AHeise opened a new pull request, #25405:
URL: https://github.com/apache/flink/pull/25405
<!--
*Thank you very much for contributing to Apache Flink - we are happy that
you want to help us improve Flink. To help the community review your
contribution in the best possible way, please go through the checklist below,
which will get the contribution into a shape in which it can be best reviewed.*
*Please understand that we do not do this to make contributions to Flink a
hassle. In order to uphold a high standard of quality for code contributions,
while at the same time managing a large number of contributions, we need
contributors to prepare the contributions well, and give reviewers enough
contextual information for the review. Please also understand that
contributions that do not follow this guide will take longer to review and thus
typically be picked up with lower priority by the community.*
## Contribution Checklist
- Make sure that the pull request corresponds to a [JIRA
issue](https://issues.apache.org/jira/projects/FLINK/issues). Exceptions are
made for typos in JavaDoc or documentation files, which need no JIRA issue.
- Name the pull request in the form "[FLINK-XXXX] [component] Title of the
pull request", where *FLINK-XXXX* should be replaced by the actual issue
number. Skip *component* if you are unsure about which is the best component.
Typo fixes that have no associated JIRA issue should be named following
this pattern: `[hotfix] [docs] Fix typo in event time introduction` or
`[hotfix] [javadocs] Expand JavaDoc for PuncuatedWatermarkGenerator`.
- Fill out the template below to describe the changes contributed by the
pull request. That will give reviewers the context they need to do the review.
- Make sure that the change passes the automated tests, i.e., `mvn clean
verify` passes. You can set up Azure Pipelines CI to do that following [this
guide](https://cwiki.apache.org/confluence/display/FLINK/Azure+Pipelines#AzurePipelines-Tutorial:SettingupAzurePipelinesforaforkoftheFlinkrepository).
- Each pull request should address only one issue, not mix up code from
multiple issues.
- Each commit in the pull request has a meaningful commit message
(including the JIRA id)
- Once all items of the checklist are addressed, remove the above text and
this checklist, leaving only the filled out template below.
**(The sections below can be removed for hotfixes of typos)**
-->
## What is the purpose of the change
When a sink contains a shuffle between writer and committer, a committer may
receive committables coming from multiple subtasks. So far, we immediately
merged then on receival. However, that makes it later impossible to trace
whether we received all messages from an upstream task.
It also made rescaling of the committer awkward: during normal processing
all committables of a committer have the same subtaskId as the committer. On
downscale, these subtaskIds suddenly don't match and need to be replaced, which
we solved by merging the SubtaskCommittableManagers.
This commit decouples the collection of committables from changing the
subtaskId for emission. Committables retain the upstream subtask id in the
CommittableCollection, which survives serialization and deserialization. Only
upon emission, we substitute the subtask id with the one of the emitting
committer.
This is, in particular, useful for a global committer, where all subtasks
are collected. As a side fix, the new serialization also contains the
numberOfSubtasks such that different checkpoints may have different degree of
parallelism.
The old approach probably has edge cases where scaling a UC would result in
stalled pipelines because certain metadata doesn't match. This would not affect
pipelines which chain Writer/Committer (no channel state), Writer and
Committer have same DOP (results in a Forward channel, which doesn't use UC for
exactly these reasons), and a non-keyed shuffles (because they don't provide
any guarantees). Since a keyed shuffle must use the subtask id of the
committables, the new approach should be safe. However, since we disabled UC
entirely for sinks to adhere to the contract of notifyCheckpointComplete, this
shouldn't matter going forward. It's still important to consider these cases
though for restoring from Flink 1 checkpoints.
## Brief change log
* Do not prematurely merge CommittableManager
## Verifying this change
Covered by the following existing test suites that I partially adjusted
*
flink-runtime/src/test/java/org/apache/flink/streaming/runtime/operators/sink/committables
* flink-runtime/src/test/java/org/apache/flink/streaming/api/connector/sink2
*
flink-streaming-java/src/test/java/org/apache/flink/streaming/runtime/operators/sink
## 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, 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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
