[
https://issues.apache.org/jira/browse/KAFKA-2090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14394750#comment-14394750
]
Tim Brooks commented on KAFKA-2090:
-----------------------------------
Created reviewboard https://reviews.apache.org/r/32823/diff/
against branch origin/trunk
> Remove duplicate check to metadataFetchInProgress
> -------------------------------------------------
>
> Key: KAFKA-2090
> URL: https://issues.apache.org/jira/browse/KAFKA-2090
> Project: Kafka
> Issue Type: Improvement
> Reporter: Tim Brooks
> Priority: Minor
> Attachments: KAFKA-2090.patch
>
>
> In the NetworkClient class on a call to poll() there are three checks
> (timeToNextMetadataUpdate, timeToNextReconnectAttempt, waitForMetadataFetch)
> made to see if a metadata update should be made. If all of these == 0 then an
> update is performed.
> However, in the if block a second check is made to metadataFetchInProgress.
> This only make sense in a multithreaded environment. However, as the variable
> is not volatile I suspect this is not the case.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
