(kubectl) Creating a Deployment Mounted with an SFS Volume

Scenario

After an SFS volume is created or imported to CCE, you can mount the volume to a workload.

Notes and Constraints

The following configuration example applies to clusters of Kubernetes 1.13 or earlier.

Procedure

Use kubectl to connect to the cluster. For details, see Connecting to a Cluster Using kubectl.

Run the following commands to configure the sfs-deployment-example.yaml file, which is used to create a pod.

touch sfs-deployment-example.yaml

vi sfs-deployment-example.yaml

Example of mounting an SFS volume to a Deployment (PVC-based, shared volume):

apiVersion: apps/v1 
kind: Deployment 
metadata: 
  name: sfs-deployment-example                                # Workload name
  namespace: default 
spec: 
  replicas: 1 
  selector: 
    matchLabels: 
      app: sfs-deployment-example 
  template: 
    metadata: 
      labels: 
        app: sfs-deployment-example 
    spec: 
      containers: 
      - image: nginx 
        name: container-0 
        volumeMounts: 
        - mountPath: /tmp                                # Mount path 
          name: pvc-sfs-example 
      imagePullSecrets:
        - name: default-secret
      restartPolicy: Always 
      volumes: 
      - name: pvc-sfs-example 
        persistentVolumeClaim: 
          claimName: pvc-sfs-auto-example                # PVC name

**Table 1** Key parameters
Parent Parameter	Parameter	Description
metadata	name	Name of the pod to be created.
spec.template.spec.containers.volumeMounts	mountPath	Mount path in the container. In this example, the mount path is /tmp.
spec.template.spec.volumes.persistentVolumeClaim	claimName	Name of an existing PVC.

spec.template.spec.containers.volumeMounts.name and spec.template.spec.volumes.name must be consistent because they have a mapping relationship.

Example of mounting an SFS volume to a StatefulSet (PVC template-based, dedicated volume):

Currently, SFS file systems are sold out and cannot be exclusively used by defining the PVC template.

Example YAML:

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: deploy-sfs-nfs-rw-in
  namespace: default
  labels:
    appgroup: ''
spec:
  replicas: 2
  selector:
    matchLabels:
      app: deploy-sfs-nfs-rw-in
  template:
    metadata:
      labels:
        app: deploy-sfs-nfs-rw-in
    spec:
      containers:
        - name: container-0
          image: 'nginx:1.12-alpine-perl'
          volumeMounts:
            - name: bs-nfs-rw-mountoptionpvc
              mountPath: /aaa
      imagePullSecrets:
        - name: default-secret
  volumeClaimTemplates:
    - metadata:
        name: bs-nfs-rw-mountoptionpvc
        annotations:
          volume.beta.kubernetes.io/storage-class: nfs-rw
          volume.beta.kubernetes.io/storage-provisioner: flexvolume-huawei.com/fuxinfs
      spec:
        accessModes:
          - ReadWriteMany
        resources:
          requests:
            storage: 1Gi
  serviceName: wwww

**Table 2** Key parameters
Parent Parameter	Parameter	Description
metadata	name	Name of the created workload.
spec.template.spec.containers	image	Image of the workload.
spec.template.spec.containers.volumeMount	mountPath	Mount path in the container. In this example, the mount path is /tmp.
spec	serviceName	Service corresponding to the workload. For details about how to create a Service, see Creating a StatefulSet.

spec.template.spec.containers.volumeMounts.name and spec.volumeClaimTemplates.metadata.name must be consistent because they have a mapping relationship.

Run the following command to create the pod:

kubectl create -f sfs-deployment-example.yaml

After the creation is complete, log in to the CCE console. In the navigation pane, choose Resource Management > Storage > SFS. Click the PVC name. On the PVC details page, you can view the binding relationship between SFS and PVC.