Performing a failback
The following procedure describes how to perform a failback on a replicated virtual machine. You must always perform a failback from the destination DS-Client.
IMPORTANT: Before performing a failback, ensure that the existing source virtual machine has already been deleted and that all notifications are resolved. If you want to retain the existing source virtual machine, clone the virtual machine before deleting it. Do not delete a partially restored source virtual machine from a previously attempted failback that has been stopped or has failed.
To perform a failback:
1. On the VM Replication menu, click All Replicated Machines.
2. In the All Replicated Machines dialog box, under Items ready for failback, each row displays a target virtual machine that is being used after failover. Do one of the following:
• To perform a failback on one virtual machine, in the row that corresponds to the virtual machine, click Failback.
• To perform a failback on multiple virtual machines concurrently, select the check box in the rows that correspond to the virtual machines, and then click Trigger. You can perform a maximum of three concurrent failbacks.
NOTE: If a replication set consists of folders that contain multiple virtual machines with files located on different datastores, you should perform a failback of the multiple virtual machines concurrently to the same datastore or perform a failback of each virtual machine to a different datastore. When performing a failback to different datastores, ensure that the first failback completes before performing another failback.
3. In the Select the Source Host and Datastore for Failback dialog box, do the following for each virtual machine on which you are performing a failback:
a) Verify the original host and original datastore.
b) In the Failback to host list, select the target host for failback.
c) In the Failback to datastore list, select the target datastore for failback.
d) To start the failback process, click Failback.
During the failback process all changes to the active virtual machine are preserved and transmitted to the target host and datastore on the destination server, and the restored generation is powered on.
When the failback process completes, you can perform replication on the source DS-Client to ensure that all changed data is replicated to the destination DS-Client. Subsequent replication activities after failback will transmit only incremental changes to the destination server. The virtual machine at the destination server must not be deleted to allow for incremental replication, otherwise full replication of the virtual machine will occur.
IMPORTANT: If the failback process is stopped or fails, restart the process. Do not delete a source virtual machine that has only been partially restored in a previous failback attempt.