Best practices for using and configuring vSphere Replication can prevent your environment from possible issues during replication.

Important: You mustn't change the source VM hardware while you are in the process of configuring a replication. For example, don't add or remove hard disk to the source VM before you finish configuring the replication.

Setting the Optimal Recovering Point Objective (RPO) Time

The replication of several thousand virtual machines is a bandwidth consuming process. One of the many factors that influence bandwidth requirements for vSphere Replication is the RPO configuration for each replicated virtual machine.

You can set the RPO to 5 minutes, but you must estimate the optimal RPO time to save bandwidth for the replication, and to cover your business requirements for the protection of your VMs.

For instance, if a block changes only once per day, it is replicated only once regardless of the RPO configuration. However, if a block changes many times during the day, and the RPO is set to a low number such as 30 minutes, the block might be replicated as many as 48 times in one day.

vSphere Replication tracks larger blocks on disks over 2 TB. Replication performance on a disk over 2 TB might be different than replication performance on a disk under 2 TB for the same workload depending on how much of the disk goes over the network for a particular set of changed blocks.

A network with the appropriate bandwidth available to transfer the system’s data ingest rate is required to support the desired replication interval.

For example, if you have a dataset of 1 TB with a daily change rate of 2 GB per hour and RPO set to one hour, this means vSphere Replication must transfer 2 GB in 1 hour or 4.7 Mbps. This is the minimum theoretical bandwidth required to complete the vSphere Replication transfer.

The data change rate is not uniform throughout the day even though the above example assumes it. Use the peak data change rate in your scenario to calculate the minimum bandwidth requirement

See Calculate Bandwidth For vSphere Replication for details.

Using Multiple Point in Time (MPIT) Recovery

Each point in time snapshot consumes storage. The amount consumed depends on the data change rate in the VM. When you set multiple point in time instances for replication of a VM between two vCenter Server sites, vSphere Replication presents the retained instances as standard snapshots after recovery. The time required to consolidate snapshots after recovery, increases with the number of snapshots.

For example, if you configure a replication with 10 recovery points, then the storage consumption on the target site might increase to 11 times the original size of the source disk.

Although vSphere Replication supports up to 24 recovery points, you must set the MPIT to the lowest number of recovery points that meets your business requirements. For example, if the business requirement is for 10 recovery points, you must set up vSphere Replication to save only 10 snapshots. You can set up two recovery points per day for the last five days. As a result, the consumed storage and the time needed to consolidate the snapshots after recovery are less than if you use the maximum number of recovery points.

Configuring Quiescing

For VMs with high levels of storage I/O, quiescing of the file system and applications can take several minutes and impact the performance of the VM . When quiescing a file system and applications for Windows VMs, vSphere Replication requires a regular VM snapshot before replication. When you estimate the RPO time, consider the time and resource consumption for the quiescing and for the consolidation of the snapshots. For example, if you configure a replication of a Windows VM with an RPO of 15 minutes and quiescing is enabled, vSphere Replication generates a VM snapshot and consolidates it every 15 minutes.

Quiescing options are available only for virtual machines that support quiescing. For more information on which operating systems are supported, see the Guest OS Quiescing Support section in Compatibility Matrices for vSphere Replication 8.7.x.
Note: Quiescing for vSphere Replication and backup operations for the same virtual machine is not supported.

Configuring Replication Seeds

You can copy virtual disk files of source VMs to the target location and use these files as replication seeds. By using replication seeds, vSphere Replication reduces the amount of time and network bandwidth required for the initial full sync process. The UUID of the source and target VMDK files must match for the replication to be successful and to prevent unintentional overwrites of disk files that belong to other VMs at the target location.

Monitoring a Datastore on the Target Site

vSphere Replication requires enough disk space at the target site to replicate a VM. If the available space is not enough to save the replication files, the replication might fail. You can create an alarm that alerts you about insufficient storage capacity at the target site.