Sxnan commented on code in PR #19653:
URL: https://github.com/apache/flink/pull/19653#discussion_r884582848
##########
flink-runtime/src/main/java/org/apache/flink/runtime/taskexecutor/TaskExecutor.java:
##########
@@ -901,7 +901,12 @@ public void releaseOrPromotePartitions(
try {
partitionTracker.stopTrackingAndReleaseJobPartitions(partitionToRelease);
partitionTracker.promoteJobPartitions(partitionsToPromote);
-
+ if (establishedResourceManagerConnection != null) {
+ establishedResourceManagerConnection
+ .getResourceManagerGateway()
+ .reportClusterPartitions(
Review Comment:
The cluster partition status is currently encapsulated in the heartbeat
payload from task executors to RM. The cluster partition may not be available
at the RM when the job is finished so we should report the cluster partition
status as soon as the partition is promoted to cluster partition.
To do this in task executor, we reuse the logic to report the status of
cluster partition, which can make it much simpler than reimplementing this in
JM side. Thus, I prefer doing this in task executor. What do you think?
--
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]
