zk和kafka都是三台
org.apache.kafka.common.errors.NotLeaderForPartitionException This server is not the leader for that topic-partition.
server.properties 配置
--listing properties--
port=9092
socket. request. max. bytes=104857600
num. io. threads=60
Log. dirs=/data01/kafka-logs
socket. send. buffer. bytes=102400
zookeeper. connect=simuwel101:2181, simuwel102:2181, simuwell03:2181
log. retention. check. interval. ms=300000
broker. id=1
Log. retention. hours=72
num. network. threads=300
queued. max. requests=2000
socket. receive. buffer. bytes=102400
num. recovery. threads. per. data. dir=1
zookeeper. connection. timeout. ms=6000
num. partitions=3
delete. topic. enable=true
zookeeper 配置
tickTime=2000
initlimit =10
syncLimit=5
dataDir=/data01/zookeeper
clientPort=2181
server.1=simuwel101:2888:3888
server.2=simuwel102:2888:3888
server.3=simuweL103:2888:3888
autopurge. snapRetainCount=3
autopurge. purgeInterval=1
maxClientCnxns=1000
有消息发成功吗?
另外,simuwel101配置host了吗。
都有配置
集群确认过正常吗?
差不多每隔十几二十分钟报错一次,用zk工具连上去看,都发现kafka节点都在线,zk也确认正常
[2018-10-17 08:42:55,763] INFO [ ReplicaFetcherManager on broker 3] Added fetcher for partitions List[ Iconsumer_offsets,3, initoffset o to broker BrokerEndPoint(1, simuwell11,9092)])(kafka. server. ReplicaFet cherManager)[2018-10-17 08:42:55,778] INFO ReplicaFetcherManager on broker 3] Removed fetcher for partitions I__consumer_offsets,2
追踪下集群之间网络的情况。
你的答案