wsry commented on a change in pull request #16844:
URL: https://github.com/apache/flink/pull/16844#discussion_r690859332
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/taskmanager/Task.java
##########
@@ -972,24 +972,28 @@ private void releaseResources() {
for (ResultPartitionWriter partitionWriter :
consumableNotifyingPartitionWriters) {
taskEventDispatcher.unregisterPartition(partitionWriter.getPartitionId());
- if (isCanceledOrFailed()) {
- partitionWriter.fail(getFailureCause());
- }
}
- closeNetworkResources();
+ if (isCanceledOrFailed()) {
+ failAllResultPartitions();
+ }
+ closeAllResultPartitions();
Review comment:
In my understanding the fail method more likes release which mean the
result partition can not be consumed and the close method closes all network
resources (network buffers). We can consume a ResultPartition after the network
buffers are recycled, for example, the blocking partition.
--
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]
