Kubernetes Administration Quiz

Kubernetes Architecture

• A Kubernetes cluster can be divided into two parts: Control Plane and Nodes (or workers)
• The Control Plane is the brain of Kubernetes, responsible for controlling and managing a set of worker nodes
• Each node has a "Kubelet", responsible for communicating with the Control Plane

Kubernetes Components

• Node: A node may be a virtual or physical machine, depending on the cluster, and runs containers using Pods
• Cluster: A set of nodes that run containerized applications, allowing containers to run across multiple machines and environments
• Master Node: Hosts the control plane aspects of the cluster and is responsible for controlling and managing a set of worker nodes
• API Server: Validates and configures data for API objects, including Pods and Services, and handles authentication, authorization, and admission control
• Etcd: A distributed key-value data store that contains all Kubernetes cluster information, with multiple instances possible for cluster availability and performance
• Kubelet: Responsible for everything that happens on the node, listening to the API Server, starting containers, and reporting their statuses
• Container Runtime: Responsible for running containers, with supported runtimes including Docker, containerd, and CRI-O
• Scheduler: Watches for newly created pods and assigns nodes for them to run on, considering factors like resource requirements, hardware/software constraints, and data locality
• Controller Manager: Contains different controllers responsible for reconciling the actual state of the cluster to the desired state, with controllers for deployments, nodes, services, and more

Pods and Containers

• Pod: The most basic object in Kubernetes, encapsulating one or more containers, with each container instance in a pod recommended to be separate
• Container: Runs inside a Pod, with multiple containers possible in a single Pod

Kubernetes Roll-outs

• Create a YAML file describing the desired state configuration of the cluster
• Apply the YAML file to the cluster through kubectl
• kubectl submits the request to the kube-apiserver, which authenticates and authorizes the request
• The kube-controller-manager continuously monitors the system for new requests and works towards reconciling the system state to the desired state
• The kube-scheduler assigns nodes for pods to run on, and communicates with the kubelet to start the deployment

Setting up Kubernetes

• Local Solutions: Minikube, a utility used to run Kubernetes on a local machine, creating a single node cluster for demo Kubernetes operations
• Public Cloud Solutions: Online k8s solutions
• Installation Methods: The Hard Way, Installer-based kubeadm, Kops, Kubespray, etc., with kubeadm allowing for simple, automated installations of multi-node clusters