Return faster with shuffle blocks since we don't need the rest of the logic in update block :)

holdenk · holdenk · commit 7e32341b33f5 · 2020-06-01T17:38:02.000-07:00
diff --git a/core/src/main/scala/org/apache/spark/storage/BlockManagerMasterEndpoint.scala b/core/src/main/scala/org/apache/spark/storage/BlockManagerMasterEndpoint.scala
@@ -499,11 +499,14 @@ class BlockManagerMasterEndpoint(
         case ShuffleIndexBlockId(shuffleId, mapId, _) =>
           // Don't update the map output on just the index block
           logDebug("Received shuffle index block update for ${shuffleId} ${mapId}")
+          return true
         case ShuffleDataBlockId(shuffleId: Int, mapId: Long, reduceId: Int) =>
           logInfo("Received shuffle data block update for ${shuffleId} ${mapId}, performing update")
           mapOutputTracker.updateMapOutput(shuffleId, mapId, blockManagerId)
+          return true
         case _ =>
           logDebug(s"Unexpected shuffle block type ${blockId}")
+          return false
       }
     }
 

Original file line number	Diff line number	Diff line change
`@@ -499,11 +499,14 @@ class BlockManagerMasterEndpoint(`
`499`	`499`	`case ShuffleIndexBlockId(shuffleId, mapId, _) =>`
`500`	`500`	`// Don't update the map output on just the index block`
`501`	`501`	`logDebug("Received shuffle index block update for ${shuffleId} ${mapId}")`
	`502`	`+ return true`
`502`	`503`	`case ShuffleDataBlockId(shuffleId: Int, mapId: Long, reduceId: Int) =>`
`503`	`504`	`logInfo("Received shuffle data block update for ${shuffleId} ${mapId}, performing update")`
`504`	`505`	`mapOutputTracker.updateMapOutput(shuffleId, mapId, blockManagerId)`
	`506`	`+ return true`
`505`	`507`	`case _ =>`
`506`	`508`	`logDebug(s"Unexpected shuffle block type ${blockId}")`
	`509`	`+ return false`
`507`	`510`	`}`
`508`	`511`	`}`
`509`	`512`