如何解决Jfrog Pipeline构建基于Artifactory私有镜像的Docker镜像报错

想要基于Artifactory中存储的已有Docker镜像（用作缓存）构建新镜像，参考了Jfrog官方的《Pipeline Example: Docker Build and Push》和《DockerBuild》文档，按照示例配置了资源与流水线：

Build a Docker image with a private base image
This example builds a Docker image that relies on a private base image stored in an 
Artifactory Docker repository.

resources:
# Application source repository
- name: gosvc_app
  type: GitRepo
  configuration:
    gitProvider: myGithub
    path: myuser/myrepo                   # replace with your repository name
    branches:
      include: master

 # Docker image in an Artifactory repository
 - name: base_image
  type: Image
  configuration:
    registry: myArtifactory
    sourceRepository: docker-local        # replace with your repository name
    imageName: docker.artprod.mycompany.com/baseimage
    imageTag: latest
    autoPull: true
pipelines:
  - name: demo_pipeline
  steps:
    - name: bld_image
      type: DockerBuild
      configuration:
        dockerFileLocation: .
        dockerFileName: Dockerfile
        dockerImageName: docker.artprod.mycompany.com/gosvc       
        dockerImageTag: ${run_number}
        inputResources:
          - name: gosvc_app
          - name: base_image
        integrations:
          - name: MyArtifactory

运行时遇到两个问题：

1. 配置两个输入资源时，抛出错误：pipeline demo_pipeline has cyclic connection involving step bld_image. A pipeline can not have cyclic connections.
2. 仅配置Image资源为输入时，抛出错误：DockerBuild step bld_image must have exactly one input resource of type GitRepo

需要解决如何正确实现基于Artifactory中的Docker镜像构建新镜像的问题。

1. 修复YAML语法缩进错误 #

你的流水线配置存在YAML缩进错误，这是导致循环依赖错误的核心原因：

• pipelines下的steps必须缩进在对应的demo_pipeline条目下，当前的steps和name: demo_pipeline处于同一层级，系统解析时会识别为无效结构，进而误报循环依赖。

2. 正确配置DockerBuild步骤的输入资源 #

DockerBuild步骤要求必须有且仅有一个GitRepo类型的输入资源，同时可以添加Image类型资源作为基础镜像/缓存镜像输入，正确配置格式如下：

resources:
# 代码仓库资源（必填）
- name: gosvc_app
  type: GitRepo
  configuration:
    gitProvider: myGithub
    path: myuser/myrepo
    branches:
      include: master

# Artifactory中的基础镜像资源
- name: base_image
  type: Image
  configuration:
    registry: myArtifactory
    sourceRepository: docker-local
    imageName: docker.artprod.mycompany.com/baseimage
    imageTag: latest
    autoPull: true

pipelines:
  - name: demo_pipeline
    # steps必须缩进在demo_pipeline条目下
    steps:
      - name: bld_image
        type: DockerBuild
        configuration:
          dockerFileLocation: .
          dockerFileName: Dockerfile
          dockerImageName: docker.artprod.mycompany.com/gosvc       
          dockerImageTag: ${run_number}
          inputResources:
            - name: gosvc_app  # 必填的GitRepo资源
            - name: base_image # 基础镜像资源
          integrations:
            - name: MyArtifactory

3. 确保Dockerfile正确引用基础镜像 #

在Dockerfile中直接引用Artifactory中的基础镜像地址即可，示例：

FROM docker.artprod.mycompany.com/baseimage:latest
# 后续构建步骤

由于已将base_image资源配置为autoPull: true，流水线会自动通过MyArtifactory集成拉取该镜像到构建环境，无需额外配置拉取凭证。

4. 关键验证要点 #

• 所有YAML层级使用统一缩进（建议用2或4个空格，禁止混用制表符）
• GitRepo资源必须是DockerBuild步骤的第一个输入资源
• Image资源的registry配置需匹配平台中已创建的Artifactory集成名称
• 确认MyArtifactory集成拥有拉取docker-local仓库镜像的权限

内容的提问来源于stack exchange，提问作者Jotta E