[
https://issues.apache.org/jira/browse/FLINK-18060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17123489#comment-17123489
]
Jark Wu commented on FLINK-18060:
---------------------------------
[~pnowojski] Oh, sorry I didn't notice the inner LIMIT in the query (tests are
migrated from blink code). Then I think you are right, the root cause is in
{{AppendOnlyTopNFunction}}. I can help to prepare a pull request for this.
But I still can't understand why the join IT cases are failed, the join cases
doesn't contain LIMIT and the join operators doesn't mutate values in state.
> Add notice for heap-based state backend of state access reference problem
> -------------------------------------------------------------------------
>
> Key: FLINK-18060
> URL: https://issues.apache.org/jira/browse/FLINK-18060
> Project: Flink
> Issue Type: Bug
> Components: Documentation
> Affects Versions: 1.10.1
> Reporter: Yun Tang
> Priority: Major
> Fix For: 1.11.0
>
>
> Currently, when talking about the limitation of {{MemoryStateBackend}} and
> {{FsStateBackend}} in
> [documentation|https://ci.apache.org/projects/flink/flink-docs-stable/ops/state/state_backends.html#the-memorystatebackend],
> Flink only focus on the memory checkpoint size limit. However, current
> documentation forgets to talk about the reference problem, which added in
> {{CopyOnWriteStateMap}}:
> {noformat}
> IMPORTANT: the contracts for this class rely on the user not holding any
> references to objects returned by this map beyond the life cycle of
> per-element operations. Or phrased differently, all get-update-put operations
> on a mapping should be within one call of processElement. Otherwise, the user
> must take care of taking deep copies, e.g. for caching purposes.
> {noformat}
> This actually also applies to {{NestedStateMap}}.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
