Stress / long running test of cluster, see #2786

* akka.cluster.StressSpec
* Configurable number of nodes and duration for each step
* Report metrics and phi periodically to see progress
* Configurable payload size
* Test of various join and remove scenarios
* Test of watch
* Exercise supervision
* Report cluster stats
* Test with many actors in tree structure

Apart from the test this commit also solves some issues:

* Avoid adding back members when downed in ClusterHeartbeatSender
* Avoid duplicate close of ClusterReadView
* Add back the publish of AddressTerminated when MemberDowned/Removed
  it was lost in merge of "publish on convergence", see #2779

akka-cluster/src/multi-jvm/scala/akka/cluster/MultiNodeClusterSpec.scala

@@ -218,17 +218,17 @@ trait MultiNodeClusterSpec extends Suite with STMultiNodeSpec { self: MultiNodeS
    */
   def awaitUpConvergence(
     numberOfMembers: Int,
-    canNotBePartOfMemberRing: immutable.Seq[Address] = Nil,
+    canNotBePartOfMemberRing: Set[Address] = Set.empty,
     timeout: FiniteDuration = 20.seconds): Unit = {
     within(timeout) {
+      if (!canNotBePartOfMemberRing.isEmpty) // don't run this on an empty set
+        awaitCond(
+          canNotBePartOfMemberRing forall (address ⇒ !(clusterView.members exists (_.address == address))))
       awaitCond(clusterView.members.size == numberOfMembers)
       awaitCond(clusterView.members.forall(_.status == MemberStatus.Up))
       // clusterView.leader is updated by LeaderChanged, await that to be updated also
       val expectedLeader = clusterView.members.headOption.map(_.address)
       awaitCond(clusterView.leader == expectedLeader)
-      if (!canNotBePartOfMemberRing.isEmpty) // don't run this on an empty set
-        awaitCond(
-          canNotBePartOfMemberRing forall (address ⇒ !(clusterView.members exists (_.address == address))))
     }
   }