Skip to main content
SANtricity 11.9

How reserved capacity works

Contributors netapp-jolieg

Reserved capacity is automatically created when copy service operations, such as snapshots or asynchronous mirroring operations, are provided for your volumes.

The purpose of reserved capacity is to store data changes on these volumes, should something go wrong. Like volumes, reserved capacity is created from pools or volume groups.

Copy service objects that use reserved capacity

Reserved capacity is the underlying storage mechanism used by these copy service objects:

  • Snapshot groups

  • Read/Write snapshot volumes

  • Consistency group member volumes

  • Mirrored pair volumes

When creating or expanding these copy service objects, you must create new reserved capacity from either a pool or volume group. Reserved capacity is typically 40 percent of the base volume for snapshot operations and 20 percent of the base volume for asynchronous mirroring operations. Reserved capacity, however, varies depending on the number of changes to the original data.

Thin volumes and reserved capacity

For a thin volume, if the maximum reported capacity of 256 TiB has been reached, you cannot increase its capacity. Make sure the thin volume's reserved capacity is set to a size larger than the maximum reported capacity. (A thin volume is always thinly-provisioned, which means that the capacity is allocated as the data is being written to the volume.)

If you create reserved capacity using a thin volume in a pool, review the following actions and results on reserved capacity:

  • If a thin volume's reserved capacity fails, the thin volume itself will not automatically transition to the Failed state. However, because all I/O operations on a thin volume require access to the reserved capacity volume, I/O operations will always result in a Check Condition being returned to the requesting host. If the underlying problem with the reserved capacity volume can be resolved, the reserved capacity volume is returned to an Optimal state and the thin volume will become functional again.

  • If you use an existing thin volume to complete an asynchronous mirrored pair, that thin volume is re-initialized with a new reserved capacity volume. Only provisioned blocks on the primary side are transferred during the initial synchronization process.

Capacity alerts

The copy service object has a configurable capacity warning and alert threshold, as well as a configurable response when reserved capacity is full.

When the reserved capacity of a copy service object volume is nearing the fill point, an alert is issued to the user. By default, this alert is issued when the reserved capacity volume is 75 percent full; however, you can adjust this alert point up or down as needed. If you receive this alert, you can increase the capacity of the reserved capacity volume at that time. Each copy service object can be configured independently in this regard.

Orphaned reserved capacity volumes

An orphaned reserved capacity volume is a volume that is no longer storing data for copy service operations because its associated copy service object was deleted. When the copy service object was deleted, its reserved capacity volume should have been deleted as well. However, the reserved capacity volume failed to delete.

Because orphaned reserved capacity volumes are not accessed by any host, they are candidates for reclamation. Manually delete the orphaned reserved capacity volume so you can use its capacity for other operations.

System Manager alerts you of orphaned reserved capacity volumes with a "Reclaim unused capacity" message in the Notifications area on the Home page. You can click Reclaim unused capacity to display the Reclaim Unused Capacity dialog box, where you can delete the orphaned reserved capacity volume.

Characteristics of reserved capacity

  • Capacity allocated to reserved capacity needs to be considered during the volume creation to retain sufficient free capacity.

  • Reserved capacity can be smaller than the base volume (the minimum size is 8 MiB).

  • Some space is consumed by metadata, but it is very little (192 KiB), so it does not need to be taken into account when determining the size of reserved capacity volume.

  • Reserved capacity is not directly readable or writeable from a host.

  • Reserved capacity exists for each read/write snapshot volume, snapshot group, consistency group member volume, and mirrored pair volume.