Question 246825 in Docker Question posted in
The official documentation can be found here.

How to extract image name and tag from docker-compose build output

I have the following script that i am hoping will build and publish to my azure container registry.

echo "env override file:${1}"

cd ../.devcontainer
docker compose --env-file .env --env-file $1 build

#parse image name and tag from above output here into $docker_image and $tag

docker image tag $docker_image:$tag mycontainerreg.azurecr.io/$docker_image:latest 
docker push mycontainerreg.azurecr.io/$docker_image:latest

After the compose comand, the last output shown is :

=> => naming to docker.io/library/my_app:v1.0

I would like to be able to extract the image name and tag from that output to use in the subsequent docker commands. How can i do that?

Also what is best practice for tagging? i tag with a version number. but should i also re-tag as the latest version as im doing – as it would simplify downstream deploys?

for completeness my docker-compose relevant context:

services:
  my_app:
    container_name: my_app_${ENV}
    image: "my_app:${VERSION:-local}"

env files are .env, env_dev, etc:

VERSION=v1.0
ENV=local
USER_ID=1000
GROUP_ID=1000
USER_NAME=user1
SVC_PORT=5000
DB_PORT=5434
Tags:

2

Answers


  1. 0

    Store the output in a variable, let it be output

    output="=> => naming to docker.io/library/my_app:v1.0"

image_and_tag=$(echo "$output" | awk -F '/' '{print $NF}')
image_name=$(echo "$image_and_tag" | awk -F ':' '{print $1}')
image_tag=$(echo "$image_and_tag" | awk -F ':' '{print $2}')

echo "Image Name: $image_name"
echo "Image Tag: $image_tag"

    Regarding tagging:

    • Use specific version tags for production deployments (e.g., v1.0,
      v1.1).
    • Use a ‘latest’ tag for testing, development, and internal use, but
      avoid using it for production.
    • Periodically update the ‘latest’ tag to the latest stable version
      when appropriate, but make sure to test thoroughly before deploying
      to production.
    Login or Signup to reply.
  2. 0

    I’d just copy the image: from the Compose file, and not try to parse Compose’s output. Compose’s environment-variable substitution is a subset of POSIX shell syntax and you should be able to use the exact same string format in your script.

    docker tag my_app:${VERSION:-local} mycontainerreg.azurecr.io/$docker_image:latest

    When you set both build: and image: in the Compose file, the image: sets the name of the built image, so a better approach here might be to just set the Compose image: to what you want. Then you can docker-compose push the built image without having to manually extract its name.

    services:
  my_app:
    build: .
    image: mycontainerreg.azurecr.io/myapp:${VERSION:-latest}
    # should not usually need container_name:

    export VERSION=20230905
docker-compose build
docker-compose push

    I’d suggest that tagging the image with a date stamp, source-control commit ID, or version string is a best practice. Fixed tags like latest can cause some problems, particularly if you’re using Kubernetes, since it becomes unclear which "latest" image you’re using – if the local system already has a "latest" image then the container system usually won’t check if the upstream registry has a newer one.

    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search