Manage Nodeless Kubernetes clusters with Nirmata

Nodeless Kubernetes is a cloud-native, simple, secure, cost-effective way to consume Kubernetes on public cloud. The Nirmata Platform is a unified Kubernetes management plane that provides self-service DevOps and complete lifecycle management and operations of Kubernetes workloads and clusters. The Nirmata Platform is a great fit for consuming Nodeless Kubernetes in a seamless, secure, simple manner. This blog illustrates how to consume Nodeless Kubernetes via Nirmata.

Step 1: Create 1-worker Nodeless Kubernetes cluster

Follow instructions in this repo to create a {1 master, 1 worker} Nodeless Kubernetes cluster.

Log on to Kubernetes master, verify cluster is up.

$ ssh -i "myechuri-key2.pem" ubuntu@100.26.23.126

$ kubectl cluster-info
Kubernetes master is running at https://10.0.23.147:6443
KubeDNS is running at https://10.0.23.147:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy

To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.

$ kubectl get nodes -o wide
NAME                          STATUS   ROLES    AGE   VERSION   INTERNAL-IP   EXTERNAL-IP     OS-IMAGE             KERNEL-VERSION   CONTAINER-RUNTIME
ip-10-0-20-188.ec2.internal   Ready    worker   28d   v1.16.1   10.0.20.188   54.227.45.93    Ubuntu 16.04.6 LTS   4.4.0-1092-aws   containerd://1.2.6
ip-10-0-23-147.ec2.internal   Ready    master   28d   v1.16.1   10.0.23.147   100.26.23.126   Ubuntu 16.04.6 LTS   4.4.0-1092-aws   docker://18.9.7

Step 2: Import Nodeless cluster into Nirmata

Go to Nirmata Dashboard and initiate Create Cluster workflow.

Select Managing an existing Kubernetes Cluster option.

Fill in cluster Name and set Cloud Provider to Amazon Web Services.

Download nirmata-kube-controller.yaml.

Scp nirmata-kube-controller.yaml to a node with kubeconfig access to Nodeless cluster.

$ scp -i myechuri-key2.pem /Users/myechuri/Downloads/nirmata-kube-controller.yaml ubuntu@100.26.23.126:/home/ubuntu/nirmata/

Apply nirmata-kube-controller.yaml.

$ kubectl create -f nirmata-kube-controller.yaml 
namespace/nirmata created
serviceaccount/nirmata created
clusterrole.rbac.authorization.k8s.io/nirmata:cluster-admin created
clusterrolebinding.rbac.authorization.k8s.io/nirmata-cluster-admin-binding created
deployment.apps/nirmata-kube-controller created

Fun fact: Nirmata-kube-controller pod will run in a Nodeless compute cell!

Verify that Nirmata namespace components are up and running.

$ kubectl get all -n nirmata
NAME                                           READY   STATUS    RESTARTS   AGE
pod/nirmata-kube-controller-859c54dd97-wdn48   1/1     Running   0          3m

NAME                                      READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/nirmata-kube-controller   1/1     1            1           3m

NAME                                                 DESIRED   CURRENT   READY   AGE
replicaset.apps/nirmata-kube-controller-859c54dd97   1         1         1       3m

On Nirmata dashboard, click I have installed the Nirmata Kubernetes Cluster.

Your Nodeless k8s cluster is now fully managed by Nirmata!

Step 3: Deploy Nginx workload, manage via Nirmata dashboard

Deploy Nginx deployment with 3 replicas.

$ wget https://raw.githubusercontent.com/elotl/nodeless-nirmata/master/nginx.yaml

$ kubectl create -f nginx.yaml

All 3 replicas of Nginx will be deployed in a Nodeless fashion — just-in-time, right-sized, cost-effective compute will be started for each of the replicas, and the pods will be dispatched to the compute cells.

Nginx pods can be visualized through Nirmata dashboard.

Teardown

Follow teardown instructions from kubeadm repo.

Summary

Nirmata and Nodeless provide a compelling solution to increase your cloud efficiency and reduce costs! Nodeless extends the powerful Nirmata Kubernetes management platform to optimize the use of cloud resources and enable advanced hybrid cloud use cases with a single control plane.

References

Video: Manage Nodeless Kubernetes with Nirmata

Nodeless Kubernetes blogs

Nirmata webinars

Originally published at http://github.com.