[
https://issues.apache.org/jira/browse/HADOOP-16242?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16813959#comment-16813959
]
Da Zhou commented on HADOOP-16242:
----------------------------------
Attached some graphs of memory usage after adding buffer pool, in the test the
memory usage is reduced by 30% after adding buffer pool.
All tests passed using SharedKey:
Tests run: 40, Failures: 0, Errors: 0, Skipped: 0
Tests run: 333, Failures: 0, Errors: 0, Skipped: 21
Tests run: 190, Failures: 0, Errors: 0, Skipped: 15
> ABFS: add bufferpool to AbfsOutputStream
> ----------------------------------------
>
> Key: HADOOP-16242
> URL: https://issues.apache.org/jira/browse/HADOOP-16242
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Affects Versions: 3.2.0
> Reporter: Da Zhou
> Assignee: Da Zhou
> Priority: Major
> Attachments: HADOOP-16242-001.patch,
> WritePerformanceWithBufferPool.pdf
>
>
> In current AbfsOutPutStream, new temporary buffers always get allocated when
> uploading the data to server. These big chunk creates huge pressure on the
> GC, and in some extreme cases that GC doesn't happen in time(observed in some
> customer's env), it also lead to OutOfMemory issue easily.
> Hence we need to replace the current implementation with a bufferpool.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
