twalthr opened a new pull request, #27294:
URL: https://github.com/apache/flink/pull/27294
## What is the purpose of the change
This fixes a critical design flaw in the CompiledPlan.
The planner guarantees that all operators receive a stable UID. It does
this, by deriving the UIDs from the CompiledPlan. Thus, CompiledPlan and
savepoint/checkpoint are closely linked. The planner does not control the
operators and transformations in sources and sinks. From a planner’s
perspective, sources and sinks are black boxes. Implementers must ensure that
UIDs are assigned to all operators.
This PR ensures that the UIDs assigned in the source are properly
represented in the final JobGraph and savepoints. For backwards compatibility,
we introduce a new version for `StreamExecTableSourceScan`. As a side effect,
this is the first time that we use the versioning infrastructure of
CompiledPlan for ExecNodes. Existing tests show that those work nicely.
## Brief change log
New versions of CompiledPlan will correctly assign the UIDs to
transformations that the source assigned, old versions will still overwrite the
last transformation with the one provided by the planner.
## Verifying this change
This change added tests and can be verified as follows: Unit tests in
`TransformationsTest`, restore tests in `TableSourceScanRestoreTest`.
## 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)`: yes
- 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, 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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
