Elastic Kubernetes Service

Kiran ChavanKiran Chavan
7 min read

In case of one pod is dead, Scheduler(Through API ) is schedule one pod in deployment mention 2 pods req.

Mantra - Everything in k8s is manifests files

Whole things like pod, services, deployment,.etc Everything is K8S is either command or manifests file. All manifest file are available k8s document (Web). Ex: Pod

kubernetes.io/docs/concepts/workloads/pods/

Basically It written in YML format

apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  containers:
  - name: nginx
    image: nginx:1.14.2
    ports:
    - containerPort: 80

yaml file executed through kubectl (kubecontrol). It trigger API server.

apiVersion - As interacts with API server

kind - What I want to make?

metadata - Information of kind(first block) It's Obj. like name

spec - specification These are details

Practice in AWS

Steps

Method 1: Through Command

mkdir  k8s-practice
cd k8s-practice
vim pod.yml

apiVersion: v1
kind: Pod
metadata:
  name: nignx-pod

spec:
  containers:
    -name: nginx
     image: nginx:latest
     ports:
       - containerPort: 80

image: nginx:latest - Executes Docker run

Save the file :wq

If namespace is not assign a default name is given.

kubectl get pods

Checks the available pods.

kubectl create namespace nginx

Create namespace with nginx (Isolated group.)

Method 2: Through Manifest

vim namespace.yml

kubernetes.io/docs/tutorials/cluster-management/namespaces-walkthrough/

apiVersion: v1
kind: Namespace
metadata:
  name: development
  labels:
    name: development

label -> To categorize

kubectl get namespace
kubectl delete namespace nginx

Checks the available namespace and deletes.

kubectl apply -f namespace.yml

Creates namespace.yml file

vim pod.yml

apiVersion: v1
kind: Pod
metadata:
  name: nignx-pod
  namespace: nginx 

spec:
  containers:
    -name: nginx
     image: nginx:latest
     ports:
       - containerPort: 80

kubectl apply -f pod.yml
kubectl get pods -n nginx

Creates pod.yml file

pod is created with pre-defined namespcae. Whereas w/o flag -n(namespace) is created with default namespace as mentioned in Method-1.

kubectl delete -f pod.yml

Replica If multiple pod req. then deployment is needed.

vim deployment.yml

kubernetes.io/docs/concepts/workloads/controllers/deployment/

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: nginx
  labels:
    app: nginx
spec:
  replicas: 5
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      name: nginx-pod
      namespace: nginx
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        ports:
        - containerPort: 80

Save the file :wq

Explanation

apiVersion: apps/v1 -> Special type for deployment

kind: Deployment -> What kind of file

metadata : -> Info. above apiVersion

labels: -> Tag/ Categorize

spec: -> Specification

selector -> Criteria

matchLabels: -> Finds the pattern label

deployment.yml file is completed just above template section spec(replicas) from template onwards we are providing pod.yml file config.

In metadata and template, I included namespace as that in nginx group.

kubectl apply --validate=true --dry-run=client --filename=deployment.yml
kubectl apply -f deployment.yml
kubectl get pods -n nginx

validate command is used to check the syntax.

It will show 5 pods has deployed as written as deployment.yml

If I want to increase the pods size (replicas) vi deployment.yml file update the values of replica and save the file. and run kubectl apply -f deployment.yml, file get config. and get pods show pod running status.

To check Auto-Healing status

kubectl delete pod nginx-deploymet-7f78fb778(pod name) -n nginx
kubectl delete pod (pod name) -n nginx
kubectl get pods -n nginx

We notice that even after deletion of pods new pods are created as no. of replica mention in deployment.yml file.

kubectl describe pod (pod_name) -n nginx

Gives whole info. about that particular pod. This cmd is used during crush/downtime of pod.

kubectl get all -n nginx

It gives all the details of pod, deployment, replica set.

To Communicate with external world

kubernetes.io/docs/concepts/services-networking/service/

targetPort -> pod/Deployment (container is running)

port -> Service

In webpage, Search for Load balancer

Video Timestamp 3:05:40

vim service.yml

apiVersion: v1
kind: Service
metadata:
  name: nginx-service
  namespace: nginx
  labels: 
    app: nginx 
spec:
  selector:
    app: ngnix
  ports:
    - protocol: TCP
      port: 80
      targetPort: 80
 type: LoadBalancer

kubectl apply -f service.yml
kubectl get deployment -n nginx
kubectl get service -n nginx

Using External IP from o/p of service cmd. We can view the nginx webpage.

Fun fact

Note-app -> Docker container

Container is formed by Image

Image is formed by Docker file

If we change the image then we access the services inside that the environment.

vim deployment.yml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: nginx
  labels:
    app: nginx
spec:
  replicas: 5
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      name: nginx-pod
      namespace: nginx
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: trainwithshubham/notes-app:latest
        ports:
        - containerPort: 8000

vim service.yml

apiVersion: v1
kind: Service
metadata:
  name: nginx-service
  namespace: nginx
  labels: 
    app: nginx 
spec:
  selector:
    app: ngnix
  ports:
    - protocol: TCP
      port: 80
      targetPort: 8000
 type: LoadBalancer

Save these files :wq

kubectl apply -f deployment.yml
kubectl apply -f service.yml
kubectl get svc -n nginx

Exact external IP to check the service.

To delete eks cluster

Initially delete the resources

kubectl delete -f .
kubectl apply -f .
kubectl get all -n nginx
eksctl delete cluster --name=tws-cluster --region=us-west-2

kubectl apply -f .

It will recreate.

Check manually also whether it has deleted or not.

Hands-on & Quiz is pending.

Do it for free?

killercoda.com/

Login to killercoda

Select playground -> kubernetes 1.32

git clone github.com/LondheShubham153/aws-eks-devops-best-practices.git 
cd aws-eks-devops-best-practices/01-deploy-sample-application
kubectl create namespace eks-sample-app
kubectl apply -f .
kubectl get all -n eks-sample-app # To get external-ip

Quiz

  1. Which Kubernetes component represents the smallest deployable unit? Node Pod Deployment Service

  2. Which component in Kubernetes stores the entire cluster state and can facilitate rollbacks of deployments? etcd ConfigMaps Deployment Service

  3. Which K8s component is responsible for assigning newly created pods to specific nodes based on various criteria like resource requirements and affinity rules? kube-controller-manager kube-scheduler Kubeapi-server kube-proxy

  4. The scheduler is responsible for both deciding where a Pod will run and actually running it. True False

  5. What are the primary components of a Kubernetes master node (Control Plane)? kube-proxy,kubelet,Docker API server, kubelet, kube-proxy kube-proxy,kubelet,Docker APIServer,ControllerManager, etcd, Scheduler

  6. What is Kubernetes kubectl? A Command-line tool used to interact with k8s clusters A tool for managing k8s resources A networking abstraction to access a set of Pods A way to manages K8s nodes

  7. What is Kubernetes Horizontal Pod Autoscaler(HPA)? A tool for managing Kubernetes resources A way to automatically scale the number of pods in a k8s deployment based on resource utilization. A networking abstraction to access a set of pods A way to manages k8s nodes

  8. Which of the following is a main component of k8s architectures? API Server Container Runtime kubectl Command Line interface

  9. What is the primary purose of kubeadm? To deploy K8s cluster To manage k8s To provide security for k8s cluster To provide appln development tools

  10. Which of the following is a kubeadm cmd used to join node to a cluster? kubeadm init kubeadm join kubeadm reset kubeadm version

Answer

  1. Pods are the smallest deployable units of computing that you can create and manage in Kubernetes. A Pod is a group of one or more containers.

  2. etcd is a consistent and highly available key-value store used by K8s to save all cluster data, ensuring cluster state and configuration stability ( etcd is typically present as a present as a separate cluster of nodes or on the master node's)

  3. kube-scheduler intelligently determines the best node for a pod, considering factors like resource availability, node constraints, affinity and anti-affinity specifications and other set policies.

  4. False while the scheduler is responsible for choosing the node on which the pod will run, Kubelet is the one that actually runs the Pod.

  5. APIServer,ControllerManager, etcd, Scheduler Other options like kubeproxy and kubelet are related to worker nodes.

  6. A Command-line tool used to interact with k8s clusters enabling administratives to manages k8s resource, create, modify and delete resource and troubleshoot cluster issues.

  7. Horizontal Pod Autoscaler(HPA) is way to automatically scale the number of pods in a k8s deployment based on resource utilization, enabling administrators to ensure that their applications can handle varying levels of traffic and load

  8. API Server is the main component of K8S arch. It is responsible for accepting user requests and managing the state of the cluster.

  9. Kubeadm is a tool for quickly and easily deploying k8s cluster on existing infrastructure

  10. kubeadm join is used to join a node to a cluster, while kubeadm init is used to initialize a cluster kubeadm reset is used to reset a cluster, and kudeadm version is used to check the version of kubeadm.

1
Subscribe to my newsletter

Read articles from Kiran Chavan directly inside your inbox. Subscribe to the newsletter, and don't miss out.

DevopsAWSKubernetes#TWS

Written by

Kiran Chavan
Kiran Chavan