Simple Usage

As you know, many of contributors have built best practices into Mahjong’s building blocks. We called it “tile”

In this case, we’re going to deploy a AWS EKS cluster

You only need to define your cluster name, nodes number, instance types and k8s version. Mahjong will help you to build cluster with minimal permission, cross AZ environment

Define and build service component

Know how to use Mahjong

Mahjong use Yaml to control deployment and tiles definition, including:

Tile

A building block, defined by YAML, represents a cloud component or a combination of multiple cloud components or resources. Tile is categorized by Network, ContainerProvider, Storage, Database, Application, ContainerApplication, Analysis, ML. Application and ContainerApplication are represented through commands and files, and rest of categories are represented through Construct::CDK

Deployment

A unit of deployment, defined by YAML, and all resources defined within the scope of Tiles.

“Hu”

A high level collection of deployment units, defined by YAML, represents a full solution and includes multiple Tiles with specific definition.

high-level-arch

Define your EKS cluster

Now, let’s define your EKS cluster with a simple yaml file

create simple-eks.yaml

    apiVersion: mahjong.io/v1alpha1
    kind: Deployment
    metadata:
    name: simple-eks
    version: 0.1.0
    spec:
    template:
        tiles:
        tileEks0005:
            tileReference: Eks0
            tileVersion: 0.0.5
            inputs:
            - name: clusterName
                inputValue: mahjong-cluster101
            - name: capacity
                inputValue: 3
            - name: capacityInstance
                inputValue: m5.large
            - name: version
                inputValue: 1.16
    summary:
        description:
        outputs:
        - name: EKS Cluster Name
            value: $(tileEks0005.outputs.clusterName)
        - name: Master role arn for EKS Cluster
            value: $(tileEks0005.outputs.masterRoleARN)
        - name: The API endpoint EKS Cluster
            value: $(tileEks0005.outputs.clusterEndpoint)
        - name: Instance type of worker node
            value: $(tileEks0005.outputs.capacityInstance)
        - name: Default capacity of worker node
            value: $(tileEks0005.outputs.capacity)

        notes: []

Deploy it to AWS cloud

Run mctl deploy -f ./simple-eks.yaml

    $ mctl deploy -f ./simple-eks.yaml

You will get the following output if everything works as expected


    [ℹ]  EKS Cluster Name = mahjong-cluster101
    [ℹ]  Master role arn for EKS Cluster = arn:aws:iam::638198787577:role/Eks0StacktileEks0005-Eks0EksClusterMasterRole76926-D9OV6NASDYGC
    [ℹ]  The API endpoint EKS Cluster = https://8F4AEE06CDA95AA5B9B82016B406F53B.gr7.us-east-2.eks.amazonaws.com
    [ℹ]  Instance type of worker node = m5.large
    [ℹ]  Default capacity of worker node = 3

Finish the deployment

Now your have finished AWS EKS deployment with Mahjong

eks

Verification

Verify EKS

Once you have finished the deployment, you can check your cluster mahjong-cluster101

  • Find the master role in deployment logs

    For example:

    arn:aws:iam::638198787577:role/Eks0StacktileEks0005-Eks0EksClusterMasterRole76926-D9OV6NASDYGC

  • Run aws --region <region-code> eks update-kubeconfig --name <cluster_name> --role-arn arn:aws:iam::<aws_account_id>:role/<role_name> to save k8s logging configuration

    For example:

    aws --region us-east-2 eks update-kubeconfig--name mahjong-cluster101 --role-arn arn:aws:iam::638198787577:role/Eks0StacktileEks0005-Eks0EksClusterMasterRole76926-D9OV6NASDYGC

  • Install eksctl

    curl --silent --location "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" | tar xz -C /tmp
sudo mv /tmp/eksctl /usr/local/bin

  • Install kubectl

    curl -LO "https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/darwin/amd64/kubectl"
chmod +x ./kubectl
sudo mv ./kubectl /usr/local/bin/kubectl
kubectl version --client

  • Check the cluster

    $ kubectl get svc
NAME         TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   172.20.0.1   <none>        443/TCP   4h41m

Use the EKS

Now let’s deploy a web game with the provisioned EKS cluster

  • create a file 2048.k8s.yaml

    apiVersion: v1
kind: Service
metadata:
name: my2048-service
spec:
ports:
    - port: 80
    targetPort: 80
selector:
    app: my2048-k8s
type: LoadBalancer
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: my2048-deployment
spec:
replicas: 2
selector:
    matchLabels:
    app: my2048-k8s
template:
    metadata:
    labels:
        app: my2048-k8s
    spec:
    containers:
        - image: alexwhen/docker-2048
        name: "my2048"
        ports:
            - containerPort: 80

  • Run deployment

    $ kubectl apply -f 2048.k8s.yaml

  • Check service status

    $ kubectl get service/my2048-service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
my2048-service LoadBalancer 172.20.88.7 a6ba48df2f1774376a8e065336c8bd44-1753238188.us-east-1.elb.amazonaws.com 80:31276/TCP 4d20h

  • Access your service

    open your application endpoint in your browser

    2048