[GitHub] [kafka] dajac commented on a diff in pull request #12138: MINOR: Followers should not have any remote replica states left over from previous leadership

[GitBox <gi...@apache.org>]

dajac commented on code in PR #12138:
URL: https://github.com/apache/kafka/pull/12138#discussion_r874478979


##########
core/src/main/scala/kafka/cluster/Partition.scala:
##########
@@ -673,7 +675,8 @@ class Partition(val topicPartition: TopicPartition,
 
       updateAssignmentAndIsr(
         assignment = partitionState.replicas.asScala.iterator.map(_.toInt).toSeq,
-        isr = Set.empty[Int],
+        followers = Seq.empty,
+        isr = Set.empty,

Review Comment:
   That's a good question. Whenever the leader updates the ISR, ISR changes would not be propagated to the followers when the ZK controller is used because the leader epoch does not change. I believe that this is why we chose not to update it here because it would not reflect the real ISR. I suppose that we could revise this with KRaft because followers get all the state updates.



-- 
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: jira-unsubscribe@kafka.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org