[
https://issues.apache.org/jira/browse/KAFKA-19327?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18111482#comment-18111482
]
Doyeon Kim commented on KAFKA-19327:
------------------------------------
While prototyping adaptive splitting, I noticed that the controller batch limit
and a custom CreateTopicPolicy can both result in PolicyViolationException. If
all topics are rejected by the policy, recursive splitting could amplify the
number of requests. A configured batch limit would avoid this ambiguity, but
would introduce a new public MM2 configuration. Is adaptive splitting still
preferable for this issue?
> MM2 Source Connector Fails at Scale
> -----------------------------------
>
> Key: KAFKA-19327
> URL: https://issues.apache.org/jira/browse/KAFKA-19327
> Project: Kafka
> Issue Type: Bug
> Components: mirrormaker
> Affects Versions: 3.7.2
> Reporter: Ed Berezitsky
> Assignee: Doyeon Kim
> Priority: Major
>
> We are running MM2 to replicate about 1000+ topics with about 25K partitions
> or something like that.
> {{The MM2 Source Connector fails to start with topics=.*}}
> The Exception:
>
> {code:java}
> [2025-05-23 16:51:06,501] WARN Could not create topic
> <just_one_of_hundreds_failed_topics>.
> (org.apache.kafka.connect.mirror.MirrorSourceConnector)
> java.util.concurrent.CompletionException:
> org.apache.kafka.common.errors.PolicyViolationException: Unable to perform
> excessively large batch operation.
> at
> java.base/java.util.concurrent.CompletableFuture.encodeThrowable(Unknown
> Source)
> at
> java.base/java.util.concurrent.CompletableFuture.completeThrowable(Unknown
> Source)
> at
> java.base/java.util.concurrent.CompletableFuture$UniApply.tryFire(Unknown
> Source)
> at java.base/java.util.concurrent.CompletableFuture.postComplete(Unknown
> Source)
> at
> java.base/java.util.concurrent.CompletableFuture.completeExceptionally(Unknown
> Source)
> at
> org.apache.kafka.common.internals.KafkaCompletableFuture.kafkaCompleteExceptionally(KafkaCompletableFuture.java:49)
> at
> org.apache.kafka.common.internals.KafkaFutureImpl.completeExceptionally(KafkaFutureImpl.java:130)
> at
> org.apache.kafka.clients.admin.KafkaAdminClient$1.handleResponse(KafkaAdminClient.java:1845)
> at
> org.apache.kafka.clients.admin.KafkaAdminClient$AdminClientRunnable.handleResponses(KafkaAdminClient.java:1397)
> at
> org.apache.kafka.clients.admin.KafkaAdminClient$AdminClientRunnable.processRequests(KafkaAdminClient.java:1550)
> at
> org.apache.kafka.clients.admin.KafkaAdminClient$AdminClientRunnable.run(KafkaAdminClient.java:1473)
> at java.base/java.lang.Thread.run(Unknown Source) {code}
>
>
> MM2 Source Connector tries to create all the topics in a[ single
> request|https://github.com/apache/kafka/blob/trunk/connect/mirror/src/main/java/org/apache/kafka/connect/mirror/MirrorSourceConnector.java#L518].
> And fails...
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
