[
https://issues.apache.org/jira/browse/KAFKA-1028?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13881326#comment-13881326
]
Andrew Olson commented on KAFKA-1028:
-------------------------------------
Right, it's a broker config property (default behavior plus topic regex
override as you suggested), and request.required.acks can still be set to
whatever value is deemed appropriate.
With this new safeguard in place, when recovering from a situation when all
brokers are down, the topic is available again once the last broker to die has
been restarted. If that's not possible due to catastrophic hardware failure,
the config could be temporarily toggled (e.g.
replica.unclean.election.enable=true) to allow a different leader broker to be
elected and restore topic availability. Note that in order to preserve the
maximum number of messages, the last surviving broker to die should still be
started up first if this last resort option of unclean election has to occur.
> per topic configuration of preference for consistency over availability
> -----------------------------------------------------------------------
>
> Key: KAFKA-1028
> URL: https://issues.apache.org/jira/browse/KAFKA-1028
> Project: Kafka
> Issue Type: Improvement
> Reporter: Scott Clasen
> Assignee: Neha Narkhede
>
> As discussed with Neha on the ML.
> It should be possible to configure a topic to disallow unclean leader
> election, thus preventing the situation where committed messages can be
> discarded once a failed leader comes back online in a situation where it was
> the only ISR.
> This would open kafka to additional usecases where the possibility of
> committted messages being discarded is unacceptable.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
