wanglijie95 commented on code in PR #21388:
URL: https://github.com/apache/flink/pull/21388#discussion_r1044005122
##########
flink-runtime/src/main/java/org/apache/flink/runtime/executiongraph/EdgeManagerBuildUtil.java:
##########
@@ -197,10 +202,22 @@ private static ConsumedPartitionGroup
createAndRegisterConsumedPartitionGroupToE
ConsumedPartitionGroup consumedPartitionGroup =
ConsumedPartitionGroup.fromMultiplePartitions(
numConsumers, consumedPartitions,
intermediateResult.getResultType());
+ finishConsumablePartitions(intermediateResult, consumedPartitions,
consumedPartitionGroup);
registerConsumedPartitionGroupToEdgeManager(consumedPartitionGroup,
intermediateResult);
return consumedPartitionGroup;
}
+ private static void finishConsumablePartitions(
+ IntermediateResult intermediateResult,
+ List<IntermediateResultPartitionID> consumedPartitionId,
+ ConsumedPartitionGroup consumedPartitionGroup) {
+ for (IntermediateResultPartitionID intermediateResultPartitionID :
consumedPartitionId) {
+ if
(intermediateResult.getPartitionById(intermediateResultPartitionID).isConsumable())
{
Review Comment:
`!mustBePipelinedConsumed && isConsumable` is ok to me.
I think it's a good idea to change to `hasDataAllProduced`, becase the
`consumable` is ambiguous. But let's wait for @zhuzhurk review as well (since
the `consumable` was originally introduced by his change).
--
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]
