Skip to main content
EARLY ACCESS

This feature is available as Early Access (EA) and should not be used in production.

Version: 3.6

FlashArray/FlashBlade Driver

Portworx Enterprise includes a FA/FB driver that manages FlashArray Direct Access (FADA), FlashBlade Direct Access (FBDA), and FlashArray File volumes by using Kubernetes-native resources, including persistent volume claims (PVCs), VolumeSnapshot resources, and custom resources.

Enabling the FA/FB driver is optional. If you do not enable it, you can continue to manage Direct Access volumes by using the existing Portworx Enterprise functionality.

note

The FA/FB driver is a component of Portworx Enterprise and cannot be upgraded independently.

Prepare your Environment​

Enable the FA/FB driver​

To enable the FA/FB driver, add the following annotations to the StorageCluster spec during new Portworx installation or to an existing Portworx cluster:

metadata:
  annotations:
    portworx.io/pure-csi-integration: "true"

Verify​

When the FA/FB driver is enabled, it manages the FADA, FBDA, and FA File volumes. To verify that the FA/FB driver is enabled, run the following commands:

  1. For an existing Portworx cluster, view the running jobs. For a new Portworx installation, skip this step.

    kubectl get jobs -n <portworx> | grep pure-csi-migrator
    NAME                     STATUS     COMPLETIONS   DURATION   AGE
    post-pure-csi-migrator   Complete   1/1           5s         41s
    pre-pure-csi-migrator    Complete   1/1           8s         51s

    These are the Kubernetes Jobs that will be perform migration for existing Portworx cluster where:

    • pre-pure-csi-migrator: Reads existing metadata from KVDB and old Portworx Enterprise style state to create the new Kubernetes CustomResources (e.g. PureVolume, PureSnapshot, PureStorageCluster, and more) that the FA/FB driver needs. It is responsible for the "heavy" migration work like enumerating volumes, snapshots, cloudsnap metadata, and writing them into K8s CRs.

    • post-pure-csi-migrator: Runs after the main metadata migration is done and after the new px-pure-csi-* pods are coming up. It performs post-processing and reconciliation steps so the new driver and operator see a consistent state. This runs a final checks that all volumes/snapshots migrated correctly into CRs.

    The migration is only run when you have an existing Portworx cluster. For new Portworx Enterprise installation with the FA/FB driver enabled, the migration does not happen.

  2. View the daemonset for new and existing Portworx cluster:

    kubectl get daemonset -n <portworx> | grep px-pure-csi-node
    NAME                 DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR        AGE
    px-pure-csi-node     6         6         6       6            6           kubernetes.io/os=linux   34s

    The DaemonSet px-pure-csi-node here is scheduled on 6 nodes. All 6 pods are running, updated, and ready. Once migration finishes, these pods are responsible for mounting the attached volumes to the directories exposed and used by the pods.

  3. View the purestoragecluster for new and existing Portworx cluster:

    kubectl get purestoragecluster -n portworx
    NAME       CLUSTER UUID   STATUS    VERSION     AGE
    test-stc   xxxxxx  xxxx  Running     3.6.0     75s

    The FA/FB driver custom resource that is recognized and reconciled by the Portworx Operator has running status.

Additional features with the FA/FB driver​

The Portworx Volume names are usually integers but when the FA/FB driver is enabled, the FADA and FBDA volume names can be a UUID or same as the PV name. The FA/FB driver provides the following additional features:

Enable additional FA/FB driver features​

To enable support of these additional features with Portworx Enterprise, add the following annotations to StorageClass when you Create a CSI storage class for encrypted PVCs and when you Setup StorageClass to authenticate requests

csi.storage.k8s.io/node-stage-secret-name: ${pvc.name}
csi.storage.k8s.io/node-stage-secret-namespace: ${pvc.namespace}

Limitations when the FA/FB driver is enabled​

The following features that are available with Portworx Enterprise 3.6.0 are not supported when the FA/FB driver is enabled for FADA and FBDA volumes:

  • Custom disk tag support for FlashArray Direct Access: The support for FlashArray Direct Access Volume tagging feature is not available when the FA/FB driver is enabled. This feature enables custom metadata on FADA volumes that allows easier filtering and sorting of volumes, and adding more identifying information to them.
  • Storage of the FlashArray and FlashBlade secret in Vault: When the FA/FB driver is enabled, storing Everpure FlashArray and FlashBlade credentials in Vault is not supported. Credentials will have to be directly stored in the configuration files.
  • Support for the multiple secret providers: Configuring multiple secrets providers like using Kubernetes Secrets and Vault within a single deployment is not supported when the FA/FB driver is enabled.

Reference​

Last updated on
In this topic: