First we have to set some env variables. like these:

 $ export PROJECT_ID=blissful-canyon-363***
 $ export REGION=us-west1
 

After that we setup gcloud's project ID and create a docker repo.

 $ gcloud config set project $PROJECT_ID
 $ gcloud artifacts repositories create hello-repo \
    --repository-format=docker \
    --location=$REGION \
    --description="Docker repository"
 

Before we push container to that repo, we want to build an docker image. Run this command:

 $ git clone https://github.com/GoogleCloudPlatform/kubernetes-engine-samples
 $ cd kubernetes-engine-samples/hello-app
 $ docker build -t ${REGION}-docker.pkg.dev/${PROJECT_ID}/hello-repo/hello-app:v1 .
 

For now we have got an image whose tag is ${REGION}-docker.pkg.dev/${PROJECT_ID}/hello-repo/hello-app:v1.

Get auth from the repo (like `dockder login`):

 $ gcloud auth configure-docker ${REGION}-docker.pkg.dev
 

Push image to that repo:

 $ docker push ${REGION}-docker.pkg.dev/${PROJECT_ID}/hello-repo/hello-app:v1
 

Now images in repo are prepared. We have to create a k8s cluster for running the image.

Setup k8s cluster to Autopilot and create one:

 $ gcloud config set compute/region ${REGION}
 $ gcloud container clusters create-auto hello-cluster
 

Running the following command will get the output:

 $ kubectl get nodes
 NAME                                           STATUS   ROLES    AGE   VERSION
 gk3-hello-cluster-default-pool-c4c5959a-4pbd   Ready    <none>   24m   v1.22.12-gke.2300
 gk3-hello-cluster-default-pool-d3b8c13d-9q98   Ready    <none>   19m   v1.22.12-gke.2300
 gk3-hello-cluster-default-pool-d3b8c13d-n7df   Ready    <none>   24m   v1.22.12-gke.2300
 

Since the cluster has been created we have to create a Deployment on it.

Run this command to get credentials from the created cluster:

 $ gcloud container clusters get-credentials hello-cluster 
 

Create a deployment to run the image in docker repo:

 $ kubectl create deployment hello-app --image=${REGION}-docker.pkg.dev/${PROJECT_ID}/hello-repo/hello-app:v1
 

Create the base replicas to 3:

 $ kubectl scale deployment hello-app --replicas=3
 

Setup auto-scaling policy from min 1 to max 5:

 $ kubectl autoscale deployment hello-app --cpu-percent=80 --min=1 --max=5
 

Finally we check the pods that were already deployed.

 $ kubectl get pods
 NAME                         READY   STATUS    RESTARTS   AGE
 hello-app-7d989d7dcc-l5jsh   1/1     Running   0          25m
 

But how can we access the docker service? Here we need to expose the cluster as a service.

Run this command to expose the service:

 $ kubectl expose deployment hello-app --name=hello-app-service --type=LoadBalancer --port 80 --target-port 8080
 

It maps the port from internal 8080 to external 80 and generates a public IP via LoadBalancer server.

Run this command to get the service IP:

 $ kubectl get service
 NAME                TYPE           CLUSTER-IP     EXTERNAL-IP     PORT(S)        AGE
 hello-app-service   LoadBalancer   10.104.2.106   34.82.180.103   80:32221/TCP   6m7s
 

Copy the external IP to browser as input and you will see the output of that docker service.

Until now all done well.

Return to home | Generated on 10/08/22