KubeCluster() fails with deploy_mode="local": 'Scheduler' object has no attribute 'service'

[bnaul]

After #525, I am seeing the following when trying to start a (classic) KubeCluster with deploy_mode="local":

  File "/usr/local/lib/python3.10/site-packages/dask_kubernetes/classic/kubecluster.py", line 503, in __init__
    super().__init__(**self.kwargs)
  File "/usr/local/lib/python3.10/site-packages/distributed/deploy/spec.py", line 275, in __init__
    self.sync(self._start)
  File "/usr/local/lib/python3.10/site-packages/distributed/utils.py", line 338, in sync
    return sync(
  File "/usr/local/lib/python3.10/site-packages/distributed/utils.py", line 405, in sync
    raise exc.with_traceback(tb)
  File "/usr/local/lib/python3.10/site-packages/distributed/utils.py", line 378, in f
    result = yield future
  File "/usr/local/lib/python3.10/site-packages/tornado/gen.py", line 762, in run
    value = future.result()
  File "/usr/local/lib/python3.10/site-packages/dask_kubernetes/classic/kubecluster.py", line 648, in _start
    self.scheduler.service.metadata.name,
AttributeError: 'Scheduler' object has no attribute 'service'. Did you mean: 'services'?

It seems that self.scheduler here is the distributed.Scheduler object, not the Pod. Without the deploy_mode kwarg it works fine.

cc @jacobtomlinson

[bnaul]

I'm guessing something like this would work instead?

ipdb>  f"localhost:{self.scheduler.services['dashboard'].port}"
'localhost:8787'

not sure there's a more elegant way than gating on if self._deploy_mode == "local":...?

[jacobtomlinson]

Thanks for raising this @bnaul.

We definitely shouldn't be trying to port forward the dashboard in local mode. I think it would be fine to add a check here.

Do you have any interest in raising a PR?