raboof/pekko
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions 10

=clu #17253 Improve cluster startup thread usage

Browse source 
When using a dispatcher (default or separate cluster dispatcher)
with less than 5 threads the Cluster extension initialization
could deadlock.

It was reproducable by adding a sleep before the Await of GetClusterCoreRef
in the Cluster extension constructor. The reason was that other cluster actors were
started too early and they also tried to get the Cluster extension and thereby blocking
dispatcher threads.

Note that the Cluster extension is started via ClusterActorRefProvider before
ActorSystem.apply returns.

The improvement is to start the cluster child actors lazily when the
GetClusterCoreRef is received.
This commit is contained in:
Patrik Nordwall 2015-08-27 08:58:52 +02:00
parent bc568e2a1b
commit 737a50ebf3
4 changed files with 120 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

7
project/MiMa.scala
View file

@ -545,7 +545,12 @@ object MiMa extends AutoPlugin {
ProblemFilters.exclude[MissingMethodProblem]("akka.japi.Pair.toString"),
// #17805
ProblemFilters.exclude[MissingMethodProblem]("akka.actor.ActorCell.clearActorFields")
ProblemFilters.exclude[MissingMethodProblem]("akka.actor.ActorCell.clearActorFields"),
// internal changes introduced by #17253
ProblemFilters.exclude[IncompatibleResultTypeProblem]("akka.cluster.ClusterDaemon.coreSupervisor"),
ProblemFilters.exclude[MissingMethodProblem]("akka.cluster.ClusterCoreSupervisor.publisher"),
ProblemFilters.exclude[IncompatibleResultTypeProblem]("akka.cluster.ClusterCoreSupervisor.coreDaemon")
)
}