Disaster Recovery Testing. The Plan Is Only Real Once You Have Tried It
Most businesses have a backup. Far fewer know how long it would take to be working again. Testing is how you find out.
A disaster recovery plan describes how you get the business working again after something serious: a fire, a flood, ransomware, a failed server or a lost internet connection. Writing the plan is the easy part. The uncomfortable truth is that an untested plan usually has gaps, such as a missing password, a backup that will not restore or a step nobody remembered.
Two numbers shape any plan. The recovery time objective, or RTO, is how long the business can be without a system before the impact becomes unacceptable. The recovery point objective, or RPO, is how much data you can afford to lose, measured in time. If you back up nightly, your RPO is up to a day. If the accounts system must be back within four hours and the backup takes eight to restore, the plan does not meet the objective, however neat the document is.
Begin by listing your systems in order of importance, and agree RTO and RPO for each with the people who rely on them. Email, line of business software, files, phones and internet access are the usual candidates. Then write the steps for bringing each back, including who does what, where credentials and licence keys are kept, and who to call at your suppliers.
There are three levels of testing. A tabletop exercise gathers the key people for an hour and talks through a scenario, such as the office being unavailable on a Monday morning. It is cheap and reveals confusion about roles. A restore test takes real backups and restores a file, a mailbox or a whole server to a separate location to prove it works and to measure how long it takes. A full failover exercise switches live operations to the recovery environment, which is more demanding and is usually reserved for critical systems.
Test at least once a year, and again after major changes such as a new server, a move to the cloud or a change of backup product. After each test, write down what went wrong and fix it, then update the plan. Keep a printed copy of the essentials somewhere accessible, because the plan is not much use if it lives only on the server that has failed.
Dragonfly IT can run a recovery test for you, show what really happens and help you set realistic recovery targets. Contact us to arrange one.
