[
https://issues.apache.org/jira/browse/TAJO-48?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13649046#comment-13649046
]
Jihoon Son commented on TAJO-48:
--------------------------------
I've already try to test my patch, but I can't since there are some problems
during the test.
The problem is that datanodes cannot register to the namenode in MiniDFSCluster.
I think that this is due to an wrong configuration.
Currently, I'm investigating what causes this problem.
> BufferUnderflowException occurs during the initialization of RowFile
> --------------------------------------------------------------------
>
> Key: TAJO-48
> URL: https://issues.apache.org/jira/browse/TAJO-48
> Project: Tajo
> Issue Type: Bug
> Components: storage
> Reporter: Jihoon Son
> Assignee: Jihoon Son
> Attachments: TAJO-48.Jihoon.130504.patch.txt
>
>
> The RowFile uses the sync marker to find the start position of a split.
> When a RowFile is initialized, it searches the sync marker from the start of
> the split. During the searching of the sync marker, the underlying buffer
> should be updated when the buffer doesn't have more bytes.
--
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