Authentication service failover automatically redirects an authentication request to a secondary server if the primary server fails because of a hardware or software problem or if the server is temporarily shut down.
An authentication context must first be created on an instance of Access Manager where the authentication service is available. If this instance of Access Manager is not available, an authentication context can then be created on a different instance of Access Manager through the authentication failover mechanism. The authentication context will check for server availability in the following order:
The authentication service URL is passed to the AuthContext API. For example:
If this API is used, it will only use the server referenced by the URL. No failover will occur even if the authentication service is available on that server.
The authentication context will check the server defined in the com.iplanet.am.server* attribute of the AMConfig.properties file.
If step 2 fails, then the authentication context queries the platform list from a server where the Naming service is available This platform list is automatically created when multiple instances of Access Manager are installed (generally, for failover purposes) sharing a one instance of Directory Server.
For example, if the platform list contains URLs for Server1, Server2 and Server3, then the authentication context will loop through Server1 , Server2 and Server3 until authentication succeeds on one of them.
The platform list may not always be obtained from the same server, as it depends on the availability of the Naming service. Furthermore, Naming service failover may occur first. Multiple Naming service URLs are specified in the com.iplanet.am.naming.url property (in AMConfing.properties ). The first available Naming service URL will be used to identify the server, which will contain the list of servers (in its platform server list) on which authentication failover will occur.