Wallaroo Community: A Quick Setup Guide for AWS Kubernetes Environment

AWS Install Demo

Wallaroo’s simple yet efficient cloud-based AI platform is available for multiple environment setups. Here we will show you how you can set up your Kubernetes cloud environment for installing Wallaroo Community using AWS (Amazon Web Services).

Prerequisites for AWS before Starting

Here are some of the requirements that must be met for you to install Wallaroo Community:

• Require a registered AWS account.
• Enable Amazon Elastic Kubernetes Service (EKS).
• Enable Role-Based Access Control (RBAC).
• Have a Wallaroo Community license file.
• Install AWS CLI.
• Install eksctl to setup their Kubernetes environment.
• Install kubectl and kots on your local system.

Setting Up AWS for your Kubernetes Environment

For this tutorial, we are going to be setting up the main pool, postgres, and machine types most suitable for the task. The Kubernetes device driver should be containerd and c5.4xlarge for recommended AWS machine type.

Step 1: Creating an eksctl Configuration File

We will have to create an AWS EKS cluster. Start with naming your AWS cluster. Here we’ll go with: wallarooCEAWS. The cluster will contain the cluster name, region, and minimum machine type for each of the nodepools and other configurations.

For Wallaroo Community, having only static nodepools, the configuration file will look like this:

apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
  name: wallarooCEAWSDocTest
  region: us-east-2
  version: "1.21"

nodeGroups:
  - name: mainpool
    instanceType: m5.2xlarge
    desiredCapacity: 1
    containerRuntime: containerd
    amiFamily: AmazonLinux2
  - name: postgres
    instanceType: m5.2xlarge
    desiredCapacity: 1
    containerRuntime: containerd
    amiFamily: AmazonLinux2
    taints:
      - key: postgres
        value: "true"
        effect: NoSchedule
  - name: engine-lb
    instanceType: c5.4xlarge
    containerRuntime: containerd
    amiFamily: AmazonLinux2
    taints:
      - key: engine-lb
        value: "true"
        effect: NoSchedule
    iam:
      withAddonPolicies:
        autoScaler: true
  - name: engine
    instanceType: c5.2xlarge
    containerRuntime: containerd
    amiFamily: AmazonLinux2
    taints:
      - key: engine
        value: "true"
        effect: NoSchedule
    iam:
      withAddonPolicies:
        autoScaler: true

Step 2: Creating the Cluster

After your eksctl configuration file is ready, you can go ahead and run it with the following command to create your cluster:

eksctl create cluster -f aws_community.yaml

Here, “aws_community.yaml” is the name of your configuration file. So make sure to replace the information correctly.

If you have entered the commands correctly, your Kubernetes configuration should be applied to your local terminal environment automatically. This command can take from 10 to 30 minutes to complete depending on your AWS setup connection. The most important thing is to make sure there were no errors in the process.

Step 3: Setup Verification

In the previous step, the Kubernetes credentials will get copied into your local system. So now you can verify if your setup was successful or not. For that, enter the command next:

kubectl get nodes

Step 4: Installing Wallaroo Community into your Kubernetes Environment

Go to where your Wallaroo Community is installed and start by creating a namespace. The command will be:

kubectl create namespace wallaroo

Now you can enter the final command before successfully installing Wallaroo Community into your Kubernetes environment.

kubectl kots install wallaroo/ce — namespace wallaroo

This will download all the necessary containers into your Kubernetes environment, and you can get started by opening your workshop and account on Wallaroo Community.

The information in the commands can be replaced with your requirements as long as it meets the prerequisites we mentioned. You can find ctl-based configuration files and commands for AWS setup if you want to quicken the process or learn more about the steps on the Wallaroo documentation site.