You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@kafka.apache.org by "Daniel (Jira)" <ji...@apache.org> on 2020/10/18 18:02:00 UTC

[jira] [Created] (KAFKA-10617) Cached zkVersion 102 not equal to that in zookeeper, skip updating ISR (kafka.cluster.Partition)

Daniel created KAFKA-10617:
------------------------------

             Summary: Cached zkVersion 102 not equal to that in zookeeper, skip updating ISR (kafka.cluster.Partition)
                 Key: KAFKA-10617
                 URL: https://issues.apache.org/jira/browse/KAFKA-10617
             Project: Kafka
          Issue Type: Bug
          Components: controller
    Affects Versions: 2.4.1
            Reporter: Daniel


This may be related to:

https://issues.apache.org/jira/browse/KAFKA-2729

however all evidence on that, indicates this was fixed ~2.2.0, ad we are seeing this in 2.4.1

 

This hit us in middle of night, are tps moving into the incident was very low..~15tps, after the incident due to offset going back in time r tps shotup to 40k for the next 2 hours, prior to us shutting down the cluster, resetting offset position.

This bug led to our offsets going back in time, causing us to re-process messages from 'days' ago.

 

 

 

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)