hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tsz Wo Nicholas Sze (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8804) Erasure Coding: use DirectBufferPool in DFSStripedInputStream for buffer allocation
Date Fri, 31 Jul 2015 23:49:05 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8804?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14650017#comment-14650017
] 

Tsz Wo Nicholas Sze commented on HDFS-8804:
-------------------------------------------

Some comments on the patch:
- Should getParityBuffer() be synchronized?  It seems that some code path from pread is not
synchronized.
- close() should check whether curStripeBuf == null since close() can be called multiple times.

Some other suggestions can be implemented later:
* It is better to have multiple small data/parity buffers with size == cellSize so that it
is more efficient for reusing the buffers.
* Should DirectBufferPool be singleton?  So that the pool can be shared.

> Erasure Coding: use DirectBufferPool in DFSStripedInputStream for buffer allocation
> -----------------------------------------------------------------------------------
>
>                 Key: HDFS-8804
>                 URL: https://issues.apache.org/jira/browse/HDFS-8804
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Jing Zhao
>            Assignee: Jing Zhao
>         Attachments: HDFS-8804.000.patch
>
>
> Currently we directly allocate direct ByteBuffer in DFSStripedInputstream for the stripe
buffer and the buffers holding parity data. It's better to get ByteBuffer from DirectBufferPool.



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

Mime
View raw message