K8S从入门到放弃系列-(15)Kubernetes集群Ingress部署

时间:2023-12-13 10:17:50

Ingress是kubernetes集群对外提供服务的一种方式.ingress部署相对比较简单,官方把相关资源配置文件,都已经集合到一个yml文件中(mandatory.yaml),镜像地址也修改为quay.io。

1、部署

官方地址:https://github.com/kubernetes/ingress-nginx

1.1 下载部署文件:

## mandatory.yaml为ingress所有资源yml文件的集合
### 若是单独部署,需要分别下载configmap.yaml、namespace.yaml、rbac.yaml、service-nodeport.yaml、with-rbac.yaml
[root@k8s-master01 ingress-master]# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/mandatory.yaml
### service-nodeport.yaml为ingress通过nodeport对外提供服务,注意默认nodeport暴露端口为随机,可以编辑该文件自定义端口
[root@k8s-master01 ingress-master]# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/static/provider/baremetal/service-nodeport.yaml

1.2 应用yml文件创建ingress资源

[root@k8s-master01 ingress-master]# kubectl apply -f mandatory.yaml 
namespace/ingress-nginx created
configmap/nginx-configuration created
configmap/tcp-services created
configmap/udp-services created
serviceaccount/nginx-ingress-serviceaccount created
clusterrole.rbac.authorization.k8s.io/nginx-ingress-clusterrole created
role.rbac.authorization.k8s.io/nginx-ingress-role created
rolebinding.rbac.authorization.k8s.io/nginx-ingress-role-nisa-binding created
clusterrolebinding.rbac.authorization.k8s.io/nginx-ingress-clusterrole-nisa-binding created
deployment.apps/nginx-ingress-controller created
[root@k8s-master01 ingress-master]# kubectl apply -f service-nodeport.yaml
service/ingress-nginx created

1.3 查看资源创建

[root@k8s-master01 ingress-master]# kubectl get pods -n ingress-nginx -o wide
NAME                                        READY   STATUS    RESTARTS   AGE   IP             NODE         NOMINATED NODE   READINESS GATES
nginx-ingress-controller-86449c74bb-cbkgp   1/1     Running   0          19s   10.254.88.48   k8s-node02   <none>           <none>
### 通过创建的svc可以看到已经把ingress-nginx service在主机映射的端口为33848(http),45891(https)
[root@k8s-master01 ingress-master]# kubectl get svc -n ingress-nginx
NAME            TYPE       CLUSTER-IP       EXTERNAL-IP   PORT(S)                      AGE
ingress-nginx   NodePort   10.254.102.184   <none>        80:33848/TCP,443:45891/TCP   43s

说明: 
  Ingress Contronler 通过与 Kubernetes API 交互,动态的去感知集群中 Ingress 规则变化,然后读取它,按照自定义的规则,规则就是写明了哪个域名对应哪个service,生成一段 Nginx 配置,再写到 Nginx-ingress-control的 Pod 里,这个 Ingress Contronler 的pod里面运行着一个nginx服务,控制器会把生成的nginx配置写入/etc/nginx.conf文件中,然后 reload 一下 使用配置生效。以此来达到域名分配置及动态更新的问题。

2、验证

2.1 创建svc及后端deployment

[root@k8s-master01 ingress-master]# cat test-ingress-pods.yml
apiVersion: v1
kind: Service
metadata:
name: myapp-svc
namespace: default
spec:
selector:
app: myapp
env: test
ports:
- name: http
port:
targetPort:
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: myapp-test
spec:
replicas:
selector:
matchLabels:
app: myapp
env: test
template:
metadata:
labels:
app: myapp
env: test
spec:
containers:
- name: myapp
image: nginx:1.15-alpine
ports:
- name: httpd
containerPort: 80
## 查看pod资源部署
[root@k8s-master01 ingress-master]# kubectl get pods
NAME                             READY   STATUS    RESTARTS   AGE
myapp-test-66cf5bf7d5-5cnjv      1/1     Running   0          3m39s
myapp-test-66cf5bf7d5-vdkml      1/1     Running   0          3m39s
## 查看svc
[root@k8s-master01 ingress-master]# kubectl get svc
NAME           TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
myapp-svc      ClusterIP   10.254.155.238   <none>        80/TCP           4m40s

2.2 创建ingress规则

## ingress规则中,要指定需要绑定暴露的svc名称
[root@k8s-master01 ingress-master]# cat test-ingress-myapp.yml
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: ingress-myapp
namespace: default
annotations:
kubernetes.io/ingress.class: "nginx"
spec:
rules:
- host: www.tchua.top
http:
paths:
- path:
backend:
serviceName: myapp-svc
servicePort: 80
[root@k8s-master01 ingress-master]# kubectl apply -f test-ingress-myapp.yml
[root@k8s-master01 ingress-master]# kubectl get ingress
NAME             HOSTS           ADDRESS   PORTS   AGE
ingress-myapp    www.tchua.top             80      13s

2.3 在win主机配置hosts域名解析

## 这里随机解析任一台节点主机都可以

172.16.11.123 www.tchua.top

然后主机浏览器访问http://www.tchua.top:33848,这里访问时需要加上svc映射到主机时随机产生的nodePort端口号。

K8S从入门到放弃系列-(15)Kubernetes集群Ingress部署

总结:

  1、上面我们创建一个针对于nginx的deployment资源,pod为2个;

  2、为nginx的pod暴露service服务,名称为myapp-svc

  3、通过ingress把nginx暴露出去

这里对于nginx创建的svc服务,其实在实际调度过程中,流量是直接通过ingress然后调度到后端的pod,而没有经过svc服务,svc只是提供一个收集pod服务的作用。

3、Ingress高可用

  上面我们只是解决了集群对外提供服务的功能,并没有对ingress进行高可用的部署,Ingress高可用,我们可以通过修改deployment的副本数来实现高可用,但是由于ingress承载着整个集群流量的接入,所以生产环境中,建议把ingress通过DaemonSet的方式部署集群中,而且该节点打上污点不允许业务pod进行调度,以避免业务应用与Ingress服务发生资源争抢。然后通过SLB把ingress节点主机添为后端服务器,进行流量转发。

##修改mandatory.yaml
### 主要修改pod相关
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
  name: nginx-ingress-controller
  namespace: ingress-nginx
  labels:
    app.kubernetes.io/name: ingress-nginx
    app.kubernetes.io/part-of: ingress-nginx
spec:
  selector:
    matchLabels:
      app.kubernetes.io/name: ingress-nginx
      app.kubernetes.io/part-of: ingress-nginx
  template:
    metadata:
      labels:
        app.kubernetes.io/name: ingress-nginx
        app.kubernetes.io/part-of: ingress-nginx
      annotations:
        prometheus.io/port: "10254"
        prometheus.io/scrape: "true"
    spec:
      serviceAccountName: nginx-ingress-serviceaccount
      hostNetwork: true
      dnsPolicy: ClusterFirstWithHostNet
      nodeSelector:
        vanje/ingress-controller-ready: "true"
      tolerations:
      - key: "node-role.kubernetes.io/master"
        operator: "Equal"
        value: ""
        effect: "NoSchedule"
      containers:
        - name: nginx-ingress-controller
          image: quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.25.0
          args:
            - /nginx-ingress-controller
            - --configmap=$(POD_NAMESPACE)/nginx-configuration
            - --tcp-services-configmap=$(POD_NAMESPACE)/tcp-services
            - --udp-services-configmap=$(POD_NAMESPACE)/udp-services
            - --publish-service=$(POD_NAMESPACE)/ingress-nginx
            - --annotations-prefix=nginx.ingress.kubernetes.io
          securityContext:
            allowPrivilegeEscalation: true
            capabilities:
              drop:
                - ALL
              add:
                - NET_BIND_SERVICE
            # www-data -> 33
            runAsUser: 33
          env:
            - name: POD_NAME
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: POD_NAMESPACE
              valueFrom:
                fieldRef:
                  fieldPath: metadata.namespace
          ports:
            - name: http
              containerPort: 80
            - name: https
              containerPort: 443
          livenessProbe:
            failureThreshold: 3
            httpGet:
              path: /healthz
              port: 10254
              scheme: HTTP
            initialDelaySeconds: 10
            periodSeconds: 10
            successThreshold: 1
            timeoutSeconds: 10
          readinessProbe:
            failureThreshold: 3
            httpGet:
              path: /healthz
              port: 10254
              scheme: HTTP
            periodSeconds: 10
            successThreshold: 1
            timeoutSeconds: 10

 修改参数如下:

  kind: Deployment #修改为DaemonSet
  replicas: 1 #注销此行,DaemonSet不需要此参数
  hostNetwork: true #添加该字段让docker使用物理机网络,在物理机暴露服务端口(80),注意物理机80端口提前不能被占用
  dnsPolicy: ClusterFirstWithHostNet #使用hostNetwork后容器会使用物理机网络包括DNS,会无法解析内部service,使用此参数让容器使用K8S的DNS
  nodeSelector:vanje/ingress-controller-ready: "true" #添加节点标签
  tolerations: 添加对指定节点容忍度

这里我在2台master节点部署(生产环境不要使用master节点,应该部署在独立的节点上),因为我们采用DaemonSet的方式,所以我们需要对2个节点打标签以及容忍度。

## 给节点打标签
[root@k8s-master01 ingress-master]# kubectl label nodes k8s-master02 vanje/ingress-controller-ready=true
[root@k8s-master01 ingress-master]# kubectl label nodes k8s-master03 vanje/ingress-controller-ready=true
## 节点打污点
### master节点我之前已经打过污点,如果你没有打污点,执行下面2条命令。此污点名称需要与yaml文件中pod的容忍污点对应
[root@k8s-master02 ~]# kubectl taint nodes k8s-master02 node-role.kubernetes.io/master=:NoSchedule
[root@k8s-master03 ~]# kubectl taint nodes k8s-master03 node-role.kubernetes.io/master=:NoSchedule

3.2)创建资源

[root@k8s-master01 ingress-master]# kubectl apply -f mandatory.yaml
## 查看资源分布情况
### 可以看到两个ingress-controller已经根据我们选择,部署在2个master节点上
[root@k8s-master01 ingress-master]# kubectl get pod -n ingress-nginx -o wide
NAME                             READY   STATUS    RESTARTS   AGE    IP              NODE           NOMINATED NODE   READINESS GATES
nginx-ingress-controller-298dq   1/1     Running   0          134m   172.16.11.122   k8s-master03   <none>           <none>
nginx-ingress-controller-sh9h2   1/1     Running   0          134m   172.16.11.121   k8s-master02   <none>           <none>

3.3)测试

这里直接使用上面创建的pod及对应svc测试即可,另外注意一点,因为我们创建的ingress-controller采用的时hostnetwork模式,所以无需在创建ingress-svc服务来把端口映射到节点主机上。

## 创建pod及svc
[root@k8s-master01 ingress-master]# kubectl test-ingress-pods.yml
## 创建ingress规则
[root@k8s-master01 ingress-master]# kubectl test-ingress-myapp.yml

在win主机上直接解析,IP地址为k8s-master03/k8s-master02 任意节点ip即可,访问的时候也无需再加端口

K8S从入门到放弃系列-(15)Kubernetes集群Ingress部署