[
https://issues.apache.org/jira/browse/HDFS-1675?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13004377#comment-13004377
]
Hairong Kuang commented on HDFS-1675:
-------------------------------------
I feel that it is better to put RBW in tmp first and then move to rbw when copy
is done. Otherwise, if there is an error during copy then copy aborts, lease
recovery will make a wrong decision deciding replica length and then truncate
visible bytes.
> Transfer RBW between datanodes
> ------------------------------
>
> Key: HDFS-1675
> URL: https://issues.apache.org/jira/browse/HDFS-1675
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: data-node
> Reporter: Tsz Wo (Nicholas), SZE
> Assignee: Tsz Wo (Nicholas), SZE
> Attachments: h1675_20110228.patch, h1675_20110228b.patch,
> h1675_20110308.patch
>
>
> This is the step \(*) described
> [here|https://issues.apache.org/jira/browse/HDFS-1606?focusedCommentId=12991321&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-12991321].
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
