changing NFSServer name will keep PVC in Pending State

[Escaflow]

Is this a bug report or feature request?

• Bug Report

Deviation from expected behavior:
PersistentVolumeClaim stays Pending if NFSServer name is not rook-nfs.

Expected behavior:
PersistentVolumeClaim get Bound if NFSServer name is not rook-nfs.

How to reproduce it (minimal and precise):
Create all resources from the documentation till the PersistentVolumeClaim
kubectl apply -f common.yaml
kubectl apply -f operator.yaml
kubectl apply -f webhook.yaml

Create a PVC for the NFSServer:

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: nfs-demo-claim
  namespace: rook-nfs
spec:
  storageClassName: csi-cinder-classic
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 3Gi

Create the PVC for the pod that is using the NFSServer:

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: rook-nfs-pv-claim
spec:
  storageClassName: "rook-nfs-demo"
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 1Mi

Now create the NFSServer and StorageClass, this will work:

---
apiVersion: nfs.rook.io/v1alpha1
kind: NFSServer
metadata:
  name: rook-nfs
  namespace: rook-nfs
spec:
  replicas: 1
  exports:
    - name: demo
      server:
        accessMode: ReadWrite
        squash: "none"
      persistentVolumeClaim:
        claimName: nfs-demo-claim
  annotations:
    rook: nfs
---
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  labels:
    app: rook-nfs
  name: rook-nfs-demo
parameters:
  exportName: demo
  nfsServerName: rook-nfs
  nfsServerNamespace: rook-nfs
provisioner: nfs.rook.io/rook-nfs-provisioner
reclaimPolicy: Delete # Retain
volumeBindingMode: Immediate

This won't work:

---
apiVersion: nfs.rook.io/v1alpha1
kind: NFSServer
metadata:
  name: foo-bar
  namespace: rook-nfs
spec:
  replicas: 1
  exports:
    - name: demo
      server:
        accessMode: ReadWrite
        squash: "none"
      persistentVolumeClaim:
        claimName: nfs-demo-claim
  annotations:
    rook: nfs
---
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  labels:
    app: rook-nfs
  name: rook-nfs-demo
parameters:
  exportName: demo
  nfsServerName: foo-bar
  nfsServerNamespace: rook-nfs
provisioner: nfs.rook.io/rook-nfs-provisioner
reclaimPolicy: Delete # Retain
volumeBindingMode: Immediate

To get logs, use kubectl -n <namespace> logs <pod name>
rook-nfs-operator:

2021-07-19 17:24:54.978595 I | nfs-operator: Initialize status state
2021-07-19 17:24:55.022047 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result created
2021-07-19 17:24:55.038120 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result created
2021-07-19 17:24:55.054868 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result created
2021-07-19 17:24:55.063750 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result unchanged
2021-07-19 17:24:55.063921 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result unchanged
2021-07-19 17:24:55.070239 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result unchanged
2021-07-19 17:24:55.079261 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result unchanged
2021-07-19 17:24:55.079437 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result unchanged
2021-07-19 17:24:55.088250 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result unchanged
2021-07-19 17:25:05.064736 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result unchanged
2021-07-19 17:25:05.064909 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result unchanged
2021-07-19 17:25:05.072739 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result unchanged
2021-07-19 17:25:15.074214 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result unchanged
2021-07-19 17:25:15.074405 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result unchanged
2021-07-19 17:25:15.080917 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result unchanged
2021-07-19 17:25:15.114188 I | nfs-operator: Reconciling NFSServer ConfigMapOperation.Result unchanged
2021-07-19 17:25:15.114430 I | nfs-operator: Reconciling NFSServer ServiceOperation.Result unchanged
2021-07-19 17:25:15.121743 I | nfs-operator: Reconciling NFSServer StatefulSetOperation.Result unchanged
2021-07-19 17:27:35.943748 I | nfs-operator: Deleting NFSServer foo-bar in rook-nfs namespace

rook-nfs-webhook:

2021-07-19 17:24:54.967696 I | nfs-webhook: validate createnamefoo-bar
2021-07-19 17:24:54.984059 I | nfs-webhook: validate updatenamefoo-bar
2021-07-19 17:27:35.962977 I | nfs-webhook: validate updatenamefoo-bar
2021-07-19 17:27:35.981433 I | nfs-webhook: validate updatenamefoo-bar

Environment:

• OS: Ubuntu 18.04.5
• Kernel (e.g. uname -a): 4.15.0-147-generic
• Cloud provider or hardware configuration: OVH
• Rook version: v1.6.7
• Kubernetes version (use kubectl version): v1.20.2
• Kubernetes cluster type (e.g. Tectonic, GKE, OpenShift): OpenShift

[nohitme]

I was having the exact same issue. I started with a non "rook-nfs" server name, and was scratching my head for the problem until I saw this..

This effectively makes it impossible to deploy more than one NFSServer.

[bobbui]

i have exact same issue, reproduced it twice and im pretty sure this issue is legit.