# Installation on Amazon Elastic Kubernetes Service (EKS)

# Prerequisites

  • AWS CLI (opens new window)
  • AWS account
  • kubectl
  • A domain or the ability to purchase one. Can use route 53 for this inside AWS if doing it all inline
  • helm2 client

# Overview

The steps below walk you though installing the Entando platform in an EKS cluster. Generally the steps are:

  • Configure an IAM role to allow kubernetes to manage the cluster
  • Create an EKS cluster with 5 nodes (to allow expansion for microservices)
  • Install nginx as an ingress controller in the cluster
  • Register a domain (if you don't already have one) and configure it for wildcard subdomains.
  • Install Entando

If you're already comfortable setting up an EKS cluster and installing nginx then you may be able to skip to setting up Entando.

# Cluster Setup

These steps will use the AWS console to create the cluster. If you’re already familiar with creating an EKS cluster and assigning nodes to it via the AWS cli then you can use the cli process for cluster creation as well.

# Setup and Connect to the Cluster

  1. Login to AWS as a non-super admin user

    • If you don’t have a user besides the super admin it is recommended that you create one. Clusters created using the super admin for your account will have some restrictions that may complicate your installation.
    • Your user will need access to EKS and at least the ability to create a cluster. You may need additional policies for Route53 and other services depending on your exact configuration.
  2. Create an IAM role for the cluster so that AWS can provision assets. See https://docs.aws.amazon.com/eks/latest/userguide/worker_node_IAM_role.html (opens new window) for more details.

    • Select IAM from services
    • Select Create role
    • Select AWS Service box at the top for the type of trusted entity
    • Click EKS from the main list
    • Click EKS - Cluster under Select your use case
    • Click Next:Permissions
    • A Policy of AmazonEKSClusterPolicy should already be present
    • Click Next: Tags
    • (Optional) Add tags if you want
    • Click Next: Review
    • Name your role (you’ll need this later), e.g. my-eks-role
  3. Refine the role to enable Nodegroup management and to add ELB access so that the cluster can deploy a load balancer for nginx.

    • Go to IAM -> Roles -> your role.
    • Under permissions click Attach policies
    • Add a policy of AmazonEKSWorkerNodePolicy
    • Add a policy of AmazonEKS_CNI_Policy
    • Add a policy of AmazonEC2ContainerRegistryReadOnly
    • Add a policy of ElasticLoadBalancingFullAccess
  4. Go to Services and select Elastic Kubernetes Service

  5. Create an EKS Cluster

    • Add a cluster name (e.g. cluster-1) and click Create EKS cluster
    • For Cluster Service Role, select the role you created above, e.g. my-eks-role. If you choose a different role it must have ELB permissions so the cluster can create a load balancer in Networking (Step 2).
    • Click Next
    • Use the defaults for Networking (Step 2) and click Next
    • Use the defaults for Configure Logging (Step 3) and click Next.
    • Review your settings and then click Create. Cluster provisioning usually takes between 10 and 15 minutes.
    • See https://docs.aws.amazon.com/eks/latest/userguide/create-cluster.html (opens new window) for more information on cluster creation.
  6. Add a node group to the cluster

    • Go to Services -> Elastic Kubernetes Service -> Clusters -> Click on your cluster name.
    • Go to the Compute tab
    • Click Add Node Group
    • Name: give your group a name, e.g. node-1
    • Node IAM Role: Select the cluster role you created above. If the role doesn't appear, verify that you added the extra policies to the role.
    • Subnets - VPC subnets should already be setup and selected.
    • Select Allow remote access to nodes. Follow the links to create a new SSH key pair if you don't already have one.
    • Click Next
    • AMI type: Amazon Linux 2
    • Instance type: t3.medium
    • Click Next
    • Set Maximum size to 5. This will be over-resourced for a Getting Started experience but will leave capacity for adding microservices to your cluster without modifying the Nodegroup.
    • Click Next
    • Review your settings and then click Create
  7. Connect kubectl to the cluster

    • Note: If this is a brand new setup you will need to login using the user you used to create your cluster in the console in the steps above. Make sure the users match.
      • aws-configure (and then provide the Access key, etc.)
    • aws eks --region region-code update-kubeconfig --name cluster_name
    • More details and troubleshooting https://docs.aws.amazon.com/eks/latest/userguide/create-kubeconfig.html (opens new window)
    • Your current context should now be configured for your AWS cluster:
    $  kubectl config current-context
    arn:aws:eks:us-east-2:483173223614:cluster/cluster-1

# Install the NGINX Ingress Controller

  1. Add the NGINX controller for ingress. This depends on your role having permissions for ELB.
    • For basic nginx ingress install run this command
    kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/controller-v0.34.1/deploy/static/provider/aws/deploy.yaml
  1. Get the ELB external URL for your nginx install
    • Run: kubectl get services -n ingress-nginx
    • Get the value of the external address (EXTERNAL-IP) for the ingress-nginx-controller:
NAME                                 TYPE           CLUSTER-IP       EXTERNAL-IP                        
ingress-nginx-controller             LoadBalancer   10.100.102.83    ad234bd11a1ff4dadb44639a6bbf707e-0e0a483d966405ee.elb.us-east-2.amazonaws.com
  1. Determine the domain to use for your cluster. The goal here is to provide a way to route wildcard DNS traffic to the different parts of the apps and this can’t be done directly on the name for the ELB.
    • Option 1. Use a domain you already have available. You'll need to route traffic on that domain to the external cluster address noted in step 9.
      • For an existing domain you can add a wildcard subdomain via a CNAME, e.g. CNAME *.mysubdomain.domain.com <EXTERNAL-ADDRESS>. THe details will vary depending on your DNS registry.
    • Option 2. Register a domain in route 53.
      • Add wildcard dns resolution in route 53 to the ELB address attached to nginx above.
      • Note: The value in your A record will automatically include dualstack. This allows the ELB to serve both IPV4 and IPV6 traffic
    • If you register a new domain use nslookup or dig to make sure the DNS changes have propagated. This can take hours.

# Verify the NGINX Ingress Install

We recommend setting up a test application so you can easily verify the ingress is working in your cluster. See this page for those steps. You can use your local kubectl.

# Install the Entando Custom Resource Definitions (CRDs)

Once per cluster you need to deploy the Entando Custom Resources.

  1. Download the Custom Resource Definitions (CRDs) and unpack them
curl -L -C - https://raw.githubusercontent.com/entando/entando-releases/v6.2.0/dist/qs/custom-resources.tar.gz | tar -xz
  1. Install the Entando CRDs: sudo kubectl create -f dist/crd

# Deploy Your Entando Application

You can now deploy your application to Amazon EKS.

  1. Download and unpack the entando-helm-quickstart release here: https://github.com/entando-k8s/entando-helm-quickstart/releases (opens new window)
    • See the included README file for more information on the following steps.
curl -sfL https://github.com/entando-k8s/entando-helm-quickstart/archive/v6.2.0.tar.gz | tar xvz
  1. Edit values.yaml in the root directory:

    • Set supportOpenshift: false
    • Set ENTANDO_DEFAULT_ROUTING_SUFFIX to the URL of your external domain:
      • For example: ENTANDO_DEFAULT_ROUTING_SUFFIX: entando-aws-test.org
      • This assumes you have enabled wildcard dns address resolution above.
    • Configure nginx as the ingress controller and enable file system groups for persistent volume access:
      • ENTANDO_INGRESS_CLASS: "nginx"
      • ENTANDO_REQUIRES_FILESYSTEM_GROUP_OVERRIDE: "true"
    • See Appendix B for an example values.yaml
  2. Create the Entando namespace: kubectl create namespace entando

  3. Run helm to generate the template file:

helm template my-eks-app --namespace=entando ./ > my-eks-app.yaml
  1. Deploy Entando via kubectl create -f my-eks-app.yaml
  2. Watch Entando startup kubectl get pods -n entando --watch
  3. Check for the Entando ingresses using kubectl describe ingress -n entando
  4. Access your app on the url for the ingress of the app builder, e.g. http://quickstart-entando.mysubdomain.domain.com/entando-de-app

# Appendix A - Troubleshooting

IAM And Roles

NGINX

 Warning  SyncLoadBalancerFailed   38m                 service-controller  (combined from similar events): Error syncing load balancer: failed to ensure load balancer: error creating

# Appendix B - Example values.yaml file for Helm Quickstart

In the example below the application will deploy with embedded databases and will use nginx as the ingress controller. Replace <YOUR-DOMAIN> with the domain you've configured for your cluster.

app:
 name: quickstart
 dbms: none
operator:
 supportOpenshift: false
 env:
   ENTANDO_DOCKER_IMAGE_VERSION_FALLBACK: 6.0.0
   #ENTANDO_DOCKER_REGISTRY_OVERRIDE: docker.io # Remove comment if you want to always use a specific docker registry
   #ENTANDO_DOCKER_IMAGE_ORG_OVERRIDE: entando # Remove the comment if you want to always use a specific docker organization
   ENTANDO_DEFAULT_ROUTING_SUFFIX: <YOUR-DOMAIN>
   ENTANDO_POD_READINESS_TIMEOUT_SECONDS: "1000"
   ENTANDO_POD_COMPLETION_TIMEOUT_SECONDS: "1000"
   ENTANDO_DISABLE_KEYCLOAK_SSL_REQUIREMENT: "true"
   ENTANDO_K8S_OPERATOR_IMPOSE_DEFAULT_LIMITS: "false"
   ENTANDO_REQUIRES_FILESYSTEM_GROUP_OVERRIDE: "true"
   ENTANDO_INGRESS_CLASS: "nginx"
 tls:
   caCrt:
   tlsCrt:
   tlsKey:
deployPDA: false