User guide

3. VSR presents the following pop-up message:
Important: This action will place live machines onto your network with the same
networking configuration as your original systems. Please confirm: is this the action
you wish to perform?
The options are OK to confirm the recovery and Cancel to abort the request.
4. After the customer clicks OK on the pop-up message, the customer's recovery
begins.
a. The VSR system suspends the customer's replication.
b. The VSR system analyzes the customer's stored replica to determine its:
v Operating system
v Patch level
v Amount of CPU, RAM and Disk the application was using when its last
snapshot was taken
c. The VSR system then launches a VM with those characteristics, and attaches
the VM to the customer's replica.
Note: For both Shared Resource and Reserved Resource Virtual Machine VMs,
this recovery process typically takes less than 15 minutes. That time starts when
the customer clicks Recover. It ends when the customer's VM starts to boot.
Failback
For the Shared Resource Virtual Machine service level and the Reserved Resource
Virtual Machine service level offerings, failback involves the activities documented
in this section. Failback is the process by which a system returns to normalcy after
failover is completed. Doing so involves use of the failback ISO image,
downloadable from the portal download page.
Before you begin
Remember: Failback is the mirror image of failover. In failover, the reference copy of
your data resides on your production server. In failback, the reference copy of your
data resides on IBM's recovery servers, which have been processing your
transactions during your outage emergency. To resume production at your facility
(or failback), VSR must first copy those transactions from IBM's recovery server to
your production server.
VSR uses a process known as a differential compare to identify exactly which
transactions are missing from your production server. Then, it reverses the normal
flow of replication, and streams all the missing changes to your server at your
production facility. Once those changes have been written, the disks are then
identical, and you are ready to resume normal production.
The biggest difference between failover and failback is control of the timing. In the
typical outage emergency (just over half of declares in the U.S. are caused by
hurricanes), a disaster occurs without warning, on a random timetable. But failback
occurs at a time that you schedule. Most businesses choose to failback when
transaction traffic is lightest, usually before dawn, before traditional business
hours, during the very early hours of the day.
This is particularly important with a recovery service that combines replication
(mirroring), and a Layer 2 network extension, like VSR. When Mirroring, a source
server and its replica have identical disk images, and thus identical IP addresses.
6 IBM SmartCloud Virtualized Server Recovery: Customer User Guide