docker - hostpath.so 中的只读文件系统无法挂载卷

Question

当我提供主机路径时，它显示它是只读文件系统，因为我是 kubernetes 的新手我没有找到任何其他方式请告诉我有没有其他方式实现卷，我正在这样做在 GKE 上
这是我的 yaml 代码

apiVersion: apps/v1
kind: Deployment
metadata:
  annotations:
    deployment.kubernetes.io/revision: "10"
  creationTimestamp: "2019-11-22T10:52:16Z"
  generation: 17
  labels:
    app: dataset
  name: dataset
  namespace: default
  resourceVersion: "283767"
  selfLink: /apis/apps/v1/namespaces/default/deployments/dataset
  uid: 26111fe8-0d16-11ea-a66e-42010aa00042
spec:
  progressDeadlineSeconds: 600
  replicas: 2
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      app: dataset
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: dataset
    spec:
      containers:
      - env:
        - name: RABBIT_MQ_HOST
          valueFrom:
            configMapKeyRef:
              key: RABBIT_MQ_HOST
              name: dataset-config
        - name: RABBIT_MQ_USER
          valueFrom:
            configMapKeyRef:
              key: RABBIT_MQ_USER
              name: dataset-config
        - name: RABBIT_MQ_PASSWORD
          valueFrom:
            configMapKeyRef:
              key: RABBIT_MQ_PASSWORD
              name: dataset-config
        - name: DATASET_DB_HOST
          valueFrom:
            configMapKeyRef:
              key: DATASET_DB_HOST
              name: dataset-config
        - name: DATASET_DB_NAME
          valueFrom:
            configMapKeyRef:
              key: DATASET_DB_NAME
              name: dataset-config
        - name: LICENSE_SERVER
          valueFrom:
            configMapKeyRef:
              key: LICENSE_SERVER
              name: dataset-config
        - name: DATASET_THUMBNAIL_SIZE
          valueFrom:
            configMapKeyRef:
              key: DATASET_THUMBNAIL_SIZE
              name: dataset-config
        - name: GATEWAY_URL
          valueFrom:
            configMapKeyRef:
              key: GATEWAY_URL
              name: dataset-config
        - name: DEFAULT_DATASOURCE_ID
          valueFrom:
            configMapKeyRef:
              key: DEFAULT_DATASOURCE_ID
              name: dataset-config
        - name: RABBIT_MQ_QUEUE_NAME
          valueFrom:
            configMapKeyRef:
              key: RABBIT_MQ_QUEUE_NAME
              name: dataset-config
        - name: RABBIT_MQ_PATTERN
          valueFrom:
            configMapKeyRef:
              key: RABBIT_MQ_PATTERN
              name: dataset-config
        image: gcr.io/gcr-testing-258008/dataset@sha256:8416ec9b023d4a4587a511b855c2735b25a16dbb1a15531d8974d0ef89ad3d73
        imagePullPolicy: IfNotPresent
        name: dataset-sha256
        resources: {}
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        volumeMounts:
        - mountPath: ./data/uploads
          name: dataset-volume-uploads
        - mountPath: ./data/thumbnails
          name: dataset-volume-thumbnails
      dnsPolicy: ClusterFirst
      restartPolicy: Always
      schedulerName: default-scheduler
      securityContext: {}
      terminationGracePeriodSeconds: 30
      volumes:
      - hostPath:
          path: /build/uploads
          type: ""
        name: dataset-volume-uploads
      - hostPath:
          path: /build/thumbnails
          type: ""
        name: dataset-volume-thumbnails
status:
  availableReplicas: 2
  conditions:
  - lastTransitionTime: "2019-11-23T07:19:13Z"
    lastUpdateTime: "2019-11-23T07:19:13Z"
    message: Deployment has minimum availability.
    reason: MinimumReplicasAvailable
    status: "True"
    type: Available
  - lastTransitionTime: "2019-11-23T06:31:03Z"
    lastUpdateTime: "2019-11-23T07:24:42Z"
    message: ReplicaSet "dataset-75b46f868f" is progressing.
    reason: ReplicaSetUpdated
    status: "True"
    type: Progressing
  observedGeneration: 17
  readyReplicas: 2
  replicas: 3
  unavailableReplicas: 1
  updatedReplicas: 1

这是我对 pod 的描述

    Path:          /build/uploads
    HostPathType:  
  dataset-volume-thumbnails:
    Type:          HostPath (bare host directory volume)
    Path:          /build/thumbnails
    HostPathType:  
  default-token-x2wmw:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-x2wmw
    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  96s                default-scheduler                                 Successfully assigned default/dataset-75b46f868f-wffm7 to gke-teric-ai-default-pool-41929025-fxnx
  Warning  BackOff    15s (x6 over 93s)  kubelet, gke-teric-ai-default-pool-41929025-fxnx  Back-off restarting failed container
  Normal   Pulled     2s (x5 over 95s)   kubelet, gke-teric-ai-default-pool-41929025-fxnx  Container image "gcr.io/gcr-testing-258008/dataset@sha256:8416ec9b023d4a4587a511b855c2735b25a16dbb1a15531d8974d0ef89ad3d73" already present on machine
  Normal   Created    2s (x5 over 95s)   kubelet, gke-teric-ai-default-pool-41929025-fxnx  Created container
  Warning  Failed     1s (x5 over 95s)   kubelet, gke-teric-ai-default-pool-41929025-fxnx  Error: failed to start container "dataset-sha256": Error response from daemon: error while creating mount source path '/build/uploads': mkdir /build/uploads: read-only file system

所以这就是问题所在，即使我动态地授予 chmod 权限，它也不允许进行写入操作。我尝试过持久卷它也不起作用，所以请告诉我我必须以哪种方式挂载卷。

Answer 1

首先，不要使用hostPath，去持久化存储

如果您需要存储一些数据，我会使用 StatefulSets 而不是 Deployment。

我能够以 root 用户身份在我的 GKE 实例上手动创建两个 hostPath。

我想您必须为 hostPath 指定类型才能创建请求目录(如果它不存在)。 type: DirectoryOrCreate您可以阅读更多关于 hostPath和可用的类型值。
此外，如果您在容器内使用用户的 hostPath 权限必须匹配节点上的所有权，因此它会变得更加复杂，当然，您可以以 root 身份运行它，但不推荐这种方式。

总而言之，只需使用 google 提供的持久存储。如果您遇到权限问题，您可能需要 init contaner更改权限或您必须设置正确 fsGroup为您的容器。