How does the policy server handle failover/failback for ODBC stores, and is the behavior configurable?
Release : ALL
Component : SITEMINDER -POLICY SERVER
By default the policy server will connect to the primary ODBC store unless it is unavailable. Regardless of whether connected to primary or secondary, a checker thread runs every 15 seconds. If the checker thread fails to connect, the node is considered down and a failover will occur. The policy server always favors the primary, so failback occurs as soon as the checker thread determines the primary is back up. The checker thread interval is not configurable and is fixed at 15 seconds.