返回到文章

采纳

编辑于

KAFKA-7870这个BUG在后续的哪个版本里修复了?

kafka

生产KAFKA集群遇到 这个BUG,现象是KAFKA集群正常。ISR也全,分区leader无法写入和消费,分区follower同步数据失败,最终异常节点内存消耗很高,重启这一个节点后,恢复正常。检查所有日志,只有follower上面有以下同步异常日志,其他的日志均未发现异常。

[2019-09-02 10:03:30,051] WARN [ReplicaFetcher replicaId=8, leaderId=2, fetcherId=0] 
Error in response for fetch request (type=FetchRequest, replicaId=8, maxWait=500, minBytes=1, maxBytes=10485760, 
fetchData={
PortrayAys-10=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[4]), 
MMS-Metric-1=(offset=749142728, logStartOffset=749142728, maxBytes=1048576, currentLeaderEpoch=Optional[0]), 
dialtest-2=(offset=15316723, logStartOffset=15316723, maxBytes=1048576, currentLeaderEpoch=Optional[4]), 
DetectoCPU-3=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[2]), 
OneMinBL-17=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[4]), 
MetricRoute-0=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[2]), 
MetricBaseData-25=(offset=649178, logStartOffset=649178, maxBytes=1048576, currentLeaderEpoch=Optional[0]), 
Argus-RawData-8=(offset=20386963624, logStartOffset=20386403279, maxBytes=1048576, currentLeaderEpoch=Optional[0]), 
MetricBaseData-15=(offset=648652, logStartOffset=648652, maxBytes=1048576, currentLeaderEpoch=Optional[0]), 
PortrayAys-20=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[2]), 
NewProxyBaseData-5=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[2]), 
Detector-To-ES-8=(offset=31787842, logStartOffset=31787842, maxBytes=1048576, currentLeaderEpoch=Optional[4]), 
AIAnomaly-2=(offset=5628, logStartOffset=5628, maxBytes=1048576, currentLeaderEpoch=Optional[2]), 
AIPortray-22=(offset=0, logStartOffset=0, maxBytes=1048576, currentLeaderEpoch=Optional[4])}, 
isolationLevel=READ_UNCOMMITTED, toForget=, metadata=(sessionId=2013330601, epoch=INITIAL)) (kafka.server.ReplicaFetcherThread)

https://issues.apache.org/jira/browse/KAFKA-7870