docker - Kubernetes:无法将未格式化的卷安装为只读

Question

我尝试将 gcePersistentDisk 用作 ReadOnlyMany，以便多个节点上的 pod 可以读取该磁盘上的数据。遵循文档here对于相同的。

为了创建并稍后格式化 gce 持久磁盘，我已按照文档 here 中的说明进行操作。 。按照这篇文档，我已经通过 sshed 进入其中一个节点并格式化了磁盘。请参阅下面的完整错误以及其他 yaml 文件。

kubectl 描述 pods -l podName

Name:               punk-fly-nodejs-deployment-5dbbd7b8b5-5cbfs
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               gke-mycluster-default-pool-b1c1d316-d016/10.160.0.12
Start Time:         Thu, 25 Apr 2019 23:55:38 +0530
Labels:             app.kubernetes.io/instance=punk-fly
                    app.kubernetes.io/name=nodejs
                    pod-template-hash=1866836461
Annotations:        kubernetes.io/limit-ranger=LimitRanger plugin set: cpu request for container nodejs
Status:             Pending
IP:
Controlled By:      ReplicaSet/punk-fly-nodejs-deployment-5dbbd7b8b5
Containers:
  nodejs:
    Container ID:
    Image:          rajesh12/smartserver:server
    Image ID:
    Port:           3002/TCP
    Host Port:      0/TCP
    State:          Waiting
      Reason:       ContainerCreating
    Ready:          False

    Restart Count:  0
    Requests:
      cpu:  100m
    Environment:
      MYSQL_HOST:           mysqlservice
      MYSQL_DATABASE:       app
      MYSQL_ROOT_PASSWORD:  password
    Mounts:
      /usr/src/ from helm-vol (rw)
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-jpkzg (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  helm-vol:
    Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
    ClaimName:  my-readonly-pvc
    ReadOnly:   true
  default-token-jpkzg:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-jpkzg
    Optional:    false
QoS Class:       Burstable
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               2m                default-scheduler                                  Successfully assigned default/punk-fly-nodejs-deployment-5dbbd7b8b5-5cbfs to gke-mycluster-default-pool-b1c1d316-d016
  Normal   SuccessfulAttachVolume  1m                attachdetach-controller                            AttachVolume.Attach succeeded for volume "pvc-9c796180-677e-11e9-ad35-42010aa0000f"
  Warning  FailedMount             10s (x8 over 1m)  kubelet, gke-mycluster-default-pool-b1c1d316-d016  MountVolume.MountDevice failed for volume "pvc-9c796180-677e-11e9-ad35-42010aa0000f" : failed to mount unformatted volume as read only
  Warning  FailedMount             0s                kubelet, gke-mycluster-default-pool-b1c1d316-d016  Unable to mount volumes for pod "punk-fly-nodejs-deployment-5dbbd7b8b5-5cbfs_default(86293044-6787-11e9-ad35-42010aa0000f)": timeout expired waiting for volumes to attach or mount for pod "default"/"punk-fly-nodejs-deployment-5dbbd7b8b5-5cbfs". list of unmounted volumes=[helm-vol]. list of unattached volumes=[helm-vol default-token-jpkzg]

readonly_pv.yaml

apiVersion: v1
kind: PersistentVolume
metadata:
  name: my-readonly-pv
spec:
  storageClassName: ""
  capacity:
    storage: 1G
  accessModes:
    - ReadOnlyMany
  gcePersistentDisk:
    pdName: mydisk0
    fsType: ext4
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: my-readonly-pvc
spec:
  accessModes:
    - ReadOnlyMany
  resources:
    requests:
      storage: 1G

deployment.yaml

  volumes:
    - name: helm-vol
      persistentVolumeClaim:
        claimName: my-readonly-pvc
        readOnly: true
  containers:
    - name: {{ .Values.app.backendName }}
      image: "{{ .Values.image.repository }}:{{ .Values.image.tagServer }}"
      imagePullPolicy: {{ .Values.image.pullPolicy }}
      env:
      - name: MYSQL_HOST
        value: mysqlservice
      - name: MYSQL_DATABASE
        value: app
      - name: MYSQL_ROOT_PASSWORD
        value: password
      ports:
        - name: http-backend
          containerPort: 3002
      volumeMounts:
        - name: helm-vol
          mountPath: /usr/src/

Answer 1

听起来像你的 PVC正在动态配置未使用 default StorageClass 格式化的新卷

您的 Pod 可能是在与 PV 不同的可用性中创建的。已提供。 gce 卷拥有多个 Pod 读取器的问题是 Pod 必须始终位于同一可用区。

一些选项:

• 只需创建并格式化 PV与您的节点位于同一可用区。

• 定义 PV 时，您可以指定 Node Affinity以确保它始终被分配给特定的节点。

• 定义 StorageClass指定文件系统

  kind: StorageClass
  apiVersion: storage.k8s.io/v1
  metadata:
    name: mysc
  provisioner: kubernetes.io/aws-ebs
  parameters:
    type: gp2
    fsType: ext4
  

  然后在 PVC 中使用它:

  apiVersion: v1
  kind: PersistentVolumeClaim
  metadata:
    name: my-pvc
  spec:
    accessModes:
      - ReadOnlyMany
    resources:
      requests:
        storage: 1G
    storageClassName: mysc
  

  该卷将自动配置和格式化。