hadoop集群上运行自定义wordcount