* Replace (deprecate) akka.cluster.auto-down config setting with
akka.cluster.auto-down-unreachable-after
* AutoDown actor that keeps track of unreachable members
and performs down from the leader node when they have been
unreachable for the specified duration
* Migration guide
* This can't go into 2.2.x since ScalaTest 1.9.2-SNAP2 has source incompatible changes and the dependecy in akka-multi-node-node-testkit would force people to upgrade.
- add “mailbox-requirement” key to dispatcher section
- split out mailbox section, add akka.actor.default-mailbox
- rewrite findMarker method and use it for Props.create() and getting
the required mailbox of an actor
- add ProducesMessageQueue trait for MailboxType so that requirements
can be checked before trying to create the actor for real
- verify actor as well as dispatcher requirements for message queue
before creation, even in remote-deployed case
- change MessageDispatcher constructor to take a Configurator, add that
to migration guide
* The problem was:
- first is leaving, second is new oldest
- two actors subscribe to cluster events, OldestChangedBuffer and ClusterSingletonManager
- ClusterSingletonManager receives MemberExited(first), and then also MemberRemoved(second)
before OldestChangedBuffer receives MemberExited(first) and delivers
OldestChanged(first->second)
- MemberRemoved(second) is the result of the cluster extension shutdown
- because ClusterSingletonManager gets the MemberRemoved(second) before the OldestChanged
it will not send the hand over data to second
- second becomes new singleton after retry period, as designed, but without hand over data
* The solution is to check the selfExited flag in Oldest state, similar to what is done
in WasOldest
* I considered the alternative to tunnel all member events through same subscriber,
but that would involve more changes to the code
if the message should be sent to a matching path on the sender node or not.
Added optional property (skipSenderNode) to PubSub.SendToAll.
Deciding if the SendToAll message should be sent to a matching path on the sender cluster node or not. + Test and Docs.
* Removed leader commands for Shutdown and Exit
* Member shutdown itself when it sees itself as Exiting
* Singleton cluster with status Exiting will shutdown itself,
in case the Exiting gossip never arrives
* Exiting member not part convergence check
* Exiting member is removed by leader (on convergence) when the
exiting member is in the unreachable set, i.e. sucessfully shutdown
* Reverted the change made for #3266, i.e. Exiting is
detected as unreachable again.
* Adjust ClusterSingletonManager to new Exiting behaviour
* Fix bug in HeartbeatSender, which caused it to continue to
send heartbeats to removed nodes, instead of rebalancing
* Refactoring of leaderActions method
* Leaving section in docs
* Assign internal upNumber when member is moved to Up
* Public API Member.isOlder
* Change cluster singleton to use oldest member instead of leader
* Update samples and docs
* Message wrappers in ClusterClient, so that DistributedPubSubMediator
is not leaking to the client api
* Register methods in ClusterReceptionistExtension, for convenience and
clarity
