Ngone51 commented on a change in pull request #31763:
URL: https://github.com/apache/spark/pull/31763#discussion_r589188420
##########
File path: core/src/main/scala/org/apache/spark/scheduler/MapStatus.scala
##########
@@ -52,6 +52,13 @@ private[spark] sealed trait MapStatus {
* partitionId of the task or taskContext.taskAttemptId is used.
*/
def mapId: Long
+
+ /**
+ * Extra metadata for map status. This could be used by different
ShuffleManager implementation
+ * to store information they need. For example, a Remote Shuffle Service
ShuffleManager could
+ * store shuffle server information and let reducer task know where to fetch
shuffle data.
+ */
+ def metadata: Option[Serializable]
Review comment:
I think we should continue the discussion based on the
[SPIP](https://docs.google.com/document/d/1d6egnL6WHOwWZe8MWv3m8n4PToNacdx7n_0iMSWwhCQ/edit#)
of SPARK-25299. I assume the SPIP of SPARK-25299 already passed since there're
already some changes merged in, which means people already discussed other
solutions previously. But sure, feel free to raise your concern if the current
design still can't cover your use case.
And we actually know what we need to do (you can also check the sub-tasks of
SPARK-25299). By further discussion, I mean, for the case like SPARK-31801, 1)
we could split it into smaller pieces since its current scope is too big (but
we need to take a look at the PR first) 2) the current implementation is still
not finalized, we can discuss other ways for SPARK-31801.
----------------------------------------------------------------
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]
