[
https://issues.apache.org/jira/browse/FLINK-2491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14699315#comment-14699315
]
Robert Metzger commented on FLINK-2491:
---------------------------------------
I quickly talked with [~StephanEwen] about this. Currently, the checkpoints are
aborted if not all tasks are online, so no checkpointing is happening if some
of the tasks have finished.
A simple workaround is keeping the source running even tough its not producing
anything.
> Operators are not participating in state checkpointing in some cases
> --------------------------------------------------------------------
>
> Key: FLINK-2491
> URL: https://issues.apache.org/jira/browse/FLINK-2491
> Project: Flink
> Issue Type: Bug
> Components: Streaming
> Affects Versions: 0.10
> Reporter: Robert Metzger
> Assignee: Márton Balassi
> Priority: Critical
> Fix For: 0.10
>
>
> While implementing a test case for the Kafka Consumer, I came across the
> following bug:
> Consider the following topology, with the operator parallelism in parentheses:
> Source (2) --> Sink (1).
> In this setup, the {{snapshotState()}} method is called on the source, but
> not on the Sink.
> The sink receives the generated data.
> The only one of the two sources is generating data.
> I've implemented a test case for this, you can find it here:
> https://github.com/rmetzger/flink/blob/para_checkpoint_bug/flink-tests/src/test/java/org/apache/flink/test/checkpointing/ParallelismChangeCheckpoinedITCase.java
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
