YutingWang98 commented on code in PR #2532:
URL: https://github.com/apache/celeborn/pull/2532#discussion_r1669086013
##########
client/src/main/scala/org/apache/celeborn/client/ChangePartitionManager.scala:
##########
@@ -47,8 +47,14 @@ class ChangePartitionManager(
// shuffleId -> (partitionId -> set of ChangePartition)
private val changePartitionRequests =
JavaUtils.newConcurrentHashMap[Int, ConcurrentHashMap[Integer,
JSet[ChangePartitionRequest]]]()
Review Comment:
Hi, I tested this patch internally, but still saw 1 shuffle read < shuffle
write with small margin. And we are still looking into it.
```
stage 0 ----Shuffle Write Size / Records: 20.8 TiB / 1965259729
stage 1 ---- Shuffle Read Size / Records: 20.8 TiB / 1965259725
```
Do we need to use concurrent hash set for the ChangePartitionRequest as
well? or any other places that could have caused the small shuffle loss? thanks
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
