Why doesn't a port open when I raise a service with docker-compose?

Question

linux

Hello, I need help.
I'm trying to deploy a 3-container system using docker-compose.
Containers:

1. Node.JS application listening on ports 44343 and 17117.
   
2. Redis from Docker Hub
   
3. Application listening on port 19833
   

When I run the whole structure through docker-compose container 1 cannot reach the port of container 3 (I check via telnet). Thus everything correctly responds. Also container 1 connects fine to radish port.
If I run container 3 separately, with port forwarding through the -p flag, I can telnet from the parent system without any problems.

# cat docker-compose.yml 
version: "2"
services:
    nodejs:
        image: nodejs
        ports:
        - "80:44343"
        links:
        - c
        - redis
    c:
        image: c
    redis:
        image: redis

# docker ps
CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS                              NAMES
a6b573d13236        nodejs              "npm start"              7 minutes ago       Up 7 minutes        17117/tcp, 0.0.0.0:80->44343/tcp   p_nodejs_1
c572e763618f        c       "/bin/sh -c ./src/..."   7 minutes ago       Up 7 minutes        19833/tcp                          p_c_1
4b085da7750c        redis               "docker-entrypoint..."   7 minutes ago       Up 7 minutes        6379/tcp                           p_redis_1

# docker -v
Docker version 17.09.0-ce, build afdb6d4

# docker-compose -v
docker-compose version 1.16.1, build 6d1ac219

# nmap 0.0.0.0 -p 19833

Starting Nmap 6.40 ( http://nmap.org ) at 2017-11-06 17:12 MSK
Nmap scan report for 0.0.0.0
Host is up (680s latency).
PORT      STATE SERVICE
19833/tcp open  unknown

Nmap done: 1 IP address (1 host up) scanned in 0.06 seconds

# docker exec -it p_nodejs_1 bash
# nmap c -p 19833

Starting Nmap 7.01 ( https://nmap.org ) at 2017-11-06 14:15 UTC
Nmap scan report for c (172.18.0.3)
Host is up (0.000046s latency).
rDNS record for 172.18.0.3: pool_c_1.pool_default
PORT      STATE  SERVICE
19833/tcp closed unknown
MAC Address: 02:42:AC:12:00:03 (Unknown)

Nmap done: 1 IP address (1 host up) scanned in 0.48 seconds

The port is available from the host machine, and the docker is closed from inside the network. Redis is working correctly.

# nmap redis -p 6379

Starting Nmap 7.01 ( https://nmap.org ) at 2017-11-06 14:21 UTC
Nmap scan report for redis (172.18.0.2)
Host is up (0.000094s latency).
rDNS record for 172.18.0.2: pool_redis_1.pool_default
PORT     STATE SERVICE
6379/tcp open  unknown
MAC Address: 02:42:AC:12:00:02 (Unknown)

Nmap done: 1 IP address (1 host up) scanned in 0.48 seconds

# docker network inspect p_default

[
    {
        "Name": "p_default",
        "Id": "ec82da66485580f74c8897ad40db613b7997f8fd0ccf18fd0d025861fc3d96c4",
        "Created": "2017-11-06T17:41:07.187705953+03:00",
        "Scope": "local",
        "Driver": "bridge",
        "EnableIPv6": false,
        "IPAM": {
            "Driver": "default",
            "Options": null,
            "Config": [
                {
                    "Subnet": "172.18.0.0/16",
                    "Gateway": "172.18.0.1"
                }
            ]
        },
        "Internal": false,
        "Attachable": true,
        "Ingress": false,
        "ConfigFrom": {
            "Network": ""
        },
        "ConfigOnly": false,
        "Containers": {
            "1f7c97272f8c8d0635fe66714f9359f27f6eb24a1a3372162e18956c1a874542": {
                "Name": "pool_redis_1",
                "EndpointID": "6c565c1b720671c8cca12c32e357127de5acffbaf30c9fc3e7dea75040b62d45",
                "MacAddress": "02:42:ac:12:00:02",
                "IPv4Address": "172.18.0.2/16",
                "IPv6Address": ""
            },
            "6c5dadbdd2b56935cd2dc4206d2d16a430ca336c2bfed7437e18cc6ac2ae9384": {
                "Name": "pool_c_1",
                "EndpointID": "36fe780b504b79d0e4013de82276ca91b744204e613e6f2c168e02d16989e0e6",
                "MacAddress": "02:42:ac:12:00:03",
                "IPv4Address": "172.18.0.3/16",
                "IPv6Address": ""
            },
            "d45c570bde8a9b99fe4f3ef75d47ff9cee5cc1cbb654567714436d602c8cdbd8": {
                "Name": "pool_nodejs_1",
                "EndpointID": "ac55025cfa9189c0f2d6e247f3ea7c3aec668e4665c1490a65f2473240566ca8",
                "MacAddress": "02:42:ac:12:00:04",
                "IPv4Address": "172.18.0.4/16",
                "IPv6Address": ""
            }
        },
        "Options": {},
        "Labels": {
            "com.docker.compose.network": "default",
            "com.docker.compose.project": "p"
        }
    }
]

Answer 1

It turned out that I myself was a dunce and the docker had nothing to do with it. The bottom line is this:
a service was started in container "c" that listened on port 19833, but only with ip 127.0.0.1. Accordingly, it is impossible to connect from any other machine. Since it was not possible to force the application to listen from the 0.0.0.0 address, I solved the problem with a crutch. Installed the redir package and through it configured traffic redirection from the external port to the internal 127.0.0.1

Answer 2

You need to tweak docker-compose.yml a bit

version: '3.1'

services:
  nodejs:
        image: nodejs
        ports:
        - "44343:44343"
        - "17117:17117"
        depends_on:
        -c 
        - redis
        links:
        - c:custom_app
        - redis:redis
    c:
        image: c
        ports:
        - "19833:19833"
    redis:
        image: redis

Inside your first container that you need to reach redis from, use custom_appand redisto connect to the appropriate containers.