Creating a Kubernetes cluster on AWS using kops

This article will help you to set up a Kubernetes cluster on AWS using kops. A command-line utility that helps you create your own master and worker nodes on AWS.

The Major difference from using the eksctl is — eksctl uses the EKS managed service that manages the master nodes for you, but here you manage the master nodes as well.

Requirements

a free-tier AWS account on which your user has the necessary permissions on IAM.
an access key for the above mention account which can be downloaded from AWS Console
AWS CLI 2 configured on your machine using AWS configure — kops will use this to connect to AWS Console
Install kops on your machine

Getting Started

kops much like terraform maintains a state for your cluster, the state is stored remotely in this case, an S3 bucket is needed. Follow the snippet below to create a bucket that will hold your kops state. A single bucket can hold multiple states and states of multiple clusters.

REGION="us-east-1"
BUCKET_NAME="sanjaybalaji-kops-store"

# create a bucket in a specified region
aws s3api create-bucket --bucket ${BUCKET_NAME} --region=${REGION}

# enable bucket versioning so that your changes are not lost
aws s3api put-bucket-versioning --bucket ${BUCKET_NAME} --region=${REGION} --versioning-configuration Status=Enabled

# validate that bucket is created successfully
aws s3 ls

Creating the cluster config

Once you are done with step 1, run the following to create your cluster configuration

# this can be used to avoid passing --state to all kops commands
export KOPS_STATE_STORE=s3://${BUCKET_NAME}

# this command creates a cluster configuration of a particular version

kops create cluster --name sanjaybalaji.k8s.local --zones eu-west-1a --master-size t2.micro --node-size t2.micro --master-count 1 --node-count 2 --kubernetes-version 1.18.0 --api-loadbalancer-class network

This command only creates the cluster config, it does not provide your cluster.

The os, size, and count can be altered, you can spread the cluster across zones by passing in a CSV of other az’s — note that every zone needs one master minimum. The API load balancer class is set to Classic by default, but that might cause an authentication issue with certain versions of Kubernetes, so we are using the Network type load balancer.

Check and Deploy your cluster

To review your cluster configuration, you can run the following

kops edit cluster --name sanjaybalaji.k8s.local

This will give you a configuration that is editable. Once you are sure of your config, you can move on to deployment

kops update cluster --name sanjaybalaji.k8s.local --yes

This will take a few minutes to provision the cluster by creating the VCN, subnets, nodes, security groups etc.

Validate and Test your cluster

To validate your cluster you can run the following command, which will make sure that your cluster is in the expected state.

kops validate cluster — name sanjaybalaji.k8s.local — wait 10m

In case your validate fails even after 10–15 mins with an Unathorized error, run this command kops export kubecfg –admin and then re-run the validate command.

To try a test run, you can create an nginx pod — and access http://localhost:8080 and you should see an Nginx welcome page.

kubectl run nginx --image nginx

kubectl port-forward nginx 8080:80

Cleanup

Finally, To delete your cluster and all of it’s supporting resources, simply run the following command

kops delete cluster --name sanjaybalaji.k8s.local --yes

For the S3 bucket, you would have to clean it and delete it from the AWS Console GUI.