Skip to content

Kubernetes a.k.a. K8s Concepts

What is Kubernetes?

  • K&s is the leading container orchestration tool
  • Designed as a loosely coupled collection of components centered around deploying, maintaining and scaling workloads
  • Vendor-neutral : Runs on all cloud providers
  • Backed by a huge community

What K&s can do

  • Service discovery and load balancing
  • Storage orchestration : Local or Cloud based
  • Automated rollouts and rollbacks
  • Self-healing
  • Secret and configuration management
  • Use the same API across on-premise and every cloud providers

What K&s can't do

  • Does not deploy source code
  • Does not build your application
  • Does not provide application-level services : Message buses, databases, caches, etc

K&s Architecture

K&s Architecture

 --------------------
|  Cluster           |
|  ---------------   |
| |  Node         |  |
| |  ---------    |  |
| | |  Pod    |   |  |
| |  ---------    |  |
|  ---------------   |
 --------------------

Local K8s

Docker Desktop

  • Requires virtualization
  • Docker Desktop
  • MicroK8s
  • Minikube
  • Runs over Docker Desktop
  • Kind
  • Limited to 1 node
  • Docker Desktop
  • Multiple nodes
  • MicroK8s
  • Kind
  • Minikube

Windows

  • Docker Desktop is currently the only way to run both Linux and Windows containers
  • Docker Desktop can run on Hyper-V or WSL 2 (Win 10 2004)
  • If Hyper-V is enabled, you can't run another hypervisor at the same time
  • You can install Minikube on Hyper-V or VirtualBox

Minikube

  • Does not require Docker Desktop
  • Installs on Linux, macOS and Windows
  • https://kubernetes.io/docs/tasks/tools/install-minikube/
  • An Hypervisor like VirtualBox is required
  • install using brew in macOS

Test Kubernetes Locally

kubectl cluster-info

K8s CLI & Context

K8s API

K8s API

  • Kubernetes API server is a service running on the master node.
  • It exposes a rest API that is the only point of communication for kubernetes clusters
  • We define the desired state in yaml files
  • Let's say we want to run X number of instances of a container in the cluster
  • We then send that desired state to the cluster via the rest API
  • Other applications like a web dashboard can also communicate with the rest API to display the cluster state

K8s CLI

  • kubectl is the CLI to interact with the cluster
  • Communicates with the apiserver
  • Configuration stored locally 
    ${HOME}/.kube/config
C:\Users\{USER}\.kube\config

K8s Context

  • A context is a group of access parameters to a K8s cluster
  • Contains a Kubernetes cluster, a user, and a namespace
  • The current context is the cluster that is currently the default for kubectl
  • All kubectl commands run against that cluster

kubectl - Context Cheat Sheet

kubectl config current-context               # Get the current context
kubectl config get-contexts                  # List all context
kubectl config use-context [contextName]     # Change/Set the current context
kubectl config delete-context [contextName]  # Delete a context from the config file

kubectx - Quickly swithc context

  • Instead of typing kubectl config use-context minikube
  • Simply type kubectx [contextName]
  • macOS brew install kubectx
  • Ubuntu sudo apt install kubectx (https://github.com/ahmetb/kubect)

2 ways to create resources

The declarative way vs the imperative way

Imperative

  • Using kubectl commands, issue a series of commands to create resources
  • Great for learning, testing and troubleshooting
  • It's like code

Declarative

  • Using kubectl and YAML manifests defining the resources that you need
  • Reproducible, repeatable
  • Can be saved in source control
  • It's like data that can be parsed and modified

What's a YAML file?

  • It's a text file that contains properties that define the resource
  • It has some required properties like the API version the object kind that defines the type of object you want to create
  • Root level required properties
  • apiVersion
    • Api version of the object
  • kind
    • type of object
  • metadata.name
    • unique name for the object
  • metadata.namespace
    • scoped environment name (will default to current)
  • spec
    • object specifications or desired state
  • You can use the kubectl create command to send the information to the kubernetes cluster 
    kubectl create -f [YAML file]

Declarative vs Imperative example : L18-04