Chapter 14. Creating exports using NFS
This section describes how to create exports using NFS that can then be accessed externally from the OpenShift cluster.
Follow the instructions below to create exports and access them externally from the OpenShift Cluster:
14.1. Enabling the NFS feature Copy linkLink copied to clipboard!
In order to use the NFS feature, it needs to be enabled in the cluster.
Prerequisites
- OpenShift Data Foundation is installed and running in the openshift-storage namespace.
- The OpenShift Data Foundation installation includes a CephFilesystem.
Procedure
Run the following commands to enable the NFS feature:
$ oc --namespace openshift-storage patch storageclusters.ocs.openshift.io ocs-storagecluster --type merge --patch '{"spec": {"nfs":{"enable": true}}}'
$ oc --namespace openshift-storage patch configmap rook-ceph-operator-config --type merge --patch '{"data":{"ROOK_CSI_ENABLE_NFS": "true"}}'
Verification steps
NFS installation and configuration is complete when the following conditions are met:
-
The CephNFS resource named
ocs-storagecluster-cephnfshas a status of Ready. Check all
csi-nfsplugin-*pods are running:oc -n openshift-storage describe cephnfs ocs-storagecluster-cephnfsoc -n openshift-storage get pod | grep csi-nfspluginOutput will be multiple pods. For example:
csi-nfsplugin-47qwq 2/2 Running 0 10s csi-nfsplugin-77947 2/2 Running 0 10s csi-nfsplugin-ct2pm 2/2 Running 0 10s csi-nfsplugin-provisioner-f85b75fbb-2rm2w 2/2 Running 0 10s csi-nfsplugin-provisioner-f85b75fbb-8nj5h 2/2 Running 0 10s
14.2. Creating NFS exports Copy linkLink copied to clipboard!
NFS exports are created by creating a Persistent Volume Claim (PVC) against the ocs-storagecluster-ceph-nfs StorageClass.
You can create NFS PVCs two ways:
Create NFS PVC using a yaml.
The following is an example PVC.
volumeMode: Block will not work for NFS volumes.
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: <desired_name>
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 1Gi
storageClassName: ocs-storagecluster-ceph-nfs
- <desired_name>
-
Specify a name for the PVC, for example,
my-nfs-export.
The export is created once the PVC reaches the Bound state.
Create NFS PVCs from the OpenShift Container Platform web console.
Prerequisites
- Ensure that you are logged into the OpenShift Container Platform web console and the NFS feature is enabled for the storage cluster.
Procedure
-
In the OpenShift Web Console, click Storage
Persistent Volume Claims - Set the Project to openshift-storage.
Click Create PersistentVolumeClaim.
-
Specify Storage Class,
ocs-storagecluster-ceph-nfs. -
Specify the PVC Name, for example,
my-nfs-export. - Select the required Access Mode.
- Specify a Size as per application requirement.
Select Volume mode as
Filesystem.Note:
Blockmode is not supported for NFS PVCs-
Click Create and wait until the PVC is in
Boundstatus.
-
Specify Storage Class,
14.3. Consuming NFS exports in-cluster Copy linkLink copied to clipboard!
Kubernetes application pods can consume NFS exports created by mounting a previously created PVC.
You can mount the PVC one of two ways:
Using a YAML:
Below is an example pod that uses the example PVC created in Section 14.2, “Creating NFS exports”:
apiVersion: v1
kind: Pod
metadata:
name: nfs-export-example
spec:
containers:
- name: web-server
image: nginx
volumeMounts:
- name: nfs-export-pvc
mountPath: /var/lib/www/html
volumes:
- name: nfs-export-pvc
persistentVolumeClaim:
claimName: <pvc_name>
readOnly: false
- <pvc_name>
-
Specify the PVC you have previously created, for example,
my-nfs-export.
Using the OpenShift Container Platform web console.
Procedure
-
On the OpenShift Container Platform web console, navigate to Workloads
Pods. - Click Create Pod to create a new application pod.
- Under the metadata section add a name. For example, nfs-export-example, with namespace as openshift-storage.
Under the spec: section, add containers: section with image and volumeMounts sections:
apiVersion: v1 kind: Pod metadata: name: nfs-export-example namespace: openshift-storage spec: containers: - name: web-server image: nginx volumeMounts: - name: <volume_name> mountPath: /var/lib/www/htmlFor example:
apiVersion: v1 kind: Pod metadata: name: nfs-export-example namespace: openshift-storage spec: containers: - name: web-server image: nginx volumeMounts: - name: nfs-export-pvc mountPath: /var/lib/www/htmlUnder the spec: section, add volumes: section to add the NFS PVC as a volume for the application pod:
volumes: - name: <volume_name> persistentVolumeClaim: claimName: <pvc_name>For example:
volumes: - name: nfs-export-pvc persistentVolumeClaim: claimName: my-nfs-export
14.4. Consuming NFS exports externally from the OpenShift cluster Copy linkLink copied to clipboard!
NFS clients outside of the OpenShift cluster can mount NFS exports created by a previously-created PVC.
Procedure
After the
nfsflag is enabled, singe-server CephNFS is deployed by Rook. You need to fetch the value of theceph_nfsfield for thenfs-ganeshaserver to use in the next step:$ oc get pods -n openshift-storage | grep rook-ceph-nfs$ oc describe pod <name of the rook-ceph-nfs pod> | grep ceph_nfsFor example:
$ oc describe pod rook-ceph-nfs-ocs-storagecluster-cephnfs-a-7bb484b4bf-bbdhs | grep ceph_nfs ceph_nfs=my-nfsExpose the NFS server outside of the OpenShift cluster by creating a Kubernetes LoadBalancer Service. The example below creates a LoadBalancer Service and references the NFS server created by OpenShift Data Foundation.
apiVersion: v1 kind: Service metadata: name: rook-ceph-nfs-ocs-storagecluster-cephnfs-load-balancer namespace: openshift-storage spec: ports: - name: nfs port: 2049 type: LoadBalancer externalTrafficPolicy: Local selector: app: rook-ceph-nfs ceph_nfs: <my-nfs> instance: aReplace
<my-nfs>with the value you got in step 1.Collect connection information. The information external clients need to connect to an export comes from the Persistent Volume (PV) created for the PVC, and the status of the LoadBalancer Service created in the previous step.
Get the share path from the PV.
Get the name of the PV associated with the NFS export’s PVC:
$ oc get pvc <pvc_name> --output jsonpath='{.spec.volumeName}' pvc-39c5c467-d9d3-4898-84f7-936ea52fd99dReplace
<pvc_name>with your own PVC name. For example:oc get pvc pvc-39c5c467-d9d3-4898-84f7-936ea52fd99d --output jsonpath='{.spec.volumeName}' pvc-39c5c467-d9d3-4898-84f7-936ea52fd99dUse the PV name obtained previously to get the NFS export’s share path:
$ oc get pv pvc-39c5c467-d9d3-4898-84f7-936ea52fd99d --output jsonpath='{.spec.csi.volumeAttributes.share}' /0001-0011-openshift-storage-0000000000000001-ba9426ab-d61b-11ec-9ffd-0a580a800215
Get an ingress address for the NFS server. A service’s ingress status may have multiple addresses. Choose the one desired to use for external clients. In the example below, there is only a single address: the host name
ingress-id.somedomain.com.$ oc -n openshift-storage get service rook-ceph-nfs-ocs-storagecluster-cephnfs-load-balancer --output jsonpath='{.status.loadBalancer.ingress}' [{"hostname":"ingress-id.somedomain.com"}]
Connect the external client using the share path and ingress address from the previous steps. The following example mounts the export to the client’s directory path
/export/mount/path:$ mount -t nfs4 -o proto=tcp ingress-id.somedomain.com:/0001-0011-openshift-storage-0000000000000001-ba9426ab-d61b-11ec-9ffd-0a580a800215 /export/mount/pathIf this does not work immediately, it could be that the Kubernetes environment is still taking time to configure the network resources to allow ingress to the NFS server.