Yes, that's related to the zombie consumer. Basically, if a consumer is in
a long GC, it may be taken out of the group. However, when the consumer
wakes up from the GC, it may still be able to consume data that it's no
longer responsible for. We protect this a little bit by not allowing a
zombie co
Thanks for the notes, Jun. There was a mention of single-writer behavior, if I
understood correctly, in the context of the new consumer, and I was wondering
if that's specific to the new consumer work or if it's going to have any
dependency or relation to KIP-27.
-Flavio
> On 18 Aug 2015, at 2
The the following are the notes for today's KIP discussion.
* client-side assignment strategy: We discussed concerns about rebalancing
time due to metadata inconsistency, especially when lots of topics are
subscribed. Will discuss a bit more on the mailing list.
* CopyCat data api: The discussion
Hi, Everyone,
We will have a Kafka KIP meeting tomorrow at 11am PST. If you plan to
attend but haven't received an invite, please let me know. The following is
the agenda.
Agenda
1. Discuss the proposal on client-side assignment strategy (
https://cwiki.apache.org/confluence/display/KAFKA/Kafka+C
