No Error When ElasticSearch HTTP Port Already in Use

I installed ES via Homebrew at one point is set it to start with launchd. Then I decided not to use it for the moment and thought I disabled it. In fact I never properly disabled it, so there was an ES instance accepting HTTP requests on port 9200 on my machine that I didn't notice.

However the point is there were no errors from Nuxeo at all (i.e. in server.log) during startup or any other time. Whenever I tried to access the embedded ES via HTTP I was really talking to my standalone cluster, I just didn't notice at first.

I'm not even 100% sure why Nuxeo worked properly but it did. Search was ok for example. Maybe the network stack is not involved when Nuxeo talks to the embedded ES?

Anyway some feedback in the logs would be cool to save some time