2 Replies Latest reply on Aug 30, 2009 10:46 AM by juangon

    richfaces robustness inside cluster - instance restart

    juangon

      Hi!

      I was testing if my richfaces application (with STATE_SAVING_METHOD set to "client") would recover from an instance restart inside a cluster without the users to be realised this happened. I used a4j:keepAlive for several managed beans with request scope.

      Theorically with STATE_SAVING_METHOD in client mode, all the view state is on the client side so if I restart an instance cluster, when client clicks an executes an AJAX call all the view state should be recovered from client.

      But this didn't work for me with some components, like rich:datascroller. . If I click in a page number in a rich:datascroller inside a xhtml with a rich:datatable and a rich:datascroller, the same page is rendered without any change (page #1 of datascroller). But if I click in a commandLink to render that first page, then the page I clicked before is rendered correctly (for example, second page).
      I am using Sun JSF RI 1.2. May the reason of this is the

      com.sun.faces.enableViewStateIdRendering = false
      

      context-param?

      Perhaps, should I have to set the rich:datascroller
      ajaxSingle
      
      to true?
      Thanks for your attention!