Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-issues@hadoop.apache.org
Date: Thu, 5 Mar 2015 00:07:38 +0000 (UTC)
From: "Jing Zhao (JIRA)" <jira@apache.org>
To: hdfs-issues@hadoop.apache.org
Message-ID: <JIRA.12777008.1424739345000.4336.1425514058431@Atlassian.JIRA>
In-Reply-To: <JIRA.12777008.1424739345000@Atlassian.JIRA>
References: <JIRA.12777008.1424739345000@Atlassian.JIRA>
 <JIRA.12777008.1424739345232@arcas>
Subject: [jira] [Commented] (HDFS-7826) Erasure Coding: Update INodeFile
 quota computation for striped blocks
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HDFS-7826?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14347810#comment-14347810 ] 

Jing Zhao commented on HDFS-7826:
---------------------------------

Some comments about the current patch:
# Instead of defining a separate {{computeStripedFileSize}} method, we can add the striped block support directly into {{computeFileSize(boolean, boolean)}}. Note that {{getBlocks}} can handle both the contiguous and striped blocks. But for the last striped UC block, if usePreferredBlockSize4LastUcBlock is true, we may   need to add {{(preferred block size) * (total number of blocks in the striped block group)}}.
# For quota usage computation of striped blocks, we may need to add some new logic. For a m+k striped block (i.e., m data blocks and k parity blocks), its {{numBytes}} field, as well as its file size, should be based on the total size of the m data blocks. However, the quota usage should include both the data blocks and parity blocks. Since our current stripe chunk size is fixed (64KB), we should be able to compute this number based on {{m}}, {{k}}, and {{numBytes}}.

> Erasure Coding: Update INodeFile quota computation for striped blocks
> ---------------------------------------------------------------------
>
>                 Key: HDFS-7826
>                 URL: https://issues.apache.org/jira/browse/HDFS-7826
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Jing Zhao
>            Assignee: Kai Sasaki
>         Attachments: HDFS-7826.1.patch
>
>
> Currently INodeFile's quota computation only considers contiguous blocks (i.e., {{INodeFile#blocks}}). We need to update it to support striped blocks.


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