Alexander Klimetschek wrote:
On Thu, May 29, 2008 at 2:18 PM, Erik Buene <[EMAIL PROTECTED]> wrote:
But this would counteract the high availability aspect of clustering, as it
would introduce single point of failure on this node. I guess it would be
possible to set up hearbeat or some other way of telling if other nodes are
up and delegate the responsibility to take the task, but it would really
complicate the system.
You could annotate the nodes that "throw" the events with an
application / cluster node identifier so that the event listener can
tell local and remote events apart.
In most cases (but not always) this would be nt:files updated by webdav.
Do you have any idea of how I could inject the needed information to the
node there? Maybe I would need to write a custom IOHandler?
Or use some locking mechanism that only one event listener actually
handles the events.
I guess this could work, but if one instance in the cluster is busy and
can't process the event until processing is done by another instance the
node would be unlocked, how would I know that it is already processed.
Event does not have a timestamp when it was generated, and there could
be many events originating from the same node.
--
==================
Erik Buene
Senior Developer
Idium AS
