Covering DNS's critical role in Active Directory: AD-integrated zones, the _msdcs zone, SRV records, the DC locator, conditional forwarders, and DNS troubleshooting steps with nslookup and dcdiag.

In episode 3, you understood AD's physical structure — sites, subnets, and replication. Now we touch the component that most often makes AD "not work" without ever breaking: DNS. If clients can't log in, the number one cause is almost always DNS, not AD itself.
Remember the analogy from episode 0: DNS is the phone book. Imagine a phone book that's wrong — the numbers are incorrect or pages are missing. The house (AD) still exists and works, but nobody can find it. Most mysterious AD problems are rooted in something this small.
AD is completely dependent on DNS for two big things:
Unlike other directory services that can run without DNS, AD actually uses DNS as its locator service. The domain controller list, service locations, even AD sites are discovered through DNS. Without correct DNS, a domain controller can be perfectly healthy yet never found by anyone.
When a computer wants to join a domain or a user wants to log in, a process called DC locator runs. The flow roughly looks like this:
ad.example.com.ad.example.com?This process is "site-aware": with subnets mapped (episode 3), the client first looks for a DC in its own site before going to another site. If DNS is wrong or an SRV record is missing, the client never reaches step 4 — and the errors that appear are often confusing, like "domain not found" while the DC is actually online.
Normal DNS zones are stored in a zone file on a single DNS server. AD introduced AD-integrated zones: zone data is stored in the Active Directory partition instead of a text file. The consequences are significant:
Because replication is automatic, an AD-integrated zone is the right default choice when DNS is installed together with AD.
When AD is installed, besides the normal domain zone, a special zone named _msdcs.ad.example.com is created. This is a forest-wide zone that stores all the critical SRV records and must replicate across the entire forest — not just one domain.
SRV records (Service Records) tell clients which services are available and where. Their format is distinctive: the service name followed by _tcp or _udp, then the domain name. Key examples:
| SRV Record | Service | Function |
|---|---|---|
_ldap._tcp.dc._msdcs.ad.example.com | LDAP | Finds a DC for LDAP authentication |
_kerberos._tcp.dc._msdcs.ad.example.com | Kerberos | Finds the KDC for Kerberos tickets |
_gc._tcp.ad.example.com | Global Catalog | Finds a Global Catalog server |
_kpasswd._tcp.ad.example.com | Password change | Finds the password change service |
Each SRV record contains the service port and the host providing it. Clients use these records to find DCs; if a record is missing or wrong, the DC might as well not exist.
Besides SRV, a DC registers several other records automatically when it joins AD:
| Record Type | Example | Function |
|---|---|---|
| A (Host) | DC01 → 192.168.10.10 | Translates the DC's hostname to an IP |
| SRV | _ldap._tcp.dc._msdcs... | Service locations (LDAP, Kerberos, GC) |
| CNAME | Records for specific connections | Additional aliases for certain services |
| Dynamic registration | Created automatically | Records updated by the DC itself when changes occur |
This registration is dynamic — DCs and domain-joined computers register their own records. That's why you should never disable dynamic updates in an AD zone; manually frozen records go stale and start pointing to wrong addresses.
A conditional forwarder is a DNS setting that answers: for a specific domain, forward queries to a specific DNS server. It's used when your domain needs to resolve names in another domain whose DNS you don't control — for example, another forest's domain or an external vendor.
Unlike a global forwarder (all unknown queries get forwarded), a conditional forwarder is specific per domain, so it's more targeted. It can be configured from DNS Manager or PowerShell:
Set-DnsServerConditionalForwarderZone -Name "partner.example.net" -MasterServers "10.0.0.53" -UseRecursionWithout a correct conditional forwarder, forest trusts or integration with other systems often fail with name-resolution errors that are hard to trace.
Most DNS problems in AD can be traced with two tools: nslookup to inspect records, and dcdiag for a comprehensive diagnosis. Verify DC SRV records:
nslookup -type=SRV _ldap._tcp.dc._msdcs.ad.example.comCorrect output shows one or more DC hosts complete with port and weight. If the answer is "Non-existent domain" or empty, the zone or SRV records are at fault.
For a comprehensive diagnosis, use dcdiag, which runs a suite of tests including DNS:
dcdiag /test:dnsBesides those two tools, watch the DNS Server event log for traces of failed updates, and always start from the most basic question: "Does the SRV record actually exist in the right zone?" The answer is often there. Another effective troubleshooting pattern: force the DC to re-register its records by restarting the Netlogon service on the DC.
In episode 4 you understand why DNS is AD's lifeline: DNS is the locator service clients use to find DCs, AD-integrated zones replicate DNS data together with AD, the _msdcs zone stores critical service SRV records, and conditional forwarders bridge other namespaces.
Key takeaways:
_ldap, _kerberos, _gc, and _kpasswd SRV records are the key to the DC locator process.nslookup -type=SRV and dcdiag /test:dns are the first tools for DNS troubleshooting.In the next episode, all this theory will be proven: installing Active Directory Domain Services — from prerequisites, installing the AD DS role via PowerShell, promoting the first domain controller with Install-ADDSForest, to verifying with dcdiag and ntdsutil. Make sure your lab's DNS is ready, because this is the episode where your first domain is born!