docker笔记26-ingress资源和ingress controller

     Ingress:就是能利用 Nginx(不常用)、Haproxy(不常用)、Traefik(常用)、Envoy(常用) 啥的负载均衡器暴露集群内服务的工具。

     Ingress为您提供七层负载均衡能力,您可以通过 Ingress 配置提供外部可访问的 URL、负载均衡、SSL、基于名称的虚拟主机等。作为集群流量接入层,Ingress 的高可靠性显得尤为重要。

    小知识:我们把k8s里面的pod服务发布到集群外部,可以用ingress,也可以用NodePort。

    externalLB:外部的负载均衡器

    service site:只是用来给pod分组归类的。


[root@master manifests]# kubectl explain ingress

    创建名称空间:


[root@master manifests]# kubectl create namespace ingress-nginx
namespace/dev created
[root@master manifests]# kubectl get ns
NAME          STATUS    AGE
default       Active    17d
ingress-nginx   Active    8s
kube-public   Active    17d
kube-system   Active    17d

    访问 https://github.com/kubernetes/ingress-nginx,进入deploy目录,里面就有我们要用的yaml文件。


 各文件的作用:
configmap.yaml:提供configmap可以在线更行nginx的配置
default-backend.yaml:提供一个缺省的后台错误页面 404
namespace.yaml:创建一个独立的命名空间 ingress-nginx
rbac.yaml:创建对应的role rolebinding 用于rbac
tcp-services-configmap.yaml:修改L4负载均衡配置的configmap
udp-services-configmap.yaml:修改L4负载均衡配置的configmap
with-rbac.yaml:有应用rbac的nginx-ingress-controller组件

    访问https://kubernetes.github.io/ingress-nginx/deploy/#generice-deployment,里面是ingress的部署文档


[root@master ~]# mkdir  ingress-nginx
[root@master ~]# cd ingress-nginx

 部署ingress方法一(分步部署):

    下载如下配置文件:


[root@master ingress-nginx]# for file in namespace.yaml configmap.yaml rbac.yaml tcp-services-configmap.yaml with-rbac.yaml udp-services-configmap.yaml; do wget  https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/$file; done


[root@master ingress-nginx]# ls
configmap.yaml  namespace.yaml  rbac.yaml  tcp-services-configmap.yaml  udp-services-configmap.yaml  with-rbac.yaml

1、创建名称空间:


[root@master ingress-nginx]# kubectl apply -f namespace.yaml 
namespace/ingress-nginx configured

2、把剩下的ymal文件全应用


[root@master ingress-nginx]# kubectl  apply -f ./
configmap/nginx-configuration created
namespace/ingress-nginx configured
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
configmap/tcp-services created
configmap/udp-services created
deployment.extensions/nginx-ingress-controller created

[root@master ingress-nginx]# kubectl get pods -n ingress-nginx -w
NAME                                        READY     STATUS              RESTARTS   AGE
default-http-backend-6586bc58b6-qd9fk       0/1       running    0          4m
nginx-ingress-controller-6bd7c597cb-zcbbz   0/1       running   0          1m

    可以看到ingress-nginx名称空间里面有两个pod都处于running状态

部署ingress方法二(一键部署):

只下载mandatory.yaml文件,因为这个文件里面包含了上面所有yaml文件里面的内容。这是一键部署。


[root@master ingress-nginx]# wget 
https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/mandatory.yaml

[root@master ingress-nginx]#kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/mandatory.yaml

[root@master ~]# kubectl get pods -n ingress-nginx -w  -o wide
NAME                                        READY     STATUS    RESTARTS   AGE       IP            NODE      NOMINATED NODE
default-http-backend-6586bc58b6-qd9fk       1/1       Running   0          11h       10.244.1.95   node1     
nginx-ingress-controller-6bd7c597cb-jlqzp   1/1       Running   3          11h       10.244.1.96   node1     

     可以看到ingress-nginx名称空间里面有两个pod都处于running状态

安装service-nodeport

    上面我们把ingress-nginx部署到了1号node上。接下来我们还需要部署一个service-nodeport服务,才能实现把集群外部流量接入到集群中来。


[root@master ingress]# wget https://raw.githubusercontent.com/kubernetes/ingress-nginx/master/deploy/provider/baremetal/service-nodeport.yaml

    我们为了不让service nodeport自动分配端口,我们自己指定一下nodeport,修改文件中加两个nodePort参数,最终如下:


[root@master ingress]# cat 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
    nodePort: 30080
  - name: https
    port: 443
    targetPort: 443
    protocol: TCP
    nodePort: 30443
  selector:
    app.kubernetes.io/name: ingress-nginx
    app.kubernetes.io/part-of: ingress-nginx

[root@master ingress]# kubectl apply -f service-nodeport.yaml 
service/ingress-nginx created

[root@master ingress]# kubectl get svc -n ingress-nginx
NAME                   TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                      AGE
default-http-backend   ClusterIP   10.110.74.183           80/TCP                       12h
ingress-nginx          NodePort    10.102.78.188           80:30080/TCP,443:30443/TCP   2m

    上面我看到80对应30080,,43对应30443

    我们直接通过node1节点的ip就可以访问到应用:


[root@master ingress]# curl  http://172.16.1.101:30080
default backend - 404


定义myapp service


[root@master manifests]# mkdir /root/manifests/ingress

[root@master ~]# kubectl explain service.spec.ports

[root@master ingress]# cat deploy-demo.yaml 
apiVersion: v1
kind: Service
#必须设置为无头service
metadata:
 name: myapp
 namespace: default
spec:
  selector:
    app: myapp
    release: canary
  ports:
  - name: http
    targetPort: 80   #这是容器port
    port: 80  #这是service port
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: myapp-deploy
  namespace: default
spec:
  replicas: 2
  selector: #标签选择器
    matchLabels: #匹配的标签为
      app: myapp
      release: canary
  template:
    metadata:
      labels:
        app: myapp #和上面的myapp要匹配
        release: canary
    spec:
      containers:
      - name: myapp
        image: ikubernetes/myapp:v1
        ports:
        - name: http
          containerPort: 80

[root@master ingress]# kubectl apply -f deploy-demo.yaml 
service/myapp created
deployment.apps/myapp-deploy unchanged

[root@master ingress]# kubectl get svc
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)    AGE
kubernetes   ClusterIP   10.96.0.1               443/TCP    17d
myapp        ClusterIP   10.108.177.62           80/TCP     1m

[root@master ingress]# kubectl get pods
NAME                            READY     STATUS             RESTARTS   AGE
myapp-deploy-69b47bc96d-79fqh   1/1       Running            0          1d
myapp-deploy-69b47bc96d-tc54k   1/1       Running            0          1d


把myapp service通过ingress发布出去

     下面我们再定义一个清单文件,把myapp应用通过Ingress(相当于nginx的反向代理功能)发布出去:


[root@master ingress]# cat ingress-myapp.yaml 
apiVersion: extensions/v1beta1 
kind: Ingress
metadata:
  name: ingress-myapp
  namespace: default #要和deployment和要发布的service处于同一个名称空间
  annotations: #这个注解说明我们要用到的ingress-controller是nginx,而不是traefic,enjoy
    kubernetes.io/ingress.class: "nginx"
spec:
  rules:
  - host: myapp.zhixin.com #表示访问这个域名,就会转发到后端myapp管理的pod上的服务:
    http:
      paths: 
      - path:
        backend:
          serviceName: myapp
          servicePort: 80

[root@master ingress]# kubectl apply -f ingress-myapp.yaml 
ingress.extensions/ingress-myapp created

[root@master ingress]# kubectl get ingress
NAME            HOSTS              ADDRESS   PORTS     AGE
ingress-myapp   myapp.zhixin.com             80        8m

[root@master ingress]# kubectl describe ingress

[root@master ingress]# kubectl get pods -n ingress-nginx 
NAME                                        READY     STATUS    RESTARTS   AGE
default-http-backend-6586bc58b6-qd9fk       1/1       Running   0          12h
nginx-ingress-controller-6bd7c597cb-jlqzp   1/1       Running   3          12h

    进入ingress-controller交互式命令行里面,可以清晰的看到nginx是怎么反向代理我们myapp.zhixin.com的:


[root@master ingress]# kubectl exec -n ingress-nginx -it nginx-ingress-controller-6bd7c597cb-jlqzp -- /bin/sh
$ cat nginx.conf
## start server myapp.zhixin.com
server {
server_name myapp.zhixin.com ;
listen 80;
set $proxy_upstream_name "-";
location / {
set $namespace      "default";
set $ingress_name   "ingress-myapp";
set $service_name   "myapp";
set $service_port   "80";
........

    测试,下面我们把myapp.zhixin.com域名解析到node1 ip 172.16.1.101上。


[root@master ingress]# curl myapp.zhixin.com:30080
Hello MyApp | Version: v1 | Pod Name


把tomcat service通过ingress发布出去(新例子)


[root@master ingress]# cat tomcat-demo.yaml 
apiVersion: v1
kind: Service
#必须设置为无头service
metadata:
 name: tomcat
 namespace: default
spec:
  selector:
    app: tomcat
    release: canary
  ports:
  - name: http
    targetPort: 8080   #这是容器port
    port: 8080  #这是service port
  - name: ajp
    targetPort: 8009
    port: 8009
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: tomcat-deploy
  namespace: default
spec:
  replicas: 2
  selector: #标签选择器
    matchLabels: #匹配的标签为
      app: tomcat
      release: canary
  template:
    metadata:
      labels:
        app: tomcat #和上面的myapp要匹配
        release: canary
    spec:
      containers:
      - name: tomcat
        image: tomcat:8.5.34-jre8-alpine  #在https://hub.docker.com/r/library/tomcat/tags/上面找
        ports:
        - name: http
          containerPort: 8080
        - name: ajp
          containerPort: 8009

[root@master ingress]# kubectl apply -f tomcat-demo.yaml 
service/tomcat created
deployment.apps/tomcat-deploy created

[root@master ingress]# kubectl get svc
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)             AGE
kubernetes   ClusterIP   10.96.0.1               443/TCP             17d
tomcat       ClusterIP   10.109.76.87            8080/TCP,8009/TCP   1m

[root@master ingress]# kubectl get pods
NAME                             READY     STATUS             RESTARTS   AGE
tomcat-deploy-64c4d54df4-68sk8   1/1       Running            0          51s
tomcat-deploy-64c4d54df4-7b58g   1/1       Running            0          51s



[root@master ingress]# cat ingress-tomcat.yaml 
apiVersion: extensions/v1beta1 
kind: Ingress
metadata:
  name: ingress-tomcat
  namespace: default #要和deployment和要发布的service处于同一个名称空间
  annotations: #这个注解说明我们要用到的ingress-controller是nginx,而不是traefic,enjoy
    kubernetes.io/ingress.class: "nginx"
spec:
  rules:
  - host: tomcat.zhixin.com #表示访问这个域名,就会转发到后端myapp管理的pod上的服务:
    http:
      paths: 
      - path:
        backend:
          serviceName: tomcat
          servicePort: 8080

[root@master ingress]# kubectl apply -f ingress-tomcat.yaml 
ingress.extensions/ingress-myapp configured

[root@master ingress]# kubectl get ingress
NAME             HOSTS               ADDRESS   PORTS     AGE
ingress-tomcat   tomcat.zhixin.com             80        11s

[root@master ingress]# kubectl describe ingress ingress-tomcat
Name:             ingress-tomcat
Namespace:        default
Address:          
Default backend:  default-http-backend:80 ()
Rules:
  Host               Path  Backends
  ----               ----  --------
  tomcat.zhixin.com  
                        tomcat:8080 ()
Annotations:
  kubernetes.io/ingress.class:  nginx
Events:
  Type    Reason  Age   From                      Message
  ----    ------  ----  ----                      -------
  Normal  CREATE  1m    nginx-ingress-controller  Ingress default/ingress-tomcat

    把tomcat.zhixin.com解析到node1上节点物理ip(我的是172.16.1.101)

    测试,可以看到tomcat欢迎界面:


[root@master ingress]# curl tomcat.zhixin.com:30080


使用https访问(新例子)

1、先做个自签的证书(我们这里不演示CA的例子)


[root@master ingress]# openssl genrsa -out tls.key 2048

[root@master ingress]# openssl req -new -x509 -key tls.key  -out tls.crt -subj /C=CN/ST=Beijing/O=DevOps/CN=tomcat.zhixin.com

2、通过secret把证书注入到pod中。


[root@master ingress]# kubectl create secret tls tomcat-infress-secret --cert=tls.crt --key=tls.key 
secret/tomcat-infress-secret created

[root@master ingress]# kubectl get secret
NAME                    TYPE                                  DATA      AGE
default-token-5r85r     kubernetes.io/service-account-token   3         17d
tomcat-ingress-secret   kubernetes.io/tls                     2         41s

[root@master ingress]# kubectl describe secret tomcat-ingress-secret 
Name:         tomcat-ingress-secret
Namespace:    default
Labels:       
Annotations:  
Type:  kubernetes.io/tls
Data
====
tls.crt:  1245 bytes
tls.key:  1679 bytes

3、配置ingress为tls方式


[root@master ingress]# cat ingress-tomcat-tls.yaml 
apiVersion: extensions/v1beta1 
kind: Ingress
metadata:
  name: ingress-tomcat-tls
  namespace: default #要和deployment和要发布的service处于同一个名称空间
  annotations: #这个注解说明我们要用到的ingress-controller是nginx,而不是traefic,enjoy
    kubernetes.io/ingress.class: "nginx"
spec:
  tls:
  - hosts:
    - tomcat.zhixin.com
    secretName: tomcat-ingress-secret #kubectl get secret命令查到的名字
  rules:
  - host: tomcat.zhixin.com #表示访问这个域名,就会转发到后端myapp管理的pod上的服务:
    http:
      paths: 
      - path:
        backend:
          serviceName: tomcat
          servicePort: 8080
[root@master ingress]# kubectl get ingress
NAME                 HOSTS               ADDRESS   PORTS     AGE
ingress-tomcat       tomcat.zhixin.com             80        2h
ingress-tomcat-tls   tomcat.zhixin.com             80, 443   3m

[root@master ingress]# kubectl describe ingress ingress-tomcat-tls  
Name:             ingress-tomcat-tls
Namespace:        default
Address:          
Default backend:  default-http-backend:80 ()
TLS:
  tomcat-ingress-secret terminates tomcat.zhixin.com
Rules:
  Host               Path  Backends
  ----               ----  --------
  tomcat.zhixin.com  
                        tomcat:8080 ()
Annotations:
  kubectl.kubernetes.io/last-applied-configuration:  {"apiVersion":"extensions/v1beta1","kind":"Ingress","metadata":{"annotations":{"kubernetes.io/ingress.class":"nginx"},"name":"ingress-tomcat-tls","namespace":"default"},"spec":{"rules":[{"host":"tomcat.zhixin.com","http":{"paths":[{"backend":{"serviceName":"tomcat","servicePort":8080},"path":null}]}}],"tls":[{"hosts":["tomcat.zhixin.com"],"secretName":"tomcat-ingress-secret"}]}}
  kubernetes.io/ingress.class:  nginx
Events:
  Type    Reason  Age   From                      Message
  ----    ------  ----  ----                      -------
  Normal  CREATE  4m    nginx-ingress-controller  Ingress default/ingress-tomcat-tls

4、连如ingress-controller查看nginx.conf的配置


[root@master ingress]# kubectl get pods -n ingress-nginx
NAME                                        READY     STATUS    RESTARTS   AGE
default-http-backend-6586bc58b6-qd9fk       1/1       Running   0          16h
nginx-ingress-controller-6bd7c597cb-jlqzp   1/1       Running   3          16h

[root@master ingress]# kubectl exec -n ingress-nginx -it nginx-ingress-controller-6bd7c597cb-jlqzp -- /bin/sh
$ $ cat nginx.conf
	## start server tomcat.zhixin.com
	server {
		server_name tomcat.zhixin.com ;
		
		listen 80;
		
		set $proxy_upstream_name "-";
		
		listen 443  ssl http2;

    看到有listen 443了。

5、测试https


[root@master ingress]# curl https://tomcat.zhixin.com:30443

部署方法三(利用ingress 的80端口)

    前面两种部署方法,是用node ip + 非80端口,访问k8s集群内部的服务。可是,我们实际生产中更希望的是node ip + 80端口的方式,访问k8s集群内的服务。我感觉这个方法最好,下面就就介绍这个方法。

    这部分内容参考的是博文http://blog.51cto.com/devingeng/2149377

    下载地址 

https://github.com/kubernetes/ingress-nginx/archive/nginx-0.11.0.tar.gz


ingress-nginx文件位于deploy目录下,各文件的作用:
configmap.yaml:提供configmap可以在线更行nginx的配置
default-backend.yaml:提供一个缺省的后台错误页面 404
namespace.yaml:创建一个独立的命名空间 ingress-nginx
rbac.yaml:创建对应的role rolebinding 用于rbac
tcp-services-configmap.yaml:修改L4负载均衡配置的configmap
udp-services-configmap.yaml:修改L4负载均衡配置的configmap
with-rbac.yaml:有应用rbac的nginx-ingress-controller组件

    修改with-rbac.yaml


apiVersion: extensions/v1beta1
kind: Daemonset
metadata:
  name: nginx-ingress-controller
  namespace: ingress-nginx 
spec:
  selector:
    matchLabels:
      app: ingress-nginx
  template:
    metadata:
      labels:
        app: ingress-nginx
      annotations:
        prometheus.io/port: '10254'
        prometheus.io/scrape: 'true'
    spec:
      serviceAccountName: nginx-ingress-serviceaccount      
      hostNetwork: true
      containers:
        - name: nginx-ingress-controller
          image: quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.11.0
          args:
            - /nginx-ingress-controller
            - --default-backend-service=$(POD_NAMESPACE)/default-http-backend
            - --configmap=$(POD_NAMESPACE)/nginx-configuration
            - --tcp-services-configmap=$(POD_NAMESPACE)/tcp-services
            - --udp-services-configmap=$(POD_NAMESPACE)/udp-services
            - --annotations-prefix=nginx.ingress.kubernetes.io
          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: 1
          readinessProbe:
            failureThreshold: 3
            httpGet:
              path: /healthz
              port: 10254
              scheme: HTTP
            periodSeconds: 10
            successThreshold: 1
            timeoutSeconds: 1
      nodeSelector:
        custom/ingress-controller-ready: "true"

需要修改的地方:

kind: DaemonSet:官方原始文件使用的是deployment,replicate 为 1,这样将会在某一台节点上启动对应的nginx-ingress-controller pod。外部流量访问至该节点,由该节点负载分担至内部的service。测试环境考虑防止单点故障,改为DaemonSet然后删掉replicate ,配合亲和性部署在制定节点上启动nginx-ingress-controller pod,确保有多个节点启动nginx-ingress-controller pod,后续将这些节点加入到外部硬件负载均衡组实现高可用性。

hostNetwork: true:添加该字段,暴露nginx-ingress-controller pod的服务端口(80)

nodeSelector: 增加亲和性部署,有custom/ingress-controller-ready 标签的节点才会部署该DaemonSet

为需要部署nginx-ingress-controller的节点设置lable


kubectl label nodes vmnode2 custom/ingress-controller-ready=true
kubectl label nodes vmnode3 custom/ingress-controller-ready=true
kubectl label nodes vmnode4 custom/ingress-controller-ready=true

加载yaml文件


kubectl apply  -f namespace.yaml
kubectl apply -f default-backend.yaml
kubectl apply -f configmap.yaml
kubectl apply -f tcp-services-configmap.yaml
kubectl apply -f udp-services-configmap.yaml
kubectl apply -f rbac.yaml
kubectl apply -f with-rbac.yaml

查看pod是否正常创建

##下载镜像可能会比较慢,等待一会所有pod都是Running状态,按Ctrl + c 退出


[root@vmnode1 deploy]# kubectl get pods --namespace=ingress-nginx  --watch
NAME                                    READY     STATUS    RESTARTS   AGEdefault-
http-backend-6c59748b9b-hc8q9   1/1       Running   0          6m
nginx-ingress-controller-7fmlp          1/1       Running   1          13d
nginx-ingress-controller-j95fb          1/1       Running   1          13d
nginx-ingress-controller-ld2jw          1/1       Running   1          13d

测试ingress 

创建一个tomcat的Service


[root@k8s-master1 test]# cat mytomcat.yaml 
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: mytomcat
spec:
  replicas: 2
  template:
    metadata:
      labels:
        run: mytomcat
    spec:
      containers:
      - name: mytomcat
        image: tomcat
        ports:
        - containerPort: 8080
---
apiVersion: v1
kind: Service
metadata:
  name: mytomcat
  labels:
    run: mytomcat
spec:
  type: NodePort
  ports:
  - port: 8080
    targetPort: 8080
  selector:
    run: mytomcat

[root@k8s-master1 test]# kubectl apply -f mytomcat.yaml

    配置ingress转发文件:


[root@k8s-master1 test]# cat test-ingress.yaml 
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: test-ingress
  namespace: default
spec:
  rules:
  - host: test.zhixin.com
    http:
      paths:
      - path: /
        backend:
          serviceName: mytomcat
          servicePort: 8080

host: 对应的域名 

path: url上下文 

backend:后向转发 到对应的 serviceName: servicePort:


[root@k8s-master1 test]# kubectl apply -f test-ingress.yaml 
ingress.extensions/test-ingress created

nginx-ingress-controller运行在node1和nod2两个节点上。如果网络中有dns服务器,在dns中把这两个域名映射到nginx-ingress-controller运行的任意一个节点上,如果没有dns服务器只能修改host文件了。 

正规的做法是在node1和node2这两个节点上安装keepalive,生成一个vip。在dns上把域名和vip做映射。

我这里直接在node1节点上操作了:

我这里node1节点的ip是172.16.22.201;node2节点的ip是172.16.22.202


[root@k8s-master1 test]# echo " 172.16.22.201 test.zhixin.com" >> /etc/hosts
[root@k8s-master1 test]# echo "172.16.22.202 test.zhixin.com" >> /etc/hosts

    然后访问测试:


    看到,我们把域名test.zhixin.com绑定到Node节点的ip补上,然后我们直接访问http://test.zhixin.com,就能访问到k8s集群里面的pod服务。















来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/28916011/viewspace-2214747/,如需转载,请注明出处,否则将追究法律责任。

转载于:http://blog.itpub.net/28916011/viewspace-2214747/

你可能感兴趣的:(docker笔记26-ingress资源和ingress controller)