If your business runs Active Directory on a single domain controller, one failed disk, bad patch or ransomware infection can stop everyone from logging in. The fix is a second domain controller, placed on separate hardware, running DNS and the Global Catalog, with the key roles spread between the two and a recovery routine that you have actually tested. Our managed services team builds and monitors this setup for Irish organisations as part of looking after their infrastructure.
This guide explains the idea in plain terms, then walks through the steps, including where each one tends to go wrong.
TL;DR
- A domain controller (DC) is the server that checks who is allowed to log in and applies your Group Policy. With only one, its failure disrupts authentication, policy and access to resources.
- Run at least two DCs per domain, on different physical hosts, each providing DNS and the Global Catalog.
- Spread the FSMO roles (the special single-owner jobs in Active Directory) and document how to move them before an outage.
- Back up at least two DCs, monitor replication, and rehearse recovery in an isolated environment.
- More DCs than you need adds management work and attack surface, so two is the right start for most small and medium businesses.
Table of Contents
- What domain controller redundancy means
- When you need a second domain controller
- How to deploy two domain controllers without a single point of failure
- How replication keeps the copies in step
- What happens when a DC fails
- Checking that it works
- A worked example
- Perspective
- How we can help
- FAQ
- Sources
What domain controller redundancy means
Active Directory is the directory that holds your users, computers and permissions. A domain controller is a server that holds a copy of it and answers every login request. Redundancy means having several DCs that replicate the directory data between them, so that if one fails another keeps answering. OVHcloud describes this as multiple active servers sharing replicated data so users can still log in and find resources when one server goes down.
The practical cost of not having it is easy to picture. Without a second DC, a single failure can disrupt authentication, Group Policy application and access to critical resources. Staff cannot sign in to new sessions, shared drives and applications that rely on Active Directory stop responding, and your team spends the day on recovery rather than work.
When you need a second domain controller
The short answer is: as soon as Active Directory matters to how your business runs. If people log in to Windows PCs with a company account, use file shares, run Microsoft 365 with synchronised identities or rely on Group Policy, you are in that position. Guidance for production environments is to add another writable DC when only one exists, and to run Active Directory-integrated DNS on both where appropriate.
How many is enough? Microsoft’s design guidance is to start with the minimum number of controllers the workload needs and increase from there, because each extra controller adds attack surface and management overhead. For a small or medium business in one office, that minimum is two.
Startups and cloud-first firms
If your team is small and mostly in Microsoft 365, you may still have on-premises Active Directory for file servers, printers or legacy applications. The same rule applies while it exists. Our Microsoft 365 and cloud services work includes planning how identity should look as you move to the cloud, so you do not build redundancy for a system you are about to retire, or retire one you still depend on.
How to deploy two domain controllers without a single point of failure
Step 1: Put them on different hosts
Two DCs on the same physical server are one failure waiting to happen. Microsoft’s guidance for virtualised DCs is to deploy at least two per domain on different virtualisation hosts and, if possible, on hardware in different regions. Where it goes wrong: a second virtual machine that quietly lives on the same host, same storage and same power feed as the first. Check the placement, not just the server count.
If you run a hypervisor cluster, keep one DC outside it. A cluster-wide outage can otherwise stop the cluster from starting because it needs the directory to come up first, a pattern described for Azure Local that applies to any cluster that authenticates against its own domain controllers.
Step 2: Give each DC DNS and Global Catalog
Active Directory depends on DNS to be found. A second DC that does not answer DNS queries is little use when the first goes down. Microsoft recommends that all domain controllers provide both DNS and Global Catalog services to keep the domain highly available. The Global Catalog is the searchable index of the whole directory that logins and applications use.
Then make sure your PCs and servers know about both. Point DHCP and server network settings at both DCs for DNS. The common failure here is a second DC that exists, but every device still points only at the first.
Step 3: Spread the FSMO roles
FSMO (Flexible Single-Master Operations) roles are five jobs in Active Directory that only one controller holds at a time, such as handing out account identifiers. Microsoft Press advises that once a domain has more than one DC, you should manually distribute the FSMO roles across controllers so one failure does not take them all away.
Also write down how you would seize a role if its holder died permanently. Document the seizure procedure before an outage, because nobody wants to research it for the first time during one.
Step 4: Plan sites if you have more than one location
If you have several offices linked by VPN or WAN, use multiple Active Directory sites so that staff log in to the nearest DC and replication follows sensible paths. Sites suit distinct locations and WAN or VPN links where client locality matters. For a remote branch with limited physical security, a read-only domain controller can reduce credential exposure.
How replication keeps the copies in step
You do not build replication links by hand. The Knowledge Consistency Checker runs on every DC and generates and adjusts the replication topology automatically. When a connection fails because a DC is down, it builds temporary connections to other available partners so that replication carries on.
Two settings are worth checking. Enable Strict Replication Consistency through Group Policy on domain controllers, which helps prevent lingering objects reappearing from a DC that was offline too long. And keep the directory itself tidy: stale accounts and computers make troubleshooting harder, which is why we use a controlled approach to Active Directory cleanup.
What happens when a DC fails
With redundancy in place, users continue accessing resources during component failures, and spreading DCs across locations protects against the loss of a whole site. Sign-ins carry on against the surviving DC, and you replace the failed one at a sensible pace instead of in an emergency.
The recovery rule is counter-intuitive. Microsoft’s forest recovery guidance says restoring an Active Directory backup should be the last option. For a single failed DC, the route is to remove it, wait a few hours for replication and promote a fresh one. That is exactly why a second DC matters: the directory survives on the healthy copy, and you rebuild from it rather than rolling back to an old backup. The same thinking applies to any rollback; planning recovery as a last resort means rehearsing the steps and knowing what each one changes before you need them.
Checking that it works
Redundancy you have not tested is an assumption. A routine we recommend has four parts.
Check replication
Use Repadmin, Microsoft’s command-line tool for replication. Running repadmin /showrepl displays replication errors, and the tool gives you information to monitor and resolve failures. If you ask how to force replication between two DCs, start here: find out why it is failing before pushing it.
Check health
Microsoft’s guidance includes regular patching, checking the NTDS.DIT database with DCDIAG and reviewing system event logs. Patching matters in both directions: a bad patch can take out a DC, and an unpatched one is an easy target, so patch the two DCs on different days. Our server patch management plan shows how we approach that.
Back up more than one DC
Microsoft recommends backing up at least two DCs for each domain, one of which holds an operations master role. If DCs run as virtual machines, use hypervisor features such as VM-GenerationID to prevent database corruption when backups are restored. Schedule system-state backups daily and keep a copy off site. Our business backup and disaster recovery plans cover that side.
Rehearse recovery
Restoration drills should be practised periodically in an isolated test environment so you know the method works during a real failure. Switch off one DC in a quiet window, confirm that logins, DNS and file access carry on, then bring it back. Our guide to testing disaster recovery sets out how to do this without guesswork.
A worked example
Take a 40-person Wicklow firm with one physical server running Active Directory, file shares and the accounting application. The server’s disk controller fails on a Monday morning. With one DC, nobody can sign in until the hardware is repaired and the directory is restored from backup, which could take a day or more.
Now take the same firm with a second DC as a virtual machine on a separate host, both running DNS and Global Catalog, and DHCP handing out both addresses. The failure happens, staff keep signing in, and the IT team replaces the failed server without time pressure. Moving the FSMO roles, if the failed server held them, is a documented step rather than a research project.
Perspective
In our experience of looking after infrastructure for Irish businesses, a second domain controller is one of the best-value pieces of resilience a company can add. It does not need new software or a large budget, but it does need to be placed, configured and monitored properly.
We would rather see two well-managed DCs than four neglected ones. A DC that nobody checks quietly stops replicating, and the first time anyone finds out is during the outage. That is why we pair redundancy with proactive monitoring, patching and tested backups, so the second server is genuinely ready.
Our approach is to treat identity as part of the wider infrastructure: networking, security, backup and disaster recovery, and your Microsoft 365 plans together. With 30 years of experience, we act as an extension of your own team, so that your staff can focus on the business and not on the server room.
Carl
How we can help
We design, monitor and support Active Directory environments as part of our managed services, with Silver, Gold and Platinum Support levels, proactive monitoring and both remote and onsite help. We also provide backup and disaster recovery and IT consultancy, so one team can plan the second domain controller, place it correctly, protect it and test it with you.
If you run a single domain controller today, get in touch with LANCAST and we will review your setup with you.
FAQ
How many domain controllers does a small business need?
Two per domain is the sensible starting point. Microsoft advises starting with the minimum number the workload needs and adding more only when load or locations call for it, since each extra controller adds management and attack surface.
Can I run the second domain controller as a virtual machine?
Yes. Place it on a different virtualisation host from the first, and ideally in a different location. Keep one DC outside any cluster that depends on the domain to start.
Does a second domain controller replace backups?
No. A second DC keeps logins running when one fails, while backups protect against corruption, deletion and disasters affecting both. Microsoft recommends backing up at least two DCs per domain, one holding an operations master role.
How do I check that my domain controllers are replicating?
Run repadmin /showrepl from an administrative command prompt. It displays replication errors between partners, and DCDIAG gives a wider health check.
Sources
- What is Server Redundancy? OVHcloud Ireland
- Implementing Redundancy to Prevent Active Directory Failures
- Active Directory Replication Topology: 2-DC Guide
- Active Directory Design Considerations and Best Practices, Microsoft Learn
- Virtualizing domain controllers with Hyper-V, Microsoft Learn
- Azure Local: Redundancy
- Install a Replica Windows Server 2012 Domain Controller, Microsoft Learn
- Active Directory, Microsoft Press Store
- Active Directory Replication Concepts, Microsoft Learn
- Active Directory Forest Recovery, Microsoft Learn
- Rollback and Recovery: A Practical Guide for Integration Leaders
- Guidance for troubleshooting Active Directory replication, Microsoft Learn
- Active Directory: Automate System State Backup, Microsoft Learn
- Best Practices for Active Directory Setup and Configuration