New replications perform a complete initial synchronization, copying the entire source data from the vApp or virtual machine (VM) to a datastore in the destination site. Using a replication seed lowers the network data traffic and the required time for the initial synchronization while briefly consuming double space.
Due to the size of the vApp or VM or to the network bandwidth, an initial full synchronization might take a long time. To reduce the initial synchronization time, you transfer the source vApp or VM to the destination site. Use removable media, failover of a previous replication, or other means of data transfer. Then, in the destination site, configure a replication using the vApp or the VM copy as a replication seed.
Destination Datastore Space Consumption
To be able to create the independent disk for the replication, when starting a replication with or without seed requires at least as much space as the source VM capacity in a single compliant destination datastore.
To start a replication using a seed VM requires twice the same storage space. The double space requirement lasts for a short period of time between the independent disk creation and the removal of the seed VM.
Using a seed VM lowers the network traffic, not the datastore usage, and requires as much free space, as for replicating from scratch, even though the space is only briefly reserved and might not even get fully utilized.
After VMware Cloud Director Availability collects the storage consumption and updates the independent disk, the disk usage with the respective quota reservation might shrink. Shrinking is due to reporting the actual usage, instead of the total disk capacity.
Use a VM as a Replication Seed
To use a VM as a seed, in the destination site, select a VM that has an identical disk configuration with the seed VM. The size and number of disks, and their assignment to disk controllers and bus nodes must match the replication source and the seed VM.
For example, if a replication source VM has two 4 GB disks, one of them assigned to SCSI controller 0 at bus number 0, the second one to SCSI controller 1 at bus number 2. Your seed VM must have the same hardware configuration - two 4 GB disks, at SCSI 0:0 and at SCSI 1:2.
The disks in the source virtual machine must match the disks in the seed VM. Else the reverse replication fails with a Disks of provided seed VM don't match the disks of the source VM message. For more information, see Selecting disks for replication.
Use a vApp as the Replication Seed
To use a vApp as a seed, in the destination site, select a vApp that has an identical VM set with the seed vApp. The VMs in the seed vApp must have a matching name to the VMs in the source site vApp. Each VM in the seed vApp, must meet the prerequisites to be a seed VM of the VM with the same name in the source site.
After you start a replication, in the VMware Cloud Director™ inventory, the seed vApp is empty and you can manually copy the vApp settings and metadata that are not replicated from the source site. The seed vApp remains available as an empty copy and you can remove it at your discretion.
Create a Replication Seed
Use one of the following methods for creating a seed vApp or VM in the destination site.
- Offline data transfer: Export the VM as an OVF package and a Cloud service administrator imports the package to your cloud organization.
- Clone a VM: Create a seed vApp or VM by cloning the vApp or VM from the destination site. VMware Cloud Director Availability calculates the checksum and exchanges the different blocks from the replication source to the seed vApp or VM.
- Failover data from a previous replication: Set up a replication, fail over to the destination site and continue using the on-premises workload. At a later point, you protect it in the destination site by using the VM that you failed over earlier as a seed.
- Copy over the network: Copy a source VM to the cloud organization and transfer the source data to the destination site by using other means than VMware Cloud Director Availability.
