hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roman Shaposhnik (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-2757) hive can't find hadoop executor scripts without HADOOP_HOME set
Date Tue, 24 Apr 2012 19:19:34 GMT

     [ https://issues.apache.org/jira/browse/HIVE-2757?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Roman Shaposhnik updated HIVE-2757:
-----------------------------------

    Attachment: HIVE-2757.patch.txt

Here's an example patch that is not meant for inclusion, but rather to generate discussion
whether such an approach would be acceptable.

Basically, the fundamental problem is that Hive java code can be used in other projects (like
Oozie) and hence it can't rely on launcher shell scripts always passing correct set of properties
along based on querying the environment at shell level.

This unique feature of Hive makes discovery of Hadoop executor script happen at the level
of Java code. The patch contains a very naive attempt at doing that while maintaining backward
compatibility with Hadoop 0.20.X and older releases.  The most notable feature that is still
missing is an ability to discover Hadoop that's part of the user's PATH. Before I implement
that, however, I'd like to ask whether exec'ing via ProcessBuilder won't be a better option,
rather than me manually tying to parse PATH (error prone).

Please let me know what you think.

P.S. I have also took the liberty of removing ConfVars.HADOOPCONF since I don't think it is
used anymore.
                
> hive can't find hadoop executor scripts without HADOOP_HOME set
> ---------------------------------------------------------------
>
>                 Key: HIVE-2757
>                 URL: https://issues.apache.org/jira/browse/HIVE-2757
>             Project: Hive
>          Issue Type: Bug
>          Components: CLI
>    Affects Versions: 0.8.0
>            Reporter: Roman Shaposhnik
>         Attachments: HIVE-2757.patch.txt
>
>
> The trouble is that in Hadoop 0.23 HADOOP_HOME has been deprecated. I think it would
be really nice if bin/hive can be modified to capture the which hadoop
> and pass that as a property into the JVM.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message