KAFKA-18056: Fixed bug in handling commitAsync responses #17909
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
AndrewJSchofield
requested changes
Nov 22, 2024
| @@ -421,6 +421,10 @@ public void onComplete(Map<TopicIdPartition, Set<Long>> offsetsMap, Exception ex | |||
| Set<Long> mergedOffsets = new HashSet<>(); | |||
| mergedOffsets.addAll(oldOffsets); | |||
| mergedOffsets.addAll(newOffsets); | |||
| if (mergedOffsets.size() < (oldOffsets.size() + newOffsets.size())) { | |||
There was a problem hiding this comment.
I think this is quite a dangerous thing to leave in here. In principle, a test could rely upon repeated delivery of a record (timeout for example) which could actually call the callback twice. I'd remove this because I predict it will be the source of a flaky test later on.
There was a problem hiding this comment.
Ah right makes sense, sometimes genuinely we can have the callback called twice. I have removed it now, thanks.
...ts/src/main/java/org/apache/kafka/clients/consumer/internals/ShareConsumeRequestManager.java
Outdated
Show resolved
Hide resolved
omkreddy
approved these changes
Nov 26, 2024
chiacyu
pushed a commit
to chiacyu/kafka
that referenced
this pull request
Nov 30, 2024
There was a bug in handling the ShareAcknowledgeResponse for commitAsync(). Currently after we receive a response, we send out a background event to the application thread to update the acknowledgement commit callbacks for EVERY TopicIdPartition. The map that was sent was not cleared after sending the event. This meant we ended up sending responses for partitions that were already sent in the previous event. So there will be duplicate calls to the callback. The PR fixes the bug and adds a unit test for the same. Reviewers: Andrew Schofield <[email protected]>, Manikumar Reddy <[email protected]>
tedyu
pushed a commit
to tedyu/kafka
that referenced
this pull request
Jan 6, 2025
There was a bug in handling the ShareAcknowledgeResponse for commitAsync(). Currently after we receive a response, we send out a background event to the application thread to update the acknowledgement commit callbacks for EVERY TopicIdPartition. The map that was sent was not cleared after sending the event. This meant we ended up sending responses for partitions that were already sent in the previous event. So there will be duplicate calls to the callback. The PR fixes the bug and adds a unit test for the same. Reviewers: Andrew Schofield <[email protected]>, Manikumar Reddy <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What
There was a bug in handling the
ShareAcknowledgeResponseforcommitAsync(). Currently after we receive a response, we send out a background event to the application thread to update the acknowledgement commit callbacks for EVERYTopicIdPartition.The map that was sent was not cleared after sending the event. This meant we ended up sending responses for partitions that were already sent in the previous event. So there will be duplicate calls to the callback.
The PR fixes the bug and adds a unit test for the same.