Protection from Failures and Expected Behaviors

Configuring High Availability for Identity Management Components 8-99 Oracle RAC database has been configured in a JDBC multi data source to protect the instances from Oracle RAC node failure. The instances in the WLS_ODS2 Managed Server on IDMHOST2 and the instances in the WLS_ODS1 Managed Server on IDMHOST1 are configured as the CLUSTER_ODS cluster. ■ A WebLogic Administration Server has been installed. Under normal operations, this is the passive Administration Server. You will make this Administration Server active if the Administration Server on IDMHOST1 becomes unavailable.

8.6.2.1.1 Starting and Stopping the Cluster In a high availability architecture, Oracle

Directory Integration Platform and Oracle Directory Services Manager are deployed on an Oracle WebLogic Cluster that has at least two servers as a part of the cluster. By default, the WebLogic Server starts, stops and monitors the applications. By default, both the Oracle Directory Integration Platform and Oracle Directory Services Manager applications leverage the high availability features of the underlying WebLogic Clusters. In case of hardware or other failures, session state is available to other cluster nodes that can resume the work of the failed node. In a high availability environment, WebLogic Node Manager is configured to monitor the WebLogic servers. In case of failure, Node Manager restarts the WebLogic Server. If Node Manager cannot restart the server, then the front-ending load balancing router detects failure of a WebLogic instance in the Cluster and routes traffic to surviving instances.

8.6.2.2 Protection from Failures and Expected Behaviors

This section discusses protection from different types of failure in an Oracle Directory Services Manager active-active cluster.

8.6.2.2.1 Process Failure In a high availability environment, the Oracle Directory

Services Manager applications are deployed on an Oracle WebLogic Server cluster comprised of at least two WebLogic instances. By default, the Oracle Directory Services Manager applications leverage the high availability features of the underlying WebLogic Clusters. In case of hardware or other failures, session state is available to other cluster nodes that can resume the work of the failed node. In addition, in a high availability environment, WebLogic Node Manager is configured to monitor the WebLogic servers. In case of failure, Node Manager restarts the WebLogic Server. If Node Manager cannot restart the server, then mod_wl_ohs, which is configured as a part of Oracle HTTP Server, routes the request to the surviving WebLogic instance. OPMN monitors the Oracle HTTP Server processes and restarts the process in case of failure. If OPMN is unable to restart the HTTP process, the front-ending load balancing router detects the failure of an Oracle HTTP Server instance and routes traffic to surviving instances. Oracle Directory Services Manager maintains a session state, but in case of failure, the session state information is not carried over to the surviving node.

8.6.2.2.2 Expected Client Application Behavior When Failure Occurs Oracle Directory

Services Manager failover is not transparent. You have to reestablish the connection during an Oracle WebLogic Server instance failover using Oracle Directory Services Manager. 8-100 Oracle Fusion Middleware High Availability Guide

8.6.2.2.3 Expected Dependency Failure Oracle Directory Services Manager requires the

WebLogic Managed Server to be available during startup. If it is not available, Oracle Directory Services Manager fails to start.

8.6.2.3 Oracle Directory Services Manager Prerequisites