Thanks Christo,
For the bug-fix, it doesn't cherry-pick cleanly, so we prepared a PR for
it: https://github.com/apache/kafka/pull/21423
-Matthias
On 2/9/26 12:01 AM, David Jacot via dev wrote:
Hey all,
While I agree that KAFKA-20115 is not a regression in 4.2, I think that
releasing 4.2 without the fix would be really bad as it impacts all groups
using the classic protocol. They experience long rebalances when the
underlying __consumer_offsets hosting them is reassigned in the cluster,
which is common with tools such as Cruise Control. It is pretty disruptive.
Best,
David
On Sun, Feb 8, 2026 at 12:32 PM Christo Lolov <[email protected]>
wrote:
Hello Mickael,
In my opinion KAFKA-20115 is a blocker and necessitates another RC once
merged. My hope is that one will be the last RC for 4.2.0. Let me know if
you disagree with the classification.
Hello Matthias,
Before I prepare a new RC, I will cherry-pick the changes.
Best,
Christo
On 7 Feb 2026, at 00:03, Matthias J. Sax <[email protected]> wrote:
Hello,
If we do a new RC, can we merge and cherry-pick two minor docs PRs?
- https://github.com/apache/kafka/pull/21366
- https://github.com/apache/kafka/pull/21397
Both already approved -- did just hold off to merge to `trunk`, to not
later forget to do the cherry-pick.
I also just approved a bug-fix for
https://issues.apache.org/jira/browse/KAFKA-19960 (
https://github.com/apache/kafka/pull/21365) which is a critical bug. Also
ok to only get into 4.2.1, but of course better to get into 4.2.0 directly
if there is no objection?
-Matthias
On 2/6/26 6:59 AM, Mickael Maison wrote:
Hi Christo,
Can you clarify how you've triaged
https://issues.apache.org/jira/browse/KAFKA-20115?
Is it a regression that requires a new RC or not?
Thanks,
Mickael
On Tue, Feb 3, 2026 at 7:00 PM Sean Quah via dev <[email protected]>
wrote:
Hi all,
Unfortunately we discovered a regression in 4.2 today:
https://issues.apache.org/jira/browse/KAFKA-20115
The impact is that when we have a group coordinator leadership change
in
the middle of consumer group rebalancing, we can leave parked
JoinGroup and
SyncGroup requests hanging until clients time out the request. I think
the
timeout interval is the rebalance timeout/max.poll.interval.ms which
is 5
minutes long by default, so it can be fairly disruptive to consumer
applications.
A fix is in progress at https://github.com/apache/kafka/pull/21396.
Thanks,
Sean
On Mon, Feb 2, 2026 at 4:42 PM Christo Lolov <[email protected]>
wrote:
Hello Kafka users, developers and client-developers,
This is the RC3 candidate for release of Apache Kafka 4.2.0.
This release has many exciting changes:
* Kafka Queues (Share Groups) is now production-ready with new
features like the RENEW acknowledgement type for extended processing
times, adaptive batching for coordinators, and comprehensive lag
metrics.
* Kafka Streams brings the server-side rebalance protocol to GA with a
limited feature set, adds dead letter queue support in exception
handlers, introduces anchored wall-clock punctuation for deterministic
scheduling, and gives users full control over whether to send a leave
group request on closing.
* This release also delivers significant improvements to consistency
and observability: CLI tools now feature standardized arguments like
--bootstrap-server across all tools, metric naming has been corrected
to follow the kafka.COMPONENT convention, and new idle ratio metrics
provide better visibility into controller and MetadataLoader
performance.
* Security is enhanced with a new allowlist connector client
configuration override policy, while thread-safety improvements to
RecordHeader eliminate concurrency risks.
* Additional highlights include external schema support in
JsonConverter for reduced message sizes, dynamic configuration for
remote log manager thread pools, and rack ID exposure in the Admin API
for consumer and share group members.
Release notes for the 4.2.0 release:
https://dist.apache.org/repos/dist/dev/kafka/4.2.0-rc3/RELEASE_NOTES.html
*** Please download, test and vote by Thursday, February 5, 5pm UTC
Kafka's KEYS file containing PGP keys we use to sign the release:
https://kafka.apache.org/KEYS
* Release artifacts to be voted upon (source and binary):
https://dist.apache.org/repos/dist/dev/kafka/4.2.0-rc3/
* Docker release artifacts to be voted upon:
apache/kafka:4.2.0-rc3
apache/kafka-native:4.2.0-rc3
* Maven artifacts to be voted upon:
https://repository.apache.org/content/groups/staging/org/apache/kafka/
* Javadoc:
https://dist.apache.org/repos/dist/dev/kafka/4.2.0-rc3/javadoc/index.html
* Tag to be voted upon (off 4.2 branch) is the 4.2.0 tag:
https://github.com/apache/kafka/releases/tag/4.2.0-rc3
* Documentation:
* Protocol:
https://github.com/apache/kafka-site/pull/785
* Successful CI builds for the 4.2 branch:
Unit/integration tests:
https://github.com/apache/kafka/actions/runs/21591368372
System tests: I have started the tests in my environment and will
share
test
results in the next ~24 hours.
* Successful Docker Image Github Actions Pipeline for 4.2 branch:
Docker Build Test Pipeline (JVM):
https://github.com/apache/kafka/actions/runs/21591979775
Docker Build Test Pipeline (Native):
https://github.com/apache/kafka/actions/runs/21592431252
Thanks,
Christo
