Dell SE500W User Manual 2

Dell™ PowerEdge™ Cluster

SE500W Systems

Installation and

Troubleshooting Guide

www.dell.com | support.dell.com

Notes, Notices, and Cautions

NOTE: A NOTE indicates important information that helps you make better use of your computer.

NOTICE: A NOTICE indicates either potential damage to hardware or loss of data and tells you how to avoid the problem.

CAUTION: A CAUTION indicates a potential for property damage, personal injury, or death.

Abbreviations and Acronyms

For a complete list of abbreviations and acronyms, see "Abbreviations and Acronyms."

__________________

Information in this document is subject to change without notice.
© 2005 Dell Inc. All rights reserved.

Reproduction in any manner whatsoever without the written permission of Dell Inc. is strictly forbidden.
Trademarks used in this text: Dell, the DELL logo, PowerEdge, PowerVault, and Dell OpenManage are trademarks of Dell Inc.;

Microsoft, Windows, and Windows NT are registered trademarks of Microsoft Corporation, and Windows Server is a trademark of Microsoft
Corporation.

Other trademarks and trade names may be used in this document to refer to either the entities claiming the marks and names or their products.
Dell Inc. disclaims any proprietary interest in trademarks and trade names other than its own.

June 2005

Contents

1 Introduction

Intended Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Obtaining Technical Assistance

Overview

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Virtual Servers and Resource Groups
Quorum Resource

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Shared Storage Systems

PowerEdge™ Cluster SE500W Solution

Operating Systems
Storage System

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Hardware and Software Technologies
Cluster Components

System Requirements

Cluster Nodes
Cluster storage

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Other Documents You May Need

. . . . . . . . . . . . . . . . . . . . . . . . . 10

. . . . . . . . . . . . . . . . . . . 10

. . . . . . . . . . . . . . . . . . . . . . . . . . 11

. . . . . . . . . . . . . . . . . . . . . 11

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

. . . . . . . . . . . . . . . . . . . 13

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

. . . . . . . . . . . . . . . . . . . . . . . . 16

2 Cabling Your Cluster Hardware

Cabling for the Cluster SE500W Solution . . . . . . . . . . . . . . . . . . . . 17

Cabling One PowerVault™ 22xS Shared Storage System

to a Cluster SE500W

Cabling Two PowerVault 22xS Storage Systems

to a Cluster SE500W

Cabling Three or Four PowerVault 22xS Storage Systems

to a Cluster SE500W

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Cabling Your Public and Private Networks

Cabling Your Public Network

. . . . . . . . . . . . . . . . . . . . . . . . 22

Cabling Your Private Network
NIC Teaming

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Cabling the Mouse, Keyboard, and Monitor

. . . . . . . . . . . . . . . . . . . 21

. . . . . . . . . . . . . . . . . . . . . . . 23

. . . . . . . . . . . . . . . . . . . 24

Power Cabling for the PowerEdge Cluster SE500W Solution

. . . . . . . . . . 24

Contents 3

3 Preparing Your Systems for Clustering

Before You Begin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Installation Overview

Selecting a Domain Model

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

. . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Configuring the Cluster Nodes as Domain Controllers

Configuring the Internal Drives in Your Cluster Nodes

. . . . . . . . . . . . . 32

Installing and Configuring the Windows Operating System

Configuring Windows Networking

. . . . . . . . . . . . . . . . . . . . . . . 34

Assigning Static IP Addresses to Your Cluster
Resources and Components

Configuring IP Addresses for the Private Network
Verifying Cluster Network Communications
Configuring the Internet Connection Firewall

Installing the PERC RAID Adapters

. . . . . . . . . . . . . . . . . . . . . . . . 34

. . . . . . . . . . . . . 35

. . . . . . . . . . . . . . . . 37

. . . . . . . . . . . . . . . 37

. . . . . . . . . . . . . . . . . . . . . . . 38

Updating the Windows Driver for the PERC RAID Adapter

Installing and Configuring the Shared Storage System

Clustering PowerVault Storage Systems

. . . . . . . . . . . . . . . . . . 39

. . . . . . . . . . . . . 39

Configuring the PowerVault 22xS Storage System

for Cluster Mode
Preparing the PERC RAID Adapter for Clustering

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

. . . . . . . . . . . . . 43

Enabling the Cluster Mode Using the PERC RAID Adapter
Setting the SCSI Host Adapter IDs
Configuring and Managing Virtual Disks

. . . . . . . . . . . . . . . . . . . . . 44

. . . . . . . . . . . . . . . . . . 45

Windows 2000 and Windows Server 2003
Dynamic Disks and Volumes

. . . . . . . . . . . . . . . . . . . . . . . . 45

Naming and Formatting Drives on the
Shared Storage System

. . . . . . . . . . . . . . . . . . . . . . . . . . . 45

. . . . . . . . . . . 31

. . . . . . . . . . 32

. . . . . . . . 38

. . . . . . . . . 44

4 Contents

Installing and Configuring MSCS . . . . . . . . . . . . . . . . . . . . . . . . 48

Verifying Cluster Readiness
Creating the Quorum Resource
Configuring Cluster Networks Running Windows 2000
Configuring Cluster Networks Running Windows Server 2003

. . . . . . . . . . . . . . . . . . . . . . . . 48

. . . . . . . . . . . . . . . . . . . . . . . 49

. . . . . . . . . . 49

. . . . . . . 50

Installing and Configuring Microsoft Windows 2000
Cluster Service (MSCS)

. . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Configuring Microsoft Windows Server 2003
Cluster Service (MSCS)

Adding Cluster Nodes Using the Advanced Configuration Option
Verifying Cluster Functionality
Verifying MSCS Operation
Verifying Cluster Resource Availability
Installing Applications in the Cluster Group
Troubleshooting Failed Resources

. . . . . . . . . . . . . . . . . . . . . . . . . . . 51

. . . . . 52

. . . . . . . . . . . . . . . . . . . . . . . 53

. . . . . . . . . . . . . . . . . . . . . . . . . 53

. . . . . . . . . . . . . . . . . . . 53

. . . . . . . . . . . . . . . . 53

. . . . . . . . . . . . . . . . . . . . . 54

4 Installing Your Cluster Management Software

Microsoft Cluster Administrator . . . . . . . . . . . . . . . . . . . . . . . . . 55

Launching Cluster Administrator on a Cluster Node
Running Cluster Administrator on a Remote Console
Launching Cluster Administrator on a Remote Console
Installing Cluster Administrator for Windows Clusters

on a Remote Console

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

. . . . . . . . . . . . 55

. . . . . . . . . . . 55

. . . . . . . . . . 56

5 Using MSCS

Cluster Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Cluster Networks

Preventing Network Failure
Node-to-Node Communication

Network Interfaces

Cluster Nodes

Forming a New Cluster
Joining an Existing Cluster

Groups

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

. . . . . . . . . . . . . . . . . . . . . . . . . 58

. . . . . . . . . . . . . . . . . . . . . . . 58

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

. . . . . . . . . . . . . . . . . . . . . . . . . . . 59

. . . . . . . . . . . . . . . . . . . . . . . . . 59

Contents 5

Cluster Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Dependent Resources
Setting Resource Properties
Configuring Resource Dependencies
Setting Advanced Resource Properties
Resource Parameters
Quorum Disk (Quorum Resource)
Resource Failure
Creating a New Resource
Deleting a Resource
File Share Resource Type

. . . . . . . . . . . . . . . . . . . . . . . . . . . 61

. . . . . . . . . . . . . . . . . . . . . . . . 61

. . . . . . . . . . . . . . . . . . . 62

. . . . . . . . . . . . . . . . . . 62

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

. . . . . . . . . . . . . . . . . . . . . . 63

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

. . . . . . . . . . . . . . . . . . . . . . . . . 65

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

. . . . . . . . . . . . . . . . . . . . . . . . . . 66

Configuring Active and Passive Cluster Nodes

Failover and Failback

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Failover
Failback

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

. . . . . . . . . . . . . . . . . 67

6 Upgrading Your System to a Cluster Configuration

Certification Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Before You Begin

Configurations Using Non-Dell Products

Completing the Upgrade

Upgrading Your Operating System

Performing a Rolling Upgrade

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

. . . . . . . . . . . . . . . . . . . . 72

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

. . . . . . . . . . . . . . . . . . . . . . . . 72

. . . . . . . . . . . . . . . . . . . . . . . 73

7 Maintaining Your Cluster

Adding a Network Adapter to a Cluster Node. . . . . . . . . . . . . . . . . . 75

Changing the IP Address of a Cluster Node on the Same IP Subnet

. . . . . . 77

6 Contents

Uninstalling MSCS From Clusters Running
Windows 2000 Advanced Server

. . . . . . . . . . . . . . . . . . . . . . . . 77

Removing Nodes From Clusters Running Windows Server 2003

Running chkdsk /f on a Quorum Disk

Recovering From a Corrupt Quorum Disk

. . . . . . . . . . . . . . . . . . . . . . 78

. . . . . . . . . . . . . . . . . . . . 79

. . . . . . . . 78

Replacing a Cluster-Enabled Dell PERC RAID Adapter . . . . . . . . . . . . . 80

Replacing a Cluster Node

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

Changing the Cluster Service Account Password in
Windows Server 2003

Reformatting a Cluster Disk

Adding New Physical Drives to an Existing Shared Storage System

Rebuilding Operation in Dell OpenManage Utilities

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

. . . . . . . . . . . . . . . . . . . . . . . . . . . 82

. . . . . . 83

. . . . . . . . . . . . . . 84

Upgrading the PowerVault 22xS EMM Firmware
Using Array Manager

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

A Troubleshooting

B Cluster Data Form

PowerEdge Cluster SE500W Solution Data Form . . . . . . . . . . . . . . . . 93

Abbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . 97

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Figures

Figure 1-1. Maximum Configuration of the PowerEdge

Cluster SE500W Solution

Figure 2-1. PowerVault 22xS Back Panel

Figure 2-2. Cabling a Clustered System With One

PowerVault 22xS Storage System

Figure 2-3. Cabling Two PowerVault 22xS Storage Systems

to a PERC RAID adapter

Figure 2-4. Example of Network Cabling Connection

Figure 2-5. Power Cabling Example With Three

Power Supplies in the Systems

. . . . . . . . . . . . . 14

. . . . . . . . . . . 18

. . . . . . . . . 19

. . . . . . . . . . . . . . 20

. . . . . . 22

. . . . . . . . . . 25

Contents 7

Figure 2-6. Power Cabling Example With One Power Supply

in the Systems

. . . . . . . . . . . . . . . . . . 26

Figure 2-7. Power Cabling Example With Two Power

Supplies in the Systems

. . . . . . . . . . . . . . 27

Tables

Figure 3-1. Back-Panel Module Features and Indicators

Figure 3-2. Bus Configuration Switch Modes

Figure 3-3. Important System Warning

. . . . . . . . . 41

. . . . . . . . . . . . 43

. . . . 40

Table 1-1. Windows Operating System Features . . . . . . 12

Table 1-2. Cluster Storage Requirements

Table 1-3. Cluster Node Requirements

Table 2-1. Cluster Cabling Components

Table 2-2. Network Connections

. . . . . . . . . . 12

. . . . . . . . . . . 15

. . . . . . . . . . . 17

. . . . . . . . . . . . . . 21

Table 2-3. Private Network Hardware Components

and Connections

. . . . . . . . . . . . . . . . . 23

Table 3-1. Applications and Hardware Requiring

IP Address Assignments

Table 3-2. Examples of IP Address Assignments

Table 3-3. Split-bus Module Modes and Functions

Table 5-1. Node States and Definitions

. . . . . . . . . . . . . 34

. . . . . . 35

. . . . . 42

. . . . . . . . . . . 59

8 Contents

Table 5-2. Cluster Resources and Required

Dependencies

. . . . . . . . . . . . . . . . . . 62

Table 5-3. Resources and Configurable Parameters

Table A-1. General Cluster Troubleshooting

. . . . . . . . . 87

. . . . . 63

Introduction

This guide provides information for installing a Dell™ PowerEdge™ Cluster SE500W solution
in a corporate enterprise.

The information in this guide includes:

• Basic SCSI cluster installation procedures, which include:

– Preparing server and storage systems for clustering

– Cabling the cluster configuration

®

– Installing the Microsoft

operating system in your cluster configuration

NOTE: Reference to Windows Server 2003 in this guide implies reference to both Windows

Server 2003 Enterprise and Windows Server 2003 Enterprise x64 Editions, unless explicitly stated.

– Configuring the cluster peripherals, including PERC 4/DC or PERC 4e/DC cards and

network adapters

• Installation procedures for installing a SCSI cluster configuration in your corporate network

• Cluster upgrade and maintenance procedures

• Information about MSCS, the clustering software built into the Windows 2000 Advanced
Server, and Windows Server 2003 operating systems

• Troubleshooting procedures

• Data form for recording critical cluster configuration information

Windows® 2000 Advanced Server or Windows Server™ 2003

Intended Audience

This guide was developed for experienced IT professionals who need to install, cable, and
configure a PowerEdge Cluster SE500W solution in an enterprise environment and for trained
service technicians who perform cluster upgrade and maintenance procedures.

Obtaining More Information

See "Obtaining Technical Assistance" and "Overview" for a general description of PowerEdge
clusters and clustering technology.

See "Using MSCS" for an overview of the clustering software built into the Windows 2000
Advanced Server, Windows Server 2003 Enterprise Edition, and Windows Server 2003 x64
Edition operating systems.

Introduction 9

Obtaining Technical Assistance

Dell Enterprise Training and Certification is available; see www.dell.com/training for more
information. This service may not be offered in all locations.

Overview

Clustering uses specific hardware and software to join multiple systems together to function as a
single system and provide an automatic failover solution. If one of the clustered systems (also
known as cluster nodes, or nodes) fails, resources running on the failed system are moved (or
failed over) to one or more systems in the cluster by the Microsoft
software. MSCS is the failover software component in specific versions of the Windows

www.dell.com | support.dell.com

operating system.

When the failed system is repaired and brought back online, resources automatically transfer
back (or fail back) to the repaired system or remain on the failover system, depending on how
MSCS is configured. See "Failover and Failback" for more information.

NOTE: Reference to Windows Server 2003 in this guide implies reference to both Windows Server 2003

Enterprise and Windows Server 2003 Enterprise x64 Editions, unless explicitly stated.

Virtual Servers and Resource Groups

In a cluster environment, you do not access a physical server; you access a virtual server, which is
managed by MSCS. Each virtual server has its own IP address, name, and hard drive(s) in the
shared storage system. MSCS manages the virtual server as a resource group, which contains the
cluster resources. Ownership of virtual servers and resource groups is transparent to users. See
"Groups" for more information on resource groups.

When MSCS detects a failed node or failed application, MSCS moves the entire resource group
to another node and remaps the virtual server to the new network connection. Users of an
application in the virtual server experience only a momentary delay in accessing resources while
MSCS re-establishes a network connection to the virtual server and restarts the application.

®

Cluster Service (MSCS)

Quorum Resource

A single disk, which is designated as the quorum resource, maintains the configuration data
(including all the changes that have been applied to a cluster database) necessary for recovery
when a node fails.

The quorum resource can be any resource with the following attributes:

• Enables a single node to gain and defend its physical control of the quorum resource

• Provides physical storage that is accessible by any node in the cluster

• Uses the Microsoft Windows NT

10 Introduction

®

file system (NTFS)

See "Quorum Disk (Quorum Resource)" and the MSCS online documentation for more
information.

NOTE: PowerEdge Cluster SE500W solutions do not support the Majority Node Set (MNS) Quorum

resource type.

Shared Storage Systems

Cluster nodes can share access to external storage systems; however, only one of the nodes can
own any RAID volume in the external storage system at any time. MSCS controls which node
has access to each RAID volume in the shared storage system.

Dell OpenManage™ Array Manager or Dell OpenManage enhanced Storage Manager provides
storage management and monitoring for SCSI storage components. See your Array Manager or
OMSM documentation for more information.

PowerEdge Cluster SE500W Solution

The PowerEdge Cluster SE500W solution implements two-node clustering technology based
on the MSCS software incorporated within the Windows 2000 Advanced Server, and Windows
Server 2003 operating systems. This cluster solution provides the following benefits to meet the
needs of mission-critical network application programs:

• High availability of system services and resources to network clients

• Redundant storage for application program data

• Failure recovery for cluster-aware applications

• Flexible maintenance capabilities, allowing you to repair, maintain, or upgrade a cluster node
without taking the entire cluster offline

Each cluster node is configured with software, storage, and network resources that enable it to
monitor and interact with the other node to provide mutually redundant operation. If a cluster
node fails for any reason, virtual servers and resource groups are failed over to the healthy cluster
node. When the failed node is repaired and brought back online, the virtual servers and resource
groups are failed back to the repaired node (if desired).

The cluster nodes, therefore, operate as a single resource, rather than a collection of individual
systems. Because the cluster nodes interact in this way, each virtual server appears as a single
system to the network clients.

Operating Systems

The PowerEdge Cluster SE500W solution supports two-node cluster configurations that vary in
size and performance. Table 1-1 provides an overview of the supported Windows
operating systems.

See your operating system documentation for a complete list of features.

Introduction 11

Table 1-1. Windows Operating System Features

Windows 2000 Advanced Server Windows Server 2003 Enterprise

Supports two-node SCSI clusters Supports two-node SCSI clusters Supports two-node SCSI clusters

Supports up to 8 GB of RAM
per node

Cluster configuration and
management using
Control Panel utilities

www.dell.com | support.dell.com

NOTE: The amount of RAM supported per node also depends on your cluster platform.

NOTE: Running different operating systems in a cluster is supported only during a rolling upgrade. You

cannot upgrade to Windows Server 2003 Enterprise x64 Edition. Only a new installation is permitted for
Windows Server 2003 Enterprise x64 Edition.

NOTE: MSCS and Network Load Balancing (NLB) features cannot coexist on the same node, but can be

used together in a multitiered cluster. For more information, see the Dell PowerEdge Clusters website at
www.dell.com/ha or the Microsoft website at www.microsoft.com.

Storage System

Table 1-2 provides a list of supported storage components and the configuration requirements
for the cluster nodes and stand-alone systems connected to the storage systems.

Windows Server 2003 Enterprise

Edition

Supports up to 32 GB of RAM
per node

Cluster configuration and
management using Configure
Your Server (CYS) and Manage
Your Server (MYS) wizards

Metadirectory Services Metadirectory Services

x64 Edition

Supports up to 1 TB RAM
per node

Cluster configuration and
management using Configure
Your Server (CYS) and Manage
Your Server (MYS) wizards

Table 1-2. Cluster Storage Requirements

Hardware Components Minimum Requirement

Supported storage systems Up to four Dell PowerVault™ 22xS storage systems (for the shared disk resource)

Enclosure management
modules

Power and cooling
requirements

Hard drives At least two SCSI hard drives in each PowerVault 22xS enclosure to support

Cables Two 1-, 4-, 8-, or 12-m SCSI cables for each PowerVault 22xS storage system

12 Introduction

Two enclosure management modules (EMMs) per PowerVault
22xS enclosure

Redundant configuration: two fans and dual power supplies per PowerVault
22xS enclosure

hardware-based RAID functionality

NOTE: The minimum recommendation for an active/passive cluster

configuration is two virtual disks. If you are going to use an active/active
configuration, at least three virtual disks are recommended.

in the cluster

Hardware and Software Technologies

The PowerEdge Cluster SE500W solution implements the following hardware and
software technologies:

• Clustering technology based on the MSCS software in the Windows 2000 Advanced Server,
and Windows Server 2003 operating systems

• PERC 4/DC card or PERC 4e/DC card

• Storage management software

Clustering Technology

Clustering is the process of connecting multiple servers together to achieve higher availability
and performance. MSCS is a software component in Windows 2000, Advanced Server and an
integrated service in Windows Server 2003, which provides failover support for applications and
services running on each node.

See "Using MSCS" for more information on the Cluster Service.

NOTE: MSCS and network load balancing (NLB) features cannot coexist on the same cluster node, but

can be used together in a multi-tiered cluster configuration. For more information, see the PowerEdge
Clusters website located at www.dell.com/ha or the Microsoft website located at www.microsoft.com.

Cluster Components

• Two PowerEdge systems in a homogeneous pair.

NOTE: Dell or Microsoft can support only the specific configurations described in the Dell

PowerEdge Cluster SE500W Platform Guide.

• Up to four PowerVault 22xS storage systems, each with dual EMMs and split-bus module.

• Supported network adapters for a public and private LANs.

• Support for RAID 1, 5, 1+0, and 5+0 levels and hot spare drives. If you are using two
enclosures, Dell recommends configuring RAID 1 or 1+0 across enclosures for
additional redundancy.

NOTE: RAID 0 and independent drives are possible but are not recommended for a high-availability

system because they do not offer data redundancy if a disk failure occurs.

• PERC 4/DC or PERC 4e/DC adapter(s) for the cluster’s shared storage.

NOTE: The PowerEdge Cluster SE500W supports up to two PERC 4/DC or PERC 4e/DC adapters in a

single cluster node. Dell does not support use of PERC 4/DC and PERC 4e/DC adapters together in
the PowerEdge Cluster SE500W solution.

NOTE: PERC RAID adapter refers to a PERC 4/DC card or a PERC 4e/DC card.

• Ethernet network adapter or integrated network adapter for node-to-node
cluster communications.

• Ethernet crossover or straight cable for node-to-node cluster interconnect.

Introduction 13

Figure 1-1 shows a sample configuration of the PowerEdge Cluster SE500W components and
their interconnections. See the Dell PowerEdge Cluster SE500W Platform Guide for
system-specific configuration information.

Figure 1-1. Maximum Configuration of the PowerEdge Cluster SE500W Solution

PowerEdge systems (2)
(two PERC 4/DC or two
PERC 4e/DC controllers
per PowerEdge system)

www.dell.com | support.dell.com

PowerVault 22xS storage

systems (up to 4)

System Requirements

PowerEdge Cluster SE500W configurations require the following hardware and
software components:

• Cluster nodes

• Storage

• Interconnects (private network)

• Client network connections (public network)

• Operating system and storage management software

network cable

node-to-node cluster
(heartbeat) interconnect

SCSI cables (8)

14 Introduction

Cluster Nodes

Table 1-3 lists the hardware requirements for the cluster nodes.

Table 1-3. Cluster Node Requirements

Component Minimum Requirement

Cluster nodes Two supported PowerEdge systems running Windows 2000 Advanced

Server operating system.

or

Two supported PowerEdge systems running Windows Server 2003
operating system.

Processors At least two processors for each cluster node.

RAM At least 256 MB of RAM installed on each cluster node for Windows

2000 Advanced Server and Windows Server 2003 Enterprise Edition.

At least 512 MB RAM installed on each cluster node for Windows
Server 2003 Enterprise x64 Edition.

On-board hard drive
controller

Clustered RAID adapter Up to two PERC 4/DC or two PERC 4e/DC cards for each cluster node.

Network adapters Minimum of two network interfaces: one for the public network (client

For the internal drives, you can use any RAID controller or SCSI
adapter supported by the server platform.

Two disk drives are required for mirroring (RAID 1) and at least three
disk drives are required for disk striping with parity (RAID 5).

NOTE: Dell strongly recommends that you use hardware-based RAID or

software-based disk-fault tolerance for the internal drives. At least 4 GB
disk space is required for Windows Server 2003 Enterprise x64 Edition.

Dell recommends placing your RAID adapters on separate I/O busses to
improve availability and performance.

See the Platform Guide for more information about supported
PowerEdge systems.

LAN connections) and another for the private network (cluster
interconnect).

The network adapters installed in each cluster node must be identical
and supported by the server platform.

Cluster storage

PowerEdge Cluster SE500W configurations support up to four PowerVault 22xS storage systems
per cluster.

Introduction 15

Other Documents You May Need

CAUTION: The Product Information Guide provides important safety and regulatory information.

Warranty information may be included within this document or as a separate document.

• The

• The

• The

www.dell.com | support.dell.com

• The

• The

• The PERC 4/DC or PERC 4/eDC documentation includes information on the

• The Dell OpenManage Array Manager or Dell OMSM documentation provides instructions

• Operating system documentation describes how to install (if necessary), configure, and use

• Microsoft Windows 2000 or Windows Server 2003 Cluster Service documentation.

• The

• Documentation for any components you purchased separately provides information to

• Updates are sometimes included with the system to describe changes to the system, software,

Dell PowerEdge Cluster SE500W Systems Platform Guide

systems that support the PowerEdge Cluster SE500W configuration.

Rack Installation Guide

system into a rack.

Setting Up Your System

your system.

Users Guide

technical specifications, SCSI drivers, the System Setup program (if applicable), software
support, and the system configuration utility.

Installation and Troubleshooting Guide

describes how to troubleshoot the system and install or replace system components.

SCSI RAID controller.

for using the array management software to configure RAID systems.

the operating system software.

System Administrator’s Guide

configure and install these options.

and/or documentation.

provides information about the

included with your rack solution describes how to install your

document provides an overview of initially setting up

for your PowerEdge or PowerVault system describes system features and

for your PowerEdge or PowerVault system

provides system operation and management operation.

NOTE: Always read the updates first because they often supersede information in other documents.

• Release notes or readme files may be included to provide last-minute updates to the system
documentation or advanced technical reference material intended for experienced users
or technicians.

16 Introduction

Cabling Your Cluster Hardware

Dell™ PowerEdge™ Cluster SE500W configurations require cabling for the storage systems,
cluster interconnects, client network connections, and power connections.

Cabling for the Cluster SE500W Solution

The cluster systems and components are interconnected to provide four independent functions
as listed in Table 2-1, each of which is described in more detail throughout this section.

Table 2-1. Cluster Cabling Components

Components Description

Shared storage system Connects the host-based, RAID controller(s) to the disk

enclosure(s). This interconnect is made from the PERC 4/DC or
PERC 4e/DC card in each PowerEdge system to the Dell
PowerVault™ 22xS EMMs, using Dell-qualified U320 SCSI cables.

Cluster interconnect
(private network)

Network connection for public
traffic (public network)

Power connection Provides a connection between the power source and the power

Connects the systems to each other to exchange information and
status. This connection can be made by using a supported
Ethernet network adapter and cabling that is connected to each
cluster node.

See the Platform Guide for a list of supported network adapters for
your configuration.

Provides a connection between each cluster node and the client
network. This connection can be made using an Ethernet network
adapter and cabling that is connected to the public network.

See the Platform Guide for a list of supported network adapters for
your configuration.

supplies in your system. By using power strips or Power
Distribution Units (PDUs) and separate AC circuits, the cluster
can fully utilize the redundant power supplies.

Cabling Your Cluster Hardware 17

Cabling One PowerVault 22xS Shared Storage System to a Cluster SE500W

NOTE: See "Configuring the PowerVault 22xS Storage System for Cluster Mode" for more information

about configuring the storage systems.

NOTICE: Do not turn on the systems or the storage system(s) until the split-bus module on the back of

the PowerVault system has been set to cluster mode and all cabling is complete.

NOTICE: The asymmetrical, D-shaped SCSI connectors can be damaged if they are forced together

when they are not oriented to each other correctly. To avoid connector damage, match the connector
shapes (long side-to-long side and short side-to-short side). Insert the cable connector straight (not at an
angle) into the card connector.

1

Connect the SCSI connector A (see Figure 2-1) on the back of the PowerVault 22xS storage

www.dell.com | support.dell.com

system to the 68-pin connector on the first SCSI cable (see Figure 2-2), and then tighten the
retaining screws.

NOTE: As viewed from the back of the system, the left EMM is the primary EMM and the right EMM

is the secondary EMM.

NOTE: See "Enabling the Cluster Mode Using the PERC RAID Adapter" for more information about

cluster-enabling the card.

2

Connect the very high-density connector interface (VHDCI) connector of the first SCSI
cable to the channel 0 connector on the cluster-enabled PERC RAID adapter in the first
PowerEdge system (see Figure 2-2), and then tighten the retaining screws.

Figure 2-1. PowerVault 22xS Back Panel

redundant cooling fans (2)

power supply

SCSI connector A

primary EMM

split bus module

NOTE: Ensure that you securely tighten the retaining screws on all SCSI connectors to ensure a

reliable connection.

18 Cabling Your Cluster Hardware

redundant power supply

SCSI connector B

secondary EMM

3

Connect the VHDCI connector of the second SCSI cable (see Figure 2-2) to the channel 0
connector on the cluster-enabled PERC RAID adapter in the second PowerEdge system, and
then tighten the retaining screws.

4

Connect the SCSI connector B (see Figure 2-1) on the back of the PowerVault 22xS storage
system to the 68-pin connector on the second SCSI cable (see Figure 2-2), and tighten the
retaining screws.

NOTE: If the PowerVault 22xS storage system is disconnected from the cluster, it must be

reconnected to the same channel on the same PERC RAID adapter for proper operation.

Figure 2-2. Cabling a Clustered System With One PowerVault 22xS Storage System

VHDCI

PERC RAID
adapter

node 1

shared storage
connector A(68-pin connector)

node 2

SCSI connections from
channel 0 on each
cluster-enabled
PERC RAID adapter

PowerVault 22xS

storage system
shared storage connector B
(68-pin connector)

Cabling Your Cluster Hardware 19

Cabling Two PowerVault 22xS Storage Systems to a Cluster SE500W

Connecting the cluster to two PowerVault 22xS storage systems is similar to connecting the
cluster to a single PowerVault 22xS storage system. Connect channel 0 of the PERC RAID
adapter in each node to the back of the first storage system, as described in "Cabling One
PowerVault 22xS Shared Storage System to a Cluster SE500W." Repeat the process for channel 1
on the controller in each node using a second PowerVault 22xS storage system. See Figure 2-3.

With dual storage systems connected to a single PERC RAID adapter, mirroring disk drives
from one storage system to another is supported through RAID 1 and 1+0. To protect the
cluster applications and your data if an entire storage system fails, Dell strongly recommends
using RAID 1 or 1+0.

www.dell.com | support.dell.com

VHDCI (channels 0
and 1)

PERC RAID
adapter
channel 0

NOTE: If you have dual cluster-enabled PERC RAID adapters (total of four channels) and only two shared

storage systems, you may want to connect one storage system to each controller. If you remove the
cable connections, you must reconnect the cables as they were previously connected. To ensure that
the cables are reconnected correctly, Dell recommends that you tag or color-code the cables.

Figure 2-3. Cabling Two PowerVault 22xS Storage Systems to a PERC RAID adapter

PERC RAID adapter
channel 1

PERC RAID adapter

SCSI connections from
channel 1 on each
PERC RAID adapter

68-pin connectors (2)

20 Cabling Your Cluster Hardware

68-pin connectors (2)

Cabling Three or Four PowerVault 22xS Storage Systems to a Cluster SE500W

To connect the cluster to three or four PowerVault 22xS storage systems, repeat the process
described in the preceding section for a second controller.

NOTICE: If you have dual storage systems that are attached to a second controller, Dell supports disk

mirroring between channels on the second controller. However, Dell does not support mirroring disks on
one cluster-enabled PERC RAID adapter to disks on another cluster-enabled PERC RAID adapter.

Cabling Your Public and Private Networks

The network adapters in the cluster nodes provide at least two network connections for each
node. These connections are described in Table 2-2.

Table 2-2. Network Connections

Network Connection Description

Public network All connections to the client LAN.

At least one public network must be configured for Mixed mode for
private network failover.

Private network A dedicated connection for sharing cluster health and status

information between the cluster nodes.

Network adapters connected to the LAN can also provide redundancy at
the communications level in case the cluster interconnect fails. See your
MSCS documentation for more information on private network
redundancy.

Cabling Your Cluster Hardware 21

Figure 2-4 shows an example of network adapter cabling in which dedicated network adapters in
each node are connected to the public network and the remaining network adapters are
connected to each other (for the private network).

Figure 2-4. Example of Network Cabling Connection

www.dell.com | support.dell.com

public network

public network
adapter

cluster node 1

private network adapter

private network

cluster node 2

Cabling Your Public Network

The public network connection (client network) to the cluster nodes is provided by a network
adapter that is installed in each node. Any network adapter supported by the system running
TCP/IP may be used to connect to the public network segments. Additional network adapters
may be installed to support additional separate public network segments or to provide
redundancy for the public network.

Installing redundant network adapters provides your cluster with a failover connection to the
public network. If the primary network adapter or a switch port fails, your cluster will be able to
access the public network through the secondary network adapter until the faulty network
adapter or switch port is repaired.

Using Dual-Port Network Adapters for Your Private Network

You can configure your cluster to use the public network as a failover for private network
communications. However, if dual-port network adapters are used, do not use two ports
simultaneously to support both the public and private networks.

22 Cabling Your Cluster Hardware

Cabling Your Private Network

The private network connection to the cluster nodes is provided by a second or subsequent
network adapter that is installed in each node. This network is used for intracluster
communications. Table 2-3 lists the required hardware components and connection method for
three possible private network configurations.

Table 2-3. Private Network Hardware Components and Connections

Method Hardware Components Connection

Network switch Fast Ethernet or Gigabit

Ethernet network adapters and
switches

Point-to-Point
Fast Ethernet

Point-to-Point
Gigabit Ethernet

NOTE: On certain Microsoft® Windows® 2000 Advanced Server, Windows Server™ 2003

configurations, using an Ethernet cable in a point-to-point connection can impact node-to-node
communications. See Microsoft Knowledge Base articles 239924, 242430, 254651, and 258750 at
www.microsoft.com for more information. This issue has been corrected in Windows Server 2003.

Fast Ethernet network adapters Connect a crossover Ethernet cable between

Copper Gigabit Ethernet
network adapters

Connect standard Ethernet cables from the
network adapters in both cluster nodes to a
Fast Ethernet or Gigabit Ethernet switch.

the Fast Ethernet network adapters in both
cluster nodes.

Connect a standard Ethernet cable between
the Gigabit Ethernet network adapters in both
cluster nodes.

NIC Teaming

Network Interface Card (NIC) teaming combines two or more NICs to provide load balancing
and/or fault tolerance. Your cluster supports NIC teaming, but only in a public network;
NIC teaming is not supported in a private network.

You should use the same brand of NICs in a team, and you cannot mix brands of
teaming drivers.

Cabling Your Cluster Hardware 23

Cabling the Mouse, Keyboard, and Monitor

If you are installing a PowerEdge Cluster SE500W configuration in a Dell rack, your cluster will
require a switch box to enable the mouse, keyboard, and monitor for your cluster nodes.

See your rack installation documentation included with your rack for instructions on cabling
each cluster node’s Keyboard Video Mouse (KVM) to the mouse/keyboard/monitor switch box
in the rack.

Power Cabling for the PowerEdge Cluster SE500W Solution

Observe the following cautions when connecting the power cables to the

www.dell.com | support.dell.com

PowerEdge Cluster SE500W.

CAUTION: Although each component of the PowerEdge Cluster SE500W meets leakage current safety

requirements, the total leakage current may exceed the maximum that is permitted when the
components are used together. To meet safety requirements in the Americas (that is, the United States,
Canada, and Latin America), you must use a Type B plug and socket connection for the cluster power to
enable the appropriate level of ground protection. In Europe, you must use one or two PDUs or two Type
B plug-and- socket connections wired and installed by a qualified electrician in accordance with the
local wiring regulations.

CAUTION: Do not attempt to cable the PowerEdge Cluster SE500W to electrical power without first

planning the distribution of the cluster’s electrical load across available circuits. For operation in the
Americas, the PowerEdge Cluster SE500W requires two AC circuits with a minimum capacity of
20 amperes (A) each to handle the electrical load of the system. Do not allow the electrical load of the
system to exceed 16 A on either circuit.

CAUTION: For operation in Europe, the PowerEdge Cluster SE500W requires two circuits rated in

excess of the combined load of the attached systems. Refer to the ratings marked on the back of each
cluster component when determining the total system’s electrical load.

See your system and storage system documentation for more information about the specific
power requirements for your cluster system’s components.

Dell recommends the following guidelines to protect your cluster system from power-related failures:

• For cluster nodes with multiple power supplies, plug each power supply into a separate
AC circuit.

• Use uninterruptible power supplies (UPS).

For some environments, you may consider having backup generators and power from separate
electrical substations.

Figure 2-5, Figure 2-6, and Figure 2-7 illustrate the proper power cabling of Cluster SE500W
components consisting of two PowerEdge systems and two PowerVault 22xS storage systems.
Each component of the cluster must have power supplied by two or three separate AC circuits—
one circuit to each component power supply. Therefore, the primary power supplies of all the
PowerEdge Cluster SE500W components are grouped onto one or two circuits and the
redundant power supplies are grouped onto a different circuit.

24 Cabling Your Cluster Hardware

Figure 2-5. Power Cabling Example With Three Power Supplies in the Systems

redundant power supplies on one
primary power supplies
on two AC power strips (or
on two AC PDUs [not shown])

AC power strip (or on one AC PDU

[not shown])

CAUTION: The arrangement of the cluster components in this illustration is

intended only to demonstrate the power distribution of the components. Do not stack
components as in the configuration shown.

Cabling Your Cluster Hardware 25

Figure 2-6. Power Cabling Example With One Power Supply in the Systems

www.dell.com | support.dell.com

CAUTION: The arrangement of the cluster components in this illustration is intended

only to demonstrate the power distribution of the components. Do not stack
components as in the configuration shown.

26 Cabling Your Cluster Hardware

primary power supplies on
one AC power strip (or on
one AC PDU [not shown])

redundant power supplies on one
AC power strip (or on one AC PDU
[not shown])

Figure 2-7. Power Cabling Example With Two Power Supplies in the Systems

redundant power supplies on one
primary power supplies on
one AC power strip (or on
one AC PDU [not shown])

AC power strip (or on one AC PDU

[not shown])

CAUTION: The arrangement of the cluster components in this illustration is intended

only to demonstrate the power distribution of the components. Do not stack components
as in the configuration shown.

Cabling Your Cluster Hardware 27

www.dell.com | support.dell.com

28 Cabling Your Cluster Hardware

Preparing Your Systems for Clustering

This section provides information for preparing your server and storage system hardware for
clustering in a Dell™ PowerEdge™ Cluster SE500W configuration.

Before You Begin

1

Ensure that your site can handle the power requirements of the cluster equipment.

Contact your sales representative for information about your region's power requirements.

CAUTION: Only trained service technicians are authorized to remove and access any of the

components inside the system. See your Product Information Guide for complete information about
safety precautions, working inside the computer, and protecting against electrostatic discharge.

2

Ensure that the following components are installed in each PowerEdge system in the cluster:

• Network adapters

• PERC RAID adapters

• SCSI hard drives

• Any additional peripheral components

• RAID controllers for internal drives (optional)

3

Ensure that the following components are installed in each Dell PowerVault™ 22xS system in
the cluster. See "Installing and Configuring the Shared Storage System."

• Two enclosure management modules (EMMs)

• A split-bus module

• SCSI hard drives

4

Cable the system hardware for clustering.

See "Cabling Your Cluster Hardware" for more information.

5

Configure the storage system(s) as described in your storage system documentation.

6

Configure the PERC RAID adapters as described in your PERC RAID adapter
documentation.

7

If you are using hardware-based RAID for the internal SCSI hard drives, configure the hard
drives using the controller's BIOS utility or Dell OpenManage™ Array Manager or
Dell OMSM.

Preparing Your Systems for Clustering 29

Installation Overview

This section provides installation overview procedures for configuring your cluster running the
Microsoft

1

2

3

www.dell.com | support.dell.com

4

5

®

Windows® 2000 Advanced Server, or Windows Server™ 2003 operating system.

Ensure that your cluster meets the requirements as described in "Before You Begin."

Select a domain model that is appropriate for your corporate network and operating system.

See "Selecting a Domain Model" for more information.

Reserve static IP addresses for your cluster resources and components.

The resources and components include:

• Public network

• Private network

• Cluster virtual servers

You will use these IP addresses when you install the Windows operating system and MSCS.

See "Assigning Static IP Addresses to Your Cluster Resources and Components" for
more information.

Configure the internal hard drives in your cluster nodes.

See "Configuring the Internal Drives in Your Cluster Nodes" for more information.

Install and configure the Windows operating system.

The Windows operating system must be installed on both of the cluster nodes. Each cluster
node must have its own licensed copy of the Windows operating system and Certificate of
Authenticity attached.

See "Installing and Configuring the Windows Operating System" for more information.

6

Install or update the PERC RAID adapter drivers.

The PERC RAID adapter drivers allow your cluster nodes to communicate with the shared
storage systems.

See "Updating the Windows Driver for the PERC RAID Adapter" for more information.

7

Install and configure the storage management software.

See the documentation included with your Array Manager software or available at the
Dell Support website (located at

8

Configure the hard drives on the shared storage system(s).

See "Configuring and Managing Virtual Disks" for more information.

30 Preparing Your Systems for Clustering

support.dell.com

) for more information.

9

Configure the MSCS software.

The MSCS software is the clustering component of the Windows operating system that
provides the failover capabilities for the cluster.

See "Installing and Configuring MSCS" for more information.

10

Verify cluster functionality. Ensure that:

• Your cluster components are communicating properly with each other.

• MSCS is started.

See "Verifying Cluster Functionality" for more information.

11

Verify cluster resource availability.

Use Cluster Administrator to check the running state of each resource group.

See "Verifying Cluster Resource Availability" for more information.

The following sections provide detailed information for each step in the "Installation Overview"
that is specific to your Windows operating system.

NOTE: Dell strongly recommends that you use the "PowerEdge Cluster SE500W Solution Data Form"

during the installation of your cluster to ensure that all installation steps are completed. The data form is
located in "Cluster Data Form."

Selecting a Domain Model

On a cluster running the Windows operating system, both cluster nodes must belong to a
common domain or directory model. The following membership configurations are supported:

• Both cluster nodes are member servers in an Active Directory domain.

• Both cluster nodes are domain controllers in an Active Directory domain.

• One cluster node is a domain controller in an Active Directory, and the other cluster node is a
member server.

Configuring the Cluster Nodes as Domain Controllers

If a cluster node is not configured as a domain controller and the node cannot contact a domain
controller, the cluster node will not be able to authenticate client requests.

If a cluster node is configured as a domain controller, client access to its cluster resources can
continue if the cluster node cannot contact other domain controller(s). However, domain
controller functions can cause additional overhead, such as log on, authentication, and
replication traffic on the cluster nodes.

Preparing Your Systems for Clustering 31

Configuring the Internal Drives in Your Cluster Nodes

If you have added new hard drives to your system or are setting up the internal drives in a
RAID configuration, you must configure the RAID (if applicable) using the RAID controller’s
BIOS configuration utility or Dell OpenManage Server Assistant before you can install the
operating system. For the best balance of fault tolerance and performance, Dell recommends
using RAID 1 for the internal hard drives.

RAID 1 is commonly called mirroring because it writes identical data to two separate drives.
RAID 1 is the best choice in high-availability applications that require high performance or
relatively low data capacity.

See the documentation for your specific RAID controller for more information on

www.dell.com | support.dell.com

RAID configurations.

NOTE: If you are not going to use a hardware-based RAID controller, Dell recommends using the

Windows Disk Management tool or Dell OpenManage Array Manager or Dell OMSM to provide
software-based redundancy for the Windows system partitions.

Installing and Configuring the Windows Operating System

Before you install the Windows operating system on your cluster nodes, ensure that your cluster
meets the requirements listed in "Before You Begin." After you complete these preliminary
procedures, install and configure the operating system on your cluster nodes.

The following is an overview sequence for installing the Windows operating system and the
cluster management software on the cluster solution.

NOTICE: Windows standby mode and hibernation mode are not supported in cluster configurations. Do

not enable either mode.

1

Ensure that your cluster configuration meets the requirements as listed in "Before You Begin."

2

Cable your cluster hardware.

NOTE: Do not connect your cluster nodes to the shared storage systems yet.

See "Cabling Your Cluster Hardware" for instructions about how to cable your system
hardware for cluster configurations, and information about cluster cables, connectors,
and components.

3

Install and configure the Windows 2000 Advanced Server or Windows Server 2003 operating
system with the latest service pack on each node.

See the

4

If you are installing Windows Server 2003, go to step 5.

If you are installing Windows 2000 Advanced Server, select the option to install the Cluster
Service files when prompted. You will configure the Cluster Service later.

32 Preparing Your Systems for Clustering

Platform Guide

for more information about the latest supported service pack.

5

Ensure that the network adapter drivers installed on each cluster node are the latest
supported version.

6

Configure the public and private network adapter interconnects in each node, and place the
interconnects on separate IP subnetworks using static IP addresses. See "Configuring
Windows Networking."

Public

refers to the network adapter that connects the cluster nodes to the client network.

Private

refers to the dedicated network (cluster interconnect) that connects the cluster nodes

to each other.

7

Install or update the driver for the PERC RAID adapter in each node and ensure that each
PERC card has a different SCSI ID. See "Preparing the PERC RAID Adapter for Clustering"
and "Installing the PERC RAID Adapters."

8

Shut down both nodes and connect each node to shared storage. See "Cabling Your Cluster
Hardware."

9

Turn on one node and configure shared storage using Dell Storage Management or the
PERC RAID adapter BIOS utility. See "Installing and Configuring the Shared Storage System."

10

Reboot node 1.

11

From node 1, write the disk signature and then partition, format, and assign drive letters and
volume labels to the hard drives in the storage system using the

Management

application or Array Manager. See "Naming and Formatting Drives on the

Windows Disk

Shared Storage System."

12

On node 1, verify disk access and functionality on all shared disks.

13

Shut down node 1.

14

Verify disk access by performing the following steps on the other node:

a

Turn on the node.

b

Use the PERC RAID adapter BIOS utility to resolve the Non-Volatile Random-Access
Memory (NVRAM) mismatch. See your PERC RAID adapter documentation.

c

Modify the drive letters to match the drive letters on node 1. See "Cluster Data Form."

This procedure allows the Windows operating system to mount the volumes.

d

Close and reopen

e

Verify that Windows can see the file systems and the volume labels correctly.

15

Turn on node 1.

16

Install and configure the Cluster Service. See "Installing and Configuring Microsoft Windows

Disk Management

.

2000 Cluster Service (MSCS)" and "Configuring Microsoft Windows Server 2003 Cluster
Service (MSCS)."

17

Install and set up your application programs (optional).

18

Record the cluster configuration using the "Cluster Data Form" (optional).

Preparing Your Systems for Clustering 33

Configuring Windows Networking

You must configure the public and private networks in each node before you install the Cluster
Service. The following sections introduce you to some principles and procedures necessary to
the networking prerequisites.

Assigning Static IP Addresses to Your Cluster Resources and Components

A static IP address is an Internet address that a network administrator assigns exclusively to a
system or a resource. The address assignment remains in effect until it is changed by the
network administrator.

The IP address assignments for the public LAN segments will depend on the configuration of

www.dell.com | support.dell.com

your environment. If the IP assignments are set up correctly, all of the network adapter resources
will respond to ping commands and appear online before and after you install MSCS. If the
IP assignments are not set up correctly, the cluster nodes may not be able to communicate with
the domain. See "Troubleshooting" for more information.

PowerEdge Cluster configurations running the Windows operating system require static
IP addresses assigned to hardware and software applications in your cluster, as listed in Table 3-1.

Table 3-1. Applications and Hardware Requiring IP Address Assignments

Application/Hardware Description

Cluster IP address The cluster IP address is used for cluster management and must

Cluster-aware applications
running on the cluster

correspond to the cluster name. Because each server has at least two
network adapters, the minimum number of static IP addresses required
for a cluster configuration is five (one for each network adapter and one
for the cluster). Additional static IP addresses are required when MSCS
is configured with application programs that require IP addresses, such
as file sharing.

These applications include Microsoft SQL Server, Enterprise Edition
Microsoft Exchange Server, and Internet Information Server (IIS). For
example, Microsoft SQL Server, Enterprise Edition requires at least one
static IP address for the virtual server. (Microsoft SQL Server does not
use the cluster's IP address.) Also, each IIS Virtual Root or IIS Server
instance configured for failover needs a unique static IP address.

34 Preparing Your Systems for Clustering

Table 3-1. Applications and Hardware Requiring IP Address Assignments (continued)

Application/Hardware Description

Cluster node network adapters The network adapters are used to connect to the public and

private networks.

For cluster operation, two network adapters are required: one network
adapter for the public network (LAN/WAN) and another network
adapter for the private network (sharing heartbeat information between
the cluster nodes).

See "Cabling Your Cluster Hardware" for more information about cluster
interconnect options.

NOTE: To ensure cluster operations during a DHCP server failure,

Dell recommends using static IP addresses for your cluster.

NOTE: On certain Windows 2000 Advanced Server configurations, using an

Ethernet cable in a point-to-point connection can impact node-to-node
communications. See Microsoft Knowledge Base articles 239924, 242430,
254651, and 258750 at www.microsoft.com for more information. This issue
has been corrected in Windows Server 2003.

Configuring IP Addresses for the Private Network

Dell recommends using the static IP address assignments for the network adapters used for the
private network (cluster interconnect). The IP addresses in Table 3-2 are used as examples only.

Table 3-2. Examples of IP Address Assignments

Usage Cluster Node 1 Cluster Node 2

Public network static IP address (for client and
domain controller communications)

Public network subnet mask 255.255.255.0 255.255.255.0

Private network static IP address cluster
interconnect (for node-to-node communications)

Private network subnet mask 255.255.255.0 255.255.255.0

Default gateway 192.168.1.1 192.168.1.1

Windows Internet Naming Service (WINS)
servers

Domain Naming Server (DNS) servers Primary

192.168.1.101 192.168.1.102

10.0.0.1 10.0.0.2

Primary

192.168.1.11

Secondary

192.168.1.12

192.168.1.21

Secondary

192.168.1.22

Preparing Your Systems for Clustering 35

Primary

192.168.1.11

Secondary

192.168.1.12

Primary

192.168.1.21

Secondary

192.168.1.22

NOTE: Dell recommends that you do not configure Default Gateway, NetBIOS, WINS, and DNS on your

private network. If you are running Windows 2000 Advanced Server or Windows Server 2003 disable
NetBIOS on your private network.

If multiple cluster interconnect network adapters are connected to a network switch, ensure that
all of the private network’s network adapters have a unique address. You can continue the IP
address scheme in Table 3-2 with 10.0.0.3, 10.0.0.4, and so on for the private network’s network
adapters of the other clusters connected to the same switch.

Additional fault tolerance for the LAN segments can be achieved by using network adapters that
support adapter teaming or by having multiple LAN segments. To avoid communication
problems in the private network, Dell recommends that you
adapters for the cluster interconnect.

www.dell.com | support.dell.com

Creating Separate Subnets for the Public and Private Networks

The public and private network’s network adapters installed in the same cluster node must
reside on separate IP subnetworks. Therefore, the private network used to exchange heartbeat
information between the cluster nodes must have a separate IP subnet or a different network ID
than the public network, which is used for client connections.

Setting the Network Interface Binding Order for Clusters Running Windows 2000

1

On the Windows 2000 desktop, right-click

Network and Dial-up Connections

The
network interfaces.

2

Click the

The

Advanced Settings

Advanced

window appears, displaying all available

menu and then click

window appears.

My Network Places

Advanced Settings

use dual-port network

, and then click

Properties

.

.

3

In the

Adapters and Bindings

connections are designated for
of the list.

To change the connection order:

a

Click

Public

or

Private

b

Click the up-arrow or down-arrow to move the connection type to the top or bottom of

Connections

the

c

Click OK.

d

Close the

Setting the Network Interface Binding Order for Clusters Running Windows Server 2003

1

Click the

2

Click the

The

Start

Advanced

Advanced Settings

box.

Network and Dial-up Connections

button, select

menu, and then click

36 Preparing Your Systems for Clustering

tab in the

Client access only

.

Control Panel

window appears.

Connections

box, ensure that the

or

All communications

window.

, and double-click

Advanced Settings

Public

are at the top

Network Connections

.

.

3

In the

Adapters and Bindings

tab, ensure that the

Private

and

Public

connections are at the

top of the list.

To change the connection order:

a

Click

Public

or

Private

.

b

Click the up-arrow or down-arrow to move the connection to the top or bottom of the

Connections

c

Click OK.

d

Close the

Using Dual-Port Network Adapters for the Private Network

box.

Network Connections

window.

You can configure your cluster to use the public network as a failover for private network
communications. However, dual-port network adapters are not supported in the
private network.

Verifying Cluster Network Communications

To ensure proper cluster operations, the cluster nodes must be able to communicate with each
other through the private network (cluster interconnect). This communication involves the
exchange of heartbeat messages, whereby the two cluster nodes inquire about each other's
status, or "health," and acknowledge each inquiry.

To verify network communications between the cluster nodes:

1

Open a command prompt on each cluster node.

2

At the prompt, type:

ipconfig /all

3

Press <Enter>.

All known IP addresses for each local server appear on the screen.

4

Issue the

Ensure that each local server responds to the

ping

command from each remote system.

ping

command.

Configuring the Internet Connection Firewall

The Microsoft Windows Server 2003 Enterprise x64 Edition and Windows Server 2003
Enterprise Edition operating systems with Service Pack 1 include an enhanced Internet
Connection Firewall that can be configured to block incoming network traffic to a PowerEdge
system. To prevent the Internet Connection Firewall from disrupting cluster communications,
additional configuration settings are required for PowerEdge systems that are configured as
cluster nodes in an MSCS cluster.

Preparing Your Systems for Clustering 37

Certain network communications are necessary for cluster operations, applications and services
hosted by the cluster, and clients accessing those services. If the Internet Connection Firewall is
enabled on the cluster nodes, install and run the Security Configuration Wizard, and then
configure access for the cluster service and for any applications or services hosted by the cluster
and the operating system.

See the following Microsoft Base articles located at the Microsoft Support website at
support.microsoft.com for more information:

• KB883398 — Internet Connection Firewall

• KB832017 — Network ports used by the Windows Server 2003 operating system

www.dell.com | support.dell.com

Installing the PERC RAID Adapters

For systems with dual PERC RAID adapters, Dell recommends installing the cards on separate
Peripheral Component Interconnect (PCI) buses. Placing the cards on separate buses improves
availability and performance.

See the Platform Guide for more information about your system's PCI bus configuration.

Updating the Windows Driver for the PERC RAID Adapter

Windows automatically installs a compatible driver for the PERC RAID adapters. The following
procedure describes how to update the Windows driver to the PERC family driver.

NOTICE: To ensure proper functionality of the PERC RAID adapter for cluster or internal drive use, this

driver must be updated. Perform the procedure at the completion of the Windows installation, when the

system boots for the first time, and prior to configuring the shared drives.

If a PERC RAID adapter driver CD was not shipped with your system, go to the Dell Support
website at support.dell.com to download the latest Windows driver for the PERC RAID adapter.

To update the default driver to a PERC RAID adapter driver, perform the following steps:

1

Click the
click

Computer Management

2

Select

System Tools

RAID controllers

3

Right-click the PERC RAID adapter, select

Update Driver

4

Click

Next

5

Select

Display a list of known drivers for this device…

6

Click

Have Disk

Dell's updated driver, specify the location of the driver (A:> or D:>), and then click

7

Select the appropriate controller (PERC RAID adapter), and then click

8

Click

Next

Start

button, select

, select

. One or more PERC RAID adapters are listed.

to start the Windows Device Driver wizard.

to proceed to the

, insert the diskette or the

to begin the installation.

Programs

.

Device Manager

Install Hardware Device Drivers

, select

Administrative Tools

, and click the plus sign (+) to expand

Properties

Dell OpenManage Server Assistant

, select the

and then click

Driver

dialog box.

, and

tab, and then click

Next

.

CD that contains

OK

Next

.

SCSI and

.

38 Preparing Your Systems for Clustering

9

When the installation is complete, click

10

Click

Close

to exit the

11

Click

Yes

to restart the system.

12

Repeat this procedure for cluster node 2.

Properties

window.

Finish

to exit the wizard.

Installing and Configuring the Shared Storage System

Clustering PowerVault Storage Systems

If you are upgrading an existing PowerVault 22xS storage system to meet the cluster
requirements for the shared storage system, you may need to install additional hard drives
and/or EMMs in the shared storage system. The size and number of drives you add depends on
the RAID level you want to use, the number of hard drives installed in your system, and the
number of application programs you want to run in your cluster environment. For information
about installing hard drives in the PowerVault 22xS storage system, see the Dell PowerVault 220S
and 221S System Installation and Troubleshooting Guide.

NOTE: In cluster mode, the last slot (SCSI ID 15) in the PowerVault 22xS is not used; SCSI ID 15 is used

for the primary EMM.

Configuring the PowerVault 22xS Storage System for Cluster Mode

To ensure that both servers recognize all the drives in the storage system, you must set the
split-bus configuration switch to cluster mode on the PowerVault 22xS storage system before
powering up the storage system.

To configure the PowerVault 22xS storage system in cluster mode, perform the following steps:

1

Set the bus configuration switch on the split-bus module to cluster mode (down position).
See Figure 3-1.

For more information about the split-bus module, see "Split-Bus Module."

2

Install the split-bus module in the PowerVault 22xS storage system.

3

Install the two EMMs in the PowerVault 22xS storage system.

See "Enclosure Management Module (EMM)" for basic information about EMMs; see the

Dell PowerVault 220S and 221S Installation and Troubleshooting Guide

installing EMMs.

Split-Bus Module

Your system supports three SCSI bus modes controlled by the split-bus module:

• Joined-bus mode

• Split-bus mode

• Cluster mode

Preparing Your Systems for Clustering 39

for information about

These modes are controlled by the position of the bus configuration switch when the system is
turned on. Figure 3-2 illustrates the switch position for each mode.

Figure 3-1. Back-Panel Module Features and Indicators

bus configuration switch

power LED indicator

power LED indicator

www.dell.com | support.dell.com

fault LED indicator
active LED indicator

The only difference between cluster mode and joined-bus mode is the SCSI ID occupied by the
enclosure services processor. When cluster mode is detected, the processor SCSI ID changes
from 6 to 15, allowing a second initiator to occupy SCSI ID 6. As a result, SCSI ID 15 is
disabled, leaving 13 available hard drives in cluster mode. For more information about SCSI ID
assignments and cluster mode operation, see your Dell PowerVault 220S and 221S Systems
Installation and Troubleshooting Guide. See Table 3-3 for a description of the split-bus module
modes and functions.

primary
EMM

split-bus
module

power LED indicator

fault LED indicator

active LED indicator

secondary
EMM

NOTE: To change the SCSI bus mode, you must change the position of the bus configuration switch

before turning on the storage system. Using the bus configuration switch while the system is on does not

affect system operation. If you change the bus configuration switch while the system is running, the
change will not take effect until you perform the following sequence: shut down the nodes, reboot the
storage system, and then turn on the nodes.

40 Preparing Your Systems for Clustering

Figure 3-2. Bus Configuration Switch Modes

joined-bus mode

split-bus mode

cluster mode

Preparing Your Systems for Clustering 41

Table 3-3. Split-bus Module Modes and Functions

Mode Position of Bus

Joined-bus mode Up Low Voltage Differential (LVD) termination on the

www.dell.com | support.dell.com

Split-bus mode Center LVD termination on the split-bus module is enabled

Cluster mode Down LVD termination is disabled and the buses are

The split-bus module has only one LED indicator (see Figure 3-1 for location), which is
illuminated when the module is receiving power.

Enclosure Management Module (EMM)

The EMM serves two primary functions in your storage system:

• SCSI bus expansion — Acts as a buffer for the SCSI bus, electrically dividing the bus into two
independent segments while logically allowing all SCSI bus traffic to pass through it
transparently. The buffer improves the quality of the SCSI signals and allows longer cable
length connections.

• Management functions — Includes SCSI enclosure services (SES) and SCSI accessed
fault-tolerant enclosures (SAF-TE) reporting to the host initiator, control of all enclosure
LED indicators, and monitoring of all enclosure environmental elements such as temperature
sensors, cooling modules, and power supplies.

A system with redundant enclosure management features two EMMs that are designated as
primary and secondary and can be configured in either a cluster, joined-bus, or split-bus mode.
A nonredundant configuration consists of one EMM and one SCSI terminator card, and can be
configured in a joined-bus mode only. In a redundant system, only one EMM per SCSI bus is
active at one time, so only one EMM per SCSI bus can respond to SCSI commands from
an initiator.

Function
Configuration
Switch

split-bus module is disabled, electrically joining the

two SCSI buses to form one contiguous bus. In this

mode, neither the split-bus nor the cluster Light-

Emitting Diode (LED) indicators on the front of the

enclosure are illuminated.

and the two buses are electrically isolated, resulting in

two seven-drive SCSI buses. The split-bus LED

indicator on the front of the enclosure is illuminated

while the system is in split-bus mode.

electrically joined. The cluster LED on the front of the

enclosure is illuminated while the system is in

cluster mode.

42 Preparing Your Systems for Clustering

In joined-bus and cluster modes, if a secondary EMM receives a message that the primary EMM
has failed, the fault LED indicator on the primary EMM is illuminated and the condition is
reported back to the host initiator. The secondary EMM then becomes active and holds the
failed primary in a state of reset until it is replaced. If the primary EMM detects that the
secondary has failed, the secondary’s fault LED indicator is illuminated and the failed status is
reported back to the host initiator.

NOTE: In split-bus mode, each EMM controls half of the enclosure. If one EMM fails in split-bus mode,

the second EMM reports the failure but does not assume control of the entire SCSI bus.

The primary EMM is always plugged into the slot on the left (viewed from the back of the
system). In a redundant joined-bus configuration, the primary EMM assumes control of all the
enclosure functionality. In addition, the primary EMM is the only module that reports the status
of the system to the host initiator through SES and SAF-TE protocols. Because the secondary
EMM must assume the responsibilities of the primary in the event that the primary fails, both
the primary and secondary EMMs are continuously monitoring the status of the
system’s components.

Preparing the PERC RAID Adapter for Clustering

The warning message shown in Figure 3-3 appears on your screen when you attempt to modify
the configuration of the shared storage system on your cluster by using the PERC RAID adapter
BIOS configuration utility.

Figure 3-3. Important System Warning

!!!STOP!!!
This operation may change the configuration of disks and

can cause loss of data! Ensure:

1.Peer server is powered up for its controller NVRAM to
be updated. Otherwise, disk configuration should be read
from disk and saved to controller's NVRAM.

2.The second server must not be configuring the disks.

3.There is no I/O activity against shared drives.

Press Any Key To Continue

The warning message appears on the screen immediately after activating the PERC RAID
adapter BIOS configuration utility by pressing <Ctrl><m> during the system’s Power-On
Self-Test (POST).

Preparing Your Systems for Clustering 43

This warning message alerts you to the possibility of data loss if certain precautions are not
taken to protect the integrity of the data on your cluster.

NOTICE: To prevent data loss, your cluster must meet the conditions in the following bulleted list before

you attempt any data-destructive operation on your shared hard drives.

• Ensure that the peer system is turned on during the operation so that the PERC RAID
adapter’s NVRAM can be updated with the new configuration information. Alternately, if the
peer system is down, you must save the disk configuration to the shared storage system.
When you restart the system later, update the peer system’s NVRAM from the disk
configuration saved to the shared storage system.

• Ensure that the peer cluster node is not currently configuring the shared storage system.

www.dell.com | support.dell.com

• Ensure that I/O activity does not occur on the shared storage system during the operation.

• Ensure that your PERC RAID adapter firmware is the latest version. See your PERC RAID
adapter documentation for information on downloading the latest firmware.

Enabling the Cluster Mode Using the PERC RAID Adapter

Each PERC RAID adapter that is used to connect to a shared storage enclosure must have
cluster mode enabled using the PERC RAID adapter’s BIOS configuration utility. Enabling
cluster mode implements the additional functionality required for the controller to operate in a
cluster environment.

NOTICE: If you replace your PERC RAID adapter, ensure that you enable the cluster mode on the

replacement PERC RAID adapter and set the SCSI ID to the appropriate value (6 or 7) before you connect
the SCSI cables to the shared storage. See "Maintaining Your Cluster" for more information.

For more information about enabling cluster mode, see the PERC RAID adapter
documentation, which includes information on the SCSI host adapter.

Setting the SCSI Host Adapter IDs

After cluster mode on the PERC RAID adapter is enabled, the option to change the SCSI ID for
both of the adapter’s channels will appear. For each shared SCSI bus (a connection from a
channel on one system's PERC RAID adapter to the shared storage enclosure to a channel on
the second system's PERC RAID adapter), the SCSI IDs for each controller must be different.
The default SCSI ID for the PERC RAID adapter is ID 7; thus, the SCSI ID for one of the
system's PERC RAID adapters must be changed to ID 6.

For cluster configurations with two PERC RAID adapters in each node connected to shared
storage enclosures, set both controllers in one system to SCSI ID 6; that is, one node's pair of
PERC RAID adapters utilizes SCSI ID 7 (default) and the other node's pair of PERC RAID
adapters is changed to utilize SCSI ID 6.

44 Preparing Your Systems for Clustering

See the PERC RAID adapter documentation for more information about setting the SCSI host
adapter ID number.

NOTICE: If you replace a PERC RAID adapter, you must set the appropriate SCSI ID before you connect

the SCSI cables to the shared storage. See "Maintaining Your Cluster" for more information.

Configuring and Managing Virtual Disks

The hard drives in the shared storage system must be set up for clustering. The first step is to
configure the RAID levels that you will be using in your cluster. For instructions about setting
up a RAID array, see the PERC RAID adapter documentation and the Dell OpenManage Array
Manager or Dell OMSM documentation.

All virtual disks, especially if they are used for the quorum resource, should incorporate the
appropriate RAID level to ensure high availability. See "Creating the Quorum Resource" for
more information on the quorum resource.

NOTE: Dell recommends that you use a RAID level other than RAID 0 (which is commonly called

striping). RAID 0 configurations provide very high performance, but do not provide the level of availability

required for the quorum resource. See the documentation for your storage system for more information
about setting up RAID levels for the system.

In a cluster configuration, if multiple NTFS partitions are created on a single virtual disk, these
partitions will fail over together. If you plan to run cluster-aware applications on each cluster
node, you must create at least two separate virtual disks to ensure that the applications can fail
over independently.

"Naming and Formatting Drives on the Shared Storage System" describes how to assign drive
letters to the shared hard drives in a cluster installation. For instructions on partitioning and
formatting the shared storage system’s hard drives, see the appropriate operating system
documentation and the PERC RAID adapter documentation.

Windows 2000 and Windows Server 2003 Dynamic Disks and Volumes

The Windows operating system does not support dynamic disks (upgraded disks) or volumes as
shared cluster storage. If the shared cluster storage is configured as a dynamic disk, the Cluster
Configuration wizard is not able to discover the disks, preventing the cluster and network clients
from accessing the disks.

Naming and Formatting Drives on the Shared Storage System

After the virtual disks are created, write the disk signature, assign drive letters to the virtual
disks, and then format the drives as NTFS drives. Format the drives and assign drive letters from
only one cluster node.

NOTICE: Accessing the hard drives from multiple cluster nodes may corrupt the file system.

Preparing Your Systems for Clustering 45

Assigning Drive Letters and Mount Points

A mount point is a drive attached to an empty folder on an NTFS volume. A mount point
functions the same as a normal drive but is assigned a label or name instead of a drive letter. Using
mount points, a cluster can support more shared disks than the number of available drive letters.

The cluster installation procedure does not automatically add the mount point into the disks
managed by the cluster. To add the mount point to the cluster, create a physical disk resource in
the cluster resource group for each mount point. Ensure that the new physical disk resource is in
the same cluster resource group and is dependent on the root disk.

NOTE: Mount points are only supported in MSCS on the Windows Server 2003 operating systems. When

mounting a drive to an NTFS volume, do not create mount points from the quorum resource or between
the clustered disks and the local disks. Mount points must be in the same cluster resource group and

www.dell.com | support.dell.com

must be dependent on the root disk.

NOTICE: If the disk letters are manually assigned from the second node, the shared disks are

simultaneously accessible from both nodes. To ensure file system integrity and prevent possible data loss
before you install the MSCS software, prevent any I/O activity to the shared drives by performing the
following procedure on one node at a time and ensuring that the other node is shut down.

Before installing MSCS, ensure that both nodes have the same view of the shared storage
systems. Because each node has access to hard drives that are in a common storage array, each
node must have identical drive letters assigned to each hard drive. Up to 22 logical drive letters
(E through Z) can be used for the shared storage systems.

NOTE: Drive letters A through D are reserved for the local system.

The number of drive letters required by individual servers in a cluster may vary.
Dell recommends that the shared drives be named in reverse alphabetical order beginning with
the letter z.

To assign drive letters and format drives on the shared storage system, perform the
following steps:

1

With node 2 shut down, open

Disk Management

on node 1.

NOTE: For Windows 2000 Advanced Server, Array Manager launches instead of Disk Management

if Array Manager is installed on your system.

2

Allow Windows to enter a signature on all new physical or logical drives.

NOTE: Do not create dynamic disks on your hard drives.

3

Locate the icon for the first unnamed, unformatted drive on the shared storage system.

4

Right-click the icon and select

If the unformatted drives are not visible, verify the following:

• The latest version of the PERC RAID adapter driver is installed.

• The storage system is properly cabled to the servers.

• The split-bus module on the PowerVault 22

46 Preparing Your Systems for Clustering

Create

from the submenu.

x

S is set to cluster mode.

5

In the dialog box, create a partition the size of the entire drive (the default) and then

OK

click

6

Click

7

With the mouse pointer on the same icon, right-click and select

Path

8

Assign a drive letter to an NTFS volume or create a mount point.

.

NOTE: The MSCS software allows only one node to access a logical drive at a time. If a logical

drive is partitioned into multiple disks, only one node is able to access all the partitions for that
logical drive. If each node must access a separate disk, two or more logical drives must be present
in the storage system.

Yes

to confirm the partition.

Change Drive Letter and

from the submenu.

To assign a drive letter to an NTFS volume:

a

Click

Edit

and select the letter you want to assign to the drive (for example, z).

b

Click OK.

c

Go to step 9.

To create a mount point:

a

Click

Add

.

b

Click

Mount in the following empty NTFS folder

c

Type the path to an empty folder on an NTFS volume, or click

d

Click OK.

e

Go to step 9.

9

Click

Yes

to confirm the changes.

10

Right-click the drive icon again and select

11

Under

Volume Label

Email_Data

12

In the dialog box, change the file system to

, enter a descriptive name for the new volume; for example,

.

Format

NTFS

.

Browse

from the submenu.

, select

Quick Format

to locate it.

, and click the

button.

Disk_Z or

Start

NOTE: The NTFS file system format is required for shared-disk resources under MSCS.

13

Click OK at the warning.

14

Click OK to acknowledge that the format is complete.

15

Click

Close

to close the dialog box.

16

Repeat step 3 through step 15 for each remaining drive.

17

Close

Disk Management

18

Shut down node 1.

19

Turn on node 2.

.

Preparing Your Systems for Clustering 47

20

On node 2, open

21

Ensure that the drive letters for node 2 are correct.

To modify the drive letters on node 2, repeat step 7 through step 9.

Installing and Configuring MSCS

MSCS is a software component in Windows 2000 and an integrated service in Windows
Server 2003. MSCS performs the basic cluster functionality, which includes membership,
communication, and failover management. When MSCS is installed properly, the service starts
on each node and responds automatically if one of the nodes fails or goes offline. To provide
application failover for the cluster, the MSCS software must be installed on both cluster nodes.

www.dell.com | support.dell.com

See "Using MSCS" for more information.

NOTE: For systems with split backplane modules installed, the cluster installation tries to use the logical

drives on the secondary backplane as cluster disks. Because these drives are not accessible to all nodes
in the cluster, ensure that they are removed from the cluster after the installation is complete.

NOTE: In Windows Server 2003, mapping a network drive to the same drive letter as a cluster disk

resource renders the cluster disk inaccessible from Windows Explorer on the host. Ensure that mapped
network drives and cluster disks are never assigned the same drive letter.

Verifying Cluster Readiness

To ensure that your server and storage systems are ready for MSCS installation, ensure that
these systems are functioning correctly and verify the following:

• All cluster servers are able to log on to the same domain.

• The shared disks are partitioned and formatted, and the same drive letters that reference
logical drives on the shared storage system are used on each node.

• For each attached PowerVault 22
before power-up.

• Cluster mode is enabled on all PERC RAID adapters connected to shared storage.

• The controller’s SCSI IDs (6 or 7) on each node are different.

• All peer PERC RAID adapters are connected to the same PowerVault system through the
same channel number.

• All IP addresses and network names for each system node are communicating with each other
and the rest of the network. The private IP addresses should not be accessible from the LAN.

Disk Management

x

S storage system, the split-bus module is set to cluster mode

.

48 Preparing Your Systems for Clustering

Creating the Quorum Resource

When you install Windows 2000 Advanced Server in your cluster, the software installation
wizard prompts you for a location to install the quorum resource. When you install Windows
Server 2003 in your cluster, the software installation wizard automatically selects an NTFS disk
as the quorum disk, which you can modify later using Cluster Administrator. To prevent
quorum disk corruption, Dell and Microsoft recommend that you do not place applications or
data on the disk.

Creating a Partition for the Quorum Disk Resource

Dell recommends creating a separate partition—approximately 1 GB in size—for the
quorum resource.

When you create the partition for the quorum resource:

• Format the partition with NTFS.

• Use the partition exclusively for your quorum logs.

• Do not store any application data or user data on the quorum resource partition.

• To properly identify the quorum resource, Dell recommends that you assign the drive letter Q
to the quorum resource partition.

Dell does not recommend using the remainder of the virtual disk for other cluster resources. If
you do use the space for cluster resources, be aware that when you create two volumes
(partitions) on a single virtual disk, they will both fail over together if a server fails.

Preventing Quorum Disk Failure

Because the quorum disk plays a crucial role in cluster operation, losing a quorum disk causes
the entire cluster to fail. To prevent cluster failure, configure the quorum disk on a RAID
volume in the shared storage system.

NOTICE: Dell recommends that you use a RAID level other than RAID 0, which is commonly called

striping. RAID 0 configurations provide very high performance, but they do not provide the level of
availability required for the quorum resource.

Configuring Cluster Networks Running Windows 2000

When you install and configure a cluster running Windows 2000 Advanced Server, the software
installation wizard prompts you to identify the public and private network segments connected
to your cluster nodes. Dell recommends the following configuration, which provides added fault
tolerance for the private network:

1

Set the private network (cluster interconnect) to

2

Name this network

Private

.

Use for internal communications only

Preparing Your Systems for Clustering 49

.

3

Set the client system’s public network segment(s) to

This setting provides a redundant path for the cluster-to-cluster communication in the event
the private network fails.

4

For all subsequent network adapters, set each network adapter to

communications

5

Set the priority of the networks so that the network you designated as
priority for internal communications.

You can set the priority of the networks when you install MSCS or when using Cluster
Administrator software.

.

All communications

Client use only

Private

.

or

All

has the highest

www.dell.com | support.dell.com

Configuring Cluster Networks Running Windows Server 2003

When you install and configure a cluster running Windows Server 2003, the software
installation wizard automatically assigns and configures the public and private networks for your
cluster. You can rename a network, allow or disallow the cluster to use a particular network, or
modify the network role using Cluster Administrator. Dell recommends that you configure at
least one network for the cluster interconnect (private network) and one network for
all communications.

Installing and Configuring Microsoft Windows 2000 Cluster Service (MSCS)

After the Windows 2000 setup files are copied to the cluster node, Windows 2000 reboots and
enters the second phase of installation.

During this phase, the Windows 2000 setup prompts you for the product key, licensing mode,
system name, and additional information before starting the Windows 2000 Components
program. From this dialog box, you can select the Windows 2000 core components for
installation. Selecting Cluster Service copies the required files to the system.

If your operating system was preinstalled by Dell or you did not select Cluster Service when you
installed the operating system, you can install the Cluster Service later by running the
Add/Remove Components from the Control Panel.

After the Windows 2000 setup completes and all clustering prerequisites are met, run the
Cluster Configuration wizard to complete the installation, add the other node to the cluster,
and configure cluster resources.

To access the Cluster Configuration wizard:

1

From node 1, click the

2

In the

Run

C:\windows\cluster\cluscfg.exe

Start

button and select

box, type the following:

Run

.

3

Click OK.

50 Preparing Your Systems for Clustering

4

Follow the directions in the

5

Add the remaining node to the cluster:

a

From node 2, click the

b

In the

Run

box, type

c

Click OK.

d

When prompted, select

Cluster Configuration Wizard

Start

button and select

Run

window.

.

C:\windows\cluster\cluscfg.exe

The second or next node in the cluster

.

.

You are joining an existing cluster.

Configuring Microsoft Windows Server 2003 Cluster Service (MSCS)

The cluster setup files are automatically installed on the system disk. To create a new cluster:

From either node, click the

1

double-click

2

From the

3

In the

The

New Server Cluster Wizard

4

Click

5

Follow the procedures in the wizard, and then click

6

Add the second node to the cluster.

a

Turn on the remaining node.

b

Click the

Cluster Administrator

File

menu, select

Action

box of the

Next

to continue.

Start

button, select

Cluster Administrator

c

d

e

From the

In the

In the

File

menu, select

Action

box of the

Cluster or server name

an available cluster from the list, and then click

Add Nodes Wizard

The

Start

button, select

Programs→ Administrative Tools,

.

Open Connection

.

Open Connection to Cluster

appears.

Finish

Programs→ Administrative Tools,

.

Open Connection

.

Open Connection to Cluster

box, type the name of the cluster or click

OK

window appears.

, select

Create new cluster

.

, select

.

and then

.

and double-click

Add nodes to cluster

Browse

to select

.

• If the Add Nodes Wizard

• If the Add Nodes Wizard

does not

generates a cluster feasibility error, go to "Adding Cluster

generate a cluster feasibility error, go to step f.

Nodes Using the Advanced Configuration Option."

f

Click

Next

to continue.

g

Follow the procedures in the wizard, and then click

Finish

.

Preparing Your Systems for Clustering 51

Adding Cluster Nodes Using the Advanced Configuration Option

If you are adding additional nodes to the cluster using the Add Nodes Wizard and the nodes are
not configured with identical internal storage devices, the wizard may generate one or more
errors while checking cluster feasibility in the Analyzing Configuration menu. If this situation
occurs, select Advanced Configuration Option in the Add Nodes Wizard to add the nodes
to the cluster.

To add the nodes using the Advanced Configuration Option:

1

From the

2

In the

Action

OK

click

The

www.dell.com | support.dell.com

3

4

5

6

7

8

9

10

11

Add Nodes Wizard

Click

Next

In the

Select Computers

In the

Enter the object names to select (examples)

to add to the cluster, with each system name separated by a semicolon.

Click

Check Names

The Add Nodes Wizard verifies and underlines each valid system name.

Click OK.

In the

Select Computer

In the

Advanced Configuration Options

and then click

In the

Add Nodes

In the

Analyzing Configuration

cluster configuration.

If Cluster Administrator discovers a problem with the cluster configuration, a warning icon
appears in
if needed.

File

menu in Cluster Administrator, select

box of the

Open Connection to Cluster

.

window appears.

.

menu, click

Browse

.

s menu, click

Add

window, click

OK

.

window, click

Next

.

menu, Cluster Administrator analyzes the

Checking cluster feasibility

. Click the plus sign (+) to review any warnings,

Open Connection

, select

Add nodes to cluster

.

, and

.

, type the names of one to seven systems

.

Advanced (minimum) configuration

,

12

Click

Next

to continue.

13

In the

Password

field of the

used to run the Cluster Service, and click

The

Proposed Cluster Configuration

settings for your cluster.

52 Preparing Your Systems for Clustering

Cluster Service Account

Next

.

menu appears with a summary with the configuration

menu, type the password for the account

14

Click

Next

to continue.

The new systems (hosts) are added to the cluster. When completed,

15

16

Adding Nodes to the Cluster

in the

NOTE: This process may take several minutes to complete.

Click

Next

to continue.

In the

Completing the Add Nodes Wizard

menu.

, click

Finish

.

Tasks completed

appears

Verifying Cluster Functionality

To verify cluster functionality, monitor the cluster network communications to ensure that your
cluster components are communicating properly with each other. Also, verify that MSCS is
running on the cluster nodes.

Verifying MSCS Operation

After you install MSCS, verify that the service is operating properly.

1

Click the

2

In the

•In the

•In the

•In the

Start

Services

Name

Status

Startup Type

button and select

window, verify the following:

column,

column, Cluster Service is set to

column, Cluster Service is set to

Programs→

Cluster Service

appears.

Administrative Tools

Started

.

Automatic

, and then select

.

Services

Verifying Cluster Resource Availability

In the context of clustering, a resource is a basic unit of failover management. Application
programs are made up of resources that are grouped together for recovery purposes. All recovery
groups, and therefore the resources that comprise the recovery groups, must be online (or in a
ready state) for the cluster to function properly.

To verify that the cluster resources are online:

1

Start

Cluster Administrator

2

Click the

Administrator

3

Open a connection to the cluster and observe the running state of each resource group. If a
group has failed, one or more of its resources might be offline.

Start

button and select

.

on the monitoring node.

Programs→

Administrative Tools (Common)→

Cluster

.

Installing Applications in the Cluster Group

The Cluster Group contains a network name and IP address resource, which is used to manage
the cluster. Because the Cluster Group is dedicated to cluster management and for best cluster
performance, Dell recommends that you do not install applications in this group.

Preparing Your Systems for Clustering 53

Troubleshooting Failed Resources

Troubleshooting the failed resources is beyond the scope of this document, but examining the
properties of each resource and ensuring that the specified parameters are correct are the first
two steps in this process. In general, if a resource is offline, it can be brought online by
right-clicking the resource and selecting Bring Online from the pull-down menu. See the
documentation and online help for Windows 2000 Advanced Server or Windows Server 2003 for
information about troubleshooting resource failures.

Obtaining More Information

See Microsoft's online help for configuring the Cluster Service.

www.dell.com | support.dell.com

See "Using MSCS" for more information about the Cluster Service.

54 Preparing Your Systems for Clustering

Installing Your Cluster Management Software

This section provides information on configuring and administering your cluster using
Microsoft
cluster management.

®

Cluster Administrator. Microsoft provides Cluster Administrator as a built-in tool for

Microsoft Cluster Administrator

Cluster Administrator is Microsoft's tool for configuring and administering a cluster. The following
procedures describe how to run Cluster Administrator locally on a cluster node and how to install
the tool on a remote console.

Launching Cluster Administrator on a Cluster Node

1

Click the

2

Select

3

Select

Running Cluster Administrator on a Remote Console

You can administer and monitor the Cluster Service remotely by installing the
Windows Administration Tools package and Cluster Administrator on a remote console (or
management station) running the Microsoft Windows
Administrator is part of the Administration Tools package, which is included with the following
operating systems

• Windows 2000 Advanced Server

• Windows Server™ 2003 Enterprise Edition

• Windows Server 2003 Enterprise x64 Edition

The Windows 2000 Administration Tools can only be installed on systems running Windows 2000.
Additionally, the Windows 2003 Administrative Tools can only be installed on systems running
Windows XP (with Service Pack 1 or later) and Windows Server 2003.

Start

button and select

Administrative Tools

Cluster Administrator

:

.

.

Programs

.

®

operating system. Cluster

Installing Your Cluster Management Software 55

To install Cluster Administrator and the Windows Administration Tools package on a remote
console

www.dell.com | support.dell.com

, perform the following steps:

1

Select a system that you wish to configure as the remote console.

2

Identify the operating system that is currently running on the selected system.

3

Insert the appropriate operating system CD into the system’s CD drive:

•

Microsoft Windows 2000 Advanced Server CD

•

Windows Server 2003 Enterprise Edition CD

•

Windows Server 2003 Enterprise x64 Edition CD

4

Open an Explorer window, navigate to the system’s CD drive and double-click the \i386
directory.

5

If you inserted the

Edition CD

If you inserted the

WADMINPAK.MSI

6

Follow the instructions on your screen to complete the installation.

Launching Cluster Administrator on a Remote Console

Perform the following steps on the remote console:

1

Ensure that the Windows Administrative Tools package was installed on the system.

2

Click the

3

Select

Administrative Tools

4

Select

Cluster Administrator

Windows 2000 Advanced Server CD

, double-click

Windows Server 2003 Enterprise x64 Edition CD

.

Start

button and select

ADMINPAK.MSI

Programs

.

.

or

Windows Server 2003 Enterprise

.

, double-click

.

Installing Cluster Administrator for Windows Clusters on a Remote Console

You cannot install the Windows 2000 or Windows Server 2003 Administration Tools package on
clients running any version of Windows NT
Windows Server 2003, Enterprise Edition cluster can be remotely administered using the Cluster
Administrator included with Windows NT 4.0 operating systems with limited support.

See your Windows NT 4.0 operating system documentation for more information about the
installation of Cluster Administrator on a remote client.

Cluster Administration and Monitoring

When using Cluster Administrator provided with Windows NT 4.0 operating systems on a
system running Windows NT 4.0, Cluster Administrator may generate error messages if the
software detects Windows 2000 or Windows Server 2003 cluster resources. Dell strongly
recommends using client systems running Windows 2000 or Windows Server 2003 with the
appropriate Administrator Pack for cluster administration and monitoring.

56 Installing Your Cluster Management Software

®

4.0. However, a Windows 2000 Advanced Server or

Using MSCS

This section provides information about Microsoft® Cluster Service (MSCS). This section is
intended to be an overview of MSCS and provides information about the following:

• Cluster objects

• Cluster networks

• Network interfaces

• Cluster nodes

• Groups

• Cluster resources

• Failover and failback

For information about specific MSCS procedures, see the MSCS online help.

NOTE: In this guide and in other cluster documentation, the quorum resource is also referred to as the

quorum disk.

Cluster Objects

Cluster objects are the physical and logical units managed by MSCS. Each object is associated
with the following:

• One or more properties, or attributes, that define the object and its behavior within the
cluster

• A set of cluster control codes used to manipulate the object's properties

• A set of object management functions used to manage the object through MSCS

Cluster Networks

A network performs one of the following roles in a cluster:

• A network that carries internal cluster communication

• A public network that provides client systems with access to cluster application services

• A public-and-private network that carries both internal cluster communication and connects
client systems to cluster application services

• Neither a public nor private network that carries traffic unrelated to cluster operation

Using MSCS 57

Preventing Network Failure

The Cluster Service uses all available private and public-and-private networks for internal
communication. Configure multiple networks as private or public-and-private to protect the
cluster from a single network failure. If there is only one such network available and it fails, the
cluster nodes stop communicating with each other. When two nodes are unable to
communicate, they are partitioned, and the Cluster Service automatically shuts down on one
node. While this shutdown guarantees the consistency of application data and the cluster
configuration, it can make cluster resources unavailable.

For example, if each node has only one network adapter, and the network cable on one of the
nodes fails, each node (because it is unable to communicate with the other) attempts to take
control of the quorum disk. There is no guarantee that the node with a functioning network

www.dell.com | support.dell.com

connection will gain control of the quorum disk. If the node with the failed network cable gains
control, the entire cluster is unavailable to network clients. To avoid this problem, ensure that
all nodes have at least two networks and are configured to use both networks for the private
network (internal communications).

Node-to-Node Communication

The Cluster Service does not use public-only networks for internal communication. For
example, a cluster has Network A configured as private and Network B configured as public. If
Network A fails, the Cluster Service does not use Network B because it is public; the nodes stop
communicating, and one node terminates its Cluster Service.

Network Interfaces

The Microsoft® Windows® operating system keeps track of all network adapters in a server
cluster. This tracking system allows you to view the state of all cluster network interfaces from a
cluster management application, such as Cluster Administrator.

Cluster Nodes

A cluster node is a system in a server cluster that has a working installation of the Windows
operating system and the Cluster Service.

Cluster nodes have the following characteristics:

• Every node is attached to one or more cluster storage devices. Each cluster storage device
attaches to one or more disks. The disks store all of the cluster's configuration and resource
data. Each disk can be owned by only one node at any point in time, but ownership can be
transferred between nodes. The result is that each node has access to all cluster
configuration data.

• Every node communicates with the other nodes in the cluster through one or more network
adapters that attach nodes to networks.

58 Using MSCS

• Every node in the cluster is aware of another system joining or leaving the cluster.

• Every node in the cluster is aware of the resources that are running on all nodes in the cluster.

• All nodes in the cluster are grouped under a common cluster name, which is used when
accessing and managing the cluster.

Table 5-1 defines various states of a node that can occur in cluster operation.

Table 5-1. Node States and Definitions

State Definition

Down The node is not actively participating in cluster operations.

Joining The node is in the process of becoming an active participant in the

cluster operations.

Paused The node is actively participating in cluster operations but cannot take ownership of

resource groups and cannot bring resources online.

Up The node is actively participating in all cluster operations, including hosting

cluster groups.

Unknown The state cannot be determined.

When the Cluster Service is installed for the first time on a node, the administrator must
choose whether that node forms its own cluster or joins an existing cluster. When the
Cluster Service is started on a node, that node searches for other active nodes on networks
enabled for internal cluster communications.

Forming a New Cluster

If a node cannot join a cluster, the node attempts to form the cluster by gaining control of the
quorum disk. If the node gains control of the quorum disk, the node forms the cluster and uses
the recovery logs in the quorum disk to update its cluster database. The Cluster Service
maintains a consistent, updated copy of the cluster database on all active nodes.

Joining an Existing Cluster

A node can join an existing cluster if it can communicate with another cluster node. If a cluster
exists and the joining node finds an active node, it attempts to join that node's cluster. If it
succeeds, the Cluster Service then validates the node's name and verifies version compatibility.
If the validation process succeeds, the node joins the cluster. The node is updated with the
latest copy of the cluster database.

Using MSCS 59

Groups

A group is a collection of cluster resources with the following characteristics:

• All of the resources in the group are moved to the alternate node when one resource in a
group fails and it is necessary to move the resource to an alternate node.

• A group is always owned by one node at any point in time, and a resource is always a member
of a single group. Therefore, all of a group's resources reside on the same node.

Groups enable resources to be combined into larger logical units. Typically a group is made up of
related or dependent resources, such as applications and their associated peripherals and data.
However, groups can also be established with resources that are unrelated and nondependent to
balance the load or for administrative convenience.

www.dell.com | support.dell.com

Every group maintains a prioritized list of the nodes that can and should act as its host. The
preferred nodes list is generated by the Cluster Service. Cluster Service produces a list of
preferred nodes for a group from the list of possible owners that is maintained by the group's
resources and can be modified by an Administrator.

To maximize the processing power of a cluster, establish at least as many groups as there are
nodes in the cluster.

Cluster Resources

A cluster resource is any physical or logical component that has the following characteristics:

• Can be brought online and taken offline

• Can be managed in a server cluster

• Can be hosted (owned) by only one node at a time

To manage resources, the Cluster Service communicates to a resource Dynamic Link Libraries
(DLL) through a Resource Monitor. When the Cluster Service makes a request of a resource,
the Resource Monitor calls the appropriate entry-point function in the resource DLL to check
and control the resource's state.

60 Using MSCS

Dependent Resources

A dependent resource requires—or depends on—another resource to operate. For example, if a
Generic Application resource requires access to clustered physical storage, it would depend on a
physical disk resource. A resource can specify one or more resources on which it is dependent; it
can also specify a list of nodes on which it is able to run.

The following terms describe resources in a dependency relationship:

•

Dependent resource

Dependency

•

•

Dependency tree

resource B, resource B depends on resource C, and so on.

Resources in a dependency tree obey the following rules:

– A dependent resource and all of its dependencies must be in the same group.

– The Cluster Service takes a dependent resource offline before any of its dependencies are

taken offline and brings a dependent resource online after all its dependencies are online,
as determined by the dependency hierarchy.

— A resource on which another resource depends.

— A resource that depends on other resources (the dependencies).

— A series of dependency relationships such that resource A depends on

Setting Resource Properties

Using the resource Properties dialog box, you can perform the following tasks:

• View or change the resource name

• View or change the resource description and possible owners

• Assign a separate memory space for the resource

• View the resource type, group ownership, and resource state

• View which node currently owns the resource

• View pre-existing dependencies and modify resource dependencies

• Specify whether to restart a resource and the settings used to restart the resource (if required)

• Check the online state of the resource by configuring the
intervals in the Cluster Service

• Specify the time requirement for resolving a resource in a pending state (

Offline Pending

• Set specific resource parameters

The General, Dependencies, and Advanced tabs are the same for every resource. Some resource
types support additional tabs.

Properties of a cluster object should not be updated on multiple nodes simultaneously. See the
MSCS online documentation for more information.

) before the Cluster Service places the resource in

Looks Alive

and

Is Alive

Online Pending

Offline

or

Failed

polling

or

status

Using MSCS 61

Configuring Resource Dependencies

Groups function properly only if resource dependencies are configured correctly. The Cluster
Service uses the dependencies list when bringing resources online and offline. For example, if a
group in which a physical disk and a file share are located is brought online, the physical disk
containing the file share must be brought online before the file share.

Table 5-2 shows resources and their dependencies. The resources in the right column must be
configured before you create the resource.

Table 5-2. Cluster Resources and Required Dependencies

Resource Required Dependencies

www.dell.com | support.dell.com

File share Network name (only if configured as a distributed file system [DFS] root)

IP address None

Network name IP address that corresponds to the network name

Physical disk None

Setting Advanced Resource Properties

You can configure the advanced resource properties using the Advanced tab in the resource
Properties dialog box. Use the Advanced tab to have the Cluster Service perform the following

tasks:

• Restart a resource or allow the resource to fail.

– To restart the resource, select

– To fail over the resource group to another cluster node when the resource fails, select

Affect the group

do not select
cluster node.

The

Threshold

resource before the resource fails over to a healthy cluster node.

Affect the group

and then enter the appropriate values in

Affect the group

, the resource group will not fail over to the healthy

(if applicable).

Threshold

and

Period

. If you

value determines the number of attempts by the Cluster Service to restart the

The

• Adjust the time parameters for
(

detailed check of the resource

• Select the default number for the resource type.

To apply the default number, select

• Specify the time parameter for a resource in a pending state (

Pending

62 Using MSCS

Period

value assigns a time requirement for the

Looks Alive

(general check of the resource) or

) to determine if the resource is in the online state.

Use resource type value

) to resolve its status before moving the resource to

Threshold

.

Offline

value to restart the resource.

Online Pending

or

Failed

Is Alive

or

status.

Offline

Resource Parameters

The Parameters tab in the Properties dialog box is available for most resources. Table 5-3 lists
each resource and its configurable parameters.

Table 5-3. Resources and Configurable Parameters

Resource Configurable Parameters

File share Share permissions and number of simultaneous users

Share name (clients will detect the name in their browse or explore lists)

Share comment

Shared file path

IP address IP address

Subnet mask

Network parameters for the IP address resource (specify the correct cluster
network)

Network name System name

Physical disk Drive for the physical disk resource (the drive cannot be changed after the

resource is created)

Quorum Disk (Quorum Resource)

The quorum resource is a common resource in the cluster that is accessible by all of the cluster
nodes. Normally a physical disk on the shared storage, the quorum resource maintains data
integrity, cluster unity, and cluster operations—such as forming or joining a cluster—by
performing the following tasks:

•

Enables a single node to gain and defend its physical control of the quorum resource

When the cluster is formed or when the cluster nodes fail to communicate, the quorum
resource guarantees that only one set of active, communicating nodes is allowed to form
a cluster.

•

Maintains cluster unity —

The quorum resource allows cluster nodes that can communicate
with the node containing the quorum resource to remain in the cluster. If a cluster node fails
for any reason and the cluster node containing the quorum resource is unable to
communicate with the remaining nodes in the cluster, MSCS automatically shuts down the
node that does not control the quorum resource.

•

Stores the most current version of the cluster configuration database and state
data

— If a cluster node fails, the configuration database helps the cluster recover a failed

resource or recreates the cluster in its current configuration.

The only type of resource supported by the Cluster Service that can act as a quorum resource is
the physical disk resource. However, developers can create their own quorum disk types for any
resources that meet the arbitration and storage requirements.

—

Using MSCS 63

Using the Quorum Disk for Cluster Integrity

The quorum disk is also used to ensure cluster integrity by performing the following functions:

• Maintaining the cluster node database

• Ensuring cluster unity

When a node joins or forms a cluster, the Cluster Service must update the node's private copy of
the cluster database. When a node joins an existing cluster, the Cluster Service can retrieve the
data from the other active nodes. However, when a node forms a cluster, no other node is
available. The Cluster Service uses the quorum disk's recovery logs to update the node's cluster
database, thereby maintaining the correct version of the cluster database and ensuring that the
cluster is intact.

www.dell.com | support.dell.com

For example, if node 1 fails, node 2 continues to operate, writing changes to the cluster
database. Before you can restart node 1, node 2 fails. When node 1 becomes active, it updates
its private copy of the cluster database with the changes made by node 2 using the quorum
disk’s recovery logs to perform the update.

To ensure cluster unity, the operating system uses the quorum disk to ensure that only one set of
active, communicating nodes is allowed to operate as a cluster. A node can form a cluster only if
it can gain control of the quorum disk. A node can join a cluster or remain in an existing cluster
only if it can communicate with the node that controls the quorum disk.

For example, if the private network (cluster interconnect) between cluster nodes 1 and 2 fails,
each node assumes that the other node has failed, causing both nodes to continue operating as
the cluster. If both nodes were allowed to operate as the cluster, the result would be two separate
clusters using the same cluster name and competing for the same resources. To solve this
problem, MSCS uses the node that owns the quorum disk to maintain cluster unity and solve
this problem. In this scenario, the node that gains control of the quorum disk is allowed to form
a cluster, and the other fails over its resources and becomes inactive.

Resource Failure

A failed resource is not operational on the current host node. At periodic intervals, the Cluster
Service checks to see if the resource appears operational by periodically invoking the Resource
Monitor. The Resource Monitor uses the resource DLL for each resource to detect if the
resource is functioning properly. The resource DLL communicates the results back through the
Resource Monitor to the Cluster Service.

Adjusting the Poll Intervals

You can specify how frequently the Cluster Service checks for failed resources by setting the
Looks Alive (general resource check) and Is Alive (detailed resource check) poll intervals. The
Cluster Service requests a more thorough check of the resource's state at each Is Alive interval
than it does at each Looks Alive interval; therefore, the Is Alive poll interval is typically longer
than the Looks Alive poll interval.

NOTE: Do not adjust the Looks Alive and Is Alive settings unless instructed by technical support.

64 Using MSCS

Adjusting the Threshold and Period Values

If the resource DLL reports that the resource is not operational, the Cluster Service attempts to
restart the resource. You can specify the number of times the Cluster Service can attempt to
restart a resource in a given time interval. If the Cluster Service exceeds the maximum number
of restart attempts (Threshold value) within the specified time period (Period value), and the
resource is still not operational, the Cluster Service considers the resource to be failed.

NOTE: See "Setting Advanced Resource Properties" to configure the Looks alive, Is alive, Threshold, and

Period values for a particular resource.

NOTE: Do not adjust the Threshold and Period values settings unless instructed by technical support.

Configuring Failover

You can configure a resource to fail over an entire group to another node when a resource in that
group fails for any reason. If the failed resource is configured to cause the group that contains
the resource to fail over to another node, Cluster Service will attempt a failover. If the number of
failover attempts exceeds the group's threshold and the resource is still in a failed state, the
Cluster Service will attempt to restart the resource. The restart attempt will be made after a
period of time specified by the resource's Retry Period On Failure property, a property common
to all resources.

When you configure the Retry Period On Failure properly, consider the following guidelines:

• Select a unit value of minutes, rather than milliseconds (the default value is milliseconds).

• Select a value that is greater or equal to the value of the resource’s restart period property.
This rule is enforced by the Cluster Service.

NOTE: Do not adjust the Retry Period On Failure settings unless instructed by technical support.

Creating a New Resource

Before you add a resource to your PowerEdge Cluster, you must verify that the following
elements exist in your cluster:

• The type of resource is either one of the basic types provided with MSCS or a custom resource
type provided by the application vendor, Microsoft, or a third party vendor.

• A group that contains the resource already exists within your cluster.

• All dependent resources have been created.

• A separate Resource Monitor—recommended for any resource that has caused problems in
the past.

To create a new resource:

1

Click the

The

2

In the console tree (usually the left pane), double-click the

Start

button and select

Cluster Administrator

Programs→

window appears.

Administrative Tools→

Groups

folder.

Cluster Administrator

Using MSCS 65

.

3

In the details pane (usually the right pane), click the group to which you want the resource
to belong.

4

On the

5

In the New Resource wizard, type the appropriate information in
click the appropriate information in

6

Click

Next

7

Add or remove possible owners of the resource, and then click

The

New Resource

dependencies

8

To

add

www.dell.com | support.dell.com

9

10

11

dependencies, under

To

remove

Repeat step 7 for any other resource dependencies, and then click

Set the resource properties.

For more information on setting resource properties, see the MSCS online help.

Deleting a Resource

1

Click the

The

Cluster Administrator

2

In the console tree (usually the left pane), click the

3

In the details pane (usually the right pane), click the resource you want to remove.

4

In the

File

When you delete a resource, Cluster Administrator also deletes all the resources that have a
dependency on the deleted resource.

File

menu, point to

.

window appears with

selections.

dependencies, under

Start

button and select

menu, click

Delete

New

, and then click

Resource type

Available resources

Available resources

Resource dependencies

Programs→

window appears.

.

Resource

, click a resource, and then click

Administrative Tools→ Cluster Administrator

.

Name

and

Description

and

Group

.

Next

.

and

Resource

, click a resource, and then click

Finish

.

Resources

folder.

Add

, and

.

Remove

.

.

File Share Resource Type

If you want to use a PowerEdge Cluster as a high-availability file server, you will need to select
the type of file share for your resource. Three ways to use this resource type are available:

•

Basic file share

•

Share subdirectories

its immediate subfolders. This method is an efficient way to create large numbers of related
file shares on a single file server. For example, you can create a file share for each user with
files on the cluster node.

•

DFS root

roots cannot be managed by this resource. A DFS root file share resource has required
dependencies on a network name and an IP address. The network name can be either the
cluster name or any other network name for a virtual server.

66 Using MSCS

— Publishes a single file folder to the network under a single name.

— Publishes several network names—one for each file folder and all of

— Creates a resource that manages a stand-alone DFS root. Fault-tolerant DFS

Configuring Active and Passive Cluster Nodes

Active nodes process application requests and provide client services. Passive nodes are backup
nodes that ensure that client applications and services are available if a hardware or software
failure occurs. Cluster configurations may include both active and passive nodes.

NOTE: Passive nodes must be configured with appropriate processing power and storage capacity to

support the resources that are running on the active nodes.

Your cluster solution supports variations of active/active (activex) and active/passive

x

(active
active or passive.

Cluster solutions running the Windows operating system support active/active and
active/passive configurations.

An active/active (active
services on each node. When an application is running on node 1, the remaining node(s) do not
have to wait for node 1 to fail. Those node(s) can run their own cluster-aware applications (or
another instance of the same application) while providing failover for the resources on node 1.
For example, multiway failover is an active/active failover solution because running applications
from a failed node can migrate to multiple active nodes in the cluster. However, you must ensure
that adequate resources are available on each node to handle the increased load if one node fails.

/passivex) configurations. The variable x indicates the number of nodes that are

x

) configuration contains virtual servers running separate applications or

•

In an active/passive (activex/passivex) configuration, one or more active cluster nodes are
processing requests for a clustered application while the passive cluster nodes only wait for
the active node(s) to fail.

Failover and Failback

This section provides information about the failover and failback capabilities of the
Cluster Service.

Failover

When an individual application or user resource (also known as a cluster resource) fails on a
cluster node, the Cluster Service will detect the application failure and try to restart the
application on the cluster node. If the restart attempt reaches a preset threshold, the Cluster
Service brings the running application offline, moves the application and its resources to
another cluster node, and restarts the application on the other cluster node(s). This process of
automatically moving resources from a failed cluster node to other healthy cluster node(s) is
called failover.

In order to fail over and fail back running applications, cluster resources are placed together in a
group so the Cluster Service can move the cluster resources as a combined unit. This process
ensures that the failover and/or failback procedures transfer all of the user resources as
transparently as possible.

Using MSCS 67

After failover, the Cluster Administrator can reset the following recovery policies:

• Application dependencies

• Application restart on the same cluster node

• Workload rebalancing (or failback) when a failed cluster node is repaired and brought
back online

Failover Process

The Cluster Service attempts to fail over a group when any of the following conditions occur:

• The node currently hosting the group becomes inactive for any reason.

• One of the resources within the group fails, and it is configured to affect the group.

www.dell.com | support.dell.com

• Failover is forced by the System Administrator.

When a failover occurs, the Cluster Service attempts to perform the following procedures:

• The group’s resources are taken offline.

The resources in the group are taken offline by the Cluster Service in the order determined by
the group'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, allowing the application to write changes to the disk before the
disk is taken offline.

• The resource is taken offline.

Cluster Service takes a resource offline by invoking, through the Resource Monitor, the
resource DLL that manages the resource. If the resource does not shut down within a
specified time limit, the Cluster Service forces the resource to shut down.

• The group is transferred to the next preferred host node.

When all of the resources are offline, the Cluster Service attempts to transfer the group to the
node that is listed next on the group's list of preferred host nodes.

For example, if cluster node 1 fails, the Cluster Service moves the resources to the next cluster
node number, which is cluster node 2.

• The group’s resources are brought back online.

If the Cluster Service successfully moves the group to another node, it tries to bring all of the
group's resources online. Failover is complete when all of the group's resources are online on
the new node.

68 Using MSCS

The Cluster Service continues to try and fail over a group until it succeeds or until the number
of attempts occurs within a predetermined time span. A group’s failover policy specifies the
maximum number of failover attempts that can occur in an interval of time. The Cluster
Service will discontinue the failover process when it exceeds the number of attempts in the
group’s failover policy.

Modifying Your Failover Policy

Because a group’s failover policy provides a framework for the failover process, make sure that
your failover policy is appropriate for your particular needs. When you modify your failover
policy, consider the following guidelines:

• Define the method in which the Cluster Service detects and responds to individual resource
failures in a group.

• Establish dependency relationships between the cluster resources to control the order in
which the Cluster Service takes resources offline.

•Specify

Time-out

–

–

Threshold

Time-out

, failover

Threshold

, and failover

Period

for your cluster resources.

controls how long the Cluster Service waits for the resource to shut down.

and

Period

control how many times the Cluster Service attempts to fail over a

resource in a particular period of time.

• Specify a

Possible owner list

for your cluster resources. The

Possible owner list

for a resource

controls which cluster nodes are allowed to host the resource.

Failback

When the System Administrator repairs and restarts the failed cluster node, the opposite
process occurs. After the original cluster node has been restarted and rejoins the cluster, the
Cluster Service will bring the running application and its resources offline, move them from the
failover cluster node to the original cluster node, and then restart the application. This process
of returning the resources back to their original cluster node is called failback.

You can configure failback to occur immediately at any given time, or not at all. However, be
sure to configure the failback time during your offpeak hours to minimize the effect on users, as
they may see a delay in service until the resources come back online.

Using MSCS 69

www.dell.com | support.dell.com

70 Using MSCS

Upgrading Your System to a Cluster Configuration

This section provides additional procedures for upgrading your non-clustered system to a
Dell™ PowerEdge™ Cluster SE500W solution with the cluster components currently installed
in your system. These cluster configurations include the Microsoft
Server, and Windows Server™ 2003 Enterprise Edition operating systems with MSCS.

Certification Requirements

All hardware components of the PowerEdge Cluster SE500W must pass Microsoft Windows
Hardware Qualification Labs (MS WHQL) certification. Each cluster configuration must be
certified and qualified against the latest MSCS certification program.

Before You Begin

Before you upgrade your non-clustered system to a Cluster SE500W solution:

• Back up the data on your cluster nodes.

• Verify that your cluster hardware and storage systems meet the minimum system
requirements for a Cluster SE500W as described in "System Requirements."

• Verify that your cluster hardware and storage systems are installed and configured properly as
explained in the following sections:

– "Cabling Your Cluster Hardware"

– "Preparing Your Systems for Clustering"

Dell certifies and supports only Cluster SE500W solutions that are configured with the Dell
products described in this guide. For a description of the PowerEdge cluster components, see
the Platform Guide.

®

Windows® 2000 Advanced

Upgrading Your System to a Cluster Configuration 71

Configurations Using Non-Dell Products

Configurations using non-Dell products, such as servers, rack cabinets, and storage systems,
have not been approved by any safety agencies. It is the customer’s responsibility to have these
systems evaluated for any safety hazards or implications by a certified safety agency.

NOTE: Your system is safety-certified as a free-standing unit and as a component for use in a rack

cabinet using the customer rack kit when both the rack cabinet and rack kit were designed for your
system. The installation of your system and rack kit in any other rack cabinet has not been approved by
any safety agencies. It is your responsibility to have the final combination of system and rack kit in a
cabinet evaluated for suitability by a certified safety agency. The manufacturer disclaims all warranties
and liability in connection with such combinations.

www.dell.com | support.dell.com

Completing the Upgrade

After you install the required hardware and network adapter upgrades, you can set up and cable
the system hardware.

The final phase for upgrading to a Cluster SE500W solution is to install and configure Windows
2000 Advanced Server, or Windows Server 2003 with MSCS.

Upgrading Your Operating System

You can upgrade your Windows 2000 Advanced Server cluster to a Windows Server 2003
Enterprise Edition cluster using one of the following methods:

•

Standard upgrade

nodes are disconnected from the client network. This procedure requires you to recreate your
cluster configuration.

See your Windows operating system documentation for performing a standard upgrade.

•

Rolling upgrade

node is connected to the client network and available to handle client requests. This
procedure
node must be configured with the appropriate resources to run all virtual servers and services
for the entire cluster while you upgrade the remaining node.

The following section explains how to perform a rolling upgrade on a two-node cluster running
Windows 2000 Advanced Server.

— Upgrading the operating system on each cluster node while all cluster

— Upgrading the operating system on each cluster node while one cluster

does not

require you to recreate your cluster configuration. However, each cluster

NOTE: Only new (full) installations of the Microsoft Windows Server 2003 Enterprise x64 Edition are

permitted. You cannot upgrade from existing operating system to the Windows Server 2003 Enterprise
x64 Edition.

72 Upgrading Your System to a Cluster Configuration

Performing a Rolling Upgrade

Before you perform a rolling upgrade:

• Ensure that your cluster nodes are running Windows 2000 Advanced Server.

• Backup your data and system states.

• Run the Check System Compatibility Wizard to determine if your cluster nodes are
configured with the appropriate resources to run Windows Server 2003 Enterprise Edition.

• Ensure that your cluster service account user privileges are set to

Act as part of operating system

Upgrading Node 1

1

Click the

2

In

Start

button and select

Cluster Administrator

The following steps refer to the node you selected as node 1.

3

Right-click a node and then click

4

Right-click a cluster group and then click

The Cluster Group is moved and restarted on node 2.

5

Repeat step 4 for the remaining cluster groups.

6

Insert the

7

Double-click

The

Microsoft Windows Server 2003 Enterprise Edition

Install Windows Server 2003 Enterprise Edition.

Windows Setup Wizard

.

Programs→ Administrative Tools→

, click one of the nodes in the cluster.

Paus e

.

Move Group

.

CD into the CD drive.

window appears.

Cluster Administrator

.

8

Follow the instructions in the Windows Setup Wizard to upgrade your operating system.

NOTE: If you are running IIS World Wide Web Publishing service on your cluster node, this service

is disabled during the upgrade to protect your system.

9

Verify that the upgraded node is functioning correctly.

a

Click the

Administrator

b

Move one or more cluster resource groups from node 2.

c

Verify that the resource group(s) can be brought online.

d

Close

10

Remove the

11

Go to "Upgrading Node 2."

NOTE: After you upgrade node 1, your cluster is running two separate operating systems. Dell

recommends that you do not modify your cluster configuration—such as adding or removing cluster
nodes or resources—until you upgrade both cluster nodes.

Start

button and select

.

Cluster Administrator

Enterprise Edition

CD from the CD drive.

Programs→ Administrative Tools→

.

Upgrading Your System to a Cluster Configuration 73

Cluster

Upgrading Node 2

1

On node 2, click the

Administrator

2

In

Cluster Administrator

3

Right-click node 2 and then click

4

Right-click a cluster group and then click

The cluster group is moved and restarted on node 1.

5

Repeat step 4 for the remaining cluster groups.

6

Insert the

7

www.dell.com | support.dell.com

Double-click

Windows Setup Wizard

The

8

Follow the instructions in the
operating system.

NOTE: If you are running IIS World Wide Web Publishing service on your cluster node, this service

is disabled during the upgrade to protect your system.

9

Verify that the upgraded node is functioning correctly.

a

Click the

Administrator

b

Move one or more cluster resource groups from node 1.

c

Verify that the resource group(s) can be brought online.

d

Close

10

In

Cluster Administrator

Start

button and select

Programs→ Administrative Tools→

.

, right click node 1, and then click

Pause Node

.

Move Group

.

Microsoft Windows Server 2003 Enterprise Edition

Install Windows Server 2003 Enterprise Edition

window appears.

Start

button and select

Windows Setup Wizard

Programs→ Administrative Tools→

window to upgrade your

.

Cluster Administrator

.

, redistribute the cluster groups to the appropriate cluster nodes.

Cluster

Resume Node

.

CD into the CD drive.

.

Cluster

74 Upgrading Your System to a Cluster Configuration

Maintaining Your Cluster

This section provides the following maintenance procedures for systems running the
Microsoft

• Adding a network adapter to a cluster node

• Changing the IP address of a cluster node

• Uninstalling MSCS

• Running

• Recovering from a corrupt quorum disk

• Replacing a cluster node

®

Windows®2000 Advanced Server, and Windows Server™ 2003 operating systems:

chkdsk /f

on a quorum disk

Adding a Network Adapter to a Cluster Node

This procedure assumes that Windows 2000 Advanced Server, or Windows Server 2003 with the
latest Windows Service Pack, and MSCS are installed on both cluster nodes.

NOTE: The IP addresses used in the following sections are examples only and do not represent of actual

addresses to use. The IP addresses are 192.168.1.101 for the network adapter in the first node and

192.168.1.102 for the network adapter in the second node. The subnet mask for both nodes
is 255.255.255.0.

1

Move all cluster resources from the cluster node you are upgrading to another node in
the cluster.

See the Cluster Service documentation for information about moving cluster resources to a
specific node.

2

Shut down the cluster node you are upgrading and install the additional network adapters in
that system.

See the system
expansion cards in your system.

Installation and Troubleshooting Guide

for instructions about installing

Maintaining Your Cluster 75

3

Boot to the Windows operating system.

Windows Plug and Play detects the new network adapter and installs the appropriate drivers.

NOTE: If Plug and Play does not detect the new network adapter, the network adapter is

not supported.

Update the network adapter drivers (if required).

a

b

After the drivers are installed, click the
double-click

c

In the

d

Right-click the new network adapter, and then select

e

www.dell.com | support.dell.com

4

Assign a unique static IP address, subnet mask, and gateway.

Ensure that the network ID portion of the new network adapter’s IP address is different from
the other adapter.

For example, if the first network adapter in the node had an address of 192.168.1.101 with a
subnet mask of 255.255.255.0, you might enter the following IP address and subnet mask for
the second network adapter:

Network Connections

Connections

box, locate the new network adapter that you installed in the system.

Start

button, select

Control Panel

, and then

.

Properties

.

IP address:

Subnet mask:

5

Click OK and exit network adapter properties.

6

On the Windows desktop, click the

Tools

7

Click the

8

Verify that a new resource called "New Cluster Network" appears in the window.

192.168.2.102

255.255.255.0

→ Cluster Administrator

Network

tab.

Start

button and select

Programs→

Administrative

.

To rename the new resource, right-click the resource and enter a new name.

9

Move all cluster resources to another cluster node.

10

Repeat step 2 through step 9 on each cluster node.

NOTE: Ensure that you assign the new network adapter the same IP address as the second

network adapter on the first node (for example, 192.168.2.101) as you did with the second node.

If the installation and IP address assignments have been performed correctly, all of the new
network adapter resources appear online and respond successfully to ping commands.

76 Maintaining Your Cluster

Changing the IP Address of a Cluster Node on the Same IP Subnet

NOTE: If you are migrating your cluster nodes to a different subnet, take all cluster resources offline and

then migrate all nodes together to the new subnet.

1

Open

Cluster Administrator

2

Stop MSCS on the cluster node.

The Cluster Administrator utility running on the second cluster node indicates that the first
node is down by displaying a red icon in the

3

Reassign the IP address.

4

If you are running DNS, verify that the DNS entries are correct (if required).

5

Restart MSCS on the cluster node.

The cluster nodes re-establish their connection and the Cluster Administrator changes the
node icon back to blue to show that the node is back online.

.

Cluster Service

window.

Uninstalling MSCS From Clusters Running Windows 2000 Advanced Server

1

Take all resource groups offline or move them to another cluster node.

2

Stop Cluster Service on the node that you want to uninstall.

3

Click the

4

Select

5

Deselect the check box for Cluster Service and click

6

Click

7

From the remaining node, click the

Tools

8

Right-click the node icon and select

9

Close

10

Restart the Cluster Service.

Start

button and select

Add/Remove Windows Components

Finish

.

→ Cluster Administrator

Cluster Administrator

Settings→

Start

.

Options→

.

Control Panel→

.

Next

.

button and select

Evict Node

Add/Remove Programs

Programs→ Administrative

.

.

Maintaining Your Cluster 77

Removing Nodes From Clusters Running Windows Server 2003

1

Take all resource groups offline or move them to another cluster node.

2

Click the

Cluster Administrator

3

In

Cluster Administrator

select

Stop Cluster Service

4

In

Cluster Administrator

select

Evict Node

www.dell.com | support.dell.com

If you cannot evict the node and the node is the last node in the cluster:

a

Open a command prompt.

b

Type the following:

cluster node <

where

5

Close

Cluster Administrator

Running chkdsk /f on a Quorum Disk

You cannot run the chkdsk command with the /f (fix) option on a device that has an open file
handle active. Because MSCS maintains an open handle on the quorum resource, you cannot
run chkdsk /f on the hard drive that contains the quorum resource.

To run chkdsk /f on a quorum resource’s hard drive:

1

Move the quorum resource temporarily to another drive:

a

Right-click the cluster name and select

b

Click the

c

Select another disk as the quorum disk and press <Enter>.

2

Run

chkdsk /f

3

Move the quorum disk back to the original drive.

Start

button, select

.

<node_name>

Quorum

on the drive that previously stored the quorum resource.

Programs→ Administrative Tools,

.

, right-click the icon of the node you want to uninstall and then

.

, right-click the icon of the node you want to uninstall and then

node_name

is the cluster node you are evicting from the cluster.

.

tab.

> /force

Properties

.

and then double-click

78 Maintaining Your Cluster

Recovering From a Corrupt Quorum Disk

The quorum disk maintains the configuration data necessary for cluster recovery when a cluster
node fails. If the quorum disk resource is unable to come online, the cluster will not start and all
of the shared drives will be unavailable. If this situation occurs, and you need to run chkdsk on
the quorum disk, you can start the cluster manually from the command line.

To start the cluster manually from a command prompt:

1

Open a command prompt window.

2

Select the cluster folder directory by typing one of the following:

cd \2000\cluster
cd \windows\cluster

3

Start the cluster in manual mode (on one node only) with no quorum logging by typing
the following:

Clussvc -debug -noquorumlogging

MSCS starts.

4

Run

chkdsk /f

To run t h e

a

Open a second command prompt window.

b

Ty p e :

on the disk designated as the quorum resource.

chkdsk /f

chkdsk /f

(for Windows 2000 Advanced Server), or

(for Windows Server 2003)

utility:

5

After the

6

Restart the cluster service.

To restart MSCS from the

a

b

c

chkdsk

utility completes, stop MSCS by pressing <Ctrl><c>.

Click the

In the

Services

Services

Start

button and select

window, right-click

console:

In the drop-down menu, click the

Programs→

Cluster Service

Start

button.

Administrative Tools→

.

Services

.

To restart MSCS from the command prompt:

a

Open the second command prompt window that you opened in step 4a.

b

Type the following:

Net Start Clussvc

The Cluster Service restarts.

See the Microsoft Knowledge Base article 258078 located at the Microsoft Support website at
www.microsoft.com for more information on recovering from a corrupt quorum disk.

Maintaining Your Cluster 79

Replacing a Cluster-Enabled Dell PERC RAID Adapter

If you have to replace a cluster-enabled PERC RAID adapter, perform the following steps:

1

Turn off the failed node.

2

Disconnect the failed PERC RAID adapter’s cable from the shared storage system.

NOTICE: If you replace your PERC RAID adapter, ensure that you enable cluster mode on the

replacement PERC RAID adapter before you connect the SCSI cables to the shared storage system.

NOTICE: A failed PERC 4/DC card must be replaced with a PERC4/DC card. A failed PERC 4e/DC card

must be replaced with a PERC 4e/DC card.

3

Replace the failed PERC RAID adapter in the system without reconnecting the cable.

www.dell.com | support.dell.com

4

Turn on the system with the replaced PERC RAID adapter and run the
BIOS configuration utility.

NOTICE: If you replace a PERC RAID adapter that will be connected to shared storage system, you must

set the appropriate SCSI ID before you connect the SCSI cables to the shared storage system.

5

Change the SCSI ID so that it differs from the SCSI ID on the peer cluster node's
PERC RAID adapter.

NOTE: See your PERC RAID adapter documentation for more information about changing the

SCSI ID. Also, see the cluster configuration tables (if you completed the information in the tables) in

the Dell PowerEdge Cluster SE500W Platform Guide.

6

Shut down the system.

7

Reconnect the system to the shared storage system.

8

Restart the system and restore the RAID configuration using configuration information
stored on the disks. See the PERC RAID adapter documentation for more information about
this procedure.

Replacing a Cluster Node

This section assumes that you have a recent tape backup of the cluster node that contains the
local registry information.

1

Ensure that the replacement cluster node is physically disconnected from the storage system.

2

Ensure that Windows and the latest Windows Service Pack are installed and configured
properly on the replacement node.

3

Install the correct network adapter drivers, assign the appropriate IP addresses, and install the
PERC RAID adapter driver.

4

Shut down the replacement node.

5

Connect the SCSI cables from each PERC RAID adapter to the Dell™ PowerVault™ 22xS
storage system.

80 Maintaining Your Cluster

6

Turn on the replacement cluster node.

If the PERC RAID adapter has been replaced, the following error message is reported:

Configuration of NVRAM and drives mismatch (Normal mismatch)
Run View/Add Configuration option of Config Utility
Press <Ctrl><H> for WebBIOS
Press A Key to Run Configuration Utility
Or <Alt><F10> to Continue

7

Press any key to enter the RAID controller's BIOS configuration utility, and select

Configure→ View/Add Configuration→ View Disk Configuration

Verify that the configuration that displays includes the existing configuration on the disks.

8

Press <Esc>, and select

9

Restart the system and allow Windows to start normally.

10

Add the replacement node to the cluster.

11

Use Cluster Administrator to verify that the node rejoins the cluster, and check the

Event Viewer

12

Reinstall any cluster applications (such as Microsoft SQL Server or Exchange Server onto the
new node, if required).

NOTE: You may need to reinstall or configure your cluster applications before moving or testing the

failover capabilities of any cluster resources to the new node.

to ensure errors were not encountered.

Yes

to save the disk configuration, and exit the configuration utility.

.

Windows

Changing the Cluster Service Account Password in Windows Server 2003

To change the cluster service account password for all nodes in a cluster running Windows
Server 2003, open a command prompt and type the following syntax:

Cluster /cluster:[

where

cluster_name

For help with changing the cluster password, type the following:

cluster /changepass /help

NOTE: Blank passwords are not allowed as cluster service account passwords in Windows Server 2003.

cluster_name

is the name of your cluster.

] /changepass

Maintaining Your Cluster 81

Reformatting a Cluster Disk

NOTE: Ensure that all client systems are disconnected from the cluster disk before you perform

this procedure.

1

Click the

2

In the

Cluster Administrator

3

In the

Groups

reformatted, and select

4

In the

Cluster Administrator

reformatting and select

5

In the

www.dell.com | support.dell.com

6

7

8

9

10

Cluster Administer

reformatting and select

Properties

The

Click the

In the

Advanced

In the

Specify value

6000000

where

6000000

Click

Apply

On the Windows desktop, right-click

Computer Management

The

Start

button and select

directory, right-click a cluster resource group that contains the disk to be

Take Offline

Bring Online

Properties

window appears.

Advanced

.

tab.

tab menu in the

field, type:

equals 6,000,000 milliseconds (or 100 minutes).

Programs→

left window pane, expand the

.

right window pane, right-click the physical disk you are

.

right window pane, right-click the physical disk you are

.

"Looks Alive" poll interval

My Computer

window appears.

Administrative Tools→

Groups

and select

directory.

box, select

Manage

Cluster Administrator

Specify value

.

.

.

11

In the

Computer Management

The physical disk information appears in the right window pane.

12

Right-click the disk you want to reformat and select

Disk Management

13

In the

File

menu, select

14

In the

"Looks Alive" poll interval

15

In the

Cluster Administrator

reformatted disk and select

16

In the

File

menu, select

82 Maintaining Your Cluster

left window pane, click

reformats the disk.

Exit

.

box, select

left window pane, right-click the cluster group that contains the

Bring Online

Exit

.

Disk Management

Format

Use value from resource type

.

.

.

and click OK.

Adding New Physical Drives to an Existing Shared Storage System

The Dell™ PowerEdge™ Cluster SE500W solutions consist of two systems that share an
external SCSI storage enclosure PowerVault 22xS storage system. Each system contains a
PERC RAID adapter with cluster-enabled firmware. The following procedure describes adding
additional storage to an existing shared storage system in the cluster configuration.

To add new physical drives to an existing shared storage system in the cluster, perform the
following steps:

1

Stop all I/O activity.

2

Ensure that both nodes are online.

3

Install the new physical hard drives into the PowerVault 22xS storage system.

CAUTION: See the Dell PowerVault 220S and 221S Installation and Troubleshooting Guide, which

provides safety instructions for installing components into the PowerVault 22xS storage system.

4

Restart node 1 and press <Ctrl><m> during the system POST to launch the PERC RAID
adapter BIOS Configuration utility.

5

Configure the virtual disks.

NOTE: See the PERC RAID adapter documentation.

6

Restart node 1.

7

After system restarts, use Disk Manager to write the disk signature, create a new partition,
assign drive letters, and format the partition with NTFS.

8

Restart node 1.

9

On node 1, use Cluster Administrator to add a new group (for example Disk Groupn:).

10

Select possible owners, but do not bring the group online yet.

11

Add a new resource (for example, Disk z:).

12

Select

Physical Disk

13

Select possible owners, and select the drive letter that you assigned to the new array.

14

Bring the new group that you just added online.

15

Reboot node 2, and ensure that node 2 is completely online before you continue.

16

To verify that the new resource group is online and the drive is accessible using the cluster
name, connect to
added disk, and use Cluster Administrator
the other cluster node.

for the type of resource, and assign it to the new group you just created.

\\clustername\n$

, where n is the drive letter you assigned to the newly

to verify that you can move the new disk group to

Maintaining Your Cluster 83

Rebuilding Operation in Dell OpenManage Utilities

For the rebuild operation, see your Dell OpenManage™ Array Manager or Dell OMSM
documentation.

If the cluster node is rebooted or power to the node is lost while a PERC RAID adapter is
rebuilding a shared array, the controller terminates the rebuild operation and identifies the hard
drive as failed. This condition also occurs if the rebuild is performed from the PERC RAID
adapter BIOS Configuration utility and the user exits the utility before the rebuild completes.
This condition occurs with all versions of the PERC RAID adapter firmware on both standard
and cluster-enabled controllers.

If the second node in the clustered configuration is turned on, it restarts the operation.

www.dell.com | support.dell.com

If the rebuild fails to complete due to a system restart, the rebuild must be reinitiated using the
PERC RAID adapter BIOS configuration utility or using the Array Manager running under the
appropriate Microsoft Windows operating system.

Upgrading the PowerVault 22xS EMM Firmware Using Array Manager

NOTE: Before upgrading the EMM firmware, ensure that all I/O activity is suspended, and shut down the

node you will not be working from when you upgrade the EMM firmware. If you do not shut down the
second node, the firmware on the EMM attached to that node may not be updated.

NOTE: Use Array Manager version 3.7 or later to perform the following procedure.

To download the PowerVault 22xS EMM firmware onto a cluster node:

Download the latest EMM firmware from

1

support.dell.com

2

Shut down node B.

3

Stop all I/O activity on node A.

4

Launch the Array Manager Console from node A by clicking the

Programs→

5

In the

Arrays

<Cluster>

x

where
(0 or 1) to which the enclosure is attached.

the

Dell Support website (located at

) to your hard drive or to a diskette.

Start

button and selecting

Dell OpenManage Applications→

directory, select

→ <Channel 0>

indicates the number associated with the controller on the system. Select the channel

PERC Subsystem 1→

or

<Channel 1>

Array Manager→

PERC RAID Adapter Controller x

.

Array Manager console.

6

If you downloaded the EMM firmware to a diskette, ensure that the diskette is inserted.

7

Right-click the enclosure icon for the desired channel, and select

You can also click the channel number and select

8

From the
that you downloaded to your hard drive or diskette.

84 Maintaining Your Cluster

Firmware Download

dialog box, click

Download Firmware

Download Firmware

Browse

and navigate to the EMM firmware

from the

.

Task Menu

.

9

Verify that the selected file is correct.

10

Click

Download Firmware

to begin the download process.

This process takes several minutes to complete.

11

When the message

12

Repeat steps 3 through 9 for each channel that has an enclosure attached.

13

To verify the firmware upgrade for each channel, right-click the channel number, select

Properties

14

Start up node B and resume I/O activity.

, and view the version information.

Firmware Downloaded Successfully

appears, click OK.

Maintaining Your Cluster 85

www.dell.com | support.dell.com

86 Maintaining Your Cluster

Troubleshooting

This appendix provides troubleshooting information for Dell™ PowerEdge™ Cluster SE500W
configurations.

Table A-1 describes general cluster problems you may encounter and the probable causes and
solutions for each problem.

Table A-1. General Cluster Troubleshooting

Problem Probable Cause Corrective Action

The RAID drives in the Dell
PowerVault™ 22xS storage
system are not accessible by one
of the cluster nodes, or the
shared storage system is not
functioning properly with the
cluster software.

A disk resource will not move
over to another node or will not
come online.

The SCSI cables are loose or
defective, or the cables
exceed the maximum
allowable length.

The PERC RAID adapters
connected to a single storage
system are not configured
consistently.

The PowerVault 22xS is not
running in cluster mode.

If the cluster has multiple
PowerVault 22xS storage
systems, the cabling between
the PERC RAID adapter and
the storage systems is wrong.

Check the cable connections or replace
the cable with a working cable. For more
information on the length of SCSI cables,
see "Cabling Your Cluster Hardware."

Ensure that the RAID configuration is
identical for each channel between the
PERC 3/DC cards connected to a shared
storage system. Ensure that cluster mode
is enabled on both PERC RAID adapters
and that their SCSI IDs are different on
each node.

Configure the PowerVault 22xS for cluster
mode. For more information, see
"Preparing Your Systems for Clustering."

Ensure that the cables attached to each
channel of the PERC RAID adapter in
each server node are connected to the
correct storage system and that the
channels on an optional second PERC
RAID adapter in each server node are
connected to the correct system.

Attach or replace the SCSI cable between
the cluster node and the shared
storage system.

Troubleshooting 87

Table A-1. General Cluster Troubleshooting (continued)

Problem Probable Cause Corrective Action

Enclosure management

Install EMMs.
modules (EMMs) are
not installed.

The PERC RAID adapter
drivers are not installed in
your Microsoft

®

Windows®

Install the drivers. See the PERC RAID

adapter documentation for more

information.
operating system.

The option to change the SCSI
IDs is not visible in the

www.dell.com | support.dell.com

PERC 3/DC BIOS.

One or more of the SCSI
controllers are not detected by
the system.

PERC RAID adapters hang
during boot.

Cluster mode is not enabled. Enabling cluster mode will permit you to

change the SCSI IDs.

The controllers for the
shared storage system have
the same SCSI ID as their
peer adapters in the other
system (that is, the same
SCSI ID as the controllers
connected to the other side
of the shared

Change one of the controller SCSI IDs so

that the ID numbers do not conflict. Set

the controller in the primary node to

SCSI ID 7 (default), and set the

controller in the secondary node to SCSI

ID 6. See the PERC RAID adapter

documentation for more information

about setting SCSI host adapter IDs.
storage system).

Array Manager and PERC
RAID adapter BIOS utility only
report 13 drives in
cluster mode.

Normal The SCSI ID limitations are imposed by

SCSI protocol. As a result of this

limitation, the last slot in the PowerVault

22xS storage system cannot be utilized in

cluster mode.

One of the nodes takes a long
time to join the cluster, or fails
to join the cluster.

The node-to-node network
has failed due to a cabling or
hardware failure.

Check the network cabling. Ensure that

the node-to-node interconnection and the

public network are connected to the

correct network adapters.

Long delays in node-to-node
communications may
be normal.

Verify that the nodes can communicate

with each other by running the ping

command from each node to the other

node. Try both the host name and IP

address when using the ping command.

One or more nodes may have
the Internet Connection
Firewall enabled, and the
RPC communication
between the nodes has been
inadvertently blocked.

Configure the Internet Connection

Firewall to allow communications that are

required by the cluster service and the

clustered applications or services. For

more information see KB883398 at

support.microsoft.com.

88 Troubleshooting

Table A-1. General Cluster Troubleshooting (continued)

Problem Probable Cause Corrective Action

You are prompted to configure
one network instead of two
during MSCS installation.

The TCP/IP configuration is
incorrect.

The node-to-node network and public
network must be assigned static IP
addresses on different subnets. See
"Assigning Static IP Addresses to Your
Cluster Resources and Components" for
information about assigning the
network IPs.

Client systems are dropping off
of the network while the cluster
is failing over.

Only one network segment
appears during Cluster Service
installation.

The private (point-to-point)
network is disconnected.

With MSCS, the service
provided by the recovery
group becomes temporarily
unavailable to clients during
failover. Clients may lose
their connection if their
attempts to reconnect to the
cluster are too infrequent or
if they end too abruptly.

Public and private networks
segments are not unique.

Ensure that both systems are powered on
so that both network adapters in the
private network are available.

The time that the service is temporarily
unavailable varies depending on the
application. Contact the application
program vendor for more information.

Place all installed network adapters in a
cluster node on separate IP networks.

Ensure that the same network segments
that were used for each network adapter
are identical on the second cluster node.

Troubleshooting 89

Table A-1. General Cluster Troubleshooting (continued)

Problem Probable Cause Corrective Action

Attempts to connect to a
cluster using Cluster
Administrator fail.

The Cluster Service has not
been started.

A cluster has not been
formed on the system.

Verify that the Cluster Service is running

and that a cluster has been formed. Use

the Event Viewer and look for the

following events logged by the Cluster

Service:

The system has just been
booted and services are
still starting.

Microsoft Cluster Service

successfully formed a cluster

on this node.

www.dell.com | support.dell.com

or

Microsoft Cluster Service

successfully joined

the cluster.

If these events do not appear in Event

Viewer, see the Microsoft Cluster Service

Administrator’s Guide for instructions on

setting up the cluster on your system and

starting the Cluster Service.

Using Microsoft
Windows NT

®

4.0 to remotely
administer a Windows 2000
Advanced Server or Windows
Server™ 2003 cluster generates
error messages.

The cluster network name is
not responding on the
network because the Internet
Connection Firewall is
enabled on one or more
nodes.

Normal. Some resources in
Windows 2000, Advanced
Server and Windows Server
2003 are not supported in
Windows NT 4.0.

Configure the Internet Connection
Firewall to allow communications that are
required by the cluster service and the
clustered applications or services. For
more information see KB883398 at
support.microsoft.com.

Dell strongly recommends that you use
Windows 2000 Professional, Windows
2000 Server, or Windows 2000 Advanced
Server for remote administration of a
cluster running Windows 2000 Advanced
Server. Similarly, you should use Windows
XP Professional or Windows Server 2003
for remote administration of a cluster
running Windows Server 2003.

MSCS does not show any
available shared disks
during installation.

The PERC RAID adapter
drivers are not installed in
the operating system.

Disks are configured as
dynamic disks.

Install the drivers. See the PERC RAID
adapter documentation for
more information.

Change disks to "basic" before cluster
installation. See "Maintaining Your
Cluster" for more information on
configuring dynamic disks as basic disks.

90 Troubleshooting

Table A-1. General Cluster Troubleshooting (continued)

Problem Probable Cause Corrective Action

One of the nodes can access
one of the shared hard drives,
but the second node cannot.

NOTE: For support and troubleshooting information for Dell PowerVault™ 220S and 221S Systems refer

to support.dell.com/support/edocs/stor-sys/spv22xs/. For support and troubleshooting information for
PERC cards refer to support.dell.com/support/edocs/storage/RAID/.

If MSCS is installed, this
situation is normal.

If MSCS is installed, only the node that
owns the disk resource will be able to
access the disk. The other node will
show the disk resource as offline in
Windows Disk Management.

Troubleshooting 91

www.dell.com | support.dell.com

92 Troubleshooting

Abbreviations and Acronyms

A

ampere(s)

API

Application Programming Interface

AC

alternating current

ACM

advanced cooling module

BBS

Bulletin Board Service

BDC

backup domain controller

BIOS

basic input/output system

bps

bits per second

BTU

British thermal unit

DHCP

dynamic host configuration protocol

DLL

dynamic link library

DNS

domain naming system

ESD

electrostatic discharge

EMM

enclosure management module

ERP

enterprise resource planning

F

Fahrenheit

FC

Fibre Channel

FCAL

Fibre Channel arbitrated loop

C

Celsius

cm

centimeter(s)

DC

direct current

DFS

distributed file system

ft

feet

FTP

file transfer protocol

g

gram(s)

GB

gigabyte

Abbreviations and Acronyms 97

Gb

gigabit

lb

pound(s)

Gb/s

gigabits per second

GUI

graphical user interface

HBA

host bus adapter

HSSDC

www.dell.com | support.dell.com

high-speed serial data connector

HVD

high-voltage differential

Hz

hertz

ID

identification

IIS

Internet Information Server

I/O

input/output

LAN

local area network

LED

light-emitting diode

LS

loop resiliency circuit/SCSI enclosure services

LVD

low-voltage differential

m

meter

MB

megabyte(s)

MB/sec

megabyte(s) per second

MHz

megahertz

MNS

Majority Node Set

IP

Internet Protocol

Kb

kilobit(s)

KB

kilobyte(s)

KVM

Keyboard Video Mouse

98 Abbreviations and Acronyms

MS WHQL

Microsoft Windows Hardware Qualification Labs

MSCS

Microsoft® Cluster Service

MSDTC

Microsoft Distributed Transaction Coordinator

NIC

Network Interface card

NLB

Network Load Balancing

RAM

random access memory

NTFS

NT File System

NVRAM

nonvolatile random-access memory

OMSM

OpenManage Enhanced Storage Management

PAE

physical address extension

PCB

printed circuit board

PDC

primary domain controller

PDU

power distribution unit

PERC

PowerEdge™ Expandable RAID Controller

PERC 4/DC

PERC fourth generation, dual channel

rpm

revolutions per minute

SAFTE

SCSI accessed fault-tolerant enclosures

SCSI

small computer system interface

sec

second(s)

SES

SCSI enclosure services

SMP

symmetric multiprocess

SNMP

Simple Network Management Protocol

TCP/IP

Transmission Control Protocol/Internet Protocol

UHDCI

ultra high-density connector interface

PERC 4e/DC

PERC fourth generation express, dual channel

PCI

Peripheral Component Interconnect

POST

power-on self-test

RAID

redundant array of independent disks

UPS

uninterruptible power supply

V

volt(s)

VHDCI

very high-density connector interface

WINS

Windows Internet Naming Service

Abbreviations and Acronyms 99

www.dell.com | support.dell.com

100 Abbreviations and Acronyms

Cluster Data Form

The configuration matrix and data form on the following pages are provided for the system
installer to record pertinent information about Dell™ PowerEdge™ Cluster SE500W
configurations. The data form is used for installing Microsoft
Server, and Windows Server™ 2003 operating systems on PowerEdge clusters.

Make a copy of the appropriate section of the data form to use for the installation or upgrade,
complete the requested information on the form, and have the completed form available if you
need to call Dell for technical assistance. If you have more than one cluster, complete a copy of
the form for each cluster.

PowerEdge Cluster SE500W Solution Data Form

You can attach the following form to the back of each cluster node or rack. The system installer
may want to use the form to record important information about the hardware on each cluster
component. Have a copy of the form available any time you call Dell for technical support.

Cluster Type PowerEdge Cluster SE500W Solution

Cluster name

Domain name

Cluster IP address

Cluster subnet mask
(same as public
network)

Cluster Service
account

Cluster Service
password

Installer

Date installed

Applications

Location

Notes

®

Windows® 2000 Advanced

Cluster Data Form 93

Node
(Server Name)

Node 1

Node 2

Network Settings TCP/IP Address Subnet Mask Private or Public?

Node 1,

adapter

www.dell.com | support.dell.com

Node 1,

adapter

Additional Node 1

network

1

network

2

network
adapter

(s)

Server Type Cluster Name Service Tag Number

Node 2,

adapter

Node 2,

adapter

Additional Node 2

network

1

network

2

network
adapter

System Storage 1 Storage 2 Storage 3 Storage 4 SCSI ID

Node 1, PERC RAID
adapter 1

Node 2, PERC RAID
adapter 1

Node 1, PERC RAID
adapter 2

Node 2, PERC RAID
adapter 2

(s)

94 Cluster Data Form

Dell PowerVault™
22xS Storage System

Storage 1

Storage 2

Storage 3

Storage 4

Component Storage 1 Storage 2 Storage 3 Storage 4

Service Tag

Description of Installed Items
(Drive letters, RAID types, applications/data)

PCI Slot
Number

PCI slot 1

PCI slot 2

PCI slot 3

PCI slot 4

PCI slot 5

PCI slot 6

PCI slot 7

PCI slot 8

PCI slot 9

PCI slot 10

PCI slot 11

Adapter Installed
(PERC RAID adapter,
network adapter, and so on)

Use (public network,
private network, shared
storage, internal drives)

PCI Slot Description

Cluster Data Form 95

www.dell.com | support.dell.com

96 Cluster Data Form