top of page

Your Backups Won’t Save You If You Can’t Restore

Nov 14
5 min read

Keeping Savannah Moving · Part 5 of 8


The backup dashboard is green. The warehouse application opens. Then a supervisor asks which orders shipped before the outage, and nobody can reconcile the restored records with the loading dock. That hypothetical situation captures a common planning mistake: defining recovery as a working computer when the business needs a trustworthy way to work.


Conceptual illustration of a dispatcher and an IT colleague checking restored shipment records against a clipboard, with a backup device, truck and port cranes nearby.
Original AI-generated editorial illustration of a hypothetical dispatch recovery rehearsal.

For a Savannah carrier, warehouse or service contractor, the useful question is specific: can the people responsible for tomorrow’s commitments complete them after the systems they depend on fail? A backup contributes to that answer. A rehearsed recovery process demonstrates it.


In Part 4, The Warning Signs Before Ransomware Strikes, we tested whether an alert reaches someone who can act. Recovery asks the next question: if systems become unavailable, can the business resume work with records it can trust?


What the maritime case tells us


In Part 1 of this series, we examined a Coast Guard account of Ryuk ransomware at an unnamed maritime facility. Primary operations stopped for more than 30 hours. The public bulletin does not describe the full restoration process or establish that a particular backup failure caused that duration. We can use the incident to ask better recovery questions without inventing its missing details.


NIST’s Guide for Cybersecurity Event Recovery emphasizes identifying and prioritizing resources, building recovery playbooks, testing them, and learning from results. For a small logistics company, that starts with a business service and the people who use it.


Choose the work that must return first


Ask an operations manager to identify one service that cannot remain unavailable for long: dispatching drivers, releasing inventory, confirming appointments or contacting customers. Write a plain description of what successful recovery would let an employee do. “The dispatch server is running” is incomplete; “the dispatcher can retrieve approved jobs, assign a driver and record completion” is testable.


Then list the dependencies that make that task possible. They might include the database, application configuration, identity service, network connection, software license, printer and a customer’s portal. Name the person who can obtain each one. A database copy cannot recover a password that nobody can retrieve or replace a software installer that was never retained.


Agree on two targets with the business owner. The recovery time objective (RTO) is how long a service can remain unavailable. The recovery point objective (RPO) describes how much recent data loss the business can tolerate, expressed in time. NIST’s contingency planning guide explains these planning measures. Treat them as targets until a test demonstrates what is achievable.


Consider a hypothetical dispatch exercise: the newest usable recovery point is 6:00 a.m., an outage begins at 8:00 a.m., and staff verify usable dispatch at 11:00 a.m. Two hours of updates need reconciliation, and the observed outage lasts three hours. If the business can tolerate only 30 minutes of lost updates, that recovery point misses the target even though the application opens.


Hypothetical timeline: newest usable recovery point at 6 a.m., outage at 8 a.m., usable dispatch verified at 11 a.m. Two hours of updates need reconciliation and observed downtime is three hours. RPO and RTO are planning targets, not measured results.
Figure 1. A hypothetical restore exercise separates the age of recovered data from the time needed to resume usable dispatch. Compare both results with the business’s targets.

Protect the recovery path itself


CISA’s ransomware guidance recommends offline, encrypted copies of critical data and regular restoration testing. A useful design also asks which compromised accounts could delete backups, alter retention or obtain the keys needed to restore them. Protect backup administration separately and document emergency access.


Review the actual settings behind an “immutable” or “protected” label. Which copies are covered, for how long, and who can change the policy? Encryption protects confidentiality; it does not by itself stop an authorized account from deleting a backup. Ask the provider to show the recovery copy that remains available if production administration is compromised.


Include cloud services in this discussion. Confirm what the provider protects, what the customer can restore, the retention period and the steps required after deletion or corruption. A synchronized folder can copy unwanted changes as well as good ones. Record the recovery options for the specific service and subscription, including access if the normal account is unavailable.


Run a small, complete recovery exercise


Choose a representative service and a safe test environment with your IT provider. Use appropriate test data or protected copies. Keep restored systems separate from production and prevent test actions from sending real dispatch instructions or customer notifications. The following exercise is an LCCS planning example to adapt to your equipment and operating requirements.


  1. Write the acceptance task before starting: for example, an authorized dispatcher must retrieve a known test order, assign it and record a status change.

  2. Select a recovery point and record why it is suitable. For a real intrusion, have responders assess it against the compromise timeline and check restored data and system images for malware or unauthorized changes. Confirm that required application files, configuration and credentials are available.

  3. Restore the dependencies and application using the documented process. Record elapsed time, missing instructions and every unplanned request for help.

  4. Have an operations employee complete the acceptance task. Check a sample of expected records and permissions, rather than relying only on an administrator’s successful login.

  5. Reconcile activity that occurred after the recovery point. Decide how staff would identify duplicate jobs, missing updates or goods moved during an outage.

  6. Record the outcome, remaining gaps, responsible owner and retest date. A failed step is useful evidence when it leads to a specific repair.


A successful exercise demonstrates the service and conditions tested. It does not prove that every system will recover, that every backup is usable, or that an actual incident will meet the same timings. Record those limits alongside the result.


The most valuable finding may be ordinary: the only person who knows a recovery password is unavailable, an installer requires a vendor callback, or a restored application points to the wrong database. Discovering those dependencies during a planned exercise lets the business address them before customers are waiting.


Trust the recovered information


NIST’s work on recovery from destructive events stresses the integrity of recovered information. In practical terms, staff need confidence that records are accurate and the recovery state is appropriate. An available file is not automatically the correct file.


During a real intrusion, incident responders also need to assess the affected systems and access before reconnection. Restoring data while leaving an attacker’s access in place can undermine the recovery. Restoring availability also cannot reverse a prior theft of information; investigation, communication and any applicable reporting remain separate work.


Prepare for the interval before systems return


Give supervisors an approved fallback for the work that can safely continue. Specify how they verify assignments, record changes, contact customers and preserve a record for later reconciliation. Identify the work that must pause when required information is unavailable. Review procedures involving terminal access or operational equipment with the relevant customer and safety owners.


A useful exercise ends with a short recovery record: the service tested, recovery point, time to usable service, business task completed, data discrepancies, and unresolved dependencies. Keep it where authorized staff can reach it during an outage.


This week, select one essential service and book a restore exercise with the person who runs it. Ask for evidence that the business task worked. That gives the owner a concrete basis for deciding what to improve next.


Next in the series: The Cybersecurity Rules Savannah Port Contractors Need to Know.


Sources and further reading


 
 
 

Comments


bottom of page