Testing and Automation: The Keys to a Successful Recovery

Automation is the key to any disaster recovery strategy.

Not only does automation make recoveries faster, it also helps to reduce the potential for manual errors and minimizes risk if your recovery staff is unavailable during an emergency situation. Therefore, it’s crucial to periodically test any DR scenario in order to confirm that the process works as expected, as well as to see if it has kept pace with any changes you’ve introduced into your production environment.

Four Ways to Test Your DR Process

Testing any DR scenario is a must, particularly when there are so many components in your data center, including your ClearPath® MCP system. Not only do you want to make sure that the DR process successfully recovers your MCP environment, you’ll also need to confirm that all production workloads function as they did before.

There are a few strategies you can employ to work towards this goal, each of which has its benefits and drawbacks.

Let’s take a look at each in a little more detail.

The first is the “peek” test. You’ll start by halting storage replication, moving a point-in-time copy of your production data to a standby system, and running a test workload against that data. If it succeeds, you can re-enable the replication. Though this process won’t disrupt your production workloads, it’s not the most exhaustive strategy. You’re only testing one part of the DR process – the data – not things like network availability or whether or not a production workload will correctly execute. Another thing to be aware of is that you may not be able to fail over automatically while the test is in process.

The second is a variation of the peek test. In this version, you’ll create a point-in-time copy of your production data and halt/load the secondary system as a variation of the primary system. Once you do this, you can run a test workload. While this method is more comprehensive, you may still not be able to automatically fail over until it completes.

The third strategy is to halt production on a weekend, execute your failover plan, run some tests, and return production back to the primary set. While you’ll benefit from testing the full complement of your failover processes, doing so may have an impact on your production environment. Oftentimes, the secondary system could be an older model or might have less capacity than the production system. These potential differences create an incentive to minimize any impact on production systems, but doing so means the tests you execute may not be as exhaustive as possible.

The last strategy is to eliminate the idea of primary and secondary sites altogether. Instead, you’ll periodically move production somewhere equally configured and just as capable of running a production workload. This isn’t so much a failover test as it is a shift in the location of a workload. And while there’s an impact to production, there is no incentive to minimize the test.

No Matter What, Automate

Regardless of which test option you choose, and when you decide to run it, automation is crucial. Each of these tests can be driven by a scheduler, such as SMA Op/Con XPS, or by specialized DR software like Unisys Business Continuity Accelerator.

Imagine a daily or weekly DR test that utilizes the two “peek” strategies described earlier. With the help of these tools, you can run an automation script that sends point-in-time copies to the DR host, restarts the host, executes a few test workloads, puts everything back, and automatically notifies you when the test successfully completes.

This periodic automatic execution can also be paired with a more rigorous failover test, such as the final two strategies discussed above. Regardless of which strategy choose, it’s always a good idea to test your process and automate it as much as possible – and make sure to use Unisys knowledge and expertise to confirm you’ve covered all of your bases.

Please visit our web site for more information about the capabilities and benefits Unisys Business Continuity Accelerator has to offer. And to see how the solution helped one Unisys client strengthen its business continuity measures, please read the Pershing Limited case study.