[
https://issues.apache.org/jira/browse/HADOOP-16548?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16929008#comment-16929008
]
Sneha Vijayarajan commented on HADOOP-16548:
--------------------------------------------
Hi [~ste...@apache.org] ,
This patch renders OutputStream Flush() API to no-op in AbfsOutputStream over a
config control.
Hflush() being the only documented API that can provide persistent data
transfer, Flush() also executing actions to persist buffered data was causing
perf issues.
OutputStreams of WASB and ADLS drivers already have the Flush() as a no-op too.
Option to enable OutputStream Flush() over config is provided incase there are
any customers who have taken a dependency on this.
Kindly request you to review the attached patch. This was tested over a storage
account on East US2 location with command line:
mvn -T 1C -Dparallel-tests=abfs -Dscale -DtestsThreadCount=8 clean verify
[INFO] Tests run: 42, Failures: 0, Errors: 0, Skipped: 0
[WARNING] Tests run: 392, Failures: 0, Errors: 0, Skipped: 25
[WARNING] Tests run: 190, Failures: 0, Errors: 0, Skipped: 124
> ABFS: Config to enable/disable flush operation
> ----------------------------------------------
>
> Key: HADOOP-16548
> URL: https://issues.apache.org/jira/browse/HADOOP-16548
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Reporter: Bilahari T H
> Assignee: Bilahari T H
> Priority: Minor
> Attachments: HADOOP-16548.001.patch
>
>
> Make flush operation enabled/disabled through configuration. This is part of
> performance improvements for ABFS driver.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
