Clustering and Load Balancing with Exchange 2003 Server

 <  Day Day Up  >  

Supported only with the Microsoft Windows 2003 Enterprise Server and DataCenter Editions, Windows 2003 fault-tolerant options such as clustering and load balancing provide Exchange mailbox clients with high availability and performance when accessing Exchange 2003 messaging systems.

This section provides information to assist administrators in understanding the functionality of Exchange 2003 in Windows 2003 clusters and NLB, environment features available with each, and best practices for determining which solution best fits your Exchange 2003 server needs.

Clustering Terminology

Before deciding which fault-tolerant option is best, here are a few terms used to describe the way clustered Exchange 2003 servers operate :

  • Active/Active In Active/Active clustering, all servers in the cluster are servicing clients at the same time. Active/Active is also referred to as shared everything architecture .

  • Active/Passive In Active/Passive clustering, only one server in the cluster can service end-users per application while the other server(s) wait until a failure occurs before they begin to service clients. Active/Passive clustering is also referred to as shared nothing architecture .

  • Exchange Virtual Server An Exchange virtual server is the resource group that contains all resources for Exchange to operate on the cluster. This includes the Exchange Services, IP Address, and NetBIOS name of the server. To enable Active/Active clustering on a 2-node cluster, one Exchange virtual server is created per node in the cluster. If only one Exchange virtual server was installed then the server would be operating in Active/Passive mode. The NetBIOS name and IP address of the cluster form the virtual server. When failover occurs, the entire Exchange virtual server fails over to the surviving node in the cluster.

  • Resource DLL or Exres.dll Exres.dll is the gateway that handles the communication from the cluster service to the Exchange services. It is responsible for reporting failures and bringing resources online and offline.

  • Heartbeat A single UDP packet sent every 500 milliseconds between nodes in the cluster across the private network relays information on the health of the nodes in the cluster and on the health of the application. If there is no response, the cluster will begin a failover.

  • Failover Failover is the process of a node in the cluster taking over responsibility for the applications on another node.

  • Failback Failback is the process of moving applications that failed over to another node in the cluster back to the original node.

  • Quorum Resource This is the shared disk that holds the cluster server's configuration information. All servers must be able to contact the quorum resource to join the cluster.

  • Resource Group A resource group is a collection of Cluster Resources. A resource group defines which items fail over to a surviving server during failover. This includes Cluster Resource items such as IP address and Net BIOS name. A resource group is owned by only one node in the cluster at a time. The resource group Cluster Group is the first resource group created on the server and contains the IP Address and Net BIOS name of the cluster.

  • Cluster Resource Cluster Resources are IP Addresses, Net BIOS name, disks, or services ”such as the Exchange System Attendant. Cluster Resources are added to Cluster Groups to form Exchange virtual servers.

  • Dependency Dependencies are specified when creating the Cluster Resources. A Cluster Resource that is specified as a dependency means that before a cluster Resource is brought online, whatever resources it's dependent on must be brought online first. For instance, a Net BIOS name is dependent on the IP Address; therefore, the IP Address must be brought online before the Net BIOS name can be brought online.

Fault-Tolerance Options

MSCS and NLB services support Exchange 2003 messaging by establishing QOS when supporting server availability with back-end Exchange 2003 servers and high performance when supporting Outlook Web Access front-end systems. For these fault-tolerant clustering technologies to be truly effective, administrators must understand the characteristics of each in order to choose which technology and configuration best fits their Exchange 2003 Server design and messaging application service needs.

Microsoft Cluster Service

The Microsoft Windows 2003 Cluster Service (MSCS) is a clustering technology that provides system-level fault tolerance by using failover. In Microsoft Exchange 2003 Server environments, cluster services are best used to provide access to mailbox information, Exchange 2003 resources such as public folders, and add-on applications such as virus wall systems and Internet mail bridgehead servers.

When a problem is encountered with a Cluster Resource, the cluster service attempts to fix the problem before failing it completely. The cluster node running the failing resource attempts to restart the resource on the same node first. If the resource cannot be restarted, the cluster fails the resource, takes the cluster group offline, and moves it to another available node, where it can then be restarted.

Several conditions can cause a cluster group to fail over. Failover can occur when an active node in the cluster loses power or network connectivity or suffers a hardware failure. When a Cluster Resource cannot remain available on an active node, the resource's group is moved to an available node, where it can then be restarted.

NOTE

In most cases, client communications can be affected by the failover process causing a short disruption in client/server communications. In some cases there may be no disruption.


Cluster nodes can monitor the status of resources running on their local system, and they can also keep track of other nodes in the cluster through private network communication messages (heartbeats). The heartbeats are used to determine the status of a node and send updates of cluster configuration changes to the cluster quorum resource.

The quorum resource contains the cluster configuration data necessary to restore a cluster to a working state. Each node in the cluster needs to have access to the quorum resource; otherwise , it will not be able to participate in the cluster. Windows Server 2003 provides three types of quorum resources, one for each cluster configuration model.

To avoid unwanted failover, power management should be disabled on each of the cluster nodes, both in the motherboard BIOS and in the Power applet in the operating system's Control Panel. Power settings that enable a monitor to shut off are okay, but the administrator must make sure that the disks are configured to never go into standby mode.

Network Load Balancing (NLB)

The second clustering technology provided with the Windows Server 2003 Enterprise and DataCenter server platforms is Network Load Balancing (NLB). )NLB clusters provide high network performance and availability by balancing client requests across several server systems. When the client load increases , NLB clusters can easily be scaled out by adding more nodes to the cluster, to maintain or provide better response time to client requests.

Network load balancing offers two great features : No proprietary hardware is needed to use load balancing, and an NLB cluster can be implemented and configured fairly easily and quickly.

NLB clusters are most effectively used to provide front-end support for Web applications. In addition they are very effective solutions when used for application functionality, such as Outlook Web Access and Terminal servers.

NLB clusters can grow to 32 nodes, and if larger cluster farms are necessary, the Microsoft Application Center server should be considered as an option for server platform support, along with technologies such as Domain Name System round- robin to meet this larger demand.

NOTE

When NLB clusters are configured, each server's configuration must be updated independently. The NLB administrator is responsible for making sure that configuration and information changes on each Exchange 2003 server are kept consistent across each node.


For more information on installing load balancing with Exchange 2003 Server, refer to the section "Configuring Network Load Balancing with OWA," later in this chapter.

Cluster Permissions with Exchange 2003 Environments

Unlike previous versions of Exchange server in clustered environments, Exchange Server 2003 does not require the cluster server account to be a Full Exchange Administrator at the Exchange organization or administrative group levels. With Exchange 2003, the account logon permissions are used to establish rights for the Cluster Administrator to manage the Exchange 2003 cluster configurations. Using logon permission, the Cluster Administrator can create, delete, or modify Exchange virtual server configuration, depending on the level in which the permissions are applied and the mode (Mixed or Native) in which clustered Exchange 2003 organization is running.

When working with Exchange 2003 Mixed and Native Mode cluster environments and multidomain models, some common best practices can be applied to help administrators understand how to set permissions to manage the Exchange cluster.

With Windows 2003 multidomain models and child domain configurations, Cluster Administrator accounts require a minimum of Administrator Only permissions at the Exchange organizational level. This permission configuration is required to configure and apply Recipient Policies and configure the responsible server for providing the organization with the proper Recipient Update Services.

In addition, when the Exchange 2003 organization is in Native Mode, if Exchange virtual servers are configured in routing groups and span multiple administrative groups, the Cluster Administrator must have Exchange Full Administrator permissions. This allows the account permissions to manage each administrative group and its Exchange virtual server members .

Management Options with Exchange Server 2003

When running Microsoft Exchange Server 2003 in clustered environments, administrators must implement clusters and load balancing technologies through Windows Server 2003. Microsoft Exchange Server 2003 does not provide any clustering technologies and management options for administering load balancing technologies.

Managing Clustering with Exchange

There are no options for managing clusters and load balancing with Exchange Server 2003. Administrators can leverage the Exchange Server 2003 Windows Management Instrumentation (WMI) to monitor the state of the Exchange Server 2003 Cluster Resources.


Administrators can manage and configure clusters and load balancing options through the Windows 2003 Cluster Administrator snap-in. For more information regarding cluster management and configuration options, see the section "Managing Exchange 2003 Clusters," later in this chapter.

 <  Day Day Up  >  


Microsoft Exchange Server 2003 Unleashed
Microsoft Exchange Server 2003 Unleashed (2nd Edition)
ISBN: 0672328070
EAN: 2147483647
Year: 2003
Pages: 393
Authors: Rand Morimoto

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net