Multiple pods ReadWriteMany mount with static provisioning

[Shaked]

What happened:

Hey, I'm trying to use https://github.com/kubernetes-csi/csi-driver-nfs with a static provisioning while I want to have multiple pods to mount a ReadWriteMany NFS server. Unfortunately the pods always get stuck in ContainerCreating with the following error:

Warning FailedMount 3m40s (x1662 over 48m) kubelet MountVolume.SetUp failed for volume "clearml-agent-caching" : rpc error: code = Aborted desc = An operation with the given Volume ID shaked-test1 already exists

What you expected to happen:

Allow all pods to mount the NFS server while all pods have access to the same files i.e /exports content should be available for all pods (therefore I don't want to use dynamic provisioning).

How to reproduce it:

Create a PV and PVC:

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: clearml-agent-claim
  namespace: clearml
spec:
  selector:
    matchLabels:
      {{- include "clearml-agent.labels" . | nindent 6 }}
  accessModes:
    - ReadWriteMany
  storageClassName: ""
  volumeName:  clearml-agent
  resources:
    requests:
      storage: 100Gi
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: clearml-agent
  labels:
    {{- include "clearml-agent.labels" . | nindent 6 }}
spec:
  capacity:
    storage: 100Gi
  accessModes:
    - ReadWriteMany
  mountOptions:
    - hard
    - cvers=4.1
  csi:
    driver: nfs.csi.k8s.io
    readOnly: false
    volumeHandle: "shaked-test1" # {{ uuidv4 }}  # make sure it's a unique id in the cluster
    volumeAttributes:
      server: clearml-agent.clearml.svc.cluster.local
      share: /

Create 2 replicas or more:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nfs-test-deployment
  labels:
    app: nfs-test
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nfs-test
  template:
    metadata:
      labels:
        app: nfs-test
    spec:
      containers:
      - name: nfs-shaked-test
        image: d3fk/nfs-client
        command: [ "sleep" ]
        args: [ "infinity" ]
        securityContext:
          privileged: true
        resources: {}
        env:
          - name: SHARE
            value: "/shakedsnfs"
          - name: SERVER
            value: "clearml-agent-caching.clearml.svc.cluster.local"
        volumeMounts:
          - mountPath: /shakedsnfs
            name: caching-nfs
            readOnly: false
      volumes:
        - name: caching-nfs
          persistentVolumeClaim:
            claimName: clearml-agent-caching-claim

Create an NFS server as in the example.

Anything else we need to know?:

Environment: GKE

• CSI Driver version: Lateset
• Kubernetes version (use kubectl version):

Client Version: version.Info{Major:"1", Minor:"19", GitVersion:"v1.19.4", GitCommit:"d360454c9bcd1634cf4cc52d1867af5491dc9c5f", GitTreeState:"clean", BuildDate:"2020-11-12T01:09:16Z", GoVersion:"go1.15.4", Compiler:"gc", Platform:"darwin/amd64"}
Server Version: version.Info{Major:"1", Minor:"18+", GitVersion:"v1.18.17-gke.1900", GitCommit:"c97b098770a3dcd2e36fdccd8ec5642a5cddbd7f", GitTreeState:"clean", BuildDate:"2021-04-14T09:22:16Z", GoVersion:"go1.13.15b4", Compiler:"gc", Platform:"linux/amd64"}

[andyzhangx]

already fixed on master branch, pls try again, make sure that latest image is downloaded on your node(set imagePullPolicy: true)