Disk Downsizing¶

Disk downsizing lets you reduce provisioned storage when your deployment instance needs less disk capacity. Omnistrate moves your data to a smaller disk and verifies the copy before restarting the affected workload.

Requirements and Supported Storage¶

• Use supported, statically provisioned Omnistrate-managed persistent block data disks on AWS EBS, GCP Persistent Disk, or Azure Disk.
• The target size must be smaller than the current capacity and comply with your Plan's parameter limits and the cloud storage type's requirements.
• Used data on each affected disk must fit on its replacement. Leave room for filesystem overhead and future data growth.

Dynamically provisioned Kubernetes volumes and scheduled-job Resources are not supported.

Downsize a Disk¶

To downsize a disk, update the storage size of your deployment instance to a smaller value within your Plan's limits.

Your Plan must expose a modifiable storage size parameter for customers to change capacity. For Compose configuration, see x-omnistrate-storage and Deployment API Params.

Before starting:

1. Create and verify a restorable snapshot or backup.
2. Check used capacity on every affected disk and choose a target with sufficient free space.
3. Schedule a maintenance window and notify customers of the expected interruption.

To apply the smaller size:

1. Open the deployment instance in the Operations Console and select Modify.
2. Set the storage size parameter to the smaller value and confirm the update.
3. Track the operation through Workflows.
4. After the workload is healthy, validate application-level data and confirm the new disk capacity.

For example, you can reduce a 100 GiB disk containing 30 GiB of data to 50 GiB if your Plan and storage type allow it. Choose a size that leaves enough space for filesystem overhead and expected data growth.

If the storage size is fixed in your Plan, publish a new Plan version with the smaller capacity and apply it through an instance upgrade or upgrade path. See Managing Upgrades. The Modify action above changes parameter values on the current Plan version.

What Happens During Downsizing¶

Cloud block storage generally supports expansion but does not support reducing an existing disk's capacity. Omnistrate processes the affected nodes one at a time, replacing each node's disks in these steps:

1. Check that the requested size and storage type are supported.
2. Stop the affected workload and check that its data fits within the requested capacity.
3. Create a smaller disk, copy the data, and verify the copy.
4. Restart the workload on the smaller disk and confirm that it is healthy.
5. Delete the old disk after health confirmation.

Subsequent backups and other storage operations use the replacement disk automatically.

Failure Handling¶

If any affected disk on the current node contains more data than the requested capacity, Omnistrate stops the operation before creating replacement disks for that node. If the capacity check, copy, verification, or restart fails, Omnistrate retains that node's old disks and attempts to restore its workload on them.

Each node's old disks are retained until its workload is healthy on the replacements.

If downsizing fails, review the workflow details and check every affected node's health and disk capacity before retrying. Contact support@omnistrate.com if a workload does not recover. Keep a restorable backup and validate your application data after the operation.