[
https://issues.apache.org/jira/browse/KAFKA-1855?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14272472#comment-14272472
]
Henri Pihkala commented on KAFKA-1855:
--------------------------------------
I tried on 0.8.1.1 and didn't see this problem at all, so something must have
changed since then.
Tracking down this bug may be challenging because it seems to appear at random
and there is no clear reason to the socket getting closed, or at least it is
not apparent from the logs. It cannot be due to network problems, as the
controller and the target broker are the same node.
Maybe just make the communication more robust? Does the controller try to
resend the message to the broker?
> Topic unusable after unsuccessful UpdateMetadataRequest
> -------------------------------------------------------
>
> Key: KAFKA-1855
> URL: https://issues.apache.org/jira/browse/KAFKA-1855
> Project: Kafka
> Issue Type: Bug
> Components: controller
> Affects Versions: 0.8.2
> Reporter: Henri Pihkala
>
> Sometimes, seemingly randomly, topic creation/initialization might fail with
> the following lines in controller.log. Other logs show no errors. When this
> happens, the topic is unusable (gives UnknownTopicOrPartition for all
> requests).
> For me this happens 5-10% of the time. Feels like it's more likely to happen
> if there is time between topic creations. Observed on 0.8.2-beta, have not
> tried previous versions.
> [2015-01-09 16:15:27,153] WARN [Controller-0-to-broker-0-send-thread],
> Controller 0 fails to send a request to broker
> id:0,host:192.168.10.21,port:9092 (kafka.controller.RequestSendThread)
> java.io.EOFException: Received -1 when reading from channel, socket has
> likely been closed.
> at kafka.utils.Utils$.read(Utils.scala:381)
> at
> kafka.network.BoundedByteBufferReceive.readFrom(BoundedByteBufferReceive.scala:54)
> at kafka.network.Receive$class.readCompletely(Transmission.scala:56)
> at
> kafka.network.BoundedByteBufferReceive.readCompletely(BoundedByteBufferReceive.scala:29)
> at kafka.network.BlockingChannel.receive(BlockingChannel.scala:108)
> at
> kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:146)
> at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60)
> [2015-01-09 16:15:27,156] ERROR [Controller-0-to-broker-0-send-thread],
> Controller 0 epoch 6 failed to send request
> Name:UpdateMetadataRequest;Version:0;Controller:0;ControllerEpoch:6;CorrelationId:48;ClientId:id_0-host_192.168.10.21-port_9092;AliveBrokers:id:0,host:192.168.10.21,port:9092;PartitionState:[40963064-cdd2-4cd1-937a-9827d3ab77ad,0]
> ->
> (LeaderAndIsrInfo:(Leader:0,ISR:0,LeaderEpoch:0,ControllerEpoch:6),ReplicationFactor:1),AllReplicas:0)
> to broker id:0,host:192.168.10.21,port:9092. Reconnecting to broker.
> (kafka.controller.RequestSendThread)
> java.nio.channels.ClosedChannelException
> at kafka.network.BlockingChannel.send(BlockingChannel.scala:97)
> at
> kafka.controller.RequestSendThread.liftedTree1$1(ControllerChannelManager.scala:132)
> at
> kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:131)
> at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:60)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
