curcur opened a new pull request #13928:
URL: https://github.com/apache/flink/pull/13928
…Recovery
## What is the purpose of the change
This PR is to disable emitting EndOfChannelRecoveryEvent for Approximate
Local Recovery. EndOfChannelRecoveryEvent is introduced in
[FLINK-19856][network] Emit EndOfChannelRecoveryEvent.
There is a problem if only part of the graph (not the entire graph) is
restarted:
source -> map (1/2) -> sink
-> map (2/2) ->
Let's say map(1/2) fails and only "map (1/2) -> sink" would restart. There
is no way map (2/2) can send the EndOfChannelRecoveryEvent to the sink, because
map (2/2)'s produced partition does not fail.
In this case, sink would fail to send resume consumption notification back
to map (1/2) because it waits for map (2/2) to send the
EndOfChannelRecoveryEvent, which will never happen.
## Brief change log
Override `finishReadRecoveredState` method in
PipelinedApproximateSubpartition to do nothing.
## Verifying this change
ITCase
## 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: (no)
- The S3 file system connector: (no)
## Documentation
- Does this pull request introduce a new feature? (no)
----------------------------------------------------------------
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]
