Hi,
I've developed a distributed scheduler which adds adaptation capabilities
to the system. I'm able to define new assignments in a distributed manner
and the system automatically synchronize itself, launching or terminating
worker processes as defined by the assignments.

I'm having some troubles since, after the distributed reassignment, some
bolts can't receive tuples, even if on the destination node the supervisor
has read new assignments and had launched the new worker process.

Is there someone who can help me?
Thanks

Reply via email to