It's resolved by implementing the domain event class tag by specifying the
runtime class as classOf[] instead of using classTag from scala.reflect
On Saturday, July 15, 2017 at 11:35:04 PM UTC+7, Hengky Sucanda wrote:
>
> Hi Guys, i am experiencing NPE when starting a persistent FSM that is
>
Akka 2.5.3
On Saturday, July 15, 2017 at 11:35:04 PM UTC+7, Hengky Sucanda wrote:
>
> Hi Guys, i am experiencing NPE when starting a persistent FSM that is
> cluster sharded:
>
> a.a.OneForOneStrategy - null
> java.lang.NullPointerException: null
> at
>
Hi Guys, i am experiencing NPE when starting a persistent FSM that is
cluster sharded:
a.a.OneForOneStrategy - null
java.lang.NullPointerException: null
at
akka.persistence.fsm.PersistentFSM$$anonfun$receiveRecover$1.applyOrElse(PersistentFSM.scala:121)
~[akka-persistence_2.11-2.5.3.jar:na]
Good to know that typed actors are going to use bounded mail boxes.
Just curious, don't people face issues in production systems because of
unbounded mail boxes? Our performance tests on the logger were showing that
once mail box accumulates lots of messages, it puts unreasonable load on
memory,
Hi Mainak,
Thanks for the background to your questions. It sounds like interesting
research.
I would not go down the path of pool routers. It's pretty difficult to
implement custom things with them and we are slowly moving away from remote
deployment (used by pools). E.g. Akka Typed will not
The current default, more than anything, is bound to be the default since
it was the default since the inception of Akka and going for a bounded one
by default now could make some users unhappy.
Having that said, we're actively looking forward to pick a bounded impl for
Akka Typed. Then people
Hi,
We had some custom requirements for building a logging library wrapper for
our project. The project includes multiple subsystems, each developing
components which are essentially actors. All modules/actors are supposed to
be using our logging interface.
After some research, we found