[ 
https://issues.apache.org/jira/browse/ARTEMIS-3290?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Francesco Nigro updated ARTEMIS-3290:
-------------------------------------
    Description: 
{{JournalImpl::checkKnownRecordID}}'s slow path is performing a RTT on the 
journal `appendExecutor` to check for record existence while journal is 
processing some record eg under a burst of append records.

Considered that {{checkKnownRecordID}} is always called before a real async 
append, that's going to happen vs the same executor, probably worth inlining 
the {{checkKnownRecordID}} slow path into the subsequent append, saving 1 RTT 
and improving append latency under heavy load.

 

  was:
{{JournalImpl::checkKnownRecordID}}'s slow path is performing a RTT on the 
journal `appendExecutor` to check for record existence while journal is 
processing some record eg under a burst of append records.

Considered that {{checkKnownRecordID}} is always called before a real async 
append, that's going to happen vs the same executor, probably worth inlining 
the {{checkKnownRecordID}} slow path into the subsequent append happening on 
the append executor, saving 1 RTT and improving append latency under heavy load.

 


> JournalImpl::checkKnownRecordID can save RTT on append executor
> ---------------------------------------------------------------
>
>                 Key: ARTEMIS-3290
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-3290
>             Project: ActiveMQ Artemis
>          Issue Type: Improvement
>            Reporter: Francesco Nigro
>            Assignee: Francesco Nigro
>            Priority: Major
>
> {{JournalImpl::checkKnownRecordID}}'s slow path is performing a RTT on the 
> journal `appendExecutor` to check for record existence while journal is 
> processing some record eg under a burst of append records.
> Considered that {{checkKnownRecordID}} is always called before a real async 
> append, that's going to happen vs the same executor, probably worth inlining 
> the {{checkKnownRecordID}} slow path into the subsequent append, saving 1 RTT 
> and improving append latency under heavy load.
>  



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

Reply via email to