服务健康检测

最新推荐文章于 2024-07-25 22:31:22 发布

自由的常青树

最新推荐文章于 2024-07-25 22:31:22 发布

阅读量369

点赞数 6

分类专栏： k8s 文章标签： kubernetes 云原生

本文链接：https://blog.csdn.net/m0_63256209/article/details/136052472

版权

k8s 专栏收录该内容

3 篇文章 0 订阅

订阅专栏

容器生命周期钩子

容器启动后触发（PostStart）

容器启动时，会执行这个钩子函数

# yaml描述文件
apiVersion: v1
kind: Pod
metadata:
  name: nginx-hook
  labels:
    app: myapp-hook
spec:
  containers:
    - name: nginx-hook
      image: nginx:alpine
      ports:
        - containerPort: 80
      lifecycle:
        postStart:
          exec:
            command: [/bin/sh,-c,'echo k8s 666 >> /uuuuusr/share/nginx/html/index.html']
  restartPolicy: OnFailure



# 创建资源访问测试
kubectl apply -f start-test.yaml 
pod/nginx-hook created


# 查看该pod的ip地址
kubectl get pods -owide
NAME         READY   STATUS    RESTARTS   AGE   IP               NODE         NOMINATED NODE   READINESS GATES
nginx-hook   1/1     Running   0          4s    172.20.139.188   10.0.0.203   <none>           <none>


# 访问测试
curl 172.20.139.188
k8s 666


# 主动修改错误的钩子命令，演示，钩子，命令执行失败的话，会一直重启容器
            command: [/bin/sh,-c,'echo k8s 666 >> /uuuuuuusr/share/nginx/html/index.html']


# 此时容器状态未成功运行
kubectl get pods
NAME         READY   STATUS      RESTARTS   AGE
nginx-hook   0/1     Completed   0          63s

重启策略	说明
Always	当容器失效时，由kubelet自动重启该容器
OnFailure	当容器终止运行且退出码不为0时，由kubelet自动重启该容器
Never	不论容器运行状态如何，kubelet都不会重启该容器

容器停止前触发（PreStop）

容器退出前，会执行这个钩子函数

# 创建了挂载目录，当pod删除后数据将持久化保留在宿主机上
apiVersion: v1
kind: Pod
metadata:
  name: nginx-hook
  labels:
    app: myapp-hook
spec:
  volumes:
    - name: nginxlog
      hostPath:
        path: /nginx_log/

  restartPolicy: OnFailure
  containers:
    - name: nginx-hook
      image: nginx:alpine
      ports:
        - containerPort: 80
      lifecycle:
        postStart:
          exec:
            command: [/bin/sh,-c,'echo k8s 666 > /usr/share/nginx/html/index.html']
        
        preStop:
          exec:
            command: [/bin/sh,-c,'echo 再见！ >> /var/log/nginx/prestop-test.log']
      volumeMounts:
        - name: nginxlog
          mountPath: /var/log/nginx/


# 删除pod
kubectl delete pods nginx-hook 
pod "nginx-hook" deleted


# 容器退出后，查看宿主机上的文件
cat /nginx_log/prestop-test.log 
再见！

Pod健康探针

存活探针 liveness

nginx运行后，得先curl -I 确保状态码是2xx，才得知nginx这个进程是正确运行了
livenessprobe探针，cat命令执行失败，文件路径不存在，导致健康探针失败
如果存活探针失败，Kubernetes 将会根据定义的重启策略自动重启容器，以尝试恢复应用程序的健康状态

apiVersion: v1
kind: Pod
metadata:
  name: nginx-hook
  labels:
    app: myapp-hook
spec:
  volumes:
    - name: nginx-html
      hostPath:
        path: /nginx-html

  restartPolicy: OnFailure
  containers:
    - name: nginx-hook
      image: nginx:alpine
      ports:
        - containerPort: 80
      livenessProbe:
        exec:
          command:
            - cat
            - /usr/share/nginx/html/index.html
        initialDelaySeconds: 3 # 3秒后探针
        periodSeconds: 5   # 隔5秒探一次
      volumeMounts:
        - name: nginx-html
          mountPath: /usr/share/nginx/html


# 创建pod
kubectl create -f start-test.yaml 
pod/nginx-hook created


# 由于宿主机上文件路径不存在，导致健康探针不正确，kill，重启容器
kubectl get pods
NAME         READY   STATUS      RESTARTS   AGE
nginx-hook   0/1     Completed   3          76s


# 由于设置的重启策略为OnFailure，终止运行且退出码不为0时才会重启
kubectl get pods nginx-hook -oyaml | grep -i restart
  restartPolicy: OnFailure
    restartCount: 3


# 在宿主机上创建该文件再次查看pod
cat /nginx-html/index.html 
66

kubectl get pods
NAME         READY   STATUS    RESTARTS   AGE
nginx-hook   1/1     Running   0          3s


# 此时访问pod的ip等于访问宿主机的映射目录的index.html
curl 172.20.139.134
66

就绪探针 Readiness

httpGet：返回的状态码为2xx,3xx则为成功
需要定义好/read文件，确保nginx可以读取到即可正常running
如果就绪探针失败，Kubernetes 将会从服务负载均衡的池中剔除该容器，不会将流量路由到该容器，直到探测成功

# 创建readiness，生成pod
apiVersion: v1
kind: Pod
metadata:
  name: nginx-hook
  labels:
    app: myapp-hook
spec:
  volumes:
    - name: nginx-html
      hostPath:
        path: /nginx-html

  restartPolicy: OnFailure
  containers:
    - name: nginx-hook
      image: nginx:alpine
      livenessProbe:
        httpGet:
          path: /index.html
          port: 80

        initialDelaySeconds: 3 # 3秒后探针
        periodSeconds: 5   # 隔5秒探一次

      volumeMounts:
        - name: nginx-html
          mountPath: /usr/share/nginx/html

      readinessProbe:
        httpGet:
          path: /read
          port: 80
        initialDelaySeconds: 10    # 容器启动后多久开始探测
        timeoutSeconds: 2          # 表示容器必须在2s内做出相应反馈给probe，否则视为探测失败
        periodSeconds: 30          # 探测周期，每30s探测一次
        successThreshold: 1        # 连续探测1次成功表示成功
        failureThreshold: 3        # 连续探测3次失败表示失败


# 此时pod的流量被断开，这时候即使服务出错，对外界来说也是感知不到的
kubectl get pods
NAME         READY   STATUS    RESTARTS   AGE
nginx-hook   0/1     Running   0          75s


# 让pod正常运行，提供访问
# 在宿主机映射目录内添加yaml文件中定义的read即可
cat /nginx-html/read 
read read test


# 再次查看pod状态并访问测试
kubectl get pods -owide
NAME         READY   STATUS    RESTARTS   AGE     IP               NODE         NOMINATED NODE   READINESS GATES
nginx-hook   1/1     Running   0          6m53s   172.20.139.135   10.0.0.203   <none>           <none>

# 访问测试
curl 172.20.139.135/read
read read test