[
https://issues.apache.org/jira/browse/AMBARI-3976?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13870818#comment-13870818
]
Vitaly Brodetskyi commented on AMBARI-3976:
-------------------------------------------
[~sposetti]/[~mahadev], the same situation we have not only for -j (java-home)
but for -i (jdk-location) too. Should ambari-server throw exception, if path
not exist?
> ambari-server setup -j option installs JDK to indicated location if it does
> not exist there
> -------------------------------------------------------------------------------------------
>
> Key: AMBARI-3976
> URL: https://issues.apache.org/jira/browse/AMBARI-3976
> Project: Ambari
> Issue Type: Bug
> Reporter: Jonathan Maron
> Assignee: Vitaly Brodetskyi
>
> As part of the savanna install of an ambari server I've added code to point
> to an already installed JDK location:
> 2013-12-03 05:41:46.557 18771 DEBUG savanna.utils.remote [-] [dc1-master-001]
> Executing "ambari-server setup -s -j /opt/jdk1.6.0_31 > /dev/null 2>&1"
> _log_command /root/dev/savanna/savanna/utils/remote.py:384
> The interesting thing I've discovered is that if the JDK does not exist in
> that location the JDK is retrieved and installed in that location by ambari.
> I can't find any documentation indicating that behavior. Is it valid for me
> to assume this is expected behavior that will continue to be supported? It
> actually would simplify things for us (no need to check for the existence of
> the JDK in expected location on a remote VM prior to installation), but in
> that case I imagine this aspect of the feature should be documented and
> maintained going forward.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
