Purpose
Make sure business data can be recovered after deletion, failure, ransomware or disaster, within the time and data loss the business can accept.
Scope
Covers servers, databases, cloud services and shared files that hold business data. Personal devices, one-off migrations and archiving for legal retention are covered by their own procedures.
Roles and responsibilities
| Role | Responsibility |
|---|---|
| Backup administrator | Configures and monitors backups, handles failures and performs restores. |
| Data owner | Decides how important each data set is and how much loss is acceptable. |
| IT manager | Approves the policy, reviews the monthly report and reports risks to management. |
| Reviewer | Checks the restore tests and confirms the results are recorded. |
Prerequisites
- An inventory of systems and data sets, with an owner for each
- A recovery point objective and recovery time objective agreed with each data owner
- Backup software or the cloud provider's backup service, with admin accounts held by named people
- Storage for local backups and a separate location or provider for the off-site copy
- A place to store encryption keys that is not on the systems being backed up
Procedure
- List what must be backed up. With the data owners, list systems, databases, cloud accounts, configuration and shared files. Mark which data is critical, and note where each lives.
- Set recovery objectives. For each data set write the longest acceptable data loss (recovery point) and the longest acceptable downtime (recovery time). These decide how often you back up and how you restore.
- Choose the schedule and retention. Set the frequency from the recovery point, for example hourly or daily for critical data, and the retention, such as daily copies for a month and monthly copies for a year, as your needs and rules dictate.
- Follow the 3-2-1 rule. Keep three copies of important data: the live copy and two backups, on two different kinds of storage, with one copy off site and, where you can, one that cannot be changed or deleted for a set time.
- Encrypt and protect the backups. Encrypt backups in transit and at rest, store keys separately, use dedicated backup accounts with multi-factor authentication and do not let ordinary user accounts delete backups.
- Configure the jobs. Set up the jobs, include databases using their own export or snapshot method so files are consistent, and write down the job names, schedules and destinations in the inventory.
- Monitor every job. Send success and failure results to a mailbox or channel that someone reads. A backup that fails silently is the same as no backup.
- Handle failures. On a failure, the administrator reruns the job the same day, finds the cause, and tells the data owner if the recovery point objective has been missed.
- Test a restore every month. Restore a chosen file, database or server into a test location, check that the data opens and is complete, time the restore and compare it with the recovery time objective. Rotate which data you test so all of it is covered in a year.
- Record the test. Write the date, what was restored, who did it, how long it took and the outcome in the restore log. The reviewer signs it. Fix and re-test any failure.
- Review the policy. At least once a year and after any major change of systems, review the inventory, objectives, schedule and retention, remove backups of retired systems according to your policy and update this document.
Revision history
| Version | Date | Author | Change |
|---|---|---|---|
| 1.0 | [Date] | [Name] | First version |