[
https://issues.apache.org/jira/browse/HADOOP-6605?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12840623#action_12840623
]
Steve Loughran commented on HADOOP-6605:
----------------------------------------
# Unless there is a need for JDK stuff, all that's needed is java on the path.
Maybe check for it being an odd java and warn if its an unsupported version
(e.g. gcj), but even there, that gets annoying, like Log4J telling you off for
not having a log4j config.
# Not happy with telling people to d/l from sun. It's oracle now, URLs have
changed and anyway, well managed clusters don't have people downloading and
installing by hand.
Ant's startup script is probably the most thorough in find-and-bind, handles
cygwin, IBM JVMs, etc
[http://svn.apache.org/viewvc/ant/core/trunk/src/script/ant?view=markup]
but: people are scared of touching that code, it's brittle code that needs to
be tested everywhere before you can be sure you've made changes.
Adding this stuff is taking on maintenance problems, forever. I'd consider it
very carefully.
> Improve JAVA_HOME detection and handling in hadoop-config
> ---------------------------------------------------------
>
> Key: HADOOP-6605
> URL: https://issues.apache.org/jira/browse/HADOOP-6605
> Project: Hadoop Common
> Issue Type: Improvement
> Affects Versions: 0.20.2
> Reporter: Chad Metcalf
> Assignee: Chad Metcalf
> Attachments: HADOOP-6605.patch
>
>
> JAVA_HOME can be set in hadoop-env.sh and hadoop-config.sh should respect
> this. hadoop-config will source hadoop-env after it determines a config
> directory. If JAVA_HOME still isn't set, it will try to guess one from a list
> of Linux usual locations. If one still isn't set it will quit with a verbose
> error message.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
