Before getting started, you need to click on the 'fork' button in the upper right corner this repository.
After the fork is complete, click the green "clone or download" link, and copy the clone url.
Next, clone your fork of the repository:
USERNAME=<GITHUB_USERNAME>
git clone https://github.com/${USERNAME}/gcb-gke-codelab.git
cd gcb-gke-codelab
Create a new GCP project and capture the projectId in the PROJECT_ID env var:
PROJECT_NAME=gke-pipeline-${USERNAME}
PROJECT_ID=$(gcloud projects create --name "${PROJECT_NAME}" --format='value(projectId)')
At the prompt type 'y' to accept generated project id.
Set the default project:
gcloud config set project ${PROJECT_ID}
Set the default zone:
COMPUTE_ZONE=us-west1-c
gcloud config set compute/zone ${COMPUTE_ZONE}
Ensure the default credentials are available on your local machine:
gcloud auth application-default login
Enable Billing on your project:
On Mac:
open https://console.developers.google.com/project/${PROJECT_ID}/settings
Use xdg-open with Linux, start with Windows
Enable the required GCP APIs:
gcloud services enable --async \
container.googleapis.com \
cloudapis.googleapis.com \
cloudbuild.googleapis.com \
sourcerepo.googleapis.com \
compute.googleapis.com \
storage-component.googleapis.com \
containerregistry.googleapis.com \
logging.googleapis.com
Use gcloud services list --enabled to check progress
Create the cluster:
CLUSTER_NAME=production
gcloud container clusters create ${CLUSTER_NAME} --async
It can take up to five minutes to provision the Kubernetes clusters. Use the gcloud command to check the status of each cluster:
gcloud container clusters list
Grant the Container Builder service account developer access to the GKE API to fetch credentials and apply changes to the cluster.
PROJECT_NUMBER=$(gcloud projects describe ${PROJECT_ID} --format='value(projectNumber)')
gcloud projects add-iam-policy-binding ${PROJECT_NUMBER} \
--member=serviceAccount:${PROJECT_NUMBER}@cloudbuild.gserviceaccount.com \
--role=roles/container.developer
The project is setup at this point. Now what???
APP_NAME=my-app
Wait for cluster to be running to execute this section
Get cluster credentials and make available to kubectl
gcloud container clusters get-credentials ${CLUSTER_NAME} --zone ${COMPUTE_ZONE}
mkdir k8s
Create a Deployment file
kubectl run ${APP_NAME} --image=gcr.io/${PROJECT_NAME}/${APP_NAME} --port=3000 --env="GET_HOSTS_FROM=dns" --labels="app=${APP_NAME}" --dry-run -o yaml > k8s/deployment.yaml
Create a Service file
kubectl create service loadbalancer ${APP_NAME} --tcp=80:3000 --dry-run -o yaml > k8s/service.yaml
Create Service and Deployment resources
kubectl apply -f k8s/
touch cloudbuild.yaml
Open cloudbuild.yaml in your favorite editor
Copy/paste the following into cloudbuild.yaml:
Add docker build step:
steps:
# Build the image
- id: 'Build docker image'
name: 'gcr.io/cloud-builders/docker'
args: [ 'build', '-t', 'gcr.io/$PROJECT_ID/$_APP_NAME:$SHORT_SHA', '.' ]
Push image to registry:
# Push updated image
- id: 'Push image to registry'
name: 'gcr.io/cloud-builders/docker'
args: [ 'push', 'gcr.io/$PROJECT_ID/$_APP_NAME:$SHORT_SHA' ]
Patch Deployment manifest file:
# Patch the Deployment
- id: 'Patch manifest with new image'
name: 'gcr.io/cloud-builders/kubectl'
entrypoint: 'sh'
env:
- 'CLOUDSDK_COMPUTE_ZONE=${_CLOUDSDK_COMPUTE_ZONE}'
- 'CLOUDSDK_CONTAINER_CLUSTER=${_CLOUDSDK_CONTAINER_CLUSTER}'
args:
- '-c'
- |
gcloud container clusters get-credentials \
--project="${PROJECT_ID}" --zone="${_CLOUDSDK_COMPUTE_ZONE}" "${_CLOUDSDK_CONTAINER_CLUSTER}"
cat <<EOF > patch.yaml
spec:
template:
spec:
containers:
- name: ${_APP_NAME}
image: gcr.io/${PROJECT_ID}/${_APP_NAME}:${SHORT_SHA}
EOF
kubectl patch --local -o yaml \
-f k8s/deployment.yaml \
-p "$(cat patch.yaml)" \
> deployment.yaml
mv deployment.yaml k8s/deployment.yaml
Apply change to the GKE cluster:
# Apply change
- id: 'Apply update to cluster'
name: 'gcr.io/cloud-builders/kubectl'
args: [ 'apply', '-f', 'k8s/']
env:
- 'CLOUDSDK_COMPUTE_ZONE=${_CLOUDSDK_COMPUTE_ZONE}'
- 'CLOUDSDK_CONTAINER_CLUSTER=${_CLOUDSDK_CONTAINER_CLUSTER}'
Tell the build that you created an image:
# Associate image that was pushed to GCR with the build history UI
images:
- 'gcr.io/$PROJECT_ID/$_APP_NAME:$SHORT_SHA'
Save the file
Manually trigger a build from the CLI
gcloud container builds submit . \
--config cloudbuild.yaml \
--substitutions \
_APP_NAME=${APP_NAME},_CLOUDSDK_COMPUTE_ZONE=${COMPUTE_ZONE},_CLOUDSDK_CONTAINER_CLUSTER=${CLUSTER_NAME},SHORT_SHA=xxx
List Service
kubectl get service ${APP_NAME}
Note the external IP address once it's provisioned. Save it for later.
URL=http://<service-external-ip>
See if it worked:
open ${URL}
We're going to create a build trigger that will kick off the build and deploy pipeline on changes to source code.
echo ${APP_NAME} ${COMPUTE_ZONE} ${CLUSTER_NAME}
open https://console.cloud.google.com/gcr/triggers?project=${PROJECT_ID}
Follow prompts to OAuth into GitHub and select your repo.
Create trigger with following:
| Field | Value |
|---|---|
| Name | Deploy on push to master |
| Trigger type | branch |
| Branch | master |
| Build configuration | cloudbuild.yaml |
| cloudbuild.yaml location | cloudbuild.yaml |
| Substitution variables | _CLOUDSDK_CONTAINER_CLUSTER: <CLUSTER_NAME> |
| _CLOUDSDK_COMPUTE_ZONE: <COMPUTE_ZONE> | |
| _APP_NAME: <_APP_NAME> |
Save Trigger
Update the hello message at app/lib/get-welcome-message.js
git add .
git commit -m 'your message'
git push
A new build should kick off. When complete you should see your new welcome message.
Check build status
open https://console.cloud.google.com/gcr/builds?project=${PROJECT_ID}
Check that your change is live
open ${URL}
This was intended as a quickstart codelab for familiarizing yourself with GKE, GCB, and setting up a CICD pipeline.
For a more real-world pipeline, see this pipeline tutorial or evolve this project and implement some of the following:
- Browse the supported and community contributed builder images to get a sense of available functionality
- Speed up your build by pulling in your previously built image and doing a --cache-from build
- Commit updates to manifest files back to repo to keep as source of truth. Use hub CLI tool to commit changes back to repo from build.
- Create multiple cluster like qa, staging, and prod, and create additional build triggers connecting them, e.g. deploy to qa on push to a feature branch, push to prod on git tag.
- Separate infrastructure manifest files from application code
- Use PRs as manual gates for promoting changes to prod without rebuilding
- See Kelsey Hightower's production-ready pipeline tutorial for an example of these and other best practices
Delete the GCP Project:
gcloud projects delete ${PROJECT_ID}
Delete the GitHub repo
REPO=gcb-gke-codelab
curl -X DELETE "https://api.github.com/repos/${USERNAME}/${repo}"