CHAPTER 7

           MICROSOFT EXCHANGE 2000 SERVER IN

                         CLUSTERED ENVIRONMENTS

 

 

Clustering is an advanced technology that allows you to design high availability solutions for your

enterprise-level applications.  Clustering is often brought into play in business-critical services such

as messaging in an enterprise environment.

 

 

Lesson 1:  Cluster Service Architecture

 

A cluster is basically a group of servers that appear as a single virtual server.  A virtual server

corresponds to a generic IP address and a network name, and allows your users to access all

the resources in the cluster, including Exchange 2000 Services, without having to know the names

of the individual nodes.

 

By grouping two or more computers together in a cluster, you can minimize system downtime

caused by software, network, and hardware failures.

 

 

Server Clusters in Exchange 2000 Environments

 

Clustering is a mature technology available for all popular network operating systems, including

Solaris and Windows 2000.

 

Despite assumptions and desires, it is impossible to achieve these goals with Windows 2000

Cluster service.  Although it is not a 100% perfect solution, clustering does reduce the number

of potential single points of failure and thereby improves service availability.

 

The advantage of clustering, however, is that your users can reconnect almost immediately.

 

 

Advantages of Server Clusters

 

Despite their limitations, clusters can significantly improve system availability in case of hardware

or software failures or during planned maintenance.

 

IMPORTANT  To take advantage of four-node clustering and 64GB memory support, you must

install Windows 2000 Datacenter Server.

 

 

=======================================================================

 

winexc7.html                                                   PAGE 2                                                         2002/06/10

 

 

 

 

The advantages of clustered Windows 2000 systems are:

 

 

 

Cluster Hardware Configuration

 

A shared Small Computer System Interface (SCSI) bus and two network cards are required in

addition to typical server equipment, such as a local disk where the operating system must be

installed.

 

The private network adapter is not a requirement, but it is necessary if you want to create a

cluster with complete hardware redundancy.

 

The nodes in a cluster are connected by up to three physical connections:

 

Shared Storage bus.  Connects all nodes to the disks (or RAID storage systems) where all

clustered data must reside.

 

Public network connection.  Connects client computers to the nodes in the cluster.

 

Private network Connection.  Connects the nodes in a cluster and ensures that the nodes

will be able to communicate with each other in the event of an outage of the public local

area network (LAN).  The private LAN is optional but highly recommended.  Low-cost

Ethernet cards are sufficient to accommodate the minimal traffic of the cluster communication.

 

 

TIP  It is advisable to purchase complete cluster sets from reliable hardware vendors instead

of configuring the cluster hardware manually.

 

NOTE:  Only one node can gain access to a particular disk at any given time even through the

disks on the shared SCSI bus are physically connected to all cluster nodes.

 

 

Windows 2000 Clustering Architecture

 

To act as one unit, clustering Windows 2000 computers must cooperate with each other closely. 

The Cluster service consists of several internal elements and relies on additional external

components to handle the required tasks.

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 3                                                         2002/06/10

 

 

 

 

Resource Monitoring and Failover Initiation

 

The health of each cluster resource is monitored by the Resource Monitor, which is implemented

n a separate process communicating with the Cluster service via remote procedure calls (RPCs).

 

 

Joining a Cluster and Event Handling

 

When you start a server that is part of a cluster, Windows 2000 is booted as usual, mounting and

configuring local, noncluster resources.  After all, each node in a Windows 2000 cluster is also a

Windows 2000 Advanced Server or Datacenter Server in itself.

 

From the point of view of other nodes in the cluster, a node may be in one of three states:  offline,

online, or paused.

 

 

Lesson 2:  Load Planning and Failover Strategies

 

Microsoft Outlook 2000 and Exchange 2000 Server are designed to overcome these shortcomings

in a smart way.  For example, you don’t need to restart Outlook 2000 to reconnect.  A simple

switch to another folder in your mailbox (let’s say from Inbox to Contacts) does the job.  Exchange

2000 Enterprise Server, again, supports multiple storage groups in the information store, which is

the basis of static load balancing in clustered Exchange 2000 Systems.

 

 

Load-Balancing Clusters

 

To best utilize the hardware resources available in a cluster, most organizations implement combined

application servers that provide more than one kind of client/server services to their users.

 

Active/Active Clustering

 

It may be desirable to dedicate individual clusters to one application type.  For instance, you might

want to configure a two-mode cluster for Microsoft SQL 2000 Server and another four-node

luster for Exchange 2000 Server only.  Keep in mind that each virtual server requires access to its

own disk resources, meaning one or more dedicated sets of physical disks.

 

NOTE:  Exchange 2000 Enterprise Server supports active/active clustering.

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 4                                                         2002/06/10

 

 

 

To maximize the use of all available server in a cluster while maintaining a failover solution, you can

configure multiple virtual servers and distribute them across your node.  However, only one

resource group and therefore one virtual server can own a resource.

 

TIP:  Theoretically, you can configure one virtual Exchange 2000 Server per physical disk, including

the disk containing the quorum resource.  However, Microsoft does not recommend adding

Exchange 2000 Server services to the virtual server representing the cluster (that is, owning the

quorum disk).  Defining dedicated virtual servers for Exchange 2000 simplifies service maintenance,

such as taking a virtual Exchange 2000 system offline.

 

 

Full Load with Hot Spare

 

When all nodes are online, Node D does not own a virtual Exchange 2000 server.  This is the hot

spare assuming the role of an Exchange 2000 server in case another node in the cluster fails or is

unavailable for maintenance reasons.  Provided that you don’t run other applications on the hot

spare, such as SQL 2000 Server, a single node failure will not affect the system performance.

 

Full-Load Active Nodes

 

The hot spare configuration bears the disadvantage of an idle server system when every node is

operational.  If you don’t want to invest in this extra server, configure a three-node cluster with

our disks.  The disadvantage of a full-load cluster is that users might experience performance

losses if one node has to take over the load of a failed member.

 

To avoid measurable performance losses, you would have to operate the nodes below the

following theoretical limits:

 

·          Two-node clusters.  50% of the node’s power (which is a good as configurating

an active/passive cluster and using 100% of the active node’s power.

·          Three-node Clusters.  66% of the node’s power.

·          Four-node Clusters.  75% of the node’s power.

 

However, in most cases it will be acceptable to operate the nodes of a full-load cluster at their

maximum capabilities because node failures or failures due to maintenance should seldom occur

and temporary performance losses are usually not critical.  Note that central processing unit

(CPU) utilization is not a perfect measure of the node’s power; a node will begin to experience

performance degradation if operating above 70% peak CPU utilization for an extended period

of time.

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 5                                                         2002/06/10

 

 

 

Exchange 2000 and Virtual Server Limitations

 

Several Exchange 2000 Server components are not supported in a cluster, and others can only

run in an active/passive configuration.  The Message Transfer Agent (MTA), for instance, cannot

run on more than one node in the cluster, implicitly enforcing an active/passive configuration.  The

same restriction applies to the Chat Service.

 

The Exchange 2000 components supports the following cluster configuration:

 

·        Chat.  Active/passive

·        Full-Text indexing. Active/Active.

·        HTTP.  Active/Active.

·        IMAP.   Active/Active.

·        Information Store.  Active/Active

·        Instant Messaging. Not supported.

·        Key Management Service (KMS).  Not supported.

·        MTA.  Active/passive

·        Network News Transfer Protocol (NNTP).  Not supported.

·        POP3.  Active/Active.

·        Simple Mail Transfer Protocol (SMTP).  Active/Active

·        System Attendant (SA).  Active/active.

 

Important:  Exchange 2000 Server supports only one public store in a cluster.  When adding an

additional virtual server, make sure you delete the public store in the new resource group before

bringing the virtual server online.

 

 

Failover and Failback Procedures

 

Failover and failback are cluster-specific procedures to move resource groups (with all their

associated resources) between nodes. 

 

 

Failover

 

A failover can occur in two situations:  Either you trigger it manually for maintenance reasons or

the Cluster Service initiates it automatically in case of a resource failure on the node owning the

resource.

 

The Failover Manager is now responsible for deciding where to move the resource group.  It

communicates with its counterparts on the remaining active nodes to arbitrate the ownership

of the resource group.  This arbitration relies on the node preference list that you can specify

when creating resources in Cluster Administrator.

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 6                                                         2002/06/10

 

 

 

Failback

 

If you have specified a preferred owner for a resource group, and this node comes back online

again, the Failover Manager will fail back the resource group to the recovered or restarted node. 

 

By Default, resource groups are set not to fail back automatically when the original node is recovered. 

Without manual configuration of a failback policy, groups continue to run on the alternate node after

the failed node comes back online.

 

 

Lesson 3:  Clustering Support in Exchange 2000 Server

 

Many Exchange 2000 Server components support active/active or at least active/passive clustering. 

However, several components, such as the NNTP service, the KMS, or connectors to other mail

systems are not supported.

 

 

Exchange 2000 Server Clustering Installation

 

The installation of Exchange 2000 Server in a cluster is a process of four stages. 

 

You need to install Windows 2000 Advanced Server or Datacenter Server on all nodes in the

cluster.  You have to install the Cluster service on the nodes and configure the cluster environment.

Exchange Server installation.  Exchange 2000 Server must be installed on all nodes with exactly

the same parameters, and, after that, you can configure and start virtual servers in Cluster Administrator.

 

IMPORTANT:  You cannot install Exchange 2000 Server on a nonclustered server and integrate

this installation into a cluster afterward.  Furthermore, it is a good idea to test the installation in a

cluster prior to deploying Exchange 2000 Server in the production environment.  Exchange 2000

Server may not function properly with your hardware.  You need to ensure that a dedicated

Physical Disk resource is available for Exchange 2000 Server.

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 7                                                         2002/06/10

 

 

 

Installing Exchange 2000 in a Cluster

 

It is easy to install Exchange 2000 Server in a clustered system.  Setup then copies and configures

the Exchange 2000 components and resource DLLs and sets the Exchange 2000 service to start

manually.

 

IMPORTANT  Do not start or stop clustered Exchange 2000 services in the Services management

tool.  You have to use Cluster Administrator to bring clustered services online or offline.

 

 

Clustered Installation Features and Dependencies

 

During installation of Exchange 2000 Server in a cluster, resource DLLs and other components are

added to the configuration of each node.  It is important that you install only one node at a time

using the same account you used to install the Cluster service.

 

During installation, you need to place the binary files on the local system drive of each node.  The

binary files are not shared between the nodes.  Don’t forget to make sure that drive M is not in

use on any node because when you configure virtual servers later on, the Microsoft Web Storage

System will use the M drive by default.

 

 

Setup Initialization File

 

Because you must install the same Exchange 2000 components on all the nodes of the cluster, it is

a good idea to create and use a SETUP.INI file.

 

 

Exercise Summary:

 

 

Installation of Exchange 2000 Server is straightforward process.  Setup detects that it is started on

a cluster node and configures the Exchange 2000 components and services appropriately.  The

Exchange 2000 services will be configured to start manually.  Don’t start the services in a cluster

in the Services Management tool because this undermines the cluster environment.  Clustered

services are managed within Cluster Administrator.

 

IMPORTANT  Do not configure restart settings in the Services management utility for services

that have been installed in a cluster; it would interfere with the cluster management software.

Resource Groups and Virtual Servers

 

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 8                                                         2002/06/10

 

 

 

 

As soon as you have installed Exchange 2000 Server on all cluster nodes, you are ready to

configure resource groups.  Each virtual server (equivalent to a resource group) requires an IP

address and a network name.

 

Each virtual Exchange 2000 server requires one or more shared disk resources where the

Information Store databases must be placed.

 

 

Exercise Summary:

 

You can configure and maintain Exchange 2000 resources in a cluster through Cluster

Administrator.  Separate physical hard disks are required for each virtual server, as well as a

unique IP address and a network name.

 

IMPORTANT  Each resource group in a cluster running Exchange 2000 Server must contain

an IP address, network name, disk, and the Exchange SA resource.  Other resources, such as

automatically created protocol virtual servers, may be removed.

 

 

Configuring Additional Virtual Servers

 

Depending on your load-balancing and failover strategy, you might want to configure multiple

virtual Exchange 2000 server in a cluster.  Each virtual server will appear in the network as a

separate server.

 

 

Public Information Store Dependencies

 

When configuring additional virtual servers, keep in mind that you cannot configure more than

one public store on the cluster.  After adding additional virtual servers, you must delete the public

tore in new groups before bringing the resource groups online.

 

Virtual Protocol Servers

 

As soon as you have configured a virtual Exchange 2000 server, you can use the Exchange

System Manager to configure additional servers for Internet access protocols.  (see page 241

bottom for the steps).

 

 

=======================================================================

 

winexc7.html                                                   PAGE 9                                                         2002/06/10

 

 

 

Full-Text Indexing and Virtual Servers

 

Exchange 2000 Server supports full-text indexing in active/active cluster configurations. 

Consequently, the Exchange MS Search Instance resource is added to your virtual server(s)

when you add the Exchange SA resource to your resource group.  Even if you don’t plan to

use the full-text indexing feature of Exchange 2000 Server, do not delete the Exchange MS

Search Instance from your virtual server.

 

To enable full-text indexing on a cluster with the Exchange System manager, right-click the

desired store, and, from the shortcut menu, select Create Full-Text Index.

 

 

Exercise Summary:

 

Using Cluster Administrator, you can test the failover procedure for Exchange 2000 resource

easily.  Just right-click on an Exchange 2000 resource, and select Initiate Failure.

All resources in the cluster group will be taken offline and then brought back online on another

available node in the cluster.

The process of stopping and starting instances is performed automatically and takes only

seconds.

 

 

Upgrading Exchange Server 5.5 Clusters

 

A direct update of an Exchange Server 5.5 cluster to Exchange 2000 Server is not possible. 

However, you can install Exchange 2000 Server in an existing site and use the move-mailbox

Direct Cluster Update

 

Another option is to directly upgrade the existing information store database currently used in

 the Exchange Server 5.5 cluster, provided that the cluster itself is running Windows 2000 with

Service Pack 1.

 

The file used is PRIV.EDB and it is stored in \Program Files\Exchsrvr\MDBData directory.

 

 

Direct Upgrade Limitations

 

The direct upgrade approach requires you to install Exchange 2000 Server in the site of the

former Exchange Server 5.5

 

 

 

=======================================================================

 

winexc7.html                                                   PAGE 10                                                       2002/06/10

 

 

 

Chapter Summary:

 

network name, and disk resources.

cluster can contain multiple virtual servers.

concurrent access to physical disks from more than one node.

over from the failed node to another node in the cluster, but it is impossible to activate

the same virtual server on more than one node at the same time.

installed and properly configured first.  Each node must be rebooted to complete the

installation of the node.