argo的输入输出--output和input输出目录或文件到下一步骤（3）

look_w

完整示例代码如下:

apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
  generateName: output-artifacts-
spec:
  entrypoint: output-artifacts
  templates:
  - name: output-artifacts
    steps:
    - - name: generate-artifacts
        template: generate
    - - name: consume-artifacts
        template: consume
        arguments:
          artifacts:
          - name: in-artifact
            from: "{{steps.generate.outputs.artifacts.out-artifact}}"

  - name: generate
    container:
      image: docker/whalesay:latest
      command: [sh, -c]
      args: ["echo -n hello world > /tmp/hello_world.txt"]
    outputs:
      artifacts:
      - name: out-artifact
        path: /tmp/hello_world.txt
        s3:
          endpoint: s3.amazonaws.com
          bucket: my-aws-bucket-name
          key: path/in/bucket/my-input-artifact.txt

  - name: consume
    inputs:
      artifacts:
       - name: in-artifact
         path: /tmp/input.txt
    container:
      image: docker/whalesay:latest
      command: [sh, -c]
      args: ["        
        echo 'input artifact contents:' &&
        cat /tmp/input.txt
      "]



在配置文件中统一配置s3存储引擎

如果在每个使用的地方都去加s3的配置，那代码会很冗余，argo有一个统一的配置可以进行设置。
使用命令编辑配置文件如下:

kubectl edit configmap workflow-controller-configmap -n argo

    1

增加内容如下:

data:
  config: |
    artifactRepository:
      s3:
        bucket: my-aws-bucket-name
        keyPrefix: prefix/in/bucket     #optional可选
        endpoint: s3.amazonaws.com       #AWS => s3.amazonaws.com; GCS => storage.googleapis.com
        insecure: true                  #omit for S3/GCS. Needed when minio runs without TLS
        accessKeySecret:
          name: my-aws-s3-credentials
          key: accessKey
        secretKeySecret:
          name: my-aws-s3-credentials
          key: secretKey