Kubernetes Chapter 1: Foundation

Preface

Kubernetes ( The sound kubə'netis) The name comes from Greek , intend 「 helmsman 」 or 「 pilot 」, It is an origin of Google Open source projects for , Allow automated deployment 、 Manage and expand container applications , It has now become the de facto standard for container choreography .

Actually Kubernetes For short k8s A stronger sense of presence , The reason for this abbreviation is that the middle of the word happens to be 8 Letters , This is a numerical abbreviation (Numeronym) The way . Similar as internationalization( internationalization ) be called i18n、accessibility be called a11y etc. . The following abbreviations are used unless otherwise specified k8s.

Chapter one , First understand k8s The architecture of 、 Core concepts and basic usage .

Master node

Master Is the control node of the cluster , It is responsible for the management and control of the whole cluster ( Dispatch ). There are multiple components running on this node , The core is as follows :

• API The server (kube-apiserver).API Server Exposed to the outside world k8s API, Provide HTTP REST service . Certification provided 、 to grant authorization 、 Access control 、API Mechanisms such as registration and discovery
• etcd. Is a highly available key value pair storage system , be used as k8s Back end storage , All cluster configuration data is stored in it , For service discovery and cluster management , It can be said that the status of the entire cluster is saved
• Control Manager (kube-controller-manager).Controller Responsible for maintaining the status of the cluster , Like fault detection 、 Automatic extension 、 Rolling updates, etc.
• Scheduler (kube-scheduler).Scheduler Responsible for resource scheduling , According to the scheduled scheduling policy Pod Dispatch to the corresponding machine

to glance at Mater Component architecture rendering of node :

Pod What is it? It was said very soon ~

Node

Node It is the working node of the cluster , It provides CPU、 Resources such as memory and storage . Which container applications run on it are controlled by Master Node allocation . Multiple components are also running on this node , The core is as follows :

• Node agent (kubelet). It and Master Node collaboration , Realization Pod The creation of 、 start-up 、 monitor 、 restart 、 Destruction and other cluster management work
• Forwarding agent (kube-proxy). Maintain network rules on the host and perform connection forwarding , Implement load balancing and reverse proxy of services
• Docker. Responsible for operating the container ( Or say Pod)
• fluentd.fluentd Is a Daemons , It helps to provide cluster level logging Cluster level logs .

to glance at Node Component architecture rendering of node :

There are also some plug-ins marked in this figure , Such as CoreDNS、Dashboard And those who have not been replaced Ingress Controller etc. , I'll talk about it later , It's too much , Let's start with the basics .

You can see the cluster nodes through the following commands :

* kubectl get nodes
NAME             STATUS   ROLES    AGE    VERSION
docker-desktop   Ready    master   2m    v1.14.6

Of course , This is a single node cluster environment , only one Master node , But I have Node The role of nodes .

Pod

Pod yes k8s The basic deployment scheduling unit , Every Pod It can consist of one or more containers , Shared storage between containers 、 Network and other resources .Pod At a glance, this word is not very confused ？ Actually, take a look at its Chinese translation 「 Pods 」 It is easy to understand :

Peas are covered with pea pods 「 package 」 Get up , So that's one Pod. So you can put a single Pod It can be understood as running independent applications 「 Logical host 」—— It runs one or more tightly coupled application containers .