Powered by Zoomin Software. For more details please contactZoomin

Zerto Virtual Manager Administration Guide vSphere Environments

Platform
Product Type
Version
Content Type
Source
This publication
Table of Contents

Live Disaster Recovery Testing

This section describes how to use the basic Zerto recovery operations to perform live disaster recovery tests, in different situations.

When performing a live DR test you need to consider the following:

What is the purpose of the live DR test:
Do you only want to verify that the VMs can recover properly?
Do you want to conduct a full DR test that will include running user traffic against the recovered VMs?
The length of time you want to test the recovery, a few hours or several days.
Whether the changes to the recovered machine need to be retained after the test or can they be discarded?
Whether you are willing to accept temporary downtime of the application.
Whether you want to simulate an actual disaster at the protected site, for example by simulating a network outage or bringing down the protected site.

The following flowchart shows the testing decision flow:

During any live test, Zerto recommends that you only maintain one working version of the same virtual machine.

As such, the first step in any test, except for a Failover Test or Clone, is to make sure that the protected virtual machines are shut down before starting to test recovered machines.

During a Zerto Move operation the first step Zerto performs is to shut down the protected machines, to ensure data integrity.

However, a Zerto Failover operation assumes that the protected virtual machines are no longer accessible (the total site disaster scenario) and does not attempt to shut them down at the beginning of the operation.

In a live test using a Failover operation you have to manually shut down the virtual machines to be tested at the beginning of the test in order to prevent potential split-brain situations where two instances of the same applications are live at the same time.

If you want to perform a live DR test that includes a simulated disaster you can simulate the disaster, for example, by disconnecting the network between the two sites. In this type of test, once the disaster is simulated a Move operation cannot be used, since it requires both sites to be healthy, while a Failover operation can be used.

TitleResults for “How to create a CRG?”Also Available inAlert
Feedback Feedback