hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Walter Su (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8684) Erasure Coding: fix some block number calculation for striped block
Date Thu, 02 Jul 2015 03:35:04 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8684?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611411#comment-14611411
] 

Walter Su commented on HDFS-8684:
---------------------------------

bq. Before calling /processOverReplicatedBlock itself live vs expected replicas count check
is done.
Yes, that's true. However, processOverReplicatedBlock excludes decommisioned nodes to form
{{nonExcess}} collection. So {{nonExcess.size()<=sblk.getRealTotalBlockNum()}} still could
happen.

bq. As mentioned already in the comment there, this part is used to update the quota.
It's not only used for quota. The comment is incorrect. You can compare how contiguous block
do it.

{code}
if(!storedBlock.isStriped()){ ..chooseContiguous(..) } else{ chooseStriped()}
{code}
The code is ugly. It's my fault. Could you exchange the if-else statement for me taking this
opportunity? I'll very appreciate.

> Erasure Coding: fix some block number calculation for striped block
> -------------------------------------------------------------------
>
>                 Key: HDFS-8684
>                 URL: https://issues.apache.org/jira/browse/HDFS-8684
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Yi Liu
>            Assignee: Yi Liu
>         Attachments: HDFS-8684-HDFS-7285.001.patch
>
>
> in INodeFile#computeFileSize, the file size calucation for underconstruction striped
block is incorrect.
> in BlockManager#chooseExcessReplicasStriped, the {{if (nonExcess.size() <= groupSize)
{}} is incorrect.



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

Mime
View raw message