Re: kafka consumer rebalance

In old consumer, group coordination is based on zookeeper and new consumer uses inbuilt (not depending on ZK) group coordinator. As of now, auto-migration of migration of old consumers to new consumers is not available. More details here: https://issues.apache.org/jira/browse/KAFKA-4513 On Wed, A

Re: kafka consumer rebalance

Hi Can someone help me here ..! On Tue, Aug 29, 2017 at 11:57 PM, sunil kalva wrote: > Hi > Is there any change in consumer rebalancing algorithm from 0.8.2 and > 0.10.2 versions. If yes how do i do rolling upgrades of consumer client > applications, while doing rolling upgrade few of clients mi

kafka cpu load

Hi: There are 5 machines in my kafka cluster, and the cpu load in these machines will become high in every 7 hours, and each machine does. And I'm sure the cause of this phenomenon was broker process, not other processes. One attachment is screenshot of cpu load, another attachment is m

kafka consumer rebalance

Hi Is there any change in consumer rebalancing algorithm from 0.8.2 and 0.10.2 versions. If yes how do i do rolling upgrades of consumer client applications, while doing rolling upgrade few of clients might be with older algo and few might be with newer algo then rebalancing might not be working we

Kafka behavior not consistent between environments

Hey Kafka Team, We have Kafka deployed in two data centers (DC X and DC Y). DC X seems to be functioning perfectly fine while DC Y consistently produces errors from the brokers as well as the Consumer/Producer/Stream applications. The configuration is essentially the same between the DC's; the

Re: Kafka Streams - groupByKey and Count, null result on join

Ian, if your issue is indeed due to KAFKA-4601, currently the best way would be what I mentioned in that ticket, i.e. manually call `through` to do the repartition, and then from the repartition topic do the aggregation first followed by the join. It will enforce that for each incoming record, it w

Re: Independent zookeep for kafka multi node

Are you trying to create 3 separate ZK+Kafka Broker pairs? On Wed, Aug 23, 2017 at 12:12 PM, IT Consultant <0binarybudd...@gmail.com> wrote: > Hi All > > Can I have three independent zookeepers tagged to three kafka brokers > without any clustering or quorum ? > > Would it be a good idea ? > -

Re: Kafka Streams - groupByKey and Count, null result on join

Thanks for the information Guozhang. Any recommendations for handling or working around this? It's making tests very flakey. On 24 August 2017 at 23:48, Guozhang Wang wrote: > Hi Ian, > > I suspect it has something to do with your specified topology, in which it > triggers the join first, then