我正在使用Hadoop 2.6.0-cdh5.14.2，Hive 1.1.0-cdh5.14.2 在此系统中，确实存在一个具有183K +分区的巨大表，它是一个外部表，并且命令：
0: jdbc:hive2://hiveserver2.hd.docomodigital.> drop table unifieddata_work.old__raw_ww_eventsjson
不起作用，Metastore在600秒内没有回复，并且任务以错误结束。
我尝试使用以下范围删除分区：
0: jdbc:hive2://hiveserver2.hd.docomodigital.> alter table unifieddata_work.old__raw_ww_eventsjson drop PARTITION (country='ae', year='2017', month='01', day>'29', hour > '00' );
INFO : Compiling command(queryId=hive_20180821140909_ba6c4bb0-d0de-4fd3-a5ec-47e217289c6b): alter table unifieddata_work.old__raw_ww_eventsjson drop PARTITION (country='ae', year='2017', month='01', day>'29', hour > '00' )
INFO : Semantic Analysis Completed
INFO : Returning Hive schema: Schema(fieldSchemas:null, properties:null)
INFO : Completed compiling command(queryId=hive_20180821140909_ba6c4bb0-d0de-4fd3-a5ec-47e217289c6b); Time taken: 0.612 seconds
INFO : Executing command(queryId=hive_20180821140909_ba6c4bb0-d0de-4fd3-a5ec-47e217289c6b): alter table unifieddata_work.old__raw_ww_eventsjson drop PARTITION (country='ae', year='2017', month='01', day>'29', hour > '00' )
INFO : Starting task [Stage-0:DDL] in serial mode
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=01
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=02
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=03
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=04
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=05
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=06
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=07
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=08
INFO : Dropped the partition country=ae/year=2017/month=01/day=30/hour=09
... CUTTED HERE ...
它可以工作，但是Metastore发生了一些不好的事情：金丝雀停止工作。关于如何解决此问题的任何想法？有没有其他方法可以删除这么大的表？