Handling an unavailable stateful server

We can use the use-new-stateful-on-error stanza entry in the [junction] stanza of the WebSEAL configuration file to control how WebSEAL responds to a stateful server that becomes unavailable.

When use-new-stateful-on-error is set to "yes" and the original server becomes unavailable during a session, WebSEAL directs the next request of the user to a new replica server on the same stateful junction. If a new replica server is found on that stateful junction, and is responsive to the request, WebSEAL sets a new stateful cookie on the browser of the user. Subsequent requests during this same session are directed to this same new server.

When use-new-stateful-on-error is set to "no" and the original server becomes unavailable during a session, WebSEAL does not direct the subsequent requests of the user to a new replica server on the same stateful junction. Instead, WebSEAL returns an error and attempts to access the same server for subsequent requests by the user during this session. The "no" value is the default setting and maintains compatibility with versions of WebSEAL prior to version 6.0. For example:

To end the access attempts on an unresponsive server, the user must restart the browser, or the administrator must remove the unresponsive server.

This configuration item may be customized for a particular junction by adding the adjusted configuration item to a [junction:{junction_name}] stanza.

where {junction_name} refers to the junction point for a standard junction (including the leading / character) or the virtual host label for a virtual host junction. For example:

Parent topic: Stateful junctions