hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jing Zhao (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-4397) Combined diff may not be identical to the diffs applied sequentially
Date Mon, 14 Jan 2013 21:38:15 GMT

    [ https://issues.apache.org/jira/browse/HDFS-4397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13553149#comment-13553149
] 

Jing Zhao commented on HDFS-4397:
---------------------------------

Besides, I think the bug in the original code is that 

{noformat}
if (createdIndex < 0) {
  // 3.4
  create(node);
  insertDeleted(node, search(deleted, node));
} 
{noformat}

should be

{noformat}
if (createdIndex < 0) {
  // 3.4
  create(node);
  insertDeleted(postDiff.deleted.get(deletedIndex), search(deleted, node));
} 
{noformat}

I.e., the original code puts the wrong node into the deleted list. Shall we update the description
of the jira?

                
> Combined diff may not be identical to the diffs applied sequentially
> --------------------------------------------------------------------
>
>                 Key: HDFS-4397
>                 URL: https://issues.apache.org/jira/browse/HDFS-4397
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: namenode
>            Reporter: Tsz Wo (Nicholas), SZE
>            Assignee: Tsz Wo (Nicholas), SZE
>         Attachments: h4397_20120113.patch
>
>
> Let diffs
> {noformat}
> d(1,k)   = m_1 + m_2 + ... + m_k;
> d(k+1,n) = m_{k+1} + m_{k+2} + ... + m_n;
> {noformat}
> and
> {noformat}
> d(1,n)   = m_1 + m_2 + ... + m_n,
> {noformat}
> where m_j denote modifications.  Then, we may have
> {noformat}
> d(1,k) + d(k+1,n) != d(1,n).
> {noformat}

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message