假设当前线上环境我们已经有一套服务 app-old 对外提供 7 层服务，此时我们修复了一些问题，需要灰度发布上线一个新的版本 app-new，但是我们又不希望简单直接地将所有客户端流量切换到新版本 app-new 中，而是希望仅仅切换 20% 的流量到新版本 app-new 中，待运行一段时间稳定，将所有的流量切换到 app-new 服务中后，再平滑地下线掉 app-old 服务。

针对以上多种不同的应用发布需求，K8S Ingress Controller 支持了多种流量切分方式：

  1. 基于 Request Header 的流量切分，适用于灰度发布以及 AB 测试场景

  2. 基于 Cookie 的流量切分，适用于灰度发布以及 AB 测试场景

  3. 基于 Query Param 的流量切分，适用于灰度发布以及 AB 测试场景

  4. 基于服务权重的流量切分，适用于蓝绿发布场景

以下测试基于服务权重的流量切分，也可以将nginx.ingress.kubernetes.io/canary-weight: "30"改为基于 header 的流量切分。

准备老版本程序

老版本程序 app-old

app-old.yaml

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: app-old
spec:
  replicas: 2
  selector:
    matchLabels:
      run: app-old
  template:
    metadata:
      labels:
        run: app-old
    spec:
      containers:
      - image: zouhl/app:v2.1
        imagePullPolicy: Always
        name: app-old
        ports:
        - containerPort: 80
          protocol: TCP
      restartPolicy: Always
---
apiVersion: v1
kind: Service
metadata:
  name: app-old
spec:
  ports:
  - port: 80
    protocol: TCP
    targetPort: 80
  selector:
    run: app-old
  sessionAffinity: None
  type: NodePor

apiVersion: extensions/v1beta1 kind: Deployment metadata: name: app-old spec: replicas: 2 selector: matchLabels: run: app-old template: metadata: labels: run: app-old spec: containers: - image: zouhl/app:v2.1 imagePullPolicy: Always name: app-old ports: - containerPort: 80 protocol: TCP restartPolicy: Always --- apiVersion: v1 kind: Service metadata: name: app-old spec: ports: - port: 80 protocol: TCP targetPort: 80 selector: run: app-old sessionAffinity: None type: NodePor

老版本的 ingress

app-v1.yaml

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: my-app
  labels:
    app: my-app
  annotations:
    kubernetes.io/ingress.class: nginx
  namespace: default
spec:
  rules:
    - host: test.192.168.2.20.xip.io
      http:
        paths:
          - backend:
              serviceName: app-old
              servicePort: 80
            path: /

apiVersion: extensions/v1beta1 kind: Ingress metadata: name: my-app labels: app: my-app annotations: kubernetes.io/ingress.class: nginx namespace: default spec: rules: - host: test.192.168.2.20.xip.io http: paths: - backend: serviceName: app-old servicePort: 80 path: /

在 k8s 中创建

kubectl create -f app-old.yaml
kubectl create -f app-v1.yaml

kubectl create -f app-old.yaml kubectl create -f app-v1.yaml

装备新版本程序

新版本app-new.yaml

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: app-new
spec:
  replicas: 2
  selector:
    matchLabels:
      run: app-new
  template:
    metadata:
      labels:
        run: app-new
    spec:
      containers:
      - image: zouhl/app:v2.2
        imagePullPolicy: Always
        name: app-new
        ports:
        - containerPort: 80
          protocol: TCP
      restartPolicy: Always
---
apiVersion: v1
kind: Service
metadata:
  name: app-new
spec:
  ports:
  - port: 80
    protocol: TCP
    targetPort: 80
  selector:
    run: app-new
  sessionAffinity: None
  type: NodePort

apiVersion: extensions/v1beta1 kind: Deployment metadata: name: app-new spec: replicas: 2 selector: matchLabels: run: app-new template: metadata: labels: run: app-new spec: containers: - image: zouhl/app:v2.2 imagePullPolicy: Always name: app-new ports: - containerPort: 80 protocol: TCP restartPolicy: Always --- apiVersion: v1 kind: Service metadata: name: app-new spec: ports: - port: 80 protocol: TCP targetPort: 80 selector: run: app-new sessionAffinity: None type: NodePort

新版本 canary-ingress

app-v2-canary.yaml

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: my-app-canary
  labels:
    app: my-app
  annotations:
    kubernetes.io/ingress.class: nginx
    nginx.ingress.kubernetes.io/canary: "true"
    nginx.ingress.kubernetes.io/canary-weight: "30"
  namespace: default
spec:
  rules:
    - host: test.192.168.2.20.xip.io
      http:
        paths:
          - backend:
              serviceName: app-new
              servicePort: 80
            path: /

apiVersion: extensions/v1beta1 kind: Ingress metadata: name: my-app-canary labels: app: my-app annotations: kubernetes.io/ingress.class: nginx nginx.ingress.kubernetes.io/canary: "true" nginx.ingress.kubernetes.io/canary-weight: "30" namespace: default spec: rules: - host: test.192.168.2.20.xip.io http: paths: - backend: serviceName: app-new servicePort: 80 path: /

新版本 ingress yaml

app-v2.yaml

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: my-app
  labels:
    app: my-app
  annotations:
    kubernetes.io/ingress.class: nginx
  namespace: default
spec:
  rules:
    - host: test.192.168.2.20.xip.io
      http:
        paths:
          - backend:
              serviceName: app-new
              servicePort: 80
            path: /

apiVersion: extensions/v1beta1 kind: Ingress metadata: name: my-app labels: app: my-app annotations: kubernetes.io/ingress.class: nginx namespace: default spec: rules: - host: test.192.168.2.20.xip.io http: paths: - backend: serviceName: app-new servicePort: 80 path: /

发布流程

$ tree                                              
.
├── app-new.yaml
├── app-old.yaml
├── app-v1.yaml
├── app-v2-canary.yaml
└── app-v2.yaml

$ tree . ├── app-new.yaml ├── app-old.yaml ├── app-v1.yaml ├── app-v2-canary.yaml └── app-v2.yaml

app-v1 已经发布了，现在灰度发布第二版，权重为 30%，nginx.ingress.kubernetes.io/canary-weight: "30"，更多参数参考github

kubectl create -f app-new.yaml
kubectl create -f app-v2-canary.yaml

kubectl create -f app-new.yaml kubectl create -f app-v2-canary.yaml

检查

$ kubectl get ingresses.extensions   
NAME            HOSTS                       ADDRESS   PORTS   AGE
app-ingress     www.example.com                       80      109m
my-app          test.192.168.2.20.xip.io              80      25m
my-app-canary   test.192.168.2.20.xip.io              80      1s
nginx-test      nginx.192.168.2.20.xip.io             80      3h12m

$ kubectl get ingresses.extensions NAME HOSTS ADDRESS PORTS AGE app-ingress www.example.com 80 109m my-app test.192.168.2.20.xip.io 80 25m my-app-canary test.192.168.2.20.xip.io 80 1s nginx-test nginx.192.168.2.20.xip.io 80 3h12m

在后台观察，70% to v1，30% to v2

$ while sleep 0.5; do curl "test.192.168.2.20.xip.io";echo; done
{"v2.2 hostname":"app-new-658dfc9c6b-lbmvr"}
{"v2.2 hostname":"app-new-658dfc9c6b-qhwtg"}
{"v1 hostname":"app-old-64fd44b699-4hvlb"}
{"v1 hostname":"app-old-64fd44b699-zb58f"}

$ while sleep 0.5; do curl "test.192.168.2.20.xip.io";echo; done {"v2.2 hostname":"app-new-658dfc9c6b-lbmvr"} {"v2.2 hostname":"app-new-658dfc9c6b-qhwtg"} {"v1 hostname":"app-old-64fd44b699-4hvlb"} {"v1 hostname":"app-old-64fd44b699-zb58f"}

如果一切正常则可以正式发布

# delete the canary ingress
kubectl delete -f app-v2-canary.yaml
# set 100% traffic to v2
kubectl apply -f app-v2.yaml

# delete the canary ingress kubectl delete -f app-v2-canary.yaml # set 100% traffic to v2 kubectl apply -f app-v2.yaml

检查 ingress

$ kubectl get ingresses.extensions    
NAME          HOSTS                       ADDRESS   PORTS   AGE
app-ingress   www.example.com                       80      109m
my-app        test.192.168.2.20.xip.io              80      25m
nginx-test    nginx.192.168.2.20.xip.io             80      3h13m

$ while sleep 0.5; do curl "test.192.168.2.20.xip.io";echo; done
{"v2.2 hostname":"app-new-658dfc9c6b-lbmvr"}
{"v2.2 hostname":"app-new-658dfc9c6b-qhwtg"}

$ kubectl get ingresses.extensions NAME HOSTS ADDRESS PORTS AGE app-ingress www.example.com 80 109m my-app test.192.168.2.20.xip.io 80 25m nginx-test nginx.192.168.2.20.xip.io 80 3h13m $ while sleep 0.5; do curl "test.192.168.2.20.xip.io";echo; done {"v2.2 hostname":"app-new-658dfc9c6b-lbmvr"} {"v2.2 hostname":"app-new-658dfc9c6b-qhwtg"}
 

转载于:https://my.oschina.net/xiaominmin/blog/3070144

# 运维 # 网络
Logo
K8S/Kubernetes

K8S/Kubernetes社区为您提供最前沿的新闻资讯和知识内容

更多推荐

【深度】阿里巴巴万级规模 K8s 集群全局高可用体系之美

作者 |  韩堂、柘远、沉醉来源 | 阿里巴巴云原生公众号​前言台湾作家林清玄在接受记者采访的时候，如此评价自己 30 多年写作生涯：“第一个十年我才华横溢，‘贼光闪现’，令周边黯然失色；第二个十年，我终于‘宝光现形’，不再去抢风头，反而与身边的美丽相得益彰；进入第三个十年，繁华落尽见真醇，我进入了‘醇光初现’的阶段，真正体味到了境界之美”。​长夜有穷，真水无香。领略过了 K8s“身在江

avatar K8S/Kubernetes

如何基于 K8s 构建下一代 DevOps 平台？

作者 | 孙健波（天元）导读：当前云原生 DevOps 体系现状如何？面临哪些挑战？如何通过 OAM 解决云原生 DevOps 场景下的诸多问题？云原生开发应用模型 OAM(Open Application Model) 社区核心成员孙健波将为大家一一解答，并分享如何基于 OAM 和 Kubernetes 打造无限能力的下一代 DevOps 平台。什么是 DevOps？为什么基于 Kub

avatar K8S/Kubernetes

k8s 火了！

2020，上云之年，产品云端化成为一种趋势。在一线城市，很多公司都已经构建了自己的私有云环境，比如阿里云、网易云、华为云等。而Kubernetes 作为基于容器编排领域的王者，具备扩展...

avatar K8S/Kubernetes