广告 | kubernetes各版本离线安装包
祝贺istio1.0发布, 在此献上教程一份
安装
安装k8s 强势插播广告
三步安装,不多说
安装helm, 推荐生产环境用helm安装,可以调参
release地址
如我使用的2.9.1版本
yum install -y socat # 这个不装会报错
[root@istiohost ~]# wget https://storage.googleapis.com/kubernetes-helm/helm-v2.9.1-linux-amd64.tar.gz
[root@istiohost ~]# tar zxvf helm-v2.9.1-linux-amd64.tar.gz
[root@istiohost ~]# cp linux-amd64/helm /usr/bin
先创建一个service account 把管理员权限给helm:
[root@istiohost ~]# cat helmserviceaccount.yaml
apiVersion: v1
kind: ServiceAccount
metadata:
name: tiller
namespace: kube-system
---
kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
name: tiller-clusterrolebinding
subjects:
- kind: ServiceAccount
name: tiller
namespace: kube-system
roleRef:
kind: ClusterRole
name: cluster-admin
apiGroup: ""
kubectl create -f helmserviceaccount.yaml
安装helm 服务端 tiller :
helm init --service-account tiller # 如果已安装更新加 --upgrade 参数
helm list #没任何返回表示成功
安装istio
curl -L https://git.io/getLatestIstio | sh -
cd istio-1.0.0/
export PATH=$PWD/bin:$PATH
helm 2.10.0以前的版本需要装一下CRD:
kubectl apply -f install/kubernetes/helm/istio/templates/crds.yaml
kubectl apply -f install/kubernetes/helm/istio/charts/certmanager/templates/crds.yaml
安装istio, 由于你没有LB所以用NodePort代替:
helm install install/kubernetes/helm/istio --name istio --namespace istio-system --set gateways.istio-ingressgateway.type=NodePort --set gateways.istio-egressgateway.type=NodePort
安装成功:
[root@istiohost istio-1.0.0]# kubectl get pod -n istio-system
NAME READY STATUS RESTARTS AGE
istio-citadel-7d8f9748c5-ntqnp 1/1 Running 0 5m
istio-egressgateway-676c8546c5-2w4cq 1/1 Running 0 5m
istio-galley-5669f7c9b-mkxjg 1/1 Running 0 5m
istio-ingressgateway-5475685bbb-96mbr 1/1 Running 0 5m
istio-pilot-5795d6d695-gr4h4 2/2 Running 0 5m
istio-policy-7f945bf487-gkpxr 2/2 Running 0 5m
istio-sidecar-injector-d96cd9459-674pk 1/1 Running 0 5m
istio-statsd-prom-bridge-549d687fd9-6cbzs 1/1 Running 0 5m
istio-telemetry-6c587bdbc4-jndjn 2/2 Running 0 5m
prometheus-6ffc56584f-98mr9 1/1 Running 0 5m
[root@istiohost istio-1.0.0]# kubectl get svc -n istio-system
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
istio-citadel ClusterIP 10.108.253.89 8060/TCP,9093/TCP 5m
istio-egressgateway NodePort 10.96.151.14 80:30830/TCP,443:30038/TCP 5m
istio-galley ClusterIP 10.102.83.130 443/TCP,9093/TCP 5m
istio-ingressgateway NodePort 10.99.194.13 80:31380/TCP,443:31390/TCP,31400:31400/TCP,15011:31577/TCP,8060:30037/TCP,15030:31855/TCP,15031:30775/TCP 5m
istio-pilot ClusterIP 10.101.4.143 15010/TCP,15011/TCP,8080/TCP,9093/TCP 5m
istio-policy ClusterIP 10.106.221.68 9091/TCP,15004/TCP,9093/TCP 5m
istio-sidecar-injector ClusterIP 10.100.5.170 443/TCP 5m
istio-statsd-prom-bridge ClusterIP 10.107.28.242 9102/TCP,9125/UDP 5m
istio-telemetry ClusterIP 10.105.66.20 9091/TCP,15004/TCP,9093/TCP,42422/TCP 5m
prometheus ClusterIP 10.103.128.152 9090/TCP
使用教程
官网事例 Bookinfo Application
- productpage 调用details和reviews渲染页面
- details包含书本信息
- reviews 书本反馈,调用ratings服务
- ratings 书本租借信息
reviews服务有三个版本:
- V1 不请求ratings
- V2 请求ratings,返回1到5个黑星
- V3 请求ratings,返回1到5个红星
安装应用:
kubectl apply -f <(istioctl kube-inject -f samples/bookinfo/platform/kube/bookinfo.yaml)
安装完成:
[root@istiohost istio-1.0.0]# kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
details ClusterIP 10.104.66.31 9080/TCP 2m
kubernetes ClusterIP 10.96.0.1 443/TCP 4h
productpage ClusterIP 10.109.68.13 9080/TCP 2m
ratings ClusterIP 10.99.55.110 9080/TCP 2m
reviews ClusterIP 10.102.19.129 9080/TCP 2m
[root@istiohost istio-1.0.0]# kubectl get pods
NAME READY STATUS RESTARTS AGE
details-v1-fc9649d9c-dpnlp 2/2 Running 0 2m
productpage-v1-58845c779c-7g8th 2/2 Running 0 2m
ratings-v1-6cc485c997-fb7nh 2/2 Running 0 2m
reviews-v1-76987687b7-x5n7z 2/2 Running 0 2m
reviews-v2-86749dcd5-xchzb 2/2 Running 0 2m
reviews-v3-7f4746b959-nthrq 2/2 Running 0 2m
创建一个gateway,这是为了集群外可以访问
kubectl apply -f samples/bookinfo/networking/bookinfo-gateway.yaml
浏览器访问url:
47.254.28.88是我的节点ip,使用nodeport模式
http://47.254.28.88:31380/productpage
连续点击三次,你会发现右边没星星-> 黑星星-> 红星星切换,对应三个版本的review,默认策略是轮询
创建destination rules, 配置路由访问规则,现在还是轮询
kubectl apply -f samples/bookinfo/networking/destination-rule-all.yaml
智能路由
请求路由 request routing
根据版本路由
把所有路由切换到v1版本
kubectl apply -f samples/bookinfo/networking/virtual-service-all-v1.yaml
这样执行完后,不管怎么刷页面,我们都看不到星星,因为v1版本没星
可以看到destination是这样的:
http:
- route:
- destination:
host: details
subset: v1
试想如此我们做版本切换将是何等简单
根据用户路由
kubectl apply -f samples/bookinfo/networking/virtual-service-reviews-test-v2.yaml
你会发现用jason用户登录就能看到黑星星,而其它方式看到的页面都是无星星
因为这个user走了v2版本,能不强大? 那当然还能根据header什么的做路由了,就不多说了
http:
- match:
- headers:
end-user:
exact: jason
route:
- destination:
host: reviews
subset: v2
- route:
- destination:
host: reviews
subset: v1
故障注入 Fault injection
kubectl apply -f samples/bookinfo/networking/virtual-service-all-v1.yaml
kubectl apply -f samples/bookinfo/networking/virtual-service-reviews-test-v2.yaml
假设代码里有个bug,用户jason, reviews:v2 访问ratings时会卡10s, 我们任然希望端到端的测试能正常走完
kubectl apply -f samples/bookinfo/networking/virtual-service-ratings-test-delay.yaml
注入错误让jason用户有个7s的延迟
hosts:
- ratings
http:
- fault:
delay:
fixedDelay: 7s
percent: 100
match:
- headers:
end-user:
exact: jason
route:
- destination:
host: ratings
subset: v1
- route:
- destination:
host: ratings
subset: v1
这时访问页面显然会出错,因为我们希望7s内能返回,这样我们就发现了一个延迟的bug
Error fetching product reviews!
Sorry, product reviews are currently unavailable for this book.
所以我们就可能通过故障注入去发现这些异常现象
链路切换 Traffic Shifting
我们先把50%流量发送给reviews:v1 50%流量发送给v3,然后再把100%的流量都切给v3
把100%流量切到v1
kubectl apply -f samples/bookinfo/networking/virtual-service-all-v1.yaml
此时不论刷几遍,都没有星星
v1 v3各50%流量
kubectl apply -f samples/bookinfo/networking/virtual-service-reviews-50-v3.yaml
- route:
- destination:
host: reviews
subset: v1
weight: 50
- destination:
host: reviews
subset: v3
weight: 50
此时一会有星,一会没星,但是已经不是轮询算法了
全切v3
kubectl apply -f samples/bookinfo/networking/virtual-service-reviews-v3.yaml
这时不管怎么刷都是红心了