Container image publishing

เนื้อหานี้ยังไม่ได้แปลเป็นภาษาไทย แสดงเป็นภาษาอังกฤษแทน

This guide is the provider-neutral container image publishing baseline for freshly generated Cephalon apps.

Use it when you want a generated host to move from scaffolded source into a reusable build/tag/push image contract through the shipped Dockerfile and deploy/container-image/publish-image.ps1, before you hand that image to Kubernetes or another hosted container platform.

What ships now

Generated app roots from cephalon new and the app-focused dotnet new cephalon-* starters now include:

NuGet.config
.cephalon/packages/README.md
Dockerfile
deploy/container-image/README.md
deploy/container-image/publish-image.ps1

The generated publish script validates the generated app root, builds the shipped Dockerfile with one or more tags, and can optionally push those tags to the registry you choose without introducing provider-specific packaging first.

Prerequisites

a generated app root from cephalon new or dotnet new cephalon-*
Cephalon* packages available either through ./.cephalon/packages or a shared package source that replaces the generated cephalon feed entry
Docker Desktop or another compatible Docker engine on the machine that performs the build
docker login already completed for the target registry before you invoke the generated script with -Push

Preview the build and push contract

Example for a scaffolded modular-monolith app:

pwsh ./Acme.Store/deploy/container-image/publish-image.ps1 `
  -Image ghcr.io/acme/acme-store:1.0.0 `
  -AdditionalTags ghcr.io/acme/acme-store:latest `
  -Push `
  -Preview

The preview step:

validates that the generated app root still contains Dockerfile and NuGet.config
prints the docker build command it would run
prints one docker push command per image tag when -Push is supplied
does not contact the registry or start a real image build

If you need to target a specific runtime platform, also pass -Platform linux/amd64 or another Docker-compatible platform string.

Build the generated image locally

If you only want the local image first:

pwsh ./Acme.Store/deploy/container-image/publish-image.ps1 `
  -Image ghcr.io/acme/acme-store:1.0.0

That keeps the tag local so you can validate the Docker image before you publish it.

Publish the image to a registry

When you are ready to push the image:

docker login ghcr.io
pwsh ./Acme.Store/deploy/container-image/publish-image.ps1 `
  -Image ghcr.io/acme/acme-store:1.0.0 `
  -AdditionalTags ghcr.io/acme/acme-store:latest `
  -Push

The generated script:

builds the image from the generated app root and Dockerfile
applies the primary -Image tag plus any -AdditionalTags
pushes each tag to the registry when -Push is supplied

Once the push completes, reuse the same image tag with the generated Kubernetes apply script or another hosted container deployment surface.

Continue into deployment

After you have a pullable image, continue with:

Those guides keep the image reference explicit instead of rebuilding different images for each deploy target.

Optional repo-native validation script

To replay the generated container-image baseline from the Cephalon repository in one command, run:

pwsh ./scripts/validate-generated-app-container-image.ps1

That script:

scaffolds a fresh app through Cephalon.Cli
seeds ./.cephalon/packages with repo-local package artifacts
validates that the generated app root contains the shipped Docker and container-image assets
previews the generated build/push contract
builds the generated container image through the shipped publish script
by default, starts a local Docker registry and proves that the shipped publish script can push the generated tags into that registry
tears the temporary app and registry back down afterward unless -KeepOutput is supplied