How to Write a Disaster Recovery Plan (IT)

A disaster recovery plan is the documented procedure for getting your IT systems and data back after something takes them down — a dead server, a ransomware attack, a flooded comms room, a cloud outage. It answers four questions for every system: what gets recovered, in what order, how quickly, and by whom.

It is the IT-focused part of a wider business continuity plan. Here is how to build one that will actually work when you need it.

Step 1: Inventory what you have

You cannot recover what you have not written down.

  • Every system: line-of-business apps, email, file storage, accounting, the website, phones
  • Where each one lives — a physical server, a VM, a cloud service, a workstation under someone’s desk
  • The data each holds and where its backups go
  • Dependencies — system B needs the database on system A; everything needs the network and internet
  • Vendor and support contacts, licence keys, and admin credentials (stored securely, and reachable when systems are down)

Step 2: Set RTO and RPO for each system

Not every system deserves the same speed of recovery. For each one, agree with the business:

  • Recovery Time Objective — how long it can be down
  • Recovery Point Objective — how much data you can afford to lose

This is the core of the plan — it drives the backup frequency and the recovery method. Our full explainer: RTO vs RPO.

Step 3: Backup and replication strategy

Recovery is only as good as the copies you can restore from. Follow the 3-2-1 rule: three copies, two media types, one offsite, plus one immutable copy ransomware cannot reach, and zero untested restores. For low-RTO systems, add image-based backup or replication to a standby so you fail over rather than rebuild.

Step 4: Write the recovery procedures

For each system, or group of systems, a short procedure:

  • Preconditions — what must be working first (network, domain controller, internet)
  • Steps — where to restore from, to what target, in what order
  • Who does it, and who is the backup person
  • How to verify it worked — a specific check, not “it looks fine”
  • Rough time expected, so you know if it is going wrong

Include the recovery order: identity and network first, then core data, then the systems that depend on them.

Step 5: Roles and communication

  • Who declares a disaster and starts the plan
  • Who runs the technical recovery
  • Who keeps management and staff informed
  • Contact details for the IT provider, key vendors and ISPs, stored offline

Step 6: Test it

  • Restore tests — monthly or quarterly, pull real files and a database back from backup
  • Failover test — at least yearly, stand a critical system up from scratch on alternate infrastructure
  • After changes — retest when you add or move a major system
  • Record the result and the recovery time, and fix what failed

Plan outline (copy this)

  1. Purpose and scope
  2. System inventory — table: system, location, data, backup target, dependencies
  3. RTO / RPO per system
  4. Backup and replication setup
  5. Recovery procedures — one per system or group, in recovery order
  6. Roles and contacts (offline copy)
  7. Communication plan
  8. Test schedule and results log
  9. Revision history

Cloud and hybrid considerations

Cloud services have their own resilience, but you are still responsible for your data, your configuration and your accounts — and “the provider had an outage” is not a recovery plan. Keep an independent backup of cloud data (including Microsoft 365), and know how you would operate if a provider is unavailable for a day. See cloud vs on-premise for where each model’s risks sit.

Common gaps

  • Backups run but restores are never tested
  • The plan and credentials only exist on the systems they are meant to recover
  • No recovery order, so teams work on the wrong system first
  • Microsoft 365 assumed to be “backed up” — it is not, by default
  • One person knows how everything works

When to get help

Building and testing a DR plan is exactly what a managed IT provider does. We inventory your environment, set targets with you, put 3-2-1 backups and recovery in place, and test on a schedule — part of our managed IT support and backup and disaster recovery services.

Want a disaster recovery plan you have actually tested? See our backup and disaster recovery service or talk to our Edmonton team.

Frequently asked questions

What is a disaster recovery plan?

A documented plan for restoring your IT systems and data after a disruptive event – a failed server, ransomware, a site loss. It defines what gets recovered, in what order, how fast, and by whom.

What should a disaster recovery plan include?

An inventory of systems and data, recovery time and recovery point targets for each, the backup and replication setup, step-by-step recovery procedures, roles and contacts, and a testing schedule.

How is a disaster recovery plan different from a backup?

Backups are one input. The DR plan is the whole procedure – where backups are, how to rebuild or fail over each system, in what order, and how to confirm it worked. A backup with no plan and no tested restore is not disaster recovery.

How often should a disaster recovery plan be tested?

Restore tests monthly or quarterly, and a fuller failover test at least once a year. Test after any major infrastructure change. Untested recovery is an assumption, not a plan.