Here's a step-by-step guide to get your Docker container talking to the PostgreSQL instance on your Windows host:

1. Configure PostgreSQL on Windows to Allow Remote Connections #

By default, PostgreSQL restricts connections to localhost. You'll need to update two config files to open it up to your Docker container:

• Modify postgresql.conf:
  Navigate to your PostgreSQL data directory (typically C:\Program Files\PostgreSQL\<version>\data), open postgresql.conf in a text editor, and change:

  listen_addresses = 'localhost'
  

  to:

  listen_addresses = '*'
  

  This lets PostgreSQL listen for connections on all network interfaces.

• Update pg_hba.conf:
  Still in the data directory, open pg_hba.conf and add a line at the end to allow connections from your Docker network:

  host    all             all             172.17.0.0/16          md5
  

  The 172.17.0.0/16 is the default subnet for Docker's bridge network. If you're using a custom Docker network, replace this with your subnet. For testing purposes, you can use 0.0.0.0/0 to allow connections from any address (not recommended for production).

• Restart PostgreSQL:
  Open the Windows Services manager (services.msc), find your PostgreSQL service (named something like postgresql-x64-<version>), right-click it and select Restart.

Note: Ensure your PostgreSQL user has login privileges and access to the target database. You can set this in psql with:

ALTER USER your_db_user WITH LOGIN;
GRANT ALL PRIVILEGES ON DATABASE your_db_name TO your_db_user;

2. Find the Windows Host IP Accessible from Docker #

Your Docker container needs the correct IP address to reach the Windows host. The easiest method is using Docker's built-in DNS name:

• Use host.docker.internal:
  Docker for Windows provides host.docker.internal as a special DNS name that resolves to your host machine's IP. This works for both WSL2 and Hyper-V setups.

• Fallback to actual host IP:
  If host.docker.internal doesn't work, run ipconfig in a Windows command prompt. Look for the IP address associated with your active network adapter (e.g., "Ethernet adapter Ethernet" for Hyper-V, or "Ethernet adapter vEthernet (WSL)" for WSL2). Use this IP in your Django settings.

3. Update Django Database Settings #

In your Django project's settings.py, modify the DATABASES configuration to point to your Windows-hosted PostgreSQL:

DATABASES = {
    'default': {
        'ENGINE': 'django.db.backends.postgresql_psycopg2',
        'NAME': '<your_database_name>',
        'USER': '<postgresql_username>',
        'PASSWORD': '<postgresql_password>',
        'HOST': 'host.docker.internal',  # Or your actual host IP
        'PORT': '5432',  # Default PostgreSQL port
    }
}

Ensure psycopg2-binary is listed in your awf/requirements.txt (it's required for Django to connect to PostgreSQL in Python 2.7). If not, add it:

psycopg2-binary>=2.8.6

4. Allow PostgreSQL Connections Through Windows Firewall #

Windows Firewall will block incoming connections to port 5432 by default. To open it:

• Open Windows Defender Firewall with Advanced Settings
• Click Inbound Rules > New Rule
• Select Port > Next
• Choose TCP, enter 5432 as the port > Next
• Select Allow the connection > Next
• Check the network types you want (e.g., Private) > Next
• Name the rule (e.g., "PostgreSQL Docker Access") > Finish

5. Test the Connection #

Rebuild your Docker image to apply the settings changes:

docker build -t your-api-image .

Run the container as usual, then check the logs for any connection errors:

docker logs <container_name>

If you see errors like "could not connect to server", double-check:

• PostgreSQL is running and restarted after config changes
• The firewall rule is active
• The host IP in settings.py is correct
• The PostgreSQL user has permissions to connect from the container's IP

内容的提问来源于stack exchange，提问作者revy