Some Qs on the life cycle of a Spark interpreter

[Sourav Mazumder <so...@gmail.com>]

Hi,

In my understanding when I run any paragraph on any notebook using Spark
interpreter it internally starts a Spark Context/Spark Driver. And that
keeps on running.

Here are my questions on that -

1. Is the Zeppelin Interpreter process same as this Spark Context/Spark
Driver ?

1. Is there a way I can stop any such running Spark Context/Spark Driver
gracefully without restarting the Zeppelin server.

2. Many a times I'm getting the error that  says 'Connection refused at
RemoteINterpreter'. I believe in this case what is happening is the Spark
Context/Spark Driver is getting killed. So Zeppelin process is not able to
communicate with the same. Is this understanding right ?

4. Many a times I see the error that Scheduler Process is not running/got
killed . The only option in that case is to restart the Zeppelin server. Is
this Scheduler process spawns the Interpreter process ? How do we configure
the memory etc for the same ?

Any pointer would be helpful on these doubts.

Regards,
Sourav