Deploy the aws-efa-k8s-device-plugin helm chart on AWS EKS

Question

Pulumi · Accepted Answer

To deploy the AWS EFA (Elastic Fabric Adapter) Kubernetes Device Plugin using Helm on an Amazon EKS cluster, you will need to complete several steps, including:

1. **Provision an EKS Cluster**: Create a new Amazon EKS cluster where your workloads will run.
2. **Create an ECR Repository**: Although not always necessary for deploying Helm charts, an ECR (Elastic Container Registry) is useful if you plan to store and manage your own container images.
3. **Deploy the Helm Chart**: Using the Helm package manager, deploy your AWS EFA Kubernetes Device Plugin chart to the EKS cluster.

Below is a detailed Pulumi TypeScript program that accomplishes these tasks:

### Program Explanation

1. The `aws.eks.Cluster` resource is used to provision an EKS cluster. This requires specifying a name, `roleArn`, and `vpcConfig`. The `subnetIds` within `vpcConfig` are an array of subnet IDs that are associated with the EKS cluster.
   
2. Optionally, an `aws.ecr.Repository` is included, creating an ECR where you could push container images if needed.
   
3. Finally, the `kubernetes.helm.v3.Chart` resource deploys the aws-efa-k8s-device-plugin Helm chart to the EKS cluster.

### Pulumi Program

```typescript
import * as pulumi from "@pulumi/pulumi";
import * as aws from "@pulumi/aws";
import * as eks from "@pulumi/eks";
import * as k8s from "@pulumi/kubernetes";

// Create an EKS cluster
const cluster = new eks.Clubster("eksCluster", {
    // Specify the desired Kubernetes version for your cluster
    version: "1.21",
    // Define the VPC configuration for your cluster
    vpcId: "vpc-12345", // Replace with your VPC ID
    subnetIds: ["subnet-12345abc", "subnet-67890def"], // Replace with your Subnet IDs
    // Additional configurations can be added as needed
});

// You can create an ECR repository if you intend to manage custom container images
const repository = new aws.ecr.Repository("myECRRepo", {
    // Container images in ECR can be manually tagged or can be set to be immutable
    imageTagMutability: "MUTABLE", // or "IMMUTABLE"
    // Enabling scan on push ensures images are scanned for vulnerabilities upon being pushed to the repository
    imageScanningConfiguration: {
        scanOnPush: true,
    },
    // Additional configurations can be provided as necessary
});

// Deploy the Helm chart for AWS EFA Kubernetes Device Plugin
const efaHelmChart = new k8s.helm.v3.Chart("awsEfaDevicePlugin", {
    chart: "aws-efa-k8s-device-plugin", // Replace with the correct chart name if different
    version: "1.0.0", // Set the specific chart version you want to deploy
    fetchOpts: {
        repo: "https://aws.github.io/eks-charts", // Helm repo URL containing the chart
    },
    // Define chart values based on the specific configuration needs; refer to the chart's documentation for options
    values: {
        // Values that configure the AWS EFA device plugin
    },
}, { provider: cluster.provider });

// Export the cluster's kubeconfig and ECR repository URL
export const kubeconfig = cluster.kubeconfig;
export const ecrRepositoryUrl = repository.repositoryUrl;
```

When deploying EFA plugins or working with advanced network interfaces in EKS, it is often necessary to ensure that the prerequisites are met, such as having the proper EC2 instance types and the right networking setup. For the EFA device plugin specifically, you should refer to its official documentation to understand any additional requirements and configuration settings that may be necessary.

Remember, the actual chart name, version, and values to use should be verified from the official AWS EFA Helm chart documentation or the chart's `values.yaml` file.

After you've crafted your Pulumi program, you can apply it using the `pulumi up` command. This command will provision the resources as specified in the program, and Pulumi will give you a preview of the changes before they are applied.