[
https://issues.apache.org/jira/browse/CAMEL-19180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17703304#comment-17703304
]
Rahul Kumar commented on CAMEL-19180:
-------------------------------------
Also how would we go about ensuring. that if we have 2 routes that want to
publish to the same dedupe store and subscribe to it, we have different group
ids? Our kafka brokers have jks file and key store pasword authentication
mechanisms So we can't just use the route config like
#class:com.example.KafkaRepository in the configuration.
> Kafka Idempotent Repository does not give the user control over a randomized
> group id if the kafka broker requires the id to be in a specified form
> ---------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CAMEL-19180
> URL: https://issues.apache.org/jira/browse/CAMEL-19180
> Project: Camel
> Issue Type: Bug
> Environment:
> https://github.com/apache/camel/blob/main/components/camel-kafka/src/main/java/org/apache/camel/processor/idempotent/kafka/KafkaIdempotentRepository.java
> Reporter: Rahul Kumar
> Assignee: Andrea Cosentino
> Priority: Major
>
> [https://github.com/apache/camel/blob/main/components/camel-kafka/src/main/java/org/apache/camel/processor/idempotent/kafka/KafkaIdempotentRepository.java]
> In this class above: Line 288
> String groupId = UUID.randomUUID().toString();
> causes major problems for kafka brokers that expect group ids to be in a
> specified form. Impact is that {*}we cannot subscribe to get the duplicate
> messages if the app restarts as we do not have access to all the messages
> that were produced to the topic for checking for duplicates{*}. If this sort
> of control can be given to the user to set the group id, that would be very
> ideal, as I am right now *copy pasting the class and adding the logic for
> formatting the group id for our requirement to make deduplication work on our
> side.*
>
> Please let us know if there is a way we can either bypass the logic here, or
> have this released in a newer version.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
