Deployment with Kubernetes

This tutorial explains how to deploy our demo site Luna (see quick start guide) to the cloud using Kubernetes. In this tutorial we use Google Kubernetes Engine (GKE) but the setup is not limited to Google Cloud Platform (GCP) and can be used with any Kubernetes cluster like Amazon EKS, Azure Kubernetes Service, etc.

WebSight Helm chart bootstraps the CMS that supports two types of storage backends: Document Storage (MongoDB) and Segment Storage (TAR). In this tutorial, we use the MongoDB instance.

Before you start

Make sure you have:

• your GCP environment set
• Google Kubernetes Engine API enabled
• kubectl installed (if you use Docker Desktop, kubectl should already be installed)
• helm installed

Step 1: Kubernetes cluster configuration

1. Create a Kubernetes cluster consisting of a single nodes pool with the following configuration:

   • number of nodes: 3
   • zone: any zone of your choice, e.g. europe-west1-b to place the cluster in Belgium
   • machine type: e2-standard-2 (2 vCPUs, 8 GB memory)

     gcloud container clusters create my-websight-cms \
       --num-nodes 3 --zone europe-west1-b --machine-type e2-standard-2
     

   Running the command above may take a couple of minutes to finish. It will also configure kubectl to use the cluster.

Step 2: Install prerequisites

1. Install an NGINX Ingress Controller

   helm upgrade --install ingress-nginx ingress-nginx \
     --repo https://kubernetes.github.io/ingress-nginx \
     --namespace ingress-nginx --create-namespace
   

2. Create cms namespace:

   kubectl create namespace cms
   

3. Install MongoDB in the cms namespace using Helm:

   helm install mongodb oci://registry-1.docker.io/bitnamicharts/mongodb --version 14.3.0 \
     --set auth.enabled=true \
     --set auth.rootUser="mongoadmin" \
     --set auth.rootPassword="mongoadmin" \
     --set architecture="replicaset" \
     -n cms
   

   At the end of the installation, you should see the following message with MongoDB connection details:

   MongoDB can be accessed on the following DNS name(s) and ports from within your cluster:
   
   mongodb-0.mongodb-headless.cms.svc.cluster.local:27017
   mongodb-1.mongodb-headless.cms.svc.cluster.local:27017
   

   You will use the connection details in the next section.

   MongoDB

   For simplicity in this step we use Bitnami MongoDB Helm chart with default configuration for replicaset architecture (2 data-bearing members and 1 arbiter).

Step 3: Deploy CMS

1. Create my-websight-cms directory and download the following files into it:
   • Example values.yaml with the configuration for Luna demo site

     curl -O https://raw.githubusercontent.com/websight-io/charts/main/websight-cms/examples/luna-proxy/values.yaml
     

   • Nginx configuration for the CMS Proxy

     curl -O https://raw.githubusercontent.com/websight-io/charts/main/websight-cms/examples/luna-proxy/luna-site.conf.template
     

2. Find the external IP address of the Ingress Controller in your cluster (YOUR_CLUSTER_IP):

   kubectl get svc ingress-nginx-controller -n ingress-nginx -o jsonpath='{.status.loadBalancer.ingress[0].ip}'
   

3. Update the proxy configuration in values.yaml with the IP address from the previous step:

   proxy:
     enabled: true
     env:
       - name: NGINX_HOST
         value: <YOUR_CLUSTER_IP>.nip.io
     sites:
       - name: luna
         host: luna.<YOUR_CLUSTER_IP>.nip.io
         configMapKeyRef:
           name: luna-site-config
           key: luna-site.conf.template
   

4. Deploy the Nginx proxy configuration as a ConfigMap:

   kubectl create configmap luna-site-config --from-file=luna-site.conf.template -n cms
   

5. Deploy the CMS using Helm and the configuration from values.yaml (replace <YOUR_CLUSTER_IP> with the IP address from the 2nd step):

   helm upgrade --install websight-cms websight-cms \
       --repo https://websight-io.github.io/charts \
       --set cms.persistence.mode=mongo \
       --set cms.persistence.mongo.hosts='mongodb-0.mongodb-headless.cms.svc.cluster.local:27017\,mongodb-1.mongodb-headless.cms.svc.cluster.local:27017' \
       --set cms.livenessProbe.initialDelaySeconds=120 \
       --set cms.ingress.enabled=true \
       --set cms.ingress.host=cms.<YOUR_CLUSTER_IP>.nip.io \
       --namespace cms \
       -f values.yaml --wait
   

   It may take about 1-2 minute of minutes to finish. At the end of the installation, you should see the following message:

   Your release is named websight-cms.
   
   Your WebSight CE instance is starting...
   WebSight instance will be available at: 
   - CMS Panel: http://cms.35.241.128.172.nip.io
   
   Sites:
   - luna: http://luna.35.241.128.172.nip.io
   
   Happy WebSight helming!
   

Step 4: Verification

1. Check the Kubernetes Workloads Dashboard to verify that all pods are running.
2. Open the WebSight CMS admin panel on http://cms.<YOUR_CLUSTER_IP>.nip.io/ (SSL is not covered in this guide). Use the credentials wsadmin/wsadmin to log in.
3. Publish some Luna pages (see Publish demo site guide for help).
4. Open http://luna.<YOUR_CLUSTER_IP>.nip.io/ to see the demo page.

Cleanup

1. When finished, you can delete your Kubernetes cluster along with all workloads using:

   gcloud container clusters delete my-websight-cms --zone europe-west1-b