Install in a failover cluster

Before you begin, create a failover cluster with a cluster shared volume using Windows Clustering.

Note: This document covers installing PI Analysis Service in a pre-existing failover cluster. For instructions on setting up a new Microsoft Windows Failover Cluster, see the Microsoft TechNet article: Failover Clustering Overview.

To enable high availability for PI Analysis Service, you must install the service and enable Buffering Manager in each node of a failover cluster. You can install PI Analysis Service in the same failover cluster as PI Notifications Service, or you can install the services in different failover clusters.

1. In each node of the failover cluster, install PI Analysis Service by running the PI Server install kit.

   See Install PI Analysis Service.

2. From one of the cluster nodes, configure a clustered role for PI Analysis Service:

   1. From Server Manager, click Tools > Failover Cluster Manager to open Failover Cluster Manager.

   2. Click the cluster that you created to open the summary page.

   3. In the Configure pane, click Configure Role to run the High Availability Wizard.

      The wizard prompts you for the following information:

      Page	Description
      Select Role	Select Generic Service as the role to configure for high availability.
      Select Service	Select PI Analysis Service as the service to use.
      Client Access Point	In the Name field, type a unique name to identify the cluster role for PI Analysis Service. The name can be no longer than 15 characters and must be unique within the network. For example, you might type AnalysisCluster. Later, you use this name to configure the PI AF configuration database.
      Select Storage	You do not need to specify a storage volume. The clustered role can use the shared volume.
      Replicate Registry Settings	Specify a registry key for PI Analysis Service: Click Add to open the Registry Key window. Enter SOFTWARE\PISystem\Analysis Service.

   Upon completion, the wizard configures a role for the clustered service and starts the role (that is, the service).

3. Prepare the cluster to configure the PI AF database:

   1. Stop the role for the newly configured PI Analysis Service cluster.

   2. In the cluster shared volume, create a folder to store runtime information.

   This folder should be a shared location among the cluster nodes.

   For example, in the C:\ClusterStorage\Volume1 directory, you might create the subdirectory Xyz\PIAnalysis.

4. Configure the PI AF configuration database to use the failover cluster:

   1. In PI System Explorer, click Database to select and connect to the configuration database on the PI AF server specified when installing PI Analysis Service.

   2. In the Elements pane, expand OSIsoft > PIANO > AnalysisService and select ServiceConfiguration.

   3. In the ServiceConfiguration pane, click the Attributes tab.

   4. Edit the value of the ServiceConfiguration attribute to set the correct values for the RuntimeStorageFolderPath and HostName XML elements.

      XML element	Description
      RuntimeStorageFolderPath	Enter the path to the folder that you created to store runtime information, such as C:\ClusterStorage\Volume1\Xyz\PIAnalysis.
      HostName	Enter the fully qualified domain name to the cluster role that you created for PI Analysis Service. For example, for a cluster role named AnalyisCluster in the domain MyCompany.com, enter AnalysisCluster.MyCompany.com.

      You can edit the attribute value in the Value column, or double-click the value to open the Text Visualizer window, where you can view and edit the XML elements.

   5. Check in your changes to the PI AF database.

5. Restart the role for the PI Analysis Service cluster.

6. For each node in the cluster, configure Buffering Manager. See Configure buffering for analysis outputs to PI points.