Network issue

Hi,
I am trying to run a python script that uses the Dataframe library of Dask inside a docker container.

If I run the container manually by starting the container with the command docker run -it --entrypoint /bin/bash docker_image and then I run the python script inside the container, everything works perfectly.

If I start the container by using docker-py Dask is not able to start, I obtain many TCP connection errors like this one:

[34mdistributed.utils - ERROR - Timed out trying to connect to 'tcp://127.0.0.1:34017' after 10 s: Timed out trying to connect to 'tcp://127.0.0.1:34017' after 10 s: connect() didn't finish in time[0m
[34mTraceback (most recent call last):[0m
[34m File "/usr/local/lib/python3.6/dist-packages/distributed/comm/core.py", line 234, in connect[0m
[34m _raise(error)[0m
[34m File "/usr/local/lib/python3.6/dist-packages/distributed/comm/core.py", line 215, in _raise[0m
[34m raise IOError(msg)[0m
[34mOSError: Timed out trying to connect to 'tcp://127.0.0.1:34017' after 10 s: connect() didn't finish in time[0m

This is the code that I am using to start the container

container = client.containers.run(
docker_tag,
cmd,
volumes={
os.path.abspath('df_benchmark'):
{'bind': '/home/app/df_benchmark', 'mode': 'ro'},
os.path.abspath('datasets'):
{'bind': '/home/app/datasets', 'mode': 'ro'},
os.path.abspath('results'):
{'bind': '/home/app/results', 'mode': 'rw'},
},
cpuset_cpus=str(cpu_limit),
mem_limit=mem_limit,
detach=True)

Any suggestion? Maybe I have to add some extra parameters? From the documentation I can't figure out.

Thanks.

No project file or test is named. Start by reproducing the difference between manual docker run and client.containers.run using the provided command, then inspect the container networking and the TCP endpoint reported in the timeout. Done means identifying whether docker-py needs additional configuration or documenting the required usage.