Storage management is crucial for vCloud Availability for Director because storage is consumed by replicated base disks, change deltas, and point-in-time snapshots.

In vCloud Director, the tenant consumes cloud resources through Organization Virtual Data Centers (Org VDCs) that represent compute, networking, and storage resources. Storage is presented in terms of storage policies (with different performance or availability SLAs).

vCloud Availability for vCloud Director is consuming replication enabled Org VDCs, and Org VDC storage policies are presented as targets for replicated VMs. The files that represent replicated disks and multiple point-in-time (MPIT) storage snapshots of protected virtual machines, however, are not real vSphere objects, and therefore, vCloud Director has no visibility as to how much space they consume until the VM is failed over. Therefore, Org VDC storage quota management has no effect on the cold data.

 

When a VM is configured for replication, vCloud Director creates a placeholder VM initially with full disk size, which is placed by the vCloud Director placement engine to a particular datastore and checked against Org VDC storage quota. Virtual machine disks are then replaced by 4 MB dummy disks so no storage consumption is reported to Org VDC. The virtual machine folder location is used for replication data. vSphere sees the replication files as independent data (hbrdisk.RDID) that is not coupled with the VM. VMware vSphere Storage vMotion® and vCloud API move replica disk call must be used to move the replication data (see Section 5.5.2, Moving Replica Disks).