Re: [DISCUSS] KIP-1203: Allow to configure custom `ReplicaPlacer` implementation

hi Jiunn-Yang chia_03 If the configuration is public, the package needs the package-info. chia_04 Additionally, the package includes some non-public stuff, which should be moved to another package. For example, StripedReplicaPlacer. chia_05 Please list the classes needing the UNSTABLE flag B

[PR reviewers wanted] Gradle version upgrade (8 --> 9)

Hello Gradle build fans 😀 I already tagged a few of maintainers directly on PR, but don't make that stop you: please join so we can hopefully bump Kafka build to Gradle 9. PR link: https://github.com/apache/kafka/pull/19513 Regards, Dejan

[DISCUSS] KIP-1208: Add prefix to TopicBasedRemoteLogMetadataManagerConfig to enable setting admin configs

Hello everyone, I would like to start a discussion thread about KIP-1208. Please take a look and feel free to share any thought. https://cwiki.apache.org/confluence/x/vYqhFg Best, Lan

Jenkins build is still unstable: Kafka » Kafka PowerPC Daily » test-powerpc #397

See

[jira] [Created] (KAFKA-19633) Kafka Connect connectors sent out zombie records during rebalance

Shichao An created KAFKA-19633: -- Summary: Kafka Connect connectors sent out zombie records during rebalance Key: KAFKA-19633 URL: https://issues.apache.org/jira/browse/KAFKA-19633 Project: Kafka

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

Hi Sean, You're absolutely right. Thanks for summarizing it so clearly! That's also the reason I renamed the config the original name could indeed be misleading. Best, Lan At 2025-08-22 01:42:09, "Sean Quah" wrote: >Hi Lan, >Thanks for your response. I misunderstood the purpose of the p

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

Hi Chia-Ping, Thanks for review. chia_03: As the number of consumers in a cluster dynamically increases, the size of records may also change dynamically. If a large number of consumers join the cluster while the `max.buffer.size` remains set to a relatively small value, performance could poten

[jira] [Created] (KAFKA-19634) document the encoding of nullable struct

Jun Rao created KAFKA-19634: --- Summary: document the encoding of nullable struct Key: KAFKA-19634 URL: https://issues.apache.org/jira/browse/KAFKA-19634 Project: Kafka Issue Type: Improvement

[jira] [Created] (KAFKA-19632) Handle overlap batch alignment on partition reassignment

Apoorv Mittal created KAFKA-19632: - Summary: Handle overlap batch alignment on partition reassignment Key: KAFKA-19632 URL: https://issues.apache.org/jira/browse/KAFKA-19632 Project: Kafka Is

Re: Request KIP permission

You should be all set. On 8/18/25 9:01 AM, Zhiyan Tang wrote: Hi, I’d like to create a KIP but lack permissions. Could you help me with that? My account number is tzyyzt2025 Thanks

RE: Re: [DISCUSS] KIP-1187: Support to retrieve remote log size via DescribeLogDirs RPC

Hi PoAn, This is truly a valuable KIP! It will be helpful for efficient real-time storage management. The previous comments and your thoughtful responses have already refined the proposal quite well. I do have one minor scenario to bring to your attention. There could be a situation where there

Request KIP permission

Hi, I’d like to create a KIP but lack permissions. Could you help me with that? My account number is tzyyzt2025 Thanks

Scram authentication fails with redhat-jdk-17 in fips environment.

*Overview:* We’re running a Kafka cluster in *KRaft mode* with each node configured as broker,controller, using *Red Hat JDK 17*. We recently migrated the *inter.broker.protocol* and *controller.protocol* authentication mechanisms to *SCRAM-SHA-512*, and updated the necessary files: server.propert

[DISCUSS] KIP-1206: Strict max fetch records in share fetch

Hello all, I would like to start a discussion on KIP-1206: Strict max fetch records in share fetch. This KIP introduces the AcquireMode in ShareFetchRequest, which provides two options: Strict or loose. When strict mode is selected, we should only acquire records till maxFetchRecords. KIP:htt

Re: [PR] Markdown source for AK website [kafka-site]

mumrah merged PR #710: URL: https://github.com/apache/kafka-site/pull/710 -- 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: dev-unsubscr...@kafka.apache.o

Re: [DISCUSS] KIP-1161: cleanup.policy shouldn't be empty

Hi, Jiunn-Yang, Thanks for the reply. The changes look good to me and we can follow up on the 0.0.0.0 issue separately in KIP-1202. Jun On Thu, Aug 21, 2025 at 6:55 AM 黃竣陽 wrote: > Hello Jun, chia, > > > (By the way, the table in KIP-1202 has an incorrect value — null is > acceptable for both

[jira] [Created] (KAFKA-19631) Admin#describeReplicaLogDirs does not handle error correctly

Chia-Ping Tsai created KAFKA-19631: -- Summary: Admin#describeReplicaLogDirs does not handle error correctly Key: KAFKA-19631 URL: https://issues.apache.org/jira/browse/KAFKA-19631 Project: Kafka

[jira] [Resolved] (KAFKA-19618) the `record-size` and `throughput`arguments don't work in TestRaftServer

[ https://issues.apache.org/jira/browse/KAFKA-19618?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Chia-Ping Tsai resolved KAFKA-19618. Fix Version/s: 4.2.0 Resolution: Fixed > the `record-size` and `throughput`argument

[DISCUSS] KIP-1206: Strict max fetch records in share fetch

Hello all, I would like to start a discussion on KIP-1206: Strict max fetch records in share fetch. This KIP introduces the AcquireMode in ShareFetchRequest, which provides two options: Strict or loose. When strict mode is selected, we should only acquire records till maxFetchRecords. KIP:http

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

Hi Lan, Thanks for your response. I misunderstood the purpose of the proposed `max.buffer.size` configs earlier. I thought we were going to reject messages larger than the `max.buffer.size`, but that's not the case. The new configs only control the buffer caching behavior and we will continue to al

RE: Re: [DISCUSS] KIP-1187: Support to retrieve remote log size via DescribeLogDirs RPC

Hi PoAn, This is truly a valuable KIP! It will be helpful for efficient real-time storage management. The previous comments and your thoughtful responses have already refined the proposal quite well. I do have one minor scenario to bring to your attention. There could be a situation where there

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

hi Lan chia_03: Since the hard limit between `message.max.bytes` and this new config was removed, is it still necessary to make `max.buffer.size` dynamic? Users won’t encounter errors when decreasing `message.max.bytes` anymore. Best, Chia-Ping On 2025/08/17 16:37:40 Lan Ding wrote: > Hello

Re: [VOTE] 4.1.0 RC2

Hello! I carried out the following checks: * Built Kafka from source using JDK 24 * Ran unit tests * Validated signatures and checksums * Ran the quickstart tutorial * Tested disabling and reenabling of ELR * Tested ELR defaulting to enabled after an upgrade from a 4.0.0 cluster to 4.1.0 * Tested

Re: [DISCUSS] KIP-1161: cleanup.policy shouldn't be empty

Hello Jun, chia, > (By the way, the table in KIP-1202 has an incorrect value — null is > acceptable for both cases.) KIP-1202 covers the scenario where the listeners configuration is set to the special value PLAINTEXT://0.0.0.0:9092. In this case, if advertised.listeners is configured as null

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

Hi Sean, Thanks for your thoughtful feedback! sq1: Thank you for pointing this out — that was indeed a typo. The intended meaning is to decouple the buffer size from the maximum allowed message size for writing. sq2: You are absolutely right. While we could theoretically add validation during

Re: [DISCUSS] KIP-1161: cleanup.policy shouldn't be empty

2. It seems that advertised.listeners should default to null and can't be That is an inconsistency between the broker and controller, and it will be addressed by KIP-1202. `null` is valid for both, while `empty` is valid only for the controller. As Jun mentioned, the broker encounters an error i

Re: Re: Kafka support for Java 24+

Thanks for the response Stig!! Right - I had some conflicting entries in my maven pom.xml file due to which it was picking up kafka 3.8.1 instead of 4.0.0. Resolved it now. On Thu, Aug 21, 2025 at 2:23 PM Stig Rohde Døssing wrote: > You can also see this from these lines of the stack trace >

[jira] [Created] (KAFKA-19630) LSO movement archives only the first available batch not all the ones prior to the new LSO

Chirag Wadhwa created KAFKA-19630: - Summary: LSO movement archives only the first available batch not all the ones prior to the new LSO Key: KAFKA-19630 URL: https://issues.apache.org/jira/browse/KAFKA-19630

Does Kafka PVC cache authentication credentials/configurations

For our combined kraft setup, We recently migrated the kafka broker and controller authentication mechanisms as follows: Broker: from sasl.mechanism.inter.broker.protocol="PLAIN" to "SCRAM-SHA-512" Controller: from SASL_SSL to SSL. After applying these changes, we restarted the Kafka pods in Kubern

Re: Performance benchmarks results for 3.9 c.f 4.0?

hi Brebner > Brebner, Paul 於 2025年8月21日 下午3:09 寫道： > > there are a couple of results that suggest there may be some 90% latency > increase in 4.0 Out of curiosity, could you please share the references with me? Thanks, Chia-Ping

Re: Performance benchmarks results for 3.9 c.f 4.0?

Hi, I think he may be talking about this https://lists.apache.org/thread/1lym8j528wmm32csk8q4jk8vjlhzbfkt Wang Chia-Ping Tsai 於 2025年8月21日 週四 下午5:43寫道： > hi Brebner > > > Brebner, Paul 於 2025年8月21日 下午3:09 寫道： > > > > there are a couple of results that suggest there may be some 90% latency > i

[jira] [Created] (KAFKA-19629) Deadlock in Kafka Streams when processing Interactive Queries and state store updates concurrently

Evgheni Popusoi created KAFKA-19629: --- Summary: Deadlock in Kafka Streams when processing Interactive Queries and state store updates concurrently Key: KAFKA-19629 URL: https://issues.apache.org/jira/browse/KAFKA

Re: Re: Kafka support for Java 24+

You can also see this from these lines of the stack trace java.base/javax.security.auth.Subject.getSubject(Subject.java:277) org.apache.kafka.common.security.authenticator.SaslClientCallbackHandler.handle(SaslClientCallbackHandler.java:58) In Kafka 4.0.0, that line in the callback handler isn't c

Re: Re: Kafka support for Java 24+

Your stack trace suggests that you aren't actually using Kafka 4.0.0. The relevant code in 4.0.0 is here https://github.com/apache/kafka/blob/4.0.0/clients/src/main/java/org/apache/kafka/common/security/authenticator/SaslClientAuthenticator.java#L221 As you can see, the line number doesn't match

Re: [DISCUSS] KIP-1196: Introduce group.coordinator.append.max.bytes config

Hi Lan, thanks for the KIP! sq1: Can we make a distinction between the broker-level `message.max.bytes` and the topic-level `max.message.bytes`? When `__consumer_offsets` and `__share_group_state` are created, their `max.message.bytes` comes from the broker-level `message.max.bytes`. The topic-lev

Re: [VOTE] KIP-1197: Introduce new method to improve the TopicBasedRemoteLogMetadataManager's initialization

Hi All, Could anyone else kindly take a look at this KIP? It’s not a complex one — we just need to add a default interface method in order to move the PR forward. Many thanks!

Performance benchmarks results for 3.9 c.f 4.0?

Hi all, just wondering if there are any performance benchmarks available for the 4.0 release please (c.f. 3.9.X)? - there are a couple of results that suggest there may be some 90% latency increase in 4.0 so it could be useful to have reassurance that everything is ok, Regards, Paul Brebner