How do backing bean scopes work?

Question

I have some misunderstanding with JSF backing bean scope. I am new to JSF and now writing a simple project and all my beans mostly have session scope. But if my site will have many users that means my session will be very very big and kill my server. Some people have told me that the solution is use request scope beans. But, for example, when my page must be validated and if validation is failure show error messages and save all user input, what can I do in this situation? I am trying use component that use AJAX-request and hoped that my request bean will be not reconstructed, but this doesn't work (I am using <rich:datascroller> ).

I think I have big hole in my JSF understanding, I will be grateful if somebody explain what I must do in this situation or link me on some good article about bean scopes.

Answer 1

Scope defines the lifetime of the beans.

Request scope beans live during the servicing of one HTTP request, so thay are available both while you analyze the user's input and formulate the response page. So for simple validation and response I'd expect request-scoped beans to be what you need - the exception being perhaps if you send a redirect back to the browser and that submits a new request, then you may need ...

Session scoped beans live for the life of the user's session, ie. across several requests. Sessions might last for some time, but eventually the user logs out, or becomes quiscent and his session gets timed-out. So it doesn't matter how many users overall you have, just how many are active at once. It's pretty common to keep some session data around for each user (like at least who is, and perhaps his recently viewed stuff) so there's no fundamental reason to be worried by some data being kept. You just need to ensure you keep it tidy, don't keep the data for old pages very very long - perhaps just a "current data" bean or some such.