docker-proxy | Squid and Stunnel assembled as a proxy server | Proxy library

Docker itself does not use ports 80 or 443, so you have something else running there (the most common offender is going to be apache). The docs at https://ddev.readthedocs.io/en/latest/users/troubleshooting/#webserver-ports-are-already-occupied-by-another-webserver explain how to debug this. Since it says docker-proxy is running there, it probably means that you have a docker container already running and bound to port 80.

Try these things:

• ddev poweroff (to make sure all ddev things are stopped)
• docker ps -a will show you other containers that may be running
• docker rm -f $(docker ps -aq) will stop all running containers.

I think when you do these things you'll find docker not listening any more; now you just have to find out why it was listening.

As you already know, the docs for various kinds of sharing are at https://ddev.readthedocs.io/en/latest/users/topics/sharing/ and you may also be interested in the general on-server approaches in https://ddev.readthedocs.io/en/latest/users/alternate-uses/#casual-project-webhosting-on-the-internet-including-lets-encrypt

The problem was that the server was listening on localhost inside the container (source). I had to change appsettings.json to make it listen on 0.0.0.0:

Here's a Dockerfile that installs the docker-ce-cli package on the Maven image you use as your base. You don't need docker-ce or containerd. The commands to install are taken from the docker website.

You're running your app in your localhost address (127.0.0.1), run it using 0.0.0.0 as host and it should work.

The difference between localhost and 0.0.0.0 is that the former is a loopback address, while the latter is a meta-address that maps all addresses from your instance.

Working example: app.py

It looks like you have a container configured to bind mount /var/lib/docker.sock and the daemon restarted that container before creating the socket. There's been some tweeks to packaging in recent releases to reduce this chance. Otherwise you may want to mount the entire directory instead of a single file.

To fix, try stopping docker, deleting the empty directory, and restarting docker to see if the socket gets created first (it's a race condition).

You must have containers running from outside this compose setup. Use docker ps to list all running containers and stop them. Afterwards, to be sure, use docker network prune to remove orphaned networks.

After that your compose setup should start normally.

When creating the libp2p node the ip address must be 0.0.0.0 instead of 127.0.0.0 or the real server's ip.

Then, when trying to remotely reach the libp2p node, the ip has to be that of the server (the one that ping your-domain.com returns)

Imo 80% of development pain comes from too simple examples in the "documentation"

I found where the problem was.

I was using the image https://hub.docker.com/r/onlyoffice/documentserver and i set up the https config

see "Running ONLYOFFICE Docs using HTTPS" in this page : https://helpcenter.onlyoffice.com/installation/docs-community-install-docker.aspx

And on this image they were an automatic proxy to redirect https to https, so i was not linked to the nginx conf on my server it was only inside the docker-proxy

So the 2 solutions that i found:

• remove the https configuration so it would be available in http
• or binding host server port 443 (https) to the 443 port of the ONLYOFFICE container, so the redirection works

At the end... my problem was that in my Redhat installation, as opposed to my Ubuntu, we had SELinux enabled.

Disabling it finally had curl --unix-socket /run/docker.sock http://docker/images/json working from within my composer containers.

To disable Selinux: edit (you may need to impersonate as root using sudo su root) file /etc/selinux/config - replace SELINUX=enforcing with SELINUX=disabled

Restart the linux server and that's it.

Remark: This may obviously not be an acceptable solution in a production environment. If this is your case, you will need to properly configure SELinux permissions settings. I was simply assigned a task to identify why this problem was happening in one of our dev machines, so disabling it suffices my needs for now.