前言：最近在看k8s相关资料,看到健康检查这一部分,这一部分是很重要的,涉及到的内容也很多，本文章将整理我在学习中的笔记以及相应方法。
部分来源于网络并通过测试。
k8s完整教程可参考：https://www.zfcdn.xyz/showinfo-3-36255-0.html

---

说明：健康检查是在 应用层面实现的。
livenessProbe（存活检查）和readinessProbe（可读性或就绪检查）
livenessProbe：
存活探测将通过http、shell命令或者tcp等方式去检测容器中的应用是否健康，然后将检查结果返回给kubelet，如果检查容器中应用为不健康状态提交给kubelet后，kubelet将杀死当前容器病根据Pod配置清单中定义的重启策略restartPolicy来对Pod进行重启。
readinessProbe：
就绪探测也是通过http、shell命令或者tcp等方式去检测容器中的应用是否健康或则是否能够正常对外提供服务，如果能够正常对外提供服务，则认为该容器为(Ready状态)，达到(Ready状态)的Pod才可以接收请求。
如果容器或则Pod状态为(NoReady)状态，Kubernetes则会把该Pod从Service的后端endpoints Pod中去剔除。
    ExecAction（exec）：在容器中执行shell命令，命令执行后返回的状态为0则成功，表示我们探测结果正常
    HTTPGetAction（httpget）：根据容器IP、端口以及路径发送HTTP请求，返回码如果是200-400之间表示成功
    TCPSocketAction（tcpsocket）：根据容器IP地址及特定的端口进行TCP检查，端口开放表示成功
以上每种检查动作都可能有以下三种返回状态
    Success，表示通过了健康检查
    Failure，表示没有通过健康检查
    Unknown，表示检查动作失败
范例：
livenessProbe（存活检查）的三种方法：
1、exec方式的示例：
创建一个Pod——》运行Nginx容器——》首先启动nginx——》然后沉睡60秒后——〉删除nginx.pid
通过livenessProbe存活探测的exec命令判断nginx.pid文件是否存在，如果探测返回结果非0，则按照重启策略进行重启。
预期是容器真正（Ready）状态60s后，删除nginx.pid，exec命令探测生效，按照重启策略进行重启

apiVersion: v1
kind: Pod
metadata:
  name: ngx-health
spec:
  containers:
  - name: ngx-liveness
    image: nginx:latest
    command:
    - /bin/sh
    - -c
    - /usr/sbin/nginx; sleep 60; rm -rf /run/nginx.pid
    livenessProbe:
      exec:
        command: [ "/bin/sh", "-c", "test", "-e", "/run/nginx.pid" ]
  restartPolicy: Always

根据yaml文件生成：Pod

kubectl apply -f ngx-health.yaml

等待Pod Ready

[root@master ~]# kubectl get pod
NAME                     READY   STATUS              RESTARTS   AGE
nginx-85b98978db-nh2m4   1/1     Running             0          2d2h
ngx-health               0/1     ContainerCreating   0          18s

[root@master ~]# kubectl get pod
NAME                     READY   STATUS    RESTARTS   AGE
nginx-85b98978db-nh2m4   1/1     Running   0          2d2h
ngx-health               1/1     Running   0          41s

查看Pod的详细信息

#第一次查看，Pod中的容器启动成功，事件正常
kubectl describe pods/ngx-health | grep -A 10 Events
Events:
  Type    Reason     Age        From                 Message
  ----    ------     ----       ----                 -------
  Normal  Scheduled  <unknown>  default-scheduler    Successfully assigned default/ngx-health to k8s-node03
  Normal  Pulling    12s        kubelet, k8s-node03  Pulling image "nginx:latest"
  Normal  Pulled     6s         kubelet, k8s-node03  Successfully pulled image "nginx:latest"
  Normal  Created    6s         kubelet, k8s-node03  Created container ngx-liveness
  Normal  Started    5s         kubelet, k8s-node03  Started container ngx-liveness
 
#第二次查看，容器的livenessProbe探测失败，
kubectl describe pods/ngx-health | grep -A 10 Events
Events:
  Type     Reason     Age                From                 Message
  ----     ------     ----               ----                 -------
  Normal   Scheduled  <unknown>          default-scheduler    Successfully assigned default/ngx-health to k8s-node03
  Normal   Pulling    52s                kubelet, k8s-node03  Pulling image "nginx:latest"
  Normal   Pulled     46s                kubelet, k8s-node03  Successfully pulled image "nginx:latest"
  Normal   Created    46s                kubelet, k8s-node03  Created container ngx-liveness
  Normal   Started    45s                kubelet, k8s-node03  Started container ngx-liveness
  Warning  Unhealthy  20s (x3 over 40s)  kubelet, k8s-node03  Liveness probe failed:
  Normal   Killing    20s                kubelet, k8s-node03  Container ngx-liveness failed liveness probe, will be restarted
 
#第三次查看，已经重新拉取镜像，然后创建容器再启动容器
kubectl describe pods/ngx-health | grep -A 10 Events
Events:
  Type     Reason     Age                From                 Message
  ----     ------     ----               ----                 -------
  Normal   Scheduled  <unknown>          default-scheduler    Successfully assigned default/ngx-health to k8s-node03
  Warning  Unhealthy  35s (x3 over 55s)  kubelet, k8s-node03  Liveness probe failed:
  Normal   Killing    35s                kubelet, k8s-node03  Container ngx-liveness failed liveness probe, will be restarted
  Normal   Pulling    4s (x2 over 67s)   kubelet, k8s-node03  Pulling image "nginx:latest"
  Normal   Pulled     2s (x2 over 61s)   kubelet, k8s-node03  Successfully pulled image "nginx:latest"
  Normal   Created    2s (x2 over 61s)   kubelet, k8s-node03  Created container ngx-liveness
  Normal   Started    2s (x2 over 60s)   kubelet, k8s-node03  Started container ngx-liveness

通过长格式输出可以看到如下,第一次长格式输出Pod运行时间22s，重启次数为0
第二次长格式输出，运行时间是76s，Pod已经完成一次重启

kubectl get pods -o wide | grep ngx-health
ngx-health                          1/1     Running            0          22s     10.244.5.44   k8s-node03   <none>           <none>

kubectl get pods -o wide | grep ngx-health
ngx-health                          1/1     Running            1          76s     10.244.5.44   k8s-node03   <none>           <none>

第二次健康探测失败及第二次重启

kubectl describe pods/ngx-health | grep -A 10 Events
Events:
  Type     Reason     Age                 From                 Message
  ----     ------     ----                ----                 -------
  Normal   Scheduled  <unknown>           default-scheduler    Successfully assigned default/ngx-health to k8s-node03
  Normal   Pulled     58s (x2 over 117s)  kubelet, k8s-node03  Successfully pulled image "nginx:latest"
  Normal   Created    58s (x2 over 117s)  kubelet, k8s-node03  Created container ngx-liveness
  Normal   Started    58s (x2 over 116s)  kubelet, k8s-node03  Started container ngx-liveness
  Warning  Unhealthy  31s (x6 over 111s)  kubelet, k8s-node03  Liveness probe failed:
  Normal   Killing    31s (x2 over 91s)   kubelet, k8s-node03  Container ngx-liveness failed liveness probe, will be restarted
  Normal   Pulling    0s (x3 over 2m3s)   kubelet, k8s-node03  Pulling image "nginx:latest"
 
kubectl get pods -o wide | grep ngx-health
ngx-health                          1/1     Running            2          2m13s   10.244.5.44   k8s-node03   <none>           <none>

2、livenessProbe for HTTPGetAction示例：
通过容器的ip地址，端口号及路径调用HTTPGet方法，如果响应的状态码大于等于200且小于400，则认为容器健康，spec.containers.livenessProbe.httpGet字段用于定义此类检测，它的可用配置字段包括如下几个：

    host ：请求的主机地址，默认为Pod IP；也可以在httpHeaders中使用 Host: 来定义
    port ：请求的端口，必选字段，端口范围1-65535
    httpHeaders <[]Object>：自定义的请求报文首部
    path ：请求的HTTP资源路径，即URL path
    scheme：建立连接使用的协议，仅可为HTTP或HTTPS，默认为HTTP

创建一个Pod——》运行Nginx容器——》首先启动nginx——》然后沉睡60秒后——〉删除nginx.pid
通过livenessProbe存活探测的httpGet方式请求nginx项目根目录下的index.html文件，访问端口为80，访问地址默认为Pod IP，请求协议为HTTP，如果请求失败则按照重启策略进行重启。

apiVersion: v1
kind: Pod
metadata:
  name: ngx-health-tcp
spec:
  containers:
  - name: ngx-liveness-tcp
    image: nginx:latest
    command:
    - /bin/sh
    - -c
    - /usr/sbin/nginx; sleep 60; rm -rf /run/nginx.pid
    livenessProbe:
      httpGet:
        path: /index.html
        port: 80
        scheme: HTTP
  restartPolicy: Always

创建Pod资源对象

kubectl apply -f ngx-health-tcp.yaml

查看Pod运行状态：

#容器创建
kubectl get pods -o wide | grep ngx-health-tcp
ngx-health-tcp                          0/1     ContainerCreating   0          7s      <none>        k8s-node02   <none>           <none>

#容器运行成功
kubectl get pods -o wide | grep ngx-health-tcp
ngx-health-tcp                          1/1     Running            0          19s     10.244.2.36   k8s-node02   <none>           <none>

查看Pod的详细事件信息
容器镜像拉取并启动成功

kubectl describe pods/ngx-health-tcp | grep -A 10 Events
Events:
  Type    Reason     Age        From                 Message
  ----    ------     ----       ----                 -------
  Normal  Scheduled  <unknown>  default-scheduler    Successfully assigned default/ngx-health-tcp to k8s-node02
  Normal  Pulling    30s        kubelet, k8s-node02  Pulling image "nginx:latest"
  Normal  Pulled     15s        kubelet, k8s-node02  Successfully pulled image "nginx:latest"
  Normal  Created    15s        kubelet, k8s-node02  Created container ngx-liveness-tcp
  Normal  Started    14s        kubelet, k8s-node02  Started container ngx-liveness-tcp

容器ready状态后运行60s左右livenessProbe健康检测，可以看到下面已经又开始拉取镜像

kubectl describe pods/ngx-health-tcp | grep -A 15 Events
Events:
  Type    Reason     Age               From                 Message
  ----    ------     ----              ----                 -------
  Normal  Scheduled  <unknown>         default-scheduler    Successfully assigned default/ngx-health-tcp to k8s-node02
  Normal  Pulled     63s               kubelet, k8s-node02  Successfully pulled image "nginx:latest"
  Normal  Created    63s               kubelet, k8s-node02  Created container ngx-liveness-tcp
  Normal  Started    62s               kubelet, k8s-node02  Started container ngx-liveness-tcp
  Normal  Pulling    1s (x2 over 78s)  kubelet, k8s-node02  Pulling image "nginx:latest"

镜像拉取完后再次重启创建并启动了一遍，可以看到 Age 列的时间已经重新计算

kubectl describe pods/ngx-health-tcp | grep -A 15 Events
Events:
  Type    Reason     Age                From                 Message
  ----    ------     ----               ----                 -------
  Normal  Scheduled  <unknown>          default-scheduler    Successfully assigned default/ngx-health-tcp to k8s-node02
  Normal  Pulling    18s (x2 over 95s)  kubelet, k8s-node02  Pulling image "nginx:latest"
  Normal  Pulled     2s (x2 over 80s)   kubelet, k8s-node02  Successfully pulled image "nginx:latest"
  Normal  Created    2s (x2 over 80s)   kubelet, k8s-node02  Created container ngx-liveness-tcp
  Normal  Started    1s (x2 over 79s)   kubelet, k8s-node02  Started container ngx-liveness-tcp

长格式输出Pod，可以看到Pod已经重启过一次

kubectl get pods -o wide | grep ngx-health-tcp
ngx-health-tcp                          0/1     Completed          0          96s     10.244.2.36   k8s-node02   <none>           <none>
k8sops@k8s-master01:~/manifests/pod$ kubectl get pods -o wide | grep ngx-health-tcp
ngx-health-tcp                          1/1     Running            1          104s    10.244.2.36   k8s-node02   <none>           <none>

通过查看容器日志，可以看到下面的探测日志，默认10秒探测一次

[root@master ~]# kubectl logs -f pods/ngx-health-tcp
2023/03/23 10:01:52 [notice] 7#7: using the "epoll" event method
2023/03/23 10:01:52 [notice] 7#7: nginx/1.21.5
2023/03/23 10:01:52 [notice] 7#7: built by gcc 10.2.1 20210110 (Debian 10.2.1-6)
2023/03/23 10:01:52 [notice] 7#7: OS: Linux 3.10.0-1160.el7.x86_64
2023/03/23 10:01:52 [notice] 7#7: getrlimit(RLIMIT_NOFILE): 1048576:1048576
2023/03/23 10:01:52 [notice] 8#8: start worker processes
2023/03/23 10:01:52 [notice] 8#8: start worker process 9
2023/03/23 10:01:52 [notice] 8#8: start worker process 10
2023/03/23 10:01:52 [notice] 8#8: start worker process 12
2023/03/23 10:01:52 [notice] 8#8: start worker process 13
2023/03/23 10:01:52 [notice] 8#8: start worker process 14
2023/03/23 10:01:52 [notice] 8#8: start worker process 15
2023/03/23 10:01:52 [notice] 8#8: start worker process 16
2023/03/23 10:01:52 [notice] 8#8: start worker process 17
10.244.2.1 - - [23/Mar/2023:10:02:01 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:02:11 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:02:21 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:02:31 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:02:41 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:02:51 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:03:01 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:03:11 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:03:21 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"
10.244.2.1 - - [23/Mar/2023:10:03:31 +0000] "GET /index.html HTTP/1.1" 200 615 "-" "kube-probe/1.23" "-"

3、livenessProbe for TCPSocketAction示例
通过容器的IP地址和端口号进行TCP检查，如果能够建立TCP连接，则表明容器健康。相比较来说，它比基于HTTP的探测要更高效，更节约资源，但精准度略低，毕竟建立连接成功未必意味着页面资源可用，spec.containers.livenessProbe.tcpSocket字段用于定义此类检测，它主要包含以下两个可用的属性：
    host：请求连接的目标IP地址，默认为Pod IP
    port：请求连接的目标端口，必选字段
    下面是在资源清单文件中使用liveness-tcp方式的示例，它向Pod IP的80/tcp端口发起连接请求，并根据连接建立的状态判定测试结果：
创建资源配置清单
name后面的名字必须小写。

apiVersion: v1
kind: Pod
metadata:
  name: ngx-health-tcpSocket
spec:
  containers:
  - name: ngx-livenesstcpSocket
    image: nginx:latest
    command:
    - /bin/sh
    - -c
    - /usr/sbin/nginx; sleep 100; rm -rf /run/nginx.pid
    livenessProbe:
      tcpSocket:
        port: 80
  restartPolicy: Always

创建资源对象

kubectl apply -f ngx-health-tcpsocket.yaml

查看Pod创建属性信息
#容器创建并启动成功

[root@master ~]# kubectl describe pods/ngx-health-tcpsocket | grep -A 15 Events
Events:
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Normal  Scheduled  49s   default-scheduler  Successfully assigned default/ngx-health-tcpsocket to node1
  Normal  Pulling    49s   kubelet            Pulling image "nginx:latest"
  Normal  Pulled     48s   kubelet            Successfully pulled image "nginx:latest" in 793.703139ms
  Normal  Created    48s   kubelet            Created container ngx-liveness-tcpsocket
  Normal  Started    48s   kubelet            Started container ngx-liveness-tcpsocket

#在容器ready状态后100s左右Pod已经有了再次拉取镜像的动作
[root@master ~]# kubectl describe pods/ngx-health-tcpsocket | grep -A 15 Events
Events:
  Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  4m11s                default-scheduler  Successfully assigned default/ngx-health-tcpsocket to node1
  Normal   Pulled     4m10s                kubelet            Successfully pulled image "nginx:latest" in 793.703139ms
  Normal   Pulled     2m13s                kubelet            Successfully pulled image "nginx:latest" in 15.787433374s
  Warning  BackOff    31s (x2 over 33s)    kubelet            Back-off restarting failed container
  Normal   Pulling    18s (x3 over 4m11s)  kubelet            Pulling image "nginx:latest"
  Normal   Created    2s (x3 over 4m10s)   kubelet            Created container ngx-liveness-tcpsocket
  Normal   Started    2s (x3 over 4m10s)   kubelet            Started container ngx-liveness-tcpsocket
  Normal   Pulled     2s                   kubelet            Successfully pulled image "nginx:latest" in 15.821059033s
#通过长格式输出Pod，也可以看到当前Pod已经进入了完成的状态，接下来就是重启Pod
[root@master ~]#  kubectl get pods -o wide | grep ngx-health-tcpsocke
ngx-health-tcpsocket     1/1     Running            2 (96s ago)     5m14s   10.244.1.3   node1   <none>           <none>

健康检测参数
上面介绍了两种在不同时间段的探测方式，以及两种探测方式所支持的探测方法，这里介绍几个辅助参数
    initialDelaySeconds：检查开始执行的时间，以容器启动完成为起点计算
    periodSeconds：检查执行的周期，默认为10秒，最小为1秒
    successThreshold：从上次检查失败后重新认定检查成功的检查次数阈值（必须是连续成功），默认为1，也必须是1
    timeoutSeconds：检查超时的时间，默认为1秒，最小为1秒
    failureThreshold：从上次检查成功后认定检查失败的检查次数阈值（必须是连续失败），默认为1
健康检测实践
以下示例使用了就绪探测readinessProbe和存活探测livenessProbe
就绪探测配置解析：
    1、容器在启动5秒initialDelaySeconds后进行第一次就绪探测，将通过http访问探测容器网站根目录下的index.html文件，如果探测成功，则Pod将被标记为（Ready）状态。
    2、然后就绪检测通过periodSeconds参数所指定的间隔时间进行循环探测，下面我所指定的间隔时间是10秒钟，每隔10秒钟就绪探测一次。
    3、每次探测超时时间为3秒，如果探测失败1次就将此Pod从Service的后端Pod中剔除，剔除后客户端请求将无法通过Service访问到其Pod。
    4、就绪探测还会继续对其进行探测，那么如果发现此Pod探测成功1次，通过successThreshold参数设定的值，那么会将它再次加入后端Pod。
存活探测配置解析
    1、容器在启动15秒initialDelaySeconds后进行第一次存活探测，将通过tcpSocket探测容器的80端口，如果探测返回值为0则成功。
    2、每次存活探测间隔为3秒钟，每次探测超时时间为1秒，如果连续探测失败2次则通过重启策略重启Pod。
    3、检测失败后的Pod，存活探测还会对其进行探测，如果再探测成功一次，那么将认为此Pod为健康状态
1.资源配置清单

#create namespace
apiVersion: v1
kind: Namespace
metadata:
  name: nginx-health-ns
  labels:
    resource: nginx-ns
spec:

---

#create deploy and pod
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-health-deploy
  namespace: nginx-health-ns
  labels:
    resource: nginx-deploy
spec:
  replicas: 3
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      app: nginx-health
  template:
    metadata:
      namespace: nginx-health-ns
      labels:
        app: nginx-health
    spec:
      restartPolicy: Always
      containers:
      - name: nginx-health-containers
        image: nginx:1.17.1
        imagePullPolicy: IfNotPresent
        command:
        - /bin/sh
        - -c
        - /usr/sbin/nginx; sleep 60; rm -rf /run/nginx.pid
        readinessProbe:
          initialDelaySeconds: 5
          periodSeconds: 10
          successThreshold: 1
          timeoutSeconds: 3
          failureThreshold: 1
          httpGet:
            path: /index.html
            port: 80
            scheme: HTTP
        livenessProbe:
          initialDelaySeconds: 15
          periodSeconds: 3
          successThreshold: 1
          timeoutSeconds: 1
          failureThreshold: 2
          tcpSocket:
            port: 80
        resources:
          requests:
            memory: "64Mi"
            cpu: "250m"
          limits:
            memory: "128Mi"
            cpu: "500m"

---

#create service
apiVersion: v1
kind: Service
metadata:
  name: nginx-health-svc
  namespace: nginx-health-ns
  labels:
    resource: nginx-svc
spec:
   clusterIP: 10.106.189.88
   ports:
   - port: 80
     protocol: TCP
     targetPort: 80
   selector:
     app: nginx-health
   sessionAffinity: ClientIP
   type: ClusterIP

2.创建资源对象

[root@master ~]# kubectl apply -f nginx-health.yaml
namespace/nginx-health-ns created
deployment.apps/nginx-health-deploy created
service/nginx-health-svc created

3.查看创建的资源对象

[root@master ~]# kubectl get all -n nginx-health-ns -o wide
NAME                                      READY   STATUS              RESTARTS   AGE   IP       NODE    NOMINATED NODE   READINESS GATES
pod/nginx-health-deploy-585c775bf-6kqph   0/1     ContainerCreating   0          14s   <none>   node1   <none>           <none>
pod/nginx-health-deploy-585c775bf-8jvlk   0/1     ContainerCreating   0          14s   <none>   node2   <none>           <none>
pod/nginx-health-deploy-585c775bf-rszkc   0/1     ContainerCreating   0          14s   <none>   node2   <none>           <none>

NAME                       TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE   SELECTOR
service/nginx-health-svc   ClusterIP   10.106.189.88   <none>        80/TCP    14s   app=nginx-health

NAME                                  READY   UP-TO-DATE   AVAILABLE   AGE   CONTAINERS                IMAGES         SELECTOR
deployment.apps/nginx-health-deploy   0/3     3            0           14s   nginx-health-containers   nginx:1.17.1   app=nginx-health

NAME                                            DESIRED   CURRENT   READY   AGE   CONTAINERS                IMAGES         SELECTOR
replicaset.apps/nginx-health-deploy-585c775bf   3         3         0       14s   nginx-health-containers   nginx:1.17.1   app=nginx-health,pod-template-hash=585c775bf

4.查看Pod状态,目前Pod状态都没有就绪并且完成状态，准备重启

[root@master ~]# kubectl get pods -n nginx-health-ns -o wide
NAME                                  READY   STATUS    RESTARTS   AGE   IP           NODE    NOMINATED NODE   READINESS GATES
nginx-health-deploy-585c775bf-6kqph   1/1     Running   0          48s   10.244.1.4   node1   <none>           <none>
nginx-health-deploy-585c775bf-8jvlk   0/1     Running   0          48s   10.244.2.5   node2   <none>           <none>
nginx-health-deploy-585c775bf-rszkc   1/1     Running   0          48s   10.244.2.4   node2   <none>           <none>

5.目前已经有一台Pod完成重启，已准备就绪

[root@master ~]# kubectl get pods -n nginx-health-ns -o wide
NAME                                  READY   STATUS             RESTARTS      AGE     IP           NODE    NOMINATED NODE   READINESS GATES
nginx-health-deploy-585c775bf-6kqph   0/1     CrashLoopBackOff   2 (10s ago)   3m55s   10.244.1.4   node1   <none>           <none>
nginx-health-deploy-585c775bf-8jvlk   1/1     Running            2 (72s ago)   3m55s   10.244.2.5   node2   <none>           <none>
nginx-health-deploy-585c775bf-rszkc   0/1     CrashLoopBackOff   2 (12s ago)   3m55s   10.244.2.4   node2   <none>           <none>

6.三台Pod都均完成重启，已准备就绪

[root@master ~]# kubectl get pods -n nginx-health-ns -o wide
NAME                                  READY   STATUS    RESTARTS      AGE     IP           NODE    NOMINATED NODE   READINESS GATES
nginx-health-deploy-585c775bf-6kqph   1/1     Running   3 (48s ago)   4m33s   10.244.1.4   node1   <none>           <none>
nginx-health-deploy-585c775bf-8jvlk   1/1     Running   3 (32s ago)   4m33s   10.244.2.5   node2   <none>           <none>
nginx-health-deploy-585c775bf-rszkc   1/1     Running   3 (50s ago)   4m33s   10.244.2.4   node2   <none>           <none>