Starting with version 3.1, Zerto Virtual Replication supports Storage vMotion operations on recovery volumes. An administrator may wish to convert recovery VMDKs from being thick provisioned lazy zeroed or thick provisioned eager zeroed to thin provisioned, or the reverse, and will want to ensure that this scenario won’t affect ZVR processes.
Zerto Virtual Replication fully supports Storage vMotion, including the change of disk provisioning. Changing the disk provisioning during Storage vMotion of recovery volumes has no effect on ZVR processes.
For example, the following is a supported scenario:
- DRaaS customer's VM is using thick provisioned VMDKs.
- Customer sends VMDKs to CSP for seeding.
- CSP copies the VMDKs to a datastore.
- CSP attaches a VM to the VMDKs, then Storage vMotion's the VMDKs to another datastore, and in the process, converts the VMDK disk provisioning to thin provisioned.
- CSP proceeds to configure the VPG using the newly thin provisioned preseeded disks.
Note that ZVR does not actively monitor for disk thickness changes, and as such, if the Storage vMotion occurs after a VPG has already been created, there could be some unexpected outcomes, such as the following:
- In ZVR 3.1, if a recovery volume is changed from thick provisioned to thin provisioned, then ZVR’s protection against VRA freeze due to a full datastore will not function properly, as the thin-thick change isn’t monitored. This can occur in ZVR 3.1. This behavior is fixed in ZVR 3.5.
- Any “reprotect” of such a VM would restore the thickness to its original state - when the VPG was created.
- Reprotect can be caused in ZVR 3.1 when the recovery datastore is changed, and in ZVR 3.5 when both the recovery host and recovery datastores are changed.