熊鑫伟 (Xinwei Xiong) · September 1, 2023

2 min · 318 words · EN | 中

Automate Multi-Architecture Image Build for openkf and Push to Multiple Image Repositories

• https://github.com/openimsdk/openkf

Description:

To meet the requirements of a diverse set of users, we aim to automate the process of building the openkf Docker images for various architectures and push them to multiple image repositories seamlessly.

Objective:

• Automatically build Docker images of openkf for linux/amd64 and linux/arm64 architectures.
• Push the images to Docker Hub, AliYun Docker Hub, and GitHub Container Registry.

Tasks:

1. Setup Multi-Architecture Build System
   • Use GitHub Actions with QEMU and Docker Buildx to support multi-architecture builds for linux/amd64 and linux/arm64.
   • On every new release, commit to the main branch, or scheduled event, trigger the build process.
2. Support Multiple Image Repositories
   • Docker Hub: Push to openim/openkf-server.
   • AliYun Docker Hub: Push to registry.cn-hangzhou.aliyuncs.com/openimsdk/openkf-server.
   • GitHub Container Registry: Push to ghcr.io/openimsdk/openkf-server.
3. Dynamic Image Tagging
   • Use Docker Metadata Action to generate dynamic tags based on events such as scheduled triggers, branch commits, pull requests, semantic versioning, and the commit SHA.
   • Ensure the built image does not get pushed during pull request events.
4. Authentication & Security
   • Configure authentication for Docker Hub, AliYun, and GitHub Container Registry using secrets.
   • Ensure a secure and seamless push operation for each repository.
5. Notifications & Logging
   • Send notifications to the development team in case of any build or push failures through GitHub Actions.
   • Maintain logs for every build and push operation for traceability.

Acceptance Criteria:

• The openkf image should be built successfully for both linux/amd64 and linux/arm64 architectures.
• Upon successful build, the image should be available on Docker Hub, AliYun Docker Hub, and GitHub Container Registry.
• Properly tagged images based on the defined events and attributes.
• No manual intervention required during the entire process.

Additional Notes:

• The automation process is defined in the GitHub Actions workflow. Ensure to review and update the workflow as needed.
• Make sure to test the process in a separate branch or environment to avoid disruptions.

Feedback and further recommendations are welcome.