[ https://issues.apache.org/jira/browse/KAFKA-9357?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Dongjin Lee resolved KAFKA-9357. -------------------------------- Resolution: Not A Problem Could you check the resource status in node 2? (For example, open sockets) It seems like the follower brokers can't fetch the latest records from broker 2, which is taking leadership. Add to this, running Kafka cluster with only three nodes in production is not recommended; it is so vulnerable to resource problems so that these kinds of troubles occur so frequently. +1. If you are Korean, request help to the Korean user community, kafkakru. In this case, you can get better help there. > Error sending fetch request > --------------------------- > > Key: KAFKA-9357 > URL: https://issues.apache.org/jira/browse/KAFKA-9357 > Project: Kafka > Issue Type: Bug > Components: core, network > Affects Versions: 2.1.1 > Environment: centos 7.4, 128gb memory, ssd 4tb disk, intel core 32cpu > Reporter: Byoung joo, Lee > Priority: Critical > > It was 'OK' at first. > I haven't had it for more than six months, but it appears suddenly. Since > then, frequency has decreased, and now the same thing happens in one day. > 6 months-> 1 month-> 1 day-> 2 days. The same thing happens. > If so, the message send to all brokers(3 brokers) stops. > What's the solution? > I read some articles and set the replica.socket.receive.buffer.bytes = > 33554432 value, but to no avail. > > {code:java} > [2019-12-28 23:57:05,153] INFO [ReplicaFetcher replicaId=3, leaderId=2, > fetcherId=0] Error sending fetch request (sessionId=507530224, epoch=INITIAL) > to node 2: java.io.IOException: Connection to 2 was disconnected before the > response was read. (org.apache.kafka.clients.FetchSessionHandler)[2019-12-28 > 23:57:05,153] INFO [ReplicaFetcher replicaId=3, leaderId=2, fetcherId=0] > Error sending fetch request (sessionId=507530224, epoch=INITIAL) to node 2: > java.io.IOException: Connection to 2 was disconnected before the response was > read. (org.apache.kafka.clients.FetchSessionHandler)[2019-12-28 23:57:05,153] > WARN [ReplicaFetcher replicaId=3, leaderId=2, fetcherId=0] Error in response > for fetch request (type=FetchRequest, replicaId=3, maxWait=500, minBytes=1, > maxBytes=10485760, fetchData={logging.client.distribution-2=(offset=8880156, > logStartOffset=7704566, maxBytes=1048576, currentLeaderEpoch=Optional[9]), > logging.client.protocolerr-3=(offset=446249750, logStartOffset=415015689, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > tracking.amount.incoming-2=(offset=71425, logStartOffset=71425, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > logging.err.protocolretry-0=(offset=8732169, logStartOffset=8732169, > maxBytes=1048576, currentLeaderEpoch=Optional[7]), > logging.client.protocolerr-0=(offset=446281548, logStartOffset=415013596, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > logging.client.reception-4=(offset=607737872, logStartOffset=548501721, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > logging.client.reception-1=(offset=607755530, logStartOffset=548502187, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > tracking.amount.outgoing-5=(offset=9667, logStartOffset=9667, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > queuing.device.button-0=(offset=2211009211, logStartOffset=1982602128, > maxBytes=1048576, currentLeaderEpoch=Optional[13]), > queuing.app.biz-0=(offset=0, logStartOffset=0, maxBytes=1048576, > currentLeaderEpoch=Optional[7]), > logging.client.authfailure-5=(offset=1716617, logStartOffset=1461528, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > logging.client.authfailure-2=(offset=1718246, logStartOffset=1463392, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > tracking.amount.outgoing-2=(offset=9667, logStartOffset=9667, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > queuing.device.biz-0=(offset=142402382, logStartOffset=130377493, > maxBytes=1048576, currentLeaderEpoch=Optional[13]), > logging.client.distribution-5=(offset=8883392, logStartOffset=7706253, > maxBytes=1048576, currentLeaderEpoch=Optional[9]), > tracking.amount.incoming-5=(offset=71422, logStartOffset=71422, > maxBytes=1048576, currentLeaderEpoch=Optional[9])}, > isolationLevel=READ_UNCOMMITTED, toForget=, metadata=(sessionId=507530224, > epoch=INITIAL)) (kafka.server.ReplicaFetcherThread)java.io.IOException: > Connection to 2 was disconnected before the response was read at > org.apache.kafka.clients.NetworkClientUtils.sendAndReceive(NetworkClientUtils.java:97) > at > kafka.server.ReplicaFetcherBlockingSend.sendRequest(ReplicaFetcherBlockingSend.scala:97) > at > kafka.server.ReplicaFetcherThread.fetchFromLeader(ReplicaFetcherThread.scala:190) > at > kafka.server.AbstractFetcherThread.processFetchRequest(AbstractFetcherThread.scala:241) > at > kafka.server.AbstractFetcherThread.$anonfun$maybeFetch$3(AbstractFetcherThread.scala:130) > at > kafka.server.AbstractFetcherThread.$anonfun$maybeFetch$3$adapted(AbstractFetcherThread.scala:129) > at scala.Option.foreach(Option.scala:257) at > kafka.server.AbstractFetcherThread.maybeFetch(AbstractFetcherThread.scala:129) > at > kafka.server.AbstractFetcherThread.doWork(AbstractFetcherThread.scala:111) at > kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:82) > {code} > > -- This message was sent by Atlassian Jira (v8.3.4#803005)