XComp opened a new pull request #15832:
URL: https://github.com/apache/flink/pull/15832
## What is the purpose of the change
Updating the Checkpoint Metadata store of ZooKeeper and Kubernetes might
lead into situations where the data was actually written but the request still
failed (due to network issue for instance). In these cases, we end up in an
inconsistent state where a reference exists in ZooKeeper or the k8s ConfigMap
but the corresponding Checkpoint's state is discarded due to the failure. That
might cause problems during recovery.
## Brief change log
* Introduces `PossibleInconsistentStateException`
* `Fabric8FlinkKubeClient` translates network issues into
`PossibleInconsistentStateException`
* `StateHandleStore` and its implementing classes handle and forward
`PossibleInconsistentStateException`
* Extends `TestingLongStateHandleHelper` to work with reference instead of
saving the state in the handle itself. The previous implementation caused
problems when deserializing the StateHandles not referencing the actual data
but providing a copy of the state.
* `CheckpointCoordinator` handles `PossibleInconsistentStateException`
## Verifying this change
The following tests were extended:
* `CheckpointCoordinatorFailureTest` checks the failure handling for
`PossibleInconsistentStateException` and another exception
* `KubernetesStateHandleStoreTest` and `ZooKeeperStateHandleStoreTest` were
extended to cover the relevant exceptions
* `Fabric8FlinkKubeClientTest` was extended to cover both failure cases
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): no
- The public API, i.e., is any changed class annotated with
`@Public(Evolving)`: no
- The serializers: no
- The runtime per-record code paths (performance sensitive): no
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn/Mesos, ZooKeeper: yes
- The S3 file system connector: no
## Documentation
- Does this pull request introduce a new feature? no
- If yes, how is the feature documented? not applicable
--
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:
[email protected]
