Clustered Server Environment [TN]


	Applies To

	Product(s):	ProjectWise Integration Server
	Version(s):	All
	Environment:	N/A
	Area:	N/A
	Subarea:	N/A
	Original Author:	Bentley Technical Support Group

Overview

ProjectWise Integration Server supports the Cluster Service and Network Load Balancing features of the Windows 2003 Server operating system. This section discusses the basic configuration of using ProjectWise Integration Servers in a clustered environment. It is recommended that you use Bentley Professional Services if you want to use this in production.

Note: ProjectWise does not include cluster support for ProjectWise User Synchronization Server, or the Automated File Processing features delivered with ProjectWise Integration Server.

Note: For cluster-like functionality in SharePoint, you can set up a web farm. Visit Microsoft's website for more information.

Setting up your server cluster

Before installing any ProjectWise software, you must first set up your server cluster, which basically means you have to create a cluster, add it to your domain, and then specify and set up each computer that will be a part of the cluster. See Microsoft documentation for details on this configuration.
To connect ProjectWise clients to the ProjectWise Integration Server, Use the ProjectWise network control panel to point to the name of the cluster, rather than the computer on which a ProjectWise Integration Server is installed The cluster will then decide which Integration Server is to handle requests from the client.
Once your basic cluster of servers is configured, the next step is to set up a database to host the ProjectWise database. Follow your normal database setup procedures, noting that the database server must exist on a computer that is NOT part of the cluster. Once the database is set up, on each Integration Server computer in the cluster you must create an ODBC datasource that points to the database. Each of these ODBC datasources must have the same name.
Next, it is recommended when using servers in a cluster, that you also use a storage area network (SAN), to host the ProjectWise storage area. When you create the datasource, the host node you specify will still be the Integration Server (or Caching Server, if you decide to use one), but the path to where the physical files exist will point to the SAN. Note, if you do use Caching Servers in your overall ProjectWise configuration, the Caching Servers you are using must exist on computers that are NOT part of the cluster.
Remember, when setting up your cluster, the following must exist on computers outside the cluster

- the ProjectWise database
- the storage area network (SAN)
- ProjectWise Caching Server
- ProjectWise Web Server Components for Web Parts

Installing ProjectWise Integration Servers and configuring the datasource

Once all of your pre-configuration is complete, the next step is to install ProjectWise Integration Server on each computer in the cluster. All installation paths and other details must be the same on each computer. After installation, you will create a datasource on one computer, modify the dmskrnl.cfg file accordingly, and then copy that file to the rest of the Integration Servers in the cluster.

To configure a datasource to be used in a clustered environment

Install ProjectWise Integration Server on each computer in the cluster.
On all but one of these computers, stop the services for the Integration Servers.
On the computer with the service still running, install ProjectWise Administrator and use it to create your datasource.
When creating your datasource, the storage area node will default to the Integration Server computer on which you are working. In order for all servers in the cluster to access the files the Storage Area will need to be moved to dedicated Caching Server or hosted on shared storage such as a SAN or NAS. When hosting on shared storage, the Storage Properties for Storage Area show have the 'Computer Name' pointing at the cluster name and the 'Path' would point to the shared storage location. EG Computer Name set to mycluster.mydomain.com. Path set to \\sharedstorage\pw_storage_area
After you have created your datasource, stop the service for that Integration Server.
Open that Integration Server's dmskrnl.cfg file.
Find the ServerName and ServerIPAddress properties and make sure they are uncommented and are pointing to the cluster's name and IP address, respectively. For example:

[TeamMate]

...

ServerName=Cluster name (EG mycluster.mydomain.com)
ServerIPAddress=Cluster IP address (EG 192.168.1.10)
Find the FileAcccessTokenMasterKey and LoginTokenMasterKey properties under the TeamMate section, and delete everything after the = for each property, so that they now look like this:

[TeamMate]

...

FileAccessTokenMasterKey=

...

LoginTokenMasterKey=
For V8i, Pre SS1 add the additional information under the Name Resolution section
For V8i SS1 and later add the additional information to the Server Name Resolution section:

Cluster IP address = Cluster name (EG 192.168.1.10=mycluster.mydomain.com)

This is a change from XM where you did not need the Name Resolution entry.
Save and close the dmskrnl.cfg file.
Restart the service for that Integration Server.
Replace the old dmskrnl.cfg file on the rest of the Integration Server computers in the cluster with a copy of the newly modified dmskrnl.cfg file.
Start the services for the rest of the Integration Servers in the cluster.