Docker on Mac: how to automatically resolve another service’s IP and add it to the .ENV

A Docker compose command solution to retrieve information from another container service and keep it on the requesting service. The example is resolving an IP address of a MySQL database service and adds it to the apache-php .env.

My first approach was, to every time look up the (changed) IP address of the database Docker service, and manually update the .env using:

docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' SERVICENAME

This cumbersome step worked to reestablish a successful connection to the MySQL database from the PHP app using IP instead of hostname. But, as stated, this required several extra steps, and started to become annoying.

Solution: resolve the database service’s IP, and update the PHP dotEnv – all automatically!

It took quite some research and try‘n‘errors with un-/supported Unix commands in Docker service shells, but ultimately I found a much more elegant – and automatic! – approach to update the `.env` file with the database service‘s IP address!

# Database connection:
MYSQL_HOST="{inserted automatically}"
MYSQL_DATABASE="testdatabase"
MYSQL_USER="mysqluser"
MYSQL_PASSWORD="mysqlpass"

I automated these steps, by adding the following commands to my docker-compose-file
(within the apache2-php service configuration):

services:
    apache-php:
        # other settings...
        volumes:
            #- other mappings...
            - ./.env:/var/www/.env.host
        entrypoint: ["/bin/sh","-c"]
            command:
                - |
                    getent ahosts DATABASESERVER | awk '{ print $1 }' | tail -1 > /tmp/serviceip-db 2>&1
                    sed "s/^MYSQL_HOST=.*/MYSQL_HOST=$(cat /tmp/serviceip-db | tr -d '\n')/g" /var/www/.env.host > /var/www/.env 2>&1

Result: every time the apache-php Docker service is (re-)started, it will dynamically fetch & update the active MySQL-database service’s IP to the PHP .env!

(?) Explanation of the docker-compose commands

First of all, I don‘t connect the app‘s .env from the host 1:1 into my Docker webserver service, but provide it to be internally modified and copied from: /var/www/.env.host
Because my Docker container services are internally on the same network, and resolvable by hostname, I look up the database service‘s IP address and write the result into a temporary file: /tmp/serviceip-db
Now the foundation is set to merge the retrieved database service‘s IP with the required .env for the PHP app!
Using the Unix sed command, the line starting with “MYSQL_HOST…” from the original dotenv .env.host is being replaced with the file contents (read: database IP) from the /tmp/serviceip-db – and the final dotenv is created under `/var/www/.env`
Because my PHP app looks for and reads from the `.env`-file, it has now access to the correct MySQL database host IP to successfully connect! 🥳

…and of course, all other dotEnv variables are also available.

This approach with the 2 commands in docker-compose yaml can be used for various other scenarios, too.

1 thought on “Docker on Mac: how to automatically resolve another service’s IP and add it to the .ENV”

Pingback: Make the switch from MAMP to Docker: how to migrate existing web projects – Swiss Mac User 

Questions? Suggestions? Let us know with a comment!Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.