[jira] [Moved] (HDFS-4737) JVM path embedded in fuse binaries

["Aaron T. Myers (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HDFS-4737?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Aaron T. Myers moved HADOOP-9468 to HDFS-4737:
----------------------------------------------

             Assignee:     (was: Sean Mackrory)
     Target Version/s: 2.0.5-beta  (was: 2.0.5-beta)
    Affects Version/s:     (was: 2.0.4-alpha)
                       2.0.4-alpha
                  Key: HDFS-4737  (was: HADOOP-9468)
              Project: Hadoop HDFS  (was: Hadoop Common)
    
> JVM path embedded in fuse binaries
> ----------------------------------
>
>                 Key: HDFS-4737
>                 URL: https://issues.apache.org/jira/browse/HDFS-4737
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 2.0.4-alpha
>            Reporter: Sean Mackrory
>         Attachments: 0001-HADOOP-9468.-Not-embedding-JVM-library-paths-in-FUSE.patch
>
>
> When the FUSE binaries are built, the paths to libraries in the JVMs is embedded in the RPATH so that they can be found at run-time. From an Apache Bigtop perspective, this is not sufficient because the software may be run on a machine configured very differently from the one on which they were built - so a wrapper sets LD_LIBRARY_PATH according to JAVA_HOME. I recently saw an issue where the original JVM path existed, causing LD_LIBRARY_PATH to be ignored in favor of the RPATH, but it was not the JVM intended for running Hadoop (not JAVA_HOME), and this caused problems.
> I'm told that setting LD_LIBRARY_PATH is standard practice before using the fuse program anyway, and if that's the case, I think removing the RPATH from the binaries is a good idea.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira