Topic Hub

Windows Server

Active Directory, DNS, DHCP, file services, patching, hardening, backup, monitoring, and PowerShell administration. Practical guidance for sysadmins running Windows Server in on-premises, hybrid, and cloud-connected environments.

Guides, scripts and analysis

Overview

What Windows Server Is Still Used For

Microsoft 365, Entra ID, and Intune have moved many workloads to cloud management, but Windows Server remains the control plane for identity, name resolution, file access, line-of-business apps, and server-side infrastructure in many real environments.

Active Directory and Kerberos

Domain controllers still authenticate users, computers, and service accounts for domain-joined endpoints, file shares, SQL Server, legacy apps, and workloads that depend on Kerberos, LDAP, or NTLM compatibility.

Core network services

Windows Server commonly hosts internal DNS, DHCP, IPAM, NPS/RADIUS, certificate services, and routing-adjacent roles that keep branch offices and datacenters reachable.

File, print, and app hosting

SMB shares, DFS namespaces, print queues, scheduled tasks, IIS apps, and vendor line-of-business services often remain on Windows Server even when collaboration and endpoint management move to Microsoft 365.

Server-side patch and compliance scope

Intune manages Windows clients, not traditional Windows Server workloads. Server patching still usually depends on WSUS, ConfigMgr, Azure Update Manager, maintenance windows, or operational runbooks.

Hybrid identity and access bridge

Entra ID extends identity to cloud apps, but many tenants still sync from AD DS. Group Policy, AD sites, DNS, and domain controller placement continue to affect sign-in, device join, and application access.

Operational boundary for regulated systems

Some environments keep critical services on controlled networks for latency, data residency, vendor support, or isolation reasons. Windows Server is often the supported platform for those systems.

Identity

Active Directory Domain Services

AD DS is still the centre of many Windows Server estates. Treat it like critical infrastructure: design for replication, recovery, auditing, delegation, and controlled change.

Keep authentication reliable

Domain controller health

Core checks
Monitor AD replication, DNS registration, SYSVOL health, time sync, disk space, and Directory Services event logs before users notice authentication failures.
Placement
Place writable domain controllers close to sites that need low-latency sign-in, and use RODCs only where physical or administrative trust is limited.
FSMO roles
Know where FSMO roles live and document transfer or seizure steps. Most outages are not FSMO-related, but recovery becomes harder when nobody knows role ownership.
DNS dependency
AD depends heavily on DNS SRV records. Broken DNS registration can look like random logon, Group Policy, LDAP, or domain join failure.

Make administration survivable

Directory design

OU model
Design OUs around administration and policy scope, not only company org charts. Keep server, workstation, privileged, and service-account objects easy to target.
Delegation
Delegate specific OU and group tasks instead of granting broad Domain Admin rights. Privileged group membership should be short, reviewed, and alertable.
Service accounts
Use group managed service accounts where supported. For legacy accounts, document SPNs, owners, password rotation process, and dependencies before changes.
Schema and functional level changes
Treat schema changes and forest or domain functional level raises as change-controlled operations with rollback planning and verified backups.

AD plus cloud identity

Hybrid identity

Entra Connect
Hybrid tenants commonly sync users, groups, and devices from AD DS. Sync health, object filtering, UPN alignment, and duplicate attributes can all affect cloud sign-in.
Device join
Hybrid Entra join depends on AD, Entra Connect, SCP configuration, device registration endpoints, and certificate renewal. Use dsregcmd /status on affected clients.
Conditional Access impact
Cloud access policies can depend on device state that begins on-premises. Stale or duplicate AD/Entra device objects can create confusing access failures.
Admin model shift
Cloud tools do not remove AD operational duties. They add a second control plane, so changes need to be validated in both AD and Entra ID.

Infrastructure roles

DNS, DHCP, File Services, Print, and Core Roles

The boring roles are often the business-critical ones. Keep them simple, documented, monitored, and recoverable.

DNS Server

AD-integrated DNS zones, forwarders, conditional forwarders, scavenging, and secure dynamic updates directly affect authentication and application access. Test name resolution before chasing higher-level symptoms.

DHCP Server

Scope design, exclusions, reservations, relay/IP helper configuration, failover pairs, and DNS dynamic update credentials determine whether clients receive usable network configuration.

File services and SMB

SMB shares need access-based enumeration, NTFS/share permission hygiene, quota planning, shadow copies where appropriate, and clear ownership for stale data cleanup.

DFS namespaces and replication

DFS namespaces can simplify user paths, but DFSR backlog, staging size, and conflict handling need monitoring. Do not treat DFSR as a general backup system.

Print services

Print queues are still common in offices, warehouses, healthcare, and manufacturing. Driver isolation, queue naming, deployment method, and spooler hardening matter more than the role gets credit for.

Certificates, NPS, IIS, and app roles

AD CS, NPS/RADIUS, IIS, Remote Desktop Services, and vendor apps often become hidden dependencies. Record owners, ports, certificates, service accounts, and recovery steps for each.

Hybrid management

Group Policy and Hybrid Management

Windows Server estates rarely use one management plane. Group Policy, ConfigMgr, Entra ID, Intune, and Azure Arc each solve different parts of the job.

Use GPO where domain context matters

Server baselines, Windows Firewall, audit policy, user rights assignment, certificate auto-enrollment, WSUS targeting, and local security policy are common server-side GPO use cases.

Avoid duplicate authority

Hybrid-joined workstations can receive both GPO and Intune policy. Servers can also receive ConfigMgr or Arc policy. Define one owner per setting, especially update, Defender, BitLocker, and firewall controls.

Intune changes endpoint administration

Intune moves many Windows client policies to MDM and Settings Catalog. That does not remove the need to manage domain controllers, DNS, DHCP, file servers, or server patch windows.

Azure Arc and server governance

Azure Arc can surface inventory, policy, update, and monitoring workflows for servers across environments. Treat it as a management overlay, not a replacement for AD, DNS, backup, and local recovery planning.

RSAT and Windows Admin Center

RSAT remains the daily toolkit for ADUC, DNS, DHCP, GPMC, and Failover Cluster Manager. Windows Admin Center adds browser-based server management without requiring the server itself to move to the cloud.

Document the boundary

For each server class, document whether it is managed by GPO, ConfigMgr, Arc, manual PowerShell, or a vendor console. Most operational confusion starts when that boundary is assumed rather than written down.

Maintenance

Patching, Maintenance Windows, and Reboot Planning

Server patching is less about clicking install and more about dependency order, rollback, maintenance windows, and business communication.

Routine patching

Standalone and member servers

Patch source
Use WSUS, ConfigMgr, Azure Update Manager, or a documented manual process. Make sure the server is not receiving conflicting update policy from multiple sources.
Window design
Group servers by business service, criticality, dependency order, and reboot tolerance. A generic monthly server window is rarely enough for databases, app tiers, and domain controllers.
Pre-checks
Confirm backups, free disk space, pending reboot state, service owner approval, and monitoring suppression before the window begins.

Identity-sensitive

Domain controllers

Sequence
Patch one domain controller per site or tier at a time where possible. Verify replication and authentication before moving to the next server.
Health checks
Run dcdiag, repadmin /replsummary, DNS checks, and event log review before and after patching. Authentication failures after reboot are often DNS, time, or replication symptoms.
Recovery
Know the difference between restoring a failed DC from backup, rebuilding it, and performing an authoritative restore. Do not improvise during an outage.

Availability-aware

Clusters and critical apps

Cluster-Aware Updating
Use CAU or an equivalent runbook to drain roles, patch one node, reboot, validate, and move on. Never patch every node in a cluster at once.
Application order
Patch dependencies first or last according to the app owner runbook. File, database, identity, and middleware layers often have strict order requirements.
Rollback limits
Some updates are not cleanly removable. Test on non-production systems, capture VM snapshots only where supported, and have service-level recovery steps ready.

Before the May server maintenance window, cross-check the May 2026 Patch Tuesday KB and known issue notes for Windows Server build targets, WSUS notes, and Server 2025 hotpatch applicability.

Hardening

Server Hardening and Baseline Configuration

A server baseline should reduce attack surface without breaking the workload. Start with role-aware controls, test them, and track exceptions explicitly.

Role-aware baselines

Domain controllers, file servers, IIS servers, SQL hosts, and jump boxes need different baselines. Apply common controls broadly, then layer role-specific policy on top.

Local admin and privileged access

Limit local Administrators membership, use Windows LAPS where supported, separate daily and privileged accounts, and alert on membership changes to privileged groups.

Firewall and service exposure

Enable Windows Firewall, restrict inbound management ports, disable unused roles and services, and document exceptions for vendor apps rather than leaving broad allow rules.

Credential and protocol hardening

Review NTLM usage, SMB signing/encryption requirements, LDAP signing/channel binding, TLS configuration, and legacy protocol dependencies before enforcing controls.

Audit policy and logging

Apply advanced audit policy for logon, account management, directory service changes, object access where needed, PowerShell logging, and process creation with command line for high-value servers.

Configuration drift

Use GPO, ConfigMgr baselines, Desired State Configuration, Azure Policy guest configuration, or scripted checks to detect drift instead of relying on one-time build hardening.

Operations

Event Logs, Monitoring, Backup, and Disaster Recovery

The recovery plan matters more than the monitoring dashboard. Make sure alerts lead to action and backups are restored often enough to prove they work.

Event logs to watch

System, Application, Security, Directory Service, DNS Server, DFS Replication, GroupPolicy/Operational, WindowsUpdateClient/Operational, and role-specific logs should feed your monitoring or SIEM.

Practical alerting

Alert on service failures, disk exhaustion, failed backups, replication errors, certificate expiry, repeated authentication failures, unexpected reboots, and critical role-specific events.

Backup scope

Back up system state for domain controllers, application-aware data for workloads, file server volumes, certificates/private keys, GPO backups, scripts, scheduled tasks, and documented configuration.

Restore testing

A backup that has never been restored is only a hope. Test file restore, bare-metal restore where needed, system state restore, and application recovery in a lab or isolated recovery network.

AD disaster recovery

Know when to rebuild a domain controller versus restore it. Practice DSRM access, authoritative restore scenarios, accidental deletion recovery, and forest recovery documentation before a real incident.

Change and incident evidence

Preserve event logs, PowerShell transcripts, GPO change records, backup job history, and patch deployment logs. These records are often the fastest path from outage to root cause.

Administration

PowerShell, RSAT, and Day-to-Day Administration

Good Windows Server administration is a mix of consoles, scripts, remoting, documentation, and repeatable checks.

RSAT from an admin workstation

Run ADUC, DNS Manager, DHCP Manager, GPMC, Failover Cluster Manager, and Server Manager from a hardened admin workstation rather than signing into servers for routine work.

PowerShell remoting

Use PowerShell remoting and CIM sessions for repeatable server checks. Standardize modules, execution policy, transcript logging, and Just Enough Administration where appropriate.

Server Manager and Windows Admin Center

Server Manager remains useful for role visibility and multi-server tasks. Windows Admin Center gives a browser-based option for local and remote server administration.

Scheduled maintenance scripts

Automate pre-patch checks, disk cleanup reporting, service status snapshots, certificate expiry checks, and event log summaries so maintenance windows start with known state.

Least privilege workflows

Use separate admin accounts, Privileged Access Workstations where needed, role-based groups, and time-bound elevation. Avoid using Domain Admin for DNS, DHCP, file share, and print tasks.

Inventory and ownership

Maintain server role, owner, patch ring, backup policy, certificate, service account, and recovery runbook metadata. The server nobody owns becomes the incident nobody can fix.

Common problems

Where Windows Server Environments Go Wrong

Most incidents are not mysterious. They are undocumented dependencies, stale infrastructure, weak recovery testing, or overlapping management authority.

DNS breaks authentication

Clients or servers point to external DNS, stale domain controllers, or missing SRV records. AD symptoms then appear as logon failure, domain join failure, GPO failure, or application outage.

DHCP failover not actually healthy

Scopes exist on two servers, but failover replication, DNS update credentials, exclusions, or relay configuration are wrong. Clients receive leases but not the options they need.

Domain controllers patched without validation

All DCs reboot in the same window, then replication, DNS, or time sync issues surface after the change. Patch in sequence and verify health between servers.

File permissions drift for years

Nested groups, direct user ACLs, inherited permissions, and orphaned SIDs make access reviews painful. Fix ownership and group design before attempting a large cleanup.

Backups exist but restores fail

Backup jobs report success, but credentials, encryption keys, application consistency, or restore media were never tested. Recovery objectives should be proven, not assumed.

GPO, ConfigMgr, and manual settings conflict

A server receives one setting from GPO, a different one from ConfigMgr, and a manual registry change from an old runbook. Document the authority per setting and remove stale policy.

Certificates expire quietly

IIS, LDAPS, RADIUS, VPN, Wi-Fi, and internal apps depend on certificates. Expiry monitoring and ownership records prevent outages that look unrelated at first glance.

Admins manage servers interactively

Routine RDP sign-ins with broad admin rights increase risk and make changes hard to audit. Prefer RSAT, PowerShell remoting, Windows Admin Center, and privileged workflow controls.