These are chat archives for akkadotnet/

Jul 2016
Maxim Cherednik
Jul 17 2016 15:24
Hi @Aaronontheweb, while Akka.Cluster was in beta the process of graceful node shutdown was a bit verbose and there were some error messages with socket exceptions and so on. That time you were saying that you guys will handle this and remove unnecessary exception logging. So at the moment I am trying the same thing - graceful node shutdown and I see some exceptions with error level ERROR - is this ok or it wasn't fixed? Even though cluster works fine. It's just those messages which makes me worried.
Samurai Ken
Jul 17 2016 23:50

Hey all. I am exploring this "actors" thing, and had a question... in lookgin at Orleans and Akka, I want to make sure I am not fundementally makign a conceptual error. I am working on expanding the scalability and maintainability of a social application. There is already a DocumentDB record for each user with a class / object overlay that provides some linited interfaces to interact with it, a manager to instantiate it and so on.

Obviously, we have re-implemented (badly) somethign that others have done better :)

So does it make sense to put an Actor as a sort of smart cache / internaction model but still have the authoritative data for a user live in the documentDB rather than exist soley in the Actors persistent state? Or is that sort of half comitting to this.