I know this has been posted before here https://discourse.bokeh.org/t/failed-to-load-resource-net-err-connection-refused-flask-gunicorn-panel-docker/9589, but when I’ve tried to follow the suggestions I am still unable to resolve the error: Failed to load resource: net::ERR_BLOCKED_BY_CLIENT.
Like the previous post, I am also using the example set here bokeh/flask_gunicorn_embed.py at branch-2.4 · bokeh/bokeh · GitHub to try and get bokeh running on a remote computer that is connected to my network.
I have changed this line to explicitly set the port
sockets, port = bind_sockets("localhost", 5006)
I’ve also tried many different variations with setting the extra_websocket_origins parameter e.g.
bokeh_tornado = BokehTornado({'/bkapp': bkapp}, extra_websocket_origins=["*"])
bokeh_tornado = BokehTornado({'/bkapp': bkapp}, extra_websocket_origins=["localhost:8000"])
bokeh_tornado = BokehTornado({'/bkapp': bkapp}, extra_websocket_origins=["0.0.0.0:8000"])
And the Dockerfile exposes both port 8000 and 5006
FROM python:3.8
RUN apt-get update && apt-get upgrade -y
WORKDIR /tool
ADD app.py app.py
COPY templates templates
EXPOSE 8000
EXPOSE 5006
Here are the docker commands I’ve tried:
docker run -d --restart always -p 8000:8000 -p 5006:5006 -v /home/bli/miniconda3/:/home/bli/miniconda3/ sanger_flask gunicorn --bind 0.0.0.0:8000 app:app
docker run -d --restart always -p 8000:8000 -v /home/bli/miniconda3/:/home/bli/miniconda3/ sanger_flask gunicorn --bind 0.0.0.0:8000 app:app
I thought it might have been a firewall issue, but I was able to access other test websites I built with flask + docker using ports 8000, 5006, 5001 ect.
My hunch is their is still some issue is with extra_websocket_origins or I’m not correctly exposing the ports between docker and bokeh.
Any examples of how this works with docker would be wonderful.
Best,
Andrei
