【GCP】Bitbucket Pipeline CICD 至 GKE / Container Registry

1. 创建一个新的service account 并且给与以下的

• Cloud Run
• AdminKubernetes Engine
• AdminKubernetes Engine Cluster Admin
• Service Account User
• Storage Admin

2. 下载service account的json key, 以便pipeline能够使用权限登入到GCP的container registry

3. 使用这bitbucket pipelines yml 文件
[GCP_Project_ID] = GCP Project ID
[GCP_Registry_URL] = asia.gcr.io/GCP Project ID/ImageName

image: atlassian/default-image:2
options:
  docker: true
  size: 2x
pipelines:
  branches:
    test:
      - step:
          name: Build & Push
          deployment: Test
          image: google/cloud-sdk:latest
          script:
            # Build image
            - export IMAGE_NAME=asia.gcr.io/[GCP_Project_ID]:test
            - docker build -t $IMAGE_NAME -f dockerfile .

            # Gcloud auth and check
            - gcloud auth activate-service-account [GCP_ServiceAccount_Email] --key-file=deployment/gcloud.json
            - gcloud config list

            # config image registry with gcloud helper
            - gcloud auth configure-docker -q

            # push image to gcr
            - docker push $IMAGE_NAME

            # delete untagged images
            - gcloud container images list-tags [GCP_Registry_URL] --filter='-tags:*' --format='get(digest)' --limit=unlimited | xargs -I {arg} gcloud container images delete  "[GCP_Registry_URL]@{arg}" --quiet

            # Login to k8s cluster and deploy
            - export GCLOUD_CLUSTER="你的GKE Cluster名"
            - export GCLOUD_ZONE="asia-southeast1-c"
            - export GCLOUD_PROJECT="你的GCP Project ID"
            - gcloud container clusters get-credentials $GCLOUD_CLUSTER --zone=$GCLOUD_ZONE --project $GCLOUD_PROJECT
            - gcloud auth configure-docker --quiet
            - kubectl rollout restart deployment/[你的K8s Deployment Name] -n [你的namespace]
          services:
            - docker
          caches:
            - docker
  
definitions:
  services:
    docker:
      memory: 4096

4. GCP_Project_ID 可以在以下的图片内获取

参考文章： https://puuga.medium.com/dev-story-deploy-to-gcp-cloud-run-with-bitbucket-pipelines-4fef8f2ece27