docker in docker localhost networking issue

Question

I have a case where I want to do a quick healthcheck on a docker-compose up -d where I can successfully curl the endpoint being created by the compose file.

Here is what the relevant parts of my file look like:

compose:
    stage: compose
    image: tmaier/docker-compose:latest
    services:
      - docker:dind
    before_script:
      - apk update
      - apk add -y curl
      - docker info
      - docker-compose --version
      - echo $CI_JOB_TOKEN | docker login -u gitlab-ci-token --password-stdin registry.gitlab.com
    script:
      - cd docker/compose/
      - docker-compose run --use-aliases -d nginx
      - cat /etc/hosts
      - curl --fail http://localhost

I have tried all types of things like http://dockerhost, http://docker, http://nginx, --use-aliases, -p 80:80. The docker-compose run -d nginx and curl --fail http://localhost works fine on my computer, just doesn't work here, also here is some output that might help you help me.

$ cat /etc/hosts
127.0.0.1   localhost
::1 localhost ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
172.17.0.3  docker fe41d99555a4 runner-0277ea0f-project-8964623-concurrent-0-docker-0
172.17.0.4  runner-0277ea0f-project-8964623-concurrent-0
$ curl --fail http://localhost
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed

  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
curl: (7) Failed to connect to localhost port 80: Connection refused

This all works locally, but doesn't appear to work in gitlab-ci. Cross posted from https://github.com/tmaier/docker-compose/issues/14.

Per request from a commenter, here is some additional debug info:

$ docker-compose port nginx 80
0.0.0.0:80

$ docker-compose port django 8000
0.0.0.0:8000

$ docker ps
$ docker ps
CONTAINER ID        IMAGE                                                                                           COMMAND                  CREATED             STATUS                            PORTS                                      NAMES
ea9c0d0f8aba        celery_worker                                           "/start_celery.sh -d…"   4 seconds ago       Up 1 second                                                                  compose_celery-beat_1_3cc1424b6246
6f5adc306f08        celery_beat                                         "/start_celery.sh -d…"   4 seconds ago       Up 1 second                                                                  compose_celery-worker_1_82f9d734734b
4d8d18ee0aa0        nginx                                               "nginx -g 'daemon of…"   4 seconds ago       Up 1 second                       0.0.0.0:80->80/tcp, 0.0.0.0:443->443/tcp   compose_nginx_1_d6a895d7ea5a
0e3f43da2694        django                                              "/start.sh -docker"      4 seconds ago       Up 3 seconds                      0.0.0.0:8000->8000/tcp                     compose_django_1_c7d04b2e58cb
260a04a50134        redis:3.0-alpine                                        "docker-entrypoint.s…"   5 seconds ago       Up 4 seconds                      0.0.0.0:6379->6379/tcp                     compose_redis_1_d6cf5f09d98a
49097614e743        postgres:10-alpine                                      "docker-entrypoint.s…"   5 seconds ago       Up 4 seconds (health: starting)   5432/tcp                                   compose_hoarder-datastore_1_a7db557d6f4c  

Answer 1

Instead of referencing http://localhost you should use http://docker.

Answer 2

I believe you do have a race condition. The http request is being sent too soon, when the container is not ready yet.

See similar case at https://gitlab.com/gitlab-com/support-forum/issues/748#note_12330190

So please add “sleep 60” before the curl