kubernetes 就绪检测readiness probe(8)

pod被"正式的使用"之前需要做一些简单的检测pod的服务可用性，就此，readiness probe就派上用场

readiness probe

当使用kubectl get pods，在READY中的1/1,右边的1是POD内容器数量，而左边的1是就绪的个数。倘若不定义，那么在启动时候，就会立刻就绪

[root@linuxea linuxea]# kubectl get pods
NAME                             READY     STATUS    RESTARTS   AGE
client-linuxea                   1/1       Running   0          3d
linuxea.com-httpget              1/1       Running   1          3h
nginx-linuxea-5786698598-dx2jr   1/1       Running   0          1d
nginx-linuxea-5786698598-stttv   1/1       Running   0          1d
nginx-linuxea-5786698598-t6lpx   1/1       Running   0          1d
pod-demo-linuxea                 3/3       Running   0          1d

但是，这样启动就会存在问题。在就绪探测中，倘若一个pod在启动时立刻就绪，那么可能会出现问题，在容器内的程序就绪之前有很多过程可能尚未准备完成。如：展开文件，启动程序如上图所示，service通过标签选择器关联各个pod。此刻，pod_NEW符合选择器的选择条件，pod_NEW将会立刻成为service标签选择器都后端pod之一。此刻，pod_NEW并未就绪，自我初始化尚未完成。那么此刻调度时，service倘若请求到pod_NEW上就会出错。

当然，期望是在调度时，pod_NEW的应用程序已经就绪,也就说以及初始化完成为此，使用readiness probe来做就绪探测。做一个测试

测试

[root@linuxea linuxea]# cat readiness-httpget.yaml 
apiVersion: v1
kind: Pod
metadata: 
  name: linuxea.com-httpget-readiness
  namespace: default
spec:
  containers:
  - name: readiness-httpd-linuxea
    image: marksugar/nginx:1.14.a
    imagePullPolicy: IfNotPresent
    ports:
    - name: http
      containerPort: 80
    livenessProbe:
      httpGet:
        port: http
        path: /index.html
      initialDelaySeconds: 1
      periodSeconds: 3
    readinessProbe:
      httpGet:
        port: http
        path: /linuxea.html
      initialDelaySeconds: 1
      periodSeconds: 3

开始创建

[root@linuxea linuxea]# kubectl create -f readiness-httpget.yaml 
pod/linuxea.com-httpget-readiness created

已经创建成功

[root@linuxea linuxea]# kubectl get pods -o wide
NAME                             READY     STATUS    RESTARTS   AGE 
client-linuxea                   1/1       Running   0          3d  
linuxea.com-httpget              1/1       Running   2          4h  
linuxea.com-httpget-readiness    1/1       Running   0          5s  

而后删除掉/data/wwwroot/linuxea.html

[root@linuxea linuxea]# kubectl exec -it linuxea.com-httpget-readiness  -- /bin/sh
/ # ls /data/wwwroot/
index.html    linuxea.html
/ # rm -rf /data/wwwroot/linuxea.html 
/ # 

紧接着就绪检测失败，不在就绪状态

[root@linuxea linuxea]# kubectl get pods -o wide
NAME                             READY     STATUS    RESTARTS   AGE 
client-linuxea                   1/1       Running   0          3d  
linuxea.com-httpget              1/1       Running   2          4h  
linuxea.com-httpget-readiness    0/1       Running   0          52s 

此刻，可以使用kubectl describe pods linuxea.com-httpget-readiness查看

• 恢复

在将文件补齐echo `date`-linuxea >> /data/wwwroot/linuxea.html

[root@linuxea linuxea]# kubectl exec -it linuxea.com-httpget-readiness  -- /bin/sh
/ # ls /data/wwwroot/
index.html    linuxea.html
/ # rm -rf /data/wwwroot/linuxea.html 
/ # echo `date`-linuxea >> /data/wwwroot/linuxea.html

READY又恢复就绪状态

[root@linuxea linuxea]# kubectl get pods 
NAME                             READY     STATUS    RESTARTS   AGE
client-linuxea                   1/1       Running   0          3d
linuxea.com-httpget              1/1       Running   2          4h
linuxea.com-httpget-readiness    1/1       Running   0          6m

curl

[root@linuxea linuxea]# curl 172.16.3.30/linuxea.html
Mon Aug 27 13:21:06 UTC 2018-linuxea
[root@linuxea linuxea]# 

钩子

PostStart ：创建容器后，立即执行的操作，如果执行失败，容器终止，并且重启。重启与否取决于重启策略prestop：容器终止前执行命令，执行完成后终止都有三种行为：exec,httpget,tcpSocketexec：commamd命令httpget: get请求tcpsocket: socket请求

PostStart

[root@linuxea linuxea]# cat poststart-linuxea.yaml 
apiVersion: v1
kind: Pod
metadata: 
  name: linuxea.com.poststart
  namespace: default
spec:
  containers:
  - name: poststart-linuxea-com
    image: marksugar/nginx:1.14.a
    imagePullPolicy: IfNotPresent
    ports:
    - name: http
      containerPort: 80
    lifecycle:
      postStart:
        exec:
          command:
            - "/bin/sh"
            - "-c"
            - >
              if [ ! -d /data/wwwroot/linuxea.com/ ]; then
              mkdir -p /data/wwwroot/linuxea.com;
              echo `date`-linuxea.com >>/data/wwwroot/linuxea.com/index.html;
              sed -i 's#/data/wwwroot#/data/wwwroot/linuxea.com#g' /etc/nginx/nginx.conf;
              fi;
              if [ -s /data/wwwroot/linuxea.com/index.html ]; then
              chown -R www.www /data/wwwroot;
              fi;

create

[root@linuxea linuxea]# kubectl create -f poststart-linuxea.yaml 
pod/linuxea.com.poststart created

[root@linuxea linuxea]# kubectl get pods
NAME                             READY     STATUS    RESTARTS   AGE
client-linuxea                   1/1       Running   0          4d
linuxea.com-httpget              1/1       Running   2          21h
linuxea.com-httpget-readiness    1/1       Running   0          17h
linuxea.com.poststart            1/1       Running   0          3s

[root@linuxea linuxea]# kubectl get pods -o wide
NAME                             READY     STATUS    RESTARTS   AGE       IP             NODE                 NOMINATED NODE
client-linuxea                   1/1       Running   0          4d        172.16.2.252   linuxea.node-2.com   <none>
linuxea.com-httpget              1/1       Running   2          21h       172.16.2.12    linuxea.node-2.com   <none>
linuxea.com-httpget-readiness    1/1       Running   0          17h       172.16.3.30    linuxea.node-3.com   <none>
linuxea.com.poststart            1/1       Running   0          10s       172.16.1.32    linuxea.node-1.com   <none>

[root@linuxea linuxea]# curl 172.16.1.32
Tue Aug 28 06:17:41 UTC 2018-linuxea.com

通过sh登陆到pod容器内

[root@linuxea linuxea]# kubectl exec -it linuxea.com.poststart  -- /bin/sh
/ # cat /data/wwwroot/linuxea.com/index.html 
Tue Aug 28 06:17:41 UTC 2018-linuxea.com

/ # cat /etc/nginx/nginx.conf|grep root
        root   /data/wwwroot/linuxea.com;
       root /data/wwwroot/linuxea.com;