十九 服务发布Ingress
(一) Ingress Nginx Controller安装
首先安装Helm
# wget https://get.helm.sh/helm-v3.6.3-linux-amd64.tar.gz
# tar -zxvf helm-v3.0.0-linux-amd64.tar.gz
# mv linux-amd64/helm /usr/local/bin/helm
下载 Ingress Nginx Controller安装包
# helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
# helm repo update
# helm pull ingress-nginx/ingress-nginx --version 4.0.1
更改对应的配置
# tar xf ingress-nginx-4.0.1.tgz
#cd ingress-nginx
#vim values.yaml
需要修改的位置
- Controller和admissionWebhook的镜像地址,需要将公网镜像同步至公司内网镜像仓库(和课程不一致的版本,需要自行同步gcr镜像,一样的版本可以直接用registry.cn-beijing.aliyuncs.com/dotbalo/kube-webhook-certgen:v1.0和registry.cn-beijing.aliyuncs.com/dotbalo/controller:v1.0.0)
- 镜像的digest值注释
- hostNetwork设置为true
- dnsPolicy设置为 ClusterFirstWithHostNet
- NodeSelector添加ingress: "true"部署至指定节点
- 类型更改为kind: DaemonSet
-
将ingress nginx设置为默认的ingressClass
- 部署ingress,给需要部署ingress的节点上打标签
# kubectl label node k8s-node02 ingress=true
# kubectl create ns ingress-nginx
# helm install ingress-nginx -n ingress-nginx .
(二) Ingress Nginx入门使用
创建一个用于学习Ingress的Namespace,之后所有的操作都在此Namespace进行:
# kubectl create ns study-ingress
namespace/study-ingress created
创建一个简单的Nginx模拟Web服务:
# kubectl create deploy nginx --image=registry.cn-beijing.aliyuncs.com/dotbalo/nginx:1.15.12 -n study-ingress
然后创建该Web容器的Service:
# kubectl expose deploy nginx --port 80 -n study-ingress
之后创建Ingress指向上面创建的Service:
# vim web-ingress.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: nginx-ingress
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: nginx.test.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
如果apiVersion是networking.k8s.io/v1beta1,对应的配置如下:
apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
name: nginx-ingress
namespace: study-ingress
spec:
rules:
- host: nginx.test.com
http:
paths:
- backend:
serviceName: nginx
servicePort: 80
path: /
pathType: ImplementationSpecific
提示:本章内容均采用networking.k8s.io/v1版本创建Ingress资源,如果Kubernetes版本低于1.19,可以使用networking.k8s.io/v1beta1替代,配置可以参考上述的networking.k8s.io/v1beta1,只有backend配置不一样。
创建该Ingress:
# kubectl create -f web-ingress.yaml
将域名nginx.test.com即可访问Web服务器,如下图所示:
![](https://img.haomeiwen.com/i20896689/d366d9a5c43192f3.png)
(三) Ingress Nginx域名重定向Redirect
在Nginx作为代理服务器时,Redirect可用于域名的重定向,比如访问old.com被重定向到new.com。Ingress可以更简单的实现Redirect功能,接下来用nginx.redirect.com作为旧域名,baidu.com作为新域名进行演示:
# vim redirect.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
nginx.ingress.kubernetes.io/permanent-redirect: https://www.baidu.com
name: nginx-redirect
namespace: study-ingress
spec:
rules:
- host: nginx.redirect.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
使用curl访问域名nginx.redirect.com,可以看到301(请求被重定向的返回值):
# curl -I nginx.redirect.com
HTTP/1.1 301 Moved Permanently
...
Location: https://www.baidu.com
(四) Ingress Nginx前后端分离Rewrite
创建一个应用模拟后端服务:
# kubectl create deploy backend-api --image=registry.cn-beijing.aliyuncs.com/dotbalo/nginx:backend-api -n study-ingress
deployment.apps/backend-api created
创建Service暴露该应用
# kubectl expose deploy backend-api --port 80 -n study-ingress
查看该Service的地址,并且通过/api-a访问测试:
# kubectl get svc -n study-ingress
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
backend-api ClusterIP 192.168.60.190 <none> 80/TCP 5s
nginx ClusterIP 192.168.170.24 <none> 80/TCP 3h27m
# curl 192.168.60.190/api-a
<html>
<head><title>404 Not Found</title></head>
<body>
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.15.12</center>
</body>
</html>
直接访问根路径是可以的:
# curl 192.168.60.190
<h1> backend for ingress rewrite </h1>
<h2> **Path: /api-a** </h2>
通过Ingress Nginx的Rewrite功能,将/api-a重写为“/”,配置示例如下:
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
nginx.ingress.kubernetes.io/rewrite-target: /$2
name: backend-api
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: nginx.test.com
http:
paths:
- backend:
service:
name: backend-api
port:
number: 80
path: /api-a(/|$)(.*)
pathType: ImplementationSpecific
再次访问nginx.test.com/api-a即可访问到后端服务:
![](https://img.haomeiwen.com/i20896689/95f410d34e7cd20b.png)
(五) Ingress Nginx错误代码重定向
1.通过Helm进行更改
修改values.yaml如下图所示位置:
![](https://img.haomeiwen.com/i20896689/e8a6165013941a54.png)
更新ConfigMap:
config:
apiVersion: v1
client_max_body_size: 20m
custom-http-errors: "404,415,503"
更新Release:
# helm upgrade ingress-nginx -n ingress-nginx .
更新后Pod会自动重启,并且会创建一个defaultbackend:
# kubectl get po -n ingress-nginx
NAME READY STATUS RESTARTS AGE
ingress-nginx-controller-pdjvh 1/1 Running 0 95s
ingress-nginx-defaultbackend-79d64fb85f-kxlnc 1/1 Running 0 109s
更新完成以后访问一个不存在的页面,比如之前定义的nginx.test.com。 访问一个不存在的页面123,就会跳转到Error Server中的页面:
# curl nginx.test.com/123
default backend - 404
使用浏览器访问,会出现如下图所示的页面:
![](https://img.haomeiwen.com/i20896689/0a193ba311ac7fac.png)
(六) Ingress Nginx SSL
生产环境对外的服务,一般需要配置https协议,使用Ingress也可以非常方便的添加https的证书。
由于我们是学习环境,并没有权威证书,所以需要使用OpenSSL生成一个测试证书。如果是生产环境,证书为在第三方公司购买的证书,无需自行生成:
# openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout tls.key -out tls.crt -subj "/CN=nginx.test.com"
Generating a 2048 bit RSA private key
.....+++
............................+++
writing new private key to 'tls.key'
-----
# kubectl create secret tls ca-secret --cert=tls.crt --key=tls.key -n study-ingress
secret/ca-secret created
配置Ingress添加TLS配置:
# vim ingress-ssl.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
creationTimestamp: null
name: nginx-ingress
namespace: study-ingress
# annotations:
# kubernetes.io/ingress.class: nginx
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: nginx.test.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
tls:
- hosts:
- nginx.test.com
secretName: ca-secret
可以看到Ingress添加TLS配置也非常简单,只需要在spec下添加一个tls字段即可:
- hosts:证书所授权的域名列表
- secretName:证书的Secret名字
- ingressClassName: ingress class的名字,1.22+需要配置
接下来更新该Ingress即可:
# kubectl apply -f ingress-ssl.yaml
ingress.networking.k8s.io/nginx-ingress configured
使用curl进行测试,域名已经被重定向到https:
# curl http://nginx.test.com -I
HTTP/1.1 308 Permanent Redirect
...
Location: https://nginx.test.com
使用浏览器访问,会自动跳转到https,如下图所示:
![](https://img.haomeiwen.com/i20896689/ee05b75ab6c52a9c.png)
(七) Ingress Nginx匹配请求头
首先部署移动端应用:
# kubectl create deploy phone --image=registry.cn-beijing.aliyuncs.com/dotbalo/nginx:phone -n study-ingress
# kubectl expose deploy phone --port 80 -n study-ingress
Ingress实例也可以通过kubectl create进行创建,只需要一条命令即可:
# kubectl create ingress phone --rule=m.test.com/*=phone:80 -n study-ingress
部署电脑端应用:
# kubectl create deploy laptop --image=registry.cn-beijing.aliyuncs.com/dotbalo/nginx:laptop -n study-ingress
deployment.apps/laptop created
# kubectl expose deploy laptop --port 80 -n study-ingress
service/laptop exposed
# kubectl get po -n study-ingress -l app=laptop
NAME READY STATUS RESTARTS AGE
laptop-664b565969-hpxlh 1/1 Running 0 3m22s
之后创建电脑端的Ingress,注意Ingress annotations的 nginx.ingress.kubernetes.io/server-snippet配置。Snippet配置是专门用于一些复杂的Nginx配置,和Nginx配置通用。匹配移动端实例如下:
# vim laptop-ingress.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
kubernetes.io/ingress.class: nginx
nginx.ingress.kubernetes.io/server-snippet: |
set $agentflag 0;
if ($http_user_agent ~* "(Android|iPhone|Windows Phone|UC|Kindle)" ){
set $agentflag 1;
}
if ( $agentflag = 1 ) {
return 301 http://m.test.com;
}
name: laptop
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: test.com
http:
paths:
- backend:
service:
name: laptop
port:
number: 80
path: /
pathType: ImplementationSpecific
首先通过浏览器访问test.com,可以看到页面是Laptop:
![](https://img.haomeiwen.com/i20896689/15d8adc865c8c957.png)
接下来使用浏览器的开发者工具将终端类型改为iPhone,或者直接用iPhone手机访问(线上业务一般配置的都有DNS,可以直接解析域名,测试环境可能需要自己单独配置),如下图所示:
![](https://img.haomeiwen.com/i20896689/5f1263fac29a0705.png)
刷新页面会自动跳转至m.test.com,如下图所示:
![](https://img.haomeiwen.com/i20896689/da58ce4a8462fec6.png)
(八) Ingress Nginx基本认证
有些网站可能需要通过密码来访问,对于这类网站可以使用Nginx的basic-auth设置密码访问,具体方法如下,由于需要使用htpasswd工具,所以需要安装httpd:
# yum install httpd -y
使用htpasswd创建foo用户的密码:
# htpasswd -c auth foo
New password:
Re-type new password:
Adding password for user foo
# cat auth
foo:$apr1$okma2fx9$hdTJ.KFmi4pY9T6a2MjeS1
基于之前创建的密码文件创建Secret:
# kubectl create secret generic basic-auth --from-file=auth -n study-ingress
secret/basic-auth created
创建包含密码认证的Ingress:
# vim ingress-with-auth.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
# kubernetes.io/ingress.class: nginx
nginx.ingress.kubernetes.io/auth-realm: Please Input Your Username and Password
nginx.ingress.kubernetes.io/auth-secret: basic-auth
nginx.ingress.kubernetes.io/auth-type: basic
name: ingress-with-auth
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: auth.test.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
- nginx.ingress.kubernetes.io/auth-type:认证类型,可以是basic和digest
- nginx.ingress.kubernetes.io/auth-secret:密码文件的Secret名称
- nginx.ingress.kubernetes.io/auth-realm:需要密码认证的消息提醒
创建该Ingress,并访问测试:
![](https://img.haomeiwen.com/i20896689/f6839f5a7415cf5f.png)
输入密码后即可进入到页面,在此不再演示。
(九) Ingress Nginx黑/白名单
1.黑名单
配置黑名单禁止某一个或某一段IP,需要在Nginx Ingress的ConfigMap中配置,比如将192.168.10.130(多个配置逗号分隔)添加至黑名单:
# vim values.yaml
config:
block-cidrs: 192.168.10.130
滚动更新Nginx Ingress:
# helm upgrade ingress-nginx -n ingress-nginx .
使用192.168.10.130主机再次访问,发现该IP已经被禁止:
# curl auth.test.com -I
HTTP/1.1 403 Forbidden
...
Connection: keep-alive
2.白名单
白名单表示只允许某个IP可以访问,直接在yaml文件中配置即可(也可以通过ConfigMap配置),比如只允许192.168.10.128访问,只需要添加一个nginx.ingress.kubernetes.io/whitelist-source-range 注释即可:
# vim auth-whitelist.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
# kubernetes.io/ingress.class: nginx # K8s >= 1.22+ 使用ingressClassName替代
nginx.ingress.kubernetes.io/auth-realm: Please Input Your Username and Password
nginx.ingress.kubernetes.io/auth-secret: basic-auth
nginx.ingress.kubernetes.io/auth-type: basic
nginx.ingress.kubernetes.io/whitelist-source-range: 192.168.10.128
name: ingress-with-auth
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: auth.test.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
更新该Ingress:
# kubectl apply -f test-auth.yaml
ingress.networking.k8s.io/ingress-with-auth configured
此时192.168.10.128主机是可以访问的:
# curl auth.test.com -I
HTTP/1.1 401 Unauthorized
其它IP访问被禁止:
# curl auth.test.com -I
HTTP/1.1 403 Forbidden
(十) Ingress Nginx速率限制
有时候可能需要限制速率以降低后端压力,或者限制单个IP每秒的访问速率防止攻击。此时可以使用Nginx的rate limit进行配置。
首先没有加速率限制,使用ab进行访问,Failed为0:
# ab -c 10 -n 100 http://auth.test.com/ | grep requests
Complete requests: 100
Failed requests: 0
添加速率限制,限制只能有一个连接,只需要添加nginx.ingress.kubernetes.io/limit-connections为1即可:
# vim auth-rate-limit.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
kubernetes.io/ingress.class: nginx
nginx.ingress.kubernetes.io/auth-realm: Please Input Your Username and Password
nginx.ingress.kubernetes.io/auth-secret: basic-auth
nginx.ingress.kubernetes.io/auth-type: basic
nginx.ingress.kubernetes.io/limit-connections: "1"
name: ingress-with-auth
namespace: study-ingress
spec:
ingressClassName: nginx # for k8s >= 1.22+
rules:
- host: auth.test.com
http:
paths:
- backend:
service:
name: nginx
port:
number: 80
path: /
pathType: ImplementationSpecific
再次使用ab测试,Failed为67:
[root@k8s-master01 5.10]# ab -c 10 -n 100 http://auth.test.com/ | grep requests:
Complete requests: 100
Failed requests: 67
还有很多其它方面的限制,常用的配置如下:
#限制每秒的连接,单个IP:
nginx.ingress.kubernetes.io/limit-rps
#限制每分钟的连接,单个IP:
nginx.ingress.kubernetes.io/limit-rpm
#限制客户端每秒传输的字节数,单位为K,需要开启proxy-buffering:
nginx.ingress.kubernetes.io/limit-rate
# 速率限制白名单
nginx.ingress.kubernetes.io/limit-whitelist
(十一) 使用Nginx实现灰度/金丝雀发布
1.创建v1版本
首先创建模拟Production(生产)环境的Namespace和服务:
# kubectl create ns production
namespace/production created
# kubectl create deploy canary-v1 --image=registry.cn-beijing.aliyuncs.com/dotbalo/canary:v1 -n production
# kubectl expose deploy canary-v1 --port 8080 -n production
# kubectl create ingress canary-v1 --rule=canary.com/*=canary-v1:8080 -n production
使用浏览器访问该服务,可以看到Canary v1的页面:
![](https://img.haomeiwen.com/i20896689/2d61678c757d1eb0.png)
2.创建v2版本
接下来创建v2版本,充当灰度环境:
# kubectl create ns canary
namespace/canary created
创建v2版本的应用和Service:
# kubectl create deploy canary-v2 --image=registry.cn-beijing.aliyuncs.com/dotbalo/canary:v2 -n canary
# kubectl expose deploy canary-v2 --port 8080 -n canary
待程序启动完成后,通过Service访问该服务,会返回Canary v2:
# kubectl get svc -n canary
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
canary-v2 ClusterIP 192.168.181.120 <none> 8080/TCP 89s
# curl 192.168.181.120:8080
<h1>Canary v2</h1>
接下来通过Ingress控制流量。
3.Canary版本切入部分流量
创建v2版本的Ingress时,需要添加两个注释,一个是nginx.ingress.kubernetes.io/canary,表明是灰度环境,nginx.ingress.kubernetes.io/canary-weight表明切多少流量到该环境,本示例为10%:
# vim canary-v2.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
nginx.ingress.kubernetes.io/canary: "true"
nginx.ingress.kubernetes.io/canary-weight: "10"
name: canary-v2
namespace: canary
spec:
ingressClassName: nginx # for k8s >=1.22+
rules:
- host: canary.com
http:
paths:
- backend:
service:
name: canary-v2
port:
number: 8080
path: /
pathType: ImplementationSpecific
此时通过nginx.ingress.kubernetes.io/canary-weight: "10"设置的权重是10,即v1:v2为9:1。
4.测试灰度发布
接下来使用Ruby脚本进行测试,此脚本会输出v1和v2的访问次数比值:
# vim test-canary.rb
counts = Hash.new(0)
100.times do
output = `curl -s canary.com | grep 'Canary' | awk '{print $2}' | awk -F"<" '{print $1}'`
counts[output.strip.split.last] += 1
end
puts counts
安装ruby并测试
# yum install ruby -y
# ruby test-canary.rb
{"v1"=>90, "v2"=>10}
# ruby test-canary.rb
{"v1"=>92, "v2"=>8}
# ruby test-canary.rb
{"v1"=>91, "v2"=>9}
(十二) 环境清理
测试无误后,可以清理之前的学习数据:
# kubectl delete deploy,svc,ingress -n production --all
deployment.apps "canary-v1" deleted
service "canary-v1" deleted
ingress.networking.k8s.io "canary-v1" deleted
# kubectl delete deploy,svc,ingress -n canary --all
deployment.apps "canary-v2" deleted
service "canary-v2" deleted
ingress.networking.k8s.io "canary-v2" deleted
# kubectl delete deploy,svc,ingress -n study-ingress --all
deployment.apps "err-page" deleted
deployment.apps "laptop" deleted
deployment.apps "nginx" deleted
deployment.apps "phone" deleted
service "err-page" deleted
service "laptop" deleted
service "nginx" deleted
service "phone" deleted
ingress.networking.k8s.io "ingress-with-auth" deleted
ingress.networking.k8s.io "laptop" deleted
ingress.networking.k8s.io "nginx-ingress" deleted
ingress.networking.k8s.io "phone" deleted
# kubectl delete ns study-ingress production canary
namespace "study-ingress" deleted
namespace "production" deleted
namespace "canary" deleted