Re: [akka-user] Method to block Migration of Cluster Sharded Actors in a Certain State

2016-09-02 Thread Patrik Nordwall
The stop signal is that the actor is terminated, i.e. you called context.stop, so stashing is fine fre 2 sep. 2016 kl. 16:46 skrev kraythe : > So if I have an actor that while on a specific state simply stashes any > message that it doesn't handle, will it ignore the stop request until the > messa

Re: [akka-user] Method to block Migration of Cluster Sharded Actors in a Certain State

2016-09-02 Thread kraythe
So if I have an actor that while on a specific state simply stashes any message that it doesn't handle, will it ignore the stop request until the message is I stashed and processed? Although that time delay should be a max of a couple seconds it does exist. And in the meantime j don't want the a

Re: [akka-user] Method to block Migration of Cluster Sharded Actors in a Certain State

2016-08-30 Thread Patrik Nordwall
You can define a custom handOffStopMessage that your actors will receive when they are asked to stop for rebalance. You can delay the stop when receiving that message by not calling context.stop(self) until you are done. Note that if you don't stop within the handOffTimeout the rebalance will be a

[akka-user] Method to block Migration of Cluster Sharded Actors in a Certain State

2016-08-27 Thread kraythe
I have a cluster shared actor that basically has two states; IDLE and BUSY. The actor is cluster shared because I need the behavior of having only a single one of these in the cluster per user and I want all of the messages that the actor to handle to be serialized per user. The actor mostly si