[
https://issues.apache.org/jira/browse/KAFKA-4760?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tomasz Gański updated KAFKA-4760:
---------------------------------
Affects Version/s: 0.10.1.1
Component/s: tools
> Restarting one MirrorMaker causes a global pause in mirroring.
> ---------------------------------------------------------------
>
> Key: KAFKA-4760
> URL: https://issues.apache.org/jira/browse/KAFKA-4760
> Project: Kafka
> Issue Type: Improvement
> Components: tools
> Affects Versions: 0.10.1.1
> Reporter: Tomasz Gański
> Attachments: 1_mm_stop.png
>
>
> When 1 MirrorMaker is restarted in consumer group, the whole consumer group
> stop working and lead to big consumer lag.
> Observed situation: [^1_mm_stop.png]
> It becoming huge problem when you want keep system stable and guarantee that
> crucial messages come as soon as possible. In this case each deployment or
> setting change is risky.
> It would be grate if MirrorMaker could rejoin to previous assigned partitions
> and not caused consumer group rebalance.
> The rejoins to existing consumer group is based on *memberId* and *group
> generation id*. I would nice if it would be possible to make them
> persistence and keep this variable during MirrorMaker restart.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
