Seems like the thriftServer cannot connect to Zookeeper, so it cannot get lock.
This is how it the log looks when I run SparkSQL: "load data inpath "kv1.txt" into table src;" log: 14/09/16 14:40:47 INFO Driver: <PERFLOG method=acquireReadWriteLocks> 14/09/16 14:40:47 INFO ClientCnxn: Opening socket connection to server SVR4044HW2285.hadoop.lpt.qa.nt.ctripcorp.com/10.2.4.191:2181. Will not attempt to authenticate using SASL (unknown error) 14/09/16 14:40:47 INFO ClientCnxn: Socket connection established to SVR4044HW2285.hadoop.lpt.qa.nt.ctripcorp.com/10.2.4.191:2181, initiating session 14/09/16 14:40:47 INFO ClientCnxn: Session establishment complete on server SVR4044HW2285.hadoop.lpt.qa.nt.ctripcorp.com/10.2.4.191:2181, sessionid = 0x347c1b1f78d495e, negotiated timeout = 180000 14/09/16 14:40:47 INFO Driver: </PERFLOG method=acquireReadWriteLocks start=1410849647447 end=1410849647457 duration=10> You can see, between the PERFLOG of acquireReadWriteLocks, the ClientCnxn will try to connect to Zookeeper. After the connection has been successfully established, the acquireReadWriteLocks phrase can be finished. But, when I run the ThriftServer, and run the same SQL. Here is the log: 14/09/16 14:40:09 INFO Driver: <PERFLOG method=acquireReadWriteLocks> It will wait here. So I doubt, the reason why "Drop" or "Load" failed in thriftServer mode, is because of the thriftServer cannot connect to Zookeeper. -- View this message in context: http://apache-spark-user-list.1001560.n3.nabble.com/SparkSQL-1-1-hang-when-DROP-or-LOAD-tp14222p14336.html Sent from the Apache Spark User List mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@spark.apache.org For additional commands, e-mail: user-h...@spark.apache.org
