Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Hi I am getting this erro on CI when I try to register my flow (which uses the `RunNamespacedJob` task);

How can I force it to defer trying to access the cluster until its actually running?

```Traceback (most recent call last):
  File "/usr/local/lib/python3.8/runpy.py", line 185, in _run_module_as_main
    mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
  File "/usr/local/lib/python3.8/runpy.py", line 111, in _get_module_details
    __import__(pkg_name)
  File "/home/circleci/project/prefect_pipelines/__init__.py", line 8, in &lt;module&gt;
    from prefect_pipelines.pod_tasks.flow import flow as pod_tasks
  File "/home/circleci/project/prefect_pipelines/pod_tasks/flow.py", line 12, in &lt;module&gt;
    make_coal_input.run()
  File "/home/circleci/.cache/pypoetry/virtualenvs/prefect-pipelines-3aSsmiER-py3.8/lib/python3.8/site-packages/prefect/utilities/tasks.py", line 454, in method
    return run_method(self, *args, **kwargs)
  File "/home/circleci/.cache/pypoetry/virtualenvs/prefect-pipelines-3aSsmiER-py3.8/lib/python3.8/site-packages/prefect/tasks/kubernetes/job.py", line 713, in run
    client.BatchV1Api, get_kubernetes_client("job", kubernetes_api_key_secret)
  File "/home/circleci/.cache/pypoetry/virtualenvs/prefect-pipelines-3aSsmiER-py3.8/lib/python3.8/site-packages/prefect/utilities/kubernetes.py", line 54, in get_kubernetes_client
    kubernetes_api_key = Secret(kubernetes_api_key_secret).get()
  File "/home/circleci/.cache/pypoetry/virtualenvs/prefect-pipelines-3aSsmiER-py3.8/lib/python3.8/site-packages/prefect/client/secrets.py", line 131, in get
    raise ValueError(
ValueError: Secrets should only be retrieved during a Flow run, not while building a Flow.```

My hope is that since I am running on k8s it should just work :tm: and no auth will be required - but obvs CI is wanting to know how to authenticate when I register...

Anybody get any ideas here?? Really don't understand why it needs to be able to connect to the cluster when it registers the flows!?

Hi <@U01K9JXEQAJ> - can you share a min reproducible example of your code?

It looks like your task is running (the error occurs in in the `run` method) which should never be called during registration. Perhaps you're calling run somewhere in your flow script?

Yeah defo - so I instantiate in one file `my_job = RunNamespacedJob(body=the_job_sepc)`

Then `import my_job` in the flow and then call `my_job.run()`

So I should just call `my_job()` ?

When defining a flow you should never call `run` (as that runs immediately). `my_job()` would be correct, yes.