[
https://issues.apache.org/jira/browse/KAFKA-9125?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16968349#comment-16968349
]
Ismael Juma commented on KAFKA-9125:
------------------------------------
The cache is asynchronously updated and there's no guarantee that a non zero
value is the final value, so it's a bit subtle although good if we can make it
work.
> GroupMetadataManager and TransactionStateManager should use metadata cache
> instead of zkClient
> ----------------------------------------------------------------------------------------------
>
> Key: KAFKA-9125
> URL: https://issues.apache.org/jira/browse/KAFKA-9125
> Project: Kafka
> Issue Type: Sub-task
> Reporter: Viktor Somogyi-Vass
> Assignee: Viktor Somogyi-Vass
> Priority: Major
>
> Both classes query their respective topic's partition count via the zkClient.
> This however could be queried by the broker's local metadata cache.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
