3

I've tried several ports without success: 8787 is indeed busy serving rstudio. I could redirect rstudio, but shouldn't the following work?

from distributed import Client, LocalCluster
cluster = LocalCluster( scheduler_port = 8785 , n_workers = 2 )

Error:

/home/ec2-user/anaconda3/lib/python3.6/site-packages/distributed/bokeh/core.py:56: UserWarning: 
Port 8787 is already in use. 
Perhaps you already have a cluster running?
Hosting the diagnostics dashboard on a random port instead.
  warnings.warn('\n' + msg)
tornado.application - ERROR - Multiple exceptions in yield list

...

~/anaconda3/lib/python3.6/multiprocessing/forkserver.py in connect_to_new_process(self, fds)
     64             raise ValueError('too many fds')
     65         with socket.socket(socket.AF_UNIX) as client:
---> 66             client.connect(self._forkserver_address)
     67             parent_r, child_w = os.pipe()
     68             child_r, parent_w = os.pipe()

ConnectionRefusedError: [Errno 111] Connection refused
user2105469
  • 1,413
  • 3
  • 20
  • 37

1 Answers1

4

I think you need to override the diagnostics_port instead:

https://github.com/dask/distributed/blob/2ec428ae5652f7d068baeb18223fb8a04ab8804e/distributed/deploy/local.py#L45-L49

Dave Hirschfeld
  • 768
  • 2
  • 6
  • 15