[
https://issues.apache.org/jira/browse/KAFKA-2477?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14715742#comment-14715742
]
Håkon Hitland edited comment on KAFKA-2477 at 8/26/15 11:38 PM:
----------------------------------------------------------------
We use a replication factor of 3.
The only line with "start offset" that day is the one in the attached log:
[2015-08-24 18:32:32,299] WARN [ReplicaFetcherThread-3-0], Replica 3 for
partition [log.event,3] reset its fetch offset from 10200597616 to current
leader 0's start offset 10200597616 (kafka.server.ReplicaFetcherThread)
e: the leader error reads:
[2015-08-24 18:32:32,145] ERROR [Replica Manager on Broker 0]: Error when
processing fetch request for partition [log.event,3] offset 10349592111 from
follower with correlation id 141609587. Possible cause: Request for offset
10349592111 but we only have log segments in the range 10200597616 to
10349592109. (kafka.server.ReplicaManager)
was (Author: hakon):
We use a replication factor of 3.
The only line with "start offset" that day is the one in the attached log:
[2015-08-24 18:32:32,299] WARN [ReplicaFetcherThread-3-0], Replica 3 for
partition [log.event,3] reset its fetch offset from 10200597616 to current
leader 0's start offset 10200597616 (kafka.server.ReplicaFetcherThread)
> Replicas spuriously deleting all segments in partition
> ------------------------------------------------------
>
> Key: KAFKA-2477
> URL: https://issues.apache.org/jira/browse/KAFKA-2477
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.8.2.1
> Reporter: Håkon Hitland
> Attachments: kafka_log.txt
>
>
> We're seeing some strange behaviour in brokers: a replica will sometimes
> schedule all segments in a partition for deletion, and then immediately start
> replicating them back, triggering our check for under-replicating topics.
> This happens on average a couple of times a week, for different brokers and
> topics.
> We have per-topic retention.ms and retention.bytes configuration, the topics
> where we've seen this happen are hitting the size limit.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
