How to Exclude or Overwrite Specific JARs in Databricks Jars

[sahil_s_jain]

• Spark Version in Databricks 15.5 LTS: The runtime includes Apache Spark 3.5.x, which defines the SparkListenerApplicationEnd constructor as:

  public SparkListenerApplicationEnd(long time)

  This constructor takes a single long parameter.

• Conflicting Spark Library in Databricks: The error arises due to a conflicting library: ----ws_3_5--core--core-hive-2.3__hadoop-3.2_2.12_deploy.jar. This library includes a different version of the SparkListenerApplicationEnd class, which defines the constructor as:

  public SparkListenerApplicationEnd(long time, scala.Option<Object> exitCode)

  This is method is present Spark 4.0.0-preview2 version.

• Impact: At runtime, the JVM attempts to use the single-parameter constructor (<init>(J)V) but fails because the conflicting library expects the two-parameter version. This mismatch leads to the NoSuchMethodError.

How can I exclude or overwrite a specific JAR file, ws_3_5--core--core-hive-2.3__hadoop-3.2_2.12_deploy.jar, from Databricks compute dependencies when launching a Databricks compute cluster.

[Alberto_Umana]

Hi @sahil_s_jain,

How are you installing the jar file in the cluster?

[sahil_s_jain]

I am creating a uber jar of my application with spark 3.5.0 spark depedencies and using jar submit on the cluster for execution.
As the spark libraries from the above jar "----ws_3_5--core--core-hive-2.3__hadoop-3.2_2.12_deploy.jar" are getting loaded.

the constructor SparkListenerApplicationEnd(long time, scala.Option<Object> exitCode) is creating a conflict.

In my code i have used a single argument constructor which is correct as per spark 3.5.0 but the loaded class from the above Databricks jar is expecting a constructor of 2 arguments which is not correct as per spark version 3.5.0

Constructor in spark 3.5.0

public SparkListenerApplicationEnd(long time)

Constructor in "----ws_3_5--core--core-hive-2.3__hadoop-3.2_2.12_deploy.jar"

public SparkListenerApplicationEnd(long time, scala.Option<Object> exitCode)

Need you inputs on this conflict, as the class "SparkListenerApplicationEnd" packed in "----ws_3_5--core--core-hive-2.3__hadoop-3.2_2.12_deploy.jar" is not same as the spark 3.5.0.