Active Directory

Operations Guide
Part I: Active Directory Operations

Version 1.0

Developed by the Windows Resource Kits team

Microsoft Windows 2000

Microsoft Corporation
2 Contents

Acknowledgements
Program Managers: Stuart Kwan, Andreas Luther, Paul Reiner
Writers: Mary Hillman, Dave Kreitler, Merrilee McDonald, Randy McLaughlin, Andrea Weiss
Editors: Laura Graham and Justin Hall
Copy Editors: Anika Nelson and Dee Teodoro
Test Plan: Mary Hillman and Cheryl Jenkins
Testers: Justin Hall, David Stern, Matt Winberry
Lab Staff: Robert Thingwold and David Meyer
Lab Partners: Compaq, Inc. and Cisco Systems

We thank the following people for reviewing the guide and providing valuable feedback:
Tadao Arima, Bill Bagley, Duncan Bryce, J.C. Cannon, Sudarshan Chitre, Arren Conner, Joseph
Davies, Jim Dobbin, Levon Esibov, Eric Fitzgerald, David Golds, Jin Huang, Khushru Irani, J.K.
Jaganathan, Asaf Kashi, William Lees, Jonathan Liem, Doug Lindsey, Arun Nanda, Paul
O’Connell, Boyd Peterson, Paul Rich, Sanjiv Sharma, Michael Snyder, David Stern, Mark
Szalkiewics, Kahren Tevosyan, Derek Vincent
 Contents 3

Contents
Contents.........................................................................................................................3
Introduction....................................................................................................................3
Using the Microsoft Operations Framework for Active Directory Operations.......3
Audience..................................................................................................................3
Using this Guide......................................................................................................3
Overview of Active Directory Operations.....................................................................3
Planning for Active Directory Operations...............................................................3
Tools Used for Active Directory Operations...........................................................3
Operations Tasks Checklist.....................................................................................3
Monitoring Active Directory.........................................................................................3
Active Directory Backup and Restore...........................................................................3
Backing Up Active Directory and Associated Components.............................3
Performing a Non-Authoritative Restore..........................................................3
Performing an Authoritative Restore of a Subtree or Leaf Object....................3
Performing an Authoritative Restore of Entire Directory.................................3
Recovering a Domain Controller Through Reinstallation................................3
Restoring a Domain Controller Through Reinstallation and Subsequent
Restore from Backup.........................................................................................3
Managing Domain Controllers.......................................................................................3
Installing and Removing Active Directory..............................................................3
Preparing for Active Directory Installation.......................................................3
Installing Active Directory................................................................................3
Performing Active Directory Post-Installation Tasks.......................................3
Decommissioning a Domain Controller............................................................3
Renaming Domain Controllers................................................................................3
Identifying the Current Configuration of a Domain Controller........................3
Renaming a Domain Controller........................................................................3
Restoring the Original Configuration of a Domain Controller.........................3
Managing Global Catalog Servers...........................................................................3
Identifying Global Catalog Servers in a Site.....................................................3
Identifying a Site That Has No Global Catalog Servers...................................3
4 Contents

Adding the Global Catalog to a Domain Controller and Verifying

Readiness...........................................................................................................3
Removing the Global Catalog from a Domain Controller................................3
Managing Operations Masters.................................................................................3
Designating Operations Master Roles...............................................................3
Reducing the Workload on the PDC Emulator.................................................3
Decommissioning a Role Holder......................................................................3
Seizing Operations Master Roles......................................................................3
Choosing a Standby Operations Master............................................................3
Managing the Database............................................................................................3
Relocating Directory Database Files.................................................................3
Returning Unused Disk Space from the Directory Database to the
File System........................................................................................................3
Speeding Removal of an Expired-Tombstone Backlog....................................3
Managing SYSVOL.................................................................................................3
Changing the Space Allocated to the Staging Area..........................................3
Relocating the Staging Area..............................................................................3
Moving SYSVOL by Using the Active Directory Installation Wizard............3
Moving SYSVOL Manually.............................................................................3
Updating the System Volume Path...................................................................3
Restoring and Rebuilding SYSVOL.................................................................3
Managing Windows Time Service..........................................................................3
Configuring a Time Source for the Forest........................................................3
Configuring a Reliable Time Source on a Computer Other than the
PDC Emulator...................................................................................................3
Configuring a Client to Request Time from a Specific Time Source...............3
Optimizing the Polling Interval.........................................................................3
Disabling the Windows Time Service...............................................................3
Managing Long-Disconnected Domain Controllers................................................3
Preparing a Domain Controller for a Long Disconnection...............................3
Reconnecting Long-Disconnected Domain Controllers...................................3
Removing Lingering Objects from an Outdated Writable Domain
Controller..........................................................................................................3
Removing Lingering Objects from a Global Catalog Server............................3
Managing Trusts............................................................................................................3
Creating External Trusts....................................................................................3
Creating Shortcut Trusts....................................................................................3
 Contents 5

Removing Manually Created Trusts.................................................................3

Preventing Unauthorized Privilege Escalation..................................................3
Managing Sites...............................................................................................................3
Adding a New Site............................................................................................3
Adding a Subnet to the Network.......................................................................3
Linking Sites for Replication............................................................................3
Changing Site Link Properties..........................................................................3
Moving a Domain Controller to a Different Site..............................................3
Removing a Site................................................................................................3
6 Contents

Introduction
Microsoft® Windows® 2000 Active Directory provides a robust directory service environment
that requires few regularly scheduled maintenance tasks. However, you might perform some
tasks on a regular basis, including backing up the database, and adding or removing domain
controllers. You can use this guide to help you efficiently operate your Active Directory
environment.
Although this guide specifically addresses the operating phase of the IT life cycle, Microsoft
Enterprise Services Framework provides guidelines for all four phases of the life cycle. These
four phases are listed in Table 1.
Table 1 IT Life Cycle and Microsoft Enterprise Services Frameworks Assistance
For this Phase… Microsoft Enterprise Services Frameworks Provides this Assistance…
Planning Although not currently a dedicated Enterprise Services framework,
Microsoft Business Value Services provide tools to assess and plan the IT
infrastructure, prioritize projects, and make a compelling business case for
undertaking an IT project.
Preparing Microsoft Readiness Framework helps IT organizations develop individual
and organizational readiness to use Microsoft products and technologies.
Building and Microsoft Solutions Framework provides guidelines for building and
Deploying deploying a project. The phases involved in this part of the IT lifecycle
include Envisioning, Planning, Developing, and Deploying.
Operating Microsoft Operations Framework provides guidelines for managing
production systems within complex distributed IT environments.

Active Directory operations occur after you plan, prepare, and deploy your Active Directory
implementation.
 Contents 7

Note
All references to Windows 2000 include both Microsoft® Windows® 2000 Server and
Microsoft® Windows® 2000 Advanced Server, unless otherwise specified. This document
assumes that you are using Windows 2000 with Service Pack 2 (SP2) or greater.

Using the Microsoft Operations Framework

for Active Directory Operations
Microsoft Operations Framework (MOF) is a collection of best practices, principles, and models.
It provides comprehensive technical guidance for achieving reliable, available, supportable, and
manageable solutions and services built on Microsoft products and technologies. MOF bases its
recommendations on current industry best practices for IT service management, as documented
and validated by the IT Infrastructure Library (ITIL) of the Central Computer and
Telecommunications Agency (CCTA).
The MOF process model describes an operations life cycle that applies to releases of any size,
relating to any service solution. MOF identifies four main areas of operations, which are divided
into quadrants in the operations life cycle. Table 2 lists the four quadrants and the area of
operations they cover.
Table 2 MOF Operations Quadrants
Quadrant Service Mission
Operating Perform day-to-day tasks effectively and efficiently.
Supporting Resolve incidents, problems, and inquiries quickly.
Optimizing Optimize cost, performance, capacity, and availability in the delivery of IT
services and drive necessary changes, based on the data that you collect.
Changing Introduce new service solutions, technologies, systems, applications, hardware,
and processes.

This guide includes processes for operating Active Directory.

8 Contents

For more information about MOF, see the MOF link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.

Audience
This guide is for medium and large organizations that have one or more centralized IT operations
departments. It includes information that is relevant to different roles within an IT organization,
including IT Operations management and administrators. It contains high-level information that
is required in planning an Active Directory operations environment. This information requires
management-level knowledge of the technology and IT processes.
In addition, this guide contains low-level procedures that are designed for operators who have
varied levels of expertise and experience. Although the procedures provide operator guidance
from start to finish, operators must have a basic proficiency with the Microsoft Management
Console (MMC) and snap-ins, and know how to start programs and access the command line.

Using this Guide

To accommodate a wide IT audience, the operations areas are divided into the following types of
content:
 Overview, which explains what you need to consider for operating an Active
Directory component, along with a list of tasks involved in operating that component.
 Tasks, which contain the caveats that you should be aware of when performing the
task, along with a list of procedures involved in the task. For your convenience, a list
of tasks and procedures appears in alphabetical order in Appendix A.
 Procedures, which appear in full in Appendix B of this document, and are often
referred to by more than one task. All tasks in this document link to the associated
procedures.
For maximum benefit in using this guide:
 Read through the entire Operating Active Directory chapter to gain a management-
level knowledge of how to operate Active Directory.
 Ensure that you have all the tools installed where operators use them.
 Use the task lists to schedule recurring tasks.
 Create “tear sheets” for each task that operators perform within your organization.
Cut and paste the task and its related procedures into a separate document and then
either print these documents, or store them online, depending on the preference of
your organization.
 Give the operator the tear sheets for the task when a task needs to be performed,
along with information relevant to the environment (such as the name and IP address
of the domain controller involved in the task).
 Contents 9

This guide is your tool. Use it in a way that best meets the needs of your particular IT
department.

Overview of Active Directory

Operations
The goal of operations is to ensure that IT services are delivered according to service level
requirements that are agreed to by IT management and its various customer business units. The
day-to-day operations of an IT department are proactive, and require that the proper products and
services be in place to identify and prevent potential problems.

Planning for Active Directory Operations

To plan your Active Directory operations environment, you need to perform the following tasks:
 Assess the IT environment and establish a baseline.
 Determine operational needs.
 Define operations actions.

Assessing the IT Environment and Establishing a Baseline

You must have a complete and accurate idea of the details behind each service that the IT
department delivers in order to properly configure management systems and technologies, and to
collect any necessary metric data.
Review any service specifications that were produced during the deployment process, along with
any service level requirements defined in Service Level Agreements between the IT organization
and customer business units.
The following information is especially useful when planning your operations:
 Server specifications
 Network specifications
 Logical and physical architectural diagrams
 Supported applications
 User statistics and requirements
 Current thresholds and performance metrics
 Acceptable performance and outage times
This data provides a starting point to establish a baseline for the operations environment, and to
set the proper level of service.
10 Contents

Determining Operational Needs

The Active Directory operations team must establish processes for the following tasks:
 Continuous monitoring and reporting
 Auditing
 Backup and restoration
 Managing Active Directory components, including:
 Domain controllers (including issues relating to installation, global catalog servers,
operations masters, database, SYSVOL, Windows Time Service, and long-disconnected
domain controllers)
 Trusts
 Sites

Defining Operations Actions

Categorize actions that are performed during the course of day-to-day operations as follows:
 Automated actions
 Operator-driven actions
Automated Actions
Automated actions provide a time-saving method to detect and react to incidents occurring in the
production environment. Identify those tasks and procedures that you want to automate, whether
with scripts or a monitoring product such as Microsoft Operations Manager 2000 (MOM). Also
identify the triggers, such as alerts generated by MOM, which start the automated action.
An example of an automated action is configuring an agent process to respond when it detects
that the threshold for disk space has been exceeded. In this case, the agent process running on the
affected computer automatically takes action to resolve the situation, such as deleting all the files
in the Temp directory, thereby returning the system to acceptable conditions as defined in the
Service Level Agreement. The agent system also sends a message to the management server that
includes any necessary event data (the name and address of the affected system, the error
message, the results of the action taken, and so on). After the automated action resolves the
incident, the operations team can determine what, if any, further action to take. In this example,
the automated action temporarily resolves the incident, and the operations team must investigate
further to determine a permanent resolution.
Operator-Driven Actions
Operator-driven actions are those that are performed by an operator, as opposed to those
performed by an automated system. Operator-driven actions need to be defined whenever and
wherever possible, so that operators with varying degrees of skills and training can perform
specific tasks, such as changing a password, loading forms into a printer, starting or stopping
processes, and so on.
 Contents 11

Tools Used for Active Directory Operations

Active Directory operations involves using tools that are either part of the Windows 2000
operating system, the Windows 2000 Support Tools, or the Microsoft® Windows® 2000 Server
Resource Kit. Table 3 lists the tools that are used to operate Active Directory, where the tools are
found, and a brief description of the purpose of the tool.
For information about installing the Windows 2000 Support Tools and the Windows 2000
Administrative Tools Pack, see Windows 2000 Server Help.
Table 3 Tools Used in Active Directory Operations
Tool Location Function
Active Directory http://www.microsoft.com/ Migrate account and resource
Migration Tool windows2000/downloads/ domains.
(ADMT) tools/ADMT/default.asp
Active Directory Windows 2000 Administrative Administer domain trusts, add
Domains and Trusts Tools Pack user principal name suffixes, and
snap-in change the domain mode.
Active Directory Windows 2000 Install Active Directory, and
Installation Wizard promote or demote domain
controllers.
Active Directory Sites Windows 2000 Administrative Administer the replication of
and Services snap-in Tools Pack directory data.
Active Directory Users Windows 2000 Administrative Administer and publish
and Computers snap-in Tools Pack information in the directory.
ADSI Edit, MMC Windows 2000 Support Tools View, modify, and set access
snap-in control lists on objects in the
directory.
Backup Wizard Windows 2000 system tool Back up and restore data.
Control Panel Windows 2000 View and modify computer,
application, and network settings.
Dcdiag.exe Windows 2000 Support Tools Analyze the state of domain
and Windows 2000 Server controllers in a forest or
Resource Kit enterprise; assist in
troubleshooting by reporting any
problems.
DNS snap-in Windows 2000 Administrative Manage DNS.
Tools Pack
Dsastat.exe Windows 2000 Support Tools Compare directory information
on domain controllers and
detectsdifferences.
12 Contents

Event viewer Windows 2000 Administrative Monitor events recorded in event

Tools Pack logs.
Lbridge.cmd Windows 2000 Server Resource Replicate logon scripts and
Kit profiles between Windows 2000–
based domain controllers and
Windows NT 4.0–based domain
controllers.
Ldp.exe Windows 2000 Support Tools Perform LDAP operations against
Active Directory.
Linkd.exe Windows 2000 Server Resource Create, delete, update, and view
Kit the links that are stored in
junction points.
MMC Windows 2000 Create, save, and open
administrative tools (called MMC
snap-ins) that manage hardware,
software, and network
components.
Netdiag.exe Windows 2000 Server Resource Check end-to-end network
Kit and Windows 2000 Support connectivity and distributed
Tools services functions.
Netdom.exe Windows 2000 Support Tools Allow batch management of
trusts, joining computers to
domains, and verifying trusts and
secure channels.
Net use, start, stop, Windows 2000 system tool Perform common tasks on
del, copy, time network services, including
stopping, starting, and connecting
to network resources.
Nltest.exe Windows 2000 Support Tools Verify that the locator and secure
channel are functioning.
Notepad Windows 2000 Accessories View, create, and modify text
files.
Ntdsutil.exe Windows 2000 system tool Manage Active Directory,
manage single master operations,
remove metadata, create
application directory partitions.
Regedit.exe Windows 2000 system tool View and modify registry
settings.
Repadmin.exe Windows 2000 Support Tools Verify replication consistency
between replication partners,
monitor replication status, display
replication metadata, and force
 Contents 13

replication events and topology

recalculation.
Replmon.exe Windows 2000 Support Tools Display replication topology,
monitor replication status, and
force replication events and
topology recalculation.
Services snap-in Windows 2000 Administrative Start, stop, pause, or resume
Tools Pack system services on remote and
local computers, and configures
startup and recovery options for
each service.
Terminal Services Windows 2000 Access and manage computers
remotely.
W32tm Windows 2000 system tool Manage Windows Time Service.
Windows Explorer Windows 2000 Access files, Web pages, and
network locations.

Operations Tasks Checklist

Table 4 provides a quick reference for those product maintenance tasks that the operations team
must perform on a regular basis. These task lists summarize the tasks that are required to
maintain Active Directory operations.
Table 4 Active Directory Operations Tasks
Frequency Tasks
Daily. Verify that all domain controllers are communicating with the central
monitoring console or collector.
Daily. View and examine all new alerts on each domain controller, resolving
them in a timely fashion.
Daily. Resolve alerts indicating the following services are not running: FRS,
Net Logon, KDC, W32Time, ISMSERV. MOM reports these as
Active Directory Essential Services.
Daily. Resolve alerts indicating SYSVOL is not shared.
Daily. Resolve alerts indicating that the domain controller is not advertising
itself.
Daily. Resolve alerts indicating time synchronization problems.
Daily. Resolve all other alerts in order of severity. If alerts are given error,
warning, and information status similar to the event log, resolve alerts
marked error first.
14 Contents

Daily to weekly, Identify a site that has no global catalog server.

depending on
environment.
Weekly. Review the Time Synchronization Report to detect intermittent
problems and resolve time-related alerts.
Weekly. Review the Authentication Report to help resolve problems generated
by computer accounts with expired passwords.
Weekly. Review the Duplicate Service Principal Name Report to list all
security principals that have a service principal name conflict.
Weekly. Review a report of the top alerts generated by the Active Directory
monitoring indicators and resolve those items that occur most
frequently.
Weekly. Review the report that lists all trust relationships in the forest and
check for obsolete, unintended, or broken trusts.
Monthly. Verify that all domain controllers are running with the same service
pack and hot fix patches.
Monthly. Review all Active Directory reports and adjust thresholds as needed.
Examine each report and determine which reports, data, and alerts are
important for your environment and service level agreement.
Monthly. Review the Replication Monitoring Report to verify that replication
throughout the forest occurs within acceptable limits
Monthly. Review the Active Directory response time reports.
Monthly. Review the domain controller disk space reports.
Monthly. Review all performance related reports. These reports are called
Health Monitoring reports in MOM.
Monthly. Review all performance related reports for capacity planning
purposes to ensure that you have enough capacity for current and
expected growth. These reports are called Health Monitoring reports
in MOM.
Monthly. Adjust performance counter thresholds or disable rules that are not
applicable to your environment or that generate irrelevant alerts.
Monthly. Identify the global catalog servers in a site.
At least twice within Back up Active Directory and associated components.
the tombstone
lifetime.
As needed. Perform a non-authoritative restore.
As needed. Perform an authoritative restore of a subtree or leaf object.
As needed. Perform an authoritative restore of the entire directory.
 Contents 15

As needed. Recover a domain controller through reinstallation.

As needed. Restore a domain controller through reinstallation and subsequent
restore from backup.
As needed. Prepare for Active Directory Installation.
As needed. Install Active Directory.
As needed. Perform Active Directory post-installation tasks.
As needed. Decommission a domain controller.
As needed. Identify the current configuration of a domain controller.
As needed. Rename a domain controller.
As needed. Restore the original configuration of a domain controller.
As needed. Add the global catalog to a domain controller and verify global
catalog readiness.
As needed. Remove the global catalog from a domain controller.
As needed. Designate operations master roles.
As needed. Reduce the workload on a PDC emulator.
As needed. Decommission an operations master role holder.
As needed. Seize operations master roles.
As needed. Choose a standby operations master.
As needed. Relocate directory database files.
As needed. Return unused disk space from the directory database to the file
system.
As needed. Speed removal of an expired-tombstone backlog.
As needed. Change the space allocated to the Staging Area folder.
As needed. Relocate the Staging Area folder.
As needed. Move SYSVOL by using the Active Directory Installation Wizard.
As needed. Move SYSVOL manually.
As needed. Update the SYSVOL path.
As needed. Restore and rebuild SYSVOL.
As needed. Configure a time source for the forest.
As needed. Configure a reliable time source on a computer other than the PDC
emulator.
As needed. Configure a client to request time from a specific time source.
As needed. Optimize the polling interval.
16 Contents

As needed. Disable the Windows Time Service.

As needed. Prepare a domain controller for long disconnection.
As needed. Reconnect a long-disconnected domain controller.
As needed. Remove lingering objects from an outdated writable domain
controller.
As needed. Remove lingering objects from a global catalog server.
As needed. Create an external trust (between a Windows 2000 domain and a
Windows NT 4.0 domain, or between domains in different forests).
As needed. Create a shortcut trust.
As needed. Remove a manually created trust.
As needed. Prevent unauthorized privilege escalation.
As needed. Add a new site.
As needed. Add a subnet to the network.
As needed. Link sites for replication.
As needed. Change site link properties.
As needed. Move a domain controller to a different site.
As needed. Remove a site.

Monitoring Active Directory

Monitoring the distributed Active Directory service and the services that it relies upon helps
maintain consistent directory data and the needed level of service throughout the forest. You can
monitor important indicators to discover and resolve minor problems before they develop into
potentially lengthy service outages. Most large organizations with many domains or remote
physical sites require an automated monitoring system such as Microsoft Operations
Manager 2000 (MOM) to monitor important indicators. An automated monitoring system
provides the necessary consolidation and timely problem resolution to administer Active
Directory successfully.

Benefits for End-Users

Monitoring Active Directory helps resolve issues in a timely manner, and users experience the
following benefits:
 Improved reliability of productivity applications that rely on back-end servers, such
as e-mail.
 Quicker logon time and more reliable resource usage.
 Contents 17

 Decreased help desk support issues.

Benefits for Administrators

Monitoring Active Directory provides administrators with a centralized view of Active Directory
across the entire forest. By monitoring important indicators, administrators can realize the
following benefits:
 Higher customer satisfaction, because issues can be resolved before users notice
problems.
 Increased service levels, due to improved reliability and system understanding.
 Greater schedule flexibility and ability to prioritize workload, due to early
notification of problems, allowing resolution of issues while they are still a lower
priority.
 Increased ability for the system to cope with periodic service outages.
Monitoring Active Directory also assures administrators that:
 All necessary services that support Active Directory are running on each domain
controller.
 Data is consistent across all domain controllers and end-to-end replication completes
in accordance with your service level agreements.
 Lightweight Directory Access Protocol (LDAP) queries respond quickly.
 Domain controllers do not experience high CPU usage.
 The central monitoring console collects all events that can adversely affect Active
Directory.

Risks of not Monitoring Active Directory

Systematic monitoring is necessary to ensure consistent service delivery in a large environment
with many domain controllers, domains, or physical sites. As a distributed service, Active
Directory relies upon many interdependent services distributed across many devices and in many
remote locations. As you increase the size of your network to take advantage of the scalability of
Active Directory, monitoring becomes more important. It helps you avoid potentially serious
problems, including:
 Logon failure. Logon failure can occur throughout the domain or forest if a trust
relationship or name resolution fails, or if a global catalog server cannot determine
universal group membership.
 Account lockout. User and service accounts can become locked out if the PDC
emulator is unavailable in the domain or replication fails between several domain
controllers.
 Domain Controller failure. If the drive containing the Ntds.dit file runs out of disk
space, the domain controller stops functioning.
18 Contents

 Application failure. Applications that are critical to your business, such as Microsoft
Exchange or another e-mail application, can fail if address book queries into the
directory fail.
 Inconsistent directory data. If replication fails for an extended period of time,
objects (known as lingering objects and re-animated objects) can be created in the
directory and might require extensive diagnosis and time to eliminate.
 Account creation failure. A domain controller is unable to create user or computer
accounts if it exhausts its supply of relative IDs and the RID master is unavailable.
 Security policy failure. If the SYSVOL shared folder does not replicate properly,
Group Policy objects and security policies are not properly applied to clients.

Levels of Monitoring
Use a cost-benefit analysis to determine the degree or level of monitoring that you need for your
environment. Compare the cost of formalizing a monitoring solution with the costs associated
with service outages and the time that is required to diagnose and resolve problems that might
occur. The level of monitoring also depends on the size of your organization and your service
level needs.
Organizations with few domains and domain controllers, or that do not provide a critical level of
service, might only need to periodically check the health of a single domain controller by using
the built-in tools provided in Windows 2000 Server.
Larger organizations that have many domains, domain controllers, sites, or that provide a critical
service and cannot afford the cost of lost productivity due to a service outage, need to use an
enterprise-level monitoring solution such as MOM.
Enterprise-level monitoring solutions use agents or local services to collect the monitoring data
and consolidate the results on a central console. Enterprise-level monitoring solutions also take
advantage of the physical network topology to reduce network traffic and increase performance.
In a complex environment, directory administrators need enterprise-level monitoring to derive
meaningful data and to make good decisions and analysis. For more information about MOM,
see http://www.microsoft.com/mom/.

Active Directory Monitoring During the Deployment Phase

As a best practice, deploy monitoring with the first domain controller. By integrating monitoring
into the design and deployment process, you can avoid many of the problems that arise during
deployment. Because monitoring solutions require network connectivity between the monitored
servers and the management consoles, you must account for particular TCP/IP ports and
bandwidth usage.
As with any sophisticated service, implement a monitoring solution such as MOM in a lab before
you deploy it in a production environment.

Service-Level Baseline
A baseline represents service level needs as performance data. By setting thresholds to indicate
when the baseline boundaries are exceeded, your monitoring solution can generate alerts to
 Contents 19

inform the administrator of degraded performance and jeopardized service levels. For example,
you can use performance indicators to set a baseline and monitor for low disk space on the disk
drives that contain the Active Directory database and log files, and you can monitor CPU usage
of a domain controller. You can also monitor critical services running on a domain controller.
Monitoring these indicators allows the administrator to ensure adequate performance.
To determine an accurate baseline, monitor and collect data for a time period that is long enough
to represent peak and low usage. For example, monitor during the time in the morning when the
greatest number of users log on. Monitor for an interval that is long enough to span your
password change policy and any month-end or other periodic processing that you perform. Also,
collect data when network demands are low to determine this minimal level. Be sure to collect
data when your environment is functioning properly. To accurately assess what is acceptable for
your environment, remove data caused by network outages or other failures when you establish
your baseline.
The baseline that you establish for your environment can change over time as you add new
applications, users, hardware, and domain infrastructure to the environment, and as the
expectations of users change. Over time, the directory administrator might look for trends and
changes that occur, and take actions designed to meet the increased demands on the system and
maintain the desired level of service. Such actions might include fine-tuning the software
configuration and adding new hardware.
Determining the thresholds when alerts are generated to notify the administrator that the baseline
has been exceeded is a delicate balance between providing either too much information or not
enough. The vendor of your monitoring solution, such as MOM, can provide general
performance thresholds, but you must periodically adjust these thresholds to meet your service
level requirements. To adjust these thresholds, first collect and analyze the monitoring data to
determine what is acceptable or usual activity for your environment. After you gather a good data
sample and consider your service level needs, you can set meaningful thresholds that trigger
alerts.
To determine thresholds:
 For each performance indicator, collect monitoring data and determine the minimum,
maximum and average values.
 Analyze the data with respect to your service level needs.
 Adjust thresholds to trigger alerts when indicators cross the parameters for acceptable
service levels.
As you become more familiar with the monitoring solution you choose, it becomes easier to
correlate the thresholds that trigger the alerts to your service level delivery. If you are uncertain,
it is usually better to set the thresholds low to view a greater number of alerts. As you understand
the alerts you receive and determine why you receive them, you can increase the threshold at
which alerts are generated, thereby reducing the amount of information that you receive from
your monitoring solution. MOM uses thresholds that are a reasonable starting point and work for
the majority of medium-sized customers. Larger organizations might need to increase the
thresholds.
20 Contents

Requirements for Monitoring

Managing an enterprise-level directory requires monitoring many important indicators. Failure to
monitor all of the important indicators can create gaps in coverage. Use any monitoring solution
that best suits your needs, but monitor the necessary important indicators to ensure that all
aspects of Active Directory are functioning properly. MOM monitors all of the important
indicators.
For more information about monitoring Active Directory see http://www.microsoft.com/ad.
For more information about MOM, see http://www.microsoft.com/mom/.
For more information about installing MOM, see
http://www.microsoft.com/mom/docs/DeployGuide.doc.

Relationship between Monitoring and Troubleshooting

The goal of a comprehensive monitoring solution is to monitor all of the important indicators and
provide alerts that are concise, highly relevant, and lead an operator to resolve the problem.
Ideally, the monitoring solution alerts the operator only when a problem requires action. In this
case, monitoring alerts are the first indicator that a problem exists. If the operator cannot easily
resolve the problem that generated an alert, you might want to create a help desk ticket to begin
troubleshooting and root-cause analysis. Your monitoring solution can initiate your
troubleshooting processes or flowcharts.
Monitoring helps ensure that the Active Directory service is available for service requests. Active
Directory is designed to be fault tolerant and can continue to operate if individual servers are
unavailable for periodic maintenance or while operators troubleshoot them. You can assure a
high-degree of reliability by monitoring the distributed services that make up Active Directory,
and resolving issues as they develop.
In addition to providing increased service availability, the relationship between monitoring and
troubleshooting increases your understanding of the root causes of most problems that arise. As
your environment becomes more reliable, monitoring alerts more precisely indicate the cause of
new problems that arise.

Reports
Many important problems do not cause alerts, but they still require periodic attention. Your
monitoring solution might generate reports that display data over time and present patterns that
indicate problems. Review the reports to resolve issues before they generate alerts.

Frequency of Monitoring Tasks

You can perform the daily, weekly, and monthly tasks as specified in the following tables, but
you must adjust the frequency to meet the needs of your particular environment and monitoring
solution.

Daily Monitoring Tasks

Table 5 Daily Tasks and Their Importance
Tasks Importance
 Contents 21

Verify that all domain controllers are Communication failure between the domain
communicating with the central monitoring controller and the monitoring infrastructure
console or collector. prevents you from receiving alerts so you can
examine and resolve them.
View and examine all new alerts on each This precaution helps you avoid service
domain controller, resolving them in a timely outages.
fashion.
Resolve alerts indicating the following Active Directory depends on these services.
services are not running: FRS, Net Logon, They must be running on every domain
KDC, W32Time, ISMSERV. MOM reports controller.
these as Active Directory Essential Services.
Resolve alerts indicating SYSVOL is not Active Directory cannot apply Group Policy
shared. unless SYSVOL is shared.
Resolve alerts indicating that the domain Domain controllers must register DNS
controller is not advertising itself. records to be able to respond to LDAP and
other service requests.
Resolve alerts indicating time synchronization The Kerberos authentication protocol requires
problems. that time be synchronized between all domain
controllers and clients that use it.
Resolve all other alerts in order of severity. If The highest priority alerts indicate the most
alerts are given error, warning, and serious risk to your service level..
information status similar to the event log,
resolve alerts marked error first.

Weekly Monitoring Tasks

Table 6 Weekly Tasks and Their Importance
Tasks Importance
Review the Time Synchronization Report to The Kerberos authentication protocol requires
detect intermittent problems and resolve time- that time be synchronized between all domain
related alerts. controllers and clients that use it.
Review the Authentication Report to help Expired passwords must be reset to allow the
resolve problems generated by computer computers to authenticate and participate in
accounts with expired passwords. the domain.
Review the Duplicate Service Principal Name User or computer accounts cannot be
Report to list all security principals that have authenticated or log on if they share an SPN
a service principal name conflict. with another account.
Review a report of the top alerts generated by Report shows alerts that occur most often.
the Active Directory monitoring indicators Focusing on the top alert generators
and resolve those items that occur most significantly reduces the number of alerts seen
frequently. by the operator.
Review the report that lists all trust Authentication between domains or forests
22 Contents

relationships in the forest and check for requires trust relationships.

obsolete, unintended, or broken trusts.

Monthly Monitoring Tasks

Table 7 Monthly Tasks and Their Importance
Tasks Importance
Verify that all domain controllers are running Potential issues can arise if distributed
with the same service pack and hot fix services are running with different versions of
patches. software.
Review all Active Directory reports and Examining the data that is relevant to your
adjust thresholds as needed. Examine each environment allows you to determine the
report and determine which reports, data, and thresholds that trigger the alerts to your
alerts are important for your environment and service level delivery.
service level agreement.
Review the Replication Monitoring Report to Timely replication helps assure that you meet
verify that replication throughout the forest your service level agreements.
occurs within acceptable limits
Review the Active Directory response time Services must respond quickly for the system
reports. to function properly and applications such as
e-mail to work properly.
Review the domain controller disk space The drives containing the Active Directory
reports. database and log files must have sufficient
free space to accommodate growth and
routine processing.
Review all performance-related reports. These These reports can help you determine the
reports are called Health Monitoring reports baseline for your environment and adjust
in MOM. thresholds.
Review all performance-related reports for These reports help you track growth trends in
capacity planning purposes to ensure that you your environment and plan for future
have enough capacity for current and hardware and software needs.
expected growth. These reports are called
Health Monitoring reports in MOM.
Adjust performance counter thresholds or Monitoring indicators must be adjusted to suit
disable rules that are not applicable to your your environment. The goal is to provide
environment or that generate irrelevant alerts. alerts that are concise, highly relevant, and
lead an operator to resolve the problem.
 Contents 23

Active Directory Backup and

Restore
Active Directory is backed up as part of system state, a collection of system components that
depend on each other. You must back up and restore system state components together.
Components that comprise the system state on a domain controller include:
 System Start-up Files (boot files). These are the files required for Windows 2000
Server to start.
 System registry.
 Class registration database of Component Services. The Component Object Model
(COM) is a binary standard for writing component software in a distributed systems
environment.
 SYSVOL. The system volume provides a default Active Directory location for files
that must be shared for common access throughout a domain. The SYSVOL folder on
a domain controller contains:
 NETLOGON shared folders. These usually host user logon scripts and Group Policy
objects (GPOs) for non-Windows 2000–based network clients.
 User logon scripts for Windows 2000 Professional–based clients and clients that are
running Windows 95, Windows 98, or Windows NT 4.0.
 Windows 2000 GPOs.
 File system junctions.
 File Replication service (FRS) staging directories and files that are required to be
available and synchronized between domain controllers.
 Active Directory. Active Directory includes:
 Ntds.dit: The Active Directory database.
 Edb.chk: The checkpoint file.
 Edb*.log: The transaction logs, each 10 megabytes (MB) in size.
 Res1.log and Res2.log: Reserved transaction logs.
24 Contents
Which domain controllers to back up
Contents
Age
Note
Best
If youperformance
use Active Directory-integrated
practice states that the
DNS,
Active
thenDirectory’s
the zone datalogsisand
backed
database
up as
part of
files should
the Active
be on Directory
separate disks.
database.
If you
If have
you do configured
not use Active
your domain
Directory-integrated
controllers in this manner
DNS,you
youwill
musthave
explicitly
Active back
Directory
up thecomponents
zone files. However,
spread
if you
out on back
multiple
up the
drives,
system
such
disk
as along
D:\Winnt\NTDS
with the system
for your
state,
logs
zoneanddata
E:\Winnt\
is backed
up as part
NTDS for of
your
thedatabase.
system disk.
You do not need to specify these log and database
locations
If in order
you installed for themClustering
Windows to be backed up; the backup
or Certificate utility
Services onwill
yourautomatically
domain
locate and include
controller, they arethem when you
also backed up back upofsystem
as part systemstate.
state. Details of these
components are not discussed in this guide.

General Guidelines for Backup

The backup tool in Windows 2000 Server supports multiple types of backup: normal, copy,
incremental, differential, and daily. However, because Active Directory is backed up as part of
system state, the only type of backup available for Active Directory is normal. A normal backup
creates a backup of the entire system state while the domain controller is online. In addition, the
backup tool marks each file as a backed up file, which clears the archive attribute of the file.
Considerations for ensuring a good backup
To ensure a successful restore from backup, you must know what defines a good backup.
At a minimum, back up two domain controllers in each domain, one of which should be an
operations master role holder (excluding the relative ID (RID) master, which should not be
restored). Note that backup data from a domain controller can only be used to restore that domain
controller. You cannot use a backup of one domain controller to restore another.
A good backup includes at least the system state and the contents of the system disk. Backing up
the system disk ensures that all the required system files and folders are present so you can
successfully restore the data.

A backup that is older than the tombstone lifetime set in Active Directory is not a good backup.
At a minimum, perform at least two backups within the tombstone lifetime. The default
tombstone lifetime is 60 days. Active Directory incorporates the tombstone lifetime into the
backup and restore process as a means of protecting itself from inconsistent data.
Deleting an object from Active Directory is a two-step process. When an object is deleted in
Active Directory, the object gets converted into a tombstone, which is then replicated to the other
 Contents 25

domain controllers in the environment to inform them of the deletion. Active Directory purges
the tombstone when the tombstone lifetime is reached.
If you restore a domain controller to a state prior to the deletion of an object, and the tombstone
for that object is not replicated to the restored domain controller before the tombstone expires,
the object remains present only on the restored domain controller, resulting in inconsistent data.
Thus, you must restore the domain controller prior to expiration of the tombstone, and allow
inbound replication from a domain controller containing the tombstone to complete prior to
expiration of the tombstone.
Active Directory protects itself from restoring data older than the tombstone lifetime by
disallowing the restore. As a result, the useful life of a backup is equivalent to the tombstone
lifetime setting for the enterprise.

General Guidelines for Restore

You can start the restore process by using either the Windows 2000 Server backup utility or
another supported utility. You can perform either a non-authoritative restore or an authoritative
restore.

How to Select the Appropriate Restore Method

You select the appropriate restore method by considering:
 Circumstances and characteristics of the failure. The two major categories of failure,
from an Active Directory perspective, are Active Directory data corruption and
hardware failure. Active Directory data corruption occurs when the directory contains
corrupt data that has been replicated to all domain controllers or when a large portion
of the Active Directory hierarchy has been changed accidentally (such as deletion of
an OU) and this change has replicated to other domain controllers.
 Roles and functions of the failed server.
Non-authoritative restore of Active Directory
A non-authoritative restore returns the domain controller to its state at the time of backup, then
allows normal replication to overwrite that state with any changes that have occurred after the
backup was taken. After you restore the system state, the domain controller queries its replication
partners. The replication partners replicate any changes to the restored domain controller,
ensuring that the domain controller has an accurate and updated copy of the Active Directory
database.
Non-authoritative restore is the default method for restoring Active Directory, and you will use it
in most situations that result from Active Directory data loss or corruption. To perform a non-
authoritative restore, you must be able to start the domain controller in Directory Services
Restore Mode.
Non-authoritative restore of SYSVOL
When you non-authoritatively restore the SYSVOL, the local copy of SYSVOL on the restored
domain controller is compared with that of its replication partners. After the domain controller
restarts, it contacts its replication partners, compares SYSVOL information, and replicate the any
necessary changes, bringing it up-to-date with the other domain controllers within the domain.
26 Contents

Perform a non-authoritative restore of SYSVOL if at least one other functioning domain

controller exists in the domain. This is the default method for restoring SYSVOL and occurs
automatically if you perform a non-authoritative restore of the Active Directory.
If no other functioning domain controller exists in the domain, then perform a primary restore of
the SYSVOL. A primary restore builds a new File Replication service (FRS) database by loading
the data present under SYSVOL on the local domain controller. This method is the same as a
non-authoritative restore, except that the SYSVOL is marked primary.
Authoritative restore of Active Directory
An authoritative restore is an extension of the non-authoritative restore process. You must
perform the steps of a non-authoritative restore before you can perform an authoritative restore.
The main difference is that an authoritative restore has the ability to increment the version
number of the attributes of all objects in an entire directory, all objects in a subtree, or an
individual object (provided that it is a leaf object) to make it authoritative in the directory.
Restore the smallest unit necessary, for example, do not restore the entire directory in order to
restore a single subtree.
As with a non-authoritative restore, after a domain controller is back online, it will contact its
replication partners to determine any changes since the time of the last backup. However,
because the version number of the object attributes that you want to be authoritative will be
higher than the existing version numbers of the attribute held on replication partners, the object
on the restored domain controller will appear to be more recent and therefore will be replicated
out to the rest of the domain controllers within the environment.
Unlike a non-authoritative restore, an authoritative restore requires the use of a separate tool,
Ntdsutil.exe. No backup utilities — including the Windows 2000 Server system tools — can
perform an authoritative restore.
An authoritative restore will not overwrite new objects that have been created after the backup
was taken. You can authoritatively restore only objects from the configuration and domain-
naming contexts. Authoritative restores of schema-naming contexts are not supported.
Perform an authoritative restore when human error is involved, such as when an administrator
accidentally deletes a number of objects and that change replicates to the other domain
controllers and you cannot easily recreate the objects. To perform an authoritative restore, you
must start the domain controller in Directory Services Restore Mode.
Authoritative restore of SYSVOL
By authoritatively restoring the SYSVOL, you are specifying that the copy of SYSVOL that is
restored from backup is authoritative for the domain. After the necessary configurations have
been made, Active Directory marks the local SYSVOL as authoritative and it is replicated to the
other domain controllers within the domain.
The authoritative restore of SYSVOL does not occur automatically after an authoritative restore
of Active Directory. Additional steps are required.
As with Active Directory authoritative restore, you typically perform an authoritative restore of
SYSVOL when human error is involved and the error has replicated to other domain controllers.
 Contents 27

For example, you might perform an authoritative restore of SYSVOL if an administrator has
accidentally deleted an object that resides in SYSVOL, such as a Group Policy object.
Recover a domain controller through reinstallation
To recover a domain controller through reinstallation, you do not restore the system state from
backup media; instead, you reinstall Windows, install Active Directory, and allow replication
partners to bring the recovered domain controller up to date.
Recovering a domain controller through reinstallation can quickly return the computer to service
if the following conditions exist:
 A domain controller has failed and you cannot restart in Directory Services Restore
mode. If failure was caused by a hardware failure, you have resolved the hardware
problem (for example, by replacing the disk).
 There are other domain controllers in the domain, to serve as replication partners.
 The computer is functioning only as a domain controller (it does not run other server
services such as Exchange), and it does not contain other data that needs to be
recovered from a backup.
Restore a domain controller through reinstallation and restore from backup
This method involves first reinstalling Windows 2000, to enable you to start in Directory
Services Restore Mode. During the Windows 2000 Server setup process, you will obtain more
information about the nature of the failure and you can then determine whether you can reinstall
Windows 2000 Server into the same partition as it was previously installed or whether you will
need to re-partition the drive. After you successfully reinstall Windows 2000, you can start in
Directory Services Restore Mode and perform a normal non-authoritative restore from backup
media.
Restore a domain controller through reinstallation and restore the system state from backup if the
following conditions exist:
 A domain controller has failed and you cannot restart in Directory Services Restore
mode. If failure was caused by a hardware failure, you have resolved the hardware
problem (for example, by replacing the disk).
 You have the following information about the failed domain controller:
 Disk configuration. You need a record of the volumes and sizes of the disks and
partitions. You use this information to recreate the disk configuration in the case of a
complete disk failure. You must recreate all disk configurations prior to restoring
system state. Failure to recreate all disk configurations can cause the restore process to
fail and can prevent you from starting the domain controller following the restore.
 Computer name. You need the computer name to restore a domain controller of the
same name and avoid changing client configuration settings.
 Domain membership. You must know the domain name because even if the computer
name does not change, you might need to re-establish a new computer account.
28 Contents

 Local Administrator password. You must know the local computer’s Administrator
password that was used when the backup was created. Without it, you will not be able
to log on to the computer to establish a domain account for the computer after you
restore it. If you are not part of the domain, you will not be able to log on by using a
domain account, even if you are a domain administrator. The local Administrator
password is also required to restore the system state on a domain controller.
 The domain controller is running other server services such as Exchange, or contains
other data you must restore from a backup.
 You have a good backup, made within the tombstone lifetime.
Considerations for restoring operations masters
To restore an operations master role holder, you must perform one of the following procedures:
 Restore the failed operations master from backup.
 Seize the role to another domain controller within the environment. Seize the
operations master role only if you do not intend to restore the original role holder
from backup. For more information about seizing operations master roles, see
“Managing Operations Masters” in this guide.
Restoring the RID Master can result in Active Directory data corruption, so it is not
recommended.
Restoring the Schema Master can result in orphaned objects, so it is not recommended.
Considerations for recovering global catalog servers
To recover the global catalog server you can either:
 Restore the failed global catalog server from backup.
 Assign a new global catalog to compensate for the loss of the original.
Restoring from backup is the only way that a domain controller that was functioning as a global
catalog at the time of backup can automatically be restored to the role of global catalog.
Restoring a domain controller by reinstallation does not automatically reinstate the global catalog
role. In a multi-domain environment, be aware that restoring a global catalog server from backup
requires more time than restoring a domain controller that does not host the global catalog.
As there are no real disadvantages in configuring multiple global catalogs, you might want to
create a new global catalog in your environment if you anticipate an extended downtime for the
failed global catalog server. Creating a new global catalog server is particularly relevant if users
associated with the original global catalog server can no longer access a global catalog server, or
if the requirement for the global catalog service is significant in your environment, such as when
you are running Exchange 2000.
For more information about creating a new global catalog server, see “Managing Global Catalogs
Servers” in this guide.
 Contents 29
Impact on group membership
Note
Configuring multiple global catalogs servers in a forest increases the availability
of the system, but also increases replication traffic and database size. If you do
restore the failed domain controller and maintain its role as a global catalog
server, you might want to remove any additional global catalogs servers that you
configured during its absence.

Considerations for restoring onto different hardware

It is possible to restore a domain controller onto different hardware. However, you should
consider the following issues:
 Different hardware abstraction layers (HALs). By default, the Hal.dll is not
backed up as part of system state, however the Kernel32.dll is. Therefore, if you try
to restore a backup onto a computer that requires a different HAL (for example, to
support a multiprocessor environment) compatibility issues exist between the new
HAL and the original Kernel32.dll. To overcome this incompatibility, manually copy
the Hal.dll from the original computer and install it on the new computer. The
limitation is that the new computer can use only a single processor.
 Incompatible Boot.ini File. If you backup and restore the boot.ini file, you might
have some incompatibility with your new hardware configuration, resulting in a
failure to start. Before you restore it, ensure that the boot.ini file is correct for your
new hardware environment.
 Different Network or Video Cards. If your new hardware has a different video
adapter or multiple network adapters, then uninstall them before you restore data.
When you restart the computer; the normal Plug and Play functionality makes the
necessary changes.
 Disk Space and Partition Configuration. Partitions on the new computer must
match those on the original computer. Specifically, all the drive mappings must be
the same and the partition size must be at least equal to that on the original computer.
Considerations for authoritative restores
Performing an authoritative restore can affect group membership and passwords for trusts and
computer accounts.
By performing an authoritative restore, you risk possible loss of group membership information.
Because group membership is a multi-valued attribute, and because of how Active Directory
handles links, back links and deletions, an authoritative restore can produce varying results to
group membership. These variations are based on which objects replicate first after an
authoritative restore: the User object or the Group object.
If the un-deletion of the user replicates first, then the group membership information of both the
group (the members it contains) and the user (the groups to which the user belongs) will be
represented correctly.
If the un-deletion of the group replicates first, the replication partners will drop the addition of
the (locally) deleted user from the group membership. The only exception to this is the user’s
primary group, which is always represented correctly both from the user and group reference.
30 Contents
Impact on trusts and computer accounts
You cannot control which object replicates first after you perform an authoritative restore. If your
environment is affected by this situation, the only option is to modify the group membership
attribute of the affected groups on the domain controller where you performed the authoritative
restore.
This issue stems not from the integrity of the restored data, but from the way in which the data is
replicated. By looking at this domain controller, administrators can view the way the directory
should look and take steps to replicate the accurate directory information to the other domain
controllers within the domain.
The best way to do this is to add a fictitious user and then delete that same fictitious user to and
from each group that was involved in the authoritative restore.
A group is involved in the restore if it was either authoritatively restored itself or if it had
members restored who did not have that group defined as their primary group.
By doing this, you force the correct group membership information to be replicated out from the
source domain controller (the domain controller on which you performed the original
authoritative restore) and update the group membership information on its replication partners.
These updated objects reflect the correct memberships and also correct the information
represented in the Member of tab of the restored user objects’ properties.
You must ensure that no additions are made to group membership (for the affected groups and
users) on any of the other domain controllers within the environment.
If you do not adhere to this process, the accurate version of the directory (held on the domain
controller where the restore was performed) can become corrupted by the incorrect membership
information. If the accurate version of the directory becomes corrupted, you must either update
group membership manually or perform another authoritative restore of the objects by using the
verinc option, and perform the process again.
In Windows 2000, trust relationships and computer account passwords are negotiated at a
specified interval (by default 30 days for trust relationships and computer passwords).
When you perform an authoritative restore, you might restore previously used passwords for the
objects in the Active Directory that maintain trust relationships and computer accounts.
In the case of trust relationships, this can impact communication with other domain controllers
from other domains, causing permissions errors when users try to access resources in other
domain. To rectify this, you must remove and recreate NTLM trust relationships to
Windows 2000 or Windows NT 4.0 domains.
In the case of a computer account password, this can impact communications between the
member workstation or server and a domain controller of its domain. This effect might cause
users on Windows NT or Windows 2000 computers to have authentication difficulty due to an
invalid computer account.

Backup and Restore Tasks and Procedures

Table 8 shows the tasks and procedures for backup and restore.
 Contents 31

Table 8 Backup and Restore Tasks and Procedures

Tasks Procedures Tools Frequency
Back up  Back up system state on a  NTBackup.e At least
Active domain controller. xe twice within
Directory and  Back up system state and system the
associated disk on a domain controller. tombstone
components. lifetime
Perform a  Restart the domain controller in  NTBackup.e As needed
non- Directory Services Restore xe
authoritative Mode (locally or remotely).  Ntdsutil.exe
restore.  Restore from backup media.  Event
 Verify Active Directory restore. Viewer
 Repadmin.ex
e
Perform an  Restart in Directory Services  NTBackup.e As needed
authoritative Restore Mode. xe
restore of a  Restore from backup media for  Ntdsutil.exe
subtree or leaf authoritative restore.  Event
object.
 Restore system state to an Viewer
alternate location.  Repadmin.ex
 Perform authoritative restore of e
the subtree or leaf object.
 Restart in normal mode.
 Restore applicable portion of
SYSVOL from alternate
location.
 Verify Active Directory restore.
Perform an  Restart in Directory Services  NTBackup.e As needed
authoritative Restore Mode. xe
restore of the  Restore from backup media for  Ntdsutil.exe
entire authoritative restore.  Event
directory.
 Restore system state to an Viewer
alternate location.  Repadmin.ex
 Restore the database. e
 Restart in normal mode.
 Copy SYSVOL from alternate
location.
 Verify Active Directory restore.
32 Contents

Recover a  Clean up metadata.  Ntdsutil.exe As needed

domain  Install Windows 2000 Server.  Active
controller Directory
 Install Active Directory.
through Sites and
reinstallation. Services
 Active
Directory
Users and
Computers
 Dcpromo.ex
e
Restore a  Install Windows 2000 Server on  NTBackup.e As needed
domain the same drive letter and xe
controller partition as before the failure,
through partitioning the drive if
reinstallation necessary.
and  Restore from backup media
subsequent (non-authoritative restore).
restore from
 Verify Active Directory restore.
backup.

Backing Up Active Directory and Associated Components

To back up Active Directory and associated components on a domain controller, you can back up
only system state or you can back up both system state and the system disk.

Procedures for Backing Up Active Directory and Associated Components

Use one of the following procedures to back up Active Directory and associated components.
Procedures are explained in detail in the linked topics.
1. Back up system state.
2. Back up system state and the system disk.

Performing a Non-Authoritative Restore

Non-authoritative restore is the default method for restoring Active Directory, and you use it in
most situations that result from Active Directory data loss or corruption. You must be able to
start in Directory Services Restore Mode to perform a non-authoritative restore. After you restore
the domain controller from backup media, replication partners use the standard replication
protocols to update both the Active Directory and FRS on the restored domain controller.

Procedures for Performing a Non-Authoritative Restore

Use the following procedures to perform a non-authoritative restore of a domain controller.
Procedures are explained in detail in the linked topics.
 Contents 33

1. Restart the domain controller in Directory Services Restore Mode (locally or

remotely).
2. Restore from backup media.
3. Verify Active Directory restore.

Performing an Authoritative Restore of a Subtree or Leaf

Object
An authoritative restore of a subtree or leaf object restores that subtree or leaf and marks it as
authoritative for the directory. You begin by restoring from backup media, just as in a non-
authoritative restore, but then you perform additional steps to complete an authoritative restore.

Procedures for Authoritative Restore of a Subtree or Leaf Object

Use the following procedures to perform an authoritative restore of an Active Directory subtree
or leaf object. Procedures are explained in detail in the linked topics.
1. Restart the domain controller in Directory Services Restore Mode (locally or
remotely).
2. Restore from backup media for authoritative restore.
3. Restore system state to an alternate location.
4. Perform authoritative restore of the subtree or leaf object.
5. Restore applicable portion of SYSVOL from alternate location if necessary.
6. Verify Active Directory restore.

Performing an Authoritative Restore of Entire Directory

Authoritative restore of the entire directory is a major operation. Perform an authoritative restore
of the entire directory only after consultation with a Microsoft Support professional. Do not
perform an authoritative restore of the entire directory if only one domain controller exists in the
domain.

Procedures for Authoritative Restore of the Entire Directory

Use the following procedures to perform an authoritative restore of the entire Active Directory.
Procedures are explained in detail in the linked topics.
1. Restart the domain controller in Directory Services Restore Mode (locally or
remotely).
2. Restore from backup media.
3. Restore system state to an alternate location.
4. Perform authoritative restore of entire directory.
5. Restore SYSVOL from alternate location.
34 Contents

6. Verify Active Directory restore.

Recovering a Domain Controller Through Reinstallation

Recovering through reinstallation is the same process as creating a new domain controller. It
does not involve restoring from backup media. This method relies on Active Directory
replication to restore a domain controller to a working state, and is only valid if another healthy
domain controller exists in the same domain. This option is normally used on computers that
function only as a domain controller.

Bandwidth Considerations
The primary consideration when recovering a domain controller through replication is
bandwidth. The bandwidth required is directly proportional to the size of the Active Directory
database and the time in which the domain controller is required to be at a functioning state.
Ideally, the existing functional domain controller is located in the same Active Directory site as
the replicating domain controller (new domain controller) in order to reduce network impact and
restore duration.

Procedures for Recovering a Domain Controller Through Reinstallation

Use the following procedures to recover a domain controller. Procedures are explained in detail
in the linked topics.
1. Clean up metadata.
2. Reinstall Windows 2000 Server. (This procedure is not covered in this guide.)
3. Install Active Directory. During the installation process, replication occurs, ensuring
that the domain controller has an accurate and up to date copy of the Active
Directory. For more information about seizing operations master roles, see “Installing
Active Directory” in this guide.

Restoring a Domain Controller Through Reinstallation and

Subsequent Restore from Backup
If you cannot restart a domain controller in Directory Services Restore Mode, you can restore a
domain controller through reinstallation and subsequently restore Active Directory from backup.
This option is normally used on domain controllers that also run other services, such as
Exchange, or have other data you want to recover.

Procedures for Restoring a Domain Controller Through Reinstallation and

Subsequent Restore from Backup
To restore a domain controller through reinstallation and subsequently restore Active Directory
from backup, you must ensure that you install Windows 2000 Server on the same drive letter and
on a partition that is at least as large as the partition used before the failure. You must repartition
the drive if necessary. After you reinstall Windows 2000, perform a non-authoritative restore of
the system state and the system disk. Procedures are explained in detail in the linked topics.
 Contents 35

1. Install Windows 2000 Server on the same drive letter and partition as before the
failure. (This procedure is not covered in this guide.)
2. Restore from backup media.
3. Verify Active Directory restore.

Managing Domain Controllers

While individual domain controllers require little management, your overall operations
environment might require change-related tasks, such as adding or removing domain controllers,
or reintroducing a domain controller that has been offline for more than one replication cycle.
During your day-to-day operations, you might need to do some or all of the following:
 Install and remove Active Directory
 Rename domain controllers
 Manage global catalog servers
 Manage operations masters
 Manage the database
 Manage SYSVOL
 Manage Windows Time Service
 Manage long-disconnected domain controllers

Installing and Removing Active Directory

Only domain controllers can host Active Directory. All servers that are not domain controllers
must access the directory in the same manner as the workstations. They send requests for
information to a domain controller, which processes the request and returns the information back
to them.
Domain controllers store and maintain portions of the directory. They also have services that
allow them to directly store and retrieve information from the directory. These services are
referred to as the Active Directory. When you install Active Directory on a Windows 2000–
based server, it becomes a Windows 2000–based domain controller.
The process of removing Active Directory involves steps similar to those for installation. You
run many of the same tests before you remove the directory as you run before you install the
directory. These tests ensure that the process occurs without any problems. In the event that a
domain controller suffers a hardware failure and you plan to never return it to service, you must
take additional steps to remove it from the directory.
36 Contents

The Active Directory Installation Wizard

You install Active Directory by running the Active Directory Installation Wizard on a
Windows 2000–based server. The wizard simplifies the process by automating as much of the
installation process as possible. During the installation, the wizard asks for the name of the
domain that you want this domain controller to host, and for the location where you want to
install required files. To run the Active Directory Installation Wizard, you must be a member of
the Domain Admins group.

Active Directory Installation Prerequisites

This guide covers the installation of Active Directory in an environment that is configured
according to the best practices described in Best Practice Active Directory Design for Managing
Windows Networks and Best Practice Active Directory Deployment for Managing Windows
Networks. To download these guides, see the Active Directory link on the Web Resources page
at http://www.microsoft.com/windows/reskits/webresources. They describe the process of
planning your forests and domains and provide recommendations for deploying DNS. They also
provide guidelines for estimating the number of domains as well as the number of domain
controllers in each domain.
Before you begin your installation, the following conditions must exist in your environment:
 Your Active Directory forest must already exist. At least two properly functioning
domain controllers must reside in the forest root.
 Your Active Directory Domain must already exist. At least two properly functioning
domain controllers must reside in the domain.
 DNS must be functioning properly.
 You must use Active Directory–integrated DNS zones. You must configure at least
one domain controller as a DNS server.
 Contents 37

Note
Creating or removing a domain or forest is beyond the scope of this guide. This
guide does not cover deploying DNS into an environment that has not
previously hosted a DNS infrastructure.
For information about these options, see the Active Directory link on the
Web Resources page at
http://www.microsoft.com/windows/reskits/webresources and the Microsoft®
Windows® 2000 Server Deployment Planning Guide.

Active Directory Installation Preparation

Properly preparing for the installation of Active Directory decreases the chances of problems
during the installation process and helps you quickly complete the operation. Preparation
includes installing and configuring DNS and gathering information that you need for the
installation.
Configure all domain controllers as DNS servers. Install the DNS server service prior to
installing Active Directory. Follow the recommendations mentioned earlier so that your domain
is already configured, DNS is functioning, and you have Active Directory–integrated DNS zones.
Installing the DNS Server service prior to installing Active Directory allows the DNS Server
service to automatically start using the DNS zones that are stored on the directory after you
complete the Active Directory installation.
The installation wizard asks for specific configuration information, such as the domain
administrator's user name and password, location of the directory database and log files, and the
password needed to us Directory Services Restore Mode, before it begins installing Active
Directory. Have that information ready before you run the Active Directory Installation Wizard.
38 Contents

Note
For better performance, store the log files and the Ntds.dit file on separate hard
disks.

Active Directory Installation

During the installation, the Active Directory Installation Wizard communicates with other
domain controllers to obtain configuration information. This information can come from any
domain controller in the same domain. The Active Directory Installation Wizard also
communicates with the various operations masters so that the new domain controller can
properly join the domain and be added to the directory. For this process to succeed, the wizard
must be able to communicate with the various domain controllers involved. Test these channels
of communication prior to installing Active Directory to help ensure that the process does not
encounter problems during the installation.
After successfully testing the communication paths, the Active Directory Installation Wizard
installs Active Directory on the server to make it a domain controller. During the installation
process, the wizard asks for the information that you gathered during the preparation phase. After
the wizard finishes, it restarts the domain controller and the installation completes during the
restart process.

Active Directory Post-installation Tasks

After you complete the installation of Active Directory, perform some validation tests to ensure
that the domain controller is properly joined to the domain and is functioning as expected. The
areas you must test include:
 Site placement
 DNS configuration
 Network connectivity
 SYSVOL
 Replication
If your tests show that all of these areas are configured and functioning properly, the Active
Directory installation is successful.
 Contents 39

Active Directory Unattended Installation

You can automate the Active Directory installation process by performing an unattended
installation. You can create an answer file to answer the questions that the Active Directory
Installation Wizard asks during the installation. The installation does not require user input and
proceeds quickly.
For more information about unattended installation options, see “Using the Answer File with the
Active Directory Installation Wizard” in the Deployment Planning Guide.

Domain Controller Removal

A domain controller can be removed from a domain in one of two ways: by removing Active
Directory or by a system failure that renders the domain controller inoperable so that you cannot
restore it to service.
Active Directory removal
Similarly to how you can install Active Directory to turn a Windows 2000–based server into a
domain controller, you can remove Active Directory and turn a Windows 2000–based domain
controller back into a server. This process removes most of the references to the domain
controller from the directory. You must manually remove the server object that represents the
domain controller from the computer container after you remove Active Directory. This method
properly removes the domain controller from the directory.
Domain controller failure
A hardware failure on a domain controller can render it inoperable. If the problem is severe
enough, you might never be able to return the domain controller to service. In this case, the other
domain controllers eventually reconfigure themselves so that they can continue to replicate
directory information without the failed domain controller.
When a domain controller is removed from the domain without removing Active Directory, all
the information about that domain controller remains in the directory. You must take additional
steps to remove this information from the directory.

Active Directory Installation and Removal Management Tasks and Procedures

Table 9 shows the tasks and procedures for managing Active Directory installation and removal.
Table 9 Active Directory Installation and Removal Management Tasks and Procedures
Tasks Procedures Tools Frequency
Prepare for Active  Install the DNS Server  Control Panel As needed.
Directory Installation. service.
 Gather installation
information.
40 Contents

Install Active  Verify DNS registration  Dcdiag.exe As needed.

Directory. and functionality. and
 Verify that an IP address Netdiag.exe
maps to a subnet and  Dcpromo.exe
determine the site
association.
 Verify communication
with other domain
controllers.
 Verify the existence of
operations masters.
 Install Active Directory.
Perform Active  Determine whether a  Active As needed.
Directory post- server object has child Directory
installation tasks. objects. Sites and
 Verify the site Services
assignment of a domain  DNS snap-in
controller.  Dcdiag.exe
 Move a domain and
controller to a different Netdiag.exe
site.
 Configure DNS server
recursive name
resolution.
 Perform final DNS
configuration.
 Check the status of the
shared system volume.
 Verify DNS registration
and functionality.
 Verify domain
membership for the new
domain controller.
 Verify communication
with other domain
controllers.
 Verify replication is
functioning.
 Verify the existence of
the operations masters.
 Contents 41

Decommission a  View the current  Active As needed.

domain controller. operations master role Directory
holders. Users and
 Transfer the forest-level Computers
operations master roles.  Active
 Transfer the domain- Directory
level operations master Sites and
roles. Services
 Determine whether a  Dcdiag.exe
domain controller is a and
global catalog server. Netdiag.exe
 Verify DNS registration  Dcpromo.exe
and functionality.
 Verify communication
with other domain
controllers.
 Verify the existence of
the operations masters.
 Remove Active
Directory.
 Determine whether a
server object has child
objects.
 Delete a server object
from a site.

Preparing for Active Directory Installation

Preparation helps the Active Directory installation proceed successfully. To prepare for the
installation process, you must have the appropriate domain information and credentials available
before you start the Active Directory Installation Wizard. It is recommended that you configure
all domain controllers as DNS servers. You must have your DNS server configuration
information available for that portion of the installation process.

DNS Service Installation

Domain controllers use DNS to locate other domain controllers that are hosting Active Directory.
Configure every domain controller as a DNS server to help ensure that a DNS server is always
available. Using Active Directory–integrated DNS zones simplifies the configuration required
because you do not need to create the zone files on each DNS server. Active Directory–
integrated zones are stored in the directory and are replicated to each domain controller along
with other Active Directory data. When you start a domain controller that also runs DNS, the
DNS Server service detects the zones in the directory and uses them.
42 Contents

Before you install DNS server on a domain controller that you want to host Active Directory–
integrated zones, ensure that you already have other domain controllers functioning in the
domain with at least one configured as a DNS server that uses Active Directory–integrated zones.
For more information about DNS configuration and operations master role placement, see Best
Practice Active Directory Design for Managing Windows Networks and Best Practice Active
Directory Deployment for Managing Windows Networks. To download these guides, see the
Active Directory link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.

Active Directory Installation Information

Gather the information that you must supply to the Active Directory Installation Wizard before
you run the wizard.

Procedures for Preparing for Active Directory Installation

To prepare for the Active Directory installation, install the DNS Server service on the server that
you want to make a domain controller and gather the information that you must supply to the
Active Directory Installation Wizard.
1. Install the DNS Server service.
2. Gather installation information, including:
 The user name, password, and the domain that contains the user account that you intend
to use to run the Active Directory Installation Wizard.
 The name of the domain that you want the new domain controller to host.
 Location for the Active Directory database (Ntds.dit).
 Location for the log files.
 Location for the Shared System Volume (SYSVOL).
 The server administrator account name and password to use in Directory Services
Restore mode.

Installing Active Directory

You install Active Directory by using the Active Directory Installation Wizard (DCPromo.exe).
During installation, the wizard contacts other domain controllers for information that it needs to
complete the installation. If the wizard cannot communicate with other domain controllers, the
installation fails. To help ensure successful installation, test the communication channels prior to
running the wizard.
 Contents 43

Site Placement
During installation, the Active Directory Installation Wizard attempts to place the new domain
controller in the appropriate site. The appropriate site is determined by the domain controller's IP
address and subnet mask. The wizard uses the IP information to calculate the subnet address of
the domain controller and checks to see if a subnet object exists in the directory for that subnet
address. If the subnet object exists, the wizard uses it to place the new server object in the
appropriate site. If not, the wizard places the new server object in the same site as the domain
controller that is being used as a source to replicate the directory database to the new domain
controller. Make sure the subnet object has been created for the desired site prior to running the
wizard.

Domain Connectivity
During the installation process, the Active Directory Installation Wizard needs to communicate
with other domain controllers in order to join the new domain controller to the domain. The
wizard needs to communicate with a member of the domain to receive the initial copy of the
directory database for the new domain controller. It needs to communicate with the domain
naming master so that the new domain controller can be added to the domain. The wizard also
needs to contact the RID master so that the new domain controller can receive its RID pool, and
it needs to communicate with another domain controller in order to populate the SYSVOL shared
folder on the new domain controller. All of this communication depends on proper DNS
installation and configuration. By using Netdiag.exe and Dcdiag.exe, you can test all of these
connections prior to starting the Active Directory Installation Wizard.

The Active Directory Installation Wizard

After you have gathered all the information that you need to run the Active Directory Installation
Wizard and performed the tests to verify that all of necessary domain controllers are available,
you are ready to install Active Directory on your server and turn it into a domain controller.
You need to log on with local administrative credentials to start the wizard. Start the wizard and
supply the information you gathered earlier. If the wizard asks for information that you did not
gather, such as if you want to install DNS Server service, it is indicating that it cannot locate the
DNS servers. The wizard assumes that none exist and asks you if you want to install one.
Running the verification tests prior to using the installation wizard helps prevent this kind of
situation from happening.
During the installation process, the wizard asks for information that it needs to properly
configure the new domain controller. First, it asks is if you want to install a domain controller in
a new domain or an additional domain controller in an existing domain. Because this guide
pertains to adding domain controllers to domains that already exist, choose Additional domain
controller in an existing domain.
During the installation process, the wizard needs to communicate with other domain controllers
in order to add this new domain controller to the domain and get the appropriate information into
the Active Directory database. To maintain security, you must provide credentials that have
administrative access to the directory. Once your credentials are validated, the wizard guides you
through the following steps:
44 Contents

Note
If any of the verification tests fail, do not continue until you determine and fix
the problems. If these tests fail, the installation is also likely to fail.

 The wizard asks for a user name, password, and domain name of the account it uses
to add this domain controller to the directory.
 The wizard then asks for the name of the domain that you want this new domain
controller to host. Enter the fully qualified domain name of the appropriate domain.
 Next, the wizard asks where you want to store the Active Directory database and the
database log files. For better performance, store these files on separate hard disks.
 The wizard then asks for the location where you want to store the shared System
Volume (SYSVOL). Ensure that the location has adequate disk space. For more
information about ensuring adequate disk space for SYSVOL, see “Managing
Sysvol” later in this guide.
 The wizard then asks for the password that is assigned to the Directory Services
Restore Mode administrator account. This account is not the domain administrator
account or the local administrator account on the server, but a special account that
can only be used when the domain controller starts in Directory Services Restore
Mode.
 Before installation begins, the wizard displays a dialog box that summarizes the
information that you supplied. Verify that the information is correct before the
installation process begins.

Procedures for Installing Active Directory

1. Verify DNS registration and functionality.
2. Verify that an IP address maps to a subnet and determine the site association .
3. Verify communication with other domain controllers.
4. Verify the existence of the operations masters.
5. Install Active Directory.

Performing Active Directory Post-Installation Tasks

After completing the installation of Active Directory, perform some validation tests to ensure
that the domain controller is properly installed into the domain and is functioning as expected.
Successfully passing these tests is a good indication that the new domain controller is functioning
properly. You might also need to perform additional tasks regarding DNS configuration and
hosting the global catalog.
 Contents 45

Proper Site Placement

You must ensure that the new domain controller is located in the proper site so that after the
installation is complete, the new domain controller can locate replication partners and become
part of the replication topology. During Active Directory installation, the wizard creates a server
object for the new domain controller in the directory and attempts to place the server object in
the proper site. To place the server object, the wizard uses the current IP address and subnet
mask of the new domain controller. If the subnet associated with the domain controller's IP
address is not defined by an existing subnet object, the wizard places the new server object in the
same site as the source domain controller, which is the domain controller from which the new
domain controller downloaded a copy of the directory database. If the site is not correct, you can
use the Active Directory Sites and Services snap-in to move the server object for the domain
controller to the proper site after Active Directory installation is complete.
The last dialog box displayed by the Active Directory Installation Wizard lists the site where the
new domain controller is installed. If this is not the proper site, you need must move the server
object.
For more information about sites or to create a new site object, see "Managing Site Topology"
later in this guide.

Final DNS Configuration

If you installed the DNS server service and made this domain controller a DNS server, you might
need to perform some additional configuration of the DNS installation to ensure that it conforms
to the recommended practices. The configuration that you must perform depends upon whether
this is a new domain controller in the forest root domain or a new domain controller in a child
domain. Performing final DNS configuration helps balance the load among your DNS servers
and provides redundancy in case a DNS server becomes unavailable.
You might need to add a delegation for the new domain controller. If your forest root domain is a
child domain in your corporate DNS domain structure, you must add a delegation for the new
domain controller in the forest root's parent DNS domain. If the forest root domain has no parent
DNS domain, you do not need to add the delegation.
If the new domain controller is located in a child domain of the forest root domain, you must add
a delegation for the new domain controller to the forest root domain.
You also need to configure the DNS client settings on the new domain controller. Configure a
domain controller in the forest root domain to refer to another DNS server located nearby as its
primary DNS server and refer to itself as the secondary DNS server. If the new domain controller
is located in a child domain of the forest root domain, configure the DNS client to use its own IP
address as its primary DNS server address, and another local DNS server as the secondary server
address.
If the new domain controller is located in a child domain below the forest root, create a
secondary zone to make the process of locating domain controllers more reliable.
Whether or not the new domain controller is located in a parent or child domain, you must also
configure the DNS server to use either root hints or forwarders for recursive name resolution.
Follow the established method on your network.
46 Contents

Note
This process can take 15 minutes or longer to complete, depending on the
connection speed between the domain controller and its replication partners.

Domain Connectivity
After the Active Directory Installation Wizard finishes, the domain controller restarts and
performs a few tasks before it is ready to assume its role as a domain controller. It registers itself
with its DNS server so that other members of the domain know that it is a domain controller and
can locate it.
When a new domain controller first joins the network, it receives SYSVOL information from its
replication partners. Until it finishes the initial replication of the SYSVOL, it does not create the
NETLOGON and SYSVOL shared folders and does not start the Net Logon service, both of
which are necessary for it to assume the role of a domain controller. An event number 13516 in
the File Replication Service event log indicates that replication is complete and is working
properly. At this point, the domain controller starts the Net Logon service and the domain
controller becomes available to the domain.
Domain controllers make changes to the directory and replicate these changes among
themselves through a series of connections that are established when the domain controller joins
the network. The connections can be generated automatically or an administrator might manually
create the connections objects. If these connections are not functioning properly, the domain
controller cannot replicate changes to the other domain controllers and cannot receive changes
from other domain controllers.
To function properly, domain controllers must periodically communicate with various operations
masters. The domain controllers send password changes to the PDC emulator. They receive a
RID pool from the RID master. As their pools are depleted, the domain controller periodically
replenishes their allocations by sending requests to the RID master.
All of these features depend upon communication between the new domain controller and other
domain controllers in the domain and forest. When a new domain controller joins the network,
perform tests that verify the communication channels used by these features.

Configure Other Roles

After the domain controller is functioning properly and you complete verification tests and final
DNS configuration, configure any additional roles, such as global catalog server, on the domain
controller. For information about configuring a global catalog server, see “Managing Global
Catalog Servers” later in this guide.

Procedures for Performing Active Directory Post-Installation Tasks

To perform this task, the site object must already be defined in Active Directory Sites and
Services and you must know the site in which you want to place the server object.
1. Determine whether a server object has child objects.
2. Verify the site assignment for the domain controller.
 Contents 47

3. Move a server object to a different site if the domain controller is located in the
wrong site.
4. Configure DNS server recursive name resolution.
5. Perform final DNS configuration for a new domain controller that is located in the
forest root domain:
a. Create a delegation for the new domain controller in the parent domain of the
DNS infrastructure if a parent domain exists and a Microsoft DNS server hosts
it. If a Microsoft DNS server does not host the parent domain, follow the
procedures outlined in the vendor documentation to add the delegation for the
new domain controller.
b. Configure the DNS client settings.
– or –
Perform final DNS configuration for a new domain controller that is located in a child
domain:
c. Create a delegation for the new domain controller in the forest root domain.
d. Create a secondary zone.
e. Configure the DNS client settings.
6. Check the status of the shared system volume.
7. Verify DNS registration and functionality.
8. Verify domain membership for the new domain controller.
9. Verify communication with other domain controllers.
10. Verify replication is functioning.
11. Verify the existence of the operations masters.

Decommissioning a Domain Controller

Just as you can install Active Directory to make a Windows 2000–based server a domain
controller, you can also remove Active Directory and to make a Windows 2000–based domain
controller back into a server.
Removing Active Directory is a similar process to installing it. You use the Active Directory
Installation Wizard and it contacts other domain controllers to copy information from the domain
controller that you want to decommission. As with installation, if the domain controller cannot
contact the other domain controllers during the Active Directory removal, the process is likely to
fail. Perform the same connectivity tests prior to decommissioning a domain controller as you
perform prior to installing Active Directory.
This guide does not include procedures for decommissioning the last domain controller in a
domain. Decommissioning the last domain controller in a domain constitutes the removal of the
48 Contents

domain from the forest. For more information about removing domains, see “Removing Active
Directory” in the Windows 2000 Server Distributed Systems Guide.

Operations Master Role Transfer

During the decommissioning process, the Active Directory Installation Wizard transfers the
operations master roles to other domain controllers without any user interaction. You do not have
control over which domain controller receives the roles. The wizard transfers the roles to any
available domain controller and does not indicate which domain controller hosts them.
Because of this behavior, transfer any operations master roles prior to running the Active
Directory Installation Wizard to decommission a domain controller so you can control operations
master role placement. If you need to transfer any roles from a domain controller, understand all
the recommendations for role placement before performing the transfer. For more information
about transferring operations master roles and role placement, see "Managing Operations Master
Roles" later in this guide.

Global Catalog Removal

If you remove Active Directory from a domain controller that hosts the global catalog, the Active
Directory Installation Wizard confirms that you want to continue with removing Active
Directory. This confirmation ensures that you are aware that you are removing a global catalog
from your environment. Do not remove the last global catalog server from your environment
because users cannot logon without an available global catalog server. If you are not sure, do not
proceed with removing Active Directory until you know at least one other global catalog server
is available. For more information about removing and creating global catalog servers, see
“Managing Global Catalog Servers” later in the guide.

Domain Connectivity
During the removal of Active Directory, the Active Directory Installation Wizard must
communicate with various domain controllers. Any unreplicated changes to the directory must be
replicated to another domain controller. The wizard attempts to connect to another domain
controller to replicate these changes. The wizard must contact another domain controller so that
Active Directory can remove the domain controller from the directory database. If the domain
controller hosts any operations master roles that you chose not to transfer, the wizard must
contact another domain controller in order to transfer the operations master roles.
If the domain controller cannot contact the other domain controllers during Active Directory
removal, the decommissioning operation fails. As with the installation process, test the
communication infrastructure prior to running the installation wizard. When you remove Active
Directory, use the same connectivity tests that you use during Active Directory installation.

Active Directory Removal

After you transfer operations master roles and verify that all the necessary domain controllers are
available, you can use the Active Directory Installation Wizard to remove Active Directory.
When you run the wizard on a server that is already a domain controller, it displays the Remove
Active Directory options.
 Contents 49

Note
If any of the verification tests fail, do not continue until you determine and fix
the problems. If these tests fail, the installation is also likely to fail.

The wizard asks whether or not this is the last domain controller in the domain and requests the
password that is assigned to the local administrator account on the server after Active Directory
is removed. Note that the procedures in this guide do not pertain to removing Active Directory
from the last domain controller in the domain, because that action also deletes the domain from
the forest.

Server Object Removal

After removing Active Directory from a domain controller, the Active Directory Installation
Wizard removes information about that domain controller from the directory. Because it no
longer acts as a domain controller, the server is not part of the replication topology and the
directory does not maintain connections to it. During the decommissioning process, the Active
Directory Installation Wizard removes the server object from the Domain Controller container in
Active Directory Users and Computers and removes the connection objects associated with the
domain controller from the NTDS Settings object in Active Directory Sites and Services.
The Active Directory Installation Wizard does not delete the server object from the site object
during the removal of Active Directory because other services, such as Microsoft Operations
Manager 2000 (MOM), use this container to store their own site-specific information. After you
remove Active Directory, you can use the Active Directory Sites and Services snap-in to safely
remove the server object that represents the decommissioned domain controller in Active
Directory Sites and Services if the server object container is empty.

Procedures for Decommissioning Domain Controllers

1. View the current operations master role holders to see if any roles are assigned to this
domain controller.
2. Transfer the forest-level operations master roles to another domain controller in the
forest root domain if this domain controller hosts either the schema master or domain
naming master roles.
3. Transfer the domain-level operations master roles if this domain controller hosts the
PDC emulator, infrastructure master, or RID master.
4. Determine whether a domain controller is a global catalog server to ensure that other
domain controllers are configured as global catalog servers before you remove Active
Directory.
5. Verify DNS registration and functionality.
6. Verify communication with other domain controllers.
7. Verify the existence of the operations masters.
8. Remove Active Directory.
9. Determine whether a server object has child objects.
50 Contents

10. Delete a server object from a site.

Renaming Domain Controllers

Renaming a domain controller that is running Windows 2000 Server involves the following
steps:
1. Removing Active Directory
2. Renaming the computer
3. Reinstalling Active Directory
4. Restoring the domain controller to its original configuration
When you rename a domain controller, you must reinstall any services that cannot identify the
computer name dynamically or that can only operate on a domain controller. You do not need to
reinstall any of the services that ship with Windows 2000 Server, such as File and Print sharing
or DNS.
It is recommended that you do not rename a domain controller unless it is absolutely necessary.
For example, it would be necessary to rename a domain controller if:
 You moved the domain controller to another site and the name of the domain
controller needs to map to the naming convention of the new site.
 The name of the domain controller was chosen in error; such as when the naming
convention requires the site name and a derivative of the domain, but the name
includes the incorrect site or domain.
Because renaming a domain controller requires that Active Directory be removed and then
reinstalled on the computer, the impact on the network of renaming a domain controller is
identical to the impact of installing Active Directory to create a new domain controller or global
catalog server.
 Contents 51

Renaming Domain Controllers Tasks and Procedures

Table 10 lists the tasks and procedures for renaming domain controllers.
Table 10 Tasks and Procedures for Renaming Domain Controllers
Recommende
Tasks Procedures Tools
d Frequency
Identify the  Determine whether the domain  Active As needed.
current controller is a global catalog Directory
configuration server. Sites and
of the domain  View the operations master role Services
controller. holders.  Ntdsutil.exe
 Transfer forest-level  Services
operations master roles, if  Regedit.exe
appropriate.
 Transfer domain-level
operations master roles, if
appropriate.
 Determine whether the domain
controller is a DNS server.
 Determine the initial change
notification delay.
 Determine whether the domain
controller is a preferred
bridgehead server.
Rename the  Remove Active Directory.  DCPromo.exe As needed.
domain  Rename the member server.  System
controller. Control Panel
 Run the Active Directory
Installation Wizard.
Restore the  Configure the domain controller  Active As needed.
original as a global catalog server, if Directory
configuration appropriate. Sites and
of the domain  Transfer the domain operations Services
controller. master roles, if appropriate.  Active
 Transfer the forest operations Directory
master roles, if appropriate. Users and
Computers
 Create a delegation for the new
domain controller, if appropriate.  Active
Directory
 Create a secondary DNS zone, if
Domains and
appropriate.
Trusts
 Change the delay for initial
 Regedit.exe
notification of an intrasite
replication partner, if appropriate.
 Configure the domain controller
52 Contents

as a preferred bridgehead server,

if appropriate.

Identifying the Current Configuration of a Domain Controller

Because renaming a domain controller involves removing and reinstalling Active Directory, you
must be able to reestablish the current configuration of the domain controller after you rename it.
Before you begin, identify the current configuration of the domain controller so that you can
restore it after you reinstall Active Directory. Specifically, determine the status of the following
roles and configurations:
 Global catalog server. If the domain controller is a global catalog server, the global
catalog partial directory partitions are removed when you remove Active directory.
Therefore, after you rename the domain controller, you need to reconfigure the
domain controller as a global catalog server. For information about configuring a
domain controller as a global catalog server, see “Managing Global Catalog Servers”
in this guide.
 Operations master role holder. If the domain controller holds operations master
roles, it is recommended that you transfer the roles to the standby master for the roles
prior to removing Active Directory. If you do not transfer the roles, they are
transferred automatically, but you have no control over the placement of the roles. By
manually transferring the roles prior to removing Active Directory, you control the
role placement. For information about transferring operations master roles, see
“Managing Operations Masters” in this guide.
 DNS server. Removing Active Directory does not remove the DNS Server service if
it is installed. However, when you reinstall Active Directory, you need to reconfigure
the domain controller to assume authority for the appropriate DNS zones and to
contain all appropriate delegations. For information about configuring DNS server
after installing Active Directory, see “Managing the Installation and Removal of
Active Directory” in this guide.
 Initial change notification delay. This server-specific configuration determines how
long the domain controller waits before it signals its first replication partner that it
has changes. If you change the default initial change notification delay setting on the
domain controller, you need to reconfigure the setting when you reinstall Active
Directory.
 Preferred bridgehead server. This configuration is not recommended for domain
controllers running Windows 2000 Server. However, if the domain controller is
configured to be a preferred bridgehead server, you must reconfigure the domain
controller as a preferred bridgehead server after you reinstall Active Directory. For
more information about using preferred bridgehead servers, see “Managing Site
Topology” in this guide.
 Contents 53

Caution
The registry editor bypasses standard safeguards, allowing settings that can
damage your system, or even require you to reinstall Windows. If you must edit
the registry, back up system state first. For information about backing up
system state, see "Active Directory Backup and Restore" in this guide.

Procedures for Identifying the Current Configuration of a Domain Controller

Use the following procedures to identify the current configuration of the domain controller. You
need to reconfigure the current configuration on the renamed domain controller after you reinstall
Active Directory.
1. Determine whether the domain controller is a global catalog server .
2. View the operations master role holders. If roles are held by this domain controller,
transfer the roles to the standby operations master prior to removing Active
Directory, as follows:
 If the domain controller holds any forest-level roles, transfer forest-level operations
master roles.
 If the domain controller holds any domain-level roles, transfer domain-level operations
master roles.
3. Determine whether the domain controller is a DNS server. Make a note of the DNS
configuration so that you can reproduce it when you reinstall Active Directory.
4. Determine the initial change notification delay. If this setting has been changed from
the default on this domain controller, you need to reconfigure the setting after you
rename the server and add Active Directory.
5. Determine whether the domain controller is a preferred bridgehead server .

Renaming a Domain Controller

Before you rename a domain controller, you must remove Active Directory to return the domain
controller to member server status. Prior to performing this procedure, be sure that you have
transferred any operations master roles that are held by the domain controller.
54 Contents

Caution
The registry editor bypasses standard safeguards, allowing settings that can
damage your system, or even require you to reinstall Windows. If you must edit
the registry, back up system state first. For information about backing up
system state, see "Active Directory Backup and Restore" in this guide.

After you remove Active Directory, rename the member server and then reinstall Active
Directory on the member server to restore it to domain controller status.

Procedures for Renaming a Domain Controller

Use the following procedures to rename a domain controller. You must perform these procedures
directly on the domain controller; they cannot be performed remotely.
1. Remove Active Directory. This procedure results in the domain controller becoming
a member server in the domain.
2. Rename the member server.
3. Run the Active Directory Installation Wizard. This procedure installs Active
Directory on the member server to restore it to domain controller status.

Restoring the Original Configuration of a Domain Controller

After you have renamed a member server and returned it to domain controller status, you must
restore the original configuration of the domain controller.
If you transferred any domain operations master roles to another domain controller in the domain
prior to renaming the domain controller, you can now transfer them back to the renamed domain
controller.
If the domain controller was originally configured as a DNS server, you must restore the zone
and delegation configurations. The following instructions are based upon best practice
recommendations for DNS design, as described in “Best Practice Active Directory Design for
Managing Windows Networks” and “Best Practice Active Directory Deployment for Managing
Windows Networks” at http://windows.microsoft.com. Follow the links under Products to
Windows 2000 Server, Technical Resources, Planning & Deployment, Deploying the
 Contents 55

Caution
The registry editor bypasses standard safeguards, allowing settings that can
damage your system, or even require you to reinstall Windows. If you must edit
the registry, back up system state first. For information about backing up
system state, see "Active Directory Backup and Restore" in this guide.

Windows 2000 Server Family. If your deployment uses a different DNS design, you might not
use the delegations and secondary zones described below.
If the domain controller is located in a child domain anywhere in the forest, then you must:
 Create a delegation for the domain controller in the forest root domain.
 Create a secondary zone.
If the domain controller is located in the forest root domain and the forest root domain has a
parent domain, then you must:
 Create a delegation for the new domain controller in the parent domain.
For information about how to configure DNS servers after installing Active Directory, see
“Completing Active Directory Installation” in this guide.

Procedures for Restoring the Original Configuration of a Domain Controller

Use the following procedures to restore a domain controller to its original configuration.
1. Configure the domain controller as a global catalog server, if appropriate.
2. Transfer the domain operations master roles, if appropriate.
3. Transfer the forest operations master roles, if appropriate.
4. Create a delegation for the new domain controller, if appropriate. Perform this
procedure in the parent domain of the domain of the DNS server, if one exists.
5. Create a secondary DNS zone, if appropriate. Perform this procedure only if the DNS
server is located in a child domain, not in the forest root domain.
6. Change the delay for initial notification of an intrasite replication partner , if
appropriate.
7. Configure the domain controller as a preferred bridgehead server, if appropriate.
56 Contents

Managing Global Catalog Servers

Designate global catalog servers in sites to accommodate forest-wide directory searching and so
that Active Directory can determine universal group membership of native-mode domain clients.

Global Catalog Placement

To improve the speed of logging on and searching, place at least one global catalog server in
each site, and at least two global catalog servers if the site has multiple domain controllers. As a
best practice, make half of all domain controllers in a site global catalog servers if the site
contains more than three domain controllers. If your deployment uses a single global domain,
configure all domain controllers as global catalog servers. In a single-domain forest, configuring
all domain controllers as global catalog servers requires no additional resources.
When placing global catalog servers, primary concerns are:
 Does any site have no global catalog servers?
 What domain controllers are designated as global catalog servers in a particular site?

Initial Global Catalog Replication

When you add a global catalog server to a site, the Knowledge Consistency Checker (KCC)
updates the replication topology, after which replication of partial domain directory partitions
that are available within the site begins. Replication of partial domain directory partitions that are
available only from other sites begins at the next scheduled interval.
Adding subsequent global catalog servers within a site requires only intrasite replication and does
not affect network performance. Replication of the global catalog potentially affects network
performance only when adding the first global catalog server in the site, and the impact varies
depending on the following conditions:
 The speed and reliability of the wide area network (WAN) link or links to the site.
 The size of the forest.
For example, in a forest that has a large hub site, five domains, and thirty small branch sites
(some of which are connected by only dial-up connections), global catalog replication to the
small sites takes considerably longer than replication of one or two domains to a few well-
connected sites.

Global Catalog Readiness

After replication of the partial domain directory partitions, the domain controller advertises as a
global catalog server and begins accepting queries on ports 3268 and 3269. The requirements for
advertising as a global catalog server differ in Microsoft Windows 2000 Server Service Pack 3
(SP3) and in Windows 2000 Server Service Pack 2 (SP2). The default requirements in
Windows 2000 Server SP3 include replication of all domain directory partitions in the forest. The
default requirements in Windows 2000 Server SP2 are limited to replication of the domain
directory partitions that are local to the site. If the domain controller advertises as a global
catalog server before it has complete information from all domains in the forest, it might return
false information to applications that begin using the server for forest-wide searches.
 Contents 57

For example, Microsoft Exchange 2000 servers use the global catalog exclusively when looking
up addresses. A domain controller that advertises as a global catalog server before it contains all
partial directory partitions can cause Address Book lookup and mail delivery problems for
Exchange clients. To avoid this problem, ensure that the domain controller does not advertise as
global catalog server before it contains all partial domain directory partitions.
Premature advertisement of the global catalog is an issue only for global catalog servers that are
running Windows 2000 Server SP2, and only when you add the first global catalog server in a
site that does not include all domains. If all domains are represented in the site, or if a global
catalog server already exists in the site, then the new global catalog server always has all
domains prior to advertising as a global catalog server.

Global Catalog Removal

When you remove the global catalog, the domain controller immediately stops advertising as a
global catalog server. The KCC gradually removes the read-only replicas from the domain
controller.

Global Catalog Server Management Tasks and Procedures

Table 11 shows the tasks and procedures for managing global catalog servers.
Table 11 Global Catalog Server Management Tasks and Procedures
Tasks Procedures Tools Frequency
Identify the global  Determine whether a  Active Monthly.
catalog servers in a domain controller is a global Directory
site. catalog server. Sites and
Services
Identify a site that  Determine whether a site  Nltest.exe Daily to weekly,
has no global has at least one global depending on
catalog server. catalog server. environment.
58 Contents

Add the global Windows 2000 Server SP2:  Net stop As needed.
catalog to a domain  Stop the Net Logon service  Active
controller and verify (first global catalog server Directory
global catalog in the site only). Sites and
readiness. Services
 Configure the domain
controller as a global  Dcdiag.exe
catalog server.  Repadmin.e
 Monitor global catalog xe
replication progress (first  Ldp.exe
global catalog server in the
 DNS
site only).
 ADSI Edit
 Verify successful replication
to a domain controller.
 Verify global catalog
readiness.
 Restart the Net Logon
service, if needed.
 Restart the global catalog
server and verify global
catalog DNS registrations.
Windows 2000 Server SP3:
 Configure the domain
controller as a global
catalog server.
 Verify global catalog
readiness.
 Restart the global catalog
server and verify global
catalog DNS registrations.
Remove the global  Clear the global catalog  Active As needed.
catalog from a setting. Directory
domain controller.  Monitor global catalog Sites and
removal. Services
 Event
Viewer

Identifying Global Catalog Servers in a Site

Maintain a list of those servers that are designated as global catalog servers. Routinely check
these servers to ensure that no one has changed the designation. Check other servers to ensure
that no one has erroneously designated a global catalog server.
 Contents 59

Procedure for Identifying a Global Catalog Server

Use the following procedure to determine whether a domain controller is a global catalog server.
The procedure is explained in detail in the linked topic.
 To determine whether a domain controller is a global catalog server, check the
properties on the NTDS Settings object of the respective server object.

Identifying a Site That Has No Global Catalog Servers

To quickly identify a site that has no global catalog servers, you can perform one command
rather than check each server individually. You can perform this test any time you add a site, or
routinely if global catalog servers can potentially be removed inappropriately.

Procedure for Identifying a Site that has No Global Catalog Servers

Use the following procedure to determine whether a site has a global catalog server. The
procedure is explained in detail in the linked topic.
 To identify a site that has no global catalog servers, determine whether the site has at
least one global catalog server.

Adding the Global Catalog to a Domain Controller and

Verifying Readiness
When conditions in a site warrant adding a global catalog server, you can configure a domain
controller to be a global catalog server. Selecting the Global catalog setting on the NTDS
Settings object prompts the KCC to update the topology. After the topology is updated, then
read-only partial domain directory partitions are replicated to the designated domain controller.
When replication must occur between sites to create the global catalog, the site link schedule
determines when replication can occur.
Minimum hardware requirements for global catalog servers depend upon the numbers of users in
the site. Table 12 contains guidelines for assessing hardware requirements.
Table 12 Global Catalog Hardware Requirements
Users in site Domain controller
<= 100 One uniprocessor PIII 500, 512 MB.
101 – 500 One uniprocessor PIII 500, 512 MB.
500 – 1,000 One Dual PIII 500, 1 GB.
1,001 – 10,000 Two Quad PIII XEON, 2 GB.
> 10,000 users One Quad PIII XEON, 2 GB for every 5,000 users.
60 Contents

When configuring a global catalog server, be sure the machine has adequate hard disk space. Use
the information in Table 13 to determine how much storage to provide for the Active Directory
database.
Table 13 Global Catalog Storage Requirements for the Active Directory Database
Server Active Directory database storage requirements
Domain controller 0.4 GB of storage for each 1,000 users.
Global catalog server
 DC storage requiremen t 
 DC storage requiement s for other domains
2

For example, in a forest with two 10,000-user domains, all domain controllers need 4 GB of
storage. All global catalog servers require 6 GB of storage.
These requirements represent conservative estimates. For a more accurate determination of
storage requirements, download and run the Active Directory Sizer Tool (ADSizer.exe). You can
download the ADSizer.exe tool from the Active Directory Sizer Tool link on the Web Resources
page at http://www.microsoft.com/windows/reskits/webresources.

Occupancy Levels and Global Catalog Server Readiness

The occupancy level setting on a domain controller determines the criteria for advertising itself
as a global catalog server in DNS. If a global catalog server advertises itself before it has
synchronized all read-only directory partition replicas, clients can receive incorrect information.
The requirements of the occupancy levels are as follows (each higher level includes all levels
below it):
 0: No occupancy requirement.
 1: An inbound connection for at least one read-only directory partition in the site of
the global catalog server is added to the designated server by the KCC. Event
ID 1264 in the Directory Service log signals creation of the inbound connection.
 2: At least one read-only directory partition in the site is replicated to the global
catalog server.
 3: Inbound connections for all read-only directory partitions in the site are added by
the KCC, and at least one is replicated to the server.
 4: All read-only directory partitions in the site are replicated to the server.
 5: Inbound connections for all read-only directory partitions in the forest are added
by the KCC, and all directory partitions in the site are replicated to the server.
 6: All directory partitions in the forest are replicated to the server.
Default occupancy levels for domain controllers that are running Windows 2000 Server depend
on the Windows 2000 Server service pack release that is installed, as follows:
 Windows 2000 Server SP2 or earlier: default and maximum occupancy level = 4.
 Windows 2000 Server SP3: default and maximum occupancy level = 6.
 Contents 61

Exchange 2000 servers use the global catalog exclusively when looking up addresses. Therefore,
in addition to causing Active Directory client search problems, the condition of a global catalog
server being advertised before it receives all partial replicas can cause Address Book lookup and
mail delivery problems for Exchange clients.
The Name Service Provider Interface (NSPI) must be running on a global catalog server to
enable MAPI access to Active Directory. To enable NSPI, you must restart the global catalog
server after replication of the partial directory partitions is complete.

Verification of Global Catalog Server Readiness

A global catalog is considered ready to serve clients when the following events occur, in this
order:
 Occupancy level requirements are met by replicating read-only replicas.
 The isGlobalCatalogReady rootDSE attribute is set to TRUE.
 The Net Logon service on the domain controller has updated DNS with global-
catalog-specific SRV resource records.
At this point, the global catalog server is available to respond to requests on ports 3268 and 3269.
However, in response to various tests, the local system can indicate that it is a global catalog
server as soon as replication requirements are met, but before DNS has been updated. For a
global catalog server that is running Windows 2000 Server SP2, you must also consider the
replication requirements for the occupancy level. For the first global catalog server in a site, the
occupancy level is significant if all domains are not represented in the site.
Global Catalog Readiness in the SP2 Environment
Because the default occupancy level requirement in Windows 2000 Server SP2 is limited to
replicating only the domain directory partitions that are available in the local site, a global
catalog server in this environment might advertise itself as ready when other domains are not
present on the server. For this reason, when adding the first global catalog to a site where all
domains in the forest are not represented, you must take steps to ensure that the global catalog
server does not advertise itself before all domain directory partitions are present on the server, as
follows:
 Prior to configuring the domain controller to be a global catalog server, stop the Net
Logon service on the domain controller. If the Net Logon service is not running, then
the server cannot update DNS prematurely.
 Monitor replication until all domain directory partitions are replicated to the server.
 Verify successful replication of all domain directory partitions in the forest.
 Restart the domain controller to enable NSPI. Restarting will also start the Net Logon
service automatically.
 Verify DNS updates.
62 Contents

Global Catalog Readiness in the SP3 Environment

Because the default occupancy level requirement in Windows 2000 Server SP3 is level 6, a new
global catalog server does not advertise itself until all partial domain directory partitions in the
forest are replicated to the server. In this case, you do not have to stop the Net Logon service
before configuring the domain controller as a global catalog server. However, you do need to
restart the domain controller to enable NSPI.

Procedures for Adding the Global Catalog to a Domain Controller and Verifying
Global Catalog Readiness
Use the following procedures to add a global catalog server to a domain controller. The
procedures are explained in detail in the linked topics. Some procedures are performed only
when you are configuring the first global catalog server in the site or only when Windows 2000
Server SP2 is running on the domain controller that you are configuring.
1. Stop the Net Logon service on the domain controller (SP2 only, first global catalog
server in the site only).
2. Configure the domain controller as a global catalog server. Setting the Global
Catalog check box initiates the process of replicating all domains to the server.
3. Monitor global catalog replication progress (first global catalog server in the site
only).
4. Verify successful replication to a domain controller on the global catalog server.
Check for inbound replication of all partial domain directory partitions in the forest,
to ensure that all domain directory partitions have replicated to the global catalog
server.
5. Verify global catalog readiness. This procedure indicates that the replication
requirements have been met.
6. Restart the Net Logon service, if needed. If you are adding the first global catalog
server in a site to a domain controller that is running Windows 2000 Server SP2 and
you stopped the Net Logon service prior to adding the global catalog, then restart the
service now.
7. Restart the global catalog server and verify global catalog DNS registrationss by
checking DNS for global catalog SRV resource records.

Removing the Global Catalog from a Domain Controller

If the user population of a site decreases to the point where multiple global catalog servers are
not required, or if a global catalog server is being replaced with a more powerful machine, then
you can remove the global catalog from the domain controller.
The procedure to remove the global catalog is simply to clear the Global Catalog check box on
the NTDS Settings object properties page. As soon as you perform this step, the domain
controller stops advertising itself as a global catalog server (Net Logon de-registers the global
catalog-related records in DNS) and immediately stops accepting LDAP requests over ports 3268
and 3269.
 Contents 63

When you remove the global catalog from a domain controller, the KCC begins removing the
read-only replicas one at a time by means of an asynchronous process that removes objects
gradually over time. Each time the KCC runs (every 15 minutes by default), it attempts the
removal of the read-only replica until there are no remaining objects. At an estimated rate of
2000 objects per hour, complete removal of the global catalog from the domain controller can
take from several hours to days, depending on the size of the directory.

Procedures for Removing the Global Catalog from a Domain Controller

Use the following procedures to remove the global catalog from a domain controller. The
procedures are explained in detail in the linked topics.
1. Clear the Global Catalog setting.
2. Monitor global catalog removal in Event Viewer.

Managing Operations Masters

Operations masters keep the directory functioning properly by performing specific tasks that no
other domain controllers are permitted to perform. Because operations masters are critical to the
long-term performance of the directory, they must be available to all domain controllers and
desktop clients that require their services. Careful placement of your operations masters becomes
more important as you add more domains and sites to build your forest.

Operations Master Roles

Three operations master roles exist in each domain:
 The primary domain controller (PDC) emulator. The PDC emulator processes all
replication requests from Microsoft Windows NT 4.0 backup domain controllers and
processes all password updates for clients that are not running Active Directory–
enabled client software.
 The relative identifier (RID) master. The RID master allocates RIDs to all domain
controllers to ensure that all security principals have a unique identifier.
 The infrastructure master. The infrastructure master for a given domain maintains
a list of the security principals from other domains that are members of groups within
its domain.
In addition to the three domain-level operation master roles, two operations master roles exist in
each forest:
 The schema master, which governs all changes to the schema.
 The domain naming master, which adds and removes domains to and from the
forest.
To perform these functions, the domain controllers hosting these operations master roles must be
located in areas where network reliability is high and they need to be consistently available.
64 Contents
Insufficient
Master
Decommissioning
Incompatible
Role transfer
operations
service
configuration
role
oflevel
the
holder
domain
changes
failure
controller
Reasons to Move an Operations Master Role
Operations master role holders are placed automatically when the first domain controller in a
given domain is created. The three domain-level roles are assigned to the first domain controller
created in a domain. The two forest-level roles are assigned to the first domain controller created
in a forest.
You might need to move a master operations role to a different domain controller if the service
level becomes insufficient, if the domain controller holding the operations master role fails or is
decommissioned, or if you make incompatible configuration changes.
The PDC emulator is the operations master role that most impacts the performance of a domain
controller. For clients that do not run Active Directory client software, the PDC emulator
processes requests for password changes, replication, and user authentication. While providing
support for these clients, the domain controller continues to perform its normal services, such as
authenticating Active Directory–enabled clients. As the network grows, the volume of client
requests can increase the workload for the domain controller that hosts the PDC emulator role
and its performance can suffer. To solve this problem, you can transfer all or some of the master
operation roles to another, more powerful domain controller. You may choose to transfer the role
to another domain controller, upgrade the hardware on the original domain controller and then
transfer the role back again.
In the event of a failure, you must decide if you need to relocate the master operations roles to
another domain controller or wait for the domain controller to be returned to service. Base that
determination on the role that the domain controller hosts and the expected down time.
Before permanently taking a domain controller offline, transfer any operations master roles that
the domain controller holds to another domain controller.
Configuration changes to domain controllers or the network topology can result in the need to
transfer master operations roles. Except for the infrastructure master, you can assign operations
master roles to any domain controller regardless of any other tasks that the domain controller
performs. Do not host the infrastructure master role on a domain controller that is also acting as a
global catalog server, unless all of the domain controllers in the domain are global catalog
servers, or unless only one domain is in the forest. If the domain controller hosting the
infrastructure master role is configured to be a global catalog server, you must transfer the
infrastructure master role to another domain controller. Changes to the network topology can
result in the need to transfer operation master roles in order to keep them in a particular site.

Considerations for Moving Operations Master Roles

You can reassign an operations master role by transfer or, as a last resort, by seizure.
Role transfer is the preferred method to move an operations master role from one domain
controller to another. During a role transfer, the two domain controllers replicate to ensure that
no information is lost. After the transfer completes, the previous role holder reconfigures itself so
that it no longer attempts to perform as the operations master while the new domain controller
assumes those duties. This prevents the possibility of duplicate operations masters existing on the
network at the same time, which can lead to corruption in the directory.
 Contents 65
Role seizure
Important
If you must seize an operations master role, never reattach the previous role
holder to the network without following the procedures in this guide. Incorrectly
reattaching the previous role holder to the network can result in invalid data and
corruption of data in the directory.

Seize a role only as a last resort to assign a role to a different domain controller. Use this process
only when the previous operations master fails and remains out of service for an extended
amount of time. During a role seizure, the domain controller does not verify that replication is
updated, so recent changes can be lost. Because the previous role holder is unavailable during the
role seizure, it cannot know that a new role holder exists. If the previous role holder comes back
online it might still assume that it is the operations master. This can result in duplicate operations
master roles on the network, which can lead to corruption of data in the directory and ultimately
to the failure of the domain or forest.
To transfer a role to a new domain controller, ensure that the destination domain controller is a
direct replication partner of the previous role holder and that replication between them is up to
date and functioning properly. This minimizes the time required to complete the role transfer. If
replication is sufficiently out of date, the transfer can take a while, but it eventually finishes.

Guidelines for Role Placement

By improperly placing operations master role holders, you might prevent clients from changing
their passwords, or be unable to add domains and new objects, such as users and groups. You
might also be unable to make changes to the schema. In addition, name changes might not
properly appear within group memberships that are displayed in the user interface.
As your environment changes, you must avoid the problems associated with improperly placed
operations master role holders. Eventually, you might need to reassign the roles to other domain
controllers.
Although you can assign the forest-level and domain-level operations master roles to any domain
controller in the forest and domain respectively, improperly placing the infrastructure master role
can cause it to not function properly. Other improper configurations can increase administrative
overhead.
Requirements for infrastructure master placement
Do not place the infrastructure master on a domain controller that is also a global catalog server.
The infrastructure master updates the names of security principals from other domains that are
added to groups in its own domain. For example, if a user from one domain is a member of a
group in a second domain and the user's name is changed in the first domain, then the second
domain is not notified that the user's name must be updated in the group's membership list.
Because domain controllers in one domain do not replicate security principals to domain
controllers in another domain, the second domain never becomes aware of the change. The
infrastructure master constantly monitors group memberships, looking for security principals
from other domains. If it finds one, it checks with the security principal's domain to verify that
the information is updated. If the information is out of date the infrastructure master performs the
update and then replicates the change to the other domain controllers in its domain.
66 Contents
Forest-level role
Domain-level roleplacement
placementon
inonthe
a global
the
forest
same
catalog
root
domain
domain
server
controller
Two exceptions apply to this rule. First, if all the domain controllers are global catalog servers,
the domain controller that hosts the infrastructure master role is insignificant because global
catalogs do replicate the updated information regardless of the domain to which they belong.
Second, if the forest has only one domain, the domain controller that hosts the infrastructure
master role is not needed because security principals from other domains do not exist.
Recommendations for role placement
Although you can assign the operations master roles to any domain controller, follow these
guidelines to minimize administrative overhead and ensure the performance of Active Directory.
If a domain controller that is hosting operation master roles fails, following these guidelines also
simplifies the recovery process. Guidelines for role placement include:
 Leave the two forest-level roles on a domain controller in the forest root domain.
 Place the two forest-level roles on a global catalog server.
 Place the three domain-level roles on the same domain controller.
 Do not place the domain-level roles on a global catalog server.
 Place the domain-level roles on a higher performance domain controller.
 Adjust the workload of the operations master role holder, if necessary.
 Choose an additional domain controller as the standby operations master for the
forest-level roles and choose an additional domain controller as the standby for the
domain-level roles.
The first domain controller created in the forest is assigned the schema master and domain
naming master roles. To ease administration and backup and restore procedures, leave these roles
on the original forest root domain controller. Moving the roles to other domain controllers does
not improve performance. Separating the roles creates additional administrative overhead when
you must identify the standby operations masters and when you implement a backup and restore
policy.
Unlike the PDC emulator role, forest-level roles rarely place a significant burden on the domain
controller. Keep these roles together to provide easy, predictable management.
In addition to hosting the schema master and domain naming master roles, the first domain
controller created in a forest also hosts the global catalog. In Windows 2000 Server, you must
leave the domain naming master on a global catalog server. When the domain naming master
creates an object representing a new domain, it uses the global catalog to ensure that no other
object has the same name. The domain naming master achieves this consistency by running on a
global catalog server, which contains a partial replica of every object in the forest.
The three domain-level roles are assigned to the first domain controller created in a new domain.
Except for the forest root domain, leave the roles at that location. Keep the roles together unless
the workload on your operations master justifies the additional management burden of separating
the roles.
For the forest root domain, the first domain controller also hosts the two forest-level roles as well
as the global catalog. This additional workload requires you to take two precautionary steps to
avoid potential problems. First, the domain-level roles must not remain on a global catalog
 Contents 67
Standby operations
Domain-level
Workload adjustment
role placement
absence
master
of theon
operations
on
a global
a higher
catalog
master
performance
role
server
holder
domain controller
server. In addition, because hosting all five roles on a single domain controller can overload the
server and hurt performance, transfer the three domain-level roles to another domain controller.
Because all pre-Active Directory clients submit updates to the PDC emulator, the domain
controller holding that role uses a higher number of RIDs. Place the PDC emulator and RID
master roles on the same domain controller so these two roles interact more efficiently.
If you must separate the roles, you can still use a single standby operations master for all three
roles. However, you must ensure that the standby is a replication partner of all three of the role
holders.
Backup and restore procedures also become more complex if you separate the roles. Special care
must be taken to restore a domain controller that hosted an operations master role. By hosting the
roles on a single computer, you minimize the steps that are required to restore a role holder.
Do not host the infrastructure master on a domain controller that is acting as a global catalog
server. Because it is best to keep the three domain-level roles together, avoid putting any of them
on a global catalog server.
Host the PDC emulator role on a powerful and reliable domain controller to ensure that it is
available and capable of handling the workload. Of all the operations master roles, the PDC
emulator creates the most overhead on the server that is hosting the role. It has the most intensive
daily interaction with other systems on the network. The PDC emulator has the greatest potential
to affect daily operations of the directory.
Domain controllers can become overloaded while attempting to service client requests on the
network, manage their own resources, and handle any specialized tasks such as performing the
various operations master roles. This is especially true of the domain controller holding the PDC
emulator role. Pre-Active Directory clients and domain controllers running Windows NT 4.0 rely
more heavily on the PDC emulator than Active Directory clients and Windows 2000 Server
domain controllers. If your networking environment has pre-Active Directory clients and domain
controllers, you might need to reduce the workload of the PDC emulator.
If a domain controller begins to indicate that it is overloaded and the performance is affected, you
can reconfigure the environment so that some tasks are performed by other, less-used domain
controllers. By adjusting the domain controller's weight in the DNS environment, you can
configure the domain controller to receive fewer client requests than other domain controllers on
your network. Optionally, you can adjust the domain controller's priority in the DNS
environment so it processes client requests only if other DNS servers are unavailable. With fewer
DNS client requests to process, the domain controller can use more resources to perform
operations master services for the domain.
The standby operations master is a domain controller that you identify as the computer that
assumes the operations master role if the original computer fails. You do not need to perform any
special configuration steps or run any type of setup utilities to make a domain controller a
standby operations master. This precautionary planning step helps make your operation more
resilient if a problem arises that requires you to reassign a master operations role to a new
domain controller.
68 Contents

Ensure that the standby operations master is a direct replication partner of the actual operations
master. If the standby operations master domain controller is a direct replication partner of the
original operations master, it most likely contains the most recent changes to the domain. This
reduces the time required to transfer the role to the standby operations master and, in the case of
a failure, reduces the chances of losing information. Even if replication is not totally complete,
only few outstanding updates exist. Those outstanding updates can be replicated by a normal
replication cycle rather than requiring a full synchronization, which replicates all of the account
information in the partition. To guarantee that the two domain controllers are replication partners,
you must manually create a connection object between them. Although creating manual
connection objects is not generally recommended, in this one case it is necessary because it is so
important that these two domain controllers be replication partners.
If you must reassign the domain-level operations master roles to the standby operations master,
do not place the infrastructure master role on a global catalog server.

Ramifications of Role Seizure

If a role is seized, the new role holder is configured to host the operations master role with the
assumption that you do not intend to return the previous role holder to service. Use role seizure
only when the previous role holder is not available and you need the operations master role to
keep the directory functioning. Because the previous role holder is not available during a seizure,
you cannot reconfigure the previous role holder and inform it that another domain controller is
now hosting the operations master role.
If the previous role holder comes back online, its behavior depends on your current service pack
level. If you are running Windows 2000 Server Service Pack 2 (SP2) or earlier, the domain
controller waits for one replication cycle while it attempts to verify the current role holder. If the
previous role holder receives data that indicates that another domain controller is performing the
operations master role, it reconfigures itself so that it no longer hosts the operations master role
and Active Directory functions properly. If for any reason replication fails, it does not receive
any replicated data indicating that a new operations master exists. Whether or not replication
actually occurs, after one replication cycle it assumes that the data it has is correct. It leaves itself
configured as the current operations master and attempts to resume its duties as the operations
master role holder. This results in duplicate operations masters on the network. As shown in
Table 14, this can cause serious problems in the directory.
If you are running Windows 2000 Server Service Pack 3 (SP3), the previous role holder waits for
a full replication cycle to complete successfully before it resumes the role of operations master.
By waiting for a full replication cycle, it can see if another operations master exists before it
brings itself back online. If the previous role holder detects that another operations master exists,
it reconfigures itself so that it no longer hosts the roles in question.
To reduce risk, perform a role seizure only if the missing operations master role unacceptably
affects performance of the directory. Calculate the effect by comparing the impact of the missing
service provided by the operations master to the amount of work that is needed to bring the
previous role holder safely back online after you perform the role seizure.
 Contents 69

Active Directory continues to function when the operations master roles are not available. If the
role holder is only offline for a short period, you might not need to seize the role to a new
domain controller. Remember that returning an operation master to service after the role is seized
can have dire consequences if it is not done properly.
Table 14 Operations Master Role Functionality Risk Assessment
Consequences if Recommendation for
Operations
Role is Risk of Improper Restoration Returning to Service
Master Role
Unavailable After Seizure
Schema You cannot make Conflicting changes can be Not recommended.
master changes to the introduced to the schema if both Can lead to a
schema. schema masters attempt to modify corrupted forest and
the schema at the same time. This require rebuilding
can result in a fragmented the entire forest.
schema.
Domain You cannot add You cannot add or remove Not recommended.
naming master or remove domains or clean-up metadata. Can require
domains from the Domains might appear as though rebuilding domains.
forest. they are still in the forest even
though they are not.
PDC emulator You cannot Password validation can randomly Allowed. User
change pass or fail. Password changes authentication can be
passwords on take much longer to replicate erratic for a time, but
pre-Active throughout the domain. no permanent
Directory clients. damage occurs.
No replication to
Windows NT 4.0
backup domain
controllers.
Infrastructure Delays displaying Displays incorrect user names in Allowed. May
master updated group group membership lists in the user impact the
membership lists interface after you move users performance of the
in the user from one group to another. domain controller
interface when hosting the role, but
you move users no damage occurs to
from one group the directory.
to another.
RID master Eventually, Duplicate RID pools can be Not recommended.
domain allocated to domain controllers, Can lead to data
controllers cannot resulting in data corruption in the corruption that can
create new directory. This can lead to require rebuilding
directory objects security risks and unauthorized the domain.
as each of their access.
individual RID
pools is depleted.
70 Contents

Operations Master Role Management Tasks and Procedures

Table 15 shows the tasks and procedures for managing operations master roles.
Table 15 Operations Master Role Management Tasks and Procedures
Tasks Procedures Tools Frequency
Designate  Verify successful replication  Repadmin.exe As needed
operations to a domain controller.  Active
master roles.  Determine whether a domain Directory Sites
controller is a global catalog and Services
server.  Active
 Transfer the forest-level Directory
operations master roles. Domains and
 Transfer the domain-level Trusts
operations master roles.  Active
 View the current operations Directory Users
master role holders. and Computers
 Ntdsutil.exe
Reduce the  Change the weight for DNS  Regedit.exe As needed
workload on the SRV records in the registry. 
PDC emulator.  Change the priority for DNS
SRV records in the registry.
Decommission a  Verify successful replication  Repadmin.exe As needed
role holder. to a domain controller.  Active
 Determine whether a domain Directory Sites
controller is a global catalog and Services
server.  Active
 Transfer the forest-level Directory
operations master roles. Domains and
 Transfer the domain-level Trusts
operations master roles.  Active
 View the current operations Directory Users
master role holders. and Computers
 Ntdsutil.exe
Seize operations  Verify that a complete end-to-  Ntdsutil.exe As needed
master roles. end replication cycle had 
occurred.
 Verify successful replication
to a domain controller.
 Seize the operations master
role.
 View the current operations
master role holders.
Choose a  Determine whether a domain  Active As needed
 Contents 71

standby controller is a global catalog Directory Sites

operations server. and Services
master.  Create a connection object. 

Designating Operations Master Roles

When you create a new domain, the Active Directory Installation Wizard automatically assigns
all of the domain-level operations master roles to the first domain controller that is created in that
domain. When you create a new forest, the wizard also assigns the two forest-level operations
master roles to the first domain controller. After the domain is created and functioning, you
might transfer various operations master roles to different domain controllers to optimize
performance and simplify administration.
Transferring the forest-level and domain-level operations master roles is performed as needed
and governed by the guidelines for placing operations master roles. Before you transfer an
operations master role, use Repadmin.exe with the /showreps option to ensure that replication
between the current role holder and the domain controller assuming the role is updated.
In addition, you must determine if the domain controller that you intend to assume an operations
master role is a global catalog server. The domain naming master, a forest-level role, must also
host the global catalog. However, the infrastructure master for each domain must not host the
global catalog.
Do not change the global catalog configuration on the domain controller that you intend to
assume an operations master role unless your IT management authorizes that change. Changing
the global catalog configuration can cause changes that can take days to complete and the domain
controller might not be available during that period. Instead, transfer the operations master roles
to a different domain controller that is already properly configured.

Procedures for Designating Operations Master Roles

Procedures are explained in detail in the linked topics.
1. Verify successful replication to a domain controller.
2. Determine whether a domain controller is a global catalog server.
3. Transfer the forest-level operations master roles.
4. Transfer the domain-level operations master roles.
5. View the current operations master role holders.

Reducing the Workload on the PDC Emulator

You can configure a domain controller so that DNS sends the majority of client requests to other
domain controllers. Reducing the number of client requests helps reduce the workload on a
domain controller, giving it more time to function as an operations master, and is especially
important for the PDC emulator. Of all the operations master roles, the PDC role has the highest
72 Contents

impact on the domain controller hosting that role. You might need to take steps to keep that
domain controller from becoming overloaded.
To receive information from the domain, a client uses DNS to locate a domain controller and
then sends the request to that domain controller. By default, DNS performs rudimentary load
balancing and randomizes the distribution of client requests so they are not always sent to the
same domain controller. If too many client requests are sent to a domain controller while it
attempts to perform other duties, such as those of the PDC emulator, it can become overloaded,
which has a negative impact on performance. To reduce the number of client requests that are
processed by the PDC emulator, you can adjust its weight in the DNS environment or you can
adjust its priority in the DNS environment.

DNS Weight Registry Setting

Adjusting the weight of a domain controller to less than other domain controllers reduces the
number of clients that DNS refers to that domain controller. The default weight for all domain
controllers is 100. By reducing this value, DNS refers clients to a domain controller less
frequently based on the proportion of this value to the value of other domain controllers. For
example, to configure the system so that the domain controller hosting the PDC emulator role
receives requests only half as many times as the other domain controllers, configure the weight
of the domain controller hosting the PDC emulator role to be 50. DNS determines the weight
ratio for that domain controller to be 50/100 (50 for that domain controller and 100 for the other
domain controllers). After you reduce this ratio to 1/2, DNS refers clients to the other domain
controllers twice as often as it refers to the domain controller with the reduced weight setting. By
reducing client referrals, the domain controller receives fewer client requests and has more
resources for other tasks, such as performing the role of PDC emulator.

DNS Priority Registry Setting

Adjusting the priority of the domain controller also reduces the number of client referrals.
However, rather than reducing it proportionally to the other domain controllers, changing the
priority causes DNS to stop referring all clients to this domain controller unless all domain
controllers with a lower priority setting are unavailable.
To configure the PDC emulator in this manner, use Regedit.exe to modify the ldapsrvpriority or
ldapsrvweight registry entries.

Procedures for Reducing the Number of Client Requests Processed by the PDC
Emulator
Procedures are explained in detail in the linked topics.
1. Change the weight for DNS SRV records in the registry.
2. Change the priority for DNS SRV records in the registry.

Decommissioning a Role Holder

When you use the Active Directory Installation Wizard to decommission a domain controller that
currently hosts one or more operations master roles, the wizard reassigns the roles to a different
domain controller. When the wizard is run, it determines whether the domain controller currently
 Contents 73

hosts any operations master roles. If it detects any operations master roles, it queries the directory
for other eligible domain controllers and transfers the roles to a new domain controller. A domain
controller is eligible to host the domain-level roles if it is a member of the same domain. A
domain controller is eligible to host a forest-level role if it is a member of the same forest.
You cannot control which domain controller the wizard chooses and the wizard does not indicate
which domain controller receives the roles. Because of this behavior, it is best to transfer the
roles prior to running the wizard. That way you can control role placement and can transfer the
roles according to the recommendations discussed earlier in this guide.

Transfer to the Operations Master Standby

Transfer the operations master roles to the standby operations master. By following the
recommendations for operations master role placement, the standby operations master is a direct
replication partner and is ready to assume the roles. Remember to designate a new standby for
the domain controller that assumes the roles.

Transfer when No Standby Operations Master is Ready

If you do not follow the recommendations for role placement and you have not designated a
standby operations master, you must properly prepare a domain controller to which you intend to
transfer the operations master roles. Preparing the future role holder is the same process as
preparing a standby operations master. You must manually create a connection object to ensure
that it is a replication partner with the current role holder and that replication between the two
domain controllers is updated. To determine whether the standby domain controller received the
latest replicated updates from the current operations master, use Repadmin.exe with the
/showreps option.
In addition, you must determine whether the domain controller intended to assume an operations
master role is a global catalog server. The domain naming master, a forest-level role, must also
host the global catalog. However, the infrastructure master for each domain must not host the
global catalog.
Do not change the global catalog configuration on the domain controller that you intend to
assume an operations master role unless your IT management authorizes that change. Changing
the global catalog configuration can cause changes that can take days to complete and the domain
controller might not be available during that period. Instead, transfer the operations master roles
to a different domain controller that is already properly configured.

Procedures for Decommissioning a Role Holder

Procedures are explained in detail in the linked topics.
1. Verify successful replication to a domain controller.
2. Determine whether a domain controller is a global catalog server.
3. Transfer the forest-level operations master roles.
4. Transfer the domain-level operations master roles.
5. View the current operations master role holders.
74 Contents

Seizing Operations Master Roles

Seize an operations master role only as a last resort. If at all possible, transfer an operations
master role to a new domain controller instead. Seize an operations master role only if the current
role owner is offline and is unlikely to return to service.
Role seizure is the act of assigning an operations master role to a new domain controller without
the cooperation of the current role holder (usually because it is offline due to a hardware failure).
During role seizure, a new domain controller assumes the operations master role without
communicating with the current role holder.
Role seizure can create two conditions that can cause problems in the directory. First, the new
role holder starts performing its duties based on the data located in its current directory partition.
The new role holder might not receive changes that were made to the previous role holder before
it went offline if replication did not complete prior to the time when the original role holder went
offline. This can cause data loss or data inconsistency into the directory database.
To minimize the risk of losing data to incomplete replication, do not perform a role seizure until
enough time has passed to complete at least one complete end-to-end replication cycle across
your network. Allowing enough time for complete end-to-end replication ensures that the domain
controller that assumes the role is as up-to-date as possible.
Second, the original role holder is not informed that it is no longer the operations master role
holder, which is not a problem if the original role holder stays offline. However, if it comes back
online (for example, if the hardware is repaired or the server is restored from a backup), it might
try to perform the operations master role that it previously owned. This can result in two domain
controllers performing the same operations master role simultaneously. Depending on the role in
question and whether your environment runs Windows 2000 Server SP2 or Windows 2000
Server SP3, this can disrupt the directory service. For example, a RID master might reallocate a
duplicate RID pool resulting in corruption of data in the directory. The severity of duplicate
operations master roles varies from no visible effect to the need to rebuild the entire forest. For
more information about the risks of returning an operations master to service after the role is
seized to another domain controller, see “Ramifications of Role Seizure” earlier in this guide.
If you are seizing a role and you have not designated another domain controller as the standby
operations master, you can use Repadmin.exe with the /showreps option to identify a domain
controller that has the most recent updates from the current role holder. Seize the operations
master role to that domain controller to minimize the impact of the role seizure.

Procedures for Seizing Operations Master Roles

Procedures are explained in detail in the linked topics.
1. Verify that a complete end-to-end replication cycle has occurred. During the design
process, you calculated the maximum end-to-end replication latency. The maximum
end-to-end replication latency is the maximum amount of time it should take for
replication to take place between the two domain controllers in your enterprise that
are farthest from each other based on the topology of your network. If you verify that
replication is functioning properly and wait this amount of time without making any
 Contents 75

additional changes to the directory then you can assume that all changes have been
replicated and the domain controller is up to date.
2. Verify successful replication to a domain controller (the domain controller that will
be seizing the role).
3. Seize the operations master role.
4. View the current operations master role holders.

Choosing a Standby Operations Master

A single domain controller can act as the standby operations master for all of the operations
master roles in a domain, or you can designate a separate standby for each operations master role.
Following the recommendations, it is best to select one standby for the forest-level roles and
another standby in each domain that can be used to host the three domain-level roles if their host
fails.
No utilities or special steps are required to designate a domain controller as a standby operations
master. However, the current operations master and the standby should be well connected. This
means that the network connection between them must support at least 10 megabit transmission
rate and be available at all times. In addition, configure the current role holder and the standby as
direct replication partners by manually creating a connection object between them.
Configuring a replication partner can save some time if you must reassign any operations master
roles to the standby operations master. Before transferring a role from the current role holder to
the standby operations master, ensure that replication between the two computers is functioning
properly. Because they are replication partners, the new operations master is as updated as the
original operations master, thus reducing the time required for the transfer operation. To
determine whether the standby domain controller received the latest replicated updates from the
current operations master, use Repadmin.exe with the /showreps option.
During role transfer, the two domain controllers exchange any unreplicated information to ensure
that no transactions are lost. If the two domain controllers are not direct replication partners, a
substantial amount of information might need to be replicated before the domain controllers
completely synchronize with each other. The role transfer requires extra time to replicate the
outstanding transactions. If the two domain controllers are direct replication partners, fewer
outstanding transactions exist and the role transfer operation completes sooner.
Designating a domain controller as a standby also minimizes the risk of role seizure. By making
the operations master and the standby direct replication partners, you reduce the chance of data
loss in the event of a role seizure, thereby reducing the chances of introducing corruption into the
directory.
When you designate a domain controller as the standby, follow all recommendations that are
discussed in “Guidelines for Role Placement” earlier in this guide. To designate a standby for the
forest-level roles, choose a global catalog server so it can interact more efficiently with the
domain naming master. To designate a standby for the domain-level roles, ensure that the domain
controller is not a global catalog server so that the infrastructure master continues to function
properly if you must transfer the roles.
76 Contents

Note If you also set an alert threshold, divide the above warning thresholds in half.

Manually create a connection object between the operations master and the designated standby
operations master to ensure that replication occurs between the two domain controllers.

Procedures for Choosing a Standby Operations Master

Procedures are explained in detail in the linked topics.
1. Determine whether a domain controller is a global catalog server.
2. Create a connection object.

Managing the Database

Active Directory is stored in the Ntds.dit database file. In addition to this file, the directory uses
log files, which store transactions prior to committing them to the database file. For best
performance, store the log files and the database on separate hard drives.
The directory database is a self-maintained system. Other than regular backup, the directory
database requires no daily maintenance during ordinary operation. However, you might need to
manage the following conditions:
 Low disk space: Monitor free disk space on the partition or partitions that store the
directory database and logs. Provide warnings at the following logical-disk-space
thresholds:
 Ntds.dit partition: The greater of 20 percent of the Ntds.dit file size or 500 megabytes
(MB).
 Log file partition: The greater of 20 percent of the combined log files size or 500 MB.
 Ntds.dit and logs on the same volume: The greater of 1 gigabyte (GB) or 20 percent of
the combined Ntds.dit and log files sizes.
 Database size: During ordinary operation, the database removes expired
tombstones and defragments (consolidates) white space. This automatic online
defragmentation redistributes and retains white space for use by the database. The
following conditions might warrant taking steps to regulate database size manually:
 Temporary backlog of expired tombstones following bulk deletions: Large-scale
deletions can temporarily increase the database file size if tombstones expire in larger
numbers than garbage collection can remove in one cycle (5,000 tombstones per cycle).
After objects are deleted, their tombstones are stored in the directory for 60 days by
default and cannot be removed prior to that time. However, after the tombstone lifetime
expires, you can speed removal of the tombstone backlog by temporarily decreasing the
default garbage collection period (12 hours).
 Contents 77

 Increased white space due to large-scale deletions: If data is decreased significantly,

such as when the global catalog is removed from a domain controller, white space is not
automatically returned to the file system. Although this condition does not affect
database operation, it does result in a larger file size. You can use offline
defragmentation to decrease the size of the database file by returning white space from
the database file to the file system.
 Hardware upgrade or failure: If you need to upgrade or replace the disk on which the
database or log files are stored, move the files to a different location, either
permanently or temporarily.
For information about monitoring the database and log file partitions for low disk space, see
“Monitoring Active Directory” earlier in this guide.

General Guidelines for Directory Database Management

For all database management tasks, follow these guidelines:
 Prior to performing any procedures that affect the directory database, be sure that you
have a current system state backup. For information about performing system state
backup, see “Active Directory Backup and Restore” earlier in this guide.
 To manage the database file itself, you must take the domain controller offline by
restarting in Directory Services Restore Mode, and then use Ntdsutil.exe to manage
the file.
 To start a domain controller in Directory Services Restore Mode, you must log on to
the domain controller as the local Administrator. To remotely manage the database,
you can use Terminal Services Client to restart the domain controller in Directory
Services Restore Mode.
 NTFS Disk Compression is not supported for the database and log files.

Directory Database Management Tasks and Procedures

Table 16 shows the tasks and the procedures for managing the database.
Table 16 Directory Database Management Tasks and Procedures
Tasks Procedures Tools Frequency
Relocate  Determine the databasesize and  dir As needed.
directory location (online or offline).  Backup Wizard
database files.  Compare size of the directory  Terminal Services
database files to the volume Client
size.
 Notepad
 Back up system state.
 Ntdsutil.exe
 Restart the domain controller in
 Windows
Directory Services Restore
Explorer
Mode (locally or remotely).
 Move the directory database
files.
78 Contents

 Move the directory database

files to a local drive.
 Copy the directory database
files to a remote share and
back.
 If the path has changed, back up
system state.
Return unused  Change the garbage collection  Registry editor As needed.
disk space from logging level.  Backup Wizard
the directory  Back up system state.  net use, del, copy
database to the
 Restart the domain controller in  Ntdsutil.exe
file system.
Directory Services Restore
Mode (locally or remotely).
 Compact the directory database
offline (offline
defragmentation).
 Check database integrity.
 If no errors, perform
standard semantic database
analysis.
 If errors, perform semantic
database analysis with
fixup.
 If errors, perform database
recovery.
Speed removal  Change (decrease) the garbage  ADSI Edit As needed.
of an expired- collection period.  Registry editor
tombstone  Change (increase) the garbage  Event Viewer
backlog. collection logging level.
 Ntdsutil.exe
 Verify removal of tombstones in
the event log.
 Change (return to normal) the
garbage collection period.
 Change (return to normal) the
garbage collection logging
level.
 Compact the directory database
offline (offline
defragmentation), if needed.

Relocating Directory Database Files

The following conditions require moving database files:
 Contents 79

 Hardware maintenance: If the physical disk on which the database or log files are
stored requires upgrading or maintenance, the database files must be moved, either
temporarily or permanently.
 Low disk space: When free disk space is low on the logical drive that stores the
database file (Ntds.dit), the log files, or both, first verify that no other files are
causing the problem. If the database file or log files are the cause of the growth, then
provide more disk space by taking one of the following actions:
 Expand the partition on the disk that currently stores the database file, the log files, or
both. This procedure does not change the path to the files and does not require updating
the registry.
 Use Ntdsutil.exe to move the database file, the log files, or both to a larger existing
partition. Moving files to a different partition changes the path to the files and therefore
requires updating the registry. Ntdsutil.exe automatically updates the registry when you
use it to move database files.

Path Considerations
If the path to the database file or log files changes as a result of moving the files, be sure that
you:
 Use Ntdsutil.exe to move the files (rather than copying them) so that the registry is
updated with the new path. Even if you are moving the files only temporarily, use
Ntdsutil.exe to move files locally so that the registry is always current.
 Perform a system state backup as soon as the move is complete so that the restore
procedure uses the correct path.
 Verify that the correct permissions are applied on the destination folder following the
move. Revise permissions to those that are required to protect the database files, if
needed.

SYSVOL Considerations
If you replace or reconfigure a drive that stores the SYSVOL folder, you must first move the
SYSVOL folder manually. For information about moving SYSVOL manually, see “Managing
SYSVOL” later in this guide.

Procedures for Relocating Directory Database Files

Use the following procedures to move or copy the database file, the log files, or both. Procedures
are explained in detail in the linked topics.
1. Determine the location and size of the directory database files. Use the database size
to prepare a destination location of the appropriate size. Track the respective file sizes
during the move to ensure that you successfully move the correct files. Be sure to use
the same method to check file sizes when you compare them. The size is reported
differently, depending on whether the domain controller is online or offline, as
follows:
 Determine the database size and location online. This size is reported in bytes.
80 Contents

 Determine the database size and location offline. This size is reported in megabytes
(MB). Use this method if the domain controller is already started in Directory Services
Restore Mode.
2. Compare the size of the directory database files to the volume size. Before moving
any files in response to low disk space, verify that no other files on the volume are
responsible for the condition of low disk space.
3. Back up system state. System state includes the database file and log files as well as
SYSVOL and NETLOGON shared folders, among other things. Always ensure that
you have a current backup prior to moving database files.
4. Restart the domain controller in Directory Services Restore Mode, as follows:
 If you are logged on to the domain controller console, locally restart the domain
controller in Directory Services Restore Mode.
 If you are using Terminal Services for remote administration, modify the Boot.ini file
on the remote server so that you can remotely restart the domain controller in Directory
Services Restore Mode.
5. Move the database file, the log files, or both. Move the files to a temporary
destination if you need to reformat the original location, or to a permanent location if
you have additional disk space. Moving the files can be performed locally by using
Ntdsutil.exe or remotely (temporarily) by using a file copy, as follows:
 Move the directory database files to a local drive.
 Copy the directory database files to a remote share and back. When copying any
database files off the local computer, always copy both the database file and the log
files.
6. If the path to the database or log files has changed, back up system state so that the
restore procedure has the correct information.

Returning Unused Disk Space from the Directory Database to

the File System
During ordinary operation, the white space in the directory database file becomes fragmented.
Each time garbage collection runs (every 12 hours by default), white space is automatically
defragmented online to optimize its use within the database file. The unused disk space is thereby
maintained for the database; it is not returned to the file system.
Only offline defragmentation can return unused disk space from the directory database to the file
system. When database contents have decreased considerably through a bulk deletion (for
example, you remove the global catalog from a domain controller), if the size of the database
backup is significantly increased due to the white space, use offline defragmentation to reduce
the size of the Ntds.dit file.
You can determine how much free disk space is recoverable from the Ntds.dit file by setting the
Garbage Collection logging level in the registry. Changing the Garbage Collection logging level
 Contents 81

Caution
Setting the value of entries in the Diagnostics subkey to greater than 3 can
degrade server performance and is not recommended.

from the default value of 0 to a value of 1 results in event ID 1646 being logged in the Directory
Service log. This event describes the total amount of disk space used by the database file as well
as the amount of free disk space that is recoverable from the Ntds.dit file through offline
defragmentation.
At Garbage Collection logging level 0, only critical events and error events are logged in the
Directory Service log. At level 1, high-level events are logged as well. Events can include one
message for each major task that is performed by the service. At level 1, the following events are
logged for garbage collection:
 700 and 701: report when online defragmentation begins and ends, respectively.
 1646: reports the amount of free space available in the database out of the amount of
allocated space.
Following offline defragmentation, perform a database integrity check. The integrity
command in Ntdsutil.exe detects binary-level database corruption by reading every byte in the
database file. The process ensures that the correct headers exist in the database itself and that all
of the tables are functioning and consistent. Therefore, depending upon the size of your Ntds.dit
file and the domain controller hardware, the process might take considerable time. In testing
environments, the speed of 2 GB per hour is considered to be typical. When you run the
command, an online graph displays the percentage completed.

Procedures for Performing Offline Defragmentation

Use the following procedures to perform offline defragmentation. Procedures are explained in
detail in the linked topics.
1. Change the garbage collection logging level to 1. Check the Directory Service event
log for event ID 1646, which reports the amount of disk space that you can recover
by performing offline defragmentation.
2. Back up system state. System state includes the database file and database log files as
well as SYSVOL, NETLOGON, and the registry, among other things. Always ensure
that a current backup exists prior to defragmenting database files.
3. Take the domain controller offline, as follows:
 If you are logged on to the domain controller locally, restart the domain controller in
Directory Services Restore Mode.
 If you are using Terminal Services for remote administration, you can remotely restart
the domain controller in Directory Services Restore Mode after modifying the Boot.ini
file on the remote server.
4. Compact the directory database file (offline defragmentation) . As part of the offline
defragmentation procedure, check directory database integrity.
5. If database integrity check fails, perform semantic database analysis with fixup.
82 Contents

Note
Tombstones cannot be removed prior to expiration of the tombstone lifetime.

Speeding Removal of an Expired-Tombstone Backlog

An object that is deleted from Active Directory is stored as a tombstone, which represents the
deleted object in the directory so that the deletion is replicated. Tombstones remain in the
directory for a default period of 60 days from the time of deletion, at which point they expire and
are permanently removed by garbage collection.

Although tombstones use less space than the full object, they can affect the size of the database
temporarily following large bulk deletions. A maximum of 5,000 expired tombstones can be
deleted at one time. If the number of expired tombstones exceeds 5,000, more than one garbage
collection interval is required to clear the backlog. During the backlog, tombstones that are no
longer needed are retained, consuming database space.

Increased Rate of Tombstone Removal

The default garbage collection period is 12 hours. Temporarily decreasing the garbage collection
period (for example, to 1 hour) can help speed the removal of expired tombstones. However,
setting this period too low can also cause slow performance, so be sure to return the value to the
original setting as soon as the backlog is cleared. To reduce database size by returning the white
space left by the removed tombstones to the file system, perform offline defragmentation after
the backlog is cleared.

Logging of Tombstone Removal

The default logging level for garbage collection is 0. At this level, only errors are reported. When
garbage collection logging is set to 3, event ID 1006 reports the number of expired tombstones
removed during each garbage collection cycle.
If you want to track removal of expired tombstones, increase the logging level to 3 and decrease
the garbage collection period until the backlog is cleared, and then return the logging level and
the garbage collection period to normal.

Procedures for Regulating Directory Database Growth Caused by Tombstones

Use the following procedures to manage removal of tombstones following bulk deletions.
 Contents 83

1. Change the garbage collection period to a lower interval. Decreasing the interval
between garbage collections helps the system eliminate the tombstone backlog more
quickly.
2. Change the garbage collection logging level to 3. Increasing the logging level to 3
causes an event that reports the number of tombstones removed each time garbage
collection occurs.
3. Verify removal of tombstones in the event log. Check the Directory Service event log
for NTDS event ID 1006, which reports the number of expired tombstones removed.
When this event indicates that the number of tombstones removed is less than 5,000,
the backlog has been cleared.
4. Change the garbage collection period. When the event ID 1006 reports a number of
removed tombstones less than 5,000, you can return the interval between garbage
collections to the normal level.
5. Change the garbage collection logging level, if needed. If you no longer want
informational events logged for garbage collection, return the logging level to 0.
6. Compact the directory database file (offline defragmentation) , if needed. Clearing the
backlog does not remove the white space created by the tombstones. Only offline
defragmentation returns unused disk space to the file system.

Managing SYSVOL
The Windows 2000 Server System Volume (SYSVOL) is a collection of folders and reparse
points in the file systems that exist on each domain controller in a domain. SYSVOL provides a
standard location to store Group Policy objects (GPOs) and scripts so that the File Replication
service (FRS) can distribute them to other domain controllers and member computers in a
domain.
FRS monitors SYSVOL and if a change occurs to any file stored on SYSVOL, then FRS
automatically replicates the changed file to the SYSVOL folders on the other domain controllers
in the domain.
Computers that run Windows 2000 Server obtain GPOs, logon, logoff, startup, and shutdown
scripts from the SYSVOL shared folder. Windows NT 4.0–based domain controllers and
Windows-based clients that do not run Active Directory client software obtain GPOs and scripts
from the NETLOGON shared folder.
During the installation of Active Directory, the folders and reparse points are automatically
created in the %SystemRoot%/SYSVOL folder. FRS automatically replicates any files or GPOs
that are written to these folders to the other domain controllers in the domain, to ensure that they
are available and ready to be used when a user logs on to the domain.
The day-to-day operation of SYSVOL is an automated process that does not require any human
intervention other than watching for alerts from the monitoring system. Occasionally, you might
perform some system maintenance as you change your network. The procedures you might
perform include:
84 Contents
Capacity
Hardware
Performance
Data Organization
Maintenance
Note
If you receive indications that disk space is low, determine if the cause is
inadequate physical space on the disk, or a registry setting that allocates
inadequate disk space to SYSVOL. By modifying a setting in the registry, you
can allocate more disk space to SYSVOL rather than relocating SYSVOL or the
Staging Area. Increasing the space allocation in the registry is much faster and
easier than relocation. For more information about managing disk space, see
"Maintaining Sufficient Disk Space" later in this section.

 Relocating SYSVOL
 Relocating the Staging Area
 Changing the size of the Staging Area
These procedures involve moving SYSVOL or portions of SYSVOL to alternate locations. You
might perform these procedures to maintain capacity and performance of SYSVOL, for hardware
maintenance, or for data organization.
Depending upon the configuration of your network, SYSVOL can require much disk space to
function properly. During the initial deployment, SYSVOL might be allocated adequate disk
space to function. However, as your network grows, the required capacity can exceed the
available disk space.

Any changes made to SYSVOL are automatically replicated to the other domain controllers in
the domain. If the files stored in SYSVOL change frequently, the replication increases the input
and output for the volume where SYSVOL is located. If the volume is also host to other system
files, such as the directory database or the pagefile, then the increased input and output for the
volume can impact the performance of the server.
System maintenance, such as removal of a disk drive, can require you to relocate SYSVOL. Even
if the maintenance occurs on a different disk drive, verify that that maintenance does not affect
the system volume. Logical drive letters can change after you add and remove disks. FRS locates
SYSVOL by using pointers stored in the directory and the registry. If drive letters change after
you add or remove disk drives, be aware that these pointers are not automatically updated.
Some organizations prefer to control where specific data is stored for organizational purposes
and established backup and restore policies.
 Contents 85

Guidelines for Managing SYSVOL

To manage SYSVOL, ensure that FRS properly replicates the SYSVOL data, and provide
enough space to store SYSVOL. Implement a monitoring system that can detect low disk space
and potential FRS disruptions so that you can address those issues before the system stops
replicating. For more information about monitoring SYSVOL, see “Monitoring Active
Directory” in this guide.
Disk space maintenance
SYSVOL stores and replicates GPOs, Distributed File System (DFS) information, and scripts. As
the network grows, SYSVOL can begin to require substantial storage space. Although you do
plan for storage requirements for SYSVOL during the planning stages of deployment, you might
need to adjust the storage requirements after you deploy additional domain controllers due to
network growth and the way in which FRS replicates files.
FRS replicates files by making a temporary copy of the files in a Staging Area folder and then
sending the copies to replication partners. This method avoids problems that locked files can
cause while replication occurs. Because FRS replicates copies of the files, the original files
remain available for user access during replication. However, this method requires making a
copy of every file prior to replication. Based on the size and number of files involved, a
substantial amount of disk space might be required for temporary storage.
When the Staging Area folder runs out of disk space, FRS behaves differently depending on the
version of Windows 2000 that is running. If Windows 2000 Server Service Pack 2 (SP2) or
earlier is running, then the system will stop replicating until space is made available. If
Windows 2000 Server Service Pack 3 (SP3) is running, then FRS will detect when it is about to
run out of disk space and start removing the least recently used files to provide more space.
Although this prevents the system from halting replication, it does increase input and output for
the server’s disk and can impact performance. For more information about the changes to FRS
from Windows 2000 Server SP2 to Windows 2000 Server SP3, see KB article Q321557 in the
Microsoft Knowledge Base. To view the Microsoft Knowledge Base, see the Microsoft
Knowledge Base link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.
Both FRS and DFS use the Staging Area folder. To maintain sufficient disk space for SYSVOL,
estimate the amount of space that DFS uses as well as the space that FRS uses. For more
information about DFS, see “Distributed File System” in the Distributed Systems Guide of the
Windows 2000 Server Resource Kit.
Because the Staging Area folder holds files from all replication partners, you must consider
traffic to and from all partners when you estimate the disk space requirements for the Staging
Area folder on each computer.
If replication must occur between domain controllers that are located in different sites, remember
that FRS uses the same connection objects as Active Directory. You can configure those
connection objects so that replication can occur only during certain times of the day. Each
connection object has an associated schedule that dictates what hours of the day the connection is
available for replication. Allocate enough time in the schedule for all Active Directory replication
and all FRS replication to occur. If FRS does not complete all outstanding replication requests
86 Contents
Relocating only the Staging Area
when the schedule makes the connection available, it will hold the remaining unreplicated files
until the next time the connection becomes available. Over time, this backlog of unreplicated
files can grow to consume an enormous amount of disk space.
Additional SYSVOL recommendations
You can preserve Staging Area and bandwidth usage by following these best practices:
 Run Windows 2000 SP2 on all domain controllers that run FRS. Install
Windows 2000 SP3 as soon as possible.
 Always keep FRS service running, especially when you make bulk changes to FRS-
replicated files or files outside the tree on the same drive.
 Do not run anti-virus software against FRS-replicated directories.
 Do not enable File System Group Policy on any FRS-replicated tree.
 Watch for inconsistent directories. Duplicate folders that appear in the FRS
replication tree on multiple domain controllers can cause inconsistent directories.
Although this is not a critical problem, it can result in unanticipated behavior, such as
changes appearing to be lost. If this occurs, examine the files in these directories to
determine which directory is the proper version and then delete the duplicated
directories from the tree.
 Do not leave files open for extended amounts of time. FRS cannot replicate a file
while it is open. Avoid using elements in scripts that cause a file to be open for an
extended amount of time, such as a script that waits for user input before proceeding.
If the user is not present when the script runs, the file can remain open and cannot be
replicated until the script terminates.
 Do not attempt to relocate SYSVOL or the Staging Area if the FRS environment on
your network is unstable and you are having problems with system volumes
becoming unsynchronized among replication partners. Troubleshoot the FRS
problems and ensure that the environment is stable before attempting any relocation
operations. During all relocation operations except authoritative restore, FRS rebuilds
the SYSVOL content by replicating data from its replication partners. If FRS is not
functioning properly on the partners, their SYSVOL data may be invalid. This can
result in invalid SYSVOL data in the new location. The relocation operation can also
fail because FRS cannot replicate the necessary data from the domain controller's
replication partners.
SYSVOL and Staging Area relocation
Deployment is the best time to determine the location of SYSVOL. Consider performance and
disk capacity to determine the best location for the SYSVOL folders. During the Active
Directory installation, you must specify the location of the SYSVOL folders. After installation,
you might need to relocate SYSVOL or the Staging Area folder.
Although SYSVOL contains many folders, the Staging Area requires the most capacity because
it is used for replication. You can leave SYSVOL in its original location and relocate only the
Staging Area.
 Contents 87
Relocating SYSVOL and the Staging Area
You can relocate the entire SYSVOL folder
and its associated subtrees, including the Staging Area.
You can relocate SYSVOL by removing and reinstalling Active Directory on the domain
controller or by manually recreating SYSVOL at a new location.
Active Directory removal and reinstallation
To relocate SYSVOL, removing and reinstalling Active Directory is far easier and more reliable
than manually recreating SYSVOL at a new location, but it can also be impractical. To relocate
SYSVOL by using this method, you use the Active Directory Installation Wizard to remove
Active Directory from the domain controller then use it again to reinstall Active Directory on the
same domain controller. During the reinstallation, provide the new location for SYSVOL. The
replication process populates the folders with the appropriate files from another domain
controller. This method might not be practical to use because having a large number of objects in
your directory increases the required time for reinstallation and you might need to reinstall and
reconfigure other services if the domain controller runs additional services.
Manual SYSVOL relocation
To manually recreate the SYSVOL folder at the new location, copy the data from the existing
location to the new location and then reconfigure FRS to point to the new location. Ensure that
you properly copy all files to the new location.
Manually relocate SYSVOL only as a last resort, when you cannot remove and reinstall Active
Directory on the domain controller. If you must perform this procedure, ensure that the SYSVOL
replication between the domain controller and its replication partners is as up-to-date as possible.
If the domain controller is not replicating properly with its partners, do not attempt to recreate
SYSVOL until you determine why replication is not functioning and make the necessary fixes.
For more information about recreating SYSVOL manually, see KB article Q304300 in the
Microsoft Knowledge Base. To view the Microsoft Knowledge Base, see the Microsoft
Knowledge Base link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.

SYSVOL Management Tasks and Procedures

Table 17 shows the tasks and procedures for managing SYSVOL.
Table 17 SYSVOL Management Tasks and Procedures
Tasks Procedures Tools Frequency
Change the space  Stop the File Replication  Regedit.exe As needed
allocated to the service.
Staging Area  Change the space
folder. allocated to the Staging
Area folder.
 Start the File Replication
service
Relocate the  Identify replication  Active Directory As needed
Staging Area partners. Sites and Services
folder.
88 Contents

 Check the status of the  Dcdiag.exe

SYSVOL.  Windows Explorer
 Verify replication is  ADSI Edit
functioning.
 Regedit.exe
 Gather the SYSVOL path
information.
 Stop the File Replication
service.
 Create the new Staging
Area folder.
 Set the Staging Area path.
 Prepare a domain
controller for non-
authoritative SYSVOL
restore.
 Start the File Replication
service.
 Contents 89

Move SYSVOL  View the current  Active Directory As needed

by using the operations master role Users and
Active Directory holders. Computers
Installation  Transfer the forest-level  Active Directory
Wizard. operations master roles. Sites and Services
 Transfer the domain-level  Dcdiag.exe
operations master roles.  Netdiag.exe
 Determine whether a  DCPromo.exe
domain controller is a
 DNS snap-in
global catalog server.
 Verify DNS registration
and functionality.
 Verify communication
with other domain
controllers.
 Verify the existence of
the operations masters.
 Remove Active
Directory.
 Delete a server object
from a site.
 Verify DNS registration
and functionality.
 Install Active Directory.
 Verify the site assignment
for the domain controller.
 Move a server object to a
different site if the
domain controller is
located in the wrong site.
 Perform final DNS
configuration.
 Check the status of the
shared system volume.
 Verify DNS registration
and functionality.
 Verify domain
membership for the new
domain controller.
 Verify communication
with other domain
controllers.
 Verify replication is
functioning.
 Verify the existence of
90 Contents

Move SYSVOL  Identify replication  Active Directory As needed

manually. partners. Sites and Services
 Check the status of the  Dcdiag.exe
shared system volume.  NTBackup.exe
 Verify replication is  ADSI Edit
functioning.
 Regedit.exe
 Gather the SYSVOL path
 Linkd.exe
information.
 Stop the File Replication
service.
 Create the SYSVOL
folder structure.
 Set the SYSVOL path.
 Set the Staging Area path.
 Set the fRSRootPath.
 Prepare a domain
controller for non-
authoritative SYSVOL
restore.
 Update security on the
new SYSVOL.
 Start the File Replication
service.
 Check the status of the
SYSVOL.
Update the  Gather the SYSVOL path  Regedit.exe As needed
SYSVOL path. information.  Windows Explorer
 Stop the File Replication  ADSI Edit
service.
 Linkd.exe
 Set the SYSVOL path.
 Set the fRSRootPath.
 Set the Staging Area path.
 Start the File Replication
service.
 Contents 91

Restore and  Identify replication  Active Directory As needed

rebuild SYSVOL. partners. Sites and Services
 Check the status of the  Dcdiag.exe
SYSVOL.  Windows Explorer
 Verify replication is  Regedit.exe
functioning.
 Linkd.exe
 Restart the domain
controller in Active
Directory Restore Mode
(locally or remotely).
 Gather the SYSVOL path
information.
 Stop the File Replication
service.
 Prepare the domain
controller for non-
authoritative SYSVOL
restore.
 Import the SYSVOL
folder structure.
 Start the File Replication
service.
 Check the status of the
shared system volume.

Changing the Space Allocated to the Staging Area

The Staging Area is a folder inside the SYSVOL folder. FRS replicates files by making copies of
the files, storing these copies in the Staging Area folder, and then sending them to replication
partners. Because FRS replicates a copy of the file, the original file remains available for user
access during replication.
The Staging Area stores files prior to being replicated and stores files that it has just received
through replication. Although FRS compresses the data and attributes of the replicated files to
save space in the Staging Area folder and reduce the time that is needed to replicate the files, this
method requires making and storing a copy of every file prior to replication and can require a
substantial amount of disk space to store all of the copies.
When you examine the disk space that SYSVOL uses, you need to examine both physical disk
space and allocated disk space. Physical disk space refers to the amount of space that is available
on the disk drive. To prevent SYSVOL from using all physical disk space available on the drive,
an entry in the registry limits the amount of space that SYSVOL can use. This is the allocated
disk space.
92 Contents

The default size of the Staging Area folder is 675 megabytes(MB). The minimum size is 10 MB
and the maximum size is 2 terabytes. You can adjust the size limit of the Staging Area folder by
setting the value in kilobytes (KB) of the Staging Space Limit registry entry in
HKEY_Local_Machine\System\CurrentControlSet\Services\NtFrs\Parameters. For more
information about setting the Staging Space Limit in the registry, see KB article Q221111 in the
Microsoft Knowledge Base. To view the Microsoft Knowledge Base, see the Microsoft
Knowledge Base link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.
When the Staging Area folder runs out of disk space, FRS behaves differently depending on the
version of Windows 2000 Server that is running. If Windows 2000 Server Service Pack 2 (SP2)
or earlier is running, then FRS fills the Staging Area to the limit defined in the registry and then
suspends inbound and outbound replication until disk space is made available. In this situation,
you can avoid suspension of replication by generously estimating the amount of disk space that
SYSVOL requires.
If Windows 2000 Server Service Pack 3 (SP3) is running, then FRS fills the Staging Area to
90 percent of the limit specified in the registry and then starts removing the least recently used
files to make more space available. While this prevents FRS from suspending replication, it can
affect the performance of the domain controller. If a large number of files are constantly being
updated, then FRS constantly stages, removes, and restages files to maintain available disk space
in the Staging Area. In this case, making more space available reduces the amount of work that
the domain controller performs in order to keep FRS functioning.

Other Considerations for Estimating Required Disk Space

Both FRS and DFS use the Staging Area folder. The Staging Space Limit in the registry applies
to the sum of the space that is used by DFS and FRS. To maintain sufficient disk space for
SYSVOL, estimate the amount of space that DFS uses as well as the space that FRS uses.
If a file changes, FRS replicates the entire file and not just the change. If two replication partners
have different values set for the Staging Space Limit, the maximum size of a file that FRS can
replicate is the lower of the two values.
The Staging Area folder holds files from all replication partners. You must consider traffic to and
from all partners when you estimate the disk space requirements for the Staging Area folder in
each SYSVOL.
Active Directory replication uses connection objects to establish connections between replication
partners. FRS uses the same connections for its own replication. Two factors control the rate that
replication can take place over those connections: availability of the connection and transmission
speed of the network. Each connection object has an associated schedule that allows
administrators to dictate when the connection is available for replication. Network administrators
can limit the time that replication can take place so that processes that are more important to the
daily operation of the business can use available network bandwidth over a specific connection.
This becomes especially important if two replication partners are connected by a slow link (such
as a 128 Kbps dial-up connection). The schedule makes it possible to limit replication traffic so
that it occurs only at night or during off-peak hours.
 Contents 93

FRS stages all replication traffic and waits for the connection to become available. When the
connection is available, it begins replication and continues until it replicates all outstanding files,
or the connection becomes unavailable. If many files are awaiting replication and the network is
busy handling other traffic, then FRS might not get a chance to replicate all outstanding files
before the schedule makes the connection unavailable. If this happens, FRS holds the remaining
files until the schedule permits replication to continue. While FRS is waiting for the schedule to
permit replication, it continues to stage new files for replication. The Staging Area folder needs
enough space to store the staged files as well as to handle any backlog of files that might not get
replicated due to limited availability of the connection.

Procedures for Changing the Space Allocated to the Staging Area

Use the following procedures to change the amount of space that is allocated to the Staging Area
folder. Procedures are explained in detail in the linked topics.
1. Stop the File Replication service.
2. Change the space allocated to the Staging Area folder.
3. Start the File Replication service.

Relocating the Staging Area

The Staging Area folder is likely to use most of the disk space that is allocated to SYSVOL. This
is because the Staging Area folder stores all inbound and outbound files, and sometimes multiple
copies of those files. As the disk space requirements increase, you can allocate more space until
you reach 2 terabytes or the physical limit of the disk drive. The maximum disk space allowed
for the Staging Area is 2 terabytes. If you reach the limit of the disk drive and still have not
reached the 2 TB limit, consider relocating the Staging Area folder to a different disk that has
more space available.
By default, the Active Directory Installation Wizard installs the Staging Area folder within the
SYSVOL. The Active Directory Installation Wizard creates two folders, Staging and Staging
Area, which FRS uses for the staging process. When you relocate the Staging Area, you can
change the folder name. Ensure that you identify the proper folder in case the folder is renamed
in your environment.
Two parameters determine the location of the Staging Area. One parameter, fRSStagingPath, is
stored in the directory and contains the path to the actual location that FRS uses to stage files.
The other parameter is a junction point stored in the Staging Areas folder in SYSVOL that links
to the actual location that FRS uses to stage files. When relocating the Staging Area, you must
update these two parameters to point to the new location.

Procedures for Relocating the Staging Area Folder

Except where noted, perform these procedures on the domain controller that contains the Staging
Area folder that you want to relocate. Procedures are explained in detail in the linked topics.
1. Identify replication partners.
94 Contents

2. On the replication partners, check the status of the shared system volume. You do not
need to perform the test on every partner, but you need to perform enough tests to be
confident that the shared system volumes on the partners are healthy.
3. Verify that replication is functioning.
4. Gather the SYSVOL path information.
5. Stop the File Replication service.
6. Create the new Staging Area folder.
7. Set the Staging Area path.
8. Prepare a domain controller for non-authoritative SYSVOL restore.
9. Start the File Replication service.

Moving SYSVOL by Using the Active Directory Installation

Wizard
Relocate SYSVOL only as a last resort. The many steps involved present many opportunities to
incorrectly configure the system. If you must relocate SYSVOL, use the Active Directory
Installation Wizard because it is far easier and more reliable that manually moving SYSVOL.
The Active Directory Installation Wizard asks for the new location and then automatically
configures the system for you.
Although using the Active Directory Installation Wizard is the preferred method for relocating
SYSVOL, it is also the least practical because it involves decommissioning the domain
controller. When this process is used, the Active Directory Installation Wizard is run on the
domain controller to remove Active Directory. After Active Directory is removed, you run the
wizard again to reinstall Active Directory. During the reinstallation, the wizard asks where you
want to store SYSVOL. You enter the new location and the wizard configures it for you.
Using the Active Directory Installation Wizard to relocate SYSVOL can be too impractical for
two reasons. First, because you are removing Active Directory and then reinstalling it, you also
need to reinstall any other services that depend on Active Directory that are running on that
domain controller. This can amount to hours of additional work and an unacceptable amount of
time for the domain controller to be unavailable. Second, if a large number of objects exist in
your directory, it can take hours or even days to complete the reinstallation when the new domain
controller joins the network and completes the initial replication of the directory.
If this domain controller is not hosting any additional services that depend on the directory, and
your directory does not take an extensive amount of time to complete the initial replication to
new domain controllers, then moving SYSVOL with the Active Directory Installation Wizard
can save you time and be easier and more reliable than moving SYSVOL manually.
 Contents 95

WARNING
Note
If any
Do notofmove
the verification
SYSVOL with teststhefail,
Active
do notDirectory
continueInstallation
until you identify
Wizardandunless
fix the
you
completelyIfunderstand
problems. these tests the
fail,risks
the decommissioning
and consequencesoperation
of decommissioning
is also likelythe
to fail.
domain controller in question.

Procedures for Moving SYSVOL with the Active Directory Installation Wizard
Use the following procedures to remove and reinstall Active Directory in order to move
SYSVOL. For more information about installing and removing Active Directory, see “Managing
Installation and Removal of Active Directory” in this guide. Procedures are explained in detail in
the linked topics.
1. View the current operations master role holders to see if any roles are assigned to this
domain controller.
2. If this domain controller is listed as hosting either the schema master or domain
naming master roles, then transfer the forest-level roles to another domain controller
in the forest root domain. Any domain controller in the forest is capable of hosting
these roles but it is recommended that they remain in the forest root domain. Ensure
that you place the domain naming master role on a global catalog server.
3. If this domain controller is listed as hosting the primary domain controller (PDC)
emulator, infrastructure master or relative identifier (RID) master roles, transfer the
domain-level roles to another domain controller in the same domain. Do not place the
infrastructure master role on a global catalog server unless all of the domain
controllers host the global catalog or unless only one domain exists in the forest.
4. Determine whether a domain controller is a global catalog server and ensure that
other domain controllers are configured as global catalog servers before continuing.
5. Verify DNS registration and functionality.
6. Verify communication with other domain controllers.
7. Verify the existence of the operations masters on the network.
8. Remove Active Directory.
9. Delete the server object from a site.
96 Contents

Note
If the verification test fails, do not continue until you identify and fix the
problems. If the test fails, then installation is also likely to fail.

10. Verify DNS registration and functionality.

11. Install Active Directory. Provide the wizard with the new location for SYSVOL
when prompted.
12. Verify the site assignment for the domain controller.
13. Move a server object to a different site if the domain controller is located in the
wrong site.
14. Perform final DNS configuration for a new domain controller that is located in the
forest root domain:
a. Create a delegation for the new domain controller in the parent domain of the
DNS infrastructure if a parent domain exists and a DNS server hosts it. If a
DNS server does not host the parent domain, then follow the procedures
outlined in the vendor documentation to add the delegation for the new domain
controller.
b. Configure the DNS client settings.
–Or–
Perform final DNS configuration for a new domain controller that is located in a child
domain:
c. Create a delegation for the new domain controller in the forest root domain.
d. Create a secondary zone.
e. Configure the DNS client settings.
15. Check the status of the shared system volume.
16. Verify DNS registration and functionality.
17. Verify domain membership for the new domain controller.
18. Verify communication with other domain controllers.
19. Verify that replication is functioning.
20. Verify the existence of the operations masters.

Moving SYSVOL Manually

If you must move the entire system volume, not just the Staging Area folder, and you have
determined that moving the system volume by using the Active Directory Installation Wizard is
impractical, then you can relocate the system volume manually. Because no utilities can
automate this process, you must carefully ensure that you properly move all folders and maintain
the same level of security at the new location.
 Contents 97

Regardless of the method used to move SYSVOL, these events occur:

 The File Replication service is stopped.
 The proper folder structure is created at the new location.
 The SYSVOL path information is updated in the directory and in the registry.
 Default security settings are set on the new folder structure.
 The File Replication service is restarted.
FRS is stopped while the changes are made and then restarted after the changes are completed.
During the restart process, FRS reads the new configuration information in the directory and the
registry and reconfigures itself to use the new location.
SYSVOL uses an extensive folder structure that must be recreated accurately at the new location.
The easiest method is to copy the folder structure by using Windows Explorer. You must ensure
that you copy any folders that may have special attributes, such as hidden folders.
The folder structure also includes junction points. Junction points look like folders when they
appear in Windows Explorer but they are not really folders. Junction points contain links to other
folders. When you open a junction in Windows Explorer, you see the contents of the folder to
which the junction is linked. If you open a command prompt and display a directory listing that
contains junction points, they are designated as <JUNCTION>, while regular folders are
designated with <DIR>. Junction points behave like regular folders. When you are working in
the file system, you have no indication whether you are working with a junction or a folder.
The difference between folders and junctions appears when you copy or move a junction to a
new location. Because a junction is a link to another location, when you copy a junction to a new
location, the link still refers to the original location. SYSVOL contains two junction points that
point to folders in the SYSVOL tree. When you move the tree to a new location, you must update
the junction points to point to the new location. Otherwise, the junction points continue to point
to the original SYSVOL folders.
The registry and Active Directory store path information that FRS uses to locate the SYSVOL
and the Staging Area folders. You must update these settings to point to the new locations.
After you create the new folders and update the paths and junctions, ensure that the folders get
repopulated with the proper data. You can repopulate the files stored in SYSVOL at the new
location is done by replicating the data into the new location from one of the domain controller's
replication partners. The BURFLAGS option is set in the registry and when FRS restarts, it
replicates the data into the new folders from one of the replication partners. Because this data is
restored to the new location by means of replication, be certain that the system volumes on the
replication partners are updated and functioning properly to ensure that the data replicated into
the new folders is updated and has no errors.
98 Contents

Important
WARNING
This procedure
Remember, if the
cansystem
alter security
volumessettings.
on yourAfter
domain
youcontrollers
complete theare procedure,
becoming the
unsynchronized
security settings to
onthe
thepoint
new system
that youvolume
need toare
relocate
reset to
thethe
system
default
volumes,
settings be
that
sure to
were established
troubleshoot whentheyou
FRSinstalled
problemsActive
and resolve
Directory.
the You
issuesmust
thatreapply
cause the
any
system volumes
changes to the security
to become
settings
unsynchronized
on the systembefore
volume
youthat
attempt
you made
to relocate
since the
you
system volumes.
installed Active Directory. Failure to do so can result in unauthorized access to
Group Policy objects and logon and logoff scripts.

Procedures for Moving SYSVOL Manually

Except where noted, perform these steps on the domain controller that contains the system
volume that you want to move. Procedures are explained in detail in the linked topics.

1. Identify replication partners.

2. On the replication partners, check the status of the shared system volume. You do not
need to perform the test on every partner, but you need to perform enough tests to be
confident that the shared system volumes on the partners are healthy.
3. Verify that replication is functioning.
4. Gather the SYSVOL path information.
5. Stop the File Replication service.
6. Create the SYSVOL folder structure.
 Contents 99

7. Set the SYSVOL path.

8. Set the Staging Area path. If you have moved the Staging Area folder to a different
location already, you do not need to do this step.
9. Set the fRSRootPath.
10. Prepare a domain controller for non-authoritative SYSVOL restore.
11. Update security on the new SYSVOL.
12. Start the File Replication service.
13. Check the status of the shared system volume.

Updating the System Volume Path

Due to system maintenance, you might need to update the system volume path. When you add or
remove disk drives, the logical drive letters of the other drives on the system can change. If either
your SYSVOL or Staging Area folder is located on one of the drives whose letter changes, FRS
cannot locate them. You must update the paths that FRS uses to locate these folders to solve this
problem. To change the path for the system volume, make changes to the registry and in the
directory. Changing the Staging Area path requires a change in the directory. Both changes
require that you update the junction points. After updating the path information, you must restart
FRS so it can reinitialize with the new values.

Procedures for Updating the System Volume Path

Use the following procedures to change the amount of space that is allocated to the Staging Area
folder. Procedures are explained in detail in the linked topics.
1. Gather the System Volume path information.
2. Stop the File Replication service.
3. Set the SYSVOL path (if needed).
4. Set the fRSRootPath (if needed).
5. Set the Staging Area path (if needed).
6. Start the File Replication service.

Restoring and Rebuilding SYSVOL

In some cases, you must recreate or rebuild the SYSVOL on a single domain controller. Attempt
to rebuild SYSVOL on a single domain controller only when all other domain controllers in the
domain have a healthy and functioning SYSVOL. Do not attempt to rebuild SYSVOL until you
correct any problems that are occurring with FRS in a domain.

Procedure for Restoring and Rebuilding SYSVOL

Use these procedures only if you are working on a domain controller that does not have a
functional SYSVOL. Procedures are explained in detail in the linked topics.
100 Contents

1. Identify replication partners.

2. Choose a partner and check the status of the SYSVOL on the partner. Because you
will be copying the system volume from one of the partners, you need to make sure
that the system volume you copy from the partner is up-to-date.
3. Verify that replication is functioning on the partner.
4. Restart the domain controller that is being repaired in Directory Services Restore
Mode. If you are sitting at the console of the domain controller, locally restart a
domain controller in directory services restore mode. If you are accessing the domain
controller remotely using Terminal Services, remotely restart a domain controller in
directory services restore mode.
5. Gather the SYSVOL path information.
6. Stop the File Replication service.
7. Prepare a domain controller for non-authoritative SYSVOL restore.
8. Import the SYSVOL folder structure.
9. Start the File Replication service.
10. Check the status of the shared system volume.

Managing Windows Time Service

The Windows 2000 time service, W32Time, requires little management and is installed by
default on all Windows 2000–based computers. W32Time uses coordinated universal time
(UTC), which is based on an atomic time scale and is independent of time zone.
On computers that are joined to a domain, time synchronization occurs when the W32Time
service starts during system startup. The Net Logon service looks for a domain controller that can
authenticate and synchronize time with the client.

Time Configuration on the Forest-Root PDC Emulator

The time service uses a hierarchical relationship that controls authority and ensures common time
usage. By default, the PDC emulator in the forest root domain is the authoritative time source for
that forest.
Follow these best practices for configuring time on the forest-root PDC emulator, in this order of
preference:
 Install a hardware clock that uses the Network Time Protocol (NTP) on an internal
network, and synchronize the forest-root PDC emulator and the standby PDC
emulator to it.
 Use IPSec to securely synchronize with another network time server.
 Monitor the forest-root PDC emulator closely to ensure that its time is accurate. Do
not synchronize the forest-root PDC emulator with another computer.
 Contents 101

If none of these options are acceptable in your organization, you can synchronize with an
external reliable time source. However, this option is not recommended, as it synchronizes time
in an unauthenticated manner, potentially making time packets vulnerable to an attacker.

System Time Maintenance

Do not advance or roll back the system time on Windows 2000–based servers under any
circumstances, including attempts to:
 Test significant time and date transitions such as Year 2000 testing.
 Force the deletion of tombstones (objects that have been marked for deletion in the
Active Directory).
 Make objects on one computer override the objects on another computer.
 Extend the useful life of a system backup.
 Return a computer to an earlier system state including schema rollback.
 Incorporate test environments into production, after you test time and date transitions
on lab computers.
 Troubleshoot Active Directory or File Replication Service (FRS) issues, by
advancing the system time of a computer in an effort to make the content of one
computer authoritative over another. Advancing the time can adversely affect the
operation of the system, and it is not a useful method of resolving Active Directory or
FRS replication problems.

How advancing system time affects FRS

Advancing the system time affects FRS in the following manner:
 Active Directory prematurely deletes tombstones for deleted objects, causing
incorrect reconciliation later. When an object is deleted, it is not actually removed
from the database. It is instead marked for deletion after 60 days by default. This
tombstone is replicated to other domain controllers. When the tombstone expires, the
object is then permanently deleted. If the tombstone is deleted prematurely, then
updates from replication partners are inconsistent.
 Local file changes create change orders with event times reflecting the advanced
clock time. These change orders are inserted into the outbound log but are not sent
because the computer with the advanced clock will not join with the partners that
remain at the correct time. Later, when the time on this computer is restored to the
correct time and the computer is able to join with its outbound partners, it sends the
change orders with the advanced event time. The downstream partner ignores these
change orders because the event time is too far into the future. The result is that the
files that changed while the time was advanced are not replicated to the other
members, but remain on the computer. Furthermore, the advanced event times cause
the computer to reject updates to these files that originate from other replication
partners.
102 Contents

How advancing system time affects Active Directory

Advancing the system time affects Active Directory in the following manner:
 Replication conflicts might be incorrectly resolved. Active Directory uses the time
service to resolve replication conflicts. When the same attribute on the same object is
changed on two different servers during a latency period, the most recent change is
replicated. Thus, if you advance the time on a computer, all changes originating on
that computer appear as more recent changes and are replicated despite the fact that
they might not be the most recent changes.
 Name conflicts might be incorrectly resolved. Active Directory also uses the time
service to resolve name conflicts. When two different objects with the same name are
created on two servers, Active Directory saves the most recently created object.
Advancing the time on a computer might cause Active Directory to save the wrong
object simply because it reflects a more recent change.
 Restoring from a backup might fail. Backups are only good for the period of the
tombstone lifetime. When you back up the system state, Active Directory generates
an expiry token. The token is submitted when you restore the system state from the
backup and is used to verify that the backup is not too old. Attempting to restore a
backup after you advance the system clock might make the backup appear too old
and cause the backup to fail. Do not restore a backup that you made from a computer
with an advanced time setting.
 Link value replication is impaired. Link value replication uses a timestamp to
distinguish values. Changing the system clock hinders this mechanism.
 Kerberos authentication might fail. Kerberos authentication is based on clock
synchronization. Furthermore, the lifetimes of the Kerberos tickets are exceeded if
the clock is moved too far ahead.

Windows Time Service Management Tasks and Procedures

Table 18 lists the tasks and procedures for managing Windows Time Service.
Table 18 Windows Time Service Management Tasks and Procedures
Tasks Procedures Tools Frequency
Configure a time  Configure time on the Net time As needed
source for the forest. forest-root PDC emulator.
 Remove a time source
configure on the forest-root
PDC emulator.
Configure a reliable  Configure the selected Regedit.exe As needed
time source on a computer as a reliable time
computer other than source.
the PDC emulator.
Configure a client to  Set a manually configured Net time As needed
request time from a time source on a selected
 Contents 103

specific time source. computer.

 Remove a manually
configured time source on a
selected computer.
Optimize the polling  Change polling interval. W32tm.exe As needed
interval. Regedit.exe
Disable the Windows  Disable time service. Active As needed
Time Service. Directory Sites
and Services

Configuring a Time Source for the Forest

After initial deployment of your network, you typically only reconfigure the time service on the
PDC emulator in two situations:
 If you move the PDC emulator role to a different computer. In this case, you must
configure the time source on the new role holder.
 If you change your time source. For example, if you change from synchronizing with
an external source to an internal hardware device.

Procedures for Configuring Time on the Forest-Root PDC Emulator

To configure time service for the forest-root PDC emulator, you might need to remove an
external time source that you used previously, or, if you transferred that operations master role,
you might only need to configure the time service on the new PDC emulator. To configure time
on the forest-root PDC emulator, you can use the following procedures. Procedures are explained
in detail in the linked topics.
1. Configure time on the forest-root PDC emulator.
2. Remove a time source configured on the forest-root PDC emulator.

Configuring a Reliable Time Source on a Computer Other than

the PDC Emulator
By default, the PDC emulator in the forest root is the authoritative time source for that forest.
However, you might want to configure a different computer in your network to be authoritative
for the forest, in the following situations:
104 Contents

Note
Caution
The registry
Setting a computer
editor bypasses
that is already
standard
synchronizing
safeguards,from
allowing
the domain
settingshierarchy
that can as a
reliable time
damage your source
system,canor create
even require
loops in
you
thetosynchronization
reinstall Windows.
tree If
andyou
cause
must edit
unpredictable
the registry, back up system state first. For information about backing up
results.
system state, see "Active Directory Backup and Restore" in this guide.

 If you plan to move the PDC Operations Master role, you can configure a reliable
time source on a different computer prior to the move(s) to avoid resets or disruption
of the time service. The role of PDC emulator can move between computers, which
means that every time the role of PDC emulator moves, the new PDC emulator must
be manually configured to point to the external source, and the manual configuration
must be removed from the original PDC emulator. To avoid this process, you can set
one of the domain controllers in the parent domain as reliable and manually configure
just that computer to point to an external source. Then, no matter which computer is
the PDC emulator, the root of the time service stays the same and thus remains
properly configured.
 If you have security reasons for wanting to segregate the authoritative time computer.
When domain controllers look for a time source to synchronize with, they choose a reliable
source if one is available. It is important to note that the automatic discovery mechanism in the
time service client never chooses a computer that is not a domain controller. Clients must be
manually configured to use any server that is not a domain controller.

Procedure for Configuring a Reliable Time Source on a Computer Other

than the PDC Emulator
Although the PDC emulator in the forest root domain is the authoritative time source for that
forest, you can configure a reliable time source on a computer other than the PDC emulator.
 Configure the selected computer as a reliable time source .
 Contents 105

Note
Manually specified time sources are not authenticated, and therefore can enable
an attacker to manipulate the time source and then start Kerberos V5 replay
attacks. Also, a computer that does not synchronize with its domain controller
can have an unsynchronized time. This causes Kerberos V5 authentication to
fail, which in turn causes other actions requiring network authentication, such as
printing or file sharing, to fail. When only one computer in the forest root
domain is getting time from an external source, all computers within the forest
remain synchronized to each other, making replay attacks difficult.

Configuring a Client to Request Time from a Specific Time

Source
Certain computers do not automatically synchronize their time through the Windows 2000 time
service hierarchy, so you might want to configure these clients to request time from a specific
source. If you do not specify a source, each computer’s internal hardware clock governs its time.
The following client computers do not automatically synchronize through the time service:
 Client computers that run Windows NT 4.0
 Client computers that run UNIX
 Computers that are not members of a domain

Procedures for Configuring a Client to Request Time from a Specific Time

Source
The following procedures allow you to specify a time source for client computers that do not
automatically synchronize through the time service. Procedures are explained in detail in the
linked topics.
1. Set a manually configured time source on a selected computer.
2. Remove a manually configured time source on a selected computer.

Optimizing the Polling Interval

By default, the time service synchronizes once every 45 minutes until three successful
synchronizations occur, then once every eight hours. You might want to change this interval in
the following situations:
 If computers are polling over a paid line, you can increase the polling interval. By
polling less often, you will decrease usage of the paid line.
 If you have applications or devices that require increased time accuracy, you can
decrease the polling interval.

Procedure for Optimizing the Polling Interval

You only need to perform one procedure to disable the Windows Time service.
 Change polling interval.
106 Contents

Caution
The registry editor bypasses standard safeguards, allowing settings that can
damage your system, or even require you to reinstall Windows. If you must edit
the registry, back up system state first. For information about backing up
system state, see "Active Directory Backup and Restore" in this guide.

Disabling the Windows Time Service

If you choose to implement another time synchronization product that uses the NTP protocol,
you must disable the W32Time time service because all NTP servers need access to UDP port
123. If W32Time is running on a Windows 2000–based computer, port 123 remains occupied.

Procedure for disabling the Windows Time service

You only need to perform one procedure to disable the Windows Time service.
 Disable time service.

Managing Long-Disconnected Domain

Controllers
A disconnected domain controller is a domain controller that is not replicating. Domain
controllers can become disconnected deliberately or inadvertently. Short-term disconnections are
not problematic because Active Directory replication automatically updates domain controllers
with all changes that they have not received. However, if a domain controller must be separated
from the replication topology for several weeks, you can take preliminary steps to ensure a
smooth reconnection.
For example, when domain controllers must be moved long distances or are pre-staged and
possibly stored for a period of time prior to being shipped to a destination, you must prepare
them to ensure that no gaps occur in operations master coverage during the disconnection and
that SYSVOL is updated when you reconnect the domain controller. If you plan to disconnect a
domain controller for longer than a domain controller keeps track of object deletions, you must
 Contents 107

take additional steps to ensure directory consistency, as described in “Preparing a Domain

Controller for a Long Disconnection” later in this section.
By monitoring replication, you can detect disconnections that occur due to network failures,
service failures, or configuration errors. For information about implementing monitoring for
replication failures, see “Monitoring Active Directory” earlier in this guide.

Operations Master Considerations

If a domain controller holds an operations master role, you must transfer the role prior to
disconnecting the domain controller. Normal directory functioning depends on all roles being
active, so when you plan to disconnect the domain controller, you must first transfer any
operations master roles. Role transfer ensures that no gaps in master operations coverage occur,
which can cause directory inconsistencies. For information about transferring operations master
roles, see “Managing Operations Masters” earlier in this guide.

Active Directory Replication Considerations

Ensure that the domain controller is updated before you disconnect it. Immediately prior to
disconnecting the domain controller, force replication with all replication partners and verify that
each directory partition replicates to the domain controller that you are disconnecting. If
replication of any directory partition does not succeed, resolve the replication problem prior to
disconnecting. By ensuring that replication is up-to-date, you can maximize the possible safe
disconnection period, which cannot exceed the tombstone lifetime for the forest. For information
about estimating the maximum safe disconnection period, see “Preparing a Domain Controller
for a Long Disconnection” later in this guide.

Tombstone Lifetime and Backup Considerations

Active Directory backups are useful for recovering a domain controller for only as long as the
tombstone lifetime. When an object is deleted, Active Directory replicates the object as a
tombstone, which consists of a small subset of the attributes of the deleted object. The tombstone
is retained in Active Directory for 60 days by default, after which it is permanently removed.
Because a domain controller that is disconnected for longer than the tombstone lifetime cannot
receive deletions that occurred prior to the beginning of the tombstone lifetime, a backup that is
older than the tombstone lifetime cannot be used to restore Active Directory.
When conditions beyond your control cause a domain controller to be disconnected longer than
the tombstone lifetime, one or more objects that have been deleted from the rest of the directory
while the domain controller was offline might remain on the disconnected domain controller. The
best practice recommendation for reconciling this condition of inconsistency is to reinstall
Windows on the outdated domain controller and then reinstall Active Directory. Otherwise, the
outdated domain controller can potentially reintroduce (reanimate) objects into Active Directory
that were deleted while the domain controller was disconnected. For information about how
objects become reanimated in Active Directory, see “Reconnecting Long-Disconnected Domain
Controllers” later in this guide.
If planned domain controller disconnections are consistently lasting longer than 60 days, alert the
design team and consider extending the tombstone lifetime for the forest.
108 Contents

SYSVOL Consistency Considerations

SYSVOL is a file system folder that stores files that must be available and synchronized among
all domain controllers. SYSVOL contains the NETLOGON share, Group Policy settings, and
File Replication service (FRS) staging directories and files. SYSVOL is required for Active
Directory to function properly.
SYSVOL is replicated by the File Replication service (FRS). FRS has a fixed tombstone lifetime
of 60 days. Because you cannot change this interval, any domain controller that is disconnected
for more than 60 days potentially has an outdated SYSVOL. Updating SYSVOL requires
performing a non-authoritative restore of SYSVOL.
In addition, SYSVOL replication cannot be synchronized manually. For this reason, ensuring that
SYSVOL is updated prior to disconnecting the domain controller is more difficult than simply
updating SYSVOL when the domain controller is reconnected. Regardless of the length of the
disconnection, to ensure that SYSVOL is synchronized when the domain controller is
reconnected, prepare the domain controller to perform a non-authoritative restore of SYSVOL
prior to disconnecting it. When it restarts, non-authoritative restore of SYSVOL occurs
automatically. For information about performing non-authoritative restore of SYSVOL, see
“Restoring and Rebuilding SYSVOL” earlier in this guide.

Windows 2000 Server with SP3

Windows 2000 Server with Service Pack 3 (SP3) provides the ability to force strict replication
consistency, which prevents outdated domain controllers from reintroducing objects that no
longer exist in Active Directory. When deploying new domain controllers that are running
Windows 2000 Server SP3, modify the registry to enforce strict replication consistency. For
information about strict replication consistency, see “Removing Lingering Objects from an
Outdated Writable Domain Controller” in this guide. For information about installing domain
controllers, see “Installing and Removing Active Directory” earlier in this guide.

Best Practice Recommendations for Managing Long Disconnections

If you must disconnect a domain controller for a period of several weeks or months, follow these
recommendations:
 Prior to disconnecting, determine the maximum length of time that the domain
controller will be disconnected and subtract a generous estimate of the end-to-end
replication latency. This amount of time is the maximum period for which the domain
controller can safely be disconnected.
 Prior to disconnecting, determine the value of the tombstone lifetime for the forest. If
you estimate the maximum safe time of disconnection to be longer than the
tombstone lifetime, contact a supervisor. The design team must determine whether to
extend the tombstone lifetime or rebuild the domain controller prior to reconnecting
it.
 Prior to disconnecting, prepare the registry for automatic non-authoritative restore of
SYSVOL when the domain controller restarts.
 Contents 109

 Immediately prior to disconnecting, ensure that the domain controller replicates

successfully with all replication partners.
 When you disconnect the domain controller, attach a label to the computer that
identifies the date and time of disconnection.
 When reconnecting the domain controller, if the site contains no other domain
controller that is authoritative for the domain, time the restart of the domain
controller to coincide with the beginning of intersite replication to restore SYSVOL
as quickly as possible. If the site has one or more other domain controllers that are
authoritative for the domain, start the domain controller at any time.
 If a domain controller has been disconnected for longer than the maximum safe time
of disconnection (tombstone lifetime less end-to-end replication latency), do not
allow the domain controller to replicate. Reinstall Windows 2000 Server. This
recommendation applies to all such domain controllers, regardless of the version of
Windows 2000 Server they are running (SP3, SP2, or earlier).
 If you deploy Windows 2000 Server SP3, modify the registry to enforce strict
replication behavior at the time the domain controller is installed.
110 Contents

Tasks and Procedures for Managing Long-Disconnected Domain Controllers

Table 19 shows the tasks and procedures for managing long disconnected domain controllers,
including tasks that address removing lingering objects.
Table 19 Tasks and Procedures for Managing Long-Disconnected Domain Controllers
Tasks Procedures Tools Frequency
Prepare a  Determine the anticipated length of  ADSI Edit As needed
domain the disconnection.  Active
controller for  Determine the tombstone lifetime Directory
long for the forest. Sites and
disconnection. Services
 Determine the maximum safe
disconnection time and proceed as  Repadmin.exe
follows:  Regedit.exe
 If the estimated time of  Active
disconnection exceeds the Directory
maximum safe disconnection Domains and
time, do not proceed with the Trusts
disconnection. Contact a
 Active
supervisor.
Directory
 If the estimated time of Users and
disconnection does not exceed Computers
the maximum safe
disconnection time, proceed
with disconnection.
 View the current operations master
role holders.
 Transfer domain-level operations
master roles, if appropriate.
 Transfer forest-level operations
master roles, if appropriate.
 Prepare the domain controller for
non-authoritative SYSVOL restore.
 Synchronize replication from all
inbound (source) replication
partners.
 Verify successful replication to the
domain controller.
 Label the domain controller with the
date and time of disconnection and
the maximum safe disconnection
period.
 Contents 111

Reconnect a  Determine the tombstone lifetime  ADSI Edit As needed

long- for the forest.  Active
disconnected  Determine whether the maximum Directory
domain safe disconnection time has been Sites and
controller. exceeded, and proceed accordingly. Services
 If the maximum safe time has  Repadmin.exe
been exceeded, do not connect
the domain controller. Contact a
supervisor about reinstalling the
domain controller.
 If the maximum safe time has
not been exceeded, proceed with
reconnecting.
 If the site has one or more other
domain controllers that are
authoritative for the domain, start
the domain controller at any time.
 If domain updates are available only
from a different site:
 Determine when intersite
replication is scheduled to
begin.
 As soon as possible after the
next replication cycle begins,
start the domain controller.
 Verify successful replication on the
reconnected domain controller.
112 Contents

Remove Windows 2000 Server with SP2:  Event Viewer As needed

lingering objects  Identify a revived lingering object  Active
from an outdated and replication source on a writable Directory
writable domain domain controller. Sites and
controller. Services
 Disable outbound replication on the
outdated source domain controller.  Repadmin.exe
 Delete the object from the outdated  Dsastat.exe
source domain controller.  Active
Windows 2000 Server with SP3: Directory
 Identify and delete a known non- Users and
replicated lingering object on an Computers
outdated domain controller.
Windows 2000 Server with SP2 or
SP3, continue as follows:
 Identify unknown lingering objects
on an outdated domain controller.
 View replication metadata of the
objects.
 Delete objects created prior to
domain controller disconnection.
 Restart disabled outbound
replication (SP2 only).
 Synchronize replication from the
outdated domain controller to a
replication partner.
Remove Windows 2000 Server with SP2: Ldp.exe As needed
lingering objects  Contact Microsoft Product Support
from a global Services.
catalog server.
Windows 2000 Server with SP3:
 Establish the distinguished name
and Globally Unique Identifier
(GUID) of the object.
 Identify the GUID of a domain
controller that has a writable replica
of the domain.
 Delete the lingering object from the
global catalog server.
 Contents 113

Preparing a Domain Controller for a Long Disconnection

When you need to take a domain controller offline for a prolonged period, prepare the domain
controller by doing the following:
 Establish the maximum safe disconnection period. Determine the tombstone lifetime
interval and subtract a generous estimate of the end-to-end replication latency to
establish the maximum safe period of disconnection. Otherwise, even when the
domain controller is reconnected prior to the end of the tombstone lifetime, a
tombstone can potentially expire before reaching the reconnected domain controller.
 Verify replication success on the domain controller prior to disconnecting it. If
replication is not successful, troubleshoot and fix the problem prior to disconnecting
the domain controller.
 Modify the registry to prepare the domain controller to perform a non-authoritative
restore of SYSVOL when it restarts. SYSVOL inconsistencies are not easily
verifiable prior to disconnecting. Therefore, by setting the registry to restore
SYSVOL when the domain controller restarts, you can ensure that SYSVOL
reinitializes its membership in the replica set and updates its content at the earliest
opportunity after reconnecting the domain controller.
 When modifying the registry to restore SYSVOL, consider the following:
 If SYSVOL is the only replica set that is represented on the domain controller, modify
the global BurFlags registry entry.
 If other replica sets are represented on the domain controller and you want to update
only SYSVOL, modify the replica-set-specific BurFlags registry entry for SYSVOL.
For information about restoring SYSVOL, see “Restoring and Rebuilding SYSVOL” earlier
in this guide.
 Determine whether the domain controller holds an operations master role. If the
domain controller is an operations master, transfer the role prior to disconnecting. For
information about transferring operations master roles, see “Managing Operations
Masters” earlier in this guide.
If the length of the disconnection is predicted to be longer than the current tombstone lifetime,
consult the design team about extending the tombstone lifetime.

Procedures for Preparing a Domain Controller for Long Disconnection

Perform the following procedures prior to disconnecting a domain controller. Procedures are
explained in detail in the linked topics.
1. Determine the anticipated length of the disconnection.
2. Determine the tombstone lifetime for the forest.
3. Determine the maximum safe disconnection period by subtracting a generous
estimate of the end-to-end replication latency from the tombstone lifetime. Either find
114 Contents

Caution
The registry editor bypasses standard safeguards, allowing settings that can
damage your system, or even require you to reinstall Windows. If you must edit
the registry, back up system state first. For information about backing up system
state, see "Active Directory Backup and Restore" in this guide.

the latency estimate in the design documentation for your deployment, or request the
information from a member of the design or deployment team.
 If the anticipated time of disconnection exceeds the maximum safe disconnection
period, do not disconnect the domain controller. Contact a supervisor.
 If the estimated time of disconnection does not exceed the maximum safe disconnection
time, proceed with disconnection.
4. View the current operations master role holders to determine whether the domain
controller is an operations master role holder.
5. Transfer a domain-level operations master role, if appropriate.
6. Transfer a forest-level operations master role, if appropriate.
7. Prepare the domain controller for non-authoritative SYSVOL restore on the domain
controller that you are disconnecting. This process ensures an up-to-date SYSVOL
when the domain controller is restarted.
8. Synchronize replication from all inbound (source) replication partners. Each
connection object below the NTDS Settings object for the server you are
disconnecting represents an inbound replication partner.
9. Verify successful replication to the domain controller that you are disconnecting.
10. Label the domain controller with the date and time of disconnection and the
maximum safe disconnection period.

Reconnecting Long-Disconnected Domain Controllers

Assuming that the domain controller has not been disconnected for longer than the maximum
safe period of disconnection (tombstone lifetime minus end-to-end replication latency),
reconnecting it to the replication topology requires no special procedures. By default, the
 Contents 115

Knowledge Consistency Checker (KCC) on a domain controller runs 5 minutes after the domain
controller starts, automatically incorporating the reconnected domain controller into the
replication topology.
If you plan appropriately for disconnecting and reconnecting domain controllers, no domain
controller will be disconnected from the replication topology for longer than a tombstone
lifetime. However, if unexpected events result in a domain controller becoming outdated, do not
reconnect the domain controller. Do not attempt to remove Active Directory because this process
requires replication. To ensure directory consistency, reinstall Windows 2000 Server on the
outdated domain controller. For information about how to reinstall a domain controller that has
not replicated for longer than a tombstone lifetime, see “Recovering a Domain Controller
Through Reinstallation.”
By monitoring replication, you avoid unexpected lengthy disconnections of domain controllers.
For information about monitoring replication, see “Monitoring Active Directory” in this guide.

Long Disconnections and Tombstone Lifetime

If a domain controller remains disconnected for longer than a tombstone lifetime, an object that
has been deleted from the directory can remain on the disconnected domain controller. For this
reason, such objects are called “lingering objects.”
Lingering objects can occur in the following circumstances:
 A domain controller goes offline immediately prior to the deletion of an object on
another domain controller and remains offline for:
 A period that exceeds the tombstone lifetime.
 A period that is less than the tombstone lifetime, but replication latency exceeds the
remaining duration of the tombstone lifetime.
 A domain controller goes offline following the deletion of an object on another
domain controller but prior to receiving replication of the tombstone, and remains
offline for a period that exceeds the tombstone lifetime.
 A domain controller goes offline, an object is deleted on that domain controller, and
the object tombstone is removed by garbage collection on that domain controller
prior to the domain controller being reconnected to replication.
In the latter case, an object exists on all domain controllers in the domain (for a domain-specific
object) or forest (for a configuration or schema object) except the reconnected domain controller.
In this case, the remedy is simply to delete the object on any writable domain controller.
However, in the first two cases, if the domain controller is then reconnected to the replication
topology, objects that exist nowhere else in the forest remain on the domain controller and
potentially can be reintroduced into the directory.
If lingering objects are security principals, reintroducing them can have serious consequences.
For more information about how lingering objects are reintroduced into the directory and how to
remove them, see “Removing Lingering Objects from an Outdated Writable Domain Controller.”
116 Contents

Important
Do not use file copy utilities such as xcopy or robocopy to update an outdated
SYSVOL.

Best Practice Recommendations for Avoiding Lingering Objects

Take the following precautions to ensure that lingering objects do not occur:
 Monitor the KCC topology and replication to ensure that long disconnections are
detected. For information about monitoring the KCC and replication, see “Monitoring
Active Directory” earlier in this guide.
 Ensure that the tombstone lifetime is not lowered below the default of 60 days.
 If you know that a domain controller will be offline for longer than the tombstone
lifetime, consult the design team about increasing the tombstone lifetime to a period
that safely encompasses the offline duration plus a generous period of replication
latency.
 Install Windows 2000 Server SP3 as soon as possible and enable strict replication
consistency to ensure that lingering objects cannot replicate.

Long Disconnections and SYSVOL

If the tombstone lifetime has been extended to longer than 60 days, SYSVOL will be outdated
when you reconnect the domain controller. The recommended practice is to prepare a domain
controller for a long disconnection by modifying the registry so that SYSVOL is restored
automatically when the domain controller is restarted. To update SYSVOL as soon as possible
after reconnecting, plan the time that you restart the domain controller to optimize the replication
schedule, as follows:
 If the closest replication partner for the domain is in a different site, view site link
properties to determine the schedule and then restart the domain controller as soon as
possible after the schedule opens.
 If a replication partner for the domain is available within the site, verify replication
success on that partner prior to restarting the domain controller.
In the event that a domain controller has been disconnected for a tombstone lifetime or
longer but has already replicated, follow the instructions for detecting and removing lingering
objects in “.Removing Lingering Objects from an Outdated Writable Domain Controller.”

Procedures for Reconnecting a Long-Disconnected Domain Controller

Follow these procedures to reconnect the domain controller. Procedures are explained in detail in
the linked topics.
1. Determine the tombstone lifetime for the forest.
2. Determine whether the maximum safe disconnection time has been exceeded, and
proceed accordingly:
 Contents 117

a. If the domain controller has been disconnected for a period that exceeds the
maximum safe disconnection period, do not reconnect the domain controller.
Contact a supervisor about reinstalling the domain controller.
b. If the maximum safe time has not been exceeded, proceed with reconnecting.
3. If the site in which you are reconnecting the domain controller has one or more other
domain controllers that are authoritative for the domain, start the domain controller at
any time.
4. If the site in which you are reconnecting the domain controller has no other domain
controllers that are authoritative for the domain, proceed as follows:
a. Determine when the next intersite replication cycle is scheduled to begin by
viewing the replication properties on the site link that connects this site to the
next closest site that includes domain controllers for this domain.
b. As soon as possible after the next replication cycle begins, start the domain controller.
5. After replication is complete, verify successful replication to the domain controller (the
reconnected domain controller) of the domain, configuration, and schema directory
partitions. If the domain controller is a global catalog server, check for successful replication
of all domain directory partitions.
In the event that a domain controller has been disconnected for a tombstone lifetime or longer but
has already replicated, follow the instructions for detecting and removing lingering objects in
“Removing Lingering Objects from an Outdated Writable Domain Controller.”

Removing Lingering Objects from an Outdated Writable

Domain Controller
If a domain controller does not replicate for a period that is longer than the tombstone lifetime
and the domain controller is then reintroduced into the replication topology, objects that have
been deleted from Active Directory while the domain controller was offline can remain on the
domain controller as lingering objects.

Causes for Lingering Objects

Lingering objects can occur whenever a domain controller does not replicate for a period that
exceeds the tombstone lifetime. Unexpectedly long disconnections can be caused by the
following conditions:
 A domain controller is left in a storage room and forgotten, or shipment of the pre-
staged domain controller to its remote location takes longer than a tombstone
lifetime.
 Replication fails and monitoring is not in place. For example, if a bridgehead server
is overloaded, replication can become backlogged indefinitely.
 WAN connections are unavailable for long periods. For example, a domain controller
on board a cruise ship might be unable to replicate because the ship is at sea for
longer than the tombstone lifetime.
118 Contents

 Garbage collection tampering. For example:

 Someone changes the time on a domain controller to force garbage collection.
 Someone changes the tombstone lifetime to force garbage collection.

Indications that a Domain Controller has Lingering Objects

An outdated domain controller can store lingering objects with no noticeable effect as long as no
one updates the lingering object or tries to create an object with the same name in the domain or
the same user principal name in the forest. However, the existence of lingering objects can cause
problems, especially if the object is a security principal.
The following conditions indicate that a domain controller has lingering objects:
 A deleted user or group account does not disappear from the Global Address List on
Exchange servers. Therefore, although the account name appears in the list, attempts
to send mail result in errors.
 E-mail messages are not delivered to a user whose user object was moved between
domains. After an outdated domain controller or global catalog server becomes
reconnected, both instances of the user object appear in the global catalog. Both
objects have the same e-mail address, so e-mail messages cannot be delivered.
 A universal group that no longer exists still appears in a user's access token. Although
the group no longer exists, if a user account still has the group in its security token,
the user might have access to a resource that you intended to be unavailable to that
user.
 A new object or Exchange mailbox cannot be created when the samAccountName
attribute value of the new object is the same as a lingering object. An error reports
that the object already exists.
 Replication succeeds with “no such object” error (event ID 1388) when “loose
replication consistency” is in effect. This error indicates that the source domain
controller revived a lingering object in the directory.
 Replication fails with a “no such object” error (event ID 1084) when “strict
replication consistency” is in effect. This error indicates that the source domain
controller tried to replicate a lingering object.

Replication of Lingering Objects

If a user updates a lingering object on the outdated domain controller, the destination domain
controller that receives the request for the update cannot update the object because the object
does not exist. The destination domain controller logs an NTDS Replication error in the
Directory Service log in Event Viewer. The error that is reported depends on the type of
replication consistency that is in effect on the domain controller.
The replication response differs on domain controllers that use loose replication consistency and
domain controllers that use strict replication consistency. On domain controllers that use loose
replication consistency (the default behavior with Windows 2000 Server SP2), the destination
domain controller requests a full copy of the object from the replication source. If the object is
 Contents 119

being modified, the destination requests the full object and the object is revived in the directory.
If the object is being deleted, the destination replicates the tombstone. In either case, the NTDS
Replication event ID 1388 is logged in the Directory Service log by the destination. The error
reports that the object being updated does not exist and the domain controller does not have
enough information to create it, and so it will request a complete copy. This error alerts you to
the fact that you have at least one lingering object and gives you the information that you need in
order to locate that object and delete it if it has been revived. Deleting the revived object on a
writable domain controller removes it from the directory
Domain controllers on which strict replication consistency is enabled (configurable behavior with
Windows 2000 Server SP3) refuse replication from the outdated replication source. This action
stops replication from the outdated source and logs NTDS Replication event ID 1084 in the
Directory Service log. The error reports that the object cannot be updated and replication will not
be accepted from the source until the issue is resolved. The information in the error includes the
name, GUID, and source of the lingering object so that you can delete the object and determine
whether additional lingering objects exist on the source. For this error to be logged, however, you
must have modified the registry to implement strict replication consistency.
In both cases, you can delete the identified lingering object and then take steps to identify and
remove all additional lingering objects from the outdated domain controller.

Sequence for Removing Lingering Objects

The process for removing lingering objects from an outdated writable domain controller involves
several procedures that must be performed in sequence. After an error indicates the existence of a
lingering object, use the following general sequence to remove the lingering object and determine
whether there are other lingering objects on the source domain controller:
 Identify the domain controller that replicated the update to a lingering object. Use the
information in event ID 1388 (Windows 2000 Server with SP2) or event ID 1084
(Windows 2000 Server with SP3) to identify the source domain controller.
 Disable outbound replication on the source domain controller.
 Delete the lingering object from the source domain controller.
 Compare the database contents of the outdated source domain controller and an up-
to-date replication partner to determine whether the outdated source domain
controller contains objects that do not exist on its replication partner.
 Identify the distinguished names of the objects that exist on the outdated domain
controller but not on the replication partner.
 Examine metadata of the object to determine when it was created.
 Delete the objects that were created prior to disconnecting the domain controller.
 Restart outbound replication on the source domain controller.
Deletions of the lingering objects replicate to the other domain controllers. Any domain
controller that is running Windows 2000 Server with SP2, and that does not have the object, logs
event ID 1388. In this case, the missing object is revived as a tombstone, and replicates as such.
120 Contents

The errors on domain controllers that do not have the object can be ignored; they will cease after
the second replication cycle.
If you have domain controllers that are running Windows 2000 Server with SP3, you can set the
registry to enforce strict replication consistency, which ensures that lingering objects do not
replicate. For this reason, attempted replication of the deletions will not be accepted. You must
delete lingering objects from only the outdated domain controller. For information about setting
strict replication consistency for domain controllers that are running Windows 2000 Server with
SP3, see “Managing Active Directory Installation and Removal” in this guide.

Procedures for Removing Lingering Objects from an Outdated Writable Domain

Controller
Use the following process to identify and remove lingering objects after you have discovered an
outdated domain controller. The initial step in the process varies according to the version of
Windows 2000 Server that you are using. Procedures are explained in detail in the linked topics.
1. Identify and delete the initial occurrence of a lingering object, as follows:
For Windows 2000 Server with SP2:
a. Identify a revived lingering object and its replication source on a writable
domain controller. Event ID 1388 provides the distinguished name of an object
that has been updated on an outdated domain controller. The message also
provides the GUID of the domain controller from which the update was
replicated. Use the GUID to discover the name of the source domain controller.
Repeat this process on each source domain controller until you identify a source
domain controller that does not have the error. This domain controller is the
outdated source domain controller.
b. Disable outbound replication on the outdated source domain controller.
c. Delete the object from the outdated source domain controller.
For Windows 2000 Server with SP3:
 Identify and delete a known non-replicated lingering object on an outdated domain
controller, as identified in event ID 1084. The object and source domain controller are
named in the error message.
2. Identify unknown lingering objects on an outdated domain controller. This procedure
requires the following series of subprocedures to be performed sequentially:
a. Compare the directory databases of the outdated domain controller and the
domain controller that received the initial replication error.
b. Identify the distinguished names of the objects that exist on the outdated domain
controller but not on the partner domain controller.
 Contents 121

Note
The results of this procedure identify only objects where the numbers of objects
did not agree between domain controllers. If numbers match but an object of a
class was added on one domain controller and a different object of the same
class was deleted on the other, and these changes did not replicate, this test
cannot identify these inconsistent objects.

3. On the outdated domain controller, view the replication metadata of objects that
you identified in the previous procedure to determine whether they were created prior
to the time the domain controller was disconnected or were created during the time
that the domain controller was offline. If the newest date in the Org.Time/Date
column is older than the date on which the domain controller was disconnected, the
object is a lingering object.
4. On the outdated domain controller, delete the objects that were created prior to the
date and time that the domain controller was disconnected.
5. Restart disabled outbound replication on the outdated domain controller (SP2 only).
6. Synchronize replication from the outdated domain controller to the partner domain
controller to replicate the deletions. Use the connection object on the replication
partner that shows the name of the outdated domain controller in the From Server
column. This procedure results in error messages on domain controllers that do not
have the objects, but these messages can be ignored and will cease by the second
replication cycle.

Removing Lingering Objects from a Global Catalog Server

If you delete a lingering object on a writable domain controller, the object deletion replicates to
all writable domain controllers in the domain as well as to all global catalog servers. However, if
a global catalog server becomes outdated, lingering objects can potentially exist in a read-only
replica on the global catalog server and nowhere else, in which case you cannot delete the object
by the normal method. The recommended solution to this problem depends on the version of
Windows 2000 Server that is running on the outdated global catalog server:
 Windows 2000 Server with SP2: Contact Microsoft Product Support Services.
 Windows 2000 Server with SP3: Use Ldp.exe to identify and delete the object from
all global catalog servers that retain the object.

Causes for Lingering Objects on Global Catalog Servers

Excessively high replication load on a global catalog server, in combination with a short intersite
replication interval, can result in updates not being replicated. Global catalog servers replicate
read-only replicas of all domain directory partitions in the forest. The replication of read-only
replicas has a lower priority than the replication of writable replicas. In addition, global catalog
servers are often bridgehead servers, which adds to the replication load. If the replication load on
global catalog servers acting as bridgehead servers is too high due to an extremely short
replication interval, excessive numbers of concurrent outbound replication partners, or a
combination of both, the replication queue can become backlogged. If the condition persists,
122 Contents

read-only replicas can remain in the queue indefinitely. These conditions can result in lingering
objects on a global catalog server.
If replication of a read-only replica is stalled or the domain controller is disconnected for longer
than a tombstone lifetime, the deletion of an object from the corresponding writable directory
partition can potentially expire without ever reaching the global catalog server. In this case, the
only location of this object is in the read-only replica on the global catalog server.
As with writable domain controllers, a global catalog server that is not monitored for replication
can potentially become outdated. When appropriate monitoring is in place and sensible intersite
replication schedules are configured, global catalog servers are not susceptible to becoming
outdated. For information about monitoring replication, see “Monitoring Active Directory” in
this document. For information about scheduling replication, see “Managing Sites” in this
document.

Indications that Lingering Objects Exist on a Global Catalog Server

The following events indicate that a lingering objects exists on a global catalog server:
 A deleted user or group account does not disappear from the Global Address List on
Exchange servers.
 E-mail messages are not deliverable to a user whose Active Directory account
appears to be current.
 A new user account or Exchange mailbox cannot be created because the object
already exists, but you do not see the object in Active Directory.
 Searches that use attributes of an existing object find an object of the same name that
has been deleted from the domain but remains in an isolated global catalog server.

Sequence for Removing Lingering Objects from a Global Catalog Server

To remove a lingering object from a global catalog server, you need an attribute value to use for
the search to identify the object in the global catalog. For example, when you are trying to create
a mailbox, user account, or other object in Active Directory, and error messages indicate that the
object already exists, use the name of the object that you are trying to create. If you know that a
deleted group or user name appears in the Global Address List, use that name.
Use the following general sequence of tasks to locate and remove a lingering object from a
global catalog server:
 Use an LDAP search to establish the distinguished name and GUID of the duplicate
(lingering) object.
 Use the distinguished name to identify the domain of the object.
 Identify a writable domain controller for that domain.
 Identify the GUID of the writable domain controller.
 Delete the object from the global catalog server. This procedure requires the
preceding information.
 Repeat the previous steps for every object and global catalog server that is outdated.
 Contents 123

When deleting an object that has child objects, you must delete the child object first, then delete
the parent. You can tell from the distinguished name whether the object has parent objects.

Procedures for Removing a Lingering Object from a Global Catalog Server

Use the following procedures to identify and remove a read-only lingering object from a global
catalog server that is running Windows 2000 Server with SP3. Procedures are explained in detail
in the linked topics.
1. Establish the distinguished name and GUID of the object by searching the global
catalog on an attribute that can uniquely identify the object. From the distinguished
name, you can identify the domain by the DC= components.
2. Identify the GUID of a domain controller that has a writable replica of the domain of
the lingering object.
3. Delete the lingering object from the global catalog server. In this procedure, use the
GUID of the object and the GUID of the writable domain controller that you identify
in procedures 1 and 2.

Managing Trusts
Trusts require little management. Trust relationships between domains establish a trusted
communication path through which a computer in one domain can communicate with a computer
in the other domain. Trust relationships allow users in the trusted domain to access resources in
the trusting domain.
For example, where a one-way trust exists:
 A user who is logged on to the trusted domain can be authenticated to connect to a
resource server in the trusting domain.
 A user can use an account in the trusted domain to log on to the trusted domain from
a computer in the trusting domain.
 A user in the trusting domain can list trusted domain security principals and add them
to groups and access control lists (ACLs) on resources in the trusting domain.

General Guidelines for Trusts

When you create a Windows 2000 domain in an existing Windows 2000 forest, a trust
relationship is established automatically. These trust relationships are two-way and transitive,
and they should not be removed.
However, three types of trusts must be created manually:
 External trusts:
 Trusts between a Windows 2000 domain and a Windows NT 4.0 domain.
 Any trust between domains in different forests, whether both domains are
Windows 2000 or one is Windows 2000 and the other Windows NT 4.0.
124 Contents

 Shortcut trusts between two domains in the same forest.

 Trust relationships between a Windows 2000 domain and a non-Windows Kerberos
realm. For more information about trusts between a Windows 2000 domain and a
non-Windows Kerberos realm, see the Step-by-Step Guide to Kerberos 5 (krb5 1.0)
Interoperability link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.
You might also need to manage trusts for the following reasons:
 To remove a manually created trust.
 To configure security identifier (SID) filtering to deny one domain the right to
provide credentials for another domain. You can enable SID filtering for external
trusts, that is, trusts between domains in different forests, or between a
Windows 2000 and a Windows NT 4.0 domain.

Trust Management Tasks and Procedures

Table 20 shows the tasks and the procedures for managing trusts.
Table 20 Trust Management Tasks and Procedures
Tasks Procedures Tools Frequency
Create an external trust  Create a One-way  Active As needed
(between a Trust (MMC Directory
Windows 2000 domain Method). Domains and
and a Windows NT 4.0  Create a One-way Trusts
domain, or between Trust (Netdom.exe (Windows 2000
domains in different Method). )
forests). -Or-
 Create a Two-way
Trust (MMC  Netdom.exe
Method).
 Create a Two-way  User Manager
Trust (Netdom.exe for Domains
Method). (Windows NT 4
.0)
Create a shortcut trust.  Create a One-way  Active As needed
Trust (MMC Directory
Method). Domains and
 Create a One-way Trusts
Trust (Netdom.exe -Or-
Method).  Netdom.exe
 Create a Two-way
Trust (MMC
Method).
 Create a Two-way
Trust (Netdom.exe
Method).
 Contents 125

Remove a manually  Remove a manually  Active As needed

created trust. created trust. Directory
Domains and
Trusts
-Or-
 Netdom.exe
Prevent unauthorized  Configure SID  Netdom.exe As needed
privilege escalation. filtering.

Creating External Trusts

You create an external trust when you want to establish a trust relationship between
Windows 2000 domains that are in different forests, or between a Windows 2000 domain and a
Windows NT 4.0 domain. An external trust relationship has the following characteristics:
 It is one-way. The trust must be established manually in each direction to create a
two-way external trust relationship.
 It is nontransitive.
If you upgrade a Windows NT 4.0 domain to a Windows 2000 domain, the existing trust
relationships remain in the same state.

Methods for Creating the External Trust

 Use the procedure Create a One-way Trust - MMC Method to create a trust where
one domain trusts another to use its resources.
 Use the procedure Create a One-way Trust - Netdom.exe Method to use the
support tool Netdom.exe to create both sides of a one-way trust at once. You must
provide credentials for both domains to use the Netdom.exe method.
 Use the procedure Create a Two-way Trust - MMC Method first to create both
portions configured in one domain, and then to create both portions configured in the
other domain.
 Use the procedure Create a Two-way Trust - Netdom.exe Method to use the
support tool Netdom.exe to create both sides of the trust at once. You must provide
credentials for both domains to use the Netdom.exe method.
Requirements
 Credentials: Domain Admins
 You can create the trust when you log on to the domain, or use the Run As command
to create the trust for a different domain.
 Tools: Active Directory Domains and Trusts or Netdom.exe (Support Tools).
126 Contents

Procedures for Creating External Trusts

You can create an external trust by using one of the following methods. Procedures are explained
in detail in the linked topics.
1. Create a One-way Trust (MMC Method)
2. Create a One-way Trust (Netdom.exe Method)
3. Create a Two-way Trust (MMC Method)
4. Create a Two-way Trust (Netdom.exe Method)

Creating Shortcut Trusts

A shortcut trust relationship is a manually created trust that shortens the trust path to improve the
efficiency of users who remotely log on. A trust path is a chain of multiple trusts that enables
trust between domains that are not adjacent in the domain namespace. For example, if users in
domain A need to gain access to resources in domain C, you can create a direct link from
domain A to domain C through a shortcut trust relationship, bypassing domain B in the trust
path.
A shortcut trust relationship has the following characteristics:
 It can be established between any two domains in the same forest.
 It must be established manually in each direction.
 It is transitive.
Requirements
 Credentials: Domain Admins
 Tool: Active Directory Domains and Trusts

Procedures for Creating Shortcut Trusts

You can create a shortcut trust by using one of the following methods. Procedures are explained
in detail in the linked topics.
1. Create a One-way Trust (MMC Method)
2. Create a One-way Trust (Netdom.exe Method)
3. Create a Two-way Trust (MMC Method)
4. Create a Two-way Trust (Netdom.exe Method)

Removing Manually Created Trusts

You can remove manually created trusts, but you cannot remove the default two-way transitive
trusts between domains in a forest. It is particularly important to verify that you successfully
removed the trusts if you are planning to re-create them.
Requirements
 Contents 127

 Credentials: Domain Admins

 Tool: Active Directory Domains and Trusts or Netdom.exe.

Procedure for Removing Manually Created Trusts

You can remove a manually created trust by using one of the following methods. Procedures are
explained in detail in the linked topics.
1. Remove a manually created trust by using the Active Directory Domains and Trusts
snap-in.
2. Remove a manually created trust by using Netdom.exe.

Preventing Unauthorized Privilege Escalation

Security principals in Active Directory have an attribute called SIDHistory to which domain
administrators can add users' old SIDs. This is useful during the migration process because users
can use their old SIDs to access resources, administrators do not need to modify ACLs on large
numbers of resources. However, under some circumstances it is possible for domain
administrators to use the SIDHistory attribute to associate SIDs with new user accounts, thereby
granting themselves unauthorized rights.
You can configure SID filtering to prevent this type of attack. You might configure SID filtering
under the following circumstances:
 You have identified one or more domains in your enterprise where physical security
is lax, or where the domain administrators are less well trusted.
 You then isolate these less trustworthy domains by moving them to other forests. By
definition, all domains within a forest must be trustworthy; if a domain is deemed
less trustworthy than the others in the forest, it should not be a forest member. Once
you have moved less trustworthy domains out of the forest, establish external trusts to
these domains, and apply access control to protect resources. If you are still
concerned about SID spoofing being used for privilege escalation, then apply SID
filtering.
 Do not apply SID filtering to domains within a forest, as this removes SIDs required
for Active Directory replication, and causes authentication to fail for users from
domains that are transitively trusted through the isolated domain.

Procedure for Preventing Unauthorized Privilege Escalation

Use the following procedures to configure SID filtering. Procedures are explained in detail in the
linked topics.
1. Configure SID filtering.
2. Remove SID filtering.
128 Contents

Managing Sites
An Active Directory site object represents a collection of Internet Protocol (IP) subnets, usually
constituting a physical Local Area Network (LAN). Multiple sites are connected for replication
by site link objects.
Sites are used in Active Directory to:
 Enable clients to discover network resources (printers, published shares, domain
controllers) that are close to the physical location of the client, reducing network
traffic over Wide Area Network (WAN) links.
 Optimize replication between domain controllers.
Managing sites in Active Directory involves adding new subnet, site, and site link objects when
the network grows, as well as configuring a schedule and cost for site links. You can modify the
site link schedule, cost, or both, to optimize intersite replication. When conditions no longer
require replication to a site, you can remove the site and associated objects from Active
Directory.
Large hub-and-spoke topology management is beyond the scope of this documentation. For
information about managing Active Directory branch office deployments that include more than
200 sites, see the "Active Directory Branch Office Guide Series" at
http://www.microsoft.com/technet/win2000/win2ksrv/adguide/default.asp.
Using the SMTP intersite replication transport is beyond the scope of this documentation. For
information about SMTP replication, see "Active Directory Replication" in the Distributed
Systems Guide of the Microsoft Windows 2000 Server Resource Kit and see the "Step-by-Step
Guide to Setting up ISM-SMTP Replication." To download this guide, see the Active Directory
link on the Web Resources page at http://www.microsoft.com/windows/reskits/webresources.
Automatic site coverage is a default condition for Windows 2000 domain controllers. Operations
and guidelines documented in this guide are consistent with the enabling of automatic site
coverage.
The KCC and Replication Topology
The Knowledge Consistency Checker (KCC) uses site link configuration information to enable
and optimize replication traffic by generating a least-cost replication topology. Within a site, for
each directory partition, the KCC builds a ring topology that minimizes the number of hops
between domain controllers. Between sites, the KCC creates a spanning tree of all intersite
connections. Therefore, adding sites and domains increases the processing that is required by the
KCC. Before adding to the site topology, be sure to consider the guidelines discussed in “Adding
a New Site” later in this document.
Significant changes to site topology can affect domain controller hardware requirements. For
more information about domain controller hardware requirements, see “Domain Controller
Capacity Planning” in “Best Practice Active Directory Design for Managing Windows
Networks.” To download this guide, see the Active Directory link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.
 Contents 129

Bridgehead Server Selection

By default, bridgehead servers are automatically selected by the intersite topology generator
(ISTG) in each site. Alternatively, you can use Active Directory Sites and Services to select
preferred bridgehead servers. However, it is recommended for Windows 2000 deployments that
you do not select preferred bridgehead servers.
Selecting preferred bridgehead servers limits the bridgehead servers that the KCC can use to
those that you have selected. If you use Active Directory Sites and Services to select any
preferred bridgehead servers at all in a site, you must select as many as possible and you must
select them for all domains that must be replicated to a different site. If you select preferred
bridgehead servers for a domain and all preferred bridgehead servers for that domain become
unavailable, replication of that domain to and from that site does not occur.
If you have selected one or more bridgehead servers, removing them from the bridgehead servers
list restores the automatic selection functionality to the ISTG.

Site Management Tasks and Procedures

Table 21 shows the tasks and procedures for managing sites, as well as the tools and the
recommended frequency for performing each task. After you configure sites, subnets, and site
links for the initial deployment, most site management activity is limited to responding to
changes in network conditions.
Table 21 Site Management Tasks and Procedures
Tasks Procedures Tools Frequency
Add a new  Create a site object.  Active As needed
site.  Create a subnet object and Directory
associate it with the site. Sites and
Services
–or–
 Associate an existing subnet
object with the site.
 Create a site link object, if
appropriate.
 Remove the site from a site
link, if appropriate.
Add a subnet  Obtain the network address and  Active As needed
to the subnet mask for the subnet. Directory
network.  Create a subnet object and Sites and
associate it with a site. Services

Link sites for  Determine the names of the  Active As needed

replication. sites you are linking. Directory
 Create a site link object. Sites and
Services
 Determine the ISTG role owner
for a site.
 Generate the replication
130 Contents

topology on the ISTG, if

appropriate.
Change site  Configure the site link  Active As needed
link schedule. Directory
properties.  Configure the site link interval. Sites and
Services
 Configure the site link cost.
 Determine the ISTG role owner
for a site.
 Generate the replication
topology on the ISTG, if
appropriate.
Move a  Change the static IP address of  My Network As needed
domain the domain controller. Places
controller to a  Create a delegation for the  Active
different site. domain controller, if Directory
appropriate. Sites and
 Verify that the IP address maps Services
to a subnet and determine the  DNS snap-in
site association.
 Determine whether the server is
a preferred bridgehead server.
 Configure the domain
controller to not be a preferred
bridgehead server, if
appropriate.
 Move the server object to a
different site.
Remove a site.  Determine whether the server  Active As needed
object has child objects. Directory
 Delete the server object or Sites and
objects from the site. Services
 Delete the site link object, if
appropriate.
 Associate the subnet or subnets
with a different site.
–or–
 Delete the subnet objects.
 Delete the site object.
 Determine the ISTG role owner
for a site.
 Generate the replication
topology on the ISTG, if
appropriate.
 Contents 131

Adding a New Site

Design teams or network architects might want to add sites as part of ongoing deployment.
Although you typically create subnets to accommodate all address ranges in the network, you do
not need to create sites for every location. Generally, sites are required for those locations that
have domain controllers or other servers that run applications that depend on site topology, such
as Distributed File System (DFS). When such locations are separated from other network
locations by a WAN link, create a site object to optimize resource location, Active Directory
replication, and domain controller location for clients.
When the need for a site arises, the design team typically provides details about the placement
and configuration of site links for the new site, as well as subnet assignments or creation if
subnets are needed.
KCC calculations for generating the intersite topology for a Windows 2000 forest can cause
directory performance to suffer when the combined sites, site links, and domains exceed certain
limits. When these limits are reached, follow the site administration guidelines on the Active
Directory Branch Office Planning Guide link on the Web Resources page at
http://www.microsoft.com/windows/reskits/webresources.
As a general guideline, when any of the following conditions exist, consult your design team
before adding a new site:
 An existing site is directly connected to more than 20 sites.
 A bridgehead server has more than 20 inbound connections.
 The forest has 200 or more sites.

Procedures for Adding a New Site

Use the following procedures to add a new site. Procedures are explained in detail in the linked
topics.
1. Create a site object and add it to an existing site link.
2. Associate a range of IP addresses with the site, as follows:
 Create a subnet object or objects and associate them with the new site.
–or–
 Associate an existing subnet object with the new site.
3. Create a site link object, if appropriate, and add the new site and at least one other
site to the site link.
4. If, while performing procedure 1, you added the new site to an existing site link
temporarily in order to create the site, remove the site from that site link.
132 Contents

Adding a Subnet to the Network

If a new range of IP addresses is added to the network, create a subnet object in Active Directory
to correspond to the range of IP addresses. When you create a new subnet object, you must
associated it with a site object. You can either associate the subnet with an existing site, or create
a new site first and then create the subnet and associate it with the new site. If you are going to
create a new site for the new network segment, see “Adding a New Site.”

Procedures for Adding a Subnet

Use the following procedures to add a subnet. Procedures are explained in detail in the linked
topics.
1. Obtain the network address and subnet mask for the new subnet.
2. Create a subnet object and associate it with the appropriate site.

Linking Sites for Replication

To link sites for replication, create a site link object in the IP transport container and add two or
more sites to the link. Use a naming convention that includes the sites that you are linking. For
example, if you want to link the site named Seattle to the site named Boston, you might name the
site link SEA-BOS.
After you add two or more site names to a site link object, the bridgehead servers in the
respective sites replicate between the sites according to the replication schedule, cost, and
interval settings on the site link object. For information about modifying the default settings, see
“Changing Site Link Properties.”
At least two sites must exist when you create a site link. If you are adding a site link to connect a
new site to an existing site, create the new site first and then create the site link. For information
about creating a site, see “Adding a New Site.”

Procedures for Creating a Site Link

Use the following procedures to link sites for replication. Procedures are explained in detail in
the linked topics.
1. Determine the names of the sites you are linking.
2. Create a site link object in the IP container and add the appropriate sites to it.
3. Generate the intersite topology. By default, the KCC runs every 15 minutes to
generate the replication topology. To initiate replication topology generation
immediately, use the following procedures to refresh the intersite topology:
a. Determine the ISTG role owner for the site.
b. Generate the replication topology on the ISTG.
 Contents 133

Changing Site Link Properties

To control which sites replicate directly with each other and when, use the cost, schedule, and
interval properties on the site link object.
These settings control intersite replication as follows:
 Schedule: The time during which replication can occur (the default setting allows
replication at all times).
 Interval: The number of minutes between replication polling by intersite replication
partners within the open schedule window (default is every 180 minutes).
 Cost: The relative priority of the link (default is 100). Lower relative cost increases
the priority of the link over other higher-cost links.
Consult your design documentation for information about values to set for site link properties.

Procedures for Configuring Site Links

Use the following procedures to configure a site link. Procedures are explained in detail in the
linked topics.
1. Configure the site link schedule to identify times during which intersite replication
can occur.
2. Configure the site link interval to identify how often replication polling can occur
during the schedule window.
3. Configure the site link cost to establish a priority for replication routing.
4. Generate the intersite replication topology, if appropriate. By default, the KCC runs
every 15 minutes to generate the replication topology. To initiate intersite replication
topology generation immediately, use the following procedures to refresh the
topology:
a. Determine the ISTG role owner for the site.
b. Generate the replication topology on the ISTG.

Moving a Domain Controller to a Different Site

If you change the IP address or the subnet-to-site association of a domain controller after Active
Directory is installed on the server, the server object does not change sites automatically. You
must move it to the new site manually. When you move the server object, the Net Logon service
on the domain controller registers DNS SRV resource records for the appropriate site.

TCP/IP Settings
When you move a domain controller to a different site, if an IP address of the domain controller
is statically configured, then you must change the TCP/IP settings accordingly. The IP address of
the domain controller must map to a subnet object that is associated with the site to which you
are moving the domain controller. If the IP address of a domain controller does not match the site
134 Contents

in which the server object appears, the domain controller must communicate over a potentially
slow WAN link to locate resources rather than locating resources in its own site.
Prior to moving the domain controller, ensure that the following TCP/IP client values are
appropriate for the new location:
 IP address, including the subnet mask and default gateway.
 DNS server addresses.
 WINS server addresses (if appropriate).
If the domain controller that you are moving is a DNS server, you must also:
 Change the TCP/IP settings on any clients that have static references to the domain
controller as the preferred or alternate DNS server.
 Determine whether the parent DNS zone of any zone that is hosted by this DNS
server contains a delegation to this DNS server. If yes, update the IP address in all
such delegations. For information about creating DNS delegations, see "Performing
Active Directory Post-Installation Tasks."

Preferred Bridgehead Server Status

Before moving any server object, check the server object to see whether it is acting as a preferred
bridgehead server for the site. This condition has ISTG implications in both sites, as follows:
 Site to which you are moving the server: If you move a preferred bridgehead server
to a different site, it becomes a preferred bridgehead server in the new site. If
preferred bridgehead servers are not currently in use in this site, the ISTG behavior in
this site changes to support preferred bridgehead servers. For this reason, you must
either configure the server to not be a preferred bridgehead server (recommended), or
select additional preferred bridgehead servers in the site (not recommended).
 Site from which you are moving the server: If the server is the last preferred
bridgehead server in the original site for its domain, and if other domain controllers
for the domain are in the site, the ISTG selects a bridgehead server for the domain. If
you use preferred bridgehead servers, always select more than one server as preferred
bridgehead server for the domain. If after the removal of this domain controller from
the site multiple domain controllers remain that are hosting the same domain and
only one of them is configured as a preferred bridgehead server, either configure the
server to not be a preferred bridgehead server (recommended), or select additional
preferred bridgehead servers hosting the same domain in the site (not recommended).
 Contents 135

Note
If you select preferred bridgehead servers and all selected preferred bridgehead
servers for a domain are unavailable in the site, the ISTG does not select a new
bridgehead server. In this case, replication of this domain to and from other sites
does not occur. However, if no preferred bridgehead server is selected for a
domain or transport (through administrator error or as the result of moving the
only preferred bridgehead server to a different site), the ISTG automatically
selects a preferred bridgehead server for the domain and replication proceeds as
scheduled.

Procedures for Moving a Domain Controller to a Different Site

Use the following procedures to move a domain controller to a different site. Procedures are
explained in detail in the linked topics.
1. Change the static IP address of the domain controller. This procedure includes
changing all appropriate TCP/IP values, including preferred and alternate DNS
servers, as well as WINS servers (if appropriate). Obtain these values from the design
team.
2. Create a delegation for the domain controller, if appropriate. If the parent DNS zone
of any zone that is hosted by this DNS server contains a delegation to this DNS
server, use this procedure to update the IP address in all such delegations.
3. Verify that the IP address maps to a subnet and determine the site association to
ensure that the subnet is associated with the site to which you are moving the server
object.
4. Determine whether the server is a preferred bridgehead server .
5. If the server is a preferred bridgehead server in the current site and you do not want
the server to be a preferred bridgehead server in the new site, configure the server to
not be a preferred bridgehead server.
6. Move the server object to the new site.

Removing a Site
If domain controllers are no longer needed in a network location, you can remove them from the
site and then delete the site object. Before deleting the site, you must remove domain controllers
from the site either by removing it entirely or by moving it to a new location.
 To remove the domain controller, remove Active Directory from the server and
then delete the server object from the site in Active Directory. For information about
removing a domain controller, see “Decommissioning a Domain Controller.”
 To retain the domain controller in a different location, move the domain
controller to a different site and then move the server object to the respective site in
Active Directory. For information about moving a domain controller, see “Moving a
Domain Controller to a Different Site.”
136 Contents

Domain controllers can host other applications that depend on site topology and publish objects
as child objects of the respective server object. For example, when MOM or Message Queuing
are running on a domain controller, these applications create child objects beneath the server
object. In addition, a Message Queuing server that is not a domain controller and is configured to
be a Message Queuing Routing Server creates a server object in the Sites container. Removing
the application from the server automatically removes the child object below the respective
server object. However, the server object is not removed automatically.
When all applications have been removed from the server (no child objects appear beneath the
server object), you can remove the server object. After the application is removed from the
server, a replication cycle might be required before child objects are no longer visible below the
server object.
After you delete or move the server objects but before you delete the site object, reconcile the
following objects:
 Subnet object or objects for the site IP addresses:
 If the addresses are being reassigned to a different site, associate the subnet object or
objects with that site. Any clients using the addresses for the decommissioned site will
thereafter be assigned automatically to the other site.
 If the IP addresses will no longer be used on the network, delete the corresponding
subnet object or objects.
 Site link object or objects. You might need to delete a site link object, as follows:
 If the site you are removing is added to a site link containing only two sites, delete the
site link object.
 If the site you are removing is added to a site link that contains more than two sites, do
not delete this site link object.
Before deleting a site, obtain instructions from the design team for reconnecting any other sites
that might be disconnected from the topology by removing this site. If the site you are removing
is added to more than one site link, it might be an interim site between other sites that are added
to this site link. Deleting the site might disconnect the outer sites from each other. In this case,
the site links must be reconciled according to the instructions of the design team.

Procedures for Removing a Site

Use the following procedures to remove a site. Procedures are explained in detail in the linked
topics.
1. Determine whether the server object has child objects. If a child object appears, do
not delete the server object. If a domain controller has been decommissioned and one
or more child objects appears below the server object, replication might not have
completed. If replication has completed and child objects exist, do not delete the
server object. Contact a supervisor.
2. Delete the server objects within the Servers container of the site that you are
removing.
 Contents 137

3. Delete the site link object, if appropriate. Obtain this information from the design
team.
4. Associate the subnet or subnets with the appropriate site, if appropriate. If you no
longer want to use the IP addresses associated with the subnet object or objects,
delete the subnet objects. Obtain this information from the design team.
5. Delete the site object.
6. Generate the intersite replication topology, if appropriate. By default, the KCC runs
every 15 minutes to generate the replication topology. To initiate intersite replication
topology generation immediately, use the following procedures to refresh the
topology:
a. Determine the ISTG role owner in the site.
b. Generate the replication topology on the ISTG.
138 Contents