Cluster Client
An actor system that is not part of the cluster can communicate with actors
somewhere in the cluster via this ClusterClient
. The client can of course be part of
another cluster. It only needs to know the location of one (or more) nodes to use as initial
contact points. It will establish a connection to a ClusterReceptionist
somewhere in
the cluster. It will monitor the connection to the receptionist and establish a new
connection if the link goes down. When looking for a new receptionist it uses fresh
contact points retrieved from previous establishment, or periodically refreshed contacts,
i.e. not necessarily the initial contact points. Also, note it's necessary to change
akka.actor.provider
from akka.actor.LocalActorRefProvider
to
akka.remote.RemoteActorRefProvider
or akka.cluster.ClusterActorRefProvider
when using
the cluster client.
The receptionist is supposed to be started on all nodes, or all nodes with specified role,
in the cluster. The receptionist can be started with the ClusterReceptionistExtension
or as an ordinary actor.
You can send messages via the ClusterClient
to any actor in the cluster that is registered
in the DistributedPubSubMediator
used by the ClusterReceptionist
.
The ClusterReceptionistExtension
provides methods for registration of actors that
should be reachable from the client. Messages are wrapped in ClusterClient.Send
,
ClusterClient.SendToAll
or ClusterClient.Publish
.
1. ClusterClient.Send
The message will be delivered to one recipient with a matching path, if any such exists. If several entries match the path the message will be delivered to one random destination. The sender() of the message can specify that local affinity is preferred, i.e. the message is sent to an actor in the same local actor system as the used receptionist actor, if any such exists, otherwise random to any other matching entry.
2. ClusterClient.SendToAll
The message will be delivered to all recipients with a matching path.
3. ClusterClient.Publish
The message will be delivered to all recipients Actors that have been registered as subscribers to the named topic.
Response messages from the destination actor are tunneled via the receptionist
to avoid inbound connections from other cluster nodes to the client, i.e.
the sender()
, as seen by the destination actor, is not the client itself.
The sender()
of the response messages, as seen by the client, is preserved
as the original sender(), so the client can choose to send subsequent messages
directly to the actor in the cluster.
While establishing a connection to a receptionist the ClusterClient
will buffer
messages and send them when the connection is established. If the buffer is full
the ClusterClient
will throw akka.actor.StashOverflowException
, which can be
handled in by the supervision strategy of the parent actor. The size of the buffer
can be configured by the following stash-capacity
setting of the mailbox that is
used by the ClusterClient
actor.
akka.contrib.cluster.client {
# How often failure detection heartbeat messages should be sent
heartbeat-interval = 2s
# Number of potentially lost/delayed heartbeats that will be
# accepted before considering it to be an anomaly.
# The ClusterClient is using the akka.remote.DeadlineFailureDetector, which
# will trigger if there are no heartbeats within the duration
# heartbeat-interval + acceptable-heartbeat-pause, i.e. 15 seconds with
# the default settings.
acceptable-heartbeat-pause = 15s
mailbox {
mailbox-type = "akka.dispatch.UnboundedDequeBasedMailbox"
stash-capacity = 1000
}
}
An Example
On the cluster nodes first start the receptionist. Note, it is recommended to load the extension
when the actor system is started by defining it in the akka.extensions
configuration property:
akka.extensions = ["akka.contrib.pattern.ClusterReceptionistExtension"]
Next, register the actors that should be available for the client.
runOn(host1) {
val serviceA = system.actorOf(Props[Service], "serviceA")
ClusterReceptionistExtension(system).registerService(serviceA)
}
runOn(host2, host3) {
val serviceB = system.actorOf(Props[Service], "serviceB")
ClusterReceptionistExtension(system).registerService(serviceB)
}
On the client you create the ClusterClient
actor and use it as a gateway for sending
messages to the actors identified by their path (without address information) somewhere
in the cluster.
runOn(client) {
val c = system.actorOf(ClusterClient.props(initialContacts), "client")
c ! ClusterClient.Send("/user/serviceA", "hello", localAffinity = true)
c ! ClusterClient.SendToAll("/user/serviceB", "hi")
}
The initialContacts
parameter is a Set[ActorSelection]
, which can be created like this:
val initialContacts = Set(
system.actorSelection("akka.tcp://OtherSys@host1:2552/user/receptionist"),
system.actorSelection("akka.tcp://OtherSys@host2:2552/user/receptionist"))
You will probably define the address information of the initial contact points in configuration or system property.
A more comprehensive sample is available in the Typesafe Activator tutorial named Distributed workers with Akka and Scala! and Distributed workers with Akka and Java!.
ClusterReceptionistExtension
In the example above the receptionist is started and accessed with the akka.contrib.pattern.ClusterReceptionistExtension
.
That is convenient and perfectly fine in most cases, but it can be good to know that it is possible to
start the akka.contrib.pattern.ClusterReceptionist
actor as an ordinary actor and you can have several
different receptionists at the same time, serving different types of clients.
The ClusterReceptionistExtension
can be configured with the following properties:
# Settings for the ClusterReceptionistExtension
akka.contrib.cluster.receptionist {
# Actor name of the ClusterReceptionist actor, /user/receptionist
name = receptionist
# Start the receptionist on members tagged with this role.
# All members are used if undefined or empty.
role = ""
# The receptionist will send this number of contact points to the client
number-of-contacts = 3
# The actor that tunnel response messages to the client will be stopped
# after this time of inactivity.
response-tunnel-receive-timeout = 30s
}
Note that the ClusterReceptionistExtension
uses the DistributedPubSubExtension
, which is described
in Distributed Publish Subscribe in Cluster.
It is recommended to load the extension when the actor system is started by defining it in the
akka.extensions
configuration property:
akka.extensions = ["akka.contrib.pattern.ClusterReceptionistExtension"]
Contents