[
https://issues.apache.org/jira/browse/HADOOP-3218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12587481#action_12587481
]
Amareshwari Sriramadasu commented on HADOOP-3218:
-------------------------------------------------
This check is necessary to avoid temporary files getting created by speculative
tasks when the job is already complete.
Please see
http://issues.apache.org/jira/browse/HADOOP-2391?focusedCommentId=12566183#action_12566183
This check can be removed once HADOOP-2759 (creating a file in hdfs should not
automatically create the parent directories) goes in.
> getRecordWriter need not to check the existence of the output directory
> -----------------------------------------------------------------------
>
> Key: HADOOP-3218
> URL: https://issues.apache.org/jira/browse/HADOOP-3218
> Project: Hadoop Core
> Issue Type: Bug
> Reporter: Runping Qi
>
> I don't think the code segment in
> {code}
> if (!fs.exists(outputPath)) {
> throw new IOException("Output directory doesnt exist");
> }
> {code}
> is necessary.
> It actually cause unnecessary load to dfs namenode.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
