You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-dev@hadoop.apache.org by "Joydeep Sen Sarma (JIRA)" <ji...@apache.org> on 2010/06/11 22:56:13 UTC
[jira] Created: (MAPREDUCE-1860) LocalJobRunner should not use JT
filesystem to qualify systemdir
LocalJobRunner should not use JT filesystem to qualify systemdir
----------------------------------------------------------------
Key: MAPREDUCE-1860
URL: https://issues.apache.org/jira/browse/MAPREDUCE-1860
Project: Hadoop Map/Reduce
Issue Type: Bug
Reporter: Joydeep Sen Sarma
getSystemDir() in LocalJobRunner returns a path qualified using the default file system from the Configuration. Usually default filesystem is set to hdfs cluster. which means that when trying to use the local mode - the jobclient ends up using hdfs as system directory.
this seems like an error. localjobrunner should use local file system as it's default file system and that should be used to qualify the mapred system directory.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (MAPREDUCE-1860) LocalJobRunner should use local
filesystem to qualify systemdir
Posted by "Joydeep Sen Sarma (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAPREDUCE-1860?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Joydeep Sen Sarma resolved MAPREDUCE-1860.
------------------------------------------
Resolution: Invalid
this seems fixed via other changes in 0.21 and trunk. so attaching patch for 0.20 and closing.
> LocalJobRunner should use local filesystem to qualify systemdir
> ---------------------------------------------------------------
>
> Key: MAPREDUCE-1860
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1860
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Joydeep Sen Sarma
> Attachments: 1860.patch
>
>
> getSystemDir() in LocalJobRunner returns a path qualified using the default file system from the Configuration. Usually default filesystem is set to hdfs cluster. which means that when trying to use the local mode - the jobclient ends up using hdfs as system directory.
> this seems like an error. localjobrunner should use local file system as it's default file system and that should be used to qualify the mapred system directory.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.