Skip to content
This repository has been archived by the owner on Jan 3, 2023. It is now read-only.

Latest commit

 

History

History
273 lines (188 loc) · 8.17 KB

README.md

File metadata and controls

273 lines (188 loc) · 8.17 KB
Raw

Demo: Zone Printer

Zone Printer is a web application that displays the Google Cloud Platform (GCP) compute zone where the application is running.

How does it work?

Zone Printer application is a web server written in Go. It greets the user with the country name, flag and the zone of the server it is deployed in.

It queries the GCE metadata server to find out the zone name it is running on. Since GKE nodes are GCE instances, this application will print the name of the zone it is running on.

The Deployment consists of a container image that is listening on port 80. The Service exposes the Deployment pods on nodePort: 30061 on all the nodes in the Kubernetes cluster.

The example also includes an Ingress manifest (ingress/ingress.yaml) that is used by kubemci to provision the multi-cluster Ingress.

Step 0: Before you begin

  1. GCP Project

    Ensure that you have a GCP Project. If you don't have one already, you can create a project by following the instructions at Creating and Managing Projects.

    After your project is created, run the following commands in your terminal:

    PROJECT=<your-project-name>

  2. Set up gcloud:

    You can install gcloud by following the instructions at Installing Cloud SDK and initialize it by running the following commands:

    gcloud init # Set $PROJECT as the default project
gcloud auth login
gcloud auth application-default login

  3. Set up kubectl:

    You can install kubectl by running:

    gcloud components install kubectl

  4. Install kubemci:

    kubemci is the primary command-line tool used to create and configure a multi-cluster ingress.

    Download one of the binaries and place it in an executable path:

    For example on linux you can install kubemci by running:

    wget https://storage.googleapis.com/kubemci-release/release/latest/bin/linux/amd64/kubemci

    Ensure that the binary is executable. This can be done in the directory of the binary with the following command:

    chmod +x ./kubemci

    In google cloud shell you can move this to your local bin to automatically include it in your path

    mv ./kubemci ~/bin/kubemci

1. Create Kubernetes Clusters

kubemci requires Kubernetes clusters that are v1.8.1 or newer. (You can check your cluster version using the kubectl version command.)

You will need at least two Kubernetes clusters in two different compute zones to verify that kubemci works.

Creating new GKE clusters

The following commands will create two clusters:

  • cluster-1 in us-east4-a
  • cluster-2 in europe-west1-c

and save their connection details to a file named clusters.yaml.

# Create a cluster in us-east and get its credentials
KUBECONFIG=clusters.yaml gcloud container clusters create \
    --cluster-version=1.9 \
    --zone=us-east4-a \
    cluster-1

# Create a cluster in eu-west and get its credentials
KUBECONFIG=clusters.yaml gcloud container clusters create \
    --cluster-version=1.9 \
    --zone=europe-west1-c \
    cluster-2

Using existing GKE clusters

Note: If you don't already have Kubernetes clusters on GCP, skip to the next subsection on creating new clusters.

If you already have existing Kubernetes Engine clusters, run get-credentials with them to create a clusters.yaml file with the connection information of these clusters.

For example:

KUBECONFIG=clusters.yaml gcloud container clusters \
    get-credentials cluster-1 --zone=us-east4-a

KUBECONFIG=clusters.yaml gcloud container clusters \
    get-credentials cluster-2 --zone=europe-west1-c

# ...repeat for other clusters you would like to add to the Ingress

Using non-GKE clusters on GCE

It is possible to use Kubernetes clusters hosted on Google Compute Engine (GCE) with kubemci. You need to create a kubeconfig file that contains the connection information of the clusters you want to add to the multi-cluster Ingress.

First, make sure you have the kubeconfig entries of the clusters:

$ kubectl config get-contexts -o name

name-of-cluster-1
name-of-cluster-2
[...]

For each cluster you will use, export its configuration into a YAML file:

# do this for the cluster-1 in us-east:
kubectl config use-context [name-of-cluster-1]
kubectl config view --minify --flatten > cluster-1.yaml

# do this for the cluster-2 in eu-west:
kubectl config use-context [name-of-cluster-2]
kubectl config view --minify --flatten > cluster-2.yaml

Combine these cluster-1.yaml and cluster-2.yaml config files into one called clusters.yaml:

KUBECONFIG=cluster-1.yaml:cluster-2.yaml kubectl config view \
  --flatten > clusters.yaml

Step 2: Deploy the sample application

Deploy the application along with its NodePort Service in each of the two clusters. You can get the cluster contexts from kubectl config get-contexts and iterate through all the clusters to deploy the application manifests. This could be accomplished by running the following loop:

for ctx in $(kubectl config get-contexts -o=name --kubeconfig clusters.yaml); do
  kubectl --kubeconfig clusters.yaml --context="${ctx}" create -f manifests/
done

Step 3: Reserve a static IP address

This command reserves a static IP address for the load balancer.

ZP_KUBEMCI_IP="zp-kubemci-ip"
gcloud compute addresses create --global "${ZP_KUBEMCI_IP}"

Step 4: Use the static IP on Ingress manifest

Replace the value of kubernetes.io/ingress.global-static-ip-name annotation in ingress/ingress.yaml with the value of $ZP_KUBEMCI_IP.

sed -i -e "s/\$ZP_KUBEMCI_IP/${ZP_KUBEMCI_IP}/" ingress/ingress.yaml

Step 5: Deploy the multi-cluster Ingress with kubemci

Run kubemci to deploy the ingress/ingress.yaml and create the multi-cluster Ingress named zone-printer:

kubemci create zone-printer \
    --ingress=ingress/ingress.yaml \
    --gcp-project=$PROJECT \
    --kubeconfig=clusters.yaml

This command creates the multi-cluster Ingress.

Step 6: View multi-cluster Ingress status

You can view the status of the multi-cluster Ingress you just created using the kubemci get-status command:

$ kubemci get-status zone-printer --gcp-project=$PROJECT

Load balancer zone-printer has IPAddress 35.190.74.228 and
is spread across 2 clusters (gke_yourproject_us-east4-a_cluster-1,
gke_yourproject_europe-west1-c_cluster-2).

Step 7: Test the multi-cluster Ingress

It will take several minutes from when the Ingress and Google Cloud HTTP Load Balancer are created until they are ready. (You may see HTTP 404 or HTTP 502 errors while the load balancer is getting ready.)

Once the load balancer is ready to serve traffic, use the IP address printed in the previous step and query the server:

$ curl http://35.190.74.228

Welcome to the global site! You are being served from us-central1-b.
Congratulations

Multi-cluster Ingress has routed your traffic to a cluster in the nearest GCP region.

Step 8: Cleanup

To delete the multi-cluster Ingress you created, run:

kubemci delete zone-printer \
    --ingress=ingress/ingress.yaml \
    --gcp-project=$PROJECT \
    --kubeconfig=clusters.yaml

This will also clean up the underlying networking resources provisioned by kubemci.

If you created new Kubernetes Engine clusters for this tutorial, clean them up:

gcloud container clusters delete cluster-1 --zone=us-east4-a -q

gcloud container clusters delete cluster-2 --zone=europe-west1-c -q