Replication and Clustering in IIS

Important The clustering features described in this section are available only with Windows 2000 Advanced Server.

Windows Clustering provides two clustering technologies that can be used with IIS: Network Load Balancing clusters and Server clusters. Network Load Balancing clusters provide high scalability and availability of up to thirty-two servers. Server clusters provide high availability through fail-over clustering of two connected servers. Depending on your needs and available resources, you can set up IIS in a multiple-node Network Load Balancing cluster (for high availability and scalability) or a two-node Server cluster (for high availability only).

This documentation describes the concepts of both Network Load Balancing clusters and Server clusters, but provides only procedural information on setting up IIS in a two-node Server cluster configuration. For more information on Network Load Balancing, Server clusters, and Windows Clustering in general, see the Windows 2000 Advanced Server documentation.

Internet Information Services (IIS) is designed to integrate with the Cluster service feature of Windows 2000 Advanced Server and the Microsoft Content Replication System (CRS) feature of Microsoft Site Server. IIS also provides a command-line utility (Iissync.exe) for replicating the IIS metabase and other configuration settings (such as applications) from one server node to others. This utility provides a manual method for duplication of these settings.

Clustering Web or FTP Services

The following information describes how you can set up IIS in a Server cluster. For detailed information about Server clusters, including how to use Cluster Administrator, see the Windows 2000 Advanced Server documentation.

Every clustered IIS Web or FTP site consists of:

An IIS cluster resource.
An IP address cluster resource, on which the IIS Web or FTP site resource depends.

To set up IIS in a Server cluster

In Cluster Administrator, select the group into which you want to add the new resource. A group is the unit of failover in a clustered environment. For the first IIS resource this is typically the Cluster Group. For any other IIS resources this is typically a new group. Selecting a new group will result in the resources in that new group failing over independently.
On the File menu, point to New and then Resource. In the dialog box, type the name and description for the new resource and select the resource type IIS Server Instance from the drop-down list box. Click Next.
From the selection field, select the nodes in the cluster on which you want the resource to be available. To provide failover, both nodes should be selected; this is the default setting. Click Next.
In the navigation pane, select a dependency for the new resource. IIS cluster resources require at least one cluster IP address dependency, such as the Cluster IP Address. Click Add to select dependencies. Click Next.

Notes

The IP address dependency is required in order for failover to succeed. The IIS Server Instance must have an IP address dependency so that when IIS fails over to the second node in a cluster, the IP address that is needed to bind to the IIS Server Instance will also fail over to the second node.
If your content is stored on a clustered disk in your cluster, make sure that you have configured the IIS Server Instance as being dependent on the clustered disk. If you do not, it is possible that the clustered IIS server could come online on a node that doesn't have access to the clustered disk and IIS would not be able to access the content after a failover.
Make sure that all anonymous user names and passwords used in the Web site configurations are usable on all nodes of the cluster. All virtual directory paths should either point to a shared drive (that is, a UNC or cluster hard disk) or to identical local disks (that is, the same drive letters and directory structures on all nodes in the cluster).

Select either FTP or WWW. Select a server from the drop-down list. Click OK.

If your Web or FTP sites are bound to an IP address, use the Cluster IP Address resource and make your IIS Cluster resource dependent on that Cluster IP Address resource.

Important You must use the IIS snap-in to set Web or FTP site properties. You can also start and stop Web and FTP sites on a non-clustered computer by using the IIS snap-in, but you must use Cluster Administrator to start and stop Web or FTP sites on a clustered computer.

Replicating Configuration Settings

To replicate configuration settings from one Web or FTP server to another

On the source (master) server, at the command prompt, navigate to the utility's folder, %SystemRoot%\system32\inetsrv.
Replicate the configuration settings by typing: iissync destination, where destination is the computer name of the server on which you want to replicate configuration settings.
To replicate configuration settings from the master computer to several other computers simultaneously, list the names, separated by spaces, after the iissync command. For example iissync clusternode_2 clusternode_3 ... and so on.

Notes

You must run this utility as needed in order to maintain a reliable replication of configuration settings across both server nodes. This utility does not automatically synchronize configuration settings or recognize configuration changes when they occur.
If the destination server already contains an IP address bound to a host name and port, that IP address is not changed when the configuration is copied.
Clusters must already exist for the Iissync.exe utility to function. If the computer name used is incorrect, an error is returned and replication does not take place.

Uninstalling Cluster service

All IIS resources must be removed from a cluster before Cluster service is uninstalled. If IIS resources are not removed, you will be unable to stop or start the previously clustered IIS sites.

To do this, type the following at the command prompt for all previously clustered IIS resources:

adsutil set <service name>/<instance id>/ClusterEnabled 0