Thank you for those suggestions.
Are there specific example code repos showing either or both approaches?
On Sunday, October 2, 2016 at 5:59:01 PM UTC-4, loe...@posteo.de wrote:
>
> You could parameterize the (parent) actors with a function which creates
> child actors (and returns refs). For
You could parameterize the (parent) actors with a function which creates child
actors (and returns refs). For "normal mode” provide a function that creates
the actual child actors, for testing inject a TestProbe.ref.
Alternatively you can achieve the same with a protected factory method for
I have a simple akka system in scala which creates two actors of the same
type, which in turn create a pair of actors.
Those final (child) actors complete some processing, and reply back to the
orignal actors who spawned them.
The original actors kick off the processing with an ask (?)
It is expected that it detects crashed/stopped nodes as unreachable. When
unreachable have been downed and removed the cluster canmove joining nodes
to Up. I'm not sure I understand the question/problem.
Also, don't change heartbeat-interval. That will not make anything better.