You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hbase.apache.org by "stack (JIRA)" <ji...@apache.org> on 2018/06/24 03:17:00 UTC

[jira] [Commented] (HBASE-20780) ServerRpcConnection logging cleanup

    [ https://issues.apache.org/jira/browse/HBASE-20780?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16521356#comment-16521356 ] 

stack commented on HBASE-20780:
-------------------------------

You can see how logging is the most prominent peek in the two profiles posted.. .one with defaults  [^2.0621.2.12782.lock.svg]  and second with 2MB async WAL buffer size  [^2.0621.2M.46340.lock.svg]  (especially bad in second case).

> ServerRpcConnection logging cleanup
> -----------------------------------
>
>                 Key: HBASE-20780
>                 URL: https://issues.apache.org/jira/browse/HBASE-20780
>             Project: HBase
>          Issue Type: Sub-task
>          Components: logging, Performance
>            Reporter: stack
>            Assignee: stack
>            Priority: Major
>         Attachments: 2.0621.2.12782.lock.svg, 2.0621.2M.46340.lock.svg
>
>
> The logging we do inside in connection header parsing shows as worst offender in the perf locking profiles. Its odd, but easy to clean up. I'd doubt it makes any difference in throughput but lets get it out of the way. Let me load up a few samples of what it current looks like.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)