06-25-2021 08:48 AM
06-25-2021 08:49 AM
setting the parameter ‘spark.cleaner.ttl’ or by dividing the long running jobs into different batches and writing the intermediary results to the disk.
View solution in original post
