Facebox does not start when *_proxy environment variables are passed to the container

[jperville]

Symptoms

In http proxy environments, the machinebox/facebox docker image never displays "machine is ready" in the log. Accessing the UI works but it is impossible to use the API (eg. to teach)

How to reproduce

$ docker run --rm -ti -p 8080:8080 -e no_proxy=localhost -e http_proxy=${http_proxy} -e https_proxy=${https_proxy} -e MB_KEY="XXXXXXXXX" machinebox/facebox
2019/01/18 14:07:45 Number of MB_WORKERS is set to 4
[INFO]     starting...

    Welcome to Facebox by Machine Box
    (facebox 7f66aba)

    Visit the console to see what this box can do:
    http://localhost:8080

    If you have any questions or feedback, get in touch:
    https://machinebox.io/contact

    Please consider buying a subscription:
    https://machinebox.io/#pricing

    Report bugs and issues:
    https://github.com/machinebox/issues

    Tell us what you build on Twitter @machineboxio

# log below happens when trying to use /teach API
[ERROR]    (500) faceClient call error: box is not ready, try again later

Explanation

Looking in the http proxy logs, we can see that facebox tries to contact its workers at URL http://0.0.0.0:WORKERPORT/info .. This does not work because the http request is routed to the http proxy, which is not aware of the workers.

# tail of /var/log/squid/access.log
1547820499.053      1 10.0.3.1 TCP_MISS/503 3962 GET http://0.0.0.0:10001/info - HIER_DIRECT/0.0.0.0 text/html
1547820499.054      1 10.0.3.1 TCP_MISS/503 3962 GET http://0.0.0.0:10002/info - HIER_DIRECT/0.0.0.0 text/html
1547820499.055      1 10.0.3.1 TCP_MISS/503 3962 GET http://0.0.0.0:10004/info - HIER_DIRECT/0.0.0.0 text/html
1547820499.055      1 10.0.3.1 TCP_MISS/503 3962 GET http://0.0.0.0:10003/info - HIER_DIRECT/0.0.0.0 text/html

Workaround

I extended the machinebox/facebox docker image to use a custom entrypoint which looks like this:

#!/bin/sh

# mb-facebox never initializes if proxy_* environment variables are present
# the reason is that it tries to connect to workers using http://0.0.0.0:<workerport>/info
# which will fail if 0.0.0.0 is not listed in ${no_proxy}
if [ -n "${no_proxy}" ]
then
  no_proxy="0.0.0.0,${no_proxy}"
else
  no_proxy="0.0.0.0"
fi
export no_proxy

exec /app/facebox

[dahernan]

ohh thanks for the report and the workaround, we will try to fix it soon.

[dahernan]

That should be solved in the last pull, thanks.