The problem was simple enough that my computer was already running an instance of Postgres that I was not aware was still running (not inside Docker) on :5432, checked with:

$ lsof -n -i:5432 | grep LISTEN

So I remembered I installed it via https://gist.github.com/sgnl/609557ebacd3378f3b72, I ran

$ pg-stop

And then I had no problem connecting to the Docker instance.

Edit (2019/07/02)

This question recently passed 10,000 views, so I thought I should elaborate more on why this happened.

Usually running through docker, using python and connecting to a postgres database requires you to install psycopg2, via pip3 install psycopg2, but if you run this command you will get:

Error: pg_config executable not found.

This is because psycopg2 requires an operating system install of the postgres libraries:

yum install postgresql-develapt-get install postgresql-client

Now, on a Mac, you will need to do the same with brew:

brew install postgresql

One thing that I didn't realise, is that on Mac, doing the above will not only install required libraries, but also start a database on :5432. Because this was all done in the background, it didn't occur to me that this was the problem as none of the usual errors popped up to inform that the port was being used, etc...

In my case, the problem may be due also to the use of a super user who is not postgres. Lhe developers chose the user georchestra in the geOrchestra docker composition.Extract from the dockerfile:

environment:  - POSTGRES_USER=georchestraHEALTHCHECK --interval=30s --timeout=30s \  CMD pg_isready -U $POSTGRES_USER

consequently, you will have to use the option "-U georchestra" to connect to the container.

$ docker exec -it docker_database_1 psql -U georchestrapsql (10.5 (Debian 10.5-1.pgdg90+1))Type "help" for help.

and of course, trying to connect with the user postgres causes an error:

docker exec -it docker_database_1 psql -U postgrespsql: FATAL:  role "postgres" does not exist