Connection between TempletonJob and Worker Nodes remains in FIN_WAIT_2 state for long time

[mahender bigdata]

Hi,
We have submitted too many jobs to webhcat (templeton) reason is our HQL 
has multiple hive statements,each hive statement is submitted as a job 
causing too many job, after some times all the submitted job are in 
pending state. later after waiting for 2 hrs, all the pending jobs got 
completed. But we were not able to submit  jobs to WEBHCAT, When we 
remote head node and run netstat -a, I see too many ports open 
(incremental manner ) to templeton port. All the Connections are in 
FIN_WAIT_2, where client is not sending 'FIN' token to server. Is there 
a way to set time out for server to close this connection after some hr 
or mins. So that we dont have too many connection or ports opened and 
get into hang state.

Another Question, can we run all the hive statements in HQL file in 
single shot rather than each individual hive statement has hive job.

Thanks in advance.