k8s introduction
k8s is the full name kubernetes, and there are 8 letters between k and s, so it is referred to as k8s (this is also what other people say, I don’t know if it’s true or not)
It is an open source version of the borg architecture evolved from Google's borg architecture
Features:
Lightweight, low resource consumption (written in go language)
Elastic scaling
Load balancing
Open source
High availability
k8s components
master side
APIserver: access to a unified entrance for all services, responsible for direct communication between components
crontrollermanager: controller, to maintain the expected number of copies (there are many types of controllers, which will be described in detail later)
scheduler: scheduler, responsible for receiving tasks, selecting appropriate nodes to assign tasks
etcd: The key-value pair database stores all important information of the k8s cluster (persistence)
node end
kubelet: Directly interact with the container engine to realize the life cycle management of the container
kube-proxy: Responsible for writing rules to IPTABLES, IPVS to achieve service mapping access
COREDNS: can create a domain name ip correspondence resolution for svc in the cluster
dashboard: Provide a B/s structure access system for k8s cluster
The official INGRESS controller can only implement a four-layer proxy INGRESS can implement a seven-layer proxy
FEDETATION: Provides a unified management function for multiple K8s across cluster centers
Prometheus: Provides the monitoring capability of a k8s cluster
ELK: Provide a unified access platform for k8s cluster log analysis
Pod introduction
Pod is the smallest scheduling unit in k8s. Some people often compare pod to a pea pod. The container is the peas in the pea pod. Then there may be one or more peas. Why not use the container directly, but use the pod To encapsulate one or more containers
Let me give an example here. If there is a system of lnmp architecture, it will generally not run in one container, but in multiple containers, but in this way, the interconnection between them becomes very difficult, and the pod provides sharing. Network and storage (because the network is shared, the port cannot be repeated), so they can communicate internally through localhost. Although the network and storage are shared, the cpu and memory are not. That is to say, we can separately communicate with the pod. Restrictions on the use of resources by containers