Chapter 4 Using Kubernetes

This chapter describes how to get started using Kubernetes to deploy, maintain and scale your containerized applications. In this chapter, we describe basic usage of the kubectl command to get you started creating and managing containers and services within your environment.

The kubectl utility is fully documented in the upstream documentation at:

https://kubernetes.io/docs/reference/kubectl/overview/

4.1 About Runtime Engines

runc is the default runtime engine when you create containers. You can also use the kata-runtime runtime engine to create Kata containers. For information on Kata containers and how to create them, see Container Runtimes.

4.2 Getting Information about Nodes

To get a listing of all of the nodes in a cluster and the status of each node, use the kubectl get command. This command can be used to obtain listings of any kind of resource that Kubernetes supports. In this case, the nodes resource:

kubectl get nodes
NAME STATUS ROLES AGE VERSION control.example.com Ready master 1h v1.18.x+x.x.x.el7 worker1.example.com Ready <none> 1h v1.18.x+x.x.x.el7 worker2.example.com Ready <none> 1h v1.18.x+x.x.x.el7

You can get more detailed information about any resource using the kubectl describe command. If you specify the name of the resource, the output is limited to information about that resource alone; otherwise, full details of all resources are also printed to screen. For example:

kubectl describe nodes worker1.example.com
Name: worker1.example.com1 Roles: <none> Labels: beta.kubernetes.io/arch=amd64 beta.kubernetes.io/os=linux kubernetes.io/arch=amd64 kubernetes.io/hostname=worker1.example.com kubernetes.io/os=linux Annotations: flannel.alpha.coreos.com/backend-data: {"VtepMAC":"fe:78:5f:ea:7c:c0"} flannel.alpha.coreos.com/backend-type: vxlan flannel.alpha.coreos.com/kube-subnet-manager: true flannel.alpha.coreos.com/public-ip: 192.0.2.11 kubeadm.alpha.kubernetes.io/cri-socket: /var/run/crio/crio.sock node.alpha.kubernetes.io/ttl: 0 volumes.kubernetes.io/controller-managed-attach-detach: true ...

4.3 Running an Application in a Pod

To create a pod with a single running container, you can use the kubectl create command. For example:

kubectl create deployment --image nginx hello-world
deployment.apps/hello-world created

Substitute nginx with a container image. Substitute hello-world with a name for your deployment. Your pods are named by using the deployment name as a prefix.

Tip

Deployment, pod and service names conform to a requirement to match a DNS-1123 label. These must consist of lower case alphanumeric characters or -, and must start and end with an alphanumeric character. The regular expression that is used to validate names is '[a-z0-9]([-a-z0-9]*[a-z0-9])?'. If you use a name for your deployment that does not validate, an error is returned.

There are many additional optional parameters that can be used when you run a new application within Kubernetes. For instance, at run time, you can specify how many replica pods should be started, or you might apply a label to the deployment to make it easier to identify pod components. To see a full list of options available to you, run kubectl run --help.

To check that your new application deployment has created one or more pods, use the kubectl get pods command:

kubectl get pods
NAME READY STATUS RESTARTS AGE hello-world-5f55779987-wd857 1/1 Running 0 1m

Use kubectl describe to show a more detailed view of your pods, including which containers are running and what image they are based on, as well as which node is currently hosting the pod:

kubectl describe pods
Name: hello-world-5f55779987-wd857 Namespace: default Priority: 0 PriorityClassName: <none> Node: worker1.example.com/192.0.2.11 Start Time: Fri, 16 Aug 2019 08:48:33 +0100 Labels: app=hello-world pod-template-hash=5f55779987 Annotations: <none> Status: Running IP: 10.244.1.3 Controlled By: ReplicaSet/hello-world-5f55779987 Containers: nginx: Container ID: cri-o://417b4b59f7005eb4b1754a1627e01f957e931c0cf24f1780cd94fa9949be1d31 Image: nginx Image ID: docker-pullable://nginx@sha256:5d32f60db294b5deb55d078cd4feb410ad88e6fe7... Port: <none> Host Port: <none> State: Running Started: Mon, 10 Dec 2018 08:25:25 -0800 Ready: True Restart Count: 0 Environment: <none> Mounts: /var/run/secrets/kubernetes.io/serviceaccount from default-token-s8wj4 (ro) Conditions: Type Status Initialized True Ready True ContainersReady True PodScheduled True Volumes: default-token-s8wj4: Type: Secret (a volume populated by a Secret) SecretName: default-token-s8wj4 Optional: false QoS Class: BestEffort Node-Selectors: <none> Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s node.kubernetes.io/unreachable:NoExecute for 300s Events: ....

4.4 Scaling a Pod Deployment

To change the number of instances of the same pod that you are running, you can use the kubectl scale deployment command. For example:

kubectl scale deployment --replicas=3 hello-world
deployment.apps/hello-world scaled

You can check that the number of pod instances has been scaled appropriately:

kubectl get pods
NAME READY STATUS RESTARTS AGE hello-world-5f55779987-tswmg 1/1 Running 0 18s hello-world-5f55779987-v8w5h 1/1 Running 0 26m hello-world-5f55779987-wd857 1/1 Running 0 18s

4.5 Exposing a Service Object for an Application

Typically, while many applications may only need to communicate internally within a pod, or even across pods, you may need to expose your application externally so that clients outside of the Kubernetes cluster can interface with the application. You can do this by creating a service definition for the deployment.

To expose a deployment using a service object, you must define the service type that should be used. If you are not using a cloud-based load balancing service, you can set the service type to NodePort. The NodePort service exposes the application running within the cluster on a dedicated port on the public IP address on all of the nodes within the cluster. Use the kubectl expose deployment to create a new service. For example:

kubectl expose deployment hello-world --port 80 --type=LoadBalancer
service/hello-world exposed

Use kubectl get services to list the different services that the cluster is running, and to obtain the port information required to access the service:

kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE hello-world LoadBalancer 10.102.42.160 <pending> 80:31847/TCP 3s kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 5h13m

In this example output, you can see that traffic to port 80 inside the cluster is mapped to the NodePort 31847. The external IP that can be used to access the service is listed as <pending>, meaning that if you connect to the external IP address for any of the nodes within the cluster on the port 31847, you are able access the service.

For the sake of the example in this guide, you can open a web browser to point at any of the nodes in the cluster, such as http://worker1.example.com:31847/, and it should display the NGINX demonstration application.

4.6 Deleting a Service or Deployment

Objects can be deleted easily within Kubernetes so that your environment can be cleaned. Use the kubectl delete command to remove an object.

To delete a service, specify the services object and the name of the service that you want to remove. For example:

kubectl delete services hello-world
service "hello-world" deleted

To delete an entire deployment, and all of the pod replicas running for that deployment, specify the deployment object and the name that you used to create the deployment:

kubectl delete deployment hello-world
deployment.extensions "hello-world" deleted

4.7 Working With Namespaces

Namespaces can be used to further separate resource usage and to provide limited environments for particular use cases. By default, Kubernetes configures a namespace for Kubernetes system components and a standard namespace to be used for all other deployments for which no namespace is defined.

To view existing namespaces, use the kubectl get namespaces and kubectl describe namespaces commands.

The kubectl command only displays resources in the default namespace, unless you set the namespace specifically for a request. Therefore, if you need to view the pods specific to the Kubernetes system, you would use the --namespace option to set the namespace to kube-system for the request. For example, in a cluster with a single control plane node:

kubectl get pods --namespace=kube-system
NAME READY STATUS RESTARTS AGE coredns-5bc65d7f4b-qzfcc 1/1 Running 0 23h coredns-5bc65d7f4b-z64f2 1/1 Running 0 23h etcd-control1.example.com 1/1 Running 0 23h kube-apiserver-control1.example.com 1/1 Running 0 23h kube-controller-control1.example.com 1/1 Running 0 23h kube-flannel-ds-2sjbx 1/1 Running 0 23h kube-flannel-ds-njg9r 1/1 Running 0 23h kube-proxy-m2rt2 1/1 Running 0 23h kube-proxy-tbkxd 1/1 Running 0 23h kube-scheduler-control1.example.com 1/1 Running 0 23h kubernetes-dashboard-7646bf6898-d6x2m 1/1 Running 0 23h

4.8 Using Deployment Files

To simplify the creation of pods and their related requirements, you can create a deployment file that define all of the elements that comprise the deployment. This deployment defines which images should be used to generate the containers within the pod, along with any runtime requirements, as well as Kubernetes networking and storage requirements in the form of services that should be configured and volumes that may need to be mounted.

Deployments are described in detail at:

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