[jira] [Updated] (KAFKA-13932) Replication data loss in some cases

["Bale (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/KAFKA-13932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Bale updated KAFKA-13932:
-------------------------
    Attachment: connect-mirror-maker.properties

> Replication data loss in some cases
> -----------------------------------
>
>                 Key: KAFKA-13932
>                 URL: https://issues.apache.org/jira/browse/KAFKA-13932
>             Project: Kafka
>          Issue Type: Bug
>          Components: mirrormaker
>    Affects Versions: 2.8.0
>            Reporter: Bale
>            Priority: Minor
>         Attachments: connect-mirror-maker.properties
>
>
> Kafka version: 2.8.0
> Replication flow: A -> B
> Topic: DL2 -> A.DL2
>  
> I use mm2 to replicate cluster A data to cluster B. In some cases, I found consumer group offset is bigger than partition offset in cluster B which will cause data loss of the mirror topic.
>  
> [Steps]
> 1. Start mm2 with the attachment: connect-mirror-maker.properties.
> 2.Create topic DL2, push 5 messages and consume with group g1.
> 3.Stop mm2 and push 5 messages to the topic DL2.
> 4. After 1 hour, the data of topic DL2 is deleted.
> 5.Restart mm2 and sync consumer group offset, found the consume offset is bigger than server offset



--
This message was sent by Atlassian Jira
(v8.20.7#820007)