Shulou(Shulou.com)06/01 Report--

This article mainly introduces the relevant knowledge of K8s cluster component instance analysis, the content is detailed and easy to understand, the operation is simple and fast, and has a certain reference value. I believe everyone will gain something after reading this K8s cluster component instance analysis article. Let's take a look.

K8s is an open source project based on Google's 15 years experience in production environment, which is based on the predecessor of Borg.

K8s is committed to providing a platform for automatic deployment, extension, high availability, and running application containers across host clusters.

K8s cluster consists of master node and node node, which is a master-slave architecture. Usually, we use three servers with high availability as the master node, without business deployment (you can set whether to deploy the application on the master node) and the slave node as the work node.

The following are the cluster components and functions:

Master node: the control center of the cluster

Kube-apiserver: the control center of the cluster. The information exchange between each module needs to go through kube-apiserver. At the same time, it is also the entrance of cluster management, resource allocation and cluster security mechanism.

Controller-manager: the status manager of the cluster, which ensures that resources such as pod always achieve the desired effect. You need to communicate with apiserver to create, update or delete managed resources when needed.

Scheduler: the scheduling center of the cluster, which selects the best node to deploy pod; according to a specified series of conditions. That is to say, if we have five node nodes, when we create a pod, it is up to scheduler to decide which node the pod is deployed on.

Etcd cluster: key-value database to store cluster information. In production environment, it is best to separate etcd from cluster. It is recommended to have more than 3 nodes (odd) and install using ssd hard disk.

Node node: work node, also known as worker node

Kubelet: responsible for monitoring the pod status on the node, reporting the pod status of the node and the node, communicating with the master node, and managing pod.

Kube-proxy: responsible for communication and load balancing between pod, distributing specified traffic to backend machines.

View the kube-proxy mode of operation:

Curl 127.0.0.1:10249/proxyMode

I am using ipvs.

Ipvs: listens to master nodes to add and delete service and endpoint messages, and calls API netlink to create corresponding ipvs rules. Traffic is forwarded to the appropriate pod through ipvs rules.

Iptables: listens to master nodes to add and delete service and endpoint messages. For each service, he creates an iptables rule and proxies service's clusterIP to the corresponding pod at the back end.

Kube-system Namespace: system component pod

Calico: a CNI-compliant network plug-in that generates a unique IP address for each pod and treats each node as a router (implementation or node communication).

CoreDNS: used for service resolution within K8s cluster. You can let pod resolve the service name to an IP address, and then connect to the corresponding application through the IP address of service.

Docker: container engine, which is responsible for managing containers

Metrics-server: view pod resource usage

This is the end of the article on "K8s cluster component instance analysis". Thank you for reading! I believe you all have a certain understanding of the knowledge of "K8s cluster assembly instance analysis". If you want to learn more knowledge, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.