Dynamically Creating an SFS Turbo Subdirectory Using StorageClass

There is a minimum capacity limit when you create an SFS Turbo volume, so you are billed based on this minimum capacity rather than the capacity you have used. By default, when you mount an SFS Turbo volume to a workload, the root directory of the volume is mapped to the containers. The actual capacity used by the workload is far lower than the limit, leading to resource wastage.

The CCE Container Storage (Everest) add-on allows you to dynamically create subdirectories in an SFS Turbo file system and mount these subdirectories to containers. In this way, an SFS Turbo file system can be shared by multiple containers to increase storage efficiency.

• Only clusters of v1.15 or later are supported.
• The cluster must use the CCE Container Storage (Everest) add-on of version 1.1.13 or later.
• Kata containers are not supported.
• When the CCE Container Storage (Everest) add-on earlier than 1.2.69 or 2.1.11 is used, a maximum of 10 PVCs can be created concurrently using subdirectories. CCE Container Storage (Everest) of 1.2.69 or later or of 2.1.11 or later is recommended.
• A subPath volume is a subdirectory of an SFS Turbo file system. Increasing the capacity of a PVC of this type only changes the resource range specified by the PVC, but does not change the total capacity of the SFS Turbo file system. If the SFS Turbo file system's total resource capacity is not enough, the available capacity of the subPath volume will be restricted. To fix this, you must increase the resource capacity of the SFS Turbo file system on the SFS Turbo console.

  Deleting the subPath volume does not result in the deletion of the SFS Turbo file system.

1. Create an SFS Turbo file system in the same VPC and subnet as the cluster.
2. Create a YAML file of a storage class, for example, sfsturbo-subpath-sc.yaml.

   The following is an example:

   apiVersion: storage.k8s.io/v1
   allowVolumeExpansion: true
   kind: StorageClass
   metadata:
     name: sfsturbo-subpath-sc # Storage class name
   mountOptions:  #Mount options
   - lock
   parameters:
     csi.storage.k8s.io/csi-driver-name: sfsturbo.csi.everest.io
     csi.storage.k8s.io/fstype: nfs
     everest.io/archive-on-delete: "true"
     everest.io/share-expand-type: bandwidth
     everest.io/share-export-location: 192.168.1.1:/sfsturbo/  # Mount directory configuration
     everest.io/share-source: sfs-turbo
     everest.io/share-volume-type: STANDARD
     everest.io/volume-as: subpath
     everest.io/volume-id: 0d773f2e-1234-1234-1234-de6a35074696  # ID of an SFS Turbo volume
   provisioner: everest-csi-provisioner
   reclaimPolicy: Delete
   volumeBindingMode: Immediate
   
   

   Where:

   • name: indicates the name of the storage class.
   • mountOptions: (Optional) indicates the mount options.
     • In versions later than CCE Container Storage (Everest) 1.1.13 and earlier than 1.2.8, only the nolock parameter can be configured. By default, nolock is used for the mount operation and does not need to be configured. If nolock is set to false, the lock field is used.
     • Starting from CCE Container Storage (Everest) 1.2.8, additional mount options are supported. Do not set nolock to true. Otherwise, mount operations will fail.

       mountOptions:
       - vers=3
       - timeo=600
       - nolock
       - hard
       
       

   • everest.io/volume-as: This parameter is set to subpath to use a subPath volume.
   • everest.io/share-expand-type: (Optional) If the type of an SFS Turbo file system is SFS Turbo Standard – Enhanced or SFS Turbo Performance – Enhanced, set this parameter to bandwidth.
   • everest.io/share-export-location: indicates the mount directory. It consists of the SFS Turbo shared path and subdirectory. You can obtain the shared path on the SFS Turbo console and create the subdirectory by yourself. The PVCs created using the storage class are in this subdirectory.
   • everest.io/share-volume-type: (Optional) specifies the SFS Turbo file system type. The value can be STANDARD or PERFORMANCE. For enhanced types, this parameter must be used together with everest.io/share-expand-type (whose value should be bandwidth).
   • everest.io/zone: (Optional) Set it to the AZ where an SFS Turbo file system is located.
   • everest.io/volume-id: indicates the ID of an SFS Turbo volume. You can obtain the volume ID on the SFS Turbo console.
   • everest.io/archive-on-delete: When this parameter is set to true, and Delete is selected for Reclaim Policy, deleting a PVC deletes the subpath and archives the original PV data to a directory named archived-{$PV_NAME.TIMESTAMP}. (When a cluster is deleted, the subpath is neither deleted nor archived.) When this parameter is set to false, deleting a PVC deletes the SFS Turbo subpath without archiving the original PV data. The parameter is set to true by default.

3. Run kubectl create -f sfsturbo-subpath-sc.yaml.
4. Create a PVC YAML file named sfs-turbo-test.yaml.

   The following is an example:

   apiVersion: v1
   kind: PersistentVolumeClaim
   metadata:
     name: sfs-turbo-test  # PVC name
     namespace: default
   spec:
     accessModes:
     - ReadWriteMany
     resources:
       requests:
         storage: 50Gi
     storageClassName: sfsturbo-subpath-sc  # Storage class name
     volumeMode: Filesystem
   
   

   Where:

   • name: indicates the name of a PVC.
   • storageClassName: specifies the name of a storage class.
   • storage: In a subPath volume, modifying the value of this parameter does not impact the resource capacity of the SFS Turbo file system. A subPath volume is essentially a file path within an SFS Turbo file system. Increasing the capacity of the subPath volume in a PVC does not lead to an increase in the resources of the SFS Turbo file system.
     Note

     The capacity of a subPath volume is restricted by the overall resource capacity of the corresponding SFS Turbo file system. If the resources of the SFS Turbo file system are inadequate, you can adjust the resource capacity via the SFS Turbo console.

5. Run kubectl create -f sfs-turbo-test.yaml.

1. Create a YAML file for a Deployment, for example, deployment-test.yaml.

   The following is an example:

   apiVersion:  apps/v1 
   kind: Deployment
   metadata:
     name: test-turbo-subpath-example   # Name of the created workload
     namespace: default
     generation: 1
     labels:
       appgroup: ''
   spec: 
     replicas: 1 
     selector: 
       matchLabels: 
         app: test-turbo-subpath-example 
     template: 
       metadata: 
         labels: 
           app: test-turbo-subpath-example 
       spec: 
         containers: 
         - image: nginx:latest  # Image of the workload
           name: container-0 
           volumeMounts: 
           - mountPath: /tmp   #Mount path in a container
             name: pvc-sfs-turbo-example 
         restartPolicy: Always 
         imagePullSecrets:
         - name: default-secret
         volumes: 
         - name: pvc-sfs-turbo-example 
           persistentVolumeClaim: 
             claimName: sfs-turbo-test   # Name of an existing PVC
   
   

   Where:

   • name: indicates the name of the created workload.
   • image: specifies the image used by the workload.
   • mountPath: indicates the mount path of a container. In this example, the volume is mounted to the /tmp directory.
   • claimName: indicates the name of an existing PVC.

2. Create the Deployment.

   kubectl create -f deployment-test.yaml

1. Create a YAML file for a StatefulSet, for example, statefulset-test.yaml.

   The following is an example:

   apiVersion: apps/v1
   kind: StatefulSet
   metadata:
     name: test-turbo-subpath # Name of the created workload
     namespace: default
     generation: 1
     labels:
       appgroup: ''
   spec:
     replicas: 2
     selector:
       matchLabels:
         app: test-turbo-subpath
     template:
       metadata:
         labels:
           app: test-turbo-subpath
         annotations:
           metrics.alpha.kubernetes.io/custom-endpoints: '[{"api":"","path":"","port":"","names":""}]'
           pod.alpha.kubernetes.io/initialized: 'true'
       spec:
         containers:
           - name: container-0  
             image: 'nginx:latest'  # Image of a workload
             resources: {}
             volumeMounts:
               - name: sfs-turbo-160024548582479676
                 mountPath: /tmp  # Mount path in a container
             terminationMessagePath: /dev/termination-log
             terminationMessagePolicy: File
             imagePullPolicy: IfNotPresent
         restartPolicy: Always
         terminationGracePeriodSeconds: 30
         dnsPolicy: ClusterFirst
         securityContext: {}
         imagePullSecrets:
           - name: default-secret
         affinity: {}
         schedulerName: default-scheduler
     volumeClaimTemplates:
       - metadata:
           name: sfs-turbo-160024548582479676 
           namespace: default
           annotations: {}
         spec:
           accessModes:
             - ReadWriteMany
           resources:
             requests:
               storage: 10Gi
           storageClassName: sfsturbo-subpath-sc # Enter the name of a self-managed storage class.
     serviceName: wwww
     podManagementPolicy: OrderedReady
     updateStrategy:
       type: RollingUpdate
     revisionHistoryLimit: 10
   
   

   Where:

   • name: indicates the name of the created workload.
   • image: specifies the image used by the workload.
   • mountPath: indicates the mount path of a container. In this example, the volume is mounted to the /tmp directory.
   • spec.template.spec.containers.volumeMounts.name and spec.volumeClaimTemplates.metadata.name: must be consistent because they have a mapping relationship.
   • storageClassName: specifies the name of an on-premises storage class.

2. Create the StatefulSet.

   kubectl create -f statefulset-test.yaml