This page shows how to set the default StorageClass for a GKE on-prem cluster. See also Storage.
Overview
GKE on-prem can integrate with block or file storage by using any of the following mechanisms:
When you create a cluster, GKE on-prem creates a Kubernetes
StorageClass
named standard. This is the default StorageClass for the cluster.
To see detailed information about the standard StorageClass, enter the
following command:
kubectl --kubeconfig [CLUSTER_KUBECONFIG] get storageclass standard --output yaml
where [CLUSTER_KUBECONFIG] is the path of the kubeconfig file for your cluster.
In the output, you can see that the name of the StorageClass is standard. You
can also see the storageclass.kubernetes.io/is-default-class: "true"
annotation. This annotation identifies the StorageClass named standard as the
default StorageClass.
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
annotations:
storageclass.kubernetes.io/is-default-class: "true"
...
name: standard
...
parameters:
datastore: HOST2-DATASTORE2
diskformat: thin
fstype: ext4
provisioner: kubernetes.io/vsphere-volume
reclaimPolicy: Delete
volumeBindingMode: Immediate
Your cluster can have several StorageClass objects, but only one of them can
have the storageclass.kubernetes.io/is-default-class: "true" annotation and
therefore be the default StorageClass.
In the preceding output, you can also see that the provisioner is
kubernetes.io/vsphere-volume.
This is the provisioner that takes action when a client requests a piece of
storage of class standard.
Default StorageClass
When you request storage, you can specify a StorageClass. If you do not specify a StorageClass, the default StorageClass is used. For example, suppose you create a PersistentVolumeClaim that does not specify a StorageClass. The volume controller will fulfill the claim according to the default StorageClass.
Changing the default StorageClass
As a cluster administrator, you might want to change the default storage class. Then all requests for storage that do not specify a StorageClass will be fulfilled according to the StorageClass of your choice. This section gives the steps for changing the default.
Deploy a new storage system
Deploy a new storage system and any software components for integrating the new storage mechanism with a Kubernetes cluster. For example, you might need to install a CSI driver in the cluster.
This step depends on the storage vendor you're using. For CSI drivers, vendors should provide instructions for deploying their CSI driver to a Kubernetes cluster. A CSI driver's documentation should also include the driver-specific parameters that you provide in your StorageClass, including the provisioner name.
When you create a StorageClass for your new appliance, you should name the StorageClass after its properties (such as "fast" or "highly-replicated"), rather than after the name of the specific driver or appliance behind it. This will make it easier to have consistent storage policies across clusters and environments.
Remove the default annotation from the standard StorageClass
To open the standard StorageClass manifest in a text editor, enter the
following command:
kubectl --kubeconfig [CLUSTER_KUBECONFIG] edit storageclass standard
In the text editor, remove the
storageclass.kubernetes.io/is-default-class: "true" annotation. Close the
text editor.
To verify that the annotation was removed, enter this command:
kubectl --kubeconfig [CLUSTER_KUBECONFIG] get storageclass standard --output yaml
Create a new StorageClass
Create a manifest for a new StorageClass.
Include the storageclass.kubernetes.io/is-default-class: "true" annotation.
For example:
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
annotations:
storageclass.kubernetes.io/is-default-class: "true"
...
name: my-storage-class
...
parameters:
...
provisioner: [MY_PROVISIONER]
...
Save your manifest as a YAML file, and create the new StorageClass:
kubectl --kubeconfig [CLUSTER_KUBECONFIG] apply -f [MANIFEST_FILE]
where [MANIFEST_FILE] is the path to your new StorageClass manifest file.