Kubectl

Kubernetes consists of clusters, nodes and pods.

  • Cluster: A group of nodes, which have the same location and accessed via the same Cloud Provider’s API

  • Nodes: A worker, mostly a separate machine, consisting of various pods

  • Pod: Smallest unit in Kubernetes, defines a service, consists of one or more containers

A namespace bundles multiple pods, which may also be located on different nodes together.

Config

Get current config (from ~/.kube/config)

$ kubectl config view

Cluster

Get info about the cluster:

$ kubectl cluster-info

Node

Get all nodes in current cluster

$ kubectl get nodes

Describe node (shows included namespaces)

$ kubectl describe node <NODE>

Namespace

A namespace is a space which separates resources and workloads.

Get all namespaces in current cluster:

$ kubectl get namespaces

Note

Predefined namespaces:

  • default: The default namespace for objects with no other namespace

  • kube-node-lease: This namespace holds Lease objects associated with each node. Node leases allow the kubelet to send heartbeats so that the control plane can detect node failure.

  • kube-public: This namespace is created automatically and is readable by all users (including those not authenticated). This namespace is mostly reserved for cluster usage, in case that some resources should be visible and readable publicly throughout the whole cluster. The public aspect of this namespace is only a convention, not a requirement.

  • kube-system: The namespace for objects created by the Kubernetes system.

Describe a namespace:

$ kubectl describe namespace <NAMESPACE>

Create a namespace:

$ kubectl apply -f <path/to/namespace.yaml>

Hint

An example namespace.yaml for a development namespace:

apiVersion: v1
kind: Namespace
metadata:
  name: development

Delete a namespace (all namespaces defined in namespace.yaml:

$ kubectl delete -f <path/to/namespace.yaml>

Deployment

A deployment is a resource which manages an application on the cluster. It handles the creation of the defined pods (and possible replicas) and ensures the application runs in the desired state.

Create

Deploy an application from a single Docker image (uses single node)

$ kubectl create deployment <DEPLOYMENT_NAME> --image=<IMAGE_URL:TAG>

Describe

Get detailed information about a deployment

$ kubectl describe deployment <DEPLOYMENT_NAME>

Scaling

A deployment may have multiple replica pods running the same container to ensure the constant availability of the application or to distribute incoming traffic to multiple nodes.

Scale deployment to 4 replicas

$ kubectl scale deployments/<DEPLOYMENT_NAME> --replicas=4

Update image version

Update the used Docker image version for a deployment, initiating a rolling update: new pods using new images are created. When those running, the olds ones are terminated

$ kubectl set image deployments/<DEPLOYMENT_NAME> <IMAGE_NAME:TAG>

Check rollout status of deployment (e.g. after updating image version)

$ kubectl rollout status deployments/<DEPLOYMENT_NAME>

Revert latest rollout of deployment (e.g. a failed image version update)

$ kubectl rollout undo deployments/<DEPLOYMENT_NAME>

Services

A service is a load balancer and transfers traffic to pods. It exposes an application deployment to external traffic providing a port.

Create a new service exposing a deployment under port 8001

$ kubectl expose deployments/<DEPLOYMENT_NAME> --type="NodePort" --port 8001

Get all services within a namespace

$ kubectl get services -n <NAMESPACE>

Describe a service

Show detailed info on all services within the namespace

$ kubectl describe services -n <NAMESPACE>

Show detailed info on specific services within the namespace

$ kubectl describe service <SERVICE_NAME> -n <NAMESPACE>

Delete a service

Delete a single service (this does not affect the application deployment which uses the service)

$ kubectl delete service <SERVICE_NAME> -n <NAMESPACE>

Pod

A pod is a abstraction which represents one or more application containers.

Get names all pods of a namespace:

$ kubectl get pods -n <NAMESPACE>

Get names of all pods in all namespaces

$ kubectl get pods -A

Get wide output format of all pods of a namespace (includes cluster-internal IP address)

$ kubectl get pods -o wide

Get names of all pods using a specific label name

$ kubectl get pods -l <LABEL_NAME>

Describe a pod

Shows the containers, running inside the pod among other information.

Describe single pod within a namespace:

$ kubectl describe pod <POD> -n <NAMESPACE>

Describe all pods within a namespace:

$ kubectl describe pods -n <NAMESPACE>

Get logs

Print the latest logs of a running container inside a log (if pod only runs a single container, -c <CONTAINER> must not be specified):

$ kubectl logs -n <NAMESPACE> <POD> -c <CONTAINER>

Labels

Labels are used to tag one or multiple pods

Add another label name to a pod

$ kubectl label pod <POD_NAME> <LABEL_NAME=VALUE>

Delete a label from a pod

$ kubectl label pod <POD_NAME> <LABEL_NAME>-

ReplicaSet

A ReplicaSet is a resource, which ensures that a specific number of pod replicas are running at any given time. Commonly deployments are used to manage ReplicaSets, so it is not recommended to use them directly.

Get replica sets

Get all ReplicaSets within the namespace which manage pods containing a certain label

$ kubectl get replicaset -l <LABEL_NAME> -n <NAMESPACE>