cloud-fan commented on a change in pull request #26040: [SPARK-9853][Core]
Optimize shuffle fetch of continuous partition IDs
URL: https://github.com/apache/spark/pull/26040#discussion_r335500126
##########
File path:
core/src/test/scala/org/apache/spark/storage/ShuffleBlockFetcherIteratorSuite.scala
##########
@@ -148,6 +149,82 @@ class ShuffleBlockFetcherIteratorSuite extends
SparkFunSuite with PrivateMethodT
verify(transfer, times(1)).fetchBlocks(any(), any(), any(), any(), any(),
any())
}
+ test("fetch continuous blocks in batch successful 3 local reads + 2 remote
reads") {
+ val blockManager = mock(classOf[BlockManager])
+ val localBmId = BlockManagerId("test-client", "test-client", 1)
+ doReturn(localBmId).when(blockManager).blockManagerId
+
+ // Make sure blockManager.getBlockData would return the merged block
+ val localBlocks = Seq[BlockId](
+ ShuffleBlockId(0, 0, 0),
+ ShuffleBlockId(0, 0, 1),
+ ShuffleBlockId(0, 0, 2))
+ val mergedLocalBlocks = Map[BlockId, ManagedBuffer](
+ ShuffleBlockBatchId(0, 0, 0, 3) -> createMockManagedBuffer())
+ mergedLocalBlocks.foreach { case (blockId, buf) =>
+ doReturn(buf).when(blockManager).getBlockData(meq(blockId))
+ }
+
+ // Make sure remote blocks would return the merged block
+ val remoteBmId = BlockManagerId("test-client-1", "test-client-1", 2)
+ val remoteBlocks = Seq[BlockId](
+ ShuffleBlockId(0, 3, 0),
+ ShuffleBlockId(0, 3, 1))
+ val mergedRemoteBlocks = Map[BlockId, ManagedBuffer](
+ ShuffleBlockBatchId(0, 3, 0, 2) -> createMockManagedBuffer())
+ val transfer = createMockTransfer(mergedRemoteBlocks)
+
+ val blocksByAddress = Seq[(BlockManagerId, Seq[(BlockId, Long, Int)])](
+ (localBmId, localBlocks.map(blockId => (blockId, 1L, 0))),
+ (remoteBmId, remoteBlocks.map(blockId => (blockId, 1L, 1)))
+ ).toIterator
+
+ val taskContext = TaskContext.empty()
Review comment:
We don't need to fix it if we do
https://github.com/apache/spark/pull/26040/files#r335499627
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
