Failed to execute Spark SQL query: java.lang.ExceptionInInitializerError: null

[eloyga]

I got this error:
Execute failed: Failed to execute Spark SQL query: java.lang.ExceptionInInitializerError: null;
when I try to use any row manipulation in Spark such as: Spark Row Filter and Spark SQL Query
The same error for both nodes.
I have created the Spark Context (Livy) in AWS EMR. I am using Knime Analytics Platform 4.1. Livy 0.6.0 in EMR.

How to make sure that SQL environment is initialized?
Thank you
Eloy

[bjoern.lohrmann]

Hi @eloyga

I have tried to reproduce your issue (latest EMR version, fresh cluster with Spark, Hive, Livy, Hadoop, …) but was not successful, so maybe this is a configuration issue in your EMR cluster.

Does your cluster contain Hive? If not, then it might be worth trying to add Hive and see whether this changes anything.

Are you using any non-default spark settings? Typically those are applied either via EMR configuration or via the Create Spark Context (Livy) node (Advanced tab).

It could be useful to have a screenshot of your KNIME workflow as well as the YARN logs of the Spark application. Depending on your EMR configuration, the YARN logs can be obtained from the S3 bucket that your EMR cluster logs to.

Best,
Björn