mirror of
https://github.com/ceph/ceph-csi.git
synced 2024-11-15 02:40:23 +00:00
461effe520
specify from what ceph versions we need to run nfs service commands and when to skip it. Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
74 lines
2.3 KiB
Markdown
74 lines
2.3 KiB
Markdown
# Dynamic provisioning with NFS
|
|
|
|
The easiest way to try out the examples for dynamic provisioning with NFS, is
|
|
to use [Rook Ceph with CephNFS][rook_ceph]. Rook can be used to deploy a Ceph
|
|
cluster. Ceph is able to maintain a NFS-Ganesha service with a few commands,
|
|
making configuring the Ceph cluster a minimal effort.
|
|
|
|
## Enabling the Ceph NFS-service
|
|
|
|
Ceph does not enable the NFS-service by default. In order for Rook Ceph to be
|
|
able to configure NFS-exports, the NFS-service needs to be configured first.
|
|
|
|
- Required for Ceph v16.2.7 and below
|
|
- Optional for Ceph v16.2.8 and above
|
|
|
|
In the [Rook Toolbox][rook_toolbox], run the following commands:
|
|
|
|
```console
|
|
ceph osd pool create nfs-ganesha
|
|
ceph mgr module enable rook
|
|
ceph mgr module enable nfs
|
|
ceph orch set backend rook
|
|
```
|
|
|
|
## Create a NFS-cluster
|
|
|
|
In the directory where this `README` is located, there is an example
|
|
`rook-nfs.yaml` file. This file can be used to create a Ceph managed
|
|
NFS-cluster with the name "my-nfs".
|
|
|
|
```console
|
|
$ kubectl create -f rook-nfs.yaml
|
|
cephnfs.ceph.rook.io/my-nfs created
|
|
```
|
|
|
|
The CephNFS resource will create a NFS-Ganesha Pod and Service with label
|
|
`app=rook-ceph-nfs`:
|
|
|
|
```console
|
|
$ kubectl get pods -l app=rook-ceph-nfs
|
|
NAME READY STATUS RESTARTS AGE
|
|
rook-ceph-nfs-my-nfs-a-5d47f66977-sc2rk 2/2 Running 0 61s
|
|
$ kubectl get service -l app=rook-ceph-nfs
|
|
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
|
|
rook-ceph-nfs-my-nfs-a ClusterIP 172.30.218.195 <none> 2049/TCP 2m58s
|
|
```
|
|
|
|
## Create a StorageClass
|
|
|
|
The parameters of the StorageClass reflect mostly what CephFS requires to
|
|
connect to the Ceph cluster. All required options are commented clearly in the
|
|
`storageclass.yaml` file.
|
|
|
|
In addition to the CephFS parameters, there are:
|
|
|
|
- `nfsCluster`: name of the Ceph managed NFS-cluster (here `my-nfs`)
|
|
- `server`: hostname/IP/service of the NFS-server (here `172.30.218.195`)
|
|
|
|
Edit `storageclass.yaml`, and create the resource:
|
|
|
|
```console
|
|
$ kubectl create -f storageclass.yaml
|
|
storageclass.storage.k8s.io/csi-nfs-sc created
|
|
```
|
|
|
|
## TODO: next steps
|
|
|
|
- deploy the NFS-provisioner
|
|
- deploy the kubernetes-csi/csi-driver-nfs
|
|
- create the CSIDriver object
|
|
|
|
[rook_ceph]: https://rook.io/docs/rook/latest/ceph-nfs-crd.html
|
|
[rook_toolbox]: https://rook.io/docs/rook/latest/ceph-toolbox.html
|