Created 05-17-2016 12:47 PM
Hi Guys,
Sorry to sound dumb, but what is the use of Spark Thirft Server? We have Kerberized HDP 2.4.0 cluster. Recently installed Spark component on the HDP. Now when I am seeing the setup document, I see the option that talks about adding Spark Thrift Server component.
I googled a bit, it talks about JDBC along with thrift spark server. Not very clearly understood though.
I would like to understand more before making any changes to our Kerberized HDP 2.4 .
Many thanks.
Created 05-17-2016 01:02 PM
Spark thrift server is pretty similar to hiveserver2 thrift, rather submitting the sql queries as hive mr job it will use spark SQL engine which underline uses full spark capabilities. As an use case tools like Tableau can easily connect to spark thrift server through ODBC driver just like hiveserver2 and access the hive or spark temp tables to run the sql queries on spark framework.
Created 05-17-2016 01:02 PM
Spark thrift server is pretty similar to hiveserver2 thrift, rather submitting the sql queries as hive mr job it will use spark SQL engine which underline uses full spark capabilities. As an use case tools like Tableau can easily connect to spark thrift server through ODBC driver just like hiveserver2 and access the hive or spark temp tables to run the sql queries on spark framework.
Created 02-11-2017 12:38 AM
Question, Spark Thrift Server should stay on Master or Worker nodes ?
Created 02-02-2018 08:11 PM
with hiveserver2 you can also submit job on spark if spark is configured as execution engine in hive, right?
Created 05-17-2016 03:02 PM
Check out this article and demo for a full explanation and working example:
https://community.hortonworks.com/repos/29883/sparksql-data-federation-demo.html
Created 05-17-2016 06:46 PM
You only need to setup Spark Thrift Server if you need to provide access to SparkSQL via JDBC or ODBC. If you want to only use SparkSQL either programmatically (submit a spark app with Spark-submit) or with Spark-Shell, you don't need Spark Thrift Server.