[
https://issues.apache.org/jira/browse/KAFKA-9872?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lee chen updated KAFKA-9872:
----------------------------
Description:
Recently, I face a problem of a broker restart by server shutdown directly.
when I restart the broker, I found my consumer cannot read data.
1,My consumer group named "0",and The partitions of __consumer_offsets is 50.
2,As the Calculation results of kafka. The partition of consumer "0" is in
partition-48
3, kafkaConsumer belong to consumer group “0” cannot read datas ，but when I
change to another group which not in partition-48, all consumers are back to
normal.It seems that* the log dir of partition-48 belong to __consumer_offsets
is damaged*.
4,When I check the log dir of __consumer_offsets-48，I* found so many .snap
files. *
5,I found the nummber recorded in latest log file in __consumer_offsets-48 is
smaller than the nummber recorded in recovery-point-offset-checkpoint and
replication-offset-checkpoint
6,I use kafka-topics.sh to describe __consumer_offsets is normal,All replica
is be contained in ISR.
As my understanding，If log dir is damaged， the broker may be cannot restart.
But In this Scenes， all broker is normal.
was:
Recently, I face a problem of a broker restart by server shutdown directly.
when I restart the broker, I found my consumer cannot read data.
1,My consumer group named "0",and The partitions of __consumer_offsets is 50.
2,As the Calculation results of kafka. The partition of consumer "0" is in
partition-48
3, kafkaConsumer belong to consumer group “0” cannot read datas ，but when I
change to another group which not in partition-48, all consumers are back to
normal.It seems that* the log dir of partition-48 belong to __consumer_offsets
is damaged*.
4,When I check the log dir of __consumer_offsets-48，I* found so many .snap
files. *
5,I found the nummber recorded in latest log file in __consumer_offsets-48 is
smaller than the nummber recorded in recovery-point-offset-checkpoint and
replication-offset-checkpoint
6,I use kafka-topics.sh to describe __consumer_offsets is normal,All replica
is be contained in ISR.
As my understanding，If log dir is damaged， the broker may be cannot restart.
> After restart kafka by a abnormal way，kafkaconsumer cannot read messages
> --------------------------------------------------------------------------
>
> Key: KAFKA-9872
> URL: https://issues.apache.org/jira/browse/KAFKA-9872
> Project: Kafka
> Issue Type: Bug
> Components: core
> Affects Versions: 1.1.0, 2.0.0, 2.1.0, 2.2.0, 2.1.1, 2.3.0, 2.4.0
> Reporter: Lee chen
> Priority: Major
>
> Recently, I face a problem of a broker restart by server shutdown directly.
> when I restart the broker, I found my consumer cannot read data.
> 1,My consumer group named "0",and The partitions of __consumer_offsets is 50.
> 2,As the Calculation results of kafka. The partition of consumer "0" is in
> partition-48
> 3, kafkaConsumer belong to consumer group “0” cannot read datas ，but when I
> change to another group which not in partition-48, all consumers are back to
> normal.It seems that* the log dir of partition-48 belong to
> __consumer_offsets is damaged*.
> 4,When I check the log dir of __consumer_offsets-48，I* found so many .snap
> files. *
> 5,I found the nummber recorded in latest log file in __consumer_offsets-48
> is smaller than the nummber recorded in recovery-point-offset-checkpoint and
> replication-offset-checkpoint
> 6,I use kafka-topics.sh to describe __consumer_offsets is normal,All replica
> is be contained in ISR.
> As my understanding，If log dir is damaged， the broker may be cannot restart.
> But In this Scenes， all broker is normal.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
