[SPARK-5333][Mesos] MesosTaskLaunchData occurs BufferUnderflowException

jongyoul · jongyoul · commit c6693a8cef5d · 2015-01-21T01:11:36.000+09:00
- changed logDebug location
diff --git a/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosTaskLaunchData.scala b/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosTaskLaunchData.scala
@@ -34,8 +34,8 @@ private[spark] case class MesosTaskLaunchData(
     val dataBuffer = ByteBuffer.allocate(4 + serializedTask.limit)
     dataBuffer.putInt(attemptNumber)
     dataBuffer.put(serializedTask)
-    logDebug(s"ByteBuffer size: [${dataBuffer.remaining}]")
     dataBuffer.rewind
+    logDebug(s"ByteBuffer size: [${dataBuffer.remaining}]")
     ByteString.copyFrom(dataBuffer)
   }
 }

Original file line number	Diff line number	Diff line change
`@@ -34,8 +34,8 @@ private[spark] case class MesosTaskLaunchData(`
`34`	`34`	`val dataBuffer = ByteBuffer.allocate(4 + serializedTask.limit)`
`35`	`35`	`dataBuffer.putInt(attemptNumber)`
`36`	`36`	`dataBuffer.put(serializedTask)`
`37`		`- logDebug(s"ByteBuffer size: [${dataBuffer.remaining}]")`
`38`	`37`	`dataBuffer.rewind`
	`38`	`+ logDebug(s"ByteBuffer size: [${dataBuffer.remaining}]")`
`39`	`39`	`ByteString.copyFrom(dataBuffer)`
`40`	`40`	`}`
`41`	`41`	`}`