[
https://issues.apache.org/jira/browse/HADOOP-518?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Owen O'Malley resolved HADOOP-518.
----------------------------------
Resolution: Duplicate
This was fixed by HADOOP-1134 (block crcs).
> hadoop dfs -cp foo/bar/bad-file mumble/new-file copies a file with a bad
> checksum
> ---------------------------------------------------------------------------------
>
> Key: HADOOP-518
> URL: https://issues.apache.org/jira/browse/HADOOP-518
> Project: Hadoop
> Issue Type: Bug
> Components: dfs
> Environment: red hat
> Reporter: Dick King
> Assignee: Sameer Paranjpye
>
> I have a file that reliably generates a checksum error when it's read,
> whether by a map/reduce job as input or by a "dfs -get" command.
> However...
> if I do a "dfs -cp" from the file with the bad checksum the copy can be read
> in its entirety without a checksum error.
> I would consider it reasonable for the command to fail, or for the new file
> to be created but to also have a checksum error in the same place, but this
> behavior is unsettling.
> -dk
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
