Error: Could not find or load main class org.apache.spark.executor.CoarseGrainedExecutorBackend
This looks like mismatch between your Hadoop/Spark installation and Spark.jl’s version of base libraries. Please ask your Hadoop admin about:
- Spark version
- Hadoop version
- YARN version
If they are different from what we have, try changing the version in that file to the ones in your cluster, rebuild Spark (Pkg.build("Spark")
) and try again.
Quite unlikely to work, but you may also try solution from here, e.g.:
sess = SparkSession(master="yarn-client",
config=Dict("spark.driver.extraJavaOptions" => "-Diop.version=4.1.0.0"))