Active Directory Backup Guide: System State, Full Server and Restore Steps for Windows Server
Advertisement
Active Directory (AD) sits behind logins, Group Policy, DNS and single sign-on for most Windows-based companies, so losing it stops almost every other system. Your CRM, HRIS, ITSM and SSO tools all depend on it.
This how-to is for Windows Server administrators and IT managers. It covers backup types, command-line steps, restore methods, retention and the tooling choices that matter in 2026.
What an Active Directory backup must contain
A usable Active Directory backup must include the system state of at least one writable domain controller: the AD database (ntds.dit), the SYSVOL folder with Group Policy files, the registry, boot files and DNS data if the server hosts it. A file copy of ntds.dit alone is not a supported backup.
System state is the unit Microsoft supports for restoring a domain controller (DC). It is captured through the Volume Shadow Copy Service (VSS), which keeps the database consistent while the DC stays online.
A full server (bare-metal) backup adds the operating system and applications. Teams usually run both: system state for fast directory recovery, and full server or image backups for rebuilding hardware after ransomware or disk failure.
| Backup type | What it captures | Best used for | Limit |
|---|---|---|---|
| System state | AD database, SYSVOL, registry, boot files, DNS (if integrated) | Directory recovery on the same DC | Not a full OS image |
| Full server / bare metal | All volumes, OS, applications, system state | Rebuilding a failed or encrypted DC | Larger and slower |
| AD Recycle Bin | Deleted objects with attributes, kept online | Undoing an accidental delete in minutes | Does not undo bad attribute edits |
| Image-based VM backup | Whole DC virtual machine at the hypervisor | Fast restores of virtualized DCs | Needs VM-generation-ID aware restore |
Pick your recovery scenario before you pick a backup tool
Most Active Directory incidents fall into three scenarios: an object was deleted, a bulk change corrupted data, or the entire forest is lost. Each needs a different restore method, so map your backups to these cases first.
Scenario 1: an object was deleted
If the AD Recycle Bin is enabled, restore the object online with PowerShell, such as Get-ADObject -Filter {isDeleted -eq $true} -IncludeDeletedObjects followed by Restore-ADObject. No backup restore is needed, which is why enabling the Recycle Bin is the cheapest protection you can add.
Scenario 2: bad data was replicated
When a script overwrites attributes or deletes an organizational unit tree, replication spreads the damage. You need an authoritative restore from system state backup taken before the change, using ntdsutil in Directory Services Restore Mode (DSRM).
Scenario 3: the forest is gone
After ransomware or a total site loss, you rebuild from a clean backup of one DC per domain, then reinstall the others. Microsoft publishes a detailed forest recovery guide, and the CISA #StopRansomware Guide stresses offline, tested backups for exactly this case.
How to back up Active Directory with Windows Server Backup
Install the Windows Server Backup feature, then run wbadmin start systemstatebackup -backupTarget:E: -quiet from an elevated prompt on a writable DC. The target must be a local volume or a mapped drive, and the job takes a VSS snapshot of system state.
- Install the feature:
Install-WindowsFeature Windows-Server-Backup. - Pick a target volume that is not the system volume and has free space for at least one full system state set.
- Run the system state backup with wbadmin, or schedule it with
wbadmin enable backup -addtarget:E: -schedule:21:00 -systemState. - Confirm the result with
wbadmin get versionsand check the event log for backup success. - Copy the backup set to a second location that domain credentials cannot reach.
The command-line reference is in the Microsoft wbadmin documentation. Scheduled jobs reuse the same engine as the GUI, so you can mix both approaches across DCs.
How to restore: non-authoritative and authoritative
A non-authoritative restore brings a DC back from backup and lets replication update it from healthy partners. An authoritative restore marks restored objects as newer so they overwrite the rest of the domain. Use the second only for deliberate rollback.
Boot the DC into DSRM (bcdedit /set safeboot dsrepair, then restart) and restore with wbadmin start systemstaterecovery -version:<version>. For an authoritative restore, run ntdsutil, choose authoritative restore and restore the subtree or object. Add -authsysvol only when SYSVOL must also win replication.
Rehearse this in a lab at least twice a year. Most failed recoveries trace back to forgotten DSRM passwords, backups older than the tombstone lifetime, or no one remembering the sequence.
Backup frequency, retention and the tombstone limit
Back up at least one DC per domain daily, and two DCs including the holder of the FSMO roles for large environments. Never keep a restore point older than the forest tombstone lifetime, which is 180 days by default in current forests and 60 in older ones.
A backup older than the tombstone lifetime reintroduces "lingering objects" and cannot be safely restored. Check yours with Get-ADObject "CN=Directory Service,CN=Windows NT,CN=Services,CN=Configuration,DC=corp,DC=example" -Properties tombstoneLifetime.
- Daily system state backup of one DC per domain, twice daily for fast-changing estates.
- Weekly full server or image backup of the PDC emulator and one other DC.
- Offsite or immutable copy, kept under the 3-2-1 rule (three copies, two media types, one offsite).
- Retention shorter than the tombstone lifetime, with a quarterly restore test.
For recovery targets, use the formula below to turn business needs into a schedule.
If you back up once every 24 hours and copy offsite for 2 hours, plan for up to 26 hours of directory changes lost. Tighter RPO needs more frequent system state jobs or more DCs replicating live.
Hardening backups against ransomware
Attackers who compromise AD usually delete or encrypt the backups next. Protect them with separate credentials, network isolation, immutability and tested offline copies, as recommended in NIST contingency guidance.
The NIST SP 800-34 contingency planning guide defines backup strategy, alternate sites and recovery testing as core contingency controls. Apply them to AD by following these rules:
- Do not join the backup server to the production domain it protects, or use a dedicated, tiered admin account.
- Store at least one copy in immutable object storage or on offline media.
- Alert on VSS shadow copy deletion and on unexpected
wbadmin deleteactivity. - Document the DSRM password escrow process and test it.
Third-party backup platforms for domain controllers
Native tools are enough for a single forest with a good restore routine. Dedicated platforms add central scheduling, immutable storage, application-aware processing, reporting and faster recovery of virtualized DCs for larger teams.
SaaSPodium lists these enterprise backup products in the Backup & Archiving Software category. Confirm AD-specific recovery features with each vendor before buying:
- Veeam Backup & Replication: image-based backup of virtualized DCs with application-aware processing.
- NAKIVO Backup & Replication: VM and physical server protection with a lower entry price.
- Acronis Cyber Protect: backup combined with endpoint security, useful against ransomware.
- Commvault Complete Data Protection and Cohesity Data Cloud: enterprise platforms with policy-based retention.
- Zerto: continuous data protection for near-zero RPO of virtual machines.
For a wider buying view, read our guide to the best backup and archiving software and the Acronis vs Veeam vs Datto comparison.
Testing and monitoring your AD backups
A backup is only proven when you restore it. Test quarterly in an isolated network, record the time taken and fix gaps found, then monitor backup jobs and replication health daily.
- Restore a system state backup to an isolated VM, boot in DSRM and verify
dcdiagandrepadmin /replsummary. - Check that SYSVOL and Group Policy objects are present after the restore.
- Track backup job success, age of the newest restore point and offsite copy lag on a dashboard.
- Write the runbook, including who holds DSRM credentials and the order of FSMO role recovery.
Frequently Asked Questions
How often should I back up Active Directory?
Daily at minimum for one DC per domain, and more often when you make frequent bulk changes. Always keep restore points inside the tombstone lifetime.
Is the AD Recycle Bin a replacement for backups?
No. It recovers deleted objects quickly, but it cannot undo corrupted attributes, a deleted domain or a ransomware event. You still need system state and image backups.
Do I need a paid tool to back up Active Directory?
Not for small environments. Windows Server Backup handles system state well. Paid platforms make sense when you need central management, immutability, many DCs or tight recovery targets.
Advertisement