[
https://issues.apache.org/jira/browse/FLINK-5229?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15834202#comment-15834202
]
ASF GitHub Bot commented on FLINK-5229:
---------------------------------------
Github user StefanRRichter commented on a diff in the pull request:
https://github.com/apache/flink/pull/3179#discussion_r97285679
--- 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 --
You could use the utility method
`StateUtil.bestEffortDiscardAllStateObjects(...)` to reduce code duplication a
bit.
I think we should also discard other state objects that completed fuctures
could have created, e.g.
`StateUtil.bestEffortDiscardAllStateObjects(operatorStatesBackend);` etc. ,
what do you think?
> Cleanup StreamTaskStates if a checkpoint operation of a subsequent operator
> fails
> ----------------------------------------------------------------------------------
>
> Key: FLINK-5229
> URL: https://issues.apache.org/jira/browse/FLINK-5229
> Project: Flink
> Issue Type: Bug
> Components: State Backends, Checkpointing, TaskManager
> Affects Versions: 1.2.0, 1.1.3
> Reporter: Till Rohrmann
> Assignee: Till Rohrmann
> Fix For: 1.2.0, 1.1.4
>
>
> Due to chaining, a {{StreamTask}} needs to checkpoint multiple operators. If
> the first operators succeed in creating a checkpoint but a subsequent
> operator in the chain fails, the {{StreamTask}} has to clean up the already
> completed checkpoints. Otherwise we might end up with orphaned state data.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
