[jira] [Created] (KAFKA-6961) UnknownTopicOrPartitionException & NotLeaderForPartitionException upon replication of topics.

["kaushik srinivas (JIRA)" <ji...@apache.org>]

kaushik srinivas created KAFKA-6961:
---------------------------------------

             Summary: UnknownTopicOrPartitionException & NotLeaderForPartitionException upon replication of topics.
                 Key: KAFKA-6961
                 URL: https://issues.apache.org/jira/browse/KAFKA-6961
             Project: Kafka
          Issue Type: Bug
         Environment: kubernetes cluster kafka.
            Reporter: kaushik srinivas
         Attachments: k8s_NotLeaderForPartition.txt, k8s_replication_errors.txt

Running kafka & zookeeper in kubernetes cluster.

No of brokers : 3

No of partitions per topic : 3

creating topic with 3 partitions, and looks like all the partitions are up.

Below is the snapshot to confirm the same,

Topic:applestore  PartitionCount:3      ReplicationFactor:3   Configs:
 Topic: applestore      Partition: 0    Leader: 1001    Replicas: 1001,1003,1002        Isr: 1001,1003,1002
 Topic: applestore      Partition: 1    Leader: 1002    Replicas: 1002,1001,1003        Isr: 1002,1001,1003
 Topic: applestore      Partition: 2    Leader: 1003    Replicas: 1003,1002,1001        Isr: 1003,1002,1001
 
But, we see in the brokers as soon as the topics are created below stack traces appears,
 
error 1: 
[2018-05-28 08:00:31,875] ERROR [ReplicaFetcher replicaId=1001, leaderId=1003, fetcherId=7] Error for partition applestore-2 to broker 1003:org.apache.kafka.common.errors.UnknownTopicOrPartitionException: This server does not host this topic-partition. (kafka.server.ReplicaFetcherThread)
 
error 2 :
[2018-05-28 00:43:20,993] ERROR [ReplicaFetcher replicaId=1003, leaderId=1001, fetcherId=0] Error for partition apple-6 to broker 1001:org.apache.kafka.common.errors.NotLeaderForPartitionException: This server is not the leader for that topic-partition. (kafka.server.ReplicaFetcherThread)
 
When we tries producing records to each specific partition, it works fine and also log size across the replicated brokers appears to be equal, which means replication is happening fine.
Attaching the two stack trace files.
 
Why are these stack traces appearing ? can we ignore these stack traces if its some spam messages ?
 
 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)