> We ran onto issues with the Connect system tests caused by AsyncConsumer (https://issues.apache.org/jira/browse/KAFKA-18160).
I believe the patch for *KIP-653* is almost complete. However, it would be ideal to pass all end-to-end (e2e) tests, as it introduces significant changes. Except for *KAFKA-18160*, another failed end-to-end test is *KAFKA-18194*. The root causes have been identified, and the final step is to reach a consensus on the solution. Best, Chia-Ping Mickael Maison <mickael.mai...@gmail.com> 於 2024年12月12日 週四 上午3:20寫道： > Hi David, > > There's also KIP-653: Upgrade log4j to log4j2 which we have not been > able to merge yet. > We ran onto issues with the Connect system tests caused by > AsyncConsumer (https://issues.apache.org/jira/browse/KAFKA-18160). > > Thanks, > Mickael > > On Wed, Dec 11, 2024 at 7:41 PM José Armando García Sancio > <jsan...@confluent.io.invalid> wrote: > > > > Hi David, > > > > Happy feature freeze day! See my comments below. > > > > On Wed, Dec 11, 2024 at 3:46 AM David Jacot <dja...@confluent.io.invalid> > wrote: > > > Thanks for reaching out to me about KIP-966. It seems to me that > > > https://github.com/ahuang98/kafka/pull/2 is a pretty large change for > after > > > the feature freeze. How do you quantify the risk? > > > > You are correct that is a relatively large change and I will be the > > main reviewer for this change. I should point out that the change is > > limited to the raft module. It doesn't change any of the module's > > public interfaces so it doesn't have code impact outside of that > > module. > > > > > Do we have a way back in case we cannot make it or if we discover > issues? > > > > We'll make sure to have this feature well tested in the KRaft protocol > > tests. If there are critical issues with this feature it should be > > relatively simple to revert these changes and unblock the 4.0.0 > > release. > > > > Thanks! > > -- > > -José >
