Ubuntu 上的 Docker 网络主机

Question

我有一个 Django REST 服务和另一个 Flask 服务作为应用程序的代理。 两者都是使用自己的 Docker 容器运行的不同项目。 I'm able to POST a product on the Django service that is consumed by the Flask service, however, I cannot reach the Django service via Flask. 这些容器在同一个网络上运行，我已经尝试了 Thomasleveil 的建议，包括 qoomon 的docker-host 。 请求收到的错误与我尝试转发流量之前的错误相同。 不同之处在于，现在当我执行请求时，它会一直挂起一段时间，直到失败。

错误如下： requests.exceptions.ConnectionError: HTTPConnectionPool(host='172.17.0.1', port=8000): Max retries exceeded with url: /api/user (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f0039388340>: Failed to establish a new connection: [Errno 110] Connection timed out'))

我试图发出的请求是 /api/products/1/like 的 POST。 目前，不需要任何实体。

这是我使用 Flask 进行 POST 的方式，其中 IP 是 Docker ZA12A3079E14CED421AZBA9BBA86

@app.route("/api/products/<int:id>/like", methods=["POST"])
def like(id):
    req = requests.get("http://172.17.0.1:8000/api/user")
    json = req.json()

    try:
        product_user = ProductUser(user_id=json["id"], product=id)
        db.session.add(product_user)
        db.session.commit()

        publish("product_liked", id)
    except:
        abort(400, "You already liked this product")

    return jsonify({
        "message": "success"
    })

Django 的 docker 撰写文件（请忽略服务 tcp_message_emitter）：

version: '3.8'
services:
  backend:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python manage.py runserver 0.0.0.0:8000"
    ports:
      - 8000:8000
    volumes:
      - .:/app
    depends_on:
      - db

  queue:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python consumer.py"
    depends_on:
      - db

  db:
    image: mysql:5.7.22
    restart: always
    environment:
      MYSQL_DATABASE: admin
      MYSQL_USER: root
      MYSQL_PASSWORD: root
      MYSQL_ROOT_PASSWORD: root
    volumes:
      - .dbdata:/var/lib/mysql
    ports:
      - 33066:3306

  dockerhost:
    image: qoomon/docker-host
    cap_add:
      - NET_ADMIN
      - NET_RAW
    restart: on-failure
    networks:
      - backend

  tcp_message_emitter:
    image: alpine
    depends_on:
      - dockerhost
    command: [ "sh", "-c", "while :; do date; sleep 1; done | nc 'dockerhost' 2323 -v"]
    networks:
      - backend

networks:
  backend:
    driver: bridge

Flask 的 docker 撰写文件：

version: '3.8'
services:
  backend:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python main.py"
    ports:
      - 8001:5000
    volumes:
      - .:/app
    depends_on:
      - db

  queue:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python consumer.py"
    depends_on:
      - db

  db:
    image: mysql:5.7.22
    restart: always
    environment:
      MYSQL_DATABASE: main
      MYSQL_USER: root
      MYSQL_PASSWORD: root
      MYSQL_ROOT_PASSWORD: root
    volumes:
      - .dbdata:/var/lib/mysql
    ports:
      - 33067:3306

在这一点上，我知道我遗漏了一些细节或者我的配置有误。

您可以在这里查看存储库： https://github.com/lfbatista/ms-ecommerce

任何帮助，将不胜感激。

Answer 1

这些容器实际上不在同一个网络上。 要将来自不同docker-compose项目的两个容器放入一个网络，您需要在其中一个文件中“导入”现有网络。 以下是您的操作方法：

# first project
networks:
  internal:
  shared:
---
# second project
networks:
  internal:
  shared:
    # This is where all the magic happens:
    external: true  # Means do not create a network, import existing.
    name: admin_shared  # Name of the existing network. It's usually made of <folder_name>_<network_name> .

不要忘记将所有服务放在同一个internal网络中，否则它们将无法相互通信。 如果您忘记这样做，Docker 将创建一个<folder_name>-default网络并将任何没有明确分配网络的容器放在那里。 您可以像这样分配网络：

services:
  backend:
    ...
    networks:
      internal:
      # Since this service needs access to the service in another project
      # you put here two networks.
      shared:
        # This part is relevant for this specific question because
        # both projects has services with identical names. To avoid
        # mess with DNS names you can add an additional name to the
        # service using 'alias'. This particular service will be
        # available in shared network as 'flask-backend'.
        aliases:
        - flask-backend

  db:
    ...
    # You can also assign networks as an array if you need no extra configuration:
    networks:
      - internal

这是您存储库中的文件。 而不是 IP 地址，一项服务可以分别通过flask-backend或django-backend到达另一项。 请注意，我删除了那些奇怪的“主机网络容器”。

管理员/docker-compose.yml ：

version: '3.8'
services:
  backend:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python manage.py runserver 0.0.0.0:8000"
    ports:
      - 8000:8000
    volumes:
      - .:/app
    depends_on:
      - db
    networks:
      internal:
      shared:
        aliases:
        - django-backend

  queue:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python consumer.py"
    depends_on:
      - db
    networks:
      - internal

  db:
    image: mysql:5.7.22
    restart: always
    environment:
      MYSQL_DATABASE: admin
      MYSQL_USER: root
      MYSQL_PASSWORD: root
      MYSQL_ROOT_PASSWORD: root
    volumes:
      - .dbdata:/var/lib/mysql
    ports:
      - 33066:3306
    networks:
      - internal

networks:
  internal:
  shared:

主/docker-compose.yml ：

version: '3.8'
services:
  backend:
    build:
      context: .
      dockerfile: Dockerfile
    command: "python main.py"
    networks:
      internal:
      shared:
        aliases:
        - flask-backend
    ports:
      - 8001:5000
    volumes:
      - .:/app
    depends_on:
      - db

  queue:
    networks:
      - internal
    build:
      context: .
      dockerfile: Dockerfile
    command: "python consumer.py"
    depends_on:
      - db

  db:
    image: mysql:5.7.22
    restart: always
    networks:
      - internal
    environment:
      MYSQL_DATABASE: main
      MYSQL_USER: root
      MYSQL_PASSWORD: root
      MYSQL_ROOT_PASSWORD: root
    volumes:
      - .dbdata:/var/lib/mysql
    ports:
      - 33067:3306

networks:
  internal:
  shared:
    external: true
    name: admin_shared