Create GKE cluster and Deploy simple hello-app

In our last blog, we see the Kubernetes and it's main components.

https://sumanshunankana.hashnode.dev/kubernetes-and-its-main-components-in-a-layman-term

Now we will see all those components in practical in GKE (Google Kubernetes Engine)

Logon to https://cloud.google.com/ with your account
Navigate to Navigation menu ==> Kubernetes Engine ==> Clusters

GKE option.png

Click Create cluster

Create Cluster button.png

Leave everything as a default (just for simplicity) and click on create. After 1-2 mins cluster will be created. Here our cluster name is my-firstcluster1

cluster created.png

Click on 'connect' button to connect with the cluster, cloud shell will be open and configure kubectl command line access by running the following command:

gcloud container clusters get-credentials my-first-cluster-1 --zone us-central1-c --project qwiklabs-gcp-01-f80d9

So we created the cluster with default settings (which has 3 nodes or virtual machine). Let's see if a backend whether 3 nodes or virtual machine gets created or not.

kubectl get nodes

kubectl is the utility to interact with kubernetes cluster.

get nodes.png

vm instances nodes created.png

So we see automatically 3 vm machines created. Now lets check if any POD created. Ideally there should be NO POD, because we did not do any deployment.

kubectl get pods

get pods empty.png

So as expected, no pods

Lets check if there is any SERVICE

kubectl get service

get service empty.png

we could see default service, but not any other service for our app (because we did not do any deployment yet). So this is as expected.

Now lets do the deployment of simple hello-app, whose image is available in GCR.

kubectl create deployment hello-world --image=gcr.io/google-samples/hello-app:1.0

As we deploy our app without .yaml file, So by default it takes 1 POD

kubectl get all

get all.png

We could see 1 POD.

Now let's scale our application from 1 to 3

kubectl scale deployment hello-world --replicas 3

scale deployment.png

After scaling , we could see 3 PODS.

Now to use the app, Lets expose the app to the outer world.

kubectl expose deployment hello-world --type LoadBalancer --port 80 --target-port 8080

expose deployment and get service.png

After 1-2 minutes, SERVICE will be created which also act as a LoadBalancer and it gives us External IP to use the application. Let's browse that IP

browser output.png

We can see the same LoadBalancer in Google cloud UI Navigation menu ==> Networking ==> Load Balancer

Now, let's try to fetch the same External IP multiple times using curl command. We could see at different request, request reaches to different PODS, as shown in figure.

multiple curl.png

Let's delete one POD out of three and then try to browse same external IP

kubectl delete pod pod_name

We will notice, even after deleting one POD, we are able to access our application and after few seconds, new POD created as well (because we deleted 1 POD)

We can scale the number of PODS from UI interface as well.

Navigation menu ==> Kubernetes Engine ==> Workloads

Click on Actions ==> Scale

scale ui.png

Let's scale to 10 POD

scale ui - 1.png

Now, we could see 10 POD i.e. our application replicated to 10 PODS.

10 pods.png

So we could see -

How to create the GKE Cluster
How to do deloyment
How POD can be scaled using command or UI
How Deployment can be exposed (or Service can be created)
How we can delete the POD

Sumanshu Nankana

Sumanshu Nankana

Create GKE cluster and Deploy simple hello-app