I personally use spark submit as it's agnostic to which platform your spark clusters are working on e.g. Emr dataproc databricks etc
On Thu, 23 Feb 2017 at 08:53, nancy henry <nancyhenry6...@gmail.com> wrote: > Hi Team, > > I have set of hc.sql("hivequery") kind of scripts which i am running right > now in spark-shell > > How should i schedule it in production > making it spark-shell -i script.scala > or keeping it in jar file through eclipse and use spark-submit deploy mode > cluster? > > which is advisable? >