Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/3710#discussion_r22619853
--- Diff:
streaming/src/main/scala/org/apache/spark/streaming/scheduler/StreamingListenerBus.scala
---
@@ -17,20 +17,28 @@
package org.apache.spark.streaming.scheduler
+import java.util.concurrent.atomic.AtomicBoolean
+import java.util.concurrent.{LinkedBlockingQueue, CopyOnWriteArrayList}
+
+import scala.util.control.NonFatal
+
import org.apache.spark.Logging
-import scala.collection.mutable.{SynchronizedBuffer, ArrayBuffer}
-import java.util.concurrent.LinkedBlockingQueue
+import org.apache.spark.util.Utils
/** Asynchronously passes StreamingListenerEvents to registered
StreamingListeners. */
private[spark] class StreamingListenerBus() extends Logging {
- private val listeners = new ArrayBuffer[StreamingListener]()
- with SynchronizedBuffer[StreamingListener]
+ // `listeners` will be set up during the initialization of the whole
system and the number of
+ // listeners is small, so the copying cost of CopyOnWriteArrayList will
be little. With the help
+ // of CopyOnWriteArrayList, we can eliminate a lock during processing
every event comparing to
+ // SynchronizedBuffer.
+ private val listeners = new CopyOnWriteArrayList[StreamingListener]()
--- End diff --
is the use of copy on write here to speed up initialization, in the sense
that events posted on start up will not be blocked as we add listeners? I may
be missing something, but this speed up doesn't seem very significant to me.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
