I was playing with some new data sources in Zeppelin (master) and had an
issue on my --package declaration in zeppelin-env.sh.

First, the error stack was a little misleading, as it was only reporting
connection issues to Spark trying to retrieve paragraph status.

>From my initial investigation, we pretty much invoke the spark-submit from
interpreter.sh... and I was wondering if there is a good way to trap this
kind of issues and be able to provide a better response to the user (or
even better logging)...

One option is to have something similar to what spark does, and actually,
have a class handling some of the interpreter integration logic.

Thoughts  ? Any other possibilities ?

-- 
Luciano Resende
http://twitter.com/lresende1975
http://lresende.blogspot.com/

Reply via email to