Pyspark 集群模式异常 - Java 网关进程在向驱动程序发送其端口号之前退出

Question

在 apache 气流中，我编写了一个 PythonOperator，它使用 pyspark 在纱线集群模式下运行作业。我初始化 sparksession 对象如下。

    spark = SparkSession \
    .builder \
    .appName("test python operator") \
    .master("yarn") \
    .config("spark.submit.deployMode","cluster") \
    .getOrCreate()

但是，当我运行我的 dag 时，我得到了一个异常。

    Traceback (most recent call last):
  File "/usr/local/lib/python3.8/dist-packages/airflow/models/taskinstance.py", line 983, in _run_raw_task
    result = task_copy.execute(context=context)
  File "/usr/local/lib/python3.8/dist-packages/airflow/operators/python_operator.py", line 113, in execute
    return_value = self.execute_callable()
  File "/usr/local/lib/python3.8/dist-packages/airflow/operators/python_operator.py", line 118, in execute_callable
    return self.python_callable(*self.op_args, **self.op_kwargs)
  File "/catfish/dags/dags_dag_test_python_operator.py", line 39, in print_count
    spark = SparkSession \
  File "/usr/local/lib/python3.8/dist-packages/pyspark/sql/session.py", line 186, in getOrCreate
    sc = SparkContext.getOrCreate(sparkConf)
  File "/usr/local/lib/python3.8/dist-packages/pyspark/context.py", line 371, in getOrCreate
    SparkContext(conf=conf or SparkConf())
  File "/usr/local/lib/python3.8/dist-packages/pyspark/context.py", line 128, in __init__
    SparkContext._ensure_initialized(self, gateway=gateway, conf=conf)
  File "/usr/local/lib/python3.8/dist-packages/pyspark/context.py", line 320, in _ensure_initialized
    SparkContext._gateway = gateway or launch_gateway(conf)
  File "/usr/local/lib/python3.8/dist-packages/pyspark/java_gateway.py", line 105, in launch_gateway
    raise Exception("Java gateway process exited before sending its port number")
Exception: Java gateway process exited before sending its port number

我也设置了 PYSPARK_SUBMIT_ARGS，但它对我不起作用！

Answer 1

你需要在你的 ubuntu 容器上安装 spark。

RUN apt-get -y install default-jdk scala git curl wget
RUN wget --no-verbose https://downloads.apache.org/spark/spark-2.4.6/spark-2.4.6-bin-hadoop2.7.tgz
RUN tar xvf spark-2.4.6-bin-hadoop2.7.tgz
RUN mv spark-2.4.6-bin-hadoop2.7 /opt/spark
ENV SPARK_HOME=/opt/spark

不幸的是，您无法使用 PythonOperator 在纱线上运行 spark。我建议你使用SparkSubmitOperator 或BashOperator。