[
https://issues.apache.org/jira/browse/HDFS-7729?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14381653#comment-14381653
]
Li Bo commented on HDFS-7729:
-----------------------------
hi, Zhe
I have just uploaded a patch to HDFS-7889 and add a unit test for small cell
size. The cell size in DFSStripedOutputStream is hard coded to 1M. If you set
cell size to 1K, when executing {{encoder.encode(dataBuffers,parityBuffers)}},
the size of {{dataBuffers}} is 1k while the size of {{parityBuffers}} is
1M(because {{cellBuffers}} is created before you call {{setCellSize()}}). If we
get cell size from configuration, I think this problem will not exist.
> Add logic to DFSOutputStream to support writing a file in striping layout
> --------------------------------------------------------------------------
>
> Key: HDFS-7729
> URL: https://issues.apache.org/jira/browse/HDFS-7729
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Li Bo
> Assignee: Li Bo
> Attachments: Codec-tmp.patch, HDFS-7729-001.patch,
> HDFS-7729-002.patch, HDFS-7729-003.patch, HDFS-7729-004.patch,
> HDFS-7729-005.patch, HDFS-7729-006.patch, HDFS-7729-007.patch,
> HDFS-7729-008.patch, HDFS-7729-009.patch
>
>
> If client wants to directly write a file striping layout, we need to add some
> logic to DFSOutputStream. DFSOutputStream needs multiple DataStreamers to
> write each cell of a stripe to a remote datanode.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
