raboof/pekko
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions 11

Solve wrong barrier problem, see #2583

Browse source 
* The problem was that we didn't wait for the testconductor.shutdown Future
  to complete and therefore barriers could be triggered in unexpected order.
  The reason why we didn't await, was that during shutdown the Future was
  completed with client disconnected failure. I have fixed that and added
  await to all shutdowns.
This commit is contained in:
Patrik Nordwall 2012-10-16 17:02:13 +02:00
parent b68c7a8469
commit 5e83df74e9
8 changed files with 16 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

2
akka-cluster/src/multi-jvm/scala/akka/cluster/ClientDowningNodeThatIsUnreachableSpec.scala
View file

@ -44,7 +44,7 @@ abstract class ClientDowningNodeThatIsUnreachableSpec(multiNodeConfig: ClientDow
runOn(first) {
// kill 'third' node
testConductor.shutdown(third, 0)
testConductor.shutdown(third, 0).await
markNodeAsUnavailable(thirdAddress)
// mark 'third' node as DOWN

2
akka-cluster/src/multi-jvm/scala/akka/cluster/ClusterAccrualFailureDetectorSpec.scala
View file

@ -46,7 +46,7 @@ abstract class ClusterAccrualFailureDetectorSpec
"mark node as 'unavailable' if a node in the cluster is shut down (and its heartbeats stops)" taggedAs LongRunningTest in {
runOn(first) {
testConductor.shutdown(third, 0)
testConductor.shutdown(third, 0).await
}
enterBarrier("third-shutdown")

2
akka-cluster/src/multi-jvm/scala/akka/cluster/ConvergenceSpec.scala
View file

@ -61,7 +61,7 @@ abstract class ConvergenceSpec(multiNodeConfig: ConvergenceMultiNodeConfig)
runOn(first) {
// kill 'third' node
testConductor.shutdown(third, 0)
testConductor.shutdown(third, 0).await
markNodeAsUnavailable(thirdAddress)
}

2
akka-cluster/src/multi-jvm/scala/akka/cluster/LargeClusterSpec.scala
View file

@ -316,7 +316,7 @@ abstract class LargeClusterSpec
}
runOn(firstDatacenter) {
testConductor.shutdown(secondDatacenter, 0)
testConductor.shutdown(secondDatacenter, 0).await
}
enterBarrier("second-datacenter-shutdown")

4
akka-cluster/src/multi-jvm/scala/akka/cluster/LeaderDowningNodeThatIsUnreachableSpec.scala
View file

@ -51,7 +51,7 @@ abstract class LeaderDowningNodeThatIsUnreachableSpec(multiNodeConfig: LeaderDow
val fourthAddress = address(fourth)
runOn(first) {
// kill 'fourth' node
testConductor.shutdown(fourth, 0)
testConductor.shutdown(fourth, 0).await
enterBarrier("down-fourth-node")
// mark the node as unreachable in the failure detector
@ -81,7 +81,7 @@ abstract class LeaderDowningNodeThatIsUnreachableSpec(multiNodeConfig: LeaderDow
enterBarrier("before-down-second-node")
runOn(first) {
// kill 'second' node
testConductor.shutdown(second, 0)
testConductor.shutdown(second, 0).await
enterBarrier("down-second-node")
// mark the node as unreachable in the failure detector

4
akka-cluster/src/multi-jvm/scala/akka/cluster/LeaderElectionSpec.scala
View file

@ -18,7 +18,7 @@ case class LeaderElectionMultiNodeConfig(failureDetectorPuppet: Boolean) extends
val third = role("third")
val fourth = role("fourth")
commonConfig(debugConfig(on = true).withFallback(MultiNodeClusterSpec.clusterConfig(failureDetectorPuppet)))
commonConfig(debugConfig(on = false).withFallback(MultiNodeClusterSpec.clusterConfig(failureDetectorPuppet)))
}
class LeaderElectionWithFailureDetectorPuppetMultiJvmNode1 extends LeaderElectionSpec(failureDetectorPuppet = true)
@ -70,7 +70,7 @@ abstract class LeaderElectionSpec(multiNodeConfig: LeaderElectionMultiNodeConfig
case `controller` ⇒
val leaderAddress = address(leader)
enterBarrier("before-shutdown" + n)
testConductor.shutdown(leader, 0)
testConductor.shutdown(leader, 0).await
enterBarrier("after-shutdown" + n, "after-unavailable" + n, "after-down" + n, "completed" + n)
case `leader` ⇒

2
akka-cluster/src/multi-jvm/scala/akka/cluster/SingletonClusterSpec.scala
View file

@ -61,7 +61,7 @@ abstract class SingletonClusterSpec(multiNodeConfig: SingletonClusterMultiNodeCo
"become singleton cluster when one node is shutdown" taggedAs LongRunningTest in {
runOn(first) {
val secondAddress = address(second)
testConductor.shutdown(second, 0)
testConductor.shutdown(second, 0).await
markNodeAsUnavailable(secondAddress)

9
akka-remote-tests/src/main/scala/akka/remote/testconductor/Conductor.scala
View file

@ -23,6 +23,7 @@ import java.util.concurrent.TimeUnit.MILLISECONDS
import akka.util.{ Timeout }
import scala.reflect.classTag
import akka.ConfigurationException
import akka.AkkaException
sealed trait Direction {
def includes(other: Direction): Boolean
@ -207,7 +208,10 @@ trait Conductor { this: TestConductorExt ⇒
*/
def shutdown(node: RoleName, exitValue: Int): Future[Done] = {
import Settings.QueryTimeout
controller ? Terminate(node, exitValue) mapTo classTag[Done]
import system.dispatcher
// the recover is needed to handle ClientDisconnectedException exception,
// which is normal during shutdown
controller ? Terminate(node, exitValue) mapTo classTag[Done] recover { case _: ClientDisconnectedException ⇒ Done }
}
/**
@ -309,7 +313,7 @@ private[akka] class ServerFSM(val controller: ActorRef, val channel: Channel) ex
whenUnhandled {
case Event(ClientDisconnected, Some(s)) ⇒
s ! Status.Failure(new RuntimeException("client disconnected in state " + stateName + ": " + channel))
s ! Status.Failure(new ClientDisconnectedException("client disconnected in state " + stateName + ": " + channel))
stop()
case Event(ClientDisconnected, None) ⇒ stop()
}
@ -367,6 +371,7 @@ private[akka] class ServerFSM(val controller: ActorRef, val channel: Channel) ex
*/
private[akka] object Controller {
case class ClientDisconnected(name: RoleName)
class ClientDisconnectedException(msg: String) extends AkkaException(msg)
case object GetNodes
case object GetSockAddr
case class CreateServerFSM(channel: Channel)