When should I use Cluster Sharding Proxy and when Cluster Client?
I have a backend with sharded actors and HTTP frontend which provide
interface for them.
The operations mostly fire-and-forget semantics, so HTTP frontend does not
return any data to user, but just proxy notifications to sharded
I have a Akka Cluster application with two roles: backend and frontend.
Backend starts sharded actors and do the job. Frontend provides HTTP
interface for Backend using shard region proxy.
On backend I start shard region as follows:
ClusterSharding(system).start(
typeName = UserRegion.shardNa
I have a Akka Cluster application with two roles: backend and frontend.
Backend starts sharded actors and do the job. Frontend provides HTTP
interface for Backend using shard region proxy.
On backend I start shard region as follows:
ClusterSharding(system).start(
typeName = UserRegion.shardNa
My not clustered version of actor was:
class User(userId: String) extends PersistentActor {
override val persistenceId = userId
// ...
}
Now I decided to use Cluster Sharding so I could not pass to constructor any
userId at runtime. It looks that I should use idResolver for that purpose,
a stable module in 2.4 (which will get its first milestone
> *very soon*).
>
> On Mon, May 18, 2015 at 9:11 AM, Artyom Bolshakov > wrote:
>
>> I have the following requirements:
>>
>> A bunch of users communicate with my app via http api and sends video
>> w
I have the following requirements:
A bunch of users communicate with my app via http api and sends video watch
status (progress). Any time I should be able to determine how mach users
watching the particular video right now and how many video is watching by
particular users.
So I should maint