Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/3179#discussion_r97321764
--- Diff:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/StreamTask.java
---
@@ -954,6 +954,27 @@ public void run() {
owner.getName(),
checkpointMetaData.getCheckpointId(), asyncDurationMillis);
}
} catch (Exception e) {
+ // clean up ongoing operator snapshot results
and non partitioned state handles
+ for (OperatorSnapshotResult
operatorSnapshotResult : snapshotInProgressList) {
+ if (operatorSnapshotResult != null) {
+ try {
+
operatorSnapshotResult.cancel();
+ } catch (Exception
cancelException) {
+
e.addSuppressed(cancelException);
+ }
+ }
+ }
+
+ for (StreamStateHandle
nonPartitionedStateHandle : nonPartitionedStateHandles) {
--- End diff --
Having a second look, I think the other state objects don't have to be
cleaned up, because they are still owned by the corresponding
`OperatorSnapshotResult` which is cleaned up in case of a failure or `close`
operation.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
