Your Backup Passed but Your Recovery Failed

11/11/26

Every morning, somewhere in your environment, a backup job reports success. The dashboard is green, the summary email says “Completed,” and everyone moves on with their day.

Then a server fails, ransomware encrypts a file share, or someone deletes the wrong folder in SharePoint, and the team finally tries to restore. That is the moment many companies discover the difference between a backup that finished and a recovery that works. They are not the same thing, and the gap between them is where downtime lives.

What a Successful Backup Actually Tells You

A backup success message confirms a narrow set of facts: the job started, it read the data it was told to read, and it wrote that data to a target without throwing an error. That is useful and also far less than most people assume.

It does not confirm that the right data was selected. Nor does it confirm that the copy is readable, free of corruption, or clean of malware that was already sitting on the server when the copy was made. It does not tell you how long a restore will take, whether you still have the credentials and encryption keys to open it, or whether the application on top of that data will actually start.

Backup completion measures a process. Recovery measures an outcome. Your business only cares about the outcome.

Where Recoveries Break

Recovery failures rarely come from one dramatic flaw. They come from small gaps that never show up in a backup report:

  • A new server, database or SharePoint site went live after the backup policy was written, so it was never included.
  • The backup copy exists, but the restore runs so slowly that a four-hour outage turns into a four-day outage.
  • The database comes back, but the application will not start because a license server, service account, certificate, or configuration file was not part of the plan.
  • The encryption key or the admin password for the backup console was stored on the very server that just failed.
  • Ransomware sat quietly long enough to be copied into several generations of backups.
  • The only person who knew the restore steps left the company last year.

None of these trigger an alert. Every one of them can stop a recovery cold.

ERP Makes It Personal

For manufacturers and distributors, the system that matters most is usually the ERP. Restoring Epicor Kinetic or Epicor P21 is not a single file copy. The SQL database, application servers, task agents, integrations, reports, printers, and user access all must come back in the right order. A database backup that verifies perfectly can still leave shipping, purchasing and production waiting if the rest of the stack does not return with it.

Microsoft 365 deserves the same scrutiny. Many organizations assume Microsoft keeps everything forever. Microsoft protects the platform, but retention windows, permanently deleted items and accidental or malicious changes are still your responsibility. If you have never restored a mailbox, a OneDrive, or a Teams site from your backup, you do not yet know how that will go.

Recovery Testing Is the Real Test

Recovery testing means you actually restore, prove it works, and write down the results, on a schedule. It answers the questions a backup report never will. Can we get the data back? Can we get it back fast enough? Is it the right version? Does the application work afterward? Can the people on call today do it without the person who built it?

Two numbers anchor the conversation. Your Recovery Time Objective (RTO) is how long the business can tolerate a system being down. Your Recovery Point Objective (RPO) is how much data, measured in time, the business can afford to lose. Leadership should agree on both, rather than leaving IT to guess, and every recovery test should measure tangible results against them. If the target is four hours and the test takes eleven, you have a decision to make long before a real incident makes it for you.

A Practical Testing Rhythm

You do not need a full disaster simulation every week. A layered approach works well for most mid-sized organizations:

Each month, restore a sample of files, folders, and mailboxes and confirm that each item opens, is complete, and is the version you expected. Each quarter, restore a critical server or database, such as your ERP, into an isolated environment and make sure the application runs, users can log in, and the data is current. Twice a year, walk through a ransomware or site loss scenario as a tabletop exercise so everyone knows their role, who to call, and where the runbook lives. Once a year, run a full end-to-end recovery of a priority system and confirm that recovery time and data loss both fall within your RTO and RPO.

Document every test: what was restored, how long it took, what broke and what was fixed. That record turns recovery from a hope into evidence, and it is the kind of proof cyber insurance carriers and auditors often ask to see.

Build Backups That Are Ready to Be Restored

Testing exposes problems. A few design choices prevent many of them in the first place. Follow the 3-2-1 approach: three copies of your data, on two diverse types of storage, with one copy offsite. Keep at least one copy immutable or offline so ransomware cannot alter it. Store encryption keys, credentials, and recovery runbooks somewhere that does not depend on the systems you are trying to recover. And review backup scope every time something new goes into production, so tomorrow’s critical system is not missing from yesterday’s policy.

Change the Question

Most teams ask, “Did the backup run?” The better question is, “When did we last prove we could recover, and how long did it take?” If nobody can answer that with a date and a number, the green checkmark is giving you confidence you have not earned yet.

How 2W Tech Can Help

2W Tech helps manufacturers and distributors design backup strategies, run recovery tests and document recovery plans across Epicor, Microsoft 365 and the infrastructure that connects them. We will help you set realistic RTO and RPO targets, test against them and fix what the tests uncover before an outage does it for you.

Ready to find out whether your recovery really works? Contact us.

2W Tech. Every Solution. One Partner.

Read More:

What Is an MCP Server? The Plug That Connects AI to Your Business

ERP in Azure: Why Your Business System Belongs in the Microsoft Cloud

Back to IT News