Question

我正在Airflow中创建一个计划数据流作业的每小时任务，但是当数据流作业实际成功时，Airflow Library提供的钩子大多数时间都会崩溃。

[2018-05-25 07:05:03,523] {base_task_runner.py:98} INFO - Subtask: [2018-05-25 07:05:03,439] {gcp_dataflow_hook.py:109} WARNING -   super(GcsIO, cls).__new__(cls, storage_client))
[2018-05-25 07:05:03,721] {base_task_runner.py:98} INFO - Subtask: Traceback (most recent call last):
[2018-05-25 07:05:03,725] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/bin/airflow", line 27, in <module>
[2018-05-25 07:05:03,726] {base_task_runner.py:98} INFO - Subtask:     args.func(args)
[2018-05-25 07:05:03,729] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/bin/cli.py", line 392, in run
[2018-05-25 07:05:03,729] {base_task_runner.py:98} INFO - Subtask:     pool=args.pool,
[2018-05-25 07:05:03,731] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/utils/db.py", line 50, in wrapper
[2018-05-25 07:05:03,732] {base_task_runner.py:98} INFO - Subtask:     result = func(*args, **kwargs)
[2018-05-25 07:05:03,734] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/models.py", line 1492, in _run_raw_task
[2018-05-25 07:05:03,738] {base_task_runner.py:98} INFO - Subtask:     result = task_copy.execute(context=context)
[2018-05-25 07:05:03,740] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/contrib/operators/dataflow_operator.py", line 313, in execute
[2018-05-25 07:05:03,746] {base_task_runner.py:98} INFO - Subtask:     self.py_file, self.py_options)
[2018-05-25 07:05:03,748] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 188, in start_python_dataflow
[2018-05-25 07:05:03,751] {base_task_runner.py:98} INFO - Subtask:     label_formatter)
[2018-05-25 07:05:03,753] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 158, in _start_dataflow
[2018-05-25 07:05:03,756] {base_task_runner.py:98} INFO - Subtask:     _Dataflow(cmd).wait_for_done()
[2018-05-25 07:05:03,757] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 129, in wait_for_done
[2018-05-25 07:05:03,759] {base_task_runner.py:98} INFO - Subtask:     line = self._line(fd)
[2018-05-25 07:05:03,761] {base_task_runner.py:98} INFO - Subtask:   File "/usr/local/lib/python2.7/site-packages/airflow/contrib/hooks/gcp_dataflow_hook.py", line 110, in _line
[2018-05-25 07:05:03,763] {base_task_runner.py:98} INFO - Subtask:     line = lines[-1][:-1]
[2018-05-25 07:05:03,766] {base_task_runner.py:98} INFO - Subtask: IndexError: list index out of range

我在Airflow github repo中看到该文件并且行错误不匹配，这使我认为来自Cloud Composer的实际Airflow实例已过时。有没有办法更新它？

Answer 1

这将在1.10或2.0中解决。

看看这个PR https://github.com/apache/incubator-airflow/pull/3165

这已合并为主人。您可以使用此PR代码并创建自己的插件。

Composer - 数据流挂钩崩溃

1 个答案: