You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flume.apache.org by "Hari Shreedharan (JIRA)" <ji...@apache.org> on 2016/11/01 18:50:58 UTC

[jira] [Commented] (FLUME-3020) Improve HDFSEventSink Escape Ingestion by more then 10x by not getting InetAddress on every record

    [ https://issues.apache.org/jira/browse/FLUME-3020?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15626313#comment-15626313 ] 

Hari Shreedharan commented on FLUME-3020:
-----------------------------------------

This looks good. We could actually cache even the replacement strings in local static variables to save that cost as well (which of course is trivial compared to the lookup cost). 

> Improve HDFSEventSink Escape Ingestion by more then 10x by not getting InetAddress on every record
> --------------------------------------------------------------------------------------------------
>
>                 Key: FLUME-3020
>                 URL: https://issues.apache.org/jira/browse/FLUME-3020
>             Project: Flume
>          Issue Type: Improvement
>            Reporter: Theodore michael Malaska
>            Assignee: Theodore michael Malaska
>         Attachments: flume-3020.patch
>
>
> If you are use escaping the current code will call InetAddress on every record.  Which will result is a huge impact to performance.
> TotalTime,8403,
> totalEventTakeTime,1498,
> totalWriteTime,1981,
> totalWriterSetupTime,65,
> commitTime,201,
> flushTime,18,
> startTrans,7,
> The rest is all InetAddress



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)