spark 资源动态分配
'spark.shuffle.service.enabled': 'true', 'spark.dynamicAllocation.enabled': 'false', 'spark.dynamicAllocation.initialExecutors': 50, 'spark.dynamicAllocation.minExecutors': 1, 'spark.dynamicAllocation.maxExecutors': 125, 'spark.sql.parquet.compression.codec': 'snappy', 'spark.yarn.executor.memoryOverhead': 4096, "spark.speculation": 'true', 'spark.kryoserializer.buffer.max': '512m',
欢迎转载,不必注明出处