[cloud native learning notes] kubernetes Foundation

---

Preface

During the development in school, I seldom touch the knowledge about application deployment , Now that I know about cloud primordial, I find that my knowledge of this aspect is still in “ primeval ages ”. Because I am interested in this aspect , I'd like to write down the relevant study notes .

---

One 、 Cloud native development

Understanding the emergence of technology can lead to a deeper understanding of it , Here we can understand the emergence and development of cloud native technology from the change of the operation mode of application deployment .

~2000 year ： Physical single machine
2010~2009： Initial stage of virtualization
2010~ so far ： Virtualization maturity
2013~ so far ： Containerization
2015~ so far ： Cloud native

namely ： Traditional deployment --> Virtualization deployment --> Container deployment

Link to the original text ： Details

Two 、kubernetes What is it?

1. Container orchestration management platform ： Provide container management 、 health examination 、 Self healing and upgrading .
2. Microservice support platform ： Support service discovery 、 Deploy 、 Automatic load balancing, etc .
3. portable “ Cloud platform ”： Support “ Cloud migration ”

3、 ... and 、kubernetes Architecture and components

Link to the original text ： Details

In general ,k8s Clusters are composed of Master Nodes and Node node . All nodes exist node Components （kubelet、kube-proxy and pod）.master Nodes still exist master Components （api-server、scheduler、controller manager and etcd）.

1. master Components

It belongs to the cluster control node , Responsible for the management and control of the whole cluster , from api-server、scheduler、controller manager and etcd form .

api-server： The data interaction and communication hub between various components in the cluster . Provide cluster control and security mechanism .
scheduler： adopt api-server Of watch Interface listening New pod Copy information , And choose the most suitable one through the scheduling algorithm Node node . The scheduling algorithm supports customization .
controller manage： resources controller The core manager of , Each resource has a corresponding Controller, Ensure that every Controller The corresponding resource is always in the expected state .
etcd：k8s The main database of the cluster , Store all resource objects and their states . Data change passed api-server Conduct .

2. node Components

Belongs to the workload node , from kubelet、kube-proxy and pod form .

kubelet： Handle master Distributed to this node Upper pod Create and start / stop management tasks . towards api-server register Node Information . And regularly to Master Report node resource usage .
kube-proxy： yes service Implementation of abstract concepts , Will arrive service Request by （ Load balancing ） The policy algorithm passes endpoint Distribute to the back end pod On . Support nodeport Pattern , To access from outside the cluster service.
pod： Yes, you can. Kubernetes Created and managed in 、 The smallest deployable cell ,1 individual pod Can contain 1 One or more containers .