On 11/11/06, Piyush Purang <[EMAIL PROTECTED]> wrote:
I think we need to put in a little more thought about lifecycles first
... the states that i would love to have is
created (The application was created/initialized properly but never started)
started (The application has been asked to run)
running (The application is successfully running maybe we don't need this...)
stopped (The application was successfully started and then successfully stopped)
failed (The application failed while transiting into another state..)
Cool.
suspended (The application has been susspended because it is
undergoing maintainance work or is doing some very heavy number
crunching etc...)
Something along these lines is very important as we can have e.g., an
application filter that e.g., responds with a nice "Sorry, we're
currently out to lunch." message or redirects the caller to another
server in the cluster or....
Note also that this could interact with the hot-deployment issue.
Thanks,
John
