zentol commented on a change in pull request #9250: [FLINK-13371][coordination]
Prevent leaks of blocking partitions
URL: https://github.com/apache/flink/pull/9250#discussion_r309565233
##########
File path:
flink-runtime/src/test/java/org/apache/flink/runtime/taskexecutor/TaskExecutorPartitionLifecycleTest.java
##########
@@ -190,33 +191,62 @@ public void
testConnectionTerminationAfterExternalRelease() throws Exception {
}
@Test
- public void testPartitionReleaseAfterDisconnect() throws Exception {
+ public void testBlockingPartitionReleaseAfterDisconnect() throws
Exception {
testPartitionRelease(
(jobId, partitionId, taskExecutorGateway) ->
taskExecutorGateway.disconnectJobManager(jobId, new Exception("test")),
- true);
+ true,
+ ResultPartitionType.BLOCKING);
}
@Test
- public void testPartitionReleaseAfterReleaseCall() throws Exception {
+ public void testPipelinedPartitionNotReleasedAfterDisconnect() throws
Exception {
Review comment:
there's no good way to do that. Let's say the TE also tracks pipelined
partitions; when are they removed from the tracker? The JM never sends release
calls, and there's no callback for the release by the shuffle environment.
They'd be stuck in the tracker until the job terminates or the connection is
lost.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
