Kubernetes has become the preferred platform of choice for container orchestration and deliver maximum operational efficiency. To understand how K8s works, one must understand its most basic execution unit – a pod.

Kubernetes doesn’t run containers directly, rather through a higher-level structure called a pod. A pod has an application’s container (or, in some cases, multiple containers), storage resources, a unique network IP, and options that govern how the container(s) should run.

Pods can hold multiple containers or just one. Every container in a pod will share the same resources and network. A pod is used as a replication unit in Kubernetes; hence, it is advisable to not add too many containers in one pod. Future scaling up would lead to unnecessary and expensive duplication.

To maximize the ease and speed of Kubernetes, DevOps teams like to add automation using Jenkins CI/CD pipelines. Not only does this make the entire process of building, testing, and deploying software go faster, but it also minimizes human error. Here is how Jenkins CI/CD pipeline is used to deploy a spring boot application in K8s.

TASK on Hand:

Create a Jenkins pipeline to dockerize a spring application, build docker image, push it to the dockerhub repo and then pull the image into an AKS cluster to run it in a pod.

Complete repository:

All the files required for this task are available in this repository:


A spring-boot application, dockerfile to containerize the application.


1. Install Jenkins :

2. Connect host docker daemon to Jenkins:

  • Run the command: chmod –R Jenkins:docker filename/foldername to allow Jenkins to access docker.
  • Go to manage Jenkins from browser >Configure System and scroll to the bottom
  • Click the dropdown ‘add cloud’ and add Docker. Add the docker host URI in the format tcp://hostip:4243
  • Click verify connection to check your connection. If everything was done right, the docker version is displayed.

3. Adding global credentials:

  • Go to Credentials on the Jenkins dashboard, click global credentials, and then Add credentials.
  • Select the kind as Microsoft Azure Service Principal and enter the required ids, similarly save the docker credentials under type username with password.

4. Create the Jenkinsfile :

  • Refer to the official Jenkins documentation for the pipeline syntax, usage of Jenkinsfile, and simple examples.
  • Below is the Jenkinsfile used for this task.


NOTE: While this example uses actual id to login to Azure, its recommended to use credentials to avoid using exact parameters.

pipeline {
environment {
registryCredential = "docker"
agent any
stages {
stage(‘Build’) {
    script {
        sh 'mvn clean install'
stage(‘Load’) {
    script {
        app = docker.build("cloud007/simple-spring")
    stage(‘Deploy’) {
    script {
        docker.withRegistry( "https://registry.hub.docker.com", registryCredential ) {
        // dockerImage.push()
stage('Deploy to ACS'){
        withCredentials([azureServicePrincipal('dbb6d63b-41ab-4e71-b9ed-32b3be06eeb8')]) {
        sh 'echo "logging in" '
        sh 'az login --service-principal -u **************************** -p ********************************* -t **********************************’
        sh 'az account set -s ****************************'
        sh 'az aks get-credentials --resource-group ilink --name    mycluster'
        sh 'kubectl apply -f sample.yaml'

5. Create the Jenkins project:

  • Select New view>Pipeline and click ok.
  • Scroll to the bottom and select the definition as Pipeline from SCM.
  • Select the SCM as git and enter the git repo to be used, path to Jenkinsfile in Script path.
  • Click apply, and the Jenkins project has now been created.
  • Go to my views, select your view, and click on build to build your project.

6. Create and connect to Azure Kubernetes cluster:

  • Create an Azure Kubernetes cluster with 1-3 nodes and add its credentials to global credentials in Jenkins.
  • Install azure cli on the Jenkins host machine.
  • Use shell commands in the pipeline to log in, get-credentials, and then create a pod using the required yaml file.

YAML used:

apiVersion: apps/v1
kind: Deployment
    name: spring-helloworld
    replicas: 1
        app: spring-helloworld
        app: spring-helloworld
        - name: spring-helloworld
        image: cloud007/simple-spring:latest
        imagePullPolicy: Always
        - containerPort: 80

Here are some common problems faced during this process and the troubleshooting procedure.

  • Corrupt Jenkins exec file:
    Solved by doing an apt-purge and then apt-install Jenkins.
  • Using 32-bit VM:
    Kubectl is not supported on a 32-bit machine and hence make sure the system is 64-bit.
  • Installing azure cli manually makes it inaccessible for non-root users
    Manually installing azure cli placed it in the default directories, which were not accessible by non-root users and hence by Jenkins. So, it is recommended to install azure cli using apt.
  • Installing minikube using local cluster instead of AKS:
    Virtual box does not support nested VTx-Vtx virtualization and hence cannot run minikube. It is recommended to enable Hyperv and use HyperV as the driver to run minikube.
  • Naming the stages in the Jenkinsfile:
    Jenkins did not accept when named stage as ‘Build Docker Image’ or multiple words for some reason. Use a single word like ‘Build’, ‘Load’ etc…
  • Jenkins stopped building the project when the system ran out of memory:
    Make sure the host has at least 20 GB free in the hard disk before starting the project.
  • Jenkins couldn’t execute docker commands:
    Try the command usermog –a  –G docker Jenkins
  • Spring app not accessible from external IP:
    Created a new service with type loadbalancer, assigned it to the pod, and the application was accessible from this new external ip.

In an upcoming article we will show you how to deploy a pod containing three applications using Jenkins ci/cd pipeline and update them selectively.

Container and container orchestration have become the default system for any DevOps team that wants to scale on-demand, reduce costs, and deliver faster. And to get the best out of container technology, Kubernetes is the way to go. A recommended Kubernetes practice is to manage pods through a Deployment; this way, they can be monitored and restarted if a failure occurs.

A deployment is created by using a Kubernetes Deployment Controller object. The application (in a container) is deployed to Kubernetes by declaratively passing a desired state to the Kubernetes Deployment Controller. A K8s deployment controller object is utilized for monitoring, management of upgrade, downgrade, and scaling of services (e.g., pods) without any disruption or downtime. This is made possible because the deployment controller is the single source of truth for the sizes of new and old replica sets. It maintains multiple replica sets, and when you describe a desired state, the DC changes the actual state at the correct pace.

Here’s a detailed look at the inner workings of Kubernetes Deployment Controller

K8s deployment controller is responsible for the following functions

– Managing a set of pods in the form of Replica Sets & Hash-based labels
– Rolling out new versions of application through new Replica Sets
– Rolling back to old versions of application through old Replica Sets
– Pause & Resume Rollout/Rollback functions
– Scale-Up/Down functions

“The Kubernetes controller manager is a daemon that embeds the core control loops shipped with Kubernetes. In applications of robotics and automation, a control loop is a non-terminating loop that regulates the state of the system. In Kubernetes, a controller is a control loop that watches the shared state of the cluster through the apiserver and makes changes attempting to move the current state towards the desired state. Examples of controllers that ship with Kubernetes today are the replication controller, endpoints controller, namespace controller, and serviceaccounts controller.”

func NewControllerInitializers(loopMode ControllerLoopMode) map[string]InitFunc {
        controllers := map[string]InitFunc{}
        controllers["endpoint"] = startEndpointController
        controllers["endpointslice"] = startEndpointSliceController
        controllers["replicationcontroller"] = startReplicationController
        controllers["podgc"] = startPodGCController
        controllers["resourcequota"] = startResourceQuotaController
        controllers["namespace"] = startNamespaceController
        controllers["serviceaccount"] = startServiceAccountController
        controllers["garbagecollector"] = startGarbageCollectorController
        controllers["daemonset"] = startDaemonSetController
        controllers["job"] = startJobController
        controllers["deployment"] = startDeploymentController
        controllers["replicaset"] = startReplicaSetController
        controllers["horizontalpodautoscaling"] = startHPAController
        controllers["disruption"] = startDisruptionController
        controllers["statefulset"] = startStatefulSetController
        controllers["cronjob"] = startCronJobController
        controllers["csrsigning"] = startCSRSigningController
        controllers["csrapproving"] = startCSRApprovingController
        controllers["csrcleaner"] = startCSRCleanerController
        controllers["ttl"] = startTTLController
        controllers["bootstrapsigner"] = startBootstrapSignerController
        controllers["tokencleaner"] = startTokenCleanerController
        controllers["nodeipam"] = startNodeIpamController
        controllers["nodelifecycle"] = startNodeLifecycleController
 	if loopMode == IncludeCloudLoops {
                controllers["service"] = startServiceController
                controllers["route"] = startRouteController
                controllers["cloud-node-lifecycle"] = startCloudNodeLifecycleController
                // TODO: volume controller into the IncludeCloudLoops only set.
        controllers["persistentvolume-binder"] = startPersistentVolumeBinderController
        controllers["attachdetach"] = startAttachDetachController
        controllers["persistentvolume-expander"] = startVolumeExpandController
        controllers["clusterrole-aggregation"] = startClusterRoleAggregrationController
        controllers["pvc-protection"] = startPVCProtectionController
        controllers["pv-protection"] = startPVProtectionController
        controllers["ttl-after-finished"] = startTTLAfterFinishedController
        controllers["root-ca-cert-publisher"] = startRootCACertPublisher

        return controllers

Let’s look at inside workings of “Deployment” Controller. It watches for following object updates.

func startDeploymentController(ctx ControllerContext) (http.Handler, bool, error) {
        if !ctx.AvailableResources[schema.GroupVersionResource{Group: "apps", Version: "v1", Resource: "deployments"}] {
                return nil, false, nil
        dc, err := deployment.NewDeploymentController(
        if err != nil {
                return nil, true, fmt.Errorf("error creating Deployment controller: %v", err)
        go dc.Run(int(ctx.ComponentConfig.DeploymentController.ConcurrentDeploymentSyncs), ctx.Stop)
        return nil, true, nil

The “Deployment Controller” initializes the following Event handlers.

                AddFunc:    dc.addDeployment,
                UpdateFunc: dc.updateDeployment,
                // This will enter the sync loop and no-op, because the deployment has been deleted from the store.
                DeleteFunc: dc.deleteDeployment,
                AddFunc:    dc.addReplicaSet,
                UpdateFunc: dc.updateReplicaSet,
                DeleteFunc: dc.deleteReplicaSet,
                DeleteFunc: dc.deletePod,

Since Kubernetes uses asynchronous programming, the events are processed through work queues and workers.

func (dc *DeploymentController) addDeployment(obj interface{}) {
        d := obj.(*apps.Deployment)
        klog.V(4).Infof("Adding deployment %s", d.Name)

The items from the queue are handled by “syncDeployment” handler. Some of the functions done by the handler are shown below.

// List ReplicaSets owned by this Deployment, while reconciling ControllerRef
        // through adoption/orphaning.
        rsList, err := dc.getReplicaSetsForDeployment(d)
	// List all Pods owned by this Deployment, grouped by their ReplicaSet.
        // Current uses of the podMap are:
        // * check if a Pod is labeled correctly with the pod-template-hash label.
        // * check that no old Pods are running in the middle of Recreate Deployments.
        podMap, err := dc.getPodMapForDeployment(d, rsList)

	// Update deployment conditions with an Unknown condition when pausing/resuming
        // a deployment. In this way, we can be sure that we won't timeout when a user
        // resumes a Deployment with a set progressDeadlineSeconds.
        if err = dc.checkPausedConditions(d); err != nil {
                return err

	// rollback is not re-entrant in case the underlying replica sets are updated with a new
        // revision so we should ensure that we won't proceed to update replica sets until we
        // make sure that the deployment has cleaned up its rollback spec in subsequent enqueues.
        if getRollbackTo(d) != nil {
                return dc.rollback(d, rsList)

        scalingEvent, err := dc.isScalingEvent(d, rsList)
        if err != nil {
                return err
        if scalingEvent {
                return dc.sync(d, rsList)

        switch d.Spec.Strategy.Type {
        case apps.RecreateDeploymentStrategyType:
                return dc.rolloutRecreate(d, rsList, podMap)
        case apps.RollingUpdateDeploymentStrategyType:
                return dc.rolloutRolling(d, rsList)

Sync is responsible for reconciling deployments on scaling events or when they are paused.

func (dc *DeploymentController) sync(d *apps.Deployment, rsList []*apps.ReplicaSet) error {
        newRS, oldRSs, err := dc.getAllReplicaSetsAndSyncRevision(d, rsList, false)
        if err != nil {
                return err
        if err := dc.scale(d, newRS, oldRSs); err != nil {
                // If we get an error while trying to scale, the deployment will be requeued
                // so we can abort this resync
                return err

        // Clean up the deployment when it's paused and no rollback is in flight.
        if d.Spec.Paused && getRollbackTo(d) == nil {
                if err := dc.cleanupDeployment(oldRSs, d); err != nil {
                        return err

        allRSs := append(oldRSs, newRS)
        return dc.syncDeploymentStatus(allRSs, newRS, d)

// scale scales proportionally in order to mitigate risk. Otherwise, scaling up can increase the size
// of the new replica set and scaling down can decrease the sizes of the old ones, both of which would
// have the effect of hastening the rollout progress, which could produce a higher proportion of unavailable
// replicas in the event of a problem with the rolled out a template. Should run only on scaling events or
// when a deployment is paused and not during the normal rollout process.

func (dc *DeploymentController) scale(deployment *apps.Deployment, newRS 
*apps.ReplicaSet, oldRSs []*apps.ReplicaSet) error {

 	// If there is only one active replica set then we should scale that up to the full count of the
        // deployment. If there is no active replica set, then we should scale up the newest replica set.
        if activeOrLatest := deploymentutil.FindActiveOrLatest(newRS, oldRSs); activeOrLatest != nil {

	// If the new replica set is saturated, old replica sets should be fully scaled down.
        // This case handles replica set adoption during a saturated new replica set.
        if deploymentutil.IsSaturated(deployment, newRS) {

 // There are old replica sets with pods, and the new replica set is not saturated. 
        // We need to proportionally scale all replica sets (new and old) in case of a
        // rolling deployment.
        if deploymentutil.IsRollingUpdate(deployment) {

		// Number of additional replicas that can be either added or removed from the total
                // replicas count. These replicas should be distributed proportionally to the active
                // replica sets.
                deploymentReplicasToAdd := allowedSize - allRSsReplicas

                // The additional replicas should be distributed proportionally amongst the active
                // replica sets from the larger to the smaller in size replica set. Scaling direction
                // drives what happens in case we are trying to scale replica sets of the same size.
                // In such a case when scaling up, we should scale up newer replica sets first, and
                // when scaling down, we should scale down older replica sets first.

We hope this article helped you understand the inner workings of Kubernetes deployment controller. If you would like to learn more about Kubernetes and get certified, join our 2-day Kubernetes workshop.

Containers are being embraced at a breakneck speed – developers love them, and they are great for business because they deliver speed and scale in a cost-efficient manner. So much so, that container technology seems to be overtaking VMs – especially with container orchestration tools like Kubernetes, making them simpler to manage and extracting higher efficiency and speed from them.

Kubernetes cluster architecture

Kubernetes provides an open-source platform for simplifying multi-cloud environments. The disparities between different cloud providers are a roadblock for developers and Kubernetes helps by streamlining and standardizing container-based applications.

Kubernetes clusters are the architectural foundation that drives this simplicity and makes it possible for users to get the functionality they need at scale and with ease. Here are some of the functionalities of Kubernetes –

  • Kubernetes distributes workload efficiently across all open resources and reduces traffic spikes or outages.
  • It simplifies application deployment regardless of the size of the cluster
  • It automates horizontal scaling
  • It monitors against app failure with constant node and container health checks and performs self-healing and replication to resolve any failure issues.

All this makes the work of developers faster and frees up their time and attention from trivial repetitive tasks allowing them to build applications better and faster! For the organization, the benefits are three-fold – higher productivity, better products and, finally, cost efficiencies.

Let’s move to the specifics now and find out how to set up a Kubernetes Cluster on the RHEL 7.6 operating system on AWS.

  • You should have a VPC available.
  • A subnet within that VPC, into which you will place your cluster.
  • You should have Security Groups for the Control Plane Load Balancer and the Nodes created.
  • You should have created the Control Plane Load Balancer.
  • A bastion host, or jump box, with a public IP within your VPC from which you can secure shell into your VMs.
  • A pem file for your AWS region, which you will use to secure shell into your VMs.
Creating the IAM Roles

You will need to create 2 IAM roles: one for the Master(s), and one for the worker nodes.

Master Role

To create an IAM role, go to the IAM (Identity and Access Management) page in the AWS console. On the left-hand menu, click ‘Roles’. Then click ‘Create Role’.

Select the service that will use this role. By default, it is EC2, which is what we want. Then click “Next: Permissions”.

Click ‘Create Policy’. The Create Policy page opens in a new tab.

Click on the ‘JSON tab’. Then paste this json into it:

    "Version": "2012-10-17",
    "Statement": [
            "Action": [
            "Resource": "*",
            "Effect": "Allow"

This json defines the permissions that your master nodes will need.

Click ‘Review Policy’. Then give your policy a name and a description.

Click ‘Create Policy’ and your policy is created!

Back on the Create Role page, refresh your policy list, and filter for the policy you just created. Select it and click ‘Next: Tags’.

You should add 2 tags: Name, with a name for your role, and KubernetesCluster, with the name of the cluster that you are going to create. Click ‘Next: Review’.

Give your role a name and a description. Click ‘Create Role’ and your role is created!

Node Role

For the node role, you will follow similar steps, except that you will use the following json:

    "Version": "2012-10-17",
    "Statement": [
            "Action": [
            "Resource": "*",
            "Effect": "Allow"
Provisioning the VMs
Provisioning the Master

We will use RHEL 7.6 for our cluster because RHEL 8.0 uses iptables v1.8, and kube-proxy does not work well with iptables v1.8. However, kube-proxy works with iptables v1.4, which is installed on RHEL 7.6. We will use the x86_64 architecture.

Log into the AWS console. Go to the EC2 home page and click ‘Launch Instance’. We will search under Community AMIs for our image.

Click ‘Select’. Then choose your instance type. T2.medium should suffice for a Kubernetes master. Click ‘Next: Configure Instance Details’.

We will use only 1 instance. For an HA cluster, you will want more. Select your network and your subnet. For the purposes of this tutorial, we will enable auto-assigning a public IP.  In production, you would probably not want your master to have a public IP.  But you would need to make sure that your subnet is configured correctly with the appropriate NAT and route tables. Select the IAM role you created. Then click ‘Next: Add Storage’.

The default, 10 GB of storage, should be adequate for a Kubernetes master. Click ‘Next: Add Tags’.

We will add 3 tags: Name, with the name of your master; KubernetesCluster, with the name of your cluster; and kubernetes.io/cluster/<name of your cluster>, with the value owned. Click ‘Next: Configure Security Group’.

Select “Select an existing security group” and select the security group you created for your Kubernetes nodes. Click ‘Review and Launch’.

Click ‘Launch’. Select “Choose an Existing Key Pair”. Select the key pair from the drop-down. Check the “I acknowledge” box. You should have the private key file saved on the machine from which you plan to secure shell into your master; otherwise you will not be able to ssh into the master! Click ‘Launch Instances’ and your master is created.

Provisioning the Auto Scaling Group

Your worker nodes should be behind an Auto Scaling group. Under Auto Scaling in the left-hand menu of the AWS console, click ‘Auto Scaling Groups’. Click ‘Create Auto Scaling Group’. On the next page, click ‘Get Started’.

Under “Choose AMI”, select RHEL 7.6 x86_64 under Community AMIs, as you did for the master.

When choosing your instance type, be mindful of what applications you want to run on your Kubernetes cluster and their resource needs. Be sure to provision a size with sufficient CPUs and memory.

Under “Configure Details”, give your autoscaling group a name and select the IAM role you configured for your Kubernetes nodes.

When selecting your storage size, be mindful of the storage requirements of your applications that you want to run on Kubernetes. A database application, for example, would need plenty of storage.

Select the security group that you configured for Kubernetes nodes.

Click ‘Create Launch Configuration’. Then select your key pair as you did for the master. Click ‘Create Launch Configuration’ and you are taken to the ‘Configure Auto Scaling Group Details’ page. Give your group a name. Select a group size. For our purpose, 2 nodes will suffice. Select the same subnet on which you placed your master. Click ‘Next: Configure Scaling policies’.

For this tutorial, we will select “Keep this group at its initial size”. For a production cluster with variability in usage, you may want to use scaling policies to adjust the capacity of the group. Click ‘Next: Configure Notifications’.

We will not add any notifications in this tutorial. Click ‘Next: Configure Tags’.

We will add 3 tags: Name, with the name of your nodes; KubernetesCluster, with the name of your cluster; and kubernetes.io/cluster/<your cluster name>, with the value owned. Click ‘Review’.

Click Create Auto Scaling Group and your auto-scaling group is created!

Installing Kubernetes

Specific steps need to be followed to install Kubernetes. Run the following steps as sudo on your master(s) and worker nodes.

 # add docker repo

yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo

# install container-selinux

 yum install -y http://mirror.centos.org/centos/7/extras/x86_64/Packages/container-selinux-2.107-1.el7_6.noarch.rpm

# install docker

yum install docker-ce

# enable docker

systemctl enable --now docker

# create Kubernetes repo. The 2 urls after gpgkey have to be on 1 line.

cat <<EOF > /etc/yum.repos.d/kubernetes.repo
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg

# configure selinux

setenforce 0
sed -i 's/^SELINUX=enforcing$/SELINUX=permissive/' /etc/selinux/config

# install kubelet, kubeadm, kubectl, and Kubernetes-cni. We found that version 1.13.2 works well with RHEL 7.6.

yum install -y kubelet-1.13.2 kubeadm-1.13.2 kubectl-1.13.2 kubernetes-cni-0.6.0-0.x86_64 --disableexcludes=kubernetes –nogpgcheck

# enable kubelet

systemctl enable --now kubelet

# Run the following command as a regular user.

sudo usermod -a -G docker $USER
Creating the Kubernetes Cluster

First, add your master(s) to the control plane load balancer as follows. Log into the AWS console, EC2 service, and on the left-hand menu, under Load Balancing, click ‘Load Balancers’. Select your load balancer and click the Instances tab in the bottom window. Click ‘Edit Instances’.

Select your master(s) and click ‘Save’.

We will create the Kubernetes cluster via a config file. You will need a token, the master’s private DNS name taken from the AWS console, the Load Balancer’s IP, and the Load Balancer’s DNS name. You can generate a Kubernetes token by running the following command on a machine on which you have installed kubeadm:

kubeadm token generate

To get the load balancer’s IP, you must execute a dig command. You install dig by running the following command as sudo:

yum install bind-utils

Then you execute the following command:

dig +short <load balancer dns>

Then you create the following yaml file:

 apiVersion: kubeadm.k8s.io/v1beta1
 kind: InitConfiguration
 - groups:
   - "system:bootstrappers:kubeadm:default-node-token"
   token: "<token>"
   ttl: "0s"
   - signing
   - authentication
   name: "<master private dns>"
     cloud-provider: "aws"
 apiVersion: kubeadm.k8s.io/v1beta1
 kind: ClusterConfiguration
 kubernetesVersion: "v1.13.2"
   timeoutForControlPlane: 10m0s
   - "<Load balancer IPV4>"
     cloud-provider: "aws"
 clusterName: kubernetes
 controlPlaneEndpoint: "<load balancer DNS>:6443"
     cloud-provider: "aws"
     allocate-node-cidrs: "false"
   type: CoreDNS 

You then bootstrap the cluster with the following command as sudo:

kubeadm init --config kubeadm.yaml --ignore-preflight-errors=all

I had a timeout error on the first attempt, but the command ran successfully the second time. Make a note of the output because you will need it to configure the nodes.

You then configure kubectl as follows:

mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config

After this there are some components that need to be installed on Kubernetes on AWS:

# Grant the “admin” user complete access to the cluster

kubectl create clusterrolebinding admin-cluster-binding --clusterrole=cluster-admin --user=admin

# Add-on for networking providers, so pods can communicate. 
# Currently either calico.yaml or weave.yaml

kubectl apply -f https://aws-quickstart.s3.amazonaws.com/quickstart-vmware/scripts/weave.yaml

# Install the Kubernetes dashboard

kubectl apply -f https://aws-quickstart.s3.amazonaws.com/quickstart-vmware/scripts/dashboard.yaml

# Install the default StorageClass

kubectl apply -f https://aws-quickstart.s3.amazonaws.com/quickstart-vmware/scripts/default.storageclass.yaml

# Set up the network policy blocking the AWS metadata endpoint from the default namespace.

kubectl apply -f https://aws-quickstart.s3.amazonaws.com/quickstart-vmware/scripts/network-policy.yaml

Then you have to configure kubelet arguments:

sudo vi /var/lib/kubelet/kubeadm-flags.env

And add the following parameters:

--cloud-provider=aws --hostname-override=<the node name>

After editing the kubeadm-flags.env file:

sudo systemctl restart kubelet

Finally, you have to label your master with the provider ID. That way, any load balancers you create for this node will automatically add the node as an AWS instance:

kubectl patch node <node name> -p '{"spec":{"providerID":"aws:///<availability zone>/<instance ID>"}}'

You can join worker nodes to the cluster by running the following command as sudo, which should have been printed out after running kubeadm init on the master:

kubeadm join <load balancer dns>:6443 --token <token> --discovery-token-ca-cert-hash <discovery token ca cert hash> --ignore-preflight-errors=all

Be sure to configure kubelet arguments on each node and patch them using kubectl as you did for the master.

Your Kubernetes cluster on AWS is now ready!

Kubernetes is the reigning market leader when it comes to container orchestration! Any organization working with the container ecosystem is either already using Kubernetes or considering it. However, despite the undoubted ease and speed Kubernetes bring to the container ecosystem, they also need specialized expertise to deploy and manage.

Many organizations consider the DIY approach to Kubernetes and if you have an in-house IT team with the requisite experience or if your requirements are large enough to justify the cost of hiring a dedicated Kubernetes team – then an internal Kubernetes strategy could certainly be beneficial.

However, if you don’t fall in the category mentioned above, then managed Kubernetes is the smartest and most cost-effective way ahead. With professionals in the picture, you can be assured of getting long term strategy, seamless implementation, and dedicated on-going service, which will

  • reduce deployment time
  • provide 24×7 support
  • handle all upgrades and fixes
  • troubleshoot as and when needed

Kubernetes solution providers offer a wide range of services – from fully managed to bare bone implementation to preconfigured Kubernetes environments on SaaS models to training for your in-house staff.

Look at your operational needs and your budget and explore the market for Kubernetes services options before you pick the service and the digital partner that ticks all your boxes.   

Meanwhile, do look at our tutorial on troubleshooting Kubernetes deployments.

Kubernetes deployments issues are not always easy to troubleshoot. In some cases, the errors can be resolved easily, and, in some cases, detecting errors requires us to dig deeper and run various commands to identify and resolve the issues.

The first step is to list down all pods after installing your application. The following command lists down all pods in all namespaces.

kubectl get pods -A

If you find any issues on the pod status, you can then use kubectl describe, kubectl logs, kubectl exec commands to get more detailed information.

Debugging Pods
Pod Status Shows ImagePullBackOff or ErrImagePull

This status indicates that your pod could not run because the pod could not pull the image from the container registry. To confirm this, run the kubectl describe command along with the pod identifier to display the details.

kubectl describe pod <pod-identifier>

This command will provide more information about the issue.

  • Image name or tag incorrect.
    • Check the image name and tag and try to pull the image manually on the host using docker pull to verify.
  • Authentication failure related to Container registry.
    • Check the secrets, roles, service principal related to your container registry and try to pull the image manually on the host using docker pull to verify.
docker pull <image-name:tag> 
Pod Status Shows Waiting

This status indicates your pod has been scheduled to a worker node, but it can’t run on that machine. To confirm this, run the kubectl describe command along with the pod identifier to display the details.

kubectl describe pod <pod-identifier> -n <namespace>

The most common causes related to this issue are

  • Image name or tag incorrect.
    • Check the image name and tag and try to pull the image manually on the host using docker pull to verify.
  • Authentication failure related to Container registry.
    • Check the secrets, roles, service principal related to your container registry and try to pull the image manually on the host using docker pull to verify.
Pod Status Shows Pending or CrashLoopBackOff

This status indicates your pod could not be scheduled on a node for various reasons like resource constraints (insufficient CPU or memory resources), volume mounting issues.  To confirm this, run the kubectl describe command along with the pod identifier to display the details.

kubectl describe pod <pod-identifier> -n <namespace>

This command will provide more information about the issue. Most common issues are

  • Insufficient resources
    • If resources are insufficient, clean up your existing resources or scaling your nodes (vertically or horizontally) to increase the resources.
  • Volume mounting
    • Check your volume’s mounting definition and storage classes.
  • Using hostPort
    • When you bind a Pod to a hostPort, there are a limited number of places that a pod can be scheduled. In most cases, hostPort is unnecessary, try using a Service object to expose your pod. If you do require hostPort, then you can only schedule as many Pods as there are nodes in your Kubernetes cluster
Pod is crashing or unhealthy

Sometimes the scheduled pods are crashing or unhealthy.  Run kubectl logs to find the root cause.

kubectl logs <pod_identifier> -n <namespace>

If you have multiple containers, run the following command to find the root cause.

kubectl logs <pod_identifier> -c <container_name> -n <namespace>

If your container has previously crashed, you can access the previous container’s crash log with:

kubectl logs –previous <pod_identifier> -c <container_name> -n <namespace>

If your pod is running but with 0/1 ready state or 0/2 ready state (in case if you have multiple containers in your pod), then you need to verify the readiness. Check the health check (readiness probe) in this case.

Most common issues are

  • Application issues
    • Run the below command to check the logs.
               kubectl logs <pod_identifier> -c <container_name> -n <namespace>
  • Run the below command to verify the events.
               kubectl describe <pod_identifier> -n <namespace>
  • Readiness probe health check failed
    • Check the health check (readiness probe) in this case. Also, check the READY column of the kubectl get pods output to find out if the readiness probe is executing positively.
    • Run the below command to check the logs.
         kubectl logs <pod_identifier> -c <container_name> -n <namespace>
  • Run the below command to verify the events.
         kubectl describe <pod_identifier> -n <namespace>
  • Liveness probe health check failed
    • Check the health check (liveness probe) in this case. Also, check the RESTARTS column of the kubectl get pods output. To find out if the liveness probe is executing positively.
    • Run the below command to check the logs.
         kubectl logs <pod_identifier> -c <container_name> -n <namespace>
  • Run the below command to verify the events.
         kubectl describe <pod_identifier> -n <namespace>
Pod is running but has application issues

In some cases, the pods are running, but the output of the application is incorrect. In this case, you should run the following to find the root cause.

  • Run the below command and identify the issue.
kubectl logs <pod_identifier> -c <container_name> -n <namespace>
  • If you are interested in the last n lines of logs run
kubectl logs <pod_identifier> -c <container_name> --tail <n-lines> -n <namespace>
  • Run the commands inside the container using
kubectl exec <pod_identifier> -c <container_name> /bin/bash -n <namespace>

Run the commands like ‘curl’ or ‘ps’ ‘ls’ to troubleshoot the issue after you get into the container.

Pod is running and working but cannot access through services

In some cases, the pods are working as expected but cannot access through the services. Most common causes of this issue are

  • Service not registered properly
    • Check that the service exists and describe the service and validate the pod selectors to run the following commands.
kubectl get svc
kubectl describe svc <svc-name>
kubectl get endpoints
  • Run the following commands to verify pod selector
kubectl get pods --selector=name={name},{label-name}={label-value}
  • The service may be deployed in a different namespace.
    • Verify that the pod’s containerPort matches up with the Service’s containerPort
  • Service is registered properly but has a DNS issue
    • Get into the container using exec command and run nslookup using the following command
kubectl get endpoints
kubectl exec <pod_identifier> -c <container_name> /bin/bash
nslookup <service-name>
  • If you have any issues to run the command for curl or nslookup. Deploy debugging pod using image yauritux/busybox-curl in the same namespace to verify. Please run the following commands to verify
kubectl run --generator=run-pod/v1 -it --rm <name> --image=yauritux/busybox-curl -n <namespace>
  • Run the following to verify within the container
curl http://<servicename>
telnet <service-ip> <service-port>
nslookup <servicename>

On July 21, 2015, when Kubernetes v1.0 was released, it redefined the container technology landscape. All the bottlenecks of application deployment, scaling, and management in containers was made simpler and faster with intelligent automation.

Container technology made software development more agile and brought in resource efficiency – they made scaling smoother and faster. However, they also need to be tracked, monitored, and managed, which is where container orchestration and Kubernetes come in.

Kubernetes is an open-source container-orchestration system for automating application deployment, scaling, and management.  It was originally designed by Google and is now maintained by the Cloud Native Computing Foundation.

What does Kubernetes do?

Kubernetes allows you to leverage the full potential of your container ecosystem. With automation, it streamlines container workflow and frees up the IT team to concentrate on their core areas of application development by removing the need to manage container networking, storage, logs, alerting, etc. Overall, it automates deploying, scaling, and managing of containerized applications on a cluster of servers.

Key Benefits of Kubernetes

Flexibility for scaling – it enables horizontal infrastructure scaling by quickly adding or removing new severs. Kubernetes has the option of automating vertical scaling, too, by taking into account application provided metrics.

Health check and self-healing designed in Kubernetes allow it to maintain high availability of applications and infrastructure.

Enhanced deployment speed – with automated rollouts and rollbacks, canary deployments, and wide-ranging support for a variety of programming languages, Kubernetes speeds up the process of building, testing, and deploying new software.

Let’s understand more about Kubernetes concepts

1. Kubernetes Objects

Kubernetes contains several abstractions that represent the state of our system: deployed containerized applications and workloads, their associated network and disk resources, and other information about what our Kubernetes cluster is doing. These abstractions are represented by objects in the Kubernetes API.  The basic Kubernetes objects include:

  • Pod
  • Service
  • Volume
  • Namespace

In this blog, we will look at the Pod and Service objects.

2. Pods

A pod is a higher level of abstraction grouping containerized components.  A pod consists of one or more containers that are guaranteed to be co-located on the host machine and can share resources.  The basic scheduling unit in Kubernetes is a pod.  The host machines on which the pods are scheduled are called Nodes.

3. Pod definition yaml

Kubernetes objects are mostly created by declaring their configuration in a yaml file.
Given below is yaml file to define a simple pod.

apiVersion: v1 
kind: Pod
name: nginx 
name: nginx 
- name: nginx 
image: nginx 
- containerPort: 8080

In the above yaml file,

  1. apiVersion – denotes which version of the Kubernetes API we are using to create this object.
  2. kind – specifies what kind of object we want to create.  For Pod object, the apviVersion is always v1.
  3. metadata – has data to uniquely identify the object (name) and labels.
  4. Labels are key/value pairs that are attached to objects.  Labels are intended to be used to specify identifying attributes of objects that are meaningful and relevant to users but do not directly imply semantics to the core system.  So, in the above example, instead of “name: nginx” we can have “appname: nginx”, “name: mynginxapp” or anything we like.
  5. Spec – defines the object specification and differs for each object type.  For Pod object, the spec has an array of containers since a pod consists of one or more containers.
  6. For each container, we provide below attributes:
  • name – name of the container.  This can be different from name of pod and is not related to it.
  • Image – name of the docker image to be used to build this container
  • ports – the ports in this container to be exposed outside the pod.  Here we are running the nginx web-server on port 8080 and exposing it.

Suppose we have the above pod definition in a file named pod-definition.yaml, the pod is created by executing the below Kubernetes command:

$ kubectl create -f pod-definition.yaml

4. Pod communication and need for services

Each pod in Kubernetes is assigned a unique Pod IP address within the cluster, which allows applications to use ports without the risk of conflict. 

Within the pod, all containers can reference each other on localhost, but a container within one pod has no way of directly addressing another container within another pod; for that, it must use the Pod IP Address.

An application developer should never use the Pod IP Address though, to reference / invoke a capability in another pod, as Pod IP addresses are ephemeral – the specific pod that they are referencing may be assigned to another Pod IP address on restart.  Instead, we should use a reference to a Service, which holds a reference to the target pod at the specific Pod IP Address.

5. Services

In Kubernetes, a Service is an abstraction that defines a logical set of Pods and a policy by which to access them. The set of Pods targeted by a Service is usually determined by a selector.

Sample YAML for a service to expose the pod(s) which we created earlier is given below:

apiVersion: v1
 kind: Service
   name: my-service
     name: nginx
     - protocol: TCP
       port: 80
       targetPort: 8080
       nodePort: 30230
   type: NodePort 

This yaml file defines a service named my-service which is used to access the Pods which have a label ‘name: nginx’. 

  1. The selector field of the service must match the label field of the Pods to which we want to connect.
  2. There are 3 ports defined in above YAML file:
  • Port is the port number that makes a service visible to other services running within the same Kubernetes cluster.
  • Target Port is the port on the POD where the service is running.  This is an optional field; if not provided, Kubernetes assigns the same value as Port field
  • Node port is the port on which the service can be accessed by external users.  NodePort can only have values from 30000 to 32767.  If this optional field is not provided in the definition, Kubernetes automatically assigns a value for NodePort service.

To create the service object, enter the above yaml code in a file named service-defn.yaml and execute the command given below:

$ kubectl create -f service-defn.yaml

6. Types of Services

In the above example, we have type: Nodeport for the service.  The different values allowed for the type field are:

  1. ClusterIP: Exposes the Service on a cluster-internal IP. Choosing this value makes the Service only reachable from within the cluster. This is the default Service type.
  2. NodePort: Exposes the Service on each Node’s IP at a static port (the NodePort).   A ClusterIP Service, to which the NodePort Service routes, is automatically created.  We will be able to contact the NodePort Service, from outside the cluster, by requesting <NodeIP>:<NodePort>
  3. LoadBalancer: Exposes the Service externally using a cloud provider’s load balancer. NodePort and ClusterIP Services, to which the external load balancer routes, are automatically created.
  4. ExternalName: Maps the Service to the contents of the externalName field (e.g., foo.bar.example.com), by returning a CNAME record with its value.  No proxying of any kind is set up.  We need CoreDNS version 1.7 or higher to use the ExternalName type.

Many of you are running your mission-critical applications on containers, and if you haven’t already deployed Kubernetes to manage your container ecosystem, then chances are you soon will.

If you are considering a Kubernetes implementation, then there are several ways to go about it –

  • In-house Kubernetes deployment – if you have a large enough IT team with the requisite expertise in Kubernetes architecture and deployment, then getting your Kubernetes cluster up and running in-house is certainly a possibility. Kubernetes deployment is a complex process and requires a mix of specific skill sets. Also running and monitoring a Kubernetes platform requires the full-time services of a dedicated team, and your requirement must justify this additional cost.
  • SaaS Solutions for Kubernetes– if your business needs are specific and straightforward, then you can explore the market for pre-designed Kubernetes offerings on a SaaS payment model.
  • Fully outsourced (managed) Kubernetes services – if budget permits and your business demands, then bringing in the professionals is a safe and hassle-free solution. From infrastructure assessments to building a Kubernetes strategy to engineering, deploying, and managing enterprise-wide Kubernetes solutions – you can outsource your entire project to experts.
  • Many service providers like CloudIQ also offer day-to-day management and support as well as Kubernetes training to your IT staff to set up internal management expertise.

If the last decade of cloud has taught us anything, it is that when it comes to technology, bringing in professionals to do the job always turns out to be the best option in the long run. Kubernetes is a sophisticated platform that requires specialized competencies. Here is a look at one of our tutorials on Kubernetes Networking – how it all works under the hood.


In Kubernetes, applications run as a set of pods with their own IP address and port. Kubernetes provides an abstract way to expose the applications/pods as a network service. Various forms of the service abstractions include ClusterIP, NodePort, Load Balancer & Ingress. When service requests enters Kubernetes cluster, the service abstractions have to be directed to individual service endpoints of Pods. This data plane function is implemented using a Linux  Kernel feature – iptables.

Iptables is used to set up, maintain, and inspect the tables of IP packet filter rules in the Linux kernel. Several different tables may be defined. Each table contains a number of built-in chains and may also contain user-defined chains. Each chain is a list of rules which can match a set of packets. Each rule specifies what to do with a packet that matches. This is called a ‘target’, which may be a jump (-j) to a user-defined chain in the same table.

The service(SVC) to service endpoints(SEP) are programmed using KUBE-SERVICES user-defined chains in the NAT(Network Address Translation) table. The contents of the iptables can be extracted using “iptables-save” command

# Generated by iptables-save v1.6.0 on Mon Sep 16 08:00:17 2019
:OUTPUT ACCEPT [23:1438]
:DOCKER - [0:0]
:IP-MASQ-AGENT - [0:0]


-A PREROUTING -m comment --comment "kubernetes service portals" -j KUBE-SERVICES
-A OUTPUT -m comment --comment "kubernetes service portals" -j KUBE-SERVICES

Let’s consider the Services in the following example.

[email protected]:~$ kubectl get svc –namespace=workshop-development


Here we have the following service abstractions that are defined.




The above services have to be translated to individual service endpoints. The rules performing matching and translation are programmed using custom chains in the NAT table of Ip Tables as below.

Lets look for LoadBalancer= service

[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep
-A KUBE-SERVICES -d -p tcp -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:http loadbalancer IP" -m tcp --dport 80 -j KUBE-FW-SXB4UOYSLPHVISJM
-A KUBE-SERVICES -d -p tcp -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -m tcp --dport 443 -j KUBE-FW-JLRSZDR3OXJ4SUA2

Let’s look at the HTTPS service available on port 443.

[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep KUBE-FW-JLRSZDR3OXJ4SUA2

[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep KUBE-FW-JLRSZDR3OXJ4SUA2

-A KUBE-FW-JLRSZDR3OXJ4SUA2 -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -j KUBE-MARK-MASQ
-A KUBE-FW-JLRSZDR3OXJ4SUA2 -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -j KUBE-SVC-JLRSZDR3OXJ4SUA2
-A KUBE-FW-JLRSZDR3OXJ4SUA2 -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -j KUBE-MARK-DROP
-A KUBE-SERVICES -d -p tcp -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -m tcp --dport 443 -j KUBE-FW-JLRSZDR3OXJ4SUA2

We see below NodePort & Cluster IP translation. The service chains SVC point to two different service endpoints. In order to select between the two service endpoints, a random probability measure is calculated, and appropriate SEP service endpoints are selected.

[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep KUBE-SVC-JLRSZDR3OXJ4SUA2

-A KUBE-FW-JLRSZDR3OXJ4SUA2 -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https loadbalancer IP" -j KUBE-SVC-JLRSZDR3OXJ4SUA2
-A KUBE-NODEPORTS -p tcp -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https" -m tcp --dport 31512 -j KUBE-SVC-JLRSZDR3OXJ4SUA2
-A KUBE-SERVICES -d -p tcp -m comment --comment "workshop-development/ciq-ingress-workshop-development-nginx-ingress-controller:https cluster IP" -m tcp --dport 443 -j KUBE-SVC-JLRSZDR3OXJ4SUA2
-A KUBE-SVC-JLRSZDR3OXJ4SUA2 -m statistic --mode random --probability 0.50000000000 -j KUBE-SEP-4R3FOXQSM5T2ZADC

In the SEP service endpoints, the actual DNAT is performed.

[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep KUBE-SEP-4R3FOXQSM5T2ZADC
-A KUBE-SEP-4R3FOXQSM5T2ZADC -p tcp -m tcp -j DNAT --to-destination
-A KUBE-SVC-JLRSZDR3OXJ4SUA2 -m statistic --mode random --probability 0.50000000000 -j KUBE-SEP-4R3FOXQSM5T2ZADC
[email protected]:~$ cat ciq-dev-aks-iptables-save.output | grep KUBE-SEP-PI7R3ONIYH4XJLMW
-A KUBE-SEP-PI7R3ONIYH4XJLMW -p tcp -m tcp -j DNAT --to-destination

In this article we will discuss how to create security groups in AWS for Kubernetes. The goal is to set up a Kubernetes cluster on AWS EC2, having provisioned your virtual machines. You are going to need two security groups: one for the control plane load balancer, and another for the VMs.

Creating a Security Group through the AWS Console

Prerequisite: You should have a VPC (virtual private cloud) set up.

Log into the AWS EC2 (or VPC) console. On the left-hand menu, under Network and Security, click Security Groups.

Click on Create Security Group.

Enter a Name and a Description for your Security Group. Then select your VPC from the drop-down menu. Click Add Rule.

You will need 2 TCP ingress rules, one over port 6443, another over port 443. We are choosing to allow the Source from anywhere. In production you may want to restrict the CIDR, IP, or security group that can reach this load balancer.

We are choosing to leave the outbound rules as default, in which all outbound traffic is permitted.

Click Create and your security group is created!

Select your security group in the console. You may want to give your security group a Name (in addition to the Group Name that you specified when creating it).

But you are not done yet: you must add tags to your security group. These tags will alert AWS that this security group is to be used for Kubernetes. Click on the Tags tab at the bottom of the window. Then click Add/Edit Tags.

You will need 2 tags:
  • Name: KubernetesCluster. Value: <the name of your Kubernetes cluster>
  • Name: kubernetes.io/cluster/<the name of your Kubernetes cluster>. Value: owned

Click Save and your tags are saved!

Creating a Security Group for the Virtual Machines

Follow the steps above to create a security group for your virtual machines. Here are the ports that you will need to open for your control plane VMs:

The master node:
  1. 22 for SSH from your bastion host
  2. 6443 for the Kubernetes API Server
  3. 2379-2380 for the ETCD server
  4. 10250 for the Kubelet health check
  5. 10252 for the Kube controller manager
  6. 10255 for the read only kubelet API
The worker nodes:
  1. 22 for SSH
  2. 10250 for the kubelet health check
  3. 30000-32767 for external applications. However, it is more likely that you will expose external applications to outside the cluster via load balancers, and restrict access to these ports to within your vpc.
  4. 10255 for the read only kubelet API

We have chosen to combine the master and the worker rules into one security group for convenience. You may want to separate them into 2 security groups for extra security.

Follow the step-by-step instructions detailed above and you will have successfully created AWS Security Groups for Kubernetes.

Helm is a package manager for Kubernetes that allows developers and operators to more easily package, configure, and deploy applications and services onto Kubernetes clusters.

What Does Kubernetes Helm Solve?

Kubernetes is known as a complex platform to understand and use. Kubernetes Helm helps make Kubernetes easier and faster to use:

Increased productivity – developers can deploy a pre-tested app via a Helm chart and focus on developing their applications, instead of spending time on deploying test environments to test their Kubernetes clusters

Existing Helm Charts – allow developers to get a working database, big data platform, CMS, etc. deployed for their application with one click. Developers can modify existing charts or create their own to automate dev, test or production processes.

Easier to start with Kubernetes – it can be difficult to get started with Kubernetes and learn how to deploy production-grade applications. Helm provides one click deployment of apps, making it much easier to get started and deploy your first app, even if you don’t have extensive container experience.

Decreased complexity – deployment of Kubernetes-orchestrated apps can be extremely complex. Using incorrect values in configuration files or failing to roll out apps correctly from YAML templates can break deployments. Helm Charts allow the community to preconfigure applications, defining values that are fixed and others that are configurable with sensible defaults, providing a consistent interface for changing configuration. This dramatically reduces complexity, and eliminates deployment errors by locking out incorrect configurations.

Production ready – running Kubernetes in production with all its components (pods, namespaces, deployments, etc.) is difficult and prone to error. With a tested, stable Helm chart, users can deploy to production with confidence, and reduce the complexity of maintaining a Kubernetes App Catalog.

No duplication of effort – once a developer has created a chart, tested and stabilized it once, it can be reused across multiple groups in an organization and outside it. Previously, it was much more difficult (but not impossible) to share Kubernetes applications and replicate them between environments.

Helm provides this functionality through the following components:

  • A command line tool, helm, which provides the user interface to all Helm functionality.
  • A companion server component, tiller, that runs on your Kubernetes cluster, listens for commands from helm, and handles the configuration and deployment of software releases on the cluster.
  • The Helm packaging format, called charts.
  • An official curated charts repository with prepackaged charts for popular open-source software projects.
Installing Helm

There are two parts to Helm: The Helm client (helm) and the Helm server (Tiller).


The Helm client can be installed either from source, or from pre-built binary releases.

From the Binary Releases

Every release of Helm provides binary releases for a variety of OSes. These binary versions can be manually downloaded and installed.

Download your desired version

Unpack it (tar -zxvf helm-v2.0.0-linux-amd64.tgz)

Find the helm binary in the unpacked directory, and move it to its desired destination (mv linux-amd64/helm /usr/local/bin/helm)


Tiller, the server portion of Helm, typically runs inside of your Kubernetes cluster. But for development, it can also be run locally, and configured to talk to a remote Kubernetes cluster.

The easiest way to install tiller into the cluster is simply to run helm init. This will validate that helm’s local environment is set up correctly (and set it up if necessary). Then it will connect to whatever cluster kubectl connects to by default (kubectl config view). Once it connects, it will install tiller into the kube-system namespace.

After helm init, you should be able to run kubectl get pods –namespace kube-system and see Tiller running.


A Chart is a Helm package. It contains all of the resource definitions necessary to run an application, tool, or service inside of a Kubernetes cluster.

A Repository is the place where charts can be collected and shared.

A Release is an instance of a chart running in a Kubernetes cluster. One chart can often be installed many times into the same cluster. And each time it is installed, a new release is created.

The “helm install” command can install from several sources:

  • A chart repository
  • A local chart archive (helm install foo-0.1.1.tgz)
  • An unpacked chart directory (helm install path/to/foo)
  • A full URL (helm install https://example.com/charts/foo-1.2.3.tgz)

Helm uses a packaging format called charts. A chart is a collection of files that describe a related set of Kubernetes resources.


A chart is organized as a collection of files inside of a directory. The directory name is the name of the chart (without versioning information). Thus, a chart describing WordPress would be stored in the wordpress/ directory.

Inside of this directory, Helm will expect a structure that matches this:

     Chart.yaml          # A YAML file containing information about the chart
     LICENSE             # OPTIONAL: A plain text file containing the license for the chart
     README.md           # OPTIONAL: A human-readable README file
     requirements.yaml   # OPTIONAL: A YAML file listing dependencies for the chart
     values.yaml         # The default configuration values for this chart
     charts/             # A directory containing any charts upon which this chart depends.
     templates/          # A directory of templates that, when combined with values,
                         # will generate valid Kubernetes manifest files.
     templates/NOTES.txt # OPTIONAL: A plain text file containing short usage notes

Lets Build and publish a simple http service and say “Hello world”.

Package and publish via Helm.

     Docker: Build and publish “Hello World”

Hello world!

     rawapp-index.html hosted withby GitHub
        FROM busybox
        ADD app/index.html /www/index.html
        EXPOSE 8005
        CMD httpd -p 8005 -h /www; tail -f /dev/null
        Dockerfile hosted with by GitHub
        docker build -t hello-world .
        docker run -p 80:8005 hello-world
        ## open your browser and check http://localhost/
        docker login
        docker tag hello-world {your_dockerhub_user}/hello-world
        docker push {your_dockerhub_user}/hello-world:latest

Helm: build and install

We need helm chart files, just do:

     helm create helloworld-chart
          repository: {your_dockerhub_user}/hello-world
          tag: latest
          pullPolicy: IfNotPresent
          name: hello-world
          type: LoadBalancer
          externalPort: 80
          internalPort: 8005

Now, we need to package this helm chart

     helm package helloworld-chart --debug
        ## helloworld-chart-0.1.0.tgz file was created
        helm install helloworld-chart-0.1.0.tgz --name helloworld
        kubectl get svc --watch # wait for a IP

A chart repository is an HTTP server that houses one or more packaged charts. Any HTTP server that can serve YAML files and tar files and can answer GET requests can be used as a repository server.

CloudIQ is a leading Cloud Consulting and Solutions firm that helps businesses solve today’s problems and plan the enterprise of tomorrow by integrating intelligent cloud solutions. We help you leverage the technologies that make your people more productive, your infrastructure more intelligent, and your business more profitable. 


3520 NE Harrison Drive, Issaquah, WA, 98029


Chennai One IT SEZ,

Module No:5-C, Phase ll, 2nd Floor, North Block, Pallavaram-Thoraipakkam 200 ft road, Thoraipakkam, Chennai – 600097

© 2022 CloudIQ Technologies. All rights reserved.

Get in touch

Please contact us using the form below


    3520 NE Harrison Drive, Issaquah, WA, 98029

    +1 (206) 203-4151


    Chennai One IT SEZ,

    Module No:5-C, Phase ll, 2nd Floor, North Block, Pallavaram-Thoraipakkam 200 ft road, Thoraipakkam, Chennai – 600097