java.lang.NoSuchMethodError: 'void com.amazonaws.util.IOUtils.release(java.io.Closeable, com.amazonaws.thirdparty.apache.logging.Log)'

[YashPayU]

Unable to save spark dataframe from local to s3 bucket.

Env Setup

Base image: gcr.io/datamechanics/spark:platform-3.2.1-hadoop-3.3.1-java-11-scala-2.12-python-3.8-dm18

I've following jars available in /opt/spark/jars: 'aws-java-sdk-bundle-1.11.901.jar', 'aws-java-sdk-core-1.11.797.jar', 'aws-java-sdk-glue-1.11.797.jar', 'hadoop-aws-3.3.1.jar',

Sample code

`from pyspark.sql import SparkSession
spark = SparkSession.builder
.config("spark.hadoop.fs.s3.impl", "org.apache.hadoop.fs.s3a.S3AFileSystem")
.config("spark.dynamicAllocation.enabled", "true")
.config("spark.dynamicAllocation.maxExecutors", "4")
.config("spark.dynamicAllocation.minExecutors", "1")
.config("spark.dynamicAllocation.initialExecutors", "1")
.config("spark.sql.parquet.datetimeRebaseModeInRead", "CORRECTED")
.config("spark.sql.legacy.pathOptionBehavior.enabled", "true")
.config("spark.sql.parquet.datetimeRebaseModeInWrite", "CORRECTED")
.getOrCreate()

source_file = "/workspaces/sample/test/*"
df = spark.read.parquet(source_file)
df.write.format("parquet").mode("append").save("s3a://MY_BUCKET/MY_FOLDER/")`

ERROR: java.io.IOException: regular upload failed: java.lang.NoSuchMethodError: 'void com.amazonaws.util.IOUtils.release(java.io.Closeable, com.amazonaws.thirdparty.apache.logging.Log)'

[debora-ito]

@YashPayU "NoSuchMethodError" indicates a version conflict between core and service modules of the SDK. To avoid this error you should use all the modules of the SDK in the exact same version.

Another best practice is: instead of using the full aws-java-sdk-bundle, add only the SDK service modules you are going to use. This way you'll have a smaller dependency size (aws-java-sdk-bundle-1.11.901.jar is almost 185MB), and less dependencies mean less dependency conflicts.

[niyushabaghayi]

Thanks, removing aws-java-sdk-bundle and using more specific modules fixed my problem! 🎉