kubernetes简称k8s, 主要用途是automate deployment, scaling, and managment of containerized applications。是目前非常火的docker集群部署管理方案，适用于部署microservice架构。是google15年的经验积淀。这就要说说google为啥要用container了，因为那个时候还没有virtual machine这个技术，当然现在发现这个方向的应用前景比virtual machine集群效果要好。提到kubernetes，不能不说CNCF，CNCF出品，必属精品。现在k8s和envoy(lyft的神作)打得火热，感觉是未来云服务商的主流架构，GKE是google k8s engine也是kubernetes的原始应用，不过Azure好像现在也开始支持k8s，amazon的ecs如果不是用的k8s做后端那么也势必会迁移到k8s。
alternative solution to k8s:
- docker swarm
- hashicorp nomad
- apache mesos
- rancher
k8s的诞生
k8s的前身是google的Borg，google没有选择开源Borg，估计是因为内部依赖比较多，而是从头build了k8s，所以说k8s的架构从一开始就站在google15年大型container管理经验的肩上，根正苗红。Google 2015年发表了著名的Borg paper “Large-scale cluster management at Google with Borg”，从而将秘密昭示天下。
下面是一个趣闻，说mesos是怎么诞生的，当时伯克利的一群科学家和google做Borg的工程师酒后闲聊，弄明白了Borg的核心原理，然后就搞出了mesos，所以mesos反而是抢在了k8s的前面，提前成为了很多大公司包括twitter，verizon在内的解决方案，k8s这个正品反而由硅谷的startup率先尝试。所以江湖传言说需求大还是要靠mesos就是这么来的，但是k8s正在飞速的证明自己的成熟度。
k8s 架构
k8s could be deployed as services on VMs or bare-metal machines.
k8s 的安装，本文仅限于单机版minikube
安装minikube
https://github.com/kubernetes/minikube
安装kubectl
https://kubernetes.io/docs/tasks/tools/install-kubectl/#install-kubectl-binary-via-curl
安装virtualbox
https://www.virtualbox.org/wiki/Downloads
k8s dashboard
minikube dashboard
在dashboard里create app redis，name redis，然后查看：
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
redis-76d7657885-rk6tm 1/1 Running 0 1m
$ kubectl get deployment
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
redis 1 1 1 1 2m
$ kubectl get replicaset
NAME DESIRED CURRENT READY AGE
redis-76d7657885 1 1 1 2m
$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
minikube Ready <none> 14d v1.9.0
kubectl 应用
check redis logs并与redis进行简单交互：
$ kubectl logs redis-76d7657885-rk6tm
1:C 27 Feb 21:47:54.141 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
1:C 27 Feb 21:47:54.141 # Redis version=4.0.8, bits=64, commit=00000000, modified=0, pid=1, just started
1:C 27 Feb 21:47:54.141 # Warning: no config file specified, using the default config. In order to specify a config file use redis-server /path/to/redis.conf
1:M 27 Feb 21:47:54.146 * Running mode=standalone, port=6379.
1:M 27 Feb 21:47:54.147 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
1:M 27 Feb 21:47:54.147 # Server initialized
1:M 27 Feb 21:47:54.147 * Ready to accept connections
$ kubectl exec -ti redis-76d7657885-rk6tm -- redis-cli
127.0.0.1:6379> set foo bar
OK
127.0.0.1:6379> get foo
"bar"
查看pod的yaml输出，object存于etcd中
$get pods redis-76d7657885-rk6tm -o yaml
apiVersion: v1
kind: Pod
metadata:
creationTimestamp: 2018-02-27T21:47:43Z
generateName: redis-76d7657885-
labels:
k8s-app: redis
pod-template-hash: "3283213441"
name: redis-76d7657885-rk6tm
...
刚才是通过dashboard的create button创建的pod，下面通过yaml创建pod：
$ cat pod.yaml
apiVersion: v1
kind: Pod
metadata:
name: qingge
spec:
containers:
- name: nginx
image: nginx
$ kubectl create -f pod.yaml
pod "qingge" created
显示pods
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
qingge 0/1 ContainerCreating 0 6s
redis-76d7657885-rk6tm 1/1 Running 0 16m
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
qingge 1/1 Running 0 1m
redis-76d7657885-rk6tm 1/1 Running 0 17m
启动proxy：
$ kubectl proxy
Starting to serve on 127.0.0.1:8001
访问proxy：
$ curl localhost:8001
{
"paths": [
"/api",
"/api/v1",
"/apis",
"/apis/",
...
$ curl localhost:8001/api/v1
{
"kind": "APIResourceList",
"groupVersion": "v1",
"resources": [
{
"name": "bindings",
"singularName": "",
"namespaced": true,
"kind": "Binding",
"verbs": [
"create"
]
...
delete pods:
$ kubectl delete pods qingge
pod "qingge" deleted
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
redis-76d7657885-rk6tm 1/1 Running 0 28m
namespace:
$ kubectl get ns
NAME STATUS AGE
default Active 14d
kube-public Active 14d
kube-system Active 14d
$ kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
default redis-76d7657885-rk6tm 1/1 Running 0 33m
kube-system kube-addon-manager-minikube 1/1 Running 1 14d
kube-system kube-dns-54cccfbdf8-mhr9h 3/3 Running 3 14d
kube-system kubernetes-dashboard-77d8b98585-5r7qv 1/1 Running 1 14d
kube-system storage-provisioner 1/1 Running 1 14d
添加quota 限制：
$ kubectl create quota qingge --hard=pods=1
resourcequota "qingge" created
$ kubectl get quota
NAME AGE
qingge 12s
$ kubectl get quota qingge -o yaml
apiVersion: v1
kind: ResourceQuota
metadata:
creationTimestamp: 2018-02-27T22:24:08Z
name: qingge
namespace: default
resourceVersion: "2185"
selfLink: /api/v1/namespaces/default/resourcequotas/qingge
uid: ed5ec545-1c0c-11e8-aa17-08002725a271
spec:
hard:
pods: "1"
status:
hard:
pods: "1"
used:
pods: "1"
$ kubectl create -f pod.yaml
Error from server (Forbidden): error when creating "pod.yaml": pods "qingge" is forbidden: exceeded quota: qingge, requested: pods=1, used: pods=1, limited: pods=1
修改quota 把spec中的pods改为2，然后就可以添加pod了
$ kubectl edit resourcequota oreilly
resourcequota "qingge" edited
$ kubectl create -f pod.yaml
pod "qingge" created
观察删除redis后会自动被replica恢复：
$ kubectl delete pods redis-76d7657885-rk6tm
pod "redis-76d7657885-rk6tm" deleted
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
redis-76d7657885-lnmps 0/1 ContainerCreating 0 2s
redis-76d7657885-rk6tm 0/1 Terminating 0 43m
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
redis-76d7657885-lnmps 1/1 Running 0 6s
redis-76d7657885-rk6tm 0/1 Terminating 0 44m
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
redis-76d7657885-lnmps 1/1 Running 0 8s
添加nginx
$ more rs.yaml
apiVersion: extensions/v1beta1
kind: ReplicaSet
metadata:
name: nginx
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
name: nginx-pod
labels:
app: nginx
spec:
containers:
- name: nginx-containers
image: nginx
$ kubectl create -f rs.yaml
replicaset "nginx" created
$ kubectl get replicaset
NAME DESIRED CURRENT READY AGE
nginx 2 1 1 7s
redis-76d7657885 1 1 1 50m
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-dr4t9 1/1 Running 0 33s
redis-76d7657885-lnmps 1/1 Running 0 6m
测试删除后自动回复：
$ kubectl delete pods nginx-dr4t9
pod "nginx-dr4t9" deleted
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-g8spl 1/1 Running 0 3s
nginx-mflm7 0/1 ContainerCreating 0 2s
redis-76d7657885-lnmps 1/1 Running 0 8m
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-g8spl 1/1 Running 0 6s
nginx-mflm7 1/1 Running 0 5s
redis-76d7657885-lnmps 1/1 Running 0 8m
选择label：
$ kubectl get pods -l app=nginx
NAME READY STATUS RESTARTS AGE
nginx-g8spl 1/1 Running 0 3h
nginx-mflm7 1/1 Running 0 3h
添加label：
$ kubectl get pods -Lapp
NAME READY STATUS RESTARTS AGE APP
nginx-g8spl 1/1 Running 0 3h nginx
nginx-mflm7 1/1 Running 0 3h nginx
redis-76d7657885-lnmps 1/1 Running 0 3h
$ kubectl label pods redis-76d7657885-lnmps app=redis
pod "redis-76d7657885-lnmps" labeled
$ kubectl get pods -Lapp
NAME READY STATUS RESTARTS AGE APP
nginx-g8spl 1/1 Running 0 3h nginx
nginx-mflm7 1/1 Running 0 3h nginx
redis-76d7657885-lnmps 1/1 Running 0 3h redis
添加多个label，显示所有label
$ kubectl label pods nginx-g8spl foo=bar
pod "nginx-g8spl" labeled
$ kubectl get pods --show-labels
NAME READY STATUS RESTARTS AGE LABELS
nginx-g8spl 1/1 Running 0 3h app=nginx,foo=bar
nginx-mflm7 1/1 Running 0 3h app=nginx
redis-76d7657885-lnmps 1/1 Running 0 3h app=redis,k8s-app=redis,pod-template-hash=3283213441
思考：如果此时overwrite 一个app=nginx的pod会有什么后果：
$ kubectl label pods nginx-g8spl app=foobar --overwrite
pod "nginx-g8spl" labeled
$ kubectl get pods --show-labels
NAME READY STATUS RESTARTS AGE LABELS
nginx-g8spl 1/1 Running 0 3h app=foobar,foo=baz
nginx-mflm7 1/1 Running 0 3h app=nginx
nginx-mlhjs 1/1 Running 0 1m app=nginx
redis-76d7657885-lnmps 1/1 Running 0 3h app=redis,k8s-app=redis,pod-template-hash=3283213441
答：会额外多生成一个nginx pod，因为replicaset为label app=nginx定义的数量是2，所以它会始终维护数量为2
添加service：
$ cat svc.yaml
apiVersion: v1
kind: Service
metadata:
name: nginx
spec:
ports:
- name: main
port: 80
selector:
app: nginx
type: NodePort
$ kubectl create -f svc.yaml
service "nginx" created
查看并启动服务，会在浏览器中看到nginx的欢迎界面，port number是30978
$ kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 14d
nginx NodePort 10.110.210.218 <none> 80:30978/TCP 5s
$ minikube service nginx
$ kubectl get endpoints
NAME ENDPOINTS AGE
kubernetes 10.0.2.15:8443 14d
nginx 172.17.0.6:80,172.17.0.7:80 1h
下面来个牛逼的操作，scale：
$ kubectl scale rs nginx --replicas=5
replicaset "nginx" scaled
$ kubectl get rs
NAME DESIRED CURRENT READY AGE
nginx 5 5 5 5h
redis-76d7657885 1 1 1 5h
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-d8n5h 1/1 Running 0 13s
nginx-g8spl 1/1 Running 0 5h
nginx-mflm7 1/1 Running 0 5h
nginx-mlhjs 1/1 Running 0 1h
nginx-nc5z7 1/1 Running 0 13s
nginx-t228t 1/1 Running 0 13s
redis-76d7657885-lnmps 1/1 Running 0 5h
我已经create了一个busybox，下面通过busybox来观察nginx：
$ kubectl exec -ti busybox-7c6c97f7cb-s6trg -- /bin/sh
/ #
/ # nslookup nginx
Server: 10.96.0.10
Address 1: 10.96.0.10 kube-dns.kube-system.svc.cluster.local Name: nginx
Address 1: 10.110.210.218 nginx.default.svc.cluster.local
访问nginx：
/ # wget 10.110.210.218
Connecting to 10.110.210.218 (10.110.210.218:80)
index.html 100% |*************************************************************************************************************************| 612 0:00:00 ETA
/ # head index.html
<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
<style>
body {
width: 35em;
margin: 0 auto;
font-family: Tahoma, Verdana, Arial, sans-serif;
}
好了，今天到此为止，更多内容下回分解