Expand every Portworx storage pool in your cluster in GKE using Autopilot

Summary and Key concepts

Summary:

This article provides guidance on using Portworx Autopilot to automatically expand storage pools in a Kubernetes cluster until they reach a specified capacity. The article outlines supported cloud providers for storage pool expansion (e.g., AWS, Azure, Google Cloud Platform, VMware vSphere, FlashArray) and provides an example of an Autopilot rule. The rule monitors the size of Portworx storage pools and triggers an action to expand them by adding new drives when their size is below 400GiB. The example Autopilot rule uses the add-drive scaling method to double the pool's size while ensuring data remains in quorum during expansion.

Kubernetes Concepts:

• Custom Resource Definitions (CRD): The AutopilotRule CRD is used to define rules that automate actions like expanding storage pools.
• PersistentVolumeClaim (PVC): Autopilot automates actions on storage pools to ensure PVCs remain properly sized as demand grows.

Portworx Concepts:

• Autopilot: Automates storage management tasks, such as expanding storage pools based on predefined conditions.
• Storage Pool Expansion: Automatically scales storage pools by adding drives when pool capacity is low, ensuring no volume goes out of quorum during the expansion process.
• Storage Pool: A Portworx construct representing a collection of storage resources, which Autopilot can scale or rebalance.

You can use Autopilot to expand Portworx every storage pool in your cluster until they reach a certain capacity.

Autopilot uses Portworx APIs to expand storage pools, and these APIs also support Google Cloud Platform cloud provider.

Prerequisites

• Portworx cloud drives: Your Portworx installation must use one of the supported cloud drives where Portworx provisions the backing drives using the cloud provider
• Autopilot version: 1.3.13 or newer

Example

The following example Autopilot rule uses the auto scale type to resize all Portworx storage pools in the cluster until all of them exceed 400GiB:

note

For more information about other scale types and examples, refer to openstorage.io.action.storagepool/expand.

apiVersion: autopilot.libopenstorage.org/v1alpha1
kind: AutopilotRule
metadata:
  name: pool-expand-till-400
spec:
  enforcement: required
  ##### conditions are the symptoms to evaluate.
  conditions:
    expressions:
    # pool size is less than 400 GiB
    - key: "px_pool_stats_total_bytes/(1024*1024*1024)"
      operator: Lt
      values:
        - "400"
  ##### action to perform when condition is true
  actions:
    - name: "openstorage.io.action.storagepool/expand"
      params:
        # resize pool by scalepercentage of current size. The 100% shown below will double the current pool size.
        scalepercentage: "100"
        # when scaling, add-drive/resize-drive to the pool based on available resources
        scaletype: "auto"

Consider the key sections in this spec: conditions and actions.

The conditions section establishes threshold criteria dictating when the rule must perform its action. In this example, that criteria contains a single formula:

• The px_pool_stats_total_bytes/(1024*1024*1024) key calculates the total pool capacity in GiB
• The Lt operator sets the threshold criteria at the 400GiB value

The actions section specifies what action Portworx performs when the conditions are met. Action parameters modify action behavior, and different actions contain different action parameters. In this example, the actions section directs Portworx to do the following:

• Double the size of the pool by adding 100 percent of the scaleprecentage to the pool.
• Scale the pool automatically by either adding new disks or resizing the existing disks. Autopilot chooses the best method to expand the pool based on the current configuration and available resources without any manual intervention.

note

Autopilot expands the pools such that no volume in the system will go out of quorum. For example, if volume1 has replicas on pool1 and pool2, Autopilot first expands pool1, waits for completion, and then expands pool2.