Example Kubernetes StorageClasses for ONTAP AI Deployments

Contributors netapp-dorianh kevin-hoke mboglesby Download PDF of this page

Before you can use Trident to dynamically provision storage resources within your Kubernetes cluster, you must create one or more Kubernetes StorageClasses. The examples that follow represent different types of StorageClasses that you might want to create if you are deploying the NetApp AI Control Plane solution on an ONTAP AI pod. For more information about StorageClasses, see the Trident documentation.

  1. NetApp recommends creating a separate StorageClass for each FlexGroup-enabled Trident Backend that you created in the section Example Trident Backends for ONTAP AI Deployments, step 1. These granular StorageClasses enable you to add NFS mounts that correspond to specific LIFs (the LIFs that you specified when you created the Trident Backends) as a particular Backend that is specified in the StorageClass spec file. The example commands that follow show the creation of two StorageClasses that correspond to the two example Backends that were created in the section Example Trident Backends for ONTAP AI Deployments, step 1. For more information about StorageClasses, see the Trident documentation.

    So that a persistent volume isn’t deleted when the corresponding PersistentVolumeClaim (PVC) is deleted, the following example uses a reclaimPolicy value of Retain. For more information about the reclaimPolicy field, see the official Kubernetes documentation.

    $ cat << EOF > ./storage-class-ontap-ai-flexgroups-retain-iface1.yaml
    apiVersion: storage.k8s.io/v1
    kind: StorageClass
    metadata:
      name: ontap-ai-flexgroups-retain-iface1
    provisioner: netapp.io/trident
    parameters:
      backendType: "ontap-nas-flexgroup"
      storagePools: "ontap-ai-flexgroups-iface1:.*"
    reclaimPolicy: Retain
    EOF
    $ kubectl create -f ./storage-class-ontap-ai-flexgroups-retain-iface1.yaml
    storageclass.storage.k8s.io/ontap-ai-flexgroups-retain-iface1 created
    $ cat << EOF > ./storage-class-ontap-ai-flexgroups-retain-iface2.yaml
    apiVersion: storage.k8s.io/v1
    kind: StorageClass
    metadata:
      name: ontap-ai-flexgroups-retain-iface2
    provisioner: netapp.io/trident
    parameters:
      backendType: "ontap-nas-flexgroup"
      storagePools: "ontap-ai-flexgroups-iface2:.*"
    reclaimPolicy: Retain
    EOF
    $ kubectl create -f ./storage-class-ontap-ai-flexgroups-retain-iface2.yaml
    storageclass.storage.k8s.io/ontap-ai-flexgroups-retain-iface2 created
    $ kubectl get storageclass
    NAME                                PROVISIONER         AGE
    ontap-ai-flexgroups-retain-iface1   netapp.io/trident   0m
    ontap-ai-flexgroups-retain-iface2   netapp.io/trident   0m
  2. NetApp also recommends creating a StorageClass that corresponds to the FlexVol-enabled Trident Backend that you created in the section Example Trident Backends for ONTAP AI Deployments, step 2. The example commands that follow show the creation of a single StorageClass for FlexVol volumes.

    In the following example, a particular Backend is not specified in the StorageClass definition file because only one FlexVol-enabled Trident backend was created. When you use Kubernetes to administer volumes that use this StorageClass, Trident attempts to use any available backend that uses the ontap-nas driver.

    $ cat << EOF > ./storage-class-ontap-ai-flexvols-retain.yaml
    apiVersion: storage.k8s.io/v1
    kind: StorageClass
    metadata:
      name: ontap-ai-flexvols-retain
    provisioner: netapp.io/trident
    parameters:
      backendType: "ontap-nas"
    reclaimPolicy: Retain
    EOF
    $ kubectl create -f ./storage-class-ontap-ai-flexvols-retain.yaml
    storageclass.storage.k8s.io/ontap-ai-flexvols-retain created
    $ kubectl get storageclass
    NAME                                PROVISIONER         AGE
    ontap-ai-flexgroups-retain-iface1   netapp.io/trident   1m
    ontap-ai-flexgroups-retain-iface2   netapp.io/trident   1m
    ontap-ai-flexvols-retain            netapp.io/trident   0m
  3. NetApp also recommends creating a generic StorageClass for FlexGroup volumes. The following example commands show the creation of a single generic StorageClass for FlexGroup volumes.

    Note that a particular backend is not specified in the StorageClass definition file. Therefore, when you use Kubernetes to administer volumes that use this StorageClass, Trident attempts to use any available backend that uses the ontap-nas-flexgroup driver.

$ cat << EOF > ./storage-class-ontap-ai-flexgroups-retain.yaml
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: ontap-ai-flexgroups-retain
provisioner: netapp.io/trident
parameters:
  backendType: "ontap-nas-flexgroup"
reclaimPolicy: Retain
EOF
$ kubectl create -f ./storage-class-ontap-ai-flexgroups-retain.yaml
storageclass.storage.k8s.io/ontap-ai-flexgroups-retain created
$ kubectl get storageclass
NAME                                PROVISIONER         AGE
ontap-ai-flexgroups-retain          netapp.io/trident   0m
ontap-ai-flexgroups-retain-iface1   netapp.io/trident   2m
ontap-ai-flexgroups-retain-iface2   netapp.io/trident   2m
ontap-ai-flexvols-retain            netapp.io/trident   1m