attilapiros commented on a change in pull request #24554: [SPARK-27622][Core]
Avoiding the network when block manager fetches disk persisted RDD blocks from
the same host
URL: https://github.com/apache/spark/pull/24554#discussion_r291672475
##########
File path: core/src/main/scala/org/apache/spark/storage/BlockManager.scala
##########
@@ -946,11 +995,37 @@ private[spark] class BlockManager(
None
}
+ /**
+ * Reads the block from the local directories of another executor which runs
on the same host.
+ */
+ private[spark] def readDiskBlockFromSameHostExecutor(
+ blockId: BlockId,
+ localDirs: Array[String],
+ blockSize: Long): Option[ManagedBuffer] = {
+ val file = ExecutorDiskUtils.getFile(localDirs, subDirsPerLocalDir,
blockId.name)
+ if (file.exists()) {
+ val mangedBuffer = securityManager.getIOEncryptionKey() match {
+ case Some(key) =>
+ // Encrypted blocks cannot be memory mapped; return a special object
that does decryption
+ // and provides InputStream / FileRegion implementations for reading
the data.
+ new EncryptedManagedBuffer(
+ new EncryptedBlockData(file, blockSize, conf, key))
+
+ case _ =>
+ val transportConf = SparkTransportConf.fromSparkConf(conf, "files")
Review comment:
Good catch! Yes, it must be "shuffle" as in `NettyBlockTransferService` and
`ExternalShuffleService` too.
----------------------------------------------------------------
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]
