Query Regarding HIVE-1844.

[Mohit <mo...@huawei.com>]

Hello He Yongqiang /All,

 

I was going through the defect HIVE-1844, but I couldn't able to reproduce
the scenario using Hive 0.5 version , though I  saw some OOM consistently
while Copy Task @ server side, but the client didn't hanged.

 

As per you what could have made client hanged? In my case,  Hive client was
able to get proper response from thrift whenever OOM occurred at Server
side.

like , java.sql.SQLException: org.apache.thrift.TApplicationException :
Internal error processing execute

 

Kindly provide me pointers on reproducing it. Do I need to do more
regression on it?

 

Just a thought/observation,

And as per code change , why the OOM was caught too early (that too in the
form of Throw able, which will eat other exception as well) ?

It would have been eventually caught by
ThriftHive$Processor$execute.process() and appropriate actions would have
been taken, so I was wondering how the code change helped preventing client
hang?

 

 

Thanks and Regards,

-Mohit