[
https://issues.apache.org/jira/browse/ARTEMIS-1318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16112683#comment-16112683
]
ASF GitHub Bot commented on ARTEMIS-1318:
-----------------------------------------
Github user franz1981 commented on the issue:
https://github.com/apache/activemq-artemis/pull/1436
@clebertsuconic
Yes I need to run the whole test suite, add some others tests to see if it
is working as expected plus provides the NIO version too.
What I'm not understanding is what is happening here:
https://github.com/franz1981/activemq-artemis/blob/31d5758885214ff65aebfdeb6024f7b275397bff/artemis-journal/src/main/java/org/apache/activemq/artemis/core/journal/impl/JournalImpl.java#L2646-L2646
Seems the only visible difference when the TimedBuffer is not enable (on
NIO too)...
> Uses unbuffered MAPPED Journal if no data-sync is required
> ----------------------------------------------------------
>
> Key: ARTEMIS-1318
> URL: https://issues.apache.org/jira/browse/ARTEMIS-1318
> Project: ActiveMQ Artemis
> Issue Type: Improvement
> Reporter: Francesco Nigro
> Assignee: Francesco Nigro
>
> Disable the TimedBuffer on MAPPED journal if no data-sync in required:
> * reduced footprint (no TimedBuffer)
> * reduced CPU usage (less buffer copy and no TimedBuffer background timer
> thread)
> * reduced latencies (not timeout can influence when the write will happen)
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
