Trouble hitting a container's exposed port from a seperate container & host

I have two vagrant hosts running local on my machine and I am trying to hit a container within one host from a second container on the other host.

When I curl from within the container:

  • rg.apache.spark.SparkException: Invalid master URL: spark://tasks.501393358-spark-master:7077
  • How to serve static files from a Dockerized Python web app?
  • Migrating existing Docker containers to Docker Compose
  • Setting-up containers reverse proxy
  • Can we create a node group in docker swarm?
  • docker insecure registry daemon not using flag
  • curl search.mydomain.localhost:9090/ping

    I receive the curl: (7) Failed to connect to search.mydomain.localhost port 9090: Connection refused

    However when I curl without specifying the port:

    curl search.mydomain.localhost/ping

    I’m certain the port is properly exposed as if I try the same from within the host instead of within the container I get:

    curl search.mydomain.localhost:9090/ping

    Which shows that the service at port 9090 is exposed, however there is some networking issue with the container trying to reach it.

    A fellow dev running the same versions of vbox/vagrant/docker/docker-compose and using an identical commit of the repos has no trouble hitting the service from within the container. I’m really stumped as to what to try from here…

    I’m using the default bridge network:

    sudo brctl show
    bridge name bridge id       STP enabled interfaces
    docker0     8000.02427c9cea3c   no      veth5dc6655
    docker network inspect bridge
            "Name": "bridge",
            "Id": "e4b8df614d4b8c451cd4a26c5dda09d22d77de934a4be457e1e93d82e5321a8b",
            "Scope": "local",
            "Driver": "bridge",
            "IPAM": {
                "Driver": "default",
                "Config": [
                        "Subnet": "",
                        "Gateway": ""
            "Containers": {
                "1d67a1567ff698694b5f10ece8a62a7c2cdcfcc7fac6bc58599d5992def8df5a": {
                    "EndpointID": "4ac99ce582bfad1200d59977e6127998d940a688f4aaf4f3f1c6683d61e94f48",
                    "MacAddress": "02:42:ac:11:00:03",
                    "IPv4Address": "",
                    "IPv6Address": ""
                "3e8b6cbd89507814d66a026fd9fad26d649ecf211f1ebd72ed4689b21e781e2c": {
                    "EndpointID": "2776560da3848e661d919fcc24ad3ab80e00a0bf96673e9c1e0f2c1711d6c609",
                    "MacAddress": "02:42:ac:11:00:02",
                    "IPv4Address": "",
                    "IPv6Address": ""
            "Options": {
                "": "true",
                "": "true",
                "": "true",
                "": "",
                "": "docker0",
                "": "1500"

    I’m on Docker version 1.9.0, build 76d6bc9 and docker-compose version: 1.5.0.

    Any help would be appreciated.

  • Why can't I `docker exec` after installing the official Django docker image?
  • Docker: How to run cython_extensions?
  • Why does docker compose exit right after starting?
  • Updating a container created from a custom dockerfile
  • Vagrant docker-exec
  • Docker - Ubuntu - Nginx - MariaDB - Connection refused
  • One Solution collect form web for “Trouble hitting a container's exposed port from a seperate container & host”

    I resolved my issue which seems like it might be a bug. Essentially the container was inheriting the /etc/hosts from my local macbook, bypassing the /etc/hosts on the actual vagrant host running the container, causing my entry of “ search.mydomain.localhost” to make all connection attempts within the container redirect to itself.

    Docker will be the best open platform for developers and sysadmins to build, ship, and run distributed applications.