Disaster Recovery User's Guide

Figure 2‑5. Configure Replication - Set MPIT
Assume you have set a 15 minutes (minimum) recovery point objective (RPO) and you start the replication
at 9:00 a.m., which means data changes for that particular virtual machine are replicated every 15 minutes.
Now in the event of a failover, you can pick a specific time to recover from. For example, if the on-premises
data center has an issue and you need to failover at 11 a.m. However, at that time you find that the data may
not be suitable. If you have enabled MPIT, you can recover the data copied over at say, 10:30 a.m.
If you were to set your RPO 24 hours (maximum), then you would have the ability to restore up to 24 days
previous replication points. This allows you to failback up to 24 days. With the MPIT recovery you can pick
the specific point in time that you want to have data recovered, allowing you to successfully recover in the
event of data corruption.
Recover a Virtual Machine
Recovering a virtual machine shuts down the virtual machine at your source site (if possible) and recovers it
in the cloud. During recovery, all replication activity is stopped.
Prerequisites
You have the correct user permissions to recover the virtual machine to the cloud. See “Assign a User to the
Roles for Disaster Recovery,” on page 14 for information.
If possible, meet these prerequisites:
n
Verify that you tested the recovery before recovering the virtual machine to the cloud.
n
Verify that you cannot run a planned migration from your source site.
See Migrate a Virtual Machine to Cloud in vSphere Replication 6.0 for Disaster Recovery to Cloud for
information about planned migration from your source site.
n
Contact VMware Global Support Services and declare a disaster prior to recovering your virtual
machines.
VMware vCloud Air - Disaster Recovery User's Guide
20 VMware, Inc.