You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Steve Loughran (Jira)" <ji...@apache.org> on 2020/06/01 09:30:00 UTC

[jira] [Commented] (HADOOP-16133) S3A statistic collection underrecords bytes written in helper threads

    [ https://issues.apache.org/jira/browse/HADOOP-16133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17120888#comment-17120888 ] 

Steve Loughran commented on HADOOP-16133:
-----------------------------------------

to fix in HADOOP-16830 by only updating the counters in close() and unbuffer(). if done in the worker thread, it will be correct

> S3A statistic collection underrecords bytes written in helper threads
> ---------------------------------------------------------------------
>
>                 Key: HADOOP-16133
>                 URL: https://issues.apache.org/jira/browse/HADOOP-16133
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>    Affects Versions: 3.1.2
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>            Priority: Minor
>
> Applications collecting per-thread statistics from S3A get underreporting of bytes written, as all byte written in the worker call update those in a different thread.
> Proposed: 
> * the bytes upload statistics are uploaded in the primary thread as a block is queued for write, not after in the completion phase in the other thread
> * final {{WriteOperationsHelper.writeSuccessful()}} takes the final statistics for its own entertainment
> Really I want context-specific storage statistics.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org