就写了报告。现在我知道错了。
实际使用过程中，发现遇到表中有Date类型的column时，所有的job都会报以下错误：
2015-01-28 16:22:29,964 INFO [IPC Server handler 0 on 55187] org.apache.hadoop.mapred.TaskAttemptListenerImpl: JVM with ID : jvm_1422410148386_0003_m_000002 asked for a task
2015-01-28 16:22:29,966 INFO [IPC Server handler 0 on 55187] org.apache.hadoop.mapred.TaskAttemptListenerImpl: JVM with ID: jvm_1422410148386_0003_m_000002 given task: attempt_1422410148386_0003_m_000000_0
2015-01-28 16:22:35,516 FATAL [IPC Server handler 2 on 55187] org.apache.hadoop.mapred.TaskAttemptListenerImpl: Task: attempt_1422410148386_0003_m_000000_0 - exited : org.joda.time.LocalDateTime.parse(Ljava/lang/String;)Lorg/joda/time/LocalDateTime;
2015-01-28 16:22:35,517 INFO [IPC Server handler 2 on 55187] org.apache.hadoop.mapred.TaskAttemptListenerImpl: Diagnostics report from attempt_1422410148386_0003_m_000000_0: Error: org.joda.time.LocalDateTime.parse(Ljava/lang/String;)Lorg/joda/time/LocalDateTime;
2015-01-28 16:22:35,524 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: Diagnostics report from attempt_1422410148386_0003_m_000000_0: Error: org.joda.time.LocalDateTime.parse(Ljava/lang/String;)Lorg/joda/time/LocalDateTime;
2015-01-28 16:22:35,542 INFO [AsyncDispatcher event handler] org.apache.hadoop.mapreduce.v2.app.job.impl.TaskAttemptImpl: attempt_1422410148386_0003_m_000000_0 TaskAttempt Transitioned from RUNNING to FAIL_CONTAINER_CLEANUP
