On Mon, Aug 13, 2012 at 2:50 AM, Murali Reddy <murali.re...@citrix.com> wrote: > On 10/08/12 6:50 PM, "Chip Childers" <chip.child...@sungard.com> wrote: > >> >>Murali, >> >>I really like this approach, and will be willing to step in and >>implement the AMQP plugin after the initial framework is in place. > > Great! I will let you know once framework is in place. > >> >>As for the design itself, one thing I didn't quite wrap my head >>around: is there any provision for handling events that might be >>noticed in the environment NOT being fired off on all management >>servers? Ex: a host goes offline. Is that picked up by all >>management servers in a cluster? > > Can you please elaborate this scenario a bit? Let say if all the > management servers in the cluster pick up an event like host disconnect, > do you expect subscriber on the all management server will take any > action? > > Its conscious decision to keep event bus I am proposing to be in-memory > and scope is restricted to a single management server. My intention is > keep event bus just act as event delivery mechanism, and not as IPC for > components in a management server or across management servers. I am > assuming that a subscriber can be present on all of the management servers > in cluster and subscriber which is present on the management server where > an event occurred can take action. > >
I spent more time thinking about this... And I believe you're right. It would be the responsibility of any event handling code to either (1) ensure idempotent operations in response to an event, or (2) deal with duplicate events across the management servers in some other way. -chip
