[
https://issues.apache.org/jira/browse/MAPREDUCE-6491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14942701#comment-14942701
]
Allen Wittenauer commented on MAPREDUCE-6491:
---------------------------------------------
bq. I don't see any harm in having HADOOP_CLASSPATH in the default list for
trunk and branch-2 both since it is a classpath variable after all.
HADOOP_\*, etc are generally not safe to append to from the job in trunk. One
of the big differences in trunk vs. branch-2 is that the default settings in
hadoop-env.sh no longer ship with append. (i.e., HADOOP_FOO="settings" vs.
HADOOP_FOO="$\{HADOOP_FOO\} settings")
This was intentional to fix the "doubling" problem seen in branch-2 and earlier
when the shell scripts called themselves. (e.g., echo $\{HADOOP_FOO\} results
in "settings settings")
For tasks, there is another JIRA that is introducing HADOOP_TASK_CLASSPATH that
will allow tasks to inject classpaths into trunk in a way that the hadoop shell
commands will pick it up.
> Environment variable handling assumes values should be appended
> ---------------------------------------------------------------
>
> Key: MAPREDUCE-6491
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6491
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 2.2.0
> Reporter: Jason Lowe
> Assignee: Dustin Cote
> Attachments: YARN-2369-1.patch, YARN-2369-10.patch,
> YARN-2369-2.patch, YARN-2369-3.patch, YARN-2369-4.patch, YARN-2369-5.patch,
> YARN-2369-6.patch, YARN-2369-7.patch, YARN-2369-8.patch, YARN-2369-9.patch
>
>
> When processing environment variables for a container context the code
> assumes that the value should be appended to any pre-existing value in the
> environment. This may be desired behavior for handling path-like environment
> variables such as PATH, LD_LIBRARY_PATH, CLASSPATH, etc. but it is a
> non-intuitive and harmful way to handle any variable that does not have
> path-like semantics.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
