Veritas Storage Foundation Cluster File System High Availability supports reclamation of the unused storage on thin-reclamation capable arrays. Storage Foundation Cluster File System High Availability automatically discovers LUNs that support thin reclamation.

A Veritas File System (VxFS) file system can be mounted on a Veritas Volume Manager (VxVM) volume that is backed by a thin-capable array. The size of the VxVM volume is a virtual size, that is backed by the free storage pool. When files are created or changed, storage is physically allocated to the file system from the array. If the files on the file system are deleted or shrunk in size, the space is freed from the file system usage. However, the space is not removed from the physical allocation. Over time, the physical space allocated to the file system is greater than the actual space used by the file system. The thin LUN eventually becomes 'thick', as the physical space allocated nears the size of the LUN.

The Thin Reclamation feature provides the ability to release this unused space back to the thin pool. Storage Foundation Cluster File System High Availability uses the VxFS allocation tables to identify unused blocks. VxVM maps this information about unused blocks down to the disk, enabling VxVM to return those blocks to the free pool. If the VxFS file system is not mounted, VxVM has no visibility into the file system usage. Therefore, it is critical that the file system is mounted when you perform a reclamation. The operation of reclamation can be done on a disk group, LUN, enclosure, or file system.

VxVM reclaims space automatically when you delete a volume or remove a plex. The automatic reclamation is asynchronous, so that the space is not reclaimed at the array level immediately. The disk is marked as pending reclamation. You cannot remove a disk from VxVM until the reclamation completes. You can control the timing and frequency of the automatic reclamation.