[GitHub] [incubator-celeborn] waitinfuture commented on a diff in pull request #973: [CELEBORN-8] [ISSUE-952][FEATURE] support register shuffle task in map partition mode

[GitBox <gi...@apache.org>]

waitinfuture commented on code in PR #973:
URL: https://github.com/apache/incubator-celeborn/pull/973#discussion_r1024010098


##########
client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala:
##########
@@ -379,6 +402,33 @@ class LifecycleManager(appId: String, val conf: CelebornConf) extends RpcEndpoin
       shuffleId: Int,
       numMappers: Int,
       numReducers: Int): Unit = {
+    handleOfferAndReserveSlots(context, applicationId, shuffleId, numMappers, numReducers)

Review Comment:
   handleOfferAndReserveSlots -> offerAndReserveSlots



##########
client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala:
##########
@@ -394,7 +444,8 @@ class LifecycleManager(appId: String, val conf: CelebornConf) extends RpcEndpoin
             .values()
             .asScala
             .flatMap(_.getAllMasterLocationsWithMinEpoch(shuffleId.toString).asScala)
-            .filter(_.getEpoch == 0)
+            .filter(p =>
+              (partitionType == PartitionType.REDUCE && p.getEpoch == 0) || (partitionType == PartitionType.MAP && p.getId == partitionId))

Review Comment:
   Do we have a consumption that for MapPartition there will always be 1 epoch?



-- 
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: dev-unsubscribe@celeborn.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org