Airflow CeleryExecutor - 'int' 对象在 Celery 中没有属性 'startswith'

Question

Airflow 2.0 正在排队但未在我的开发环境中启动任务。

DAG 和 Pool 设置有效，但是当我触发它们时每个 dag 中的所有任务都是 queued，并且永远不会运行。

输入 airflow celery worker 时，出现以下错误：

Traceback (most recent call last):
  File "/usr/local/bin/airflow", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/site-packages/airflow/__main__.py", line 40, in main
    args.func(args)
  File "/usr/local/lib/python3.8/site-packages/airflow/cli/cli_parser.py", line 48, in command
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.8/site-packages/airflow/utils/cli.py", line 92, in wrapper
    return f(*args, **kwargs)
  File "/usr/local/lib/python3.8/site-packages/airflow/cli/commands/celery_command.py", line 188, in worker
    _run_worker(options=options, skip_serve_logs=skip_serve_logs)
  File "/usr/local/lib/python3.8/site-packages/airflow/cli/commands/celery_command.py", line 94, in _run_worker
    celery_app.worker_main(options)
  File "/usr/local/lib/python3.8/site-packages/celery/app/base.py", line 365, in worker_main
    return instantiate(
  File "/usr/local/lib/python3.8/site-packages/celery/bin/base.py", line 283, in execute_from_commandline
    self.maybe_patch_concurrency(argv)
  File "/usr/local/lib/python3.8/site-packages/celery/bin/base.py", line 315, in maybe_patch_concurrency
    maybe_patch_concurrency(argv, *pool_option)
  File "/usr/local/lib/python3.8/site-packages/celery/__init__.py", line 143, in maybe_patch_concurrency
    pool = _find_option_with_arg(argv, short_opts, long_opts)
  File "/usr/local/lib/python3.8/site-packages/celery/__init__.py", line 95, in _find_option_with_arg
    if arg.startswith('-'):
AttributeError: 'int' object has no attribute 'startswith'

在我的暂存、生产环境中，运行任务以及我检查 airflow celery worker 都没有问题。它会启动或警告我它已经运行（如预期的那样）。

envs 之间没有区别，但我猜测问题发生在服务器上最近部署之后。

正如我所见，celery 收到了错误的参数：

AttributeError: 'int' object has no attribute 'startswith'

但是如何跟踪哪些参数 Airflow 试图传递给 celery？我不知道如何调试它。

Answer 1

通过将 celery 升级到最新版本从 4.4.2 到 5.1.2 解决。

似乎版本 4.4.2（气流依赖之一）有一个参数错误。

如果有任何其他解决此问题的建议，请随时在此处提及。

Answer 2

对于 airflow 2.1.0+，你需要 celery 5.1.2+

文档在这里：https://airflow.apache.org/docs/apache-airflow-providers-celery/stable/index.html#pip-requirements 它解释了对 celery 5.1.2+

的需求

Airflow CeleryExecutor - 'int' 对象在 Celery 中没有属性 'startswith'

Airflow CeleryExecutor - 'int' object has no attribute 'startswith' in Celery

python

celery

airflow

airflow-scheduler

directed-acyclic-graphs