Building and Running Multi-Architecture Containers
Objective
This lab shows you how to create a multi-architecture container image for the FastAPI application using docker buildx, which is Docker CLI plugin for extended build capabilities with BuildKit. The PostgreSQL database, utilizing the official "postgres:13" image, is sourced directly from Docker Hub, eliminating the necessity to construct and deploy a custom image for it.
Prerequisites
Initial Setup
Navigate to the root directory of the python-fastapi-demo-docker project where your environment variables are sourced:
cd ~/environment/python-fastapi-demo-docker
1. Logging into Amazon ECR
From the 'python-fastapi-demo-docker' project directory, authenticate the Docker CLI to your Amazon ECR registry using:
aws ecr get-login-password \
--region ${AWS_REGION} | docker login \
--username AWS \
--password-stdin ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com
You should see the following response output: “Login Succeeded”.
Alternatively, if you're using Finch, run the following command to authenticate to your Amazon ECR registry:
aws ecr get-login-password \
--region ${AWS_REGION} | finch login \
--username AWS \
--password-stdin ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com
2. Building Multi-Architecture Docker Image for the Web Service
You can use docker buildx to build Docker images for application and database services that are compatible with multiple architectures, including Kubernetes clusters.
First, create and start new builder instances for the web service:
docker buildx create --name webBuilder
docker buildx use webBuilder
docker buildx inspect --bootstrap
The expected output should look like this:
webBuilder
[+] Building 2.7s (1/1) FINISHED
=> [internal] booting buildkit 2.6s
=> => pulling image moby/buildkit:buildx-stable-1 1.1s
=> => creating container buildx_buildkit_webbuilder0 1.5s
Name: webBuilder
Driver: docker-container
Nodes:
Name: webbuilder0
Endpoint: unix:///var/run/docker.sock
Status: running
Platforms: linux/amd64, linux/amd64/v2, linux/amd64/v3, linux/arm64, linux/riscv64, linux/ppc64le, linux/s390x, linux/386, linux/mips64le, linux/mips64, linux/arm/v7, linux/arm/v6
Alternatively, if you're using Finch, run the following command to build the multi-architecture image:
finch build \
--platform linux/amd64,linux/arm64 \
--tag ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION} \
.
3. Pushing the Image to Amazon ECR
Next, build and push the images for your web service by running the following commands:
docker buildx use webBuilder
docker buildx build --platform linux/amd64,linux/arm64 -t ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION} . --push
Alternatively, if you're using Finch, upload the images to ECR using the following command:
finch push \
--all-platforms \
${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION}
This builds Docker images based on the Dockerfile instructions and pushes them to your ECR repository. The expected output should look like this :
[+] Building 305.4s (30/30) FINISHED
=> [internal] load build definition from Dockerfile 0.1s
=> => transferring dockerfile: 1.01kB 0.0s
=> [linux/amd64 internal] load metadata for docker.io/library/python:3.9-slim-buster 2.3s
=> [linux/arm64 internal] load metadata for docker.io/library/python:3.9-slim-buster 2.2s
=> [auth] library/python:pull token for registry-1.docker.io 0.0s
=> [internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [linux/amd64 builder 1/5] FROM docker.io/library/python:3.9-slim-buster@sha256:320a7a4250aba4249f458872adecf92eea88dc 1.3s
=> => resolve docker.io/library/python:3.9-slim-buster@sha256:320a7a4250aba4249f458872adecf92eea88dc6abd2d76dc5c0f01cac9 0.2s
=> => sha256:2e1c130fa3ec1777a82123374b4c500623959f903c1dd731ee4a83e1f1b38ff2 3.14MB / 3.14MB 1.1s
=> => sha256:84c8c79126f669beec1dcf6f34cd88094471745570c19c29b465dfa7db1fdabd 243B / 243B 0.2s
=> => sha256:8d53da26040835f622504d7762fad14d226ac414efeb5363f5febebc89ff224d 11.04MB / 11.04MB 3.2s
=> => sha256:8b91b88d557765cd8c6802668755a3f6dc4337b6ce15a17e4857139e5fc964f3 27.14MB / 27.14MB 15.1s
=> => sha256:824416e234237961c9c5d4f41dfe5b295a3c35a671ee52889bfb08d8e257ec4c 2.78MB / 2.78MB 1.3s
=> => extracting sha256:8b91b88d557765cd8c6802668755a3f6dc4337b6ce15a17e4857139e5fc964f3 4.3s
=> => extracting sha256:824416e234237961c9c5d4f41dfe5b295a3c35a671ee52889bfb08d8e257ec4c 1.4s 0.0s
=> => extracting sha256:2e1c130fa3ec1777a82123374b4c500623959f903c1dd731ee4a83e1f1b38ff2 1.3s
=> [linux/arm64 builder 1/5] FROM docker.io/library/python:3.9-slim-buster@sha256:320a7a4250aba4249f458872adecf92eea88dc 0.7s
=> => resolve docker.io/library/python:3.9-slim-buster@sha256:320a7a4250aba4249f458872adecf92eea88dc6abd2d76dc5c0f01cac9 0.2s
=> [internal] load build context 0.3s
=> => transferring context: 130.52kB 0.1s
=> [linux/arm64 builder 2/5] WORKDIR /server 0.3s
=> [linux/arm64 builder 3/5] RUN apt-get update && apt-get install -y build-essential 120.4s
=> [linux/arm64 runner 3/8] RUN apt-get update && apt-get install -y netcat 71.4s
=> [linux/amd64 builder 2/5] WORKDIR /server 0.1s
=> [linux/amd64 runner 3/8] RUN apt-get update && apt-get install -y netcat 16.6s
=> [linux/amd64 builder 3/5] RUN apt-get update && apt-get install -y build-essential 50.2s
=> [linux/amd64 builder 4/5] COPY ./server/requirements.txt /server/ 0.2s
=> [linux/amd64 builder 5/5] RUN pip wheel --no-cache-dir --no-deps --wheel-dir /server/wheels -r requirements.txt 13.3s
=> [linux/amd64 runner 4/8] COPY --from=builder /server/wheels /server/wheels 0.1s
=> [linux/amd64 runner 5/8] COPY --from=builder /server/requirements.txt . 0.0s
=> [linux/amd64 runner 6/8] RUN pip install --no-cache /server/wheels/* 15.3s
=> [linux/amd64 runner 7/8] RUN pip install uvicorn 2.3s
=> [linux/amd64 runner 8/8] COPY . /server/ 0.1s
=> [linux/arm64 builder 4/5] COPY ./server/requirements.txt /server/ 0.1s
=> [linux/arm64 builder 5/5] RUN pip wheel --no-cache-dir --no-deps --wheel-dir /server/wheels -r requirements.txt 43.3s
=> [linux/arm64 runner 4/8] COPY --from=builder /server/wheels /server/wheels 0.1s
=> [linux/arm64 runner 5/8] COPY --from=builder /server/requirements.txt . 0.0s
=> [linux/arm64 runner 6/8] RUN pip install --no-cache /server/wheels/* 83.4s
=> [linux/arm64 runner 7/8] RUN pip install uvicorn 11.4s
=> [linux/arm64 runner 8/8] COPY . /server/ 0.1s
=> exporting to image 18.2s
=> => exporting layers 5.8s
=> => exporting manifest sha256:b0470562af55ca88d950a848ff258f7baed1231d11e73a93e79d9dec19c77382 0.0s
=> => exporting config sha256:f0e3c543a641a36f19e8fea24195e1aee441b00502798c607397ae93405604a3 0.0s
=> => exporting manifest sha256:0faad539de2df68a509bb9e08fbe457681cd161a0dbba98d4519b4b5c1e3cb39 0.0s
=> => exporting config sha256:d298c0b1ce0f96e0b1a4b1e05cd99706ef767045f17f118e48906273443cb88b 0.0s
=> => exporting manifest list sha256:894e90606e81e42117077fb6797a1332131dd077fa91c75ba4b498619d90e9e7 0.0s
=> => pushing layers 11.1s
=> => pushing manifest for ${AWS_ACCOUNT_ID}.dkr.ecr.us-east-1.amazonaws.com/fastapi-microservices:1.0@sha256:894e90606e81e42 1.2s
=> [auth] sharing credentials for ${AWS_ACCOUNT_ID}.dkr.ecr.us-east-1.amazonaws.com
4. Running the Services as Docker Containers
After building the images, start the application and database services in separate Docker containers using Docker Compose:
docker-compose up
Alternatively, if you're using Finch, run the following command:
finch compose up
This command initiates containers for each service as specified in the docker-compose.yml file. You can connect to the service and access FastAPI application from your browser similarly to Step #2 in section Building and Running the Docker Containers.
5. Stopping the Services and Their Containers
Stop and remove the containers of both services by pressing CTRL + C or running the following command:
docker-compose down --volumes
Alternatively, if you're using Finch, run the following command:
finch compose down
6. Rebuilding and Restarting Docker Services
If you make changes to your application, you can rebuild the multi-architecture images using the following commands:
Define a new image version
IMAGE_VERSION=1.2
Rebuild the image and push it to your ECR repository:
docker buildx use webBuilder
docker buildx build --platform linux/amd64,linux/arm64 -t ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION} . --push
Alternatively, if you're using Finch, you can run:
finch build \
--platform linux/amd64,linux/arm64 \
--tag ${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION} \
.
finch push \
--all-platforms \
${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/fastapi-microservices:${IMAGE_VERSION}
This command rebuilds the Docker image and pushes it to your ECR repository.
Cleanup
Run the following command to clean up created images:
docker rmi -f $(docker images "${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com/*" -q)
Alternatively, if you're using Finch, run the following command to clean-up created images, including dangling images:
finch rmi -f $(finch images --filter reference=${AWS_ACCOUNT_ID}.dkr.ecr.${AWS_REGION}.amazonaws.com -q)
finch rmi $(finch images -f "dangling=true" -q)
Conclusion
This lab explored the process of constructing and executing Docker containers using Docker Compose in the python-fastapi-demo-docker project. We also demonstrated how to use Docker's buildx feature to create Docker images that are compatible with multiple CPU architectures. This approach provides an efficient way to manage multi-service applications, enhancing their portability and ensuring they can run on a wider range of platforms.