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

#2778 - make thread names unique per system

Browse source 
The MonitorableThreadFactory.copy method does not take into account the
counter:AtomicLong field, which then always starts out at zero for every
new dispatcher; the PinnedDispatcher creates one dispatcher per actor,
hence resulting in identical thread names if used by multiple actors.
Solution: add the counter to the case class’ arguments
This commit is contained in:
Roland 2012-12-06 22:50:40 +01:00
parent cd0fa5aee2
commit 6450831e01
5 changed files with 9 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

2
akka-actor/src/main/scala/akka/actor/ActorSystem.scala
View file

@ -613,7 +613,7 @@ private[akka] class ActorSystemImpl(val name: String, applicationConfig: Config,
protected def createScheduler(): Scheduler =
new DefaultScheduler(
new HashedWheelTimer(log,
threadFactory.copy(threadFactory.name + "-scheduler"),
threadFactory.withName(threadFactory.name + "-scheduler"),
settings.SchedulerTickDuration,
settings.SchedulerTicksPerWheel),
log)

2
akka-actor/src/main/scala/akka/dispatch/AbstractDispatcher.scala
View file

@ -527,7 +527,7 @@ class ForkJoinExecutorConfigurator(config: Config, prerequisites: DispatcherPrer
val tf = threadFactory match {
case m: MonitorableThreadFactory ⇒
// add the dispatcher id to the thread names
m.copy(m.name + "-" + id)
m.withName(m.name + "-" + id)
case other ⇒ other
}
new ForkJoinExecutorServiceFactory(

2
akka-actor/src/main/scala/akka/dispatch/PinnedDispatcher.scala
View file

@ -20,7 +20,7 @@ class PinnedDispatcher(
_id: String,
_mailboxType: MailboxType,
_shutdownTimeout: FiniteDuration,
_threadPoolConfig: ThreadPoolConfig = ThreadPoolConfig())
_threadPoolConfig: ThreadPoolConfig)
extends Dispatcher(_prerequisites,
_id,
Int.MaxValue,

8
akka-actor/src/main/scala/akka/dispatch/ThreadPoolBuilder.scala
View file

@ -92,7 +92,7 @@ case class ThreadPoolConfig(allowCorePoolTimeout: Boolean = ThreadPoolConfig.def
val tf = threadFactory match {
case m: MonitorableThreadFactory ⇒
// add the dispatcher id to the thread names
m.copy(m.name + "-" + id)
m.withName(m.name + "-" + id)
case other ⇒ other
}
new ThreadPoolExecutorServiceFactory(tf)
@ -183,9 +183,9 @@ object MonitorableThreadFactory {
case class MonitorableThreadFactory(name: String,
daemonic: Boolean,
contextClassLoader: Option[ClassLoader],
exceptionHandler: Thread.UncaughtExceptionHandler = MonitorableThreadFactory.doNothing)
exceptionHandler: Thread.UncaughtExceptionHandler = MonitorableThreadFactory.doNothing,
protected val counter: AtomicLong = new AtomicLong)
extends ThreadFactory with ForkJoinPool.ForkJoinWorkerThreadFactory {
protected val counter = new AtomicLong
def newThread(pool: ForkJoinPool): ForkJoinWorkerThread = {
val t = wire(new MonitorableThreadFactory.AkkaForkJoinWorkerThread(pool))
@ -196,6 +196,8 @@ case class MonitorableThreadFactory(name: String,
def newThread(runnable: Runnable): Thread = wire(new Thread(runnable, name + "-" + counter.incrementAndGet()))
def withName(newName: String): MonitorableThreadFactory = copy(newName)
protected def wire[T <: Thread](t: T): T = {
t.setUncaughtExceptionHandler(exceptionHandler)
t.setDaemon(daemonic)

2
akka-cluster/src/main/scala/akka/cluster/Cluster.scala
View file

@ -95,7 +95,7 @@ class Cluster(val system: ExtendedActorSystem) extends Extension {
new DefaultScheduler(
new HashedWheelTimer(log,
system.threadFactory match {
case tf: MonitorableThreadFactory ⇒ tf.copy(name = tf.name + "-cluster-scheduler")
case tf: MonitorableThreadFactory ⇒ tf.withName(tf.name + "-cluster-scheduler")
case tf ⇒ tf
},
SchedulerTickDuration,