On Hortonworks Sandbox 2.4, Spark use of SQLServer driver results in “No Suitable Driver Found”

Question

HDP version 2.4

I am trying to write to a SQL Server database via the Microsoft SQL Server JDBC driver sqljdbc41.jar. I'm executing the following in Zeppelin:

val foo = sqlContext.jdbc("jdbc:sqlserver://X;databaseName=X;user=X;password=X", "X")

which yields the "no suitable driver found" error.

This is a common issue and so far I've tried the following, without luck, based on various remedies I've found through the interwebs:

• add the full path to the JAR for the following properties in spark-defaults.conf:

 spark.driver.extraClassPath spark.executor.extraClassPath spark.files 

• add SPARK_CLASSPATH with the same JAR path in spark_env.sh

After making these changes I've stopped and started the Spark service through the Ambari interface, and I do see the changes reflected in both the Ambari portal and within the configuration files on the VM.

I did notice that many posts indicate you still have to provide the jar on the classpath when running spark-submit, but I'm accessing through Zeppelin, so not sure how that applies.

Answer 1

Try using JDBC Interpreter with Zeppelin 0.6.0 and configure MySQL driver.

You can also try to follow http://thedataist.com/tutorial-using-apache-zeppelin-with-mysql/

Perhaps you just need to change the connection string as this post suggests.