无法在kubernetes pod中安装只读卷（使用AWS EKS中的EFS CSI驱动程序）

时间：2020-01-12 06:27:39

标签： amazon-web-services kubernetes nfs eks efs

我安装了EFS CI driver，并得到了他们的Static Provisioning示例：我能够启动附加到EFS卷上文件的Pod。我可以删除容器，然后启动另一个容器来检查该文件，并确认第一个容器写入的数据仍然存在。但是我实际上需要做的是将卷挂载为只读，而我在那里没有运气。

请注意，在成功运行该示例之后，我启动了一个EC2实例，并在其中安装了EFS文件系统，然后以只读方式添加了pod需要访问的数据。然后，我卸载了EFS文件系统并终止了实例。

使用以下配置（基于上面引用的静态配置示例），我的pod无法启动Running；它保留在ContainerCreating中。

存储类：

$ kubectl get sc efs-sc -o yaml
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"storage.k8s.io/v1","kind":"StorageClass","metadata":{"annotations":{},"name":"efs-sc"},"provisioner":"efs.csi.aws.com"}
  creationTimestamp: "2020-01-12T05:36:13Z"
  name: efs-sc
  resourceVersion: "809880"
  selfLink: /apis/storage.k8s.io/v1/storageclasses/efs-sc
  uid: 71ecce62-34fd-11ea-8a5f-124f4ee64e8d
provisioner: efs.csi.aws.com
reclaimPolicy: Delete
volumeBindingMode: Immediate

持久卷（这是集群中使用EFS存储类的唯一PV）：

$ kubectl get pv efs-pv-ro -o yaml
apiVersion: v1
kind: PersistentVolume
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"PersistentVolume","metadata":{"annotations":{},"name":"efs-pv-ro"},"spec":{"accessModes":["ReadOnlyMany"],"capacity":{"storage":"5Gi"},"csi":{"driver":"efs.csi.aws.com","volumeHandle":"fs-26120da7"},"persistentVolumeReclaimPolicy":"Retain","storageClassName":"efs-sc","volumeMode":"Filesystem"}}
    pv.kubernetes.io/bound-by-controller: "yes"
  creationTimestamp: "2020-01-12T05:36:59Z"
  finalizers:
  - kubernetes.io/pv-protection
  name: efs-pv-ro
  resourceVersion: "810231"
  selfLink: /api/v1/persistentvolumes/efs-pv-ro
  uid: 8d54a80e-34fd-11ea-8a5f-124f4ee64e8d
spec:
  accessModes:
  - ReadOnlyMany
  capacity:
    storage: 5Gi
  claimRef:
    apiVersion: v1
    kind: PersistentVolumeClaim
    name: efs-claim-ro
    namespace: default
    resourceVersion: "810229"
    uid: e0498cae-34fd-11ea-8a5f-124f4ee64e8d
  csi:
    driver: efs.csi.aws.com
    volumeHandle: fs-26120da7
  persistentVolumeReclaimPolicy: Retain
  storageClassName: efs-sc
  volumeMode: Filesystem
status:
  phase: Bound

持久卷声明（这是集群中唯一尝试使用EFS存储类的PVC：

$ kubectl get pvc efs-claim-ro -o yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"PersistentVolumeClaim","metadata":{"annotations":{},"name":"efs-claim-ro","namespace":"default"},"spec":{"accessModes":["ReadOnlyMany"],"resources":{"requests":{"storage":"5Gi"}},"storageClassName":"efs-sc"}}
    pv.kubernetes.io/bind-completed: "yes"
    pv.kubernetes.io/bound-by-controller: "yes"
  creationTimestamp: "2020-01-12T05:39:18Z"
  finalizers:
  - kubernetes.io/pvc-protection
  name: efs-claim-ro
  namespace: default
  resourceVersion: "810234"
  selfLink: /api/v1/namespaces/default/persistentvolumeclaims/efs-claim-ro
  uid: e0498cae-34fd-11ea-8a5f-124f4ee64e8d
spec:
  accessModes:
  - ReadOnlyMany
  resources:
    requests:
      storage: 5Gi
  storageClassName: efs-sc
  volumeMode: Filesystem
  volumeName: efs-pv-ro
status:
  accessModes:
  - ReadOnlyMany
  capacity:
    storage: 5Gi
  phase: Bound

这是Pod。它保留在ContainerCreating中，并且不会切换到Running：

$ kubectl get pod efs-app -o yaml
apiVersion: v1
kind: Pod
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"Pod","metadata":{"annotations":{},"name":"efs-app","namespace":"default"},"spec":{"containers":[{"args":["infinity"],"command":["sleep"],"image":"centos","name":"app","volumeMounts":[{"mountPath":"/data","name":"persistent-storage","subPath":"mmad"}]}],"volumes":[{"name":"persistent-storage","persistentVolumeClaim":{"claimName":"efs-claim-ro"}}]}}
    kubernetes.io/psp: eks.privileged
  creationTimestamp: "2020-01-12T06:07:08Z"
  name: efs-app
  namespace: default
  resourceVersion: "813420"
  selfLink: /api/v1/namespaces/default/pods/efs-app
  uid: c3b8421b-3501-11ea-b164-0a9483e894ed
spec:
  containers:
  - args:
    - infinity
    command:
    - sleep
    image: centos
    imagePullPolicy: Always
    name: app
    resources: {}
    terminationMessagePath: /dev/termination-log
    terminationMessagePolicy: File
    volumeMounts:
    - mountPath: /data
      name: persistent-storage
      subPath: mmad
    - mountPath: /var/run/secrets/kubernetes.io/serviceaccount
      name: default-token-z97dh
      readOnly: true
  dnsPolicy: ClusterFirst
  enableServiceLinks: true
  nodeName: ip-192-168-254-51.ec2.internal
  priority: 0
  restartPolicy: Always
  schedulerName: default-scheduler
  securityContext: {}
  serviceAccount: default
  serviceAccountName: default
  terminationGracePeriodSeconds: 30
  tolerations:
  - effect: NoExecute
    key: node.kubernetes.io/not-ready
    operator: Exists
    tolerationSeconds: 300
  - effect: NoExecute
    key: node.kubernetes.io/unreachable
    operator: Exists
    tolerationSeconds: 300
  volumes:
  - name: persistent-storage
    persistentVolumeClaim:
      claimName: efs-claim-ro
  - name: default-token-z97dh
    secret:
      defaultMode: 420
      secretName: default-token-z97dh
status:
  conditions:
  - lastProbeTime: null
    lastTransitionTime: "2020-01-12T06:07:08Z"
    status: "True"
    type: Initialized
  - lastProbeTime: null
    lastTransitionTime: "2020-01-12T06:07:08Z"
    message: 'containers with unready status: [app]'
    reason: ContainersNotReady
    status: "False"
    type: Ready
  - lastProbeTime: null
    lastTransitionTime: "2020-01-12T06:07:08Z"
    message: 'containers with unready status: [app]'
    reason: ContainersNotReady
    status: "False"
    type: ContainersReady
  - lastProbeTime: null
    lastTransitionTime: "2020-01-12T06:07:08Z"
    status: "True"
    type: PodScheduled
  containerStatuses:
  - image: centos
    imageID: ""
    lastState: {}
    name: app
    ready: false
    restartCount: 0
    state:
      waiting:
        reason: ContainerCreating
  hostIP: 192.168.254.51
  phase: Pending
  qosClass: BestEffort
  startTime: "2020-01-12T06:07:08Z"

我不确定subPath是否适用于此配置，但是无论subPath是否处于Pod配置中，都会发生相同的问题。

问题似乎与音量有关。如果我注释掉volumes和volumeMounts部分，则pod运行。

似乎PVC已绑定了正确的PV，但吊舱尚未启动。 我在上面的任何输出中都没有看到任何提示，但是也许我缺少了什么？

Kubernetes版本：

Client Version: version.Info{Major:"1", Minor:"14", GitVersion:"v1.14.8", GitCommit:"211047e9a1922595eaa3a1127ed365e9299a6c23", GitTreeState:"clean", BuildDate:"2019-10-15T12:11:03Z", GoVersion:"go1.12.10", Compiler:"gc", Platform:"darwin/amd64"}
Server Version: version.Info{Major:"1", Minor:"14+", GitVersion:"v1.14.9-eks-c0eccc", GitCommit:"c0eccca51d7500bb03b2f163dd8d534ffeb2f7a2", GitTreeState:"clean", BuildDate:"2019-12-22T23:14:11Z", GoVersion:"go1.12.12", Compiler:"gc", Platform:"linux/amd64"}

aws-efs-csi-driver版本：v.0.2.0。

1 个答案:

答案 0 :(得分：0)

请注意，其中一项要求是已安装版本 1.13.4 + 的Golang，但您具有 go1.12.12 。因此，您必须对其进行更新。如果要从旧版Go升级，则必须先删除现有版本。 在这里看看：upgrading-golang。

Kubernetes版本1.14和更高版本的Amazon EKS群集和辅助节点上支持此驱动程序。 Amazon EKS集群不支持Amazon EFS CSI驱动程序的Alpha功能。 无法在kubernetes容器中安装只读卷（使用AWS EKS中的EFS CSI驱动程序）。尝试将访问模式更改为：

accessModes:
 - ReadWriteMany

您可以在此处找到更多信息：efs-csi-driver。

确保在创建EFS文件系统时，可以从Kuberenetes集群访问它。这可以通过在与Kubernetes群集相同的VPC内创建文件系统或使用VPC对等来实现。

静态配置-首先需要手动创建EFS文件系统，然后可以使用驱动程序将其作为永久卷（PV）装入容器中。 挂载选项-可以在持久卷（PV）中指定挂载选项，以定义如何挂载该卷。除了常规的安装选项之外，您还可以将tls指定为安装选项，以在传输EFS文件系统时启用加密。

由于Amazon EFS是弹性文件系统，因此不会强制执行任何文件系统容量 限制。永久卷和永久卷声明中的实际存储容量值 创建文件系统时不使用。但是，由于存储容量是必填字段 在Kubernetes中，您必须指定一个有效值，例如本例中的5Gi。这个值确实 不限制您的Amazon EFS文件系统的大小

相关问题
最新问题