无法更新Kubernetes中的状态填充图

Question

我最近通过一个简单的测试项目评估了kubernetes，我能够以这样的命令来更新状态图的图像：

kubectl set image statefulset/cloud-stateful-set cloud-stateful-container=ncccloud:v716

我现在正在尝试使我们的真实系统在Kubernetes工作，即使我尝试更新映像时，Pods都不会做任何事情，即使我基本上使用了相同的命令。

它说：

statefulset.apps"云状态"图像更新

和 kubectl describe statefulset.apps/cloud-stateful-set说：

图像：ncccloud：v716"

但是kubectl describe pod cloud-stateful-set-0和kubectl describe pod cloud-stateful-set-1说：

"图像：ncccloud：最新"

ncccloud：最新是一个不起作用的图像：

$ kubectl get pods
NAME                               READY     STATUS             RESTARTS   AGE
cloud-stateful-set-0               0/1       CrashLoopBackOff   7          13m
cloud-stateful-set-1               0/1       CrashLoopBackOff   7          13m
mssql-deployment-6cd4ff766-pzz99   1/1       Running            1          55m

另一个奇怪的事情是，每当我尝试应用配置的状态满集而不是>不变的。

$ kubectl apply -f k8s/cloud-stateful-set.yaml
statefulset.apps "cloud-stateful-set" configured

这是我的云状态。yaml：

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: cloud-stateful-set
  labels:
    app: cloud
    group: service
spec:
  replicas: 2
#  podManagementPolicy: Parallel
  serviceName: cloud-stateful-set
  selector:
    matchLabels:
      app: cloud
  template:
    metadata:
      labels:
        app: cloud
        group: service
    spec:
      containers:
        - name: cloud-stateful-container
          image: ncccloud:latest
          imagePullPolicy: Never
          ports:
            - containerPort: 80
          volumeMounts:
            - name: cloud-stateful-storage
              mountPath: /cloud-stateful-data
  volumeClaimTemplates:
    - metadata:
        name: cloud-stateful-storage
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 10Mi

这是kubectl describe pod/cloud-stateful-set-1的完整输出：

Name:           cloud-stateful-set-1
Namespace:      default
Node:           docker-for-desktop/192.168.65.3
Start Time:     Tue, 02 Jul 2019 11:03:01 +0300
Labels:         app=cloud
                controller-revision-hash=cloud-stateful-set-5c9964c897
                group=service
                statefulset.kubernetes.io/pod-name=cloud-stateful-set-1
Annotations:    <none>
Status:         Running
IP:             10.1.0.20
Controlled By:  StatefulSet/cloud-stateful-set
Containers:
  cloud-stateful-container:
    Container ID:   docker://3ec26930c1a81caa39d5c5a16c4e25adf7584f90a71e0110c0b03ecb60dd9592
    Image:          ncccloud:latest
    Image ID:       docker://sha256:394427c40e964e34ca6c9db3ce1df1f8f6ce34c4ba8f3ab10e25da6e89678830
    Port:           80/TCP
    Host Port:      0/TCP
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Error
      Exit Code:    139
      Started:      Tue, 02 Jul 2019 11:19:03 +0300
      Finished:     Tue, 02 Jul 2019 11:19:03 +0300
    Ready:          False
    Restart Count:  8
    Environment:    <none>
    Mounts:
      /cloud-stateful-data from cloud-stateful-storage (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-gzxpx (ro)
Conditions:
  Type           Status
  Initialized    True
  Ready          False
  PodScheduled   True
Volumes:
  cloud-stateful-storage:
    Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
    ClaimName:  cloud-stateful-storage-cloud-stateful-set-1
    ReadOnly:   false
  default-token-gzxpx:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-gzxpx
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type     Reason                 Age                From                         Message
  ----     ------                 ----               ----                         -------
  Normal   Scheduled              19m                default-scheduler            Successfully assigned cloud-stateful-set-1 to docker-for-desktop
  Normal   SuccessfulMountVolume  19m                kubelet, docker-for-desktop  MountVolume.SetUp succeeded for volume "pvc-4c9e1796-9c9a-11e9-998f-00155d64fa03"
  Normal   SuccessfulMountVolume  19m                kubelet, docker-for-desktop  MountVolume.SetUp succeeded for volume "default-token-gzxpx"
  Normal   Pulled                 17m (x5 over 19m)  kubelet, docker-for-desktop  Container image "ncccloud:latest" already present on machine
  Normal   Created                17m (x5 over 19m)  kubelet, docker-for-desktop  Created container
  Normal   Started                17m (x5 over 19m)  kubelet, docker-for-desktop  Started container
  Warning  BackOff                4m (x70 over 19m)  kubelet, docker-for-desktop  Back-off restarting failed container

这是kubectl describe statefulset.apps/cloud-stateful-set的完整输出：

Name:               cloud-stateful-set
Namespace:          default
CreationTimestamp:  Tue, 02 Jul 2019 11:02:59 +0300
Selector:           app=cloud
Labels:             app=cloud
                    group=service
Annotations:        kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"apps/v1","kind":"StatefulSet","metadata":{"annotations":{},"labels":{"app":"cloud","group":"service"},"name":"cloud-stateful-set","names...
Replicas:           2 desired | 2 total
Pods Status:        2 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  app=cloud
           group=service
  Containers:
   cloud-stateful-container:
    Image:        ncccloud:v716
    Port:         80/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:
      /cloud-stateful-data from cloud-stateful-storage (rw)
  Volumes:  <none>
Volume Claims:
  Name:          cloud-stateful-storage
  StorageClass:
  Labels:        <none>
  Annotations:   <none>
  Capacity:      10Mi
  Access Modes:  [ReadWriteOnce]
Events:
  Type    Reason            Age   From                    Message
  ----    ------            ----  ----                    -------
  Normal  SuccessfulCreate  25m   statefulset-controller  create Pod cloud-stateful-set-0 in StatefulSet cloud-stateful-set successful
  Normal  SuccessfulCreate  25m   statefulset-controller  create Pod cloud-stateful-set-1 in StatefulSet cloud-stateful-set successful

如果很重要，我正在使用docker桌面。

Answer 1

在我的情况下，imagepullpolicy始终已经设置为：

kubectl patch statefulset web --type='json' -p='[{"op": "replace", "path": "/spec/template/spec/containers/0/image", "value":"gcr.io/google_containers/nginx-slim:0.8"}]'

在我的情况下，请参阅K8S文档：https：//kubernetes.io/docs/tutorials/stateful-application/basic-stateful-set/#rollling-update

Answer 2

在状态集yaml中，更改

imagePullPolicy: Never

to

imagePullPolicy: Always

Answer 3

我遇到了一个几乎相同的问题。我正在使用状态范围的单元舱。我已经始终设置了ImagePullPolicy。吊舱卡在退回循环中，不会使用新图像进行更新。

事实证明，陷入循环的图像规格是标签"最新标签"的图像。不存在；控制平面被卡在带有错误rpc error: code = NotFound desc = failed to pull and unpack image...

的ImagePullBackoff循环中

似乎需要删除和重新创建ImagePullbackoff状态中的POD。我相信对于crashloopbackoff状态的豆荚也是如此。

似乎也可以更新状态集的规格，而POD仍陷入了向后循环，并且不会更新。即状态满足规格和POD规格不匹配。

对我解决这种情况的是两种解决方案之一：

删除并重新创建POD(不是首选(

kubectl delete statefulset my-statefulset
kubectl apply -f my-statefulset.yml

(其中my-statefulset.yml是此POD的YAML规格(

或将吊舱缩放到零复制品，然后备份到一个(首选(

kubectl scale statefulset my-statefulset  --replicas=0 && 
kubectl scale statefulset my-statefulset  --replicas=1

此外，重新创建POD时可以添加spec.podManagementPolicy: Parallel。该策略允许创建一个新的POD，而无需等待删除卡在Crashloopbackoff中的杂物。最终，这就是我解决这个问题的未来发生的方式。

另请参见：

• kubernetes中带有crashloopbackoff状态的重新发行状态
• 如何清除crashloopbackoff