This article was previously published under Q278007
This article describes some of the new features in Windows Server 2003 clusters in comparison with Microsoft Windows 2000 Advanced Server and Microsoft Windows 2000 Datacenter Server.
In the "More Information" section, you will find a list of new features in Windows Server 2003 clusters as well as additional reference material. The following section is divided into the following categories:
Configuration and administration
Support and troubleshooting
Configuration and administration
You can have 8-node clusters in Windows Server 2003, Enterprise Edition and Windows Server 2003, Datacenter Edition. For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
288778 Maximum number of supported nodes in a cluster
There is 64-bit support (IA-64) for Windows Clustering available in Windows Advanced Server and Windows Datacenter Server.
Note Windows Clustering does not support GPT disks.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
284134 Server clusters do not support GPT shared disks
Windows Clustering is installed by default; you only need to configure a cluster by starting Cluster Administrator or by scripting the configuration with Cluster.exe.
Analysis of the nodes configuration occurs prior to the installation process so that problems are easily identified. The Analysis phase warns you if either File Server for Macintosh or Network Load Balancing is installed. If the node is configured to use DHCP for any of the network interfaces or the system has any Dynamic Disk, the Analysis phase will issue a warning because these do not work properly on a server cluster.
Terminal Server can be installed in Application mode with Windows Clustering configured. However, before you do this, review the following article when you enable Terminal Server in Application mode on a Windows Server 2003 cluster:
294926 Considerations of Terminal Services on a server cluster
The ability to change the cluster role, priority, and network name has been removed from the Cluster Configuration Wizard.
The default size of the "Reset quorum log at" parameter has been increased to 4096 Kilobytes (KB).
You no longer need to restart the computer after you install or uninstall the Windows Clustering. Windows Clustering and associated drivers can be started and stopped dynamically.
If a node is not attached to a shared network directory, it automatically configures a Local Quorum resource. With this, you can test applications on a cluster without necessarily having the Cluster service hardware.
Note You can only install a Local Quorum resource on one node, and it only supports one node (lonewolf).For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
283715 Features of the local quorum resource on Windows Server 2003 clusters
The process of uninstalling Windows Clustering is more efficient; you only need to evict the node through Cluster Administrator or Cluster.exe so that the node is no longer configured for Windows Clustering support. There is also a new switch for Cluster.exe that forces the uninstall if there is difficulty accessing Cluster Administrator:
cluster node nodename /force
For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
282227 How to uninstall the Cluster service on a Windows Server 2003 cluster
Rolling upgrades are supported from Windows 2000 to Windows Server 2003 products.
Note There is no support for rolling upgrades from a Microsoft Windows NT 4.0 cluster to a Windows Server 2003 cluster. An upgrade from Windows NT 4.0 is supported, but the cluster must be taken offline during the upgrade.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
301538 How to upgrade a Microsoft Windows NT 4.0 cluster to a Windows Server 2003 cluster
You can now configure a server cluster in the Configure Your Server Wizard, which starts during the first logon procedure, thereby simplifying configuration.
You no longer need to select which disk is going to be used as the Quorum Resource; it is automatically configured on the smallest disk that is larger then 50 megabytes (MB) and that is formatted with the NTFS File System. The option to move the Quorum Resource to another disk is available during Setup or after the cluster has been configured.
You can remotely install and configure a cluster with Cluster Administrator or Cluster.exe. Refer to the Help and Support Center under "Server Clusters" for more details.
Support and troubleshooting
A new Windows Clustering recovery tool decreases the time it takes to do a disk replacement and restore registry check point files in Windows Server 2003.For additional information, click the following article numbers to view the articles in the Microsoft Knowledge Base:
305793 How to replace a disk that is on a Windows 2000 or a Windows 2003 server custer
307469 Restore registry check points stop working after you restore a server cluster
New configuration process with increased analysis displayed on the screen and to the following log file:
When disk corruption is suspected, the Cluster service reports the results of the CHKDSK command-line utility in several places. Results are logged in the Application log and the Cluster.log file. In addition, the Cluster.log file references a log file in which detailed CHKDSK output is recorded.
For example, the Cluster.log file may reference CHKDSK issues in the following way:
For additional information about CHKDSK and the output, click the following article number to view the article in the Microsoft Knowledge Base:
272244 Location of the Chkdsk results for Windows clustering resources
The Cluster service adjust more efficiently to shared disk changes in regards to size changes and drive letter assignments. The Cluster service works directly with Volume Mount Manager and no longer directly uses the DISKINFO or DISK keys. These keys are maintained for compatibility with earlier versions of Windows Clustering. If you increase the size of a shared disk, Windows Clustering can now dynamically adjust to it, which resolves several issues with regards to bringing modified disks online.
Windows Server 2003 Clustering introduces Software Tracing that will produce more information to assist in troubleshooting Windows Clustering issues. For additional information about how to use Software Tracing to troubleshoot Windows Server 2003 cluster issues, refer to article Q301636.
The Cluster log readability is improved by adding error levels (similar to Event logs) that can help identify problem areas:
Info = Informational Entries
Warn = Warning Entries
Err = Error Entries
Cluster logs now display the server's local time in a log entry when Windows Clustering starts. This feature assists you when you are comparing Event log entries to Cluster logs.
A Windows Clustering Object file (Windows_folder\Cluster\cluster.oml) is automatically created and maintained that contains a mapping of globally unique identifiers (GUIDs) to resource name mappings.
Creating a print server on a cluster is easier with Windows Server 2003 clusters. Printer drivers are now copied over to all nodes when you install a printer on the virtual server.For additional information, click the following article numbers to view the articles in the Microsoft Knowledge Base:
302539 How to troubleshoot printing issues on a Windows Server 2003 cluster
Windows Server 2003 clusters can use 32-bit cluster resources, which makes the migration of Windows 2000 clusters to 64-bit clusters more cost effective by allowing the use 32-bit applications in the cluster.
In Windows Server 2003, there is a new quorum-capable resource name "Majority Node Set" (MNS). With MNS, you can configure a multi-node cluster without using a common shared disk.
You now have the ability to create a Generic Script resource in addition to Generic Applications and Services.
With new Message Queuing triggers resource, you can have multiple Message Queuing resources on a cluster, which allows Active/Active configurations.
Note You can only have one Message Queuing resource for each cluster group.
Distributed Transaction Coordinator (DTC) installation is simplified; just create the resource and it will run the installation and configure DTC for you on all nodes.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
301600 How to configure Microsoft Distributed Transaction Coordinator on a Windows Server 2003 cluster
Volume mount points are now supported on the shared disk and can work properly during a failover.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
280297 How to configure Volume Mount Points on a clustered server
Windows Clustering has been optimized for Storage Area Networks (SAN). The Windows Clustering device driver has been modified to do targeted device resets when arbitrating for a disk versus an entire system bus reset.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
301647 Cluster service improvements for storage area networks
Client Side Caching (CSC) is now supported for clustered File Share resources.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
285369 An update to configure client-side caching on a Windows 2000 server cluster is available
With a Windows Server 2003 cluster, you can create multiple stand-alone DFS roots. The following capabilities and features also exist:
A cluster can export multiple namespaces
Finer granularity of failover (faster failover time)
Each root can failover independently
Roots can be in different virtual servers
Active/Active configurations are supported
Windows Server 2003 is now Active Directory-aware. With Windows Server 2003:
You can publish a Network Name as a computer object in Active Directory.
Kerberos authentication for virtual servers is provided and a default location for services (such as Message Queuing) to publish service control points.
You can run Active Directory-aware applications as a clustered application.
Windows Clustering integration does not make any changes to the schema.
To have Kerberos authentication for the virtual server in a mixed mode cluster you must be running Windows 2000 Advance Server Service Pack 3 or higher; otherwise, NTLM is used for authentication.
For additional information about Windows Clustering and Kerberos authentication, click the following article numbers to view the articles in the Microsoft Knowledge Base:
302389 Description of the properties of the cluster Network Name resource in Windows Server 2003
307532 How to troubleshoot the Cluster service account when it modifies computer objects
Windows Server 2003 clusters offer enhanced cluster network integration with configuration. If network connectivity is lost, the TCP/IP stack is not unloaded, which occurred in Windows 2000 by default. With Windows Server 2003, you no longer have to set the DisableDHCPMediaSense registry key.
The internal communications (Heartbeat) in the Windows Server 2003 cluster is now on Multicast to reduce network traffic, which enhances support for server clusters with more than two nodes. If multicast communication fails for any reason, the internal communications revert to unicast. All internal communications are signed and secure.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
307962 Multicast support enabled for the cluster heartbeat
Windows Server 2003 provides enhanced logic for failover when there has been a complete loss of internal communication; the network state for public communication of all nodes is now taken into account.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
286342 Network failure detection and recovery in Windows Server 2003 clusters
The extension of a partition into free space for the shared disk is more streamlined with the included Diskpart utility.For additional information about extending a shared disk, click the following article number to view the article in the Microsoft Knowledge Base:
304736 How to extend the partition of a cluster shared disk
When you use the move group command on a cluster with more than two nodes, you can choose the Best Possible option. There is also enhanced logic for dealing with failures of nodes or resources as to which node it should fail resources over to.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
299631 Failover behavior on clusters of three or more nodes
You do not have to take the cluster nodes offline to change the password for the Cluster service account.For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
305813 How to change the Cluster service account password
You can delete resources in Cluster Administrator or with Cluster.exe without first taking them offline. The Cluster service takes them offline automatically and then deletes them.
Windows Management Instrumentation (WMI) scripting support is now provided. WMI, with regards to clustering, can:
Script ability for management operations.
Obtain access to status information for the monitoring of clusters.
Newly increased Backup and Restore functionality exists. You can actively restore the cluster configuration of the local cluster nodes, or you can restore the cluster information to all nodes in the cluster. A node restoration is also built into Automatic System Recovery (ASR). ASR restores the disk signatures for all disks (including shared), rebuilds the nodes operating system, and restores the local cluster registry.
Drive letter changes and Physical Disk resource failover is updated to the terminal server client's sessions.
The cluster service queues up changes that need to be completed if a node is offline. For example, if a node is offline and is evicted from the cluster by a remaining node, the cluster service is uninstalled the next time the evicted node attempts to join the cluster. This also holds true for applications as well. DTC automatically configures the DTC resource on an offline or newly added node to a cluster when it joins the cluster.