venkata91 commented on a change in pull request #30164:
URL: https://github.com/apache/spark/pull/30164#discussion_r518284018
##########
File path:
core/src/main/scala/org/apache/spark/storage/BlockManagerMasterEndpoint.scala
##########
@@ -360,6 +371,17 @@ class BlockManagerMasterEndpoint(
}
+ private def addMergerLocation(blockManagerId: BlockManagerId): Unit = {
+ if (!shuffleMergerLocations.contains(blockManagerId.host) &&
!blockManagerId.isDriver) {
+ val shuffleServerId = BlockManagerId(blockManagerId.executorId,
blockManagerId.host,
+ StorageUtils.externalShuffleServicePort(conf))
+ if (shuffleMergerLocations.size >= maxRetainedMergerLocations) {
+ shuffleMergerLocations -= shuffleMergerLocations.head._1
+ }
+ shuffleMergerLocations(shuffleServerId.host) = shuffleServerId
Review comment:
Yes, thats right. this is to keep track of locations where an executor
ran before that we we know the application has registered with that particular
host's External shuffle service which is needed to push the shuffle data.
----------------------------------------------------------------
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
