---
{
  "title": "Disaster recovery",
  "summary": "Restore technical services in a tested sequence with trustworthy data.",
  "topic": "grc",
  "level": "Beginner",
  "updated": "2026-10-12",
  "units": [
    3
  ],
  "aliases": [],
  "sources": [
    {
      "title": "NIST: Contingency Planning Guide",
      "url": "https://csrc.nist.gov/pubs/sp/800/34/r1/final"
    }
  ],
  "published": "2026-10-02",
  "editorialVersion": 2
}
---

## The idea

Disaster recovery restores systems and data after disruptive events. Plans identify priorities, dependencies, recovery sources, required access, procedures, and validation. Recovery objectives guide design, but tests demonstrate whether they are achievable. Include loss of a region, identity service, key, or backup access—not only one server. Recovery from a security incident also needs assurance that compromise is not reintroduced.

## In practice

A team restores an application and database in dependency order, checks data integrity and user functions, and records the actual recovery time.

## A common misunderstanding

A successful backup job is not proof that the full service can be recovered within its business requirements.

## Check your understanding

What should a recovery exercise validate?

**Think it through:** Access to recovery material, restoration steps, data quality, dependencies, functionality, and achieved timing.


## Restore a dependency chain, not an isolated server

Disaster recovery brings technical services back after disruption. A plan should account for identities, networking, configuration, data, keys, infrastructure, and application dependencies. Restoring a database onto a running machine is insufficient if the application cannot authenticate or the recovery team cannot obtain the required decryption key.

Sequence matters. Some services must be available before others can be validated. Record prerequisites, responsible operators, verification steps, and decision points. Protect recovery credentials and documentation so they remain accessible to authorized responders even when the primary environment is unavailable. A plan stored only inside the failed system may be useless at the critical moment.

## A fictional clean recovery

After a destructive incident, an academy prepares an isolated recovery environment. It selects a trustworthy recovery point, restores required infrastructure, validates application data, and checks that the conditions enabling the incident have been addressed before reconnecting. Returning an unexamined compromised image to service could recreate the problem.

Recovery testing should distinguish technical startup from business acceptance. Operators can confirm that processes run, while service owners verify that important workflows and records are usable. Record actual timings and exceptions, then compare them with recovery objectives. A rehearsal is valuable when its findings change the plan or capability.

## Walk through a dependency

Choose a fictional student-record service and list its database, identity, network, storage, and key dependencies. Put them in a plausible restoration order and identify how each will be verified. Add one failure during the rehearsal, such as an unavailable administrator account, and explain the authorized fallback. This reveals whether the plan survives realistic disruption rather than only a perfect demonstration.

## Keep exploring

- [[backups]]
- [[recovery-objectives]]
- [[eradication-and-recovery]]