Professional Documents
Culture Documents
6423A
Implementing and Managing
Windows Server®°2008 Clustering
ii Implementing and Managing Windows Server® 2008 Clustering
Information in this document, including URL and other Internet Web site references, is subject to
change without notice. Unless otherwise noted, the example companies, organizations, products,
domain names, e-mail addresses, logos, people, places, and events depicted herein are fictitious,
and no association with any real company, organization, product, domain name, e-mail address,
logo, person, place or event is intended or should be inferred. Complying with all applicable
copyright laws is the responsibility of the user. Without limiting the rights under copyright, no part
of this document may be reproduced, stored in or introduced into a retrieval system, or transmitted
in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), or for
any purpose, without the express written permission of Microsoft Corporation.
Microsoft may have patents, patent applications, trademarks, copyrights, or other intellectual
property rights covering subject matter in this document. Except as expressly provided in any
written license agreement from Microsoft, the furnishing of this document does not give you any
license to these patents, trademarks, copyrights, or other intellectual property.
The names of manufacturers, products, or URLs are provided for informational purposes only and
Microsoft makes no representations and warranties, either expressed, implied, or statutory,
regarding these manufacturers or the use of the products with any Microsoft technologies. The
inclusion of a manufacturer or product does not imply endorsement of Microsoft of the
manufacturer or product. Links may be provided to third party sites. Such sites are not under the
control of Microsoft and Microsoft is not responsible for the contents of any linked site or any link
contained in a linked site, or any changes or updates to such sites. Microsoft is not responsible for
webcasting or any other form of transmission received from any linked site. Microsoft is providing
these links to you only as a convenience, and the inclusion of any link does not imply endorsement
of Microsoft of the site or the products contained therein.
© 2008 Microsoft Corporation. All rights reserved.
Microsoft, Active Directory, Hyper-V, Internet Explorer, Outlook, PowerPoint, SQL Server, Windows,
and Windows Vista are either registered trademarks or trademarks of Microsoft Corporation in the
United States and/or other countries.
If you comply with these license terms, you have the rights below.
1. DEFINITIONS.
a. “Academic Materials” means the printed or electronic documentation such as manuals,
workbooks, white papers, press releases, datasheets, and FAQs which may be included in the
Licensed Content.
b. “Authorized Learning Center(s)” means a Microsoft Certified Partner for Learning Solutions
location, an IT Academy location, or such other entity as Microsoft may designate from time to time.
c. “Authorized Training Session(s)” means those training sessions authorized by Microsoft and
conducted at or through Authorized Learning Centers by a Trainer providing training to Students
solely on Official Microsoft Learning Products (formerly known as Microsoft Official Curriculum or
“MOC”) and Microsoft Dynamics Learning Products (formerly know as Microsoft Business Solutions
Courseware). Each Authorized Training Session will provide training on the subject matter of one
(1) Course.
d. “Course” means one of the courses using Licensed Content offered by an Authorized Learning
Center during an Authorized Training Session, each of which provides training on a particular
Microsoft technology subject matter.
e. “Device(s)” means a single computer, device, workstation, terminal, or other digital electronic or
analog device.
f. “Licensed Content” means the materials accompanying these license terms. The Licensed
Content may include, but is not limited to, the following elements: (i) Trainer Content, (ii) Student
Content, (iii) classroom setup guide, and (iv) Software. There are different and separate
components of the Licensed Content for each Course.
g. “Software” means the Virtual Machines and Virtual Hard Disks, or other software applications that
may be included with the Licensed Content.
h. “Student(s)” means a student duly enrolled for an Authorized Training Session at your location.
i. “Student Content” means the learning materials accompanying these license terms that are for
use by Students and Trainers during an Authorized Training Session. Student Content may include
labs, simulations, and courseware files for a Course.
j. “Trainer(s)” means a) a person who is duly certified by Microsoft as a Microsoft Certified Trainer
and b) such other individual as authorized in writing by Microsoft and has been engaged by an
Authorized Learning Center to teach or instruct an Authorized Training Session to Students on its
behalf.
k. “Trainer Content” means the materials accompanying these license terms that are for use by
Trainers and Students, as applicable, solely during an Authorized Training Session. Trainer Content
may include Virtual Machines, Virtual Hard Disks, Microsoft PowerPoint files, instructor notes, and
demonstration guides and script files for a Course.
l. “Virtual Hard Disks” means Microsoft Software that is comprised of virtualized hard disks (such as
a base virtual hard disk or differencing disks) for a Virtual Machine that can be loaded onto a single
computer or other device in order to allow end-users to run multiple operating systems concurrently.
For the purposes of these license terms, Virtual Hard Disks will be considered “Trainer Content”.
m. “Virtual Machine” means a virtualized computing experience, created and accessed using
Microsoft® Virtual PC or Microsoft® Virtual Server software that consists of a virtualized hardware
environment, one or more Virtual Hard Disks, and a configuration file setting the parameters of the
virtualized hardware environment (e.g., RAM). For the purposes of these license terms, Virtual Hard
Disks will be considered “Trainer Content”.
n. “you” means the Authorized Learning Center or Trainer, as applicable, that has agreed to these
license terms.
2. OVERVIEW.
Licensed Content. The Licensed Content includes Software, Academic Materials (online and
electronic), Trainer Content, Student Content, classroom setup guide, and associated media.
License Model. The Licensed Content is licensed on a per copy per Authorized Learning Center
location or per Trainer basis.
3. INSTALLATION AND USE RIGHTS.
a. Authorized Learning Centers and Trainers: For each Authorized Training Session, you
may:
i. either install individual copies of the relevant Licensed Content on classroom Devices only for
use by Students enrolled in and the Trainer delivering the Authorized Training Session, provided
that the number of copies in use does not exceed the number of Students enrolled in and the
Trainer delivering the Authorized Training Session, OR
ii. install one copy of the relevant Licensed Content on a network server only for access by
classroom Devices and only for use by Students enrolled in and the Trainer delivering the
Authorized Training Session, provided that the number of Devices accessing the Licensed
Content on such server does not exceed the number of Students enrolled in and the Trainer
delivering the Authorized Training Session.
iii. and allow the Students enrolled in and the Trainer delivering the Authorized Training Session to
use the Licensed Content that you install in accordance with (ii) or (ii) above during such
Authorized Training Session in accordance with these license terms.
i. Separation of Components. The components of the Licensed Content are licensed as a single
unit. You may not separate the components and install them on different Devices.
ii. Third Party Programs. The Licensed Content may contain third party programs. These license
terms will apply to the use of those third party programs, unless other terms accompany those
programs.
b. Trainers:
i. Trainers may Use the Licensed Content that you install or that is installed by an Authorized
Learning Center on a classroom Device to deliver an Authorized Training Session.
ii. Trainers may also Use a copy of the Licensed Content as follows:
A. Licensed Device. The licensed Device is the Device on which you Use the Licensed Content.
You may install and Use one copy of the Licensed Content on the licensed Device solely for
your own personal training Use and for preparation of an Authorized Training Session.
B. Portable Device. You may install another copy on a portable device solely for your own
personal training Use and for preparation of an Authorized Training Session.
4. PRE-RELEASE VERSIONS. If this is a pre-release (“beta”) version, in addition to the other provisions
in this agreement, these terms also apply:
a. Pre-Release Licensed Content. This Licensed Content is a pre-release version. It may not
contain the same information and/or work the way a final version of the Licensed Content will. We
may change it for the final, commercial version. We also may not release a commercial version.
You will clearly and conspicuously inform any Students who participate in each Authorized Training
Session of the foregoing; and, that you or Microsoft are under no obligation to provide them with
any further content, including but not limited to the final released version of the Licensed Content
for the Course.
b. Feedback. If you agree to give feedback about the Licensed Content to Microsoft, you give to
Microsoft, without charge, the right to use, share and commercialize your feedback in any way and
for any purpose. You also give to third parties, without charge, any patent rights needed for their
products, technologies and services to use or interface with any specific parts of a Microsoft
software, Licensed Content, or service that includes the feedback. You will not give feedback that is
subject to a license that requires Microsoft to license its software or documentation to third parties
because we include your feedback in them. These rights survive this agreement.
c. Confidential Information. The Licensed Content, including any viewer, user interface, features
and documentation that may be included with the Licensed Content, is confidential and proprietary
to Microsoft and its suppliers.
i. Use. For five years after installation of the Licensed Content or its commercial release,
whichever is first, you may not disclose confidential information to third parties. You may
disclose confidential information only to your employees and consultants who need to know
the information. You must have written agreements with them that protect the confidential
information at least as much as this agreement.
ii. Survival. Your duty to protect confidential information survives this agreement.
iii. Exclusions. You may disclose confidential information in response to a judicial or
governmental order. You must first give written notice to Microsoft to allow it to seek a
protective order or otherwise protect the information. Confidential information does not
include information that
• becomes publicly known through no wrongful act;
• you received from a third party who did not breach confidentiality obligations to
Microsoft or its suppliers; or
• you developed independently.
d. Term. The term of this agreement for pre-release versions is (i) the date which Microsoft informs
you is the end date for using the beta version, or (ii) the commercial release of the final release
version of the Licensed Content, whichever is first (“beta term”).
e. Use. You will cease using all copies of the beta version upon expiration or termination of the beta
term, and will destroy all copies of same in the possession or under your control and/or in the
possession or under the control of any Trainers who have received copies of the pre-released
version.
f. Copies. Microsoft will inform Authorized Learning Centers if they may make copies of the beta
version (in either print and/or CD version) and distribute such copies to Students and/or Trainers. If
Microsoft allows such distribution, you will follow any additional terms that Microsoft provides to you
for such copies and distribution.
5. ADDITIONAL LICENSING REQUIREMENTS AND/OR USE RIGHTS.
a. Authorized Learning Centers and Trainers:
i. Software.
ii. Virtual Hard Disks. The Licensed Content may contain versions of Microsoft XP, Microsoft
Windows Vista, Windows Server 2003, Windows Server 2008, and Windows 2000 Advanced
Server and/or other Microsoft products which are provided in Virtual Hard Disks.
A. If the Virtual Hard Disks and the labs are launched through the Microsoft
Learning Lab Launcher, then these terms apply:
Time-Sensitive Software. If the Software is not reset, it will stop running based upon the
time indicated on the install of the Virtual Machines (between 30 and 500 days after you
install it). You will not receive notice before it stops running. You may not be able to
access data used or information saved with the Virtual Machines when it stops running and
may be forced to reset these Virtual Machines to their original state. You must remove the
Software from the Devices at the end of each Authorized Training Session and reinstall and
launch it prior to the beginning of the next Authorized Training Session.
B. If the Virtual Hard Disks require a product key to launch, then these terms
apply:
Microsoft will deactivate the operating system associated with each Virtual Hard Disk.
Before installing any Virtual Hard Disks on classroom Devices for use during an Authorized
Training Session, you will obtain from Microsoft a product key for the operating system
software for the Virtual Hard Disks and will activate such Software with Microsoft using such
product key.
C. These terms apply to all Virtual Machines and Virtual Hard Disks:
You may only use the Virtual Machines and Virtual Hard Disks if you comply with
the terms and conditions of this agreement and the following security
requirements:
o You may not install Virtual Machines and Virtual Hard Disks on portable Devices or
Devices that are accessible to other networks.
o You must remove Virtual Machines and Virtual Hard Disks from all classroom Devices at
the end of each Authorized Training Session, except those held at Microsoft Certified
Partners for Learning Solutions locations.
o You must remove the differencing drive portions of the Virtual Hard Disks from all
classroom Devices at the end of each Authorized Training Session at Microsoft Certified
Partners for Learning Solutions locations.
o You will ensure that the Virtual Machines and Virtual Hard Disks are not copied or
downloaded from Devices on which you installed them.
o You will strictly comply with all Microsoft instructions relating to installation, use,
activation and deactivation, and security of Virtual Machines and Virtual Hard Disks.
o You may not modify the Virtual Machines and Virtual Hard Disks or any contents
thereof.
o You may not reproduce or redistribute the Virtual Machines or Virtual Hard Disks.
ii. Classroom Setup Guide. You will assure any Licensed Content installed for use during an
Authorized Training Session will be done in accordance with the classroom set-up guide for the
Course.
iii. Media Elements and Templates. You may allow Trainers and Students to use images, clip
art, animations, sounds, music, shapes, video clips and templates provided with the Licensed
Content solely in an Authorized Training Session. If Trainers have their own copy of the
Licensed Content, they may use Media Elements for their personal training use.
iv. iv Evaluation Software. Any Software that is included in the Student Content designated as
“Evaluation Software” may be used by Students solely for their personal training outside of the
Authorized Training Session.
b. Trainers Only:
i. Use of PowerPoint Slide Deck Templates. The Trainer Content may include Microsoft
PowerPoint slide decks. Trainers may use, copy and modify the PowerPoint slide decks only for
providing an Authorized Training Session. If you elect to exercise the foregoing, you will agree
or ensure Trainer agrees: (a) that modification of the slide decks will not constitute creation of
obscene or scandalous works, as defined by federal law at the time the work is created; and
(b) to comply with all other terms and conditions of this agreement.
ii. Use of Instructional Components in Trainer Content. For each Authorized Training
Session, Trainers may customize and reproduce, in accordance with the MCT Agreement, those
portions of the Licensed Content that are logically associated with instruction of the Authorized
Training Session. If you elect to exercise the foregoing rights, you agree or ensure the Trainer
agrees: (a) that any of these customizations or reproductions will only be used for providing an
Authorized Training Session and (b) to comply with all other terms and conditions of this
agreement.
iii. Academic Materials. If the Licensed Content contains Academic Materials, you may copy and
use the Academic Materials. You may not make any modifications to the Academic Materials
and you may not print any book (either electronic or print version) in its entirety. If you
reproduce any Academic Materials, you agree that:
• The use of the Academic Materials will be only for your personal reference or training use
• You will not republish or post the Academic Materials on any network computer or
broadcast in any media;
• You will include the Academic Material’s original copyright notice, or a copyright notice to
Microsoft’s benefit in the format provided below:
Form of Notice:
© 2008 Reprinted for personal reference use only with permission by Microsoft
Corporation. All rights reserved.
Microsoft, Windows, and Windows Server are either registered trademarks or
trademarks of Microsoft Corporation in the US and/or other countries. Other
product and company names mentioned herein may be the trademarks of their
respective owners.
6. INTERNET-BASED SERVICES. Microsoft may provide Internet-based services with the Licensed
Content. It may change or cancel them at any time. You may not use these services in any way that
could harm them or impair anyone else’s use of them. You may not use the services to try to gain
unauthorized access to any service, data, account or network by any means.
7. SCOPE OF LICENSE. The Licensed Content is licensed, not sold. This agreement only gives you some
rights to use the Licensed Content. Microsoft reserves all other rights. Unless applicable law gives you
more rights despite this limitation, you may use the Licensed Content only as expressly permitted in this
agreement. In doing so, you must comply with any technical limitations in the Licensed Content that
only allow you to use it in certain ways. You may not
• install more copies of the Licensed Content on classroom Devices than the number of Students and
the Trainer in the Authorized Training Session;
• allow more classroom Devices to access the server than the number of Students enrolled in and the
Trainer delivering the Authorized Training Session if the Licensed Content is installed on a network
server;
• copy or reproduce the Licensed Content to any server or location for further reproduction or
distribution;
• disclose the results of any benchmark tests of the Licensed Content to any third party without
Microsoft’s prior written approval;
• work around any technical limitations in the Licensed Content;
• reverse engineer, decompile or disassemble the Licensed Content, except and only to the extent
that applicable law expressly permits, despite this limitation;
• make more copies of the Licensed Content than specified in this agreement or allowed by applicable
law, despite this limitation;
• publish the Licensed Content for others to copy;
• transfer the Licensed Content, in whole or in part, to a third party;
• access or use any Licensed Content for which you (i) are not providing a Course and/or (ii) have not
been authorized by Microsoft to access and use;
• rent, lease or lend the Licensed Content; or
• use the Licensed Content for commercial hosting services or general business purposes.
• Rights to access the server software that may be included with the Licensed Content, including the
Virtual Hard Disks does not give you any right to implement Microsoft patents or other Microsoft
intellectual property in software or devices that may access the server.
8. EXPORT RESTRICTIONS. The Licensed Content is subject to United States export laws and
regulations. You must comply with all domestic and international export laws and regulations that apply
to the Licensed Content. These laws include restrictions on destinations, end users and end use. For
additional information, see www.microsoft.com/exporting.
9. NOT FOR RESALE SOFTWARE/LICENSED CONTENT. You may not sell software or Licensed
Content marked as “NFR” or “Not for Resale.”
10. ACADEMIC EDITION. You must be a “Qualified Educational User” to use Licensed Content marked as
“Academic Edition” or “AE.” If you do not know whether you are a Qualified Educational User, visit
www.microsoft.com/education or contact the Microsoft affiliate serving your country.
11. TERMINATION. Without prejudice to any other rights, Microsoft may terminate this agreement if you
fail to comply with the terms and conditions of these license terms. In the event your status as an
Authorized Learning Center or Trainer a) expires, b) is voluntarily terminated by you, and/or c) is
terminated by Microsoft, this agreement shall automatically terminate. Upon any termination of this
agreement, you must destroy all copies of the Licensed Content and all of its component parts.
12. ENTIRE AGREEMENT. This agreement, and the terms for supplements, updates, Internet-
based services and support services that you use, are the entire agreement for the Licensed
Content and support services.
13. APPLICABLE LAW.
a. United States. If you acquired the Licensed Content in the United States, Washington state law
governs the interpretation of this agreement and applies to claims for breach of it, regardless of
conflict of laws principles. The laws of the state where you live govern all other claims, including
claims under state consumer protection laws, unfair competition laws, and in tort.
b. Outside the United States. If you acquired the Licensed Content in any other country, the laws
of that country apply.
14. LEGAL EFFECT. This agreement describes certain legal rights. You may have other rights under the
laws of your country. You may also have rights with respect to the party from whom you acquired the
Licensed Content. This agreement does not change your rights under the laws of your country if the
laws of your country do not permit it to do so.
15. DISCLAIMER OF WARRANTY. The Licensed Content is licensed “as-is.” You bear the risk of
using it. Microsoft gives no express warranties, guarantees or conditions. You may have
additional consumer rights under your local laws which this agreement cannot change. To
the extent permitted under your local laws, Microsoft excludes the implied warranties of
merchantability, fitness for a particular purpose and non-infringement.
16. LIMITATION ON AND EXCLUSION OF REMEDIES AND DAMAGES. YOU CAN RECOVER FROM
MICROSOFT AND ITS SUPPLIERS ONLY DIRECT DAMAGES UP TO U.S. $5.00. YOU CANNOT
RECOVER ANY OTHER DAMAGES, INCLUDING CONSEQUENTIAL, LOST PROFITS, SPECIAL,
INDIRECT OR INCIDENTAL DAMAGES.
This limitation applies to
• anything related to the Licensed Content, software, services, content (including code) on third party
Internet sites, or third party programs; and
• claims for breach of contract, breach of warranty, guarantee or condition, strict liability, negligence,
or other tort to the extent permitted by applicable law.
It also applies even if Microsoft knew or should have known about the possibility of the damages. The
above limitation or exclusion may not apply to you because your country may not allow the exclusion or
limitation of incidental, consequential or other damages.
Please note: As this Licensed Content is distributed in Quebec, Canada, some of the clauses in
this agreement are provided below in French.
Remarque : Ce le contenu sous licence étant distribué au Québec, Canada, certaines des clauses
dans ce contrat sont fournies ci-dessous en français.
EXONÉRATION DE GARANTIE. Le contenu sous licence visé par une licence est offert « tel quel ». Toute
utilisation de ce contenu sous licence est à votre seule risque et péril. Microsoft n’accorde aucune autre
garantie expresse. Vous pouvez bénéficier de droits additionnels en vertu du droit local sur la protection dues
consommateurs, que ce contrat ne peut modifier. La ou elles sont permises par le droit locale, les garanties
implicites de qualité marchande, d’adéquation à un usage particulier et d’absence de contrefaçon sont
exclues.
LIMITATION DES DOMMAGES-INTÉRÊTS ET EXCLUSION DE RESPONSABILITÉ POUR LES
DOMMAGES. Vous pouvez obtenir de Microsoft et de ses fournisseurs une indemnisation en cas de
dommages directs uniquement à hauteur de 5,00 $ US. Vous ne pouvez prétendre à aucune indemnisation
pour les autres dommages, y compris les dommages spéciaux, indirects ou accessoires et pertes de
bénéfices.
Cette limitation concerne:
• tout ce qui est relié au le contenu sous licence , aux services ou au contenu (y compris le code)
figurant sur des sites Internet tiers ou dans des programmes tiers ; et
• les réclamations au titre de violation de contrat ou de garantie, ou au titre de responsabilité stricte,
de négligence ou d’une autre faute dans la limite autorisée par la loi en vigueur.
Elle s’applique également, même si Microsoft connaissait ou devrait connaître l’éventualité d’un tel
dommage. Si votre pays n’autorise pas l’exclusion ou la limitation de responsabilité pour les dommages
indirects, accessoires ou de quelque nature que ce soit, il se peut que la limitation ou l’exclusion ci-dessus ne
s’appliquera pas à votre égard.
EFFET JURIDIQUE. Le présent contrat décrit certains droits juridiques. Vous pourriez avoir d’autres droits
prévus par les lois de votre pays. Le présent contrat ne modifie pas les droits que vous confèrent les lois de
votre pays si celles-ci ne le permettent pas.
Implementing and Managing Windows Server® 2008 Clustering xi
Contents
Module 1: Introduction to Clusters
Lesson 1: Overview of Clusters 1-3
Lesson 2: Benefits of Using Clusters 1-11
Lesson 3: Overview of the Windows Server 2008 High Availability
Solutions 1-20
Lab: Identifying Windows Server 2008 High-Availability Solutions 1-29
Course Description
This three-day instructor-led course introduces Windows Server®°2008 clustering,
and provides students with the knowledge and skills to implement, maintain, and
troubleshoot clusters.
Audience
This course is intended for IT professional technical specialists who are
responsible for using clustering technologies to implement and maintain high-
availability solutions.
Student Prerequisites
This course requires that you meet the following prerequisites:
• Experience managing Windows Server®°2003 or Windows Server 2008, in a
networking environment
• Basic knowledge of clustering theory
• Experience managing applications and network topologies in an enterprise
environment
• Basic IT troubleshooting skills
Course Objectives
After completing this course, you will be able to:
• Provide an overview of cluster concepts.
• Describe key features and functionality of the Windows Server 2008 failover
clusters.
• Prepare to install a failover cluster.
• Describe failover cluster storage requirements.
• Configure a failover cluster.
xiv About This Course
Course Outline
The following section provides a course outline:
Module 1: Provides an overview of cluster concepts and functionality.
Module 2: Describes key features and functionality of the Windows Server 2008
failover clusters.
Module 3: Explains the technical prerequisites and planning that is required to
install a Windows failover cluster.
Module 4: Describes storage fundamentals, and how to plan and implement
storage solutions for failover clusters.
Module 5: Explains how to install and manage a failover cluster.
Module 6: Explains how to configure cluster resources, and how to cluster
common Window Server roles and applications.
Module 7: Explains how to maintain and troubleshoot failover clusters.
Module 8: Explains geographically dispersed clusters, and the challenges that they
present. This module also describes how to implement a multi-subnet cluster using
Windows Server 2008.
Module 9: Explains how to install and maintain network load balanced (NLB)
clusters.
About This Course xv
Important: At the end of each lab, you must close the virtual machine, and must
not save any changes. To close a virtual machine without saving changes, perform
the following steps for each running virtual machine:
1. Close the Virtual Machine Remote Control window.
2. In the Close box that appears, select Turn off machine
and discard changes, and then click OK.
3. Close the Lab Launcher window.
xvi About This Course
6423A-VAN-SRV3A
Member servers in Fabrikam.com
6423A-VAN-SRV3B
6423A-VAN-SRV5A
Member servers in Fabrikam.com
6423A-VAN-SRV5B
6423A-VAN-SRV6A
Member servers in Fabrikam.com
6423A-VAN-SRV36B
6423A-VAN-CORE6A
Member servers in Fabrikam.com
6423A-VAN-CORE6B
Software Configuration
The following software is installed on each virtual machine:
• Windows Server 2008 Enterprise
Classroom Setup
Each classroom computer will have the same virtual machine configured in the
same way.
This lesson provides the industry definition of clusters, and describes how clusters
can benefit an enterprise. This lesson also defines important cluster terminology,
and the cluster types and categories.
1-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
A cluster is a group of computers and storage devices that work together as a single
organized system. You can use clusters to distribute servicing load or provide high
availability of services. You can create many different types of clusters to provide
these services.
In a cluster, the component computers communicate with each other over a high-
performance, reliable network: they may share one or more common storage
devices. You use cluster configurations to address availability, scalability, and
manageability.
Key Points
Answer the questions in a classroom discussion.
1-6 Implementing and Managing Windows Server 2008 Clustering
Key Points
You must have a basic understanding of common terminology to discuss cluster
features and components.
Key Points
There are four main types of clusters that are used in the market today. The four
types are:
• Highly available (HA) clusters, which provide greater uptime for computing
services. Multiple nodes allow backup servers to be available if a running
server suffers a failure.
• Load-balanced clusters, which distribute incoming network load to multiple
servers. Resources between servers are not usually shared. These types of
clusters are primarily designed to spread the work of one server to multiple
servers. Load-balanced clusters are also called server farms.
1-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
The four types of clusters fall into two main categories: shared device, and shared
nothing clusters.
• In shared device clusters, software that runs on any node in the cluster may
access any resource (for example, a disk) that is connected to any other node
in the cluster. If two nodes must access the same data, the data must be either
read twice from the disk, or copied from one node to another. You typically
employ a locking and tracking mechanism to preserve data integrity.
• Shared nothing provide greater redundancy. Each node of the cluster owns a
subset of the resources that make up the cluster. Only one system may own
and access a particular resource at a time, although, on a failure, another node
may take ownership of the resource. Additionally, requests from clients are
automatically routed to the node that owns the resource.
1-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
Clusters provide benefits in three key areas: availability, scalability, and
manageability.
Key Points
Availability is measured in the percentage of time that resources are running and
accessible during the course of a year or other period. Generally, availability rates
do not include planned downtime. Planned downtime includes time that the
application is unavailable due to software, hardware, and other maintenance
activities.
Mean time to failure (MTTF) is the amount of time that it takes for a solution to
fail. Every component has an MTTF. For example, a hard disk might have an MTTF
of 150,000 hours, and on average, you would replace it after 150,000 hours of use.
To increase MTTF of a single hard disk, you can add multiple disks in a redundant
array. You can apply this same methodology to other components, such as
memory and power supplies.
1-14 Implementing and Managing Windows Server 2008 Clustering
Key Points
Clusters help improve:
• MTTF by spreading load across multiple servers. Redundant cluster nodes
improve the MTTF by spreading load between nodes, which reduces the
amount of time a resource is used.
• MTTR by having multiple nodes, which allows a cluster solution to respond as
soon as a failure is detected.
Key Points
Answer the questions in a classroom discussion.
Introduction to Clusters 1-17
Key Points
Clusters can address several types of scalability: scaling up, scaling out, and
consolidation. Consolidation may seem to be the opposite of scalability; however,
due to improved stability and performance of clusters, you might be able to add
more load to a cluster.
Key Points
When scaling out, you add more nodes to a cluster to improve performance across
all nodes. For example, suppose that you have a two-node cluster with both nodes
running at 75% processor utilization, and the end-user performance is impacted. If
you add another node to the cluster, you can reduce the processor utilization. This
reduction occurs because the additional node takes on part of the workload. The
result is improved performance for end users.
Key Points
Although complexity increases in a clustered environment, disaster recovery and
update management capabilities outweigh these complexities. For example:
• Disaster recovery improves in a cluster because application and hardware
failures are recovered automatically by restarting the application on a healthy
node. This enables the failed node to be repaired while the application remains
available. In a stand-alone server configuration, you might need to reinstall or
reconfigure an application on new hardware in the event of a failure.
• By using a cluster, you can perform application and operating system updates
while the application continues to run on another cluster node.
Key Points
You assign most clustered applications and their associated resources to one
cluster node at a time. This type of resource control is commonly called a shared
nothing model. In this model, only the owning node has access to the assigned
cluster resources.
If a failover cluster detects a node failure or if the node is taken offline for
maintenance while running a clustered application, the application is started on
another cluster node. To minimize the impact of the failure, client requests are
immediately redirected to the new cluster node.
Key Points
In a Windows Server 2008 failover cluster, only one node at a time controls
resources. For example, in the case of a computer running Microsoft® SQL Server®,
the SQL Server application runs on either node A or node B. Required SQL Server
resources are controlled by the node that runs SQL Server. The node that controls
the resources is called the active, node and the node that is not running any
resources is called the passive node.
1-24 Implementing and Managing Windows Server 2008 Clustering
Key Points
You can use network load balancing clusters to provide scalability for Web services
and other front-end servers such as virtual private network (VPN) servers and
firewalls. Organizations can build groups of clustered computers to support load
balancing of Transmission Control Protocol (TCP) and User Datagram Protocol
(UDP) traffic requests.
Network load balancing clusters are groups of identical (usually cloned)
computers that, because of their numbers, enhance the availability of:
• Web servers
• Microsoft Internet Security and Acceleration (ISA) servers (for proxy and
firewall servers)
• Terminal Services
• Microsoft Office Outlook® Web Access
• Other applications that receive TCP and UDP traffic
Introduction to Clusters 1-25
Key Points
Network load balancing distributes IP traffic to multiple instances of a TCP/IP
service, such as a Web server, each running on a host within the cluster. Network
load balancing transparently distributes the client requests among the hosts, and
enables the clients to access the cluster using one or more virtual IP addresses.
From the client’s point of view, the cluster appears to be a single server that
answers these client requests. As enterprise traffic increases, you can simply plug
another server into the cluster.
For example, clustered hosts work together to service network traffic from
the Internet. Each server runs a copy of an IP-based service, such as Internet
Information Services 7.0 (IIS), and network load balancing distributes the
workload among them. This speeds up normal processing so that Internet clients
receive faster turnaround on their requests.
Introduction to Clusters 1-27
Key Points
Network load balancing is usually used to provide high availability and scalability
for Web servers or File Transfer Protocol (FTP) servers. No special hardware is
required to use network load balancing, and it is available in all editions of
Windows Server 2008. Network load balancing is usually deployed in a single
network, but can also be deployed over multiple networks to provide cross-
network availability.
Key Points
You can often use failover clusters and network load balancing together to create a
comprehensive solution. For example, a Web application can have a network load-
balanced tier of Web servers that connect to a failover clustered database tier. This
type of highly available tiered configuration provides a redundant solution that can
sustain multiple failures while continuing to serve client requests.
Introduction to Clusters 1-29
Lab Objectives
• Identify solutions for Web servers
• Identify solutions for database servers
• Identify complex solutions
Scenario
In this lab, you will answer questions based on given scenarios.
1-30 Implementing and Managing Windows Server 2008 Clustering
In order to plan and implement a failover cluster, you must first understand the
terminology, and how failover clusters work. This lesson provides this information.
2-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
In a failover cluster, each node in the cluster:
• Has full connectivity and communication with the other nodes in the cluster.
• Is aware when another node joins or leaves the cluster.
• Is connected to a network through which client computers can access the
cluster.
• Is connected through a shared bus or Internet SCSI (iSCSI) connection to
shared storage.
• Is aware of the services or applications that are running locally, and the
resources that are running on all other cluster nodes.
Introduction to Windows Server 2008 Failover Clusters 2-5
Key Points
Shared storage usually refers to logical devices — normally hard drives or logical
unit numbers (LUN) — that all the cluster nodes attach to via a shared bus. This
bus is separate from the bus that contains the system and boot disks. The shared
disks store applications, file shares, and other resources that the cluster will
manage.
Key Points
Clustered services that contain an IP address resource and a network name
resource (as well as other resources,) are published to clients on the network
under a unique server name. Because these groups appear as individual logical
servers to clients, they are called a cluster instance.
Users access applications or services on an instance in the same way they would if
the application or service were on a non-clustered server. Usually, the application
or user does not know that they are connecting to a cluster, nor to which node
they are connected.
Resources are physical or logical entities, such as a file share, disk, or IP address,
which the failover cluster manages. Resources may provide a service to clients, or
be an integral part of the cluster. Resources are the most basic and smallest
configurable unit. At any given time, a resource can run only on a single node in a
cluster, and is online on a node when it is providing its service to that specific
node.
2-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
When configuring networks in failover clusters, you must dedicate a network to
connect to the shared storage. If you use iSCSI for the shared storage connection,
the network will use IP-based communications; however, you cannot use this
network for node or client communication.
You can use the private and public networks for both client and node
communications. As a best practice, you should dedicate an isolated network for
the private node communication. The public network is configured to allow client
connections to the failover cluster. The public network can provide backup for the
private network.
Introduction to Windows Server 2008 Failover Clusters 2-9
Key Points
A failover attempt consists of the following steps:
1. The cluster service takes all the resources in the instance offline in an
order that is determined by the instance’s dependency hierarchy: dependent
resources first, followed by the resources on which they depend. For example,
if an application depends on a physical disk resource, the cluster service takes
the application offline first, which allows the application to write changes to
the disk before the disk is taken offline.
2. When all the resources are offline, the cluster service attempts to transfer the
instance to the node that is listed next on the instance’s list of preferred
owners.
Introduction to Windows Server 2008 Failover Clusters 2-11
Windows Server 2008 includes several core features that make it easier to set up
and maintain a failover cluster. This lesson describes these features and how they
will affect your failover administration tasks.
Introduction to Windows Server 2008 Failover Clusters 2-13
Key Points
One of the significant new features in Windows Server 2008 failover clusters is
that all the hardware components must be marked as “Certified for Windows
Server 2008”. In addition, the complete configuration must pass all tests in the
Validate a Configuration Wizard. This is different from previous server clusters
versions, which required that all hardware be on a hardware compatibility list
(HCL), specifically a list for server clusters.
The Validate a Configuration Wizard tests include specific simulations of cluster
actions, which fall into the following categories:
• System configuration tests: Analyzes whether the selected servers meet specific
requirements, for example the requirement that the servers must run the same
operating system version and software updates.
• Network tests: Analyzes whether the cluster nodes network configuration
meets specific requirements, for example requirements for network
redundancy.
2-14 Implementing and Managing Windows Server 2008 Clustering
Key Points
The management tools install by default when you install the Failover
Clustering feature. You can also install these tools on computers running
Windows Server 2008 that do not have the Failover Clustering feature installed.
To install the management tools, choose the option to add a feature in Server
Manager. The failover cluster tools are available under the Remote Server
Administration Tools.
The failover cluster tools are also included in Remote Server Administration Tools
that you can install on computers running Windows Vista® Service Pack 1.
Question:
How would you compare the Failover Cluster Management snap-in in Windows
Server 2008 with server cluster management tools in previous Windows versions,
or with other Windows Server 2008 management tools?
Introduction to Windows Server 2008 Failover Clusters 2-17
Key Points
Networking features include the following features:
• In order to verify all nodes in a cluster, the nodes transmit and receive
heartbeats to confirm that each node is still available. The nodes use User
Datagram Protocol (UDP) unicast, versus UPD broadcast (used in legacy
clusters). The messages are sent on port 3343.
• You can now place clustered servers on different IP subnets, which reduces the
requirements for multi-site clusters.
• The Microsoft® Failover Cluster Virtual Adapter is a hidden device that is
added to each node when you install the Failover Clustering feature. The
adapter is assigned a media access control (MAC) address based of the MAC
address that is associated with the first enumerated physical network adapter
in the node.
2-18 Implementing and Managing Windows Server 2008 Clustering
Key Points
Storage features include the following features:
• Parallel SCSI, which previous versions of server clusters supported, is no
longer supported for shared disk devices in a failover cluster.
• In Windows Server 2008, the cluster disk driver is a stand-alone component
that communicates with the Partition Manager driver (PartMgr.sys), for
managing clustered disks. During the installation process, the driver is
registered as a system Plug-n-Play (PNP) device, and is started during the
normal Windows startup process.
The cluster disk driver provides improvements to disk fencing and arbitration
behavior. Disk fencing activities have been offloaded to the PartMgr.sys
driver. This change improves integration with core operating system disk
management, and reduces the chances for disk corruption. Disk arbitration
now uses persistent reservations and an improved algorithm. This works to
eliminate the possibility of the disk being in an unprotected state.
2-20 Implementing and Managing Windows Server 2008 Clustering
Key Points
Security features include the following features:
• In Windows Server 2008 failover clusters, authentication uses Kerberos
authentication, but it can fall back to NTLM authentication (version 1
or version 2), if needed. Because computer objects are created in Active
Directory® domain services (AD DS), Kerberos authentication is supported
by default, for each clustered Network Name resource.
• When you configure a failover cluster, you create both a Cluster Name Object
(CNO) and a Virtual Computer Object (VCO). The CNO is associated with the
Cluster Name core resource. A VCO is a computer object that is created in AD
DS for all other cluster Network Name resources that are created as part of
configuring a Client Access Point (CAP). After AD DS creates the CNO, it is
used as the security context for the cluster. It is used when accessing all
external resources, and in the creation of VCOs. As a result, Kerberos will be
enabled on the Network Name resource by default, and will be used as the
primary authentication mechanism.
2-22 Implementing and Managing Windows Server 2008 Clustering
The most important reason for implementing failover clusters is to provide high
availability for services and applications. Windows Server 2008 failover clusters
use a quorum to ensure that enough members are available for the cluster to
continue operating. In order to implement and manage failover clusters, you must
understand the options that are available for configuring quorums. This lesson
describes how quorum works in Windows Server 2008 failover clusters.
2-24 Implementing and Managing Windows Server 2008 Clustering
Key Points
The failover cluster quorum configuration determines the number of failed
nodes or failed storage and network components that the cluster can sustain
while continuing to function. Quorum prevents two sets of nodes from operating
simultaneously as the failover cluster. Simultaneous operation could happen when
network problems prevent one set of nodes from communicating with another set
of nodes. Without a quorum mechanism, each set of nodes could continue to
operate as a failover cluster, resulting in a partition within the cluster.
To prevent problems caused by a split in the cluster, failover clusters use a
voting algorithm to determine whether the cluster has enough votes to maintain
quorum. Because a given cluster has a specific set of nodes and a specific quorum
configuration, the cluster will know how many votes are required. If the number of
votes drops below the majority, the cluster stops running. Nodes will still listen for
the presence of other nodes, in case another node appears again on the network,
but the nodes will not function as a cluster until a consensus is reached.
Introduction to Windows Server 2008 Failover Clusters 2-25
Key Points
The Quorum Manager component determines if the current cluster membership
has quorum. If the Quorum Manager determines that quorum has not been
achieved or has been lost, it will terminate the cluster across all nodes.
The Resource Control Manager component controls the configuration and state of
resources, and resource dependency trees. It is responsible for monitoring the
online status of active resources.
Introduction to Windows Server 2008 Failover Clusters 2-27
Key Points
With the exception of the No Majority: Disk Only mode, all quorum modes
in Windows Server 2008 failover clusters are based on a simple majority vote
model. As long as a majority of the votes are available, the cluster will continue to
function. For example, if there are five votes in the cluster, the cluster will continue
to function as long as there are at least three available votes. The source of the
votes is not relevant—the vote could be a node, witness disk, or witness file share.
When only a minority of votes is available, the cluster will stop functioning.
In the No Majority: Disk Only mode, the quorum-shared disk can veto all other
possible votes. In this mode, the cluster will continue to function as long as the
quorum-shared disk and at least one node are available. If the quorum-shared disk
is not available, the cluster will stop functioning, even if all nodes are still available.
In this mode, the quorum-shared disk is a single point of failure, so this mode is
not recommended.
2-28 Implementing and Managing Windows Server 2008 Clustering
Key Points
In a Node Majority mode, only the nodes in the cluster have a vote. If there are a
total of three nodes in the cluster, the following scenarios are possible.
Key Points
In a Node and Disk Majority mode, both the nodes in the cluster and the witness
disk have a vote. If there are a total of four nodes in the cluster, the following
scenarios are possible.
Witness
Node1 Node2 Node3 Node4 disk Cluster status
Online Online Online Online Online Online
Witness
Node1 Node2 Node3 Node4 disk Cluster status
Online Offline Offline Offline Online Offline
Key Points
In a Node and File Share Majority mode, the nodes in the cluster and a file share
witness have a vote. If there are a total of three nodes in the cluster, the following
scenarios are possible.
File share
Node1 Node2 Node3 witness Cluster status
Online Online Online Online Online
File share
Node1 Node2 Node3 witness Cluster status
Offline Offline Offline Online Offline
Key Points
In a No Majority: Disk Only mode, only the shared disk has a vote. If there are a
total of three nodes in the cluster, the following scenarios are possible.
Key Points
When you configure a failover cluster in Windows Server 2008, the Installation
Wizard automatically chooses one of two possible default configurations. Modify
this setting only if you determine that a change is appropriate for your cluster, and
ensure that you understand the implications of making the change.
In addition to planning your quorum mode, you should also consider the
capacity of the nodes in your cluster, and their ability to support the services and
applications that may fail over to that node. For example, a cluster that has four
nodes and a witness disk will still have quorum after two nodes fail. However, if
you have several applications or services deployed on the cluster, each remaining
cluster node may not have the capacity to provide services.
2-36 Implementing and Managing Windows Server 2008 Clustering
Lab Objectives
• Identify clustered scenarios
Scenario
In this lab, you will answer questions based on the following scenarios.
Scenario 1
Contoso is redesigning their file server implementation. The organization
has identified high availability as an essential requirement in the new server
implementation. All of the file servers will be located in the organization’s data
center, and will be connected to a storage area network (SAN). To meet
performance requirements, the organization will require three servers to be
available. The organization has been using a SAN for several years.
Introduction to Windows Server 2008 Failover Clusters 2-37
This module outlines the requirements for failover cluster hardware and software.
Additionally, this module outlines steps to install the Failover Clustering feature,
use the Validate a Configuration Wizard to ensure proper cluster configuration,
and install the failover cluster on Windows Server® 2008 Server Core.
Preparing to Install a Failover Cluster 3-3
Windows Server 2008 failover clusters have specific recommended hardware and
software configurations that enable Microsoft® to support the cluster. Failover
clusters are intended to provide a higher level of service than stand-alone servers.
Therefore, cluster hardware requirements are often more strict than requirements
for stand-alone servers.
This lesson describes the hardware, network, storage, infrastructure, and software
requirements for Windows Server 2008 failover clusters.
3-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
Failover clusters need the following to meet availability and support requirements:
• All hardware that you select for a failover cluster must meet the “Certified for
Windows Server 2008” logo requirements. Hardware that obtains this logo is
independently tested to meet the highest technical bar for reliability,
availability, stability, security, and platform compatibility.
• You must install the same hardware on each failover cluster node. For
example, if you choose a specific model of network adapter, you should install
this adapter on each of the cluster nodes.
• After you configure the servers with the hardware, all Validate a Configuration
Wizard tests must pass before the cluster is considered a configuration that is
supported by Microsoft.
Preparing to Install a Failover Cluster 3-5
Key Points
Failover cluster network components must have both the “Certified for Windows
Server 2008” logo, and pass the Validate a Configuration Wizard tests.
Additionally:
• The network adapters in each of the nodes must be identical and have the
same IP protocol version, speed, duplex, and flow control capabilities
available.
• The networks to which you connect the nodes should be redundant so that
even a single failure allows the nodes to continue to communicate with each
other. You can use network adapter teaming to provide single network
redundancy. We recommend multiple networks to provide multiple paths
between nodes for inter-node communication; otherwise, a warning will be
generated during the validation process.
• All network adapters in a cluster network must have the same IP address
format, which means they all use static IP addresses, or all use Dynamic Host
Configuration Protocol (DHCP).
Preparing to Install a Failover Cluster 3-7
Key Points
Storage is one of the most critical components in a failover cluster and requires
that:
• The mass-storage controllers match in each server node, in model, and in
firmware versions.
• The drivers are based on the Microsoft Storport driver model, and that all
Multipath software must be based on Microsoft Multipath I/O (MPIO). MPIO
can use generic device modules that the operating system provides, or device-
specific modules that the hardware vendor develops to work with the
hardware.
• Internet SCSI-based (iSCSI-based) storage has a dedicated network that is
separate from the client access and cluster-only communications networks.
Preparing to Install a Failover Cluster 3-9
Key Points
Failover clusters depend on infrastructure services, and require that each server
node must be in the same Active Directory® domain and, if you use DNS, the
nodes should use the same Domain Name System (DNS) servers for name
resolution.
We recommend that you install the same Windows Server 2008 features and roles
on each node. In addition, you should not install the Active Directory Domain
Services (AD DS) role on any of the cluster nodes. If you install the AD DS role on
one of the nodes, you must install it on all nodes.
Key Points
Failover clusters require that each cluster node run the same edition of either
Windows Server 2008 Enterprise, or Windows Server 2008 Datacenter. Depending
on the role that will be clustered, Server Core edition may also meet the software
requirements. However, you cannot install Server Core and full editions in the
same cluster.
Each node must run the same processor architecture. This means that each node
will have the same processor family, which might be Intel Xeon processor with
Extended Memory 64Technology®, AMD Opteron AMD64, or Intel Itanium-based.
Key Points
The new support policy is different from the support policy for previous Windows
Server versions. The new policy uses the Validate a Configuration Wizard to
provide detailed configuration tests of the failover cluster. This ensures that the
cluster matches best practices, and is configured in a way that Microsoft can
support.
For more information, see "The Microsoft Support Policy for Windows
Server 2008 Failover Clusters"
Preparing to Install a Failover Cluster 3-13
After you determine the requirements for a failover cluster, you must verify that the
required components are in place. This lesson outlines additional steps that you
must take before creating a cluster. These additional steps will help ensure a stable
and successful implementation.
3-14 Implementing and Managing Windows Server 2008 Clustering
Key Points
Before you run the Validate a Configuration Wizard, review all cluster
requirements and verify that:
• All hardware and software is in place and properly configured. Windows
Failover Cluster Help provides step-by-step instructions to assist you with
preparing hardware.
• You have a plan for each clustered application, which ensures that the cluster
is configured to meet business needs.
Preparing to Install a Failover Cluster 3-15
Key Points
Windows Server 2008 creates a Clustered Name Object (CNO) during the
cluster creation process. The CNO is used to access resources such as AD DS, DNS,
and Windows Internet Name Service (WINS), and for creating cluster Network
Name resources. By default, in Windows Server 2008, clustered applications take
advantage of Kerberos authentication for resources. Windows Server 2008 failover
clusters provide support for legacy authentication, and for other applications that
require NT LAN Manager (NTLM) authentication.
Key Points
Answer the questions in a classroom discussion.
3-18 Implementing and Managing Windows Server 2008 Clustering
Key Points
You must look at all cluster configuration components to identify single points of
failure. You can remediate many single points of failure with simple solutions, such
as adding storage controllers, or network adapters, and multipathing software.
These solutions will reduce the probability that a device failure will cause a failure
in the cluster. Typically, server class hardware has options for multiple power
supplies for power redundancy, and for creating redundant array of independent
disks (RAID) sets for disk data redundancy.
Some single points of failures might be difficult or expensive to solve; in these
situations, you must weigh the benefits of eliminating the point of failure against
the cost of purchasing and maintaining the solution. For example, if a single
datacenter fails, creating a multi-site cluster can solve this problem; however, this
solution is more complicated and expensive than a single location cluster.
Preparing to Install a Failover Cluster 3-19
Key Points
Windows Server 2008 cluster service has changed from previous versions. It does
not allow legacy clusters to exist, or to be upgraded to Windows Server 2008.
However, the Migrate Services and Applications Wizard can help migrate specific
resources from legacy clusters to new Windows Server 2008 clusters. This wizard
will migrate the following resources:
• Physical Disk resource
• Network Name resource
• IP Address resource
• File Share resource (including DFS Root)
Preparing to Install a Failover Cluster 3-21
However, you will need to manually complete additional configuration steps for
the migration, such as installing supporting server roles and copying data.
Key Points
To migrate resources:
1. In the Failover Cluster Management snap-in, select or specify the cluster to
which you want to migrate settings.
2. Under Configure, click Migrate Services and Applications.
3. Read the first page of the Migrate a Cluster Wizard, and then click Next.
4. Specify the name or IP Address of the cluster or cluster node from which you
want to migrate resource groups, and then click Next.
5. Click View Report. Read the pre-migration report, which explains whether
each resource is eligible for migration, and describes additional steps to
perform after running the wizard.
Preparing to Install a Failover Cluster 3-23
Key Points
There are two main cluster load types: Active/Active, and Active/Passive.
In an Active/Active configuration, each node hosts clustered services. In an
Active/Passive configuration, at least one cluster node does not host any services.
Planning server capacity in an Active/Active configuration requires that when a
node fails, the service will be able to continue functioning on an operational node
that is running other services.
Planning server capacity in an Active/Passive configuration is dependent on the
number of failed nodes the cluster is able to sustain.
When planning server capacity, you will determine the fewest number of nodes
required to maintain the cluster operations during a failure, and to support the
applications and hardware in the cluster.
Preparing to Install a Failover Cluster 3-25
Key Points
If a node failure occurs in a two-node cluster with each node running at thirty
percent of capacity, a single node must be able to handle the additional load
from both services. To plan for this additional load, you must be sure that when
a failover does occur, both services are maintained. If you do not plan for the
additional load, and the load was sixty percent on each node in a two-node cluster,
a failover would impact both clustered applications. In contrast to a stand-alone
situation, a failure on a two-node cluster affects both the application originally
running on the failed node, and the application running on the active node.
This lesson describes how to install the Failover Clustering feature, run the
Validate a Configuration Wizard, and correct errors that the Validate a
Configuration Wizard finds.
Preparing to Install a Failover Cluster 3-27
Key Points
After you configure the hardware and have identified and addressed any single
points of failure, you can begin to configure the cluster. To do this, install the
Failover Clustering feature on each server node, and then run the Validate a
Configuration Wizard to be sure that hardware and software meets the cluster
requirements. Finally, create the cluster and configure highly available applications
and services.
Key Points
The Validate a Configuration Wizard runs tests that confirm that the hardware and
hardware settings are compatible with failover clustering. Using the wizard, you
can run the entire set of configuration tests or a subset of the tests. We recommend
that you run the tests on servers and storage devices before you configure the
failover cluster, and again after any major changes are made to the cluster. You can
access the test results in the %windir%\cluster\Reports directory.
The report from the Validate a Configuration Wizard includes one of the following
outcomes for each test:
• The tested item meets requirements for a failover cluster.
• The tested item might meet the requirements, but you receive a warning notice
suggesting that you review the hardware and software settings to be sure that
best practices are met.
Preparing to Install a Failover Cluster 3-29
Key Points
Inventory tests provide lists of information about the hardware, software, and
settings for each server that you test. For example, after you perform scheduled
maintenance, you can use the inventory test to ensure that the software updates on
each server are identical. The inventory tests can be used separate from the other
tests included in the Validate a Configuration Wizard.
Key Points
The Validate a Configuration Wizard performs the following network tests:
Test Description
Validate Cluster • Lists the cluster networks, which is the network topology as
Network seen from the perspective of the cluster.
Configuration • Validates for a particular cluster network, that all network
adapters are provided with IP addresses in the same way.
(All use static IP addresses or all use DHCP.)
• Validates for a particular cluster network, that all network
adapters use the same IP version. (All use IPv4, all use IPv6,
or all use both IPv4 and IPv6.)
3-32 Implementing and Managing Windows Server 2008 Clustering
Test Description
Key Points
The storage tests list and test the capabilities of all disks available to the cluster.
These tests are comprehensive; however, some specific tests may not run after the
cluster is running nor in a multi-site cluster.
3-34 Implementing and Managing Windows Server 2008 Clustering
Test Description
List All Disks Lists all disks that are visible to one or more tested servers. The
test lists:
• Disks that can support clustering and can be accessed by all
the servers.
• Disks on an individual server.
List Potential Cluster Lists disks that can support clustering, and are visible to all
Disks tested servers. To support clustering, the disk must be
connected through Serial Attached SCSI (SAS), iSCSI, or Fibre
Channel. In addition, the test validates that multipath I/O is
working correctly, which means that each of the disks is seen
as one disk, not two.
Validate Disk Access Validates that the latency for disk read/write operations
Latency is within an acceptable limit for a failover cluster. If disk
read/write operations take too long, one possible result is that
cluster time-outs might be triggered. Another possible result
is that the application attempting to access the disk might
appear to have failed, and the cluster might initiate a needless
failover.
Test Description
Validate Disk Failover Validates that disk failover works correctly in the cluster.
Specifically, the test validates that when a disk owned by a
clustered server is failed over, the server that takes ownership
of the disk can read it. The test also validates that information
written to the disk before the failover, is still the same after
the failover.
If disk failover occurs but the server that takes ownership of a
disk cannot read it, the cluster cannot maintain availability of
the disk. If information written to the disk is changed during
the process of failover, it might cause issues for users or
software that require this information. In either case, if the
affected disk is a witness disk, (a disk that stores cluster
configuration data and participates in quorum,) such issues
could cause the cluster to lose quorum and shut down.
Validate File System Validates that the file system on disks in shared storage is
supported by failover clusters.
Validate Multiple Validates that when multiple clustered servers arbitrate for a
Arbitration cluster disk, only one server obtains ownership.
If multiple clustered servers obtain ownership of a cluster disk
through disk arbitration, the disk might become corrupted.
Failover clusters are designed to operate in circumstances
where only one clustered server at a time owns a disk. If
multiple servers own a disk at the same time, they might
perform write operations in an uncoordinated way, possibly
corrupting the disk.
Validate SCSI Device Validates that the storage supports necessary SCSI inquiry
Vital Product Data data as well as Vital Product Data (VPD) descriptors, and that
(VPD) they are unique.
3-36 Implementing and Managing Windows Server 2008 Clustering
Test Description
Validate SCSI-3 Validates that the cluster storage uses the more recent
Persistent Reservation (SCSI-3 standard) Persistent Reserve commands, which are
different from the older SCSI-2 standard reserve/release
commands. Because the Persistent Reserve commands avoid
SCSI bus resets, they are much less disruptive than the older
reserve/release commands. Therefore, a failover cluster can
be more responsive in a variety of situations, unlike a cluster
running an earlier version of the operating system. In addition,
disks are never left in an unprotected state, which lowers the
risk of volume corruption.
Key Points
System configuration tests analyze selected servers to determine whether they are
properly configured to work together in a failover cluster.
The Validate a Configuration Wizard performs the following system configuration
tests:
Test Description
Validate Active Validates that each tested server is in the same domain and
Directory organizational unit. It also validates that all tested servers are
Configuration domain controllers, or are member servers. To change the
domain role of a server, use the Server Manager.
3-38 Implementing and Managing Windows Server 2008 Clustering
Test Description
Validate All Drivers Validates that all tested servers contain only signed drivers. If
Signed an unsigned driver is detected, the test is not considered a
failure, but a warning is issued.
Signed drivers indicate whether the drivers on your system are
original, unaltered files that either came with the operating
system, or were supplied by a vendor.
Validate Operating Validates that all tested servers have the same operating
System Versions system version.
Validate Required Validates that the services required for failover clustering are
Services running on each tested server, and are configured to start
automatically whenever the server is restarted.
Validate Same Validates that all tested servers have the same architecture. A
Processor failover cluster is supported only if the systems in it are all 32-
Architecture bit systems, or all 64-bit systems.
Validate Service Pack Validates that all tested servers have the same service packs. A
Levels failover cluster can run even if some servers have service packs
that are different than others. However, servers with different
service packs might behave differently from each other, with
unexpected results. We recommend that all servers in the
failover cluster have the same service packs.
Validate Software Validates that all tested servers have the same software
Update Levels updates. A failover cluster can run even if some servers have
different updates than others. However, servers with different
software updates might behave differently from each other,
with unexpected results. We recommend that all servers in the
failover cluster have the same software update levels.
Key Points
The Validate a Configuration report contains details about the tests that were run,
and their test results. The report also provides information about reasons for any
failures. You can also review the Failover Cluster Help checklists to be sure that all
steps and requirements have been met.
If you cannot resolve problems after reading the report, contact the vendor for
configuration guidance, or software version compatibility changes. In some cases, a
vendor may recommend a specific driver or firmware version that is known to be
stable in a clustered environment.
Questions:
What step is required before you run the Validate a Configuration Wizard?
Using a Server Core Windows Server 2008 Enterprise or Windows Server 2008
Datacenter installation to create failover clusters provides specific benefits over a
full server installation. This lesson describes the benefits of installing a failover
cluster on Server Core, and the general process for completing an installation.
3-42 Implementing and Managing Windows Server 2008 Clustering
Key Points
Installing a failover cluster on Server Core results in several key benefits, due to the
fact that Server Core only installs essential components. These benefits include:
• A reduced vulnerability attack surface, and a reduced number of required
updates.
• Fewer components to manage, and less required disk space for the installation.
• Lower resource overhead, which provides more resources for the clustered
instances.
Key Points
Because of the reduced overhead and functionality of Server Core, you can cluster
only specific roles and features.
Key Points
Installing a Server Core failover cluster is similar to installing an operating system;
however, you must administer Server Core from a command line or from a remote
server.
To install a failover cluster on Server Core, you must:
1. Install the operating system.
2. Use netsh to configure the network and Windows firewall settings.
3. Use netdom to join the servers to the domain.
4. Use Diskpart to configure the disk partitions from the command line.
5. Use ocsetup to install the Failover Clustering feature from the command line
of each node in the cluster.
Preparing to Install a Failover Cluster 3-45
Questions:
When you install the Failover Clustering feature from the command line, what
output is produced?
Why must you use the command line to manage a Server Core computer?
Lab Objectives
• Install the Failover Clustering feature.
• Validate the failover cluster.
Scenario
Fourth Coffee is deploying a new file server cluster. The two servers you will use
in the file server cluster have an operating system installed, the clustered disk
presented, and network adapters installed. You must prepare the servers for
clustering the file server role. Your task is to install the Failover Clustering feature,
and complete the Validate a Configuration Wizard.
3-48 Implementing and Managing Windows Server 2008 Clustering
Planning for shared storage is one of the most important aspects of cluster
deployment. Lack of proper storage planning can result in unstable clusters.
This module defines basic storage terminology, types of available configurations,
and information about Storage Area Networks (SAN).
Overview of Failover Cluster Storage Requirements 4-3
This lesson describes disk and connectivity types, and provides information on
Storage Area Networks (SAN), how SANs can assist with data resiliency and
replication, and the differences between SANs and Network Attached Storage
(NAS).
4-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
Windows Server®°2008 supports only Fibre Channel, Internet SCSI (iSCSI), and
Serial Attached SCSI (SAS) storage for failover clusters. Windows Server 2008
does not support Parallel SCSI for clustered disks in two-node clusters, as it did
in previous versions of Windows Server. If you create a Windows Server 2008
cluster from hardware that you used with a legacy version of Windows Server, it
is important that you verify that Windows Server 2008 supports the storage
hardware.
Key Points
Windows Server 2008 supports three types storage technology for failover clusters:
• Fibre Channel: Transmits SCSI commands over twisted-pair copper wire, or
more commonly fiber optic cables. Fibre Channel is commonly found in
speeds of 1, 2, or 4 gigabits (GB) per second. Fibre Channel requires
specialized interface cards, switches, and cabling to create the network.
• iSCSI: Transmits SCSI commands over Internet Protocol (IP). iSCSI relies on
standard Ethernet networking, and requires no specialized hardware. Each
device on an iSCSI SAN is given an IP address to communicate with other
devices on the iSCSI SAN. Although the iSCSI protocol has more overhead
than Fibre Channel, Ethernet networking speeds are now becoming affordable
at 10 GB.
• Serial Attached SCSI: Transmits SCSI commands serially in order to provide
throughput (up to 6 GB per second,) that is higher than the previous parallel
SCSI standard. Serial Attached SCSI is not backward compatible with parallel
SCSI equipment.
4-6 Implementing and Managing Windows Server 2008 Clustering
Key Points
A SAN is a set of interconnected devices, such as disks, tapes, and servers, that are
connected to a common communication and data transfer infrastructure, such as
Fibre Channel. The common communication and data transfer mechanism for a
given deployment is called the storage fabric.
The SAN enables multiple servers to access a pool of storage in which any server
can potentially access any storage unit. In this environment, management (which
determines who is authorized to access which devices,) and sequencing or
serialization guarantees, (which determine who can access which devices at what
point in time,) play a large role in providing security guarantees.
4-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
Storage Area Networks can provide advanced features such as data replication and
hardware-based snapshots. Hardware-based snapshots enable you to create copies
of the failover cluster data, which is transparent to the clustered applications.
Clustered applications, such as database and messaging servers, often store a large
amount of data on the clustered disk. These services not only require backups, but
also require effective restores in the event of data loss.
Although traditional tape-based backups continue to improve in performance,
hardware-based snapshots might be a more effective solution. In the event of data
failure, you can attach the snapshot of the data to the clustered servers in a matter
of minutes to return an application to service.
4-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
Data redundancy is important in a failover cluster where the goal for the solution
is to increase availability. You can use a variety of methods for improving storage
systems performance and reliability. The following is a list of methods for writing
data to storage:
RAID 1 Data is mirrored across two disks. Transactional performance is very good
because either disk in the mirror can respond to read requests. No parity
information needs to be calculated; therefore, disk writes are efficiently
handled. Each disk in the mirrored set must perform the same write.
4-12 Implementing and Managing Windows Server 2008 Clustering
RAID 10 Data is striped (RAID 0) across mirrored (RAID 1) sets. RAID 0+1 is not the
same as RAID 10. Transactional performance with RAID 10 is very good
because either disk in the mirror can respond to read requests. No parity
information needs to be calculated; therefore, disk writes are efficiently
handled. Each disk in the mirrored set must perform the same write.
When a disk fails in a RAID 10 array, write-performance is not affected
because there is still a member of the mirror that can accept writes. In this
situation, reads are moderately affected because only one physical disk
can respond to read requests. When the failed disk is replaced, the mirror
is again established, and the data must be copied or rebuilt.
RAID 5 Calculates parity that surviving member data can use to re-create the
data on a failed disk. Writing to a RAID 5 array causes up to four I/Os for
each I/O to be written, and the parity calculation can consume controller
or server resources. Transactional performance with RAID 5 can still be
good, particularly when using a storage controller to calculate the parity.
When a disk fails in a RAID 5 array, the array is in a degraded state,
performance is decreased, and latencies are higher. This occurs because
most arrays spread the parity information equally across all disks in the
array, and it can be combined with surviving data blocks to reconstruct
data in real time. Both reads and writes must access multiple physical
disks to reconstruct data on a lost disk; during a failure on a RAID 5 array,
this increases latency and reduces performance. When the failed disk is
replaced, the parity and surviving blocks are used to reconstruct the lost
data. This is a lengthy process that can take a number of hours to
complete.
Key Points
A SAN and a NAS differ in several key areas: a SAN is a block-level access storage
solution that uses Fibre Channel, iSCSI, or SAS, a NAS solution uses file-level
access such as Common Internet File System (CIFS), File Transfer Protocol (FTP),
network file system (NFS), and Hypertext Transfer Protocol (HTTP). Windows
Server 2008 supports SAN in a failover cluster, yet does not support NAS for
failover clusters.
4-14 Implementing and Managing Windows Server 2008 Clustering
Communication SCSI over Fibre Channel iSCSI, CIFS, NFS, FTP, and HTTP
Protocol which is SCSI over IP and
Serially Attached SCSI (SAS)
Support for Windows Several types of SAN Not supported for failover
Server 2008 Failover topologies are supported for clusters
Clustering clustered disks
Overview of Failover Cluster Storage Requirements 4-15
Key Points
Windows Server 2008 clusters offer two main components to disk and path
management: Virtual Disk Service (VDS), and Microsoft Multipath I/O (MPIO).
VDS is a set of application programming interfaces (APIs) that provide a single
interface for managing disks. This provides an end-to-end solution for managing
storage hardware and disks, and for creating volumes on those disks.
VDS uses two sets of providers to manage storage devices. The built-in VDS
software providers enable you to manage disks and volumes at the operating
system level. Hardware providers that the hardware vendor supplies enable you to
manage hardware RAID arrays. Each hardware vendor writes a VDS hardware
provider that translates the general-purpose VDS APIs into specific instructions for
their hardware. VDS simplifies storage management, because storage applications
no longer need to account for the specific hardware that is being targeted. Because
of VDS, you no longer need to use different storage applications to manage each
different storage device.
4-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
In a multipath environment, you connect two host bus adapters (HBA) or network
interface cards (NIC) to two separate fabrics or networks, so that the storage is still
accessible if there is a failure. MPIO provides multipathing and aggregation of
paths to remote disks.
MPIO storage driver provides:
• Dynamic configuration and replacement of devices: The operating system
must be able to dynamically discover and configure adapters that are
connected to the same storage media, in order to support multiple paths to
the same storage device.
• Generic device specific module: Microsoft supplies a generic device specific
module (DSM) that interacts with the multipath bus driver on behalf of the
storage device.
4-18 Implementing and Managing Windows Server 2008 Clustering
Key Points
SAN disk storage can be different than the type of connectivity that is available
from the storage to the host servers. SAN disk storage types include:
• Fibre Channel disks with dual port Fibre Channel connections to provide
connectivity redundancy; however, these are more expensive than other types
of disk-based storage. Fibre Channel disks are designed for high-duty cycle,
and a high mean time to failure (MTTF).
• Fiber Channel ATA (FATA) disks that are based on Serial Advanced
Technology Attachment (SATA) standards; however, they are connected via
Fibre Channel. These drives typically have a much lower MTTF rating than
comparably sized Fibre Channel or SAS-based disks, and are primarily for
archival data.
4-20 Implementing and Managing Windows Server 2008 Clustering
Key Points
Windows Server 2008 supports three SAN connectivity types for failover clusters:
• Fibre Channel connectivity requires specialized host adapters, cables, and
switches to provide a specialized, fast interconnect between servers and
storage. Managing Fibre Channel hardware may require training because of
differences with other SAN technologies.
• iSCSI connectivity is less expensive than Fibre Channel and requires no
specialized hardware. Because iSCSI wraps SCSI commands in TCP/IP packets
and transmits over standard Ethernet, you can expect higher latency.
• SAS connectivity is inexpensive; however, it is more limited in the number of
hosts that you can connect, and less flexible in how you can configure failover
cluster support.
4-22 Implementing and Managing Windows Server 2008 Clustering
Storage Area Networks (SAN) are the basis for most Windows Server 2008 failover
cluster storage solutions. This lesson provides details of SAN solutions, and the
components for Fibre Channel and iSCSI solutions.
Overview of Failover Cluster Storage Requirements 4-23
Key Points
Use specific terminology when discussing SANs.
4-24 Implementing and Managing Windows Server 2008 Clustering
Key Points
A storage controller is a device that houses a set of disks, and provides a single
connection (which is redundant and highly available) to a SAN fabric.
Modern controllers usually provide some level of redundancy for data. For
example, many controllers offer a wide variety of RAID levels, such as RAID 1,
RAID 5, and RAID 0+1 to ensure data availability in the event of the failure of an
individual disk drive. The controller presents a virtual view of highly available
storage devices, called logical devices or logical unit numbers (LUN), to the hosts.
A controller requires at least two ports for highly available storage controllers.
Overview of Failover Cluster Storage Requirements 4-25
Key Points
Use specific terms when discussing Fibre Channel configurations.
4-26 Implementing and Managing Windows Server 2008 Clustering
Key Points
In a Fibre Channel switched fabric (FC-SW), Fibre Channel switches connect
devices in a many-to-many design. When a host or device communicates with
another host or device, the source and target create a point-to-point connection
between them, and then communicate directly with each other. The fabric itself
routes data from the source to the target. In an FC-SW, the media is not shared.
Any device can communicate with any other device, assuming it is not busy, and
communication occurs at full bus speed regardless of whether other devices and
hosts are communicating.
When a host or device is powered on, it must first log in to the fabric. This enables
the device to determine the type of fabric (a fabric supports a specific set of
characteristics,) and provides a fabric address to a host or device. A given host or
device continues to use the same fabric address while it is logged on to the fabric,
and the fabric address is guaranteed to be unique for that fabric. In order for a host
or device to communicate with another device, it must establish a connection to
that device before transmitting data. The switches route the packets in the fabric.
Overview of Failover Cluster Storage Requirements 4-27
Key Points
In a full mesh design, each switch connects directly to every other switch. Full-
mesh design provides any-to-any connectivity, distributes traffic evenly, and
consumes low overall bandwidth.
Overview of Failover Cluster Storage Requirements 4-29
Pros/Cons Description
Key Points
The core-to-edge SAN design enables the SAN fabric to scale. The dual fabric
design provides fault tolerance by eliminating single points of failures. From a
logical perspective, the core-to-edge SAN fabric is comprised of two layers of fiber
switches: the edge and the core. Each fabric has edge switches that connect to
server hosts, and core switches that connect to the storage subsystems.
Overview of Failover Cluster Storage Requirements 4-31
Pros/Cons Description
Key Points
You use zoning to control which device Worldwide Names (WWNs) can see other
devices in a SAN. There are a number of server, storage, and switch approaches to
zoning.
On any server, there are various mechanisms that control what devices an
application can see, and whether or not the application can talk to another device.
At the lowest level, the HBA firmware or driver has a masking capability to control
whether the server can see other devices. In addition, the operating system can
control which devices it tries to mount as a storage volume. You can also use third-
party software for volume management, clustering, and file-system sharing, which
can also control application access.
Overview of Failover Cluster Storage Requirements 4-33
Key Points
You use specific terms to describe iSCSI connectivity, clients, and configuration
changes.
Key Points
Finding available storage controllers and LUNs can be difficult on a large iSCSI
network; however, there are provisions to assist in locating storage controllers.
• SendTargets retrieves a list of available targets from the target device. It
requires that you to know the IP address of the storage device that is hosting
the targets, and is only suitable in smaller iSCSI environments.
• Internet Storage Name Service (iSNS), which is a Windows Server 2008 feature
similar to DNS, enables you to locate a target on a number of target devices.
iSNS contains three distinct services:
• Name Registration Service: Enables initiators and targets to register and
query the iSNS server directory for information about initiator and target
ID and addresses.
4-36 Implementing and Managing Windows Server 2008 Clustering
This lesson outlines recommendations for Windows Server 2008 failover cluster
SAN performance and availability.
4-38 Implementing and Managing Windows Server 2008 Clustering
Key Points
To plan a failover cluster SAN that meets business needs, you must document
the availability and performance goals for the applications that you will cluster.
Availability and performance goals include application performance requirements,
and business requirements for availability of those applications.
After you determine the performance and business requirements, you identify the
hardware that meets the requirements. Make sure that each hardware component
you choose has the “Certified for Windows Server 2008” logo.
Overview of Failover Cluster Storage Requirements 4-39
Key Points
Answer the questions in a classroom discussion.
4-40 Implementing and Managing Windows Server 2008 Clustering
Key Points
Because recommendations for SAN performance can vary by vendor, we
recommend that you consult the vendor’s documentation, where appropriate. To
provide the appropriate performance for the clustered applications, review some of
the following factors that affect all storage systems:
• Verify that adequate I/O is available
• Some SAN vendors enable you to add storage controllers to increase the
amount of available throughput.
• Some SAN vendors enable you to add cache memory to the storage
controllers, which allows data to be written or retrieved from memory
instead of waiting for the hard disks to respond.
• Because all data gets written to disks eventually, be sure that you have
enough disks to sustain the data throughput so that performance does not
suffer.
Overview of Failover Cluster Storage Requirements 4-41
Key Points
If the underlying storage does not provide the availability that you require for
the failover cluster, the cluster availability goals will not be met. To provide the
appropriate availability for the clustered applications, consider some of the
common factors that affect all storage systems:
• Ensure redundant components: When available and needed to meet
availability requirements, you should consider redundant controllers, memory,
and power supplies. This also extends to the Fibre Channel or Ethernet
switches. The hosts should have multiple HBA or network adapters to provide
multipath connectivity to the storage.
• Ensure resilient configurations: Be sure that you configure the LUNs with the
appropriate level of redundancy, and the appropriate number of spare disks to
take over in the event of a failure.
• Ensure parallel configurations: Be sure that each host has the same SAN
interface adapters with the same firmware, driver, and DSM versions.
Overview of Failover Cluster Storage Requirements 4-43
In some environments, you might choose iSCSI as the SAN storage solution due to
its low cost of entry, and adequate performance. This lesson describes the iSCSI
components on a Windows Server 2008 computer, and explains how to configure
them.
4-44 Implementing and Managing Windows Server 2008 Clustering
Key Points
Windows Server 2008 includes three iSCSI components:
MPIO DSM A generic set of device-specific modules for MPIO that works
with a variety of storage vendors, and is shipped with
Windows Server 2008.
Microsoft Internet A Windows Server 2008 feature, similar to DNS, that allows a
Storage Name client to locate a target on a number of target devices.
Service
Overview of Failover Cluster Storage Requirements 4-45
Key Points
You can find iSCSI targets with hardware, third-party software, or natively on a
Windows Storage Server edition hardware appliance.
• Many storage vendors provide iSCSI targets as part of their hardware solution.
• Many solutions provide iSCSI targets by loading third-party software on an
existing operating system.
• Windows Storage Server appliance provides a hardware appliance with iSCSI
target capabilities.
4-46 Implementing and Managing Windows Server 2008 Clustering
Key Points
You use persistent bindings to configure iSCSI targets on hosts. A persistent
binding maintains a connection for cluster nodes to a target, between reboots.
You use persistent reservations to provide contention resolution when multiple
initiators have access to the same LUN. Each initiator has a specific reservation key.
When the initiator takes ownership of the LUN, it registers the assigned key to
block all other access to the LUN. The reservation keys are automatically assigned.
Question:
Questions:
When might choosing MPIO on the binding not have any affect?
Overview of Failover Cluster Storage Requirements 4-49
Lab Objectives
• Identify Fibre Channel Storage Area Network (SAN) components.
• Configuring Internet SCSI (iSCSI) storage connections.
Scenario
Use the following scenarios to complete the lab exercises.
4-50 Implementing and Managing Windows Server 2008 Clustering
Host
Storage Controller
Host
Storage Controller
Network Adapter
4-52 Implementing and Managing Windows Server 2008 Clustering
After you verify the cluster hardware and software by using the Validate a
Configuration Wizard, you are ready to create the cluster, cluster resources, and
clustered applications.
This module describes how to create a cluster, and how to perform basic
management tasks for the newly created cluster.
Configuring a Failover Cluster 5-3
This lesson describes how to create a cluster and verify cluster functionality.
5-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
After you complete the Validate a Configuration Wizard, you create a cluster
resource. To create the cluster resource, you must log on to a server with a domain
account that has local administrator permissions on each of the cluster nodes.
Start the Create a Cluster Wizard from the Failover Cluster Management
Microsoft® Management Console (MMC). The wizard prompts you for a list of
server nodes that will be members of the cluster. After entering each of the nodes,
provide a cluster name and IP addresses.
For more information, see "Failover Cluster Help: Add Resource Type
Dialog Box"
Configuring a Failover Cluster 5-5
Question:
Is there a reason that you might run the Create a Cluster Wizard from a server that
is not part of the cluster?
For more information, see "Windows Server 2008 Failover Cluster Step-
by-Step"
5-6 Implementing and Managing Windows Server 2008 Clustering
Questions:
What information should you review in the failover cluster installation report?
Why is it important to examine the installation report closely before you deploy
clustered applications?
For more information, see "Windows Server 2008 Failover Cluster Step-
by-Step"
Configuring a Failover Cluster 5-7
Built-in management tools enable you to manage a cluster. This lesson outlines the
failover cluster management options that help you maintain and configure the
cluster.
5-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
To manage the cluster, you can use either the Failover Cluster Management tool
or the cluster.exe command line tool. Typically, you will use the Failover Cluster
Management tool; however, in addition to management tasks, you might use the
command line tool for in-depth troubleshooting and logging activities.
You must use the cluster.exe tool from a computer running the Server Core
installation option of Windows Server 2008. To have the failover cluster
management tools available, you must install the Failover Clustering feature,
or the Remote Server Administration Tools, on a Windows Server 2008 or a
Windows Vista® Service Pack 1 computer.
Configuring a Failover Cluster 5-9
Key Points
You can perform several failover cluster management tasks. These tasks range from
adding and removing cluster nodes, to modifying the quorum settings.
5-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
Usually, the Create a Cluster Wizard chooses the quorum mode for your
environment based on the failover cluster configuration. However, you might
need to modify quorum settings using the Configure Cluster Quorum Wizard,
in the Failover Cluster Management tool. If you have an Exchange Server 2007
Continuous Copy Replication cluster configuration, you will need to manually set
the quorum mode to Node and File Share Majority.
Configuring a Failover Cluster 5-11
Key Points
There are three aspects to managing cluster nodes:
• You can add a node to an established failover cluster by selecting Add Node
from the Failover Cluster Management Actions pane. The Add Node Wizard
will prompt you for information about the additional node.
• Pausing a node prevents resources from being failed over or moved to the
node. You typically pause a node when a node is undergoing maintenance or
troubleshooting.
• Evicting is an irreversible process for a cluster node. Once you evict the node,
it must be re-added to the cluster. You evict nodes when a node is damaged
beyond repair, or is no longer needed in the cluster. If you evict a damaged
node, you can repair or rebuild it, and then add it back to the cluster using the
Add Node Wizard.
5-12 Implementing and Managing Windows Server 2008 Clustering
Question:
This lesson describes how to specify when and how failovers occur.
Configuring a Failover Cluster 5-15
Key Points
You need to recognize situations that trigger failover, so you can act accordingly.
For example, if an application failure causes a failover, you must determine the
root cause of the application failure before allowing the application to fail back to
the original cluster node. In all failure cases, to maintain the failover cluster’s
stability, you must repair unstable nodes before they can host resources again.
5-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
You can create a preferred owners list in order to control which nodes host
clustered applications. Using preferred owners ensures that applications run on
server nodes where they will perform the best, and not interfere with other
clustered applications.
For example, in an Active/Active cluster with two clustered file servers, one
clustered file server will have one cluster node as the preferred owner, and the
other clustered file server will have the other node as a preferred owner. This
spreads the load across the cluster nodes.
For more information, see "Failover Cluster Help: Modifying Settings for
a Clustered Service or Application"
Configuring a Failover Cluster 5-17
Key Points
You might need to adjust the preferred owners and failback settings in order to
control how the cluster responds when the application or service fails. You only
use these settings when a clustered application or service failure occurs. You
change these settings from the clustered application property sheets.
The following table includes sample failover and failback settings.
Settings Effect
General tab, Preferred owner: If the service or application fails over from Node1 to
Node1 Node2, when Node1 is again available, the service or
Failover tab, Failback setting: application will fail back to Node1.
Allow failback (Immediately)
Failover tab, Maximum failures In a six-hour period, if the application or service fails
in the specified period: 4 10 times or less, it will be restarted or failed over each
Failover tab, Period (hours): 6 time. If the application or service fails a fifth time in
the six-hour period, it will be left in the failed state.
5-18 Implementing and Managing Windows Server 2008 Clustering
Key Points
Failovers are not only caused by hardware or software failures; you might need to
perform a manual failover to maintain or troubleshoot nodes. The manual process
moves a clustered application from one node to another.
5-20 Implementing and Managing Windows Server 2008 Clustering
Question:
Lab Objectives
• Create a cluster
• Manage a failover cluster
Scenario
Fabrikam has two servers that provide clustered file services. You will complete the
deployment of Fabrikam’s failover cluster. The domain controller is available to act
as the cluster’s file share witness.
5-22 Implementing and Managing Windows Server 2008 Clustering
This module describes how to configure cluster resources and server roles using
both full server and Server Core installations.
Configuring Cluster Resources and Server Roles 6-3
Key Points
In a failover cluster, cluster resources provide functionality to the clustered
applications. Resources are the building blocks of a clustered application or a
service. When you configure resources together as a clustered application, they can
fail between nodes as a group.
Configuring Cluster Resources and Server Roles 6-5
Key Points
Windows Server®°2008 provides resources that you can use to cluster functions
and features. For example, you can cluster IP version 6 (IPv6) addresses, file
services, and virtual machines. Applications such as Microsoft Exchange Server
and Microsoft SQL Server® add special resources that support the operation of the
application.
6-6 Implementing and Managing Windows Server 2008 Clustering
Key Points
Resource dependencies perform the following:
• Determine the order in which resources can start and stop. For example, if the
network name resource is dependent on an IP address resource, the network
name must stop before the IP address stops.
• Define bindings. For example, if a network name resource is dependent on an
IP address resource, then that address will be registered in both the Domain
Name System (DNS), and Windows Internet Name Service (WINS), with that
associated network name resource.
Configuring Cluster Resources and Server Roles 6-7
Key Points
During failover, each of the resources is taken offline in the order set by the
dependencies. After all resources are offline, the cluster moves the entire group
ownership from the active node to the passive node. After the group ownership
changes, the resources are started in the reverse order, which ensures that
dependent resources are started first. The length of time for a failover depends on
the cluster node’s transaction load, and the resources that are being failed-over.
Key Points
You can customize a resource’s response to a failover. For example, if a disk
resource fails on the active node, you might want the entire group of resources to
fail over to another node. Or you might want to customize the resource policies to
attempt to restart the resource several times before failing the service or application
over to another cluster node.
For more information, see "Failover Clustering Help: User Interface: The
Failover Cluster Management Snap-In <Resource> Properties: Policies
Tab"
Configuring Cluster Resources and Server Roles 6-9
Resources are one part of a clustered service or application. You group resources to
provide a service or application on a failover cluster.
This lesson describes how to implement a clustered application.
6-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
A clustered application is a group of resources that provide a service. A file server, a
database server that provides a service, and an application, are all examples of
clustered applications.
Before you cluster a Windows Server 2008 role, you use Server Manager to install
the role on each cluster node. Upon installation, these roles add the resources that
are required to support that role.
Configuring Cluster Resources and Server Roles 6-11
Key Points
To cluster an application or service, you first create a cluster, and on each
node, install the roles that will be clustered. Then, using the Failover Cluster
Management snap-in, you cluster the role. After you cluster the application, you
can configure the failover policy and the application.
6-12 Implementing and Managing Windows Server 2008 Clustering
Key Points
Answer the questions in a classroom discussion.
Configuring Cluster Resources and Server Roles 6-13
Key Points
Although you can cluster a number of roles and features, certain roles are more
commonly clustered. These roles are more common because they are critical to
business operations. For example, a file server can store data critical to a specific
business process; therefore, it is an excellent candidate for clustering. You can also
cluster other applications that are not server roles, such as e-mail and database
services, in order to provide high availability.
Question:
Server Core provides a lower overhead environment for clustering. This leaves
more resources available, and requires fewer updates to the nodes. This lesson
describes how to implement a Server Core failover cluster, which is similar to a full
installation, but has limitations and uses different methods for the configuration.
6-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
Server Core requires less maintenance than a full installation, because it installs
fewer components. It also has a reduced attack surface, which allows fewer
components to be exploited, and requires fewer updates if vulnerabilities are
discovered. Server Core also requires less disk space and memory, which frees
more resources for the clustered applications and services.
Key Points
You can use the following command-line tools to manage Server Core:
Tool Description
OcSetup.exe Installs or removes optional Windows components.
OcList.exe Lists the server roles and optional features that are
available for use with Ocsetup.exe.
Tool Description
Key Points
A number of roles and functions can benefit from Server Core. One of the main
reasons to use Server Core is the Hyper-V feature. With its reduced memory, CPU,
and hard-drive resource overhead, Server Core can provide additional resources
to run the virtual machines. Additionally, other roles such as Print Services, File
Services, DHCP Server, and DNS Server can also benefit from running on Server
Core.
Question:
In what other instances might you want to install roles using ocSetup.exe?
Configuring Cluster Resources and Server Roles 6-21
Question:
Installing the Failover Clustering feature, creating the cluster, and clustering the file
services are some of the steps you need to take to create a failover cluster on Server
Core. What other steps do you need to take when you create a failover cluster on
Server Core?
6-22 Implementing and Managing Windows Server 2008 Clustering
Lab Objectives
• Cluster the Print Services role using Failover Cluster Management.
• Configure cluster resources.
• Cluster the File Services role on Windows Server Core.
Scenario
Use the scenarios on the following pages to complete the lab exercises.
Configuring Cluster Resources and Server Roles 6-23
f Task 10: Log on to VAN-DC1, and verify the clustered shared folder
1. Log on to VAN-DC1 as Administrator with the password Pa$$w0rd.
2. Click Start, click Run, and then type \\VAN-File01\Public.
3. In Windows Explorer, verify that the user can connect to the share.
f Task 11: Close all virtual machines and discard undo disks
1. For each running virtual machine, close the Virtual Machine Remote Control
window.
2. In the Close box, select Turn off machine and discard changes, and then
click OK.
3. Close the 6423A Lab Launcher.
MCT USE ONLY. STUDENT USE PROHIBITED
Maintaining Microsoft Failover Clusters 7-1
After deploying a failover cluster, you will need to maintain the cluster, and
troubleshoot when issues arise.
This module provides information on how to monitor cluster functionality,
troubleshoot the cluster by reviewing logs to identify and solve problems, and
adjust settings.
Maintaining Microsoft Failover Clusters 7-3
Failover clusters are more complex than stand-alone servers. This lesson describes
how to use standard cluster monitoring tools such as the Event Viewer, and the
Reliability and Performance Monitor snap-in.
7-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
A variety of tools are available to help you monitor failover clusters. You can
use standard Windows Server®°2008 tools, such as the Event Viewer and the
Reliability and Performance Monitor snap-in, to review cluster event logs, and
performance metrics. You can also use Cluster.exe and Tracerpt.exe to export data
for analysis. Additionally, you can use the MHTML-formatted cluster configuration
reports and the Validate a Configuration Wizard to troubleshoot problems with the
cluster configuration and hardware changes.
For more information, see "Failover Clustering Help: View Events and
Logs for a Failover Cluster"
Maintaining Microsoft Failover Clusters 7-5
Key Points
When problems arise in the cluster, use the Event Viewer to view events with a
Critical, Error, or Warning severity level.
Additionally, informational level events are logged to the Failover Clustering
Operations log, which can be found in the Event Viewer in the Applications and
Services Logs\Microsoft\Windows folder. Informational-level events are usually
common cluster operations, such as cluster nodes leaving and joining the cluster,
or resources going offline or coming online.
For more information, see "Failover Clustering Help: View Events and
Logs for a Failover Cluster"
7-6 Implementing and Managing Windows Server 2008 Clustering
Key Points
In previous Windows Server versions, event logs were replicated to each node in
the cluster. This simplified cluster troubleshooting, because you could review all
event logs on a single cluster node. Windows Server 2008 does not replicate the
event logs between nodes; however, the Failover Cluster Management snap-in has
a Cluster Events option that enables you to view and filter events across all cluster
nodes. This feature is helpful in correlating events across cluster nodes.
The Failover Cluster Management snap-in also provides a Recent Cluster Events
option that will query all of the Error and Warning events from all of the cluster
nodes in the last 24 hours.
You can access additional logs, such as the Debug and Analytic logs, from the
Event Viewer. To display these logs, modify the view in the top menu by selecting
Show Analytic and Debug Logs.
For more information, see "Failover Clustering Help: View Events and
Logs for a Failover Cluster"
Maintaining Microsoft Failover Clusters 7-7
Key Points
Windows event tracing is a kernel component that is available early after startup,
and late into shutdown. It is designed to allow for fast tracing and delivery of
events to trace files and to consumers. Because it is designed to be fast, it allows
only basic in-process filtering of events based on event attributes.
The event trace log contains a comprehensive accounting of the failover cluster
actions. Depending on how you want to view the data, use either Cluster.exe or
Tracerpt.exe to access the information in the event trace log.
For more information, see "Failover Clustering Help: View Events and
Logs for a Failover Cluster"
7-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
Tracerpt.exe will parse the event trace logs only on the node on which it is run. All
of the individual logs are collected in a central location. To transform the XML file
into a text file or an HTML file that can be opened in Internet Explorer, you can
parse the XML-based file using the Microsoft XSL parsing command-line utility
msxsl.exe, and an XSL style sheet.
Maintaining Microsoft Failover Clusters 7-9
Key Points
Cluster.exe provides a cluster log for the entire cluster, or for a specific node in the
cluster in a readable text format. The output from Cluster.exe is more flexible than
Tracerpt.exe, because you can export trace logs for all nodes in the cluster and save
them in a specified location. You can view the exported log using a text editor. To
run successfully, Cluster.exe must be run from an administrator-elevated
command prompt.
7-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
The Performance and Reliability Monitor snap-in allows you to:
• Trend application performance on each node. To determine how an
application is performing, you can view and trend specific information on
system resources that are being used on each node.
• Trend application failures and stability on each node. You can pinpoint when
application failures occur and correlate the application failures with other
events on the node.
• Modify trace log settings. You can start, stop, and adjust trace logs, including
their size and location.
Maintaining Microsoft Failover Clusters 7-11
This lesson describes how to backup a failover cluster, and how to perform an
authoritative or a non-authoritative restore.
7-12 Implementing and Managing Windows Server 2008 Clustering
Key Points
When you back up the cluster configuration, be aware of the following:
• Before putting a cluster into production, you must test your backup and
recovery process.
• If you choose to use Windows Server Backup, (the Windows Server 2008
backup feature), you must first add the feature. You can do this during initial
configuration tasks, or by using the Add Features Wizard in Server Manager.
Maintaining Microsoft Failover Clusters 7-13
Key Points
Windows Server Backup is the built-in backup and recovery software for Windows
Server 2008. In order to complete a successful backup, be aware of the following:
• For a backup to succeed in a failover cluster, the cluster must be running
and must have quorum. In other words, enough nodes must be running and
communicating (perhaps with a witness disk or witness file share, depending
on the quorum configuration,) that the cluster has achieved quorum.
• You must back up all clustered applications. If you cluster a Microsoft SQL
Server® database, you must have a backup plan for the databases and
configuration outside of the cluster configuration.
7-14 Implementing and Managing Windows Server 2008 Clustering
Question:
What option must you use in Windows Server Backup to back up the failover
cluster configuration?
7-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
Use a non-authoritative restore when a single node in the cluster has been
damaged or rebuilt, and the rest of the cluster is operating properly.
Perform a non-authoritative restore by restoring the system recovery (system state)
information to the damaged node. When you restart that node, it will join the
cluster and receive the latest cluster configuration automatically.
Key Points
Use an authoritative restore when the cluster configuration must be rolled back to
a previous point in time. For example, you would use an authoritative restore if an
administrator accidently removed clustered resources or modified other cluster
settings.
Perform the authoritative restore by stopping the cluster resource on each of the
nodes, and then performing a system recovery (system state) on a single node
using the command-line Windows Server Backup interface. After the restored node
restarts the cluster service, the remaining cluster nodes can also start the cluster
service.
7-18 Implementing and Managing Windows Server 2008 Clustering
Questions:
Key Points
Answer the questions in a classroom discussion.
7-20 Implementing and Managing Windows Server 2008 Clustering
Key Points
To troubleshoot a failover cluster:
• Use the Validate a Configuration Wizard to highlight configuration issues that
might cause cluster problems.
• Review cluster events and trace logs to identify application or hardware issues
that might cause an unstable cluster.
• Review hardware events and logs to help pinpoint specific hardware
components that might cause an unstable cluster.
• Review storage area network (SAN) components, switches, adapters, and
storage controllers to help identify any potential problems.
7-22 Implementing and Managing Windows Server 2008 Clustering
Key Points
When troubleshooting failover clusters, you must:
• Identify the perceived problem by gathering and documenting the symptoms
of the problem.
• Identify the scope of the problem so that you can understand what is being
affected by the problem, and what impact that effect is having on the
application and the clients.
• Gather information so that you can accurately understand and pinpoint the
possible problem. After you identify a list of possible problems, you can
prioritize them by probability, or the impact of a repair. If the problem cannot
be pinpointed, you should attempt to recreate the problem.
• Create a schedule for repairing the problem. For example, if the problem only
affects a small subset of users, you can delay the repair to an off-peak time so
that you can schedule downtime.
• Complete and test each repair one at a time so that you can identify the fix.
Maintaining Microsoft Failover Clusters 7-23
Key Points
Answer the questions in a classroom discussion.
7-24 Implementing and Managing Windows Server 2008 Clustering
Key Points
To troubleshoot SAN issues, start by checking physical connections and each of
the hardware component logs. Additionally, run the Validate a Configuration
Wizard to verify that the current cluster configuration is still supportable. When
you run the Validate a Configuration Wizard, ensure that the storage tests you
select can be run on an online failover cluster. Several of the storage tests cause
loss of service on the clustered disk when the tests are run.
Maintaining Microsoft Failover Clusters 7-25
Key Points
To troubleshoot group and resource failures:
• Use the Dependency Viewer in the Failover Cluster Management snap-in to
identify dependent resources.
• Check the Event Viewer and trace logs for errors from the dependent
resources.
• Determine if the problem only happens on a specific node, or nodes, by trying
to recreate the problem on different nodes.
7-26 Implementing and Managing Windows Server 2008 Clustering
Lab Objectives
• Monitor failover clusters.
• Perform backups on a failover cluster.
• Perform an authoritative restore on a failover cluster.
Scenario
Fabrikam has asked you to review cluster event logs and perform a test restore to
validate the backup process for their existing file cluster. The Windows Server
Backup feature is installed on both cluster nodes.
Note: Complete the steps to prepare for this lab after starting the virtual machines
and logging on as described in Exercise 1, Task 1.
Maintaining Microsoft Failover Clusters 7-27
f Task 2 : Identify the cluster events that occurred during the last restart
1. On VAN-SRV6A, open the Event Viewer from the Administrative tools.
2. In the Event Viewer, browse to Applications and Services Logs\Windows
\FailoverClustering\Operational to view the operational logs.
3. In the list of events, locate and review events with the Event IDs 1061, 1125,
and 1131.
7-30 Implementing and Managing Windows Server 2008 Clustering
Single-site failover clusters allow applications to stay available when a server fails,
or even if the cabinet that contains each node fails.
This module describes how multi-site clusters work with Windows Server®°2008,
and provides general guidelines for planning a multi-site cluster.
Implementing Multi-Site Clusters 8-3
This lesson provides information about multi-site clusters, and how they are used.
Additionally, this lesson describes the two replication types that are available for
creating a multi-site cluster.
8-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
A multi-site cluster provides highly available services in more than one location.
Multi-site clusters can solve a number of specific problems; however, they also
present specific challenges.
Implementing Multi-Site Clusters 8-5
Key Points
Answer the questions in a classroom discussion.
8-6 Implementing and Managing Windows Server 2008 Clustering
Key Points
In a multi-site cluster, each site usually has a separate storage system with
replication between the sites. Multi-site cluster storage replication allows each site
to be independent, and provides fast access to the local disk. With separate storage
systems, you cannot share a single disk between sites.
Implementing Multi-Site Clusters 8-7
Key Points
A multi-site cluster has three main advantages in a failover site compared to a
remote server:
• When a site fails, a multi-site cluster automatically fails over the clustered
service or application to another site.
• Because the cluster configuration is automatically replicated to each cluster
node in a multi-site cluster, there is less administrative overhead than a cold
standby server, which requires that you manually replicate changes.
• The automated processes in a multi-site cluster reduce the possibility of
human error, which is present in manual processes.
8-8 Implementing and Managing Windows Server 2008 Clustering
Key Points
Because of increased cost and complexity of a multi-site failover cluster, it might
not be an ideal solution for every application or business. When considering
whether to deploy a multi-site cluster, you should evaluate the importance of the
applications to the business, the type of applications, and any alternative solutions.
Some applications can provide multi-site redundancy easily with log shipping or
other processes, and can still achieve adequate availability with only a modest
increase in cost and complexity.
The complexity of a multi-site cluster requires a higher level of architectural and
hardware planning. It also requires that you develop business processes to
routinely test the cluster functionality.
Implementing Multi-Site Clusters 8-9
Key Points
In a multi-site cluster, Windows Server 2008 failover clustering does not include
built-in functionality to replicate data between sites. There are three options for
replicating data: block level hardware-based replication, software-based file
replication installed on the host, or application-based replication.
Multi-site data replication can be either synchronous or asynchronous.
Synchronous replication does not acknowledge data changes made in Site A until
the data is successfully written to Site B. With asynchronous replication, data
changes made in Site A are eventually written to Site B.
When you deploy a multi-site cluster and run the Validate a Configuration Wizard,
the disk tests will not find any shared storage, and will therefore not run. If you
follow the hardware manufacturer’s recommendations for Windows Server 2008
failover clustering hardware, Microsoft will support the solution.
8-10 Implementing and Managing Windows Server 2008 Clustering
Key Points
When you use synchronous replication, the host receives a “write complete”
response from the primary storage after the data is written successfully on both
storage systems. If the data is not written successfully to both storage systems, the
application must attempt to write to the disk again. With synchronous replication,
both storage systems are identical.
Implementing Multi-Site Clusters 8-11
Key Points
When you use asynchronous replication, the node receives a “write complete”
response from the storage after the data is written successfully on the primary
storage. The data is written to the secondary storage on a different schedule,
depending on the hardware or software vendor’s implementation.
Asynchronous replication can be storage-based, host-based, or even application-
based; however, not all forms of asynchronous replication are sufficient for a multi-
site cluster. For example, Distributed File System Replications (DFS-R) provides
file-level asynchronous replication; however, it does not support multi-site failover
clustering replication. This is because DFS-R replicates smaller documents that are
not held open continuously, and therefore, it was not designed for high-speed,
open-file replication.
8-12 Implementing and Managing Windows Server 2008 Clustering
Key Points
You use synchronous and asynchronous replication for different reasons. You
use synchronous replication when data loss cannot be tolerated. Synchronous
replication solutions cannot tolerate high-disk write latency, because the
application waits for both storage solutions to acknowledge the data writes. The
requirement for low latency disk writes also limits the distance between the storage
systems because increased distance can cause higher latency. If the disk latency is
high, the performance and even the stability of the application can be affected.
Asynchronous replication overcomes latency and distance limitations by
acknowledging local disk writes only, and by reproducing the disk write on the
remote storage system in a separate transaction. Because asynchronous replication
writes to the remote storage system after it writes to the local storage system, the
possibility of data loss during a failure is increased.
Implementing Multi-Site Clusters 8-13
Key Points
Windows Server 2008 allows cluster nodes to exist on different IP subnets, which
enables a clustered application or service to change its IP address based on the IP
subnet. DNS updates the clustered application’s DNS record so that clients can
locate the IP address change. Because clients rely on DNS to find a service or
application after a failover, you might need to adjust the DNS records’ time to live,
and the speed at which DNS data is replicated. Additionally, when cluster nodes
are in multiple sites, network latency might require that you modify the inter-node
communication (heartbeat) delay and time-out thresholds.
8-14 Implementing and Managing Windows Server 2008 Clustering
This lesson describes how to choose the correct quorum mode, and how to deploy
a multi-site cluster.
Implementing Multi-Site Clusters 8-15
Key Points
In addition to multi-subnet clustering and using inter-node communications to
accommodate network latencies, Windows Server 2008 provides new cluster
quorum modes (Node Majority and Node and File Share Majority) that help
enable multi-site clusters.
8-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
Answer the questions in a classroom discussion.
Implementing Multi-Site Clusters 8-17
Key Points
Both the Node and Disk Majority, and No Majority: Disk Only quorum modes
require a shared witness disk to provide a vote for determining quorum. You
should only use these two quorum modes if the hardware vendor specifically
recommends and supports them.
To use the Node and Disk Majority and No Majority: Disk Only modes in a multi-
site cluster, the shared disk requires that:
• You preserve the semantics of the Small Computer System Interface (SCSI)
commands across the sites, even if a complete communication failure occurs
between sites.
• You replicate the witness disk in real-time synchronous mode across all sites.
8-18 Implementing and Managing Windows Server 2008 Clustering
Key Points
If you are using Node Majority and the sites lose communication, you need a
mechanism to determine which nodes stay up, and which nodes drop out of
cluster membership. The second site requires another vote to obtain quorum after
a failure. To obtain another vote for quorum, you must join another node to the
cluster, or create a file share witness.
8-20 Implementing and Managing Windows Server 2008 Clustering
Key Points
The Node and File Share Majority mode can help maintain quorum without
adding another node to the cluster. To provide for a single site failure and allow
automatic failover, the file share witness might need to exist at a third site. In a
multi-site cluster, a single server can host the file share witness; however, you must
create a separate file share for each cluster.
Implementing Multi-Site Clusters 8-21
Key Points
In a Windows Server 2008 multi-site cluster, you can install Exchange Server 2007
Service Pack 1 in a Cluster Continuous Replication configuration. To keep the data
on both cluster nodes synchronized without needing expensive storage replication,
Exchange Server provides log shipping. You use three sites to ensure that a single
site failure does not remove the ability to maintain quorum. If you cannot use three
sites, you can manually move the file share witness to the secondary site, and
maintain quorum during a failure.
Key Points
Multi-site clusters are not appropriate for every application or every business.
When you design a multi-site solution with a hardware vendor, clearly identify the
business requirements and expectations. Additionally, carefully consider the
quorum mode you will use, and the location of the available cluster votes.
Implementing Network Load Balancing Clusters 9-1
NLB is a Windows Server 2008 solution that provides availability and scalability.
This lesson provides information about the NLB features, and explains how NLB
operates.
9-4 Implementing and Managing Windows Server 2008 Clustering
Key Points
Applications that can benefit from being deployed on a network load balancing
cluster include Web server-based applications, FTP, firewall, proxy, and virtual
private network (VPN) services.
Because any server in a Network Load Balancing cluster can respond to a client
request, both the application files and data on all servers must be exactly the same.
Usually this means that you use a separate, back-end server to store data, or
provide for a way to synchronize the data on the Web servers. However, this
requirement limits which applications are suitable for load balancing.
Note: To provide a highly available system, you should consider using a failover
cluster for the back-end server hosting the data.
Implementing Network Load Balancing Clusters 9-5
Key Points
An NLB cluster uses a virtual IP address and a virtual media access control (MAC)
address for network communication. All NLB nodes use the same virtual IP
address and virtual MAC address. When an incoming packet is addressed to the
virtual IP address and virtual MAC address, all NLB nodes receive it, but only the
appropriate node responds.
NLB nodes use a common algorithm to determine which node will respond. When
a client request arrives, all hosts simultaneously perform a mapping to determine
which host should handle the packet. The mapping uses a randomization function
that calculates a host priority based on the IP address, port, and other information.
The corresponding host accepts the client request, and the other cluster hosts
discard it.
Implementing Network Load Balancing Clusters 9-7
If the failed host begins to send heartbeat messages again, it rejoins the cluster
during the convergence. After all cluster hosts agree on the current cluster
membership, the client load is redistributed to the remaining hosts, and the
convergence completes.
Key Points
All NLB components support IPv6 addresses. You can configure the addresses
as the primary cluster IP address, the dedicated IP addresses, and the virtual IP
addresses. In addition, you can load balance IPv6 as native IPv6 and in the IPv6
over IPv4 modes.
Network Driver Interface Specification (NDIS) 6.0 is compatible with earlier NDIS
versions.
By using a callback interface, NLB can detect and notify applications during an
attack, or when a node is under excessive load. Microsoft® Internet Security and
Acceleration (ISA) Server uses this functionality if the cluster node is overloaded or
being attacked.
NLB supports rolling upgrades from Windows Server®°2003 to Windows
Server 2008.
Implementing Network Load Balancing Clusters 9-9
Key Points
Applications that work well with NLB have one or more of the following
characteristics:
• The applications do not require the same cluster host to service all client
connections. For example, most Web sites are ideal applications for an NLB
cluster because any cluster host can service the client requests. On the other
hand, applications based on Secure Sockets Layer (SSL) expect that the cluster
host that originated the response will maintain the SSL session.
• The applications do not require application state to be saved on a specific
server in the cluster, such as session state saved by Active Server Pages (ASP).
Web applications that work well on a cluster use a common database or
service to maintain session state. On the initial request, a cookie is sent to the
client. On subsequent requests, the client sends the cookie to the cluster. Any
cluster host in the cluster can restore the application session state by using the
cookie sent by the client to retrieve the session state from the session database
or service.
Implementing Network Load Balancing Clusters 9-11
Key Points
When you synchronize data between NLB nodes, there is always a small amount of
lag-time in replication. This lag-time may cause brief inconsistent results between
servers. Data synchronization is a common approach when multiple servers host a
Web site, and minor discrepancies during Web site updates are acceptable.
When you store data in a common location, the responses from the NLB nodes
are always consistent, because all nodes have the same view of the data. When
you run Web-based applications, a common storage location is required to ensure
that changes result in a consistent data structure. To do this, many Web-based
applications store data in a Microsoft SQL Server® database. You must configure all
hosts in the NLB cluster to communicate with a single database server, or with a
clustered SQL Server.
Internet Information Services (IIS) 7.0 supports shared configuration, which
enables you to store the data used by multiple Web servers on a separate file
server. You can use shared configuration in combination with NLB.
Implementing Network Load Balancing Clusters 9-13
Key Points
Answer the questions in a classroom discussion.
9-14 Implementing and Managing Windows Server 2008 Clustering
Key Points
Answer the questions in a classroom discussion.
Implementing Network Load Balancing Clusters 9-15
In order to implement NLB in Windows Server 2008, you must install the NLB
feature, and then create and configure NLB clusters using the Network Load
Balancing Manager. This lesson describes how to set up and configure NLB
clusters.
9-16 Implementing and Managing Windows Server 2008 Clustering
Key Points
We recommend that all hosts in an NLB cluster have two network adapters. This
simplifies network communication by isolating NLB cluster traffic from host traffic.
Although NLB in Windows Server 2008 supports a single network adapter for
both unicast and multicast cluster operation modes, multiple network adapters
provide enhanced redundancy and performance.
You must configure the your network adapter for NLB with TCP/IP, and all
NLB nodes must be on the same subnet. This is required because all nodes in
the cluster share a single virtual IP address. If the cluster hosts are not connected
to the same physical subnet, ensure that they are connected to a virtual subnet.
If the load balancing cluster is operating close to capacity, the loss of a server may
increase the load on remaining cluster nodes, and subsequently degrade client
performance. To avoid performance degradation due to a failed server, ensure that
each host in the load balancing cluster has sufficient capacity to handle the added
load.
Implementing Network Load Balancing Clusters 9-17
Questions:
Key Points
The cluster IP address is the virtual IP address that is assigned to the cluster. Client
requests are sent to the cluster IP address.
When you enter the fully qualified domain name (FQDN) in the cluster
parameters, the name is not automatically registered in a Domain Name System
(DNS), or used by other Windows components. Not assigning a full Internet name
setting does not affect the operation of any Windows component, including NLB.
Implementing Network Load Balancing Clusters 9-19
Will each server host the same number Determines the port rule load weight
of client connections? setting
Will the client connect to the same Determines the affinity setting
server for all connections?
Will the cluster operate the same way Determines the port rule settings
for all applications?
Key Points
We recommend that you select the unicast method for distributing client requests,
unless only one network adapter is installed in each cluster host.
Considerations for using unicast mode include:
• In order to provide peer-to-peer communication between cluster hosts, a
second network adapter is required.
• If you connect the cluster to a switch, incoming packets are sent to all the
ports on the switch, which may cause switch flooding. Outgoing packets use a
modified version of the cluster MAC address that is unique to each node. This
prevents switches from associating the cluster MAC address to a specific
switch port.
Implementing Network Load Balancing Clusters 9-21
Key Points
For each NLB node, you can configure host parameters that are specific to that
node. The host parameters are:
• Priority: Each NLB node must have a unique priority value. If the port rule
does not address traffic addressed to the cluster, the NLB node with the
lowest-priority value will respond. If you specify the same cluster host priority
for two cluster hosts, the last cluster host that starts will not join the cluster,
and an error message that describes the problem is sent to the Windows
system event log. The existing cluster hosts continue to operate as before.
• Dedicated IP address: This specifies the IP address that is used on the host for
cluster management or general computer management. NLB ensures that port
rules do not affect traffic to the dedicated address. You must list the dedicated
IP address first in the list of IP addresses, in the TCP/IP properties.
Implementing Network Load Balancing Clusters 9-23
Key Points
Port rules specify how requests to a certain IP address and port range are directed
to cluster hosts. This enables you to define different rules for different applications
running on the NLB cluster. For example, you might want to distribute requests
for a Web application on TCP port 80 evenly among servers, but direct all requests
for the Web application on TCP port 8080 to a single server.
Windows NLB uses the filtering mode in a port rule to determine how to distribute
requests among nodes in the NLB cluster. The options for filtering mode are:
• Multiple hosts: All NLB nodes respond based on the weight assigned to each
node. The weight for each node is assigned depending on the physical
capabilities of the node. If a node fails, then other nodes continue to respond.
Distributing requests among multiple nodes increases the availability and
scalability of an application.
Implementing Network Load Balancing Clusters 9-25
Important: The port rules that you apply to each cluster host must be identical,
with the exception of the load weight (in the multiple hosts filter mode,) and the
handling priority (in the single hosts filter mode). If there is a discrepancy between
port rules on existing cluster hosts, the cluster will not converge.
9-26 Implementing and Managing Windows Server 2008 Clustering
Key Points
The affinity option controls distribution of requests from a specific client among
the nodes in an NLB cluster. This setting is only relevant when you select the
multiple-hosts filtering mode. The options for affinity are:
• None: Any node can respond to any client request. This is suitable for stateless
applications such as static Web pages. The first Web page can be retrieved
from Node A, and the second Web page from Node B.
• Single: A single node responds to all requests from a single client. This is
required for state-based applications such as Web applications that you must
authenticate. The authentication state is only stored on the server that you
originally contacted. If requests are answered by any node other than the
authenticating node, the request is unsuccessful, or the client is forced to
authenticate again.
Implementing Network Load Balancing Clusters 9-27
Questions:
You want one host in the NLB cluster to handle all traffic sent to a Web application
using port 8888. How would you configure the cluster parameters?
You have deployed a new server in an existing NLB cluster and want to be sure
that this server will service twice as many client connections as the existing servers.
How would you configure the NLB cluster?
Implementing Network Load Balancing Clusters 9-29
Key Points
When you use Network Load Balancing Manager to configure cluster parameters,
you enter the cluster parameters once during the creation of the cluster. As you
add cluster hosts to the cluster, Network Load Balancing Manager automatically
configures the cluster parameters for the new cluster hosts.
9-30 Implementing and Managing Windows Server 2008 Clustering
After deployment, you must manage the NLB cluster by modifying the cluster
settings, and using the Network Load Balancing Manager to manage hosts in the
cluster. This lesson describes how to manage and troubleshoot NLB clusters.
Implementing Network Load Balancing Clusters 9-31
Questions:
What happens when you delete a host from the Network Load Balancing Manager?
After you enable logging in the Network Load Balancing Manager, what types of
events will show up in the NLB log file?
9-32 Implementing and Managing Windows Server 2008 Clustering
Question:
You must remove a host from an NLB cluster. What should you do before you
remove the host so that you do not disrupt client connections?
Note: You can perform many of these tasks by using the command-line tool
NLB.exe. For details, see "Controlling Hosts on Network Load Balancing Clusters" in
Network Load Balancing Online Help.
Implementing Network Load Balancing Clusters 9-33
Key Points
Troubleshooting an NLB cluster usually falls into one of two categories: client
connectivity issues, and NLB load balance issues. The following tools help you to
troubleshoot an NLB cluster:
• Use the standard network troubleshooting tools on the client, and use
Network Monitor on the servers to troubleshoot client connectivity issues.
• Monitor the events in the System Log in Event Viewer to troubleshoot NLB
cluster issues such as cluster convergence issues, or server performance issues.
• Use the Reliability and Performance Monitor to troubleshoot performance
issues.
9-34 Implementing and Managing Windows Server 2008 Clustering
Key Points
When troubleshooting an NLB cluster, consider the following issues:
Lab Objectives
• Prepare the NLB cluster nodes.
• Configure an NLB failover cluster.
Scenario
You are the server administrator for Fabrikam Inc., and are implementing a new
Web-based application. To increase the availability of this new application, you will
use NLB in Windows Server 2008.
9-38 Implementing and Managing Windows Server 2008 Clustering
f Task 3: Configure a Domain Name System (DNS) record for the NLB
cluster
• On VAN-DC1, in DNS, create a new host record for WebApp.fabrikam.com, in
the Fabrikam.com zone. Configure an IP address of 10.10.0.100.
Your evaluation of this course will help Microsoft understand the quality of your
learning experience.
Please work with your training provider to access the course evaluation form.
Microsoft will keep your answers to this survey private and confidential and will
use your responses to improve your future learning experience. Your open and
honest feedback is valuable and appreciated.