zoukankan html css js c++ java

Kubernetes 健康状态检查liveness和readiness

1 .健康检查

健康检查（Health Check）是让系统知道您的应用实例是否正常工作的简单方法。如果您的应用实例不再工作，则其他服务不应访问该应用或向其发送请求。相反，应该将请求发送到已准备好的应用程序实例，或稍后重试。系统还应该能够使您的应用程序恢复健康状态。

强大的自愈能力是 Kubernetes 这类容器编排引擎的一个重要特性。自愈的默认实现方式是自动重启发生故障的容器。除此之外，用户还可以利用Liveness 和 Readiness 探测机制设置更精细的健康检查，进而实现如下需求：

零停机部署。
避免部署无效的镜像。
更加安全的滚动升级。

2 .探针类型

Liveness存活性探针

Liveness探针让Kubernetes知道你的应用程序是活着还是死了。如果你的应用程序还活着，那么Kubernetes就不管它了。如果你的应用程序已经死了，Kubernetes将删除Pod并启动一个新的替换它。

Readiness就绪性探针

Readiness探针旨在让Kubernetes知道您的应用何时准备好其流量服务。 Kubernetes确保Readiness探针检测通过，然后允许服务将流量发送到Pod。如果Readiness探针开始失败，Kubernetes将停止向该容器发送流量，直到它通过。 判断容器是否处于可用Ready状态, 达到ready状态表示pod可以接受请求, 如果不健康， 从service的后端endpoint列表中把pod隔离出去。

3 .探针执行方式

HTTP

HTTP探针可能是最常见的自定义Liveness探针类型。即使您的应用程序不是HTTP服务，您也可以在应用程序内创建轻量级HTTP服务以响应Liveness探针。 Kubernetes去ping一个路径，如果它得到的是200或300范围内的HTTP响应，它会将应用程序标记为健康。否则它被标记为不健康。

httpget配置项

host：连接的主机名，默认连接到pod的IP。你可能想在http header中设置"Host"而不是使用IP。
scheme：连接使用的schema，默认HTTP。
path: 访问的HTTP server的path。
httpHeaders：自定义请求的header。HTTP运行重复的header。
port：访问的容器的端口名字或者端口号。端口号必须介于1和65535之间。

Exec

对于Exec探针，Kubernetes则只是在容器内运行命令。如果命令以退出代码0返回，则容器标记为健康。否则，它被标记为不健康。当您不能或不想运行HTTP服务时，此类型的探针则很有用，但是必须是运行可以检查您的应用程序是否健康的命令。

TCP

最后一种类型的探针是TCP探针，Kubernetes尝试在指定端口上建立TCP连接。如果它可以建立连接，则容器被认为是健康的；否则被认为是不健康的。

如果您有HTTP探针或Command探针不能正常工作的情况，TCP探测器会派上用场。例如，gRPC或FTP服务是此类探测的主要候选者。

4 .Liveness-exec样例

执行命令。容器的状态由命令执行完返回的状态码确定。如果返回的状态码是0，则认为pod是健康的，如果返回的是其他状态码，则认为pod不健康，这里不停的重启它。

#cat liveness_exec.yaml
apiVersion: v1
kind: Pod
metadata:
  labels:
    test: liveness-exec
  name: liveness-exec
spec:
  containers:
  - name: liveness-exec
    image: busybox
    args:
    - /bin/sh
    - -c
    - touch /tmp/healthy; sleep 30; rm -rf /tmp/healthy; sleep 600
    livenessProbe:
      exec:
        command:
        - cat
        - /tmp/healthy
      initialDelaySeconds: 5
      periodSeconds: 5

periodSeconds字段指定kubelet应每5秒执行一次活跃度探测。

initialDelaySeconds字段告诉kubelet它应该在执行第一个探测之前等待5秒。要执行探测，kubelet将在Container中执行命令cat /tmp/healthy。如果命令成功，则返回0，并且kubelet认为Container是活动且健康的。如果该命令返回非零值，则kubelet会终止容器并重新启动它。

5 .readiness-exec样例

apiVersion: apps/v1
kind: Deployment
metadata:
  name: busybox-deployment
  namespace: default
  labels:
    app: busybox
spec:
  selector:
    matchLabels:
      app: busybox
  replicas: 3
  template:
    metadata:
      labels:
        app: busybox

    spec:
      containers:
      - name: busybox
        image: busybox:latest
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
        args:
        - /bin/sh
        - -c
        - touch /tmp/healthy; sleep 30; rm -rf /tmp/healthy; sleep 600
        readinessProbe:
          exec:
            command:
            - cat
            - /tmp/healthy
          initialDelaySeconds: 5
          periodSeconds: 5

pod启动，创建健康检查文件，这个时候是正常的，30s后删除，ready变成0，但pod没有被删除或者重启，k8s只是不管他了，仍然可以登录

[root@k8s-master health]# kubectl get pods
NAME                                  READY   STATUS    RESTARTS   AGE
busybox-deployment-6f86ddd894-l9phc   0/1     Running   0          3m10s
busybox-deployment-6f86ddd894-lh46t   0/1     Running   0          3m
busybox-deployment-6f86ddd894-sz5c2   0/1     Running   0          3m17s

我们再登录进去，手动创建健康检查文件，健康检查通过

[root@k8s-master health]# kubectl get pods
NAME                                  READY   STATUS    RESTARTS   AGE
busybox-deployment-6f86ddd894-l9phc   1/1     Running   0          7m44s
busybox-deployment-6f86ddd894-lh46t   0/1     Running   0          7m34s
busybox-deployment-6f86ddd894-sz5c2   1/1     Running   0          7m51s
[root@k8s-master health]#
[root@k8s-master health]#
[root@k8s-master health]#
[root@k8s-master health]# kubectl exec -it busybox-deployment-6f86ddd894-lh46t  /bin/sh
/ # touch tmp/healthy
/ # [root@k8s-master health]# kubeget pods
NAME                                  READY   STATUS    RESTARTS   AGE
busybox-deployment-6f86ddd894-l9phc   1/1     Running   0          8m21s
busybox-deployment-6f86ddd894-lh46t   1/1     Running   0          8m11s
busybox-deployment-6f86ddd894-sz5c2   1/1     Running   0          8m28s
[root@k8s-master health]#

6.iveness-http样例

#cat liveness_http.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: default
  labels:
    app: nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 2
  template:
    metadata:
      labels:
        app: nginx

    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
        livenessProbe:
          httpGet:
            path: /index.html
            port: 80
            httpHeaders:
            - name: X-Custom-Header
              value: hello
          initialDelaySeconds: 5
          periodSeconds: 3

7 readiness-http样例

创建一个2个副本的deployment

# cat readiness_http.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: default
  labels:
    app: nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 2
  template:
    metadata:
      labels:
        app: nginx

    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
        readinessProbe:
          httpGet:
            path: /index.html
            port: 80
            httpHeaders:
            - name: X-Custom-Header
              value: hello
          initialDelaySeconds: 5
          periodSeconds: 3

创建一个svc能访问

# cat readiness_http_svc.yaml
apiVersion: v1
kind: Service
metadata:
  name: nginx
spec:
  type: NodePort
  ports:
    - port: 80
      nodePort: 30001
  selector:  #标签选择器
    app: nginx

服务可以访问

[root@k8s-master health]# kubectl get pods -o wide
NAME                                READY   STATUS    RESTARTS   AGE   IP            NODE         NOMINATED NODE   READINESS GATES
nginx-deployment-7db8445987-9wplj   1/1     Running   0          57s   10.254.1.81   k8s-node-1   <none>           <none>
nginx-deployment-7db8445987-mlc6d   1/1     Running   0          57s   10.254.2.65   k8s-node-2   <none>           <none>
[root@k8s-master health]# kubectl get svc
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)        AGE
kubernetes   ClusterIP   10.96.0.1       <none>        443/TCP        5d3h
nginx        NodePort    10.108.167.58   <none>        80:30001/TCP   27m
[root@k8s-master health]#
[root@k8s-master health]# curl -I 10.6.76.24:30001/index.html
HTTP/1.1 200 OK
Server: nginx/1.17.3
Date: Tue, 03 Sep 2019 04:40:05 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Tue, 13 Aug 2019 08:50:00 GMT
Connection: keep-alive
ETag: "5d5279b8-264"
Accept-Ranges: bytes

[root@k8s-master health]# curl -I 10.6.76.23:30001/index.html
HTTP/1.1 200 OK
Server: nginx/1.17.3
Date: Tue, 03 Sep 2019 04:40:11 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Tue, 13 Aug 2019 08:50:00 GMT
Connection: keep-alive
ETag: "5d5279b8-264"
Accept-Ranges: bytes

修改Nginx pod

[root@k8s-master health]# kubectl exec -it  nginx-deployment-7db8445987-9wplj /bin/bash
root@nginx-deployment-7db8445987-9wplj:/# cd /usr/share/nginx/html/
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html# ls
50x.html  index.html
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html# rm -f index.html
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html#
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html#
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html# nginx -s reload
2019/09/03 03:58:52 [notice] 14#14: signal process started
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html#
root@nginx-deployment-7db8445987-9wplj:/usr/share/nginx/html# exit
[root@k8s-master health]# kubectl get pods -o wide
NAME                                READY   STATUS    RESTARTS   AGE    IP            NODE         NOMINATED NODE   READINESS GATES
nginx-deployment-7db8445987-9wplj   0/1     Running   0          110s   10.254.1.81   k8s-node-1   <none>           <none>
nginx-deployment-7db8445987-mlc6d   1/1     Running   0          110s   10.254.2.65   k8s-node-2   <none>           <none>
[root@k8s-master health]# curl -I 10.254.1.81/index.html
HTTP/1.1 404 Not Found
Server: nginx/1.17.3
Date: Tue, 03 Sep 2019 03:59:16 GMT
Content-Type: text/html
Content-Length: 153
Connection: keep-alive

[root@k8s-master health]# kubectl describe pod nginx-deployment-7db8445987-9wplj
Events:
  Type     Reason     Age                    From                 Message
  ----     ------     ----                   ----                 -------
  Normal   Scheduled  43m                    default-scheduler    Successfully assigned default/nginx-deployment-7db8445987-9wplj tok8s-node-1
  Normal   Pulled     43m                    kubelet, k8s-node-1  Container image "nginx" already present on machine
  Normal   Created    43m                    kubelet, k8s-node-1  Created container nginx
  Normal   Started    43m                    kubelet, k8s-node-1  Started container nginx
  Warning  Unhealthy  3m47s (x771 over 42m)  kubelet, k8s-node-1  Readiness probe failed: HTTP probe failed with statuscode: 404

不再分发流量

我们把Nginx pod的index这个健康检查文件删除，并且把Nginx reload，k8s根据readiness把ready变成0，把它从集群摘除，不再分发流量，我们查看一下两个pod的日志

[root@k8s-master health]# kubectl logs  nginx-deployment-7db8445987-mlc6d | tail -10
10.254.1.0 - - [03/Sep/2019:04:43:44 +0000] "HEAD /index.html HTTP/1.1" 200 0 "-" "curl/7.29.0" "-"
10.254.1.0 - - [03/Sep/2019:04:43:45 +0000] "HEAD /index.html HTTP/1.1" 200 0 "-" "curl/7.29.0" "-"
10.254.1.0 - - [03/Sep/2019:04:43:45 +0000] "HEAD /index.html HTTP/1.1" 200 0 "-" "curl/7.29.0" "-"
10.254.2.1 - - [03/Sep/2019:04:43:46 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:43:49 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:43:52 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:43:55 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:43:58 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:44:01 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
10.254.2.1 - - [03/Sep/2019:04:44:04 +0000] "GET /index.html HTTP/1.1" 200 612 "-" "kube-probe/1.15" "-"
[root@k8s-master health]# kubectl logs  nginx-deployment-7db8445987- | tail -10
nginx-deployment-7db8445987-9wplj  nginx-deployment-7db8445987-mlc6d
[root@k8s-master health]# kubectl logs  nginx-deployment-7db8445987-9wplj | tail -10
2019/09/03 04:44:11 [error] 15#15: *939 open() "/usr/share/nginx/html/index.html" failed (2: No such file or directory), client: 10.254.1.1, server: localhost, request: "GET /index.html HTTP/1.1", host: "10.254.1.81:80"
10.254.1.1 - - [03/Sep/2019:04:44:11 +0000] "GET /index.html HTTP/1.1" 404 153 "-" "kube-probe/1.15" "-"
10.254.1.1 - - [03/Sep/2019:04:44:14 +0000] "GET /index.html HTTP/1.1" 404 153 "-" "kube-probe/1.15" "-"
2019/09/03 04:44:14 [error] 15#15: *940 open() "/usr/share/nginx/html/index.html" failed (2: No such file or directory), client: 10.254.1.1, server: localhost, request: "GET /index.html HTTP/1.1", host: "10.254.1.81:80"
2019/09/03 04:44:17 [error] 15#15: *941 open() "/usr/share/nginx/html/index.html" failed (2: No such file or directory), client: 10.254.1.1, server: localhost, request: "GET /index.html HTTP/1.1", host: "10.254.1.81:80"
10.254.1.1 - - [03/Sep/2019:04:44:17 +0000] "GET /index.html HTTP/1.1" 404 153 "-" "kube-probe/1.15" "-"
10.254.1.1 - - [03/Sep/2019:04:44:20 +0000] "GET /index.html HTTP/1.1" 404 153 "-" "kube-probe/1.15" "-"
2019/09/03 04:44:20 [error] 15#15: *942 open() "/usr/share/nginx/html/index.html" failed (2: No such file or directory), client: 10.254.1.1, server: localhost, request: "GET /index.html HTTP/1.1", host: "10.254.1.81:80"
2019/09/03 04:44:23 [error] 15#15: *943 open() "/usr/share/nginx/html/index.html" failed (2: No such file or directory), client: 10.254.1.1, server: localhost, request: "GET /index.html HTTP/1.1", host: "10.254.1.81:80"
10.254.1.1 - - [03/Sep/2019:04:44:23 +0000] "GET /index.html HTTP/1.1" 404 153 "-" "kube-probe/1.15" "-"
[root@k8s-master health]#

8 TCP liveness和readiness探针

TCP检查的配置与HTTP检查非常相似，主要对于没有http接口的pod，像MySQL，Redis，等等

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: default 
  labels:
    app: nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 1
  template:
    metadata:
      labels:
        app: nginx

    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
        livenessProbe:
          tcpSocket:
            port: 80
          initialDelaySeconds: 5
          periodSeconds: 3
        readinessProbe:
          tcpSocket:
            port: 80
          initialDelaySeconds: 5
          periodSeconds: 3

9 Probe详细配置

initialDelaySeconds：容器启动后第一次执行探测是需要等待多少秒。

periodSeconds：执行探测的频率。默认是10秒，最小1秒。

timeoutSeconds：探测超时时间。默认1秒，最小1秒。

successThreshold：探测失败后，最少连续探测成功多少次才被认定为成功。默认是1。对于liveness必须是1。最小值是1。

failureThreshold：探测成功后，最少连续探测失败多少次才被认定为失败。默认是3。最小值是1。

HTTP probe 中可以给 httpGet设置其他配置项：

使用Liveness探针时需要配置一个非常重要的设置，就是initialDelaySeconds设置。

Liveness探针失败会导致Pod重新启动。在应用程序准备好之前，您需要确保探针不会启动。否则，应用程序将不断重启，永远不会准备好！

10 健康检查在扩容中的应用readiness

对于多副本应用，当执行 Scale Up 操作时，新副本会作为 backend 被添加到 Service 的负责均衡中，与已有副本一起处理客户的请求。考虑到应用启动通常都需要一个准备阶段，比如加载缓存数据，连接数据库等，从容器启动到正真能够提供服务是需要一段时间的。我们可以通过 Readiness 探测判断容器是否就绪，避免将请求发送到还没有 ready 的 backend。

以上面readiness-http为例

# cat liveness_http.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: default
  labels:
    app: nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 2
  template:
    metadata:
      labels:
        app: nginx

    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
        livenessProbe:
          httpGet:
            path: /index.html
            port: 80
            httpHeaders:
            - name: X-Custom-Header
              value: hello
          initialDelaySeconds: 5
          periodSeconds: 3
# cat readiness_http_svc.yaml
apiVersion: v1
kind: Service
metadata:
  name: nginx
spec:
  type: NodePort
  ports:
    - port: 80
      nodePort: 30001
  selector:  #标签选择器
    app: nginx

容器启动 5 秒之后开始探测。
如果 http://[container_ip]:80/index.html 返回代码不是 200-400，表示容器没有就绪，不接收 Service web-svc 的请求。
每隔 3 秒再探测一次。
直到返回代码为 200-400，表明容器已经就绪，然后将其加入到 web-svc 的负责均衡中，开始处理客户请求。
探测会继续以 5 秒的间隔执行，如果连续发生 3 次失败，容器又会从负载均衡中移除，直到下次探测成功重新加入。
对于生产环境中重要的应用都建议配置 Health Check，保证处理客户请求的容器都是准备就绪的 Service backend。

我们手动扩容一下，在pod的健康检查没有通过之前，新起的pod就不加入集群

[root@k8s-master health]# kubectl scale deployment nginx-deployment --replicas=5
deployment.extensions/nginx-deployment scaled
[root@k8s-master health]# kubectl get pods
NAME                                READY   STATUS              RESTARTS   AGE
nginx-deployment-7db8445987-k9df8   0/1     ContainerCreating   0          3s
nginx-deployment-7db8445987-mlc6d   1/1     Running             0          3h14m
nginx-deployment-7db8445987-q5d9k   0/1     ContainerCreating   0          3s
nginx-deployment-7db8445987-w2w2t   0/1     ContainerCreating   0          3s
nginx-deployment-7db8445987-zwj8t   1/1     Running             0          8m4s

11.健康检查在滚动更新中的应用

现有一个正常运行的多副本应用，接下来对应用进行更新（比如使用更高版本的 image），Kubernetes 会启动新副本，然后发生了如下事件：

l 正常情况下新副本需要 10 秒钟完成准备工作，在此之前无法响应业务请求。

l 但由于人为配置错误，副本始终无法完成准备工作（比如无法连接后端数据库）。

因为新副本本身没有异常退出，默认的 Health Check 机制会认为容器已经就绪，进而会逐步用新副本替换现有副本，其结果就是：当所有旧副本都被替换后，整个应用将无法处理请求，无法对外提供服务。如果这是发生在重要的生产系统上，后果会非常严重。

如果正确配置了 Health Check，新副本只有通过了 Readiness 探测，才会被添加到 Service；如果没有通过探测，现有副本不会被全部替换，业务仍然正常进行。

app.v1模拟一个5个副本的应用

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: app
spec:
  replicas: 5
  template:
    metadata:
      labels:
        run: app
    spec:
      containers:
      - name: app
        image: busybox
        args:
        - /bin/sh
        - -c
        - sleep 10; touch /tmp/healthy; sleep 30000
        readinessProbe:
          exec:
            command:
            - cat
            - /tmp/healthy
          initialDelaySeconds: 10
          periodSeconds: 5

10 秒后副本能够通过 Readiness 探测

[root@k8s-master health]#  kubectl apply -f app_v1.yaml
deployment.extensions/app unchanged
[root@k8s-master health]# kubectl get pods
NAME                   READY   STATUS    RESTARTS   AGE
app-6dd7f876c4-5hvdl   1/1     Running   0          6m17s
app-6dd7f876c4-9vcp7   1/1     Running   0          6m17s
app-6dd7f876c4-k59mm   1/1     Running   0          6m17s
app-6dd7f876c4-trw8f   1/1     Running   0          6m17s
app-6dd7f876c4-wrhz8   1/1     Running   0          6m17s
[root@k8s-master health]#

接下来滚动更新应用

# cat app_v2.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: app
spec:
  replicas: 5
  template:
    metadata:
      labels:
        run: app
    spec:
      containers:
      - name: app
        image: busybox
        args:
        - /bin/sh
        - -c
        - sleep 30000
        readinessProbe:
          exec:
            command:
            - cat
            - /tmp/healthy
          initialDelaySeconds: 10
          periodSeconds: 5

[root@k8s-master health]# kubectl get pods
NAME                   READY   STATUS    RESTARTS   AGE
app-6dd7f876c4-5hvdl   1/1     Running   0          14m
app-6dd7f876c4-9vcp7   1/1     Running   0          14m
app-6dd7f876c4-k59mm   1/1     Running   0          14m
app-6dd7f876c4-trw8f   1/1     Running   0          14m
app-6dd7f876c4-wrhz8   1/1
[root@k8s-master health]#  kubectl apply -f app_v2.yaml
deployment.extensions/app configured
[root@k8s-master health]# kubectl get pods
NAME                   READY   STATUS              RESTARTS   AGE
app-6dd7f876c4-5hvdl   1/1     Running             0          14m
app-6dd7f876c4-9vcp7   1/1     Running             0          14m
app-6dd7f876c4-k59mm   1/1     Terminating         0          14m
app-6dd7f876c4-trw8f   1/1     Running             0          14m
app-6dd7f876c4-wrhz8   1/1     Running             0          14m
app-7fbf9d8fb7-g99hn   0/1     ContainerCreating   0          2s
app-7fbf9d8fb7-ltlv5   0/1     ContainerCreating   0          3s
[root@k8s-master health]# kubectl get pods
NAME                   READY   STATUS              RESTARTS   AGE
app-6dd7f876c4-5hvdl   1/1     Running             0          14m
app-6dd7f876c4-9vcp7   1/1     Running             0          14m
app-6dd7f876c4-k59mm   1/1     Terminating         0          14m
app-6dd7f876c4-trw8f   1/1     Running             0          14m
app-6dd7f876c4-wrhz8   1/1     Running             0          14m
app-7fbf9d8fb7-g99hn   0/1     ContainerCreating   0          9s
app-7fbf9d8fb7-ltlv5   0/1     Running             0          10s
[root@k8s-master health]#
[root@k8s-master health]# kubectl get pods
NAME                   READY   STATUS    RESTARTS   AGE
app-6dd7f876c4-5hvdl   1/1     Running   0          15m
app-6dd7f876c4-9vcp7   1/1     Running   0          15m
app-6dd7f876c4-trw8f   1/1     Running   0          15m
app-6dd7f876c4-wrhz8   1/1     Running   0          15m
app-7fbf9d8fb7-g99hn   0/1     Running   0          68s
app-7fbf9d8fb7-ltlv5   0/1     Running   0          69s
[root@k8s-master health]#

从 Pod 的 AGE 栏可判断，最后 2 个 Pod 是新副本，目前处于 NOT READY 状态。
旧副本从最初 5个减少到 4 个。

[root@k8s-master health]# kubectl get deployment app
NAME   READY   UP-TO-DATE   AVAILABLE   AGE
app    4/5     2            4           16m

l DESIRED 5 表示期望的状态是 5个 READY 的副本。

l UP-TO-DATE 2 表示当前已经完成更新的副本数：即 2 个新副本。

l AVAILABLE 4 表示当前处于 READY 状态的副本数：即 4个旧副本。

在我们的设定中，新副本始终都无法通过 Readiness 探测，所以这个状态会一直保持下去。

上面我们模拟了一个滚动更新失败的场景。不过幸运的是：Health Check 帮我们屏蔽了有缺陷的副本，同时保留了大部分旧副本，业务没有因更新失败受到影响。

滚动更新可以通过参数 maxSurge 和 maxUnavailable 来控制副本替换的数量。

查看全文

相关阅读:
PHP使用CURL详解
 PHP中使用sleep函数实现定时任务实例分享
 (实用篇)PHP ftp上传文件操作类
 php Pthread 多线程基本介绍
 用PHP发送POST请求
 面向对象【day08】：类的起源与metaclass（二）
面向对象【day07】：面向对象使用场景（十）
面向对象【day07】：知识点回顾（十一）
面向对象【day07】：析构函数（六）
面向对象【day07】：类的继承（七）

原文地址：https://www.cnblogs.com/yaohong/p/13040302.html