I'd recommend Lightbend Proffesional Support for such design consulting. If
you have found an issue or bottleneck in Akka you are welcome to continue
the discussion here or in the issue tracker with a reproducer of the
problem.
Thanks,
Patrik
On Wed, Jun 14, 2017 at 8:37 PM, Mainak Ghosh
Hello Patrik,
Any thoughts on this. We can setup a chat some time if that helps.
Mainak
On Thu, Jun 8, 2017 at 1:12 PM, Mainak Ghosh wrote:
> Hello Patrik,
>
> I am trying to explore the tradeoffs of instantiating a new routee when
> there is overhead of state
Hello Patrik,
I am trying to explore the tradeoffs of instantiating a new routee when
there is overhead of state reconciliation. In my application, routees
themselves can be CPU intensive, and as a result distributing the load
across multiple machine improves performance (especially when we have
I think resizer is only for local router. Also, I don't understand why
adding more actors should improve performance if the bottleneck is outside
of these actors. Have you tried by just using the local replicator from
each node without any routers. If not, please explain why that is not a
good fit
Hello Patrik and Johan,
Thanks for your reply. The goal of my work is to test the performance of
resizer block in a cluster aware situation where the routees are
replicators who are trying to reconcile state among themselves based on
application's consistency requirement. Can you please shed some
Note that the intention is to use the local `Replicator` so there is not
much point in delegating the message with a router to remote node, unless
you use nodes with different roles, such as "frontend" and "backend" and
only run distributed data on the "backend" nodes and want to send the
messages
The error message says that you have not configured serialization for the
message sample.distributeddata.ReplicatedCache.PutInCache which something
tries to send between nodes.
--
Johan
Akka Team
On Tue, May 30, 2017 at 10:37 PM, Mainak Ghosh
wrote:
> Hello,
>
> I am
Hello,
I am recently playing around with cluster aware routers and distributed
data and I have a few questions. I extended the multi jvm test code
ReplicatedCacheSpec to create a cluster aware router which creates a pool
of ReplicatedCache worker routees. Here is the
code: