pnowojski opened a new pull request #14807: URL: https://github.com/apache/flink/pull/14807
This PR depends on https://github.com/apache/flink/pull/14797 This PR fixes two bugs. First: ``` If previous checkpoint is declined, it can happen that task receives both older and newer checkpoint barrier on two different channels, before processing any checkpoint cancellation message/RPC. If the newer checkpoint barrier happens to be processed before the obsolete one incorrect `checkState` in ChannelStatePersister would cause job failure. This checkState was assuming that the previous checkpoint would have been aborted/stopped before triggering the new one, while in reality, this previous checkpoint has never been triggered on this task so it also could not have been stopped. ``` Second: ``` This commit fixes a bug where RemoteInputChannel was incorrectly deciding which buffers should be spilled, if it has received an obsoleted CheckpointBarrier, that hasn't been cancelled (yet?). ``` Both commits are tested by the existing UnalignedCheckpointITCase and some freshly added unit tests. ## Does this pull request potentially affect one of the following parts: - Dependencies (does it add or upgrade a dependency): (yes / **no**) - The public API, i.e., is any changed class annotated with `@Public(Evolving)`: (yes / **no**) - The serializers: (yes / **no** / don't know) - The runtime per-record code paths (performance sensitive): (yes / **no** / don't know) - Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn/Mesos, ZooKeeper: (**yes** / no / don't know) - The S3 file system connector: (yes / **no** / don't know) ## Documentation - Does this pull request introduce a new feature? (yes / **no**) - If yes, how is the feature documented? (**not applicable** / docs / JavaDocs / not documented) ---------------------------------------------------------------- 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: us...@infra.apache.org
