service的作用体现在两个方面:
集群内部:不断追踪pod的变化。他会更新endpoint中的pod对象,基于pod的IP地址不断变化的一种服务发现机制。
集群外部:类似负载均衡器,把流量IP+端口,不涉及转发url(http、https)。把请求转发到pod当中。
service有四种类型:
ClusterIP:创建service的默认类型
NodePort:容器端口 > service端口 > nodeport。 设定了nodeport后,每个节点都会有一个端口被打开。 端口范围:30000-32767。 访问:节点ip+30000-32767实现负载均衡
loadbalancer:云平台上的一种service服务。云平台提供负载均衡的IP地址
extrenal:域名映射。
ingress是一个API对象,通过yaml文件进行配置。
ingress的作用:是定义请求如何转发到service的规则。相当于一个配置模板
ingress通过http和https暴露集群内部的service。
ingress给service提供外部的URL域名,负载均衡以及ssl/tls(加密的https)的能力。实现一个基于域名的负载均衡。
ingress-controller的作用就是具体的实现反向代理和负载均衡的程序。对ingress定义的规则进行解析。根据ingress的配置规则进行请求转发。
ingress-controlle是以pod的方式运行在集群中
ingress-controller不是k8s自带的组件功能。他是一个统称,只要这个组件可以实现反向代理和负载均衡可以对ingress进行解析的、可以根据规则请求转发的都是ingress-controller
nginx-ingress-controller、traefik都是开源的ingress-controller
1、 定义外部流量的路由规则
2、 定义了服务的暴露方式、主机名、访问路径和其他的选项。
3、 实现负载均衡。这是由ingress-controller实现的。
1、 deployment+LoadBalancer模式:这种模式是将ingress部署在公有云。例如:华为云、阿里云、腾讯云等
在ingress的配置文件内会有一个type。type会以键值对形式设置 type: LoadBalancer
公有云平台会为 LoadBalancer 的service创建一个负载均衡器。绑定一个公网地址。通过域名指向公网地址。就可以实现集群对外暴露。
2、 DaemonSet+hostnetwork+nodeSelector模式:
其中DaemonSet会在每个节点创建一个pod。
hostnetwork表示pod和节点主机共享网络命名空间。容器内可以直接使用节点主机的IP+端口。pod中的容器可以直接访问主机上的网络资源。
nodeSelector根据标签来选择部署的节点。nginx-ingress-controller部署的节点。
缺点:直接利用节点主机的网络和端口。一个node之内部署一个ingress-controller pod。
DaemonSet+hostnetwork+nodeSelector模式性能最好,比较适合大并发的生产环境。
ingress的数据流向图:
客户端发起请求域名将先到DNS
DNS开始解析域名。映射到ingress-controller所在的节点
ingress-controller以pod形式运行在节点上。hostnetwork可以和节点主机共享网络
ingress的配置来定义URL的地址
根据ingress的标签匹配将请求转发到service
service寻找endpoints发现匹配能够转发的pod
最终还是由ingress-controller将(http/https)请求转发到不同的pod上。实现负载均衡
service和endpoints在这里起发现和监控的总用
实际的负载均衡由ingress-controller实现
实验部署:
master01 20.0.0.32
node01 20.0.0.34
node02 20.0.0.35
master01--
修改ingress的yaml文件
vim mandatory.yaml
191 #kind: Deployment
192 kind: DaemonSet
200 # replicas: 1
215 hostNetwork: true
220 test1: "true"
每台节点主机都添加nginx-ingress-controller镜像
tar -xf ingree.contro-0.30.0.tar.gz
docker load -i ingree.contro-0.30.0.tar
master01--
kubectal get pod -n ingress-nginx
kubectl label nodes node02 ingress=true
kubectl apply -f mandatory.yaml
master01---
vim nginx.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: nfs-pvc
spec:
accessModes:
- ReadWriteMany
storageClassName: nfs-client-storageclass
resources:
requests:
storage: 2Gi
---
#定义pod
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-app
labels:
app: nginx1
spec:
replicas: 3
selector:
matchLabels:
app: nginx1
template:
metadata:
labels:
app: nginx1
spec:
containers:
- name: nginx
image: nginx:1.22
volumeMounts:
- name: nfs-pvc
mountPath: /usr/share/nginx/html
volumes:
- name: nfs-pvc
persistentVolumeClaim:
claimName: nfs-pvc
---
#定义service
apiVersion: v1
kind: Service
metadata:
name: nginx-app-svc
spec:
ports:
- protocol: TCP
port: 80
targetPort: 80
selector:
app: nginx1
---
#定义ingress
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: nginx-app-ingress
spec:
rules:
- host: www.test1.com
http:
paths:
- path: /
#匹配工作目录的根目录
pathType: Prefix
#根据前缀进行匹配 只要是/开头的都可以匹配到例如/ www.test1.com/www1/www2/www3
backend:
#指定后台服务器
service:
name: nginx-app-svc
port:
number: 80
vim /etc/hosts
20.0.0.32 master01
20.0.0.34 node01
20.0.0.35 node02 www.test1.com
20.0.0.36 hub.test.com k8s5
#做域名和地址映射
k8s5---
进入挂载卷之后
echo 123 > index.html
master01---
curl www.test1.com
#测试网页是否可以访问
实验完成!
创建地址映射
回到k8s主机查看共享目录是否生成
测试访问是否成功
8181端口,nginx-controller默认配置的一个bachend。反向代理的端口。
所有请求当中。只要是不符合ingress配置的请求就会转发到8181端口
deployment+NodePort的数据流向图:
实验部署:
master01---
vim mandatory.yaml
191 kind: Deployment
215 #hostNetwork: true
200 replicas: 1
219 kubernetes.io/os: linux
220 #test1: "true"
kubectl apply -f mandatory.yaml
wget https://gitee.com/mirrors/ingress-nginx/raw/nginx-0.30.0/deploy/static/provider/baremetal/service-nodeport.yaml
#获取service.yaml文件
vim service-nodeport.yaml
apiVersion: v1
kind: Service
metadata:
name: ingress-nginx
namespace: ingress-nginx
labels:
app.kubernetes.io/name: ingress-nginx
app.kubernetes.io/part-of: ingress-nginx
spec:
type: NodePort
ports:
- name: http
port: 80
targetPort: 80
protocol: TCP
- name: https
port: 443
targetPort: 443
protocol: TCP
selector:
app.kubernetes.io/name: ingress-nginx
app.kubernetes.io/part-of: ingress-nginx
#执行这个yaml文件,会生成一个service。在ingress-nginx这个命名空间生成一个service。
#所有的controller的请求都会从这个定义的service的nodeport的端口。
#把请求转发到自定义的service的pod
kubectl apply -f service-nodeport.yaml
vim nodeport.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: nfs-pvc2
spec:
accessModes:
- ReadWriteMany
storageClassName: nfs-client-storageclass
resources:
requests:
storage: 2Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-app2
labels:
app: nginx2
spec:
replicas: 3
selector:
matchLabels:
app: nginx2
template:
metadata:
labels:
app: nginx2
spec:
containers:
- name: nginx
image: nginx:1.22
volumeMounts:
- name: nfs-pvc2
mountPath: /usr/share/nginx/html
volumes:
- name: nfs-pvc2
persistentVolumeClaim:
claimName: nfs-pvc2
---
apiVersion: v1
kind: Service
metadata:
name: nginx-app-svc1
spec:
ports:
- protocol: TCP
port: 80
targetPort: 80
selector:
app: nginx2
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: nginx-app-ingress
spec:
rules:
- host: www.test2.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: nginx-app-svc1
port:
number: 80
kubectl apply -f nodeport.yaml
k8s5---
查看挂载目录
echo 123 > index.html
master01---
vim /etc/hosts
20.0.0.32 master01
20.0.0.34 node01
20.0.0.35 node02 www.test1.com www.test2.com
20.0.0.36 hub.test.com k8s5 www.test1.com
curl www.test2.com:31456
实验完成!
nodeport不再是deployment的service创建的
而是由ingress创建的
url请求先到ingress
ingress根据标签匹配ingress-nodeport
通过标签寻找指定的service
最终由service找到deployment
核心的控制组件时nginx-ingress-controller
host----ingress的配置找到pod----controller----把请求发到pod
nodeport-----controller-----ingress----service----pod
nodeport暴露端口的方式是最简单的。nodeport多了一层net(地址转换)
并发量大的对性能会有一定影响。内部都会用nodeport
通过ingress的方式实现:一个ingress可以访问不同的主机
实验举例:
vim pod1.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: deployment1
labels:
test: nginx1
spec:
replicas: 1
selector:
matchLabels:
test: nginx1
template:
metadata:
labels:
test: nginx1
spec:
containers:
- name: nginx1
image: nginx:1.22
---
apiVersion: v1
kind: Service
metadata:
name: svc-1
spec:
ports:
- port: 80
targetPort: 80
protocol: TCP
selector:
test: nginx1
vim pod2.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: deployment1
labels:
test2: nginx2
spec:
replicas: 1
selector:
matchLabels:
test2: nginx2
template:
metadata:
labels:
test2: nginx2
spec:
containers:
- name: nginx2
image: nginx:1.22
---
apiVersion: v1
kind: Service
metadata:
name: svc-2
spec:
ports:
- port: 80
targetPort: 80
protocol: TCP
selector:
test2: nginx2
vim pod-ingress.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: ingress1
spec:
rules:
- host: www.test1.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: svc-1
port:
number: 80
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: ingress2
spec:
rules:
- host: www.test2.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: svc-2
port:
number: 80
kubectl apply -f pod.yaml
kubectl apply -f pod2.yaml
kubectl apply -f pod-ingress.yaml
vim /etc/hosts
20.0.0.32 master01
20.0.0.34 node01 www.test1.com www.test2.com
20.0.0.35 node02 www.test1.com www.test2.com
20.0.0.36 hub.test.com k8s5 www.test1.com
curl www.test1.com:31456
curl www.test2.com:31456
访问成功实验完成!
实验举例:
vim daemoset-hostnetwork-nodeselector1.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: nfs-pvc1
spec:
accessModes:
- ReadWriteMany
storageClassName: nfs-client-storageclass
resources:
requests:
storage: 2Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-app1
labels:
app: nginx1
spec:
replicas: 1
selector:
matchLabels:
app: nginx1
template:
metadata:
labels:
app: nginx1
spec:
containers:
- name: nginx
image: nginx:1.22
volumeMounts:
- name: nfs-pvc1
mountPath: /usr/share/nginx/html
volumes:
- name: nfs-pvc1
persistentVolumeClaim:
claimName: nfs-pvc1
---
apiVersion: v1
kind: Service
metadata:
name: nginx-app-svc1
spec:
ports:
- protocol: TCP
port: 80
targetPort: 80
selector:
app: nginx1
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: nginx-app-ingress1
spec:
rules:
- host: www.test1.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: nginx-app-svc1
port:
number: 80
kubectl apply -f daemoset-hostnetwork-nodeselector1.yaml
vim daemoset-hostnetwork-nodeselector2.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: nfs-pvc2
spec:
accessModes:
- ReadWriteMany
storageClassName: nfs-client-storageclass
resources:
requests:
storage: 2Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-app2
labels:
app: nginx2
spec:
replicas: 1
selector:
matchLabels:
app: nginx2
template:
metadata:
labels:
app: nginx2
spec:
containers:
- name: nginx
image: nginx:1.22
volumeMounts:
- name: nfs-pvc2
mountPath: /usr/share/nginx/html
volumes:
- name: nfs-pvc2
persistentVolumeClaim:
claimName: nfs-pvc2
---
apiVersion: v1
kind: Service
metadata:
name: nginx-app-svc2
spec:
ports:
- protocol: TCP
port: 80
targetPort: 80
selector:
app: nginx2
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: nginx-app-ingress2
spec:
rules:
- host: www.test2.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: nginx-app-svc2
port:
number: 80
kubectl apply -f daemoset-hostnetwork-nodeselector2.yaml
vim mandatory.yaml
192 kind: DaemonSet
200 #replicas: 1
215 hostNetwork: true
219 #kubernetes.io/os: linux
220 test1: "true"
kubectl apply -f mandatory.yaml
curl www.test1.com
curl www.test2.com
实验完成!!
ingress-controller两种方式:nginx-ingress-controller、traefik
ingress-controller的三种工作模式:
deployment+loadbalancer:需要云平台提供一个负载均衡的公网地址。公有云上做。需要收费
daemonset+hostnetwork+nodeselector:一般都会指定节点部署controller。缺点就是和宿主机共享网络,只能是一个controller的pod
hostnetwork会和宿主机共享网络。所以需要指定标签
deployment+NodePort:这是最常用最简单的方式。他会集中一个nodeport端口,所有ingress的请求都会转发到nodeport。然后由service将流量转发到pod
一个ingress的nodeport可以实现访问多个虚拟主机。和nginx类似。同一个端口下可以有多个域名