[
https://issues.apache.org/jira/browse/KAFKA-6625?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
barry010 updated KAFKA-6625:
----------------------------
Description:
when I upgrade kafka from 0.11.0 to 1.0.0 version，during I shut down the
broker01 with kill signal， update the code，and start the new broker01，I found
that one of my consumer group “flume.mpns” consumption rate had risen to
10M/s。
And I got the information from __consumer_offsets-63 partition，I found the most
of the partition's offset was reset in timestamp 1520316174188(2018/3/6
14:2:54)。this timestamp is between I shut down the broker01 and start the new
broker01。
I have not set the auto.offset.reset，and this topic only one consumer
group，with flume
I put the three log ，consumer log，coordinator log，one __consumer_offsets info。
was:
when I upgrade kafka from 0.11.0 to 1.0.0 version，during I shut down the
broker01 with kill signal， update the code，and start the new broker01，I found
that one of my consumer group “flume.mpns” consumerate rose to 10M/s。
And I got the information from __consumer_offsets-63 partition，I found the most
of the partition's offset was reset in timestamp 1520316174188(2018/3/6
14:2:54)。this timestamp is between I shut down the broker01 and start the new
broker01。
I have not set the auto.offset.reset，and this topic only one consumer
group，with flume
I put the three log ，consumer log，coordinator log，one __consumer_offsets info。
> kafka offset reset when I upgrade kafka from 0.11.0 to 1.0.0 version
> --------------------------------------------------------------------
>
> Key: KAFKA-6625
> URL: https://issues.apache.org/jira/browse/KAFKA-6625
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.11.0.0
> Environment: flume 1.8 with Kafka client version : 0.9.0.1
> Kafka version : 0.11.0
> Reporter: barry010
> Priority: Critical
> Attachments: consumer_offset-36.txt, flume.log.txt, server.log.txt
>
>
> when I upgrade kafka from 0.11.0 to 1.0.0 version，during I shut down the
> broker01 with kill signal， update the code，and start the new broker01，I found
> that one of my consumer group “flume.mpns” consumption rate had risen to
> 10M/s。
> And I got the information from __consumer_offsets-63 partition，I found the
> most of the partition's offset was reset in timestamp 1520316174188(2018/3/6
> 14:2:54)。this timestamp is between I shut down the broker01 and start the new
> broker01。
> I have not set the auto.offset.reset，and this topic only one consumer
> group，with flume
> I put the three log ，consumer log，coordinator log，one __consumer_offsets
> info。
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
