You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@ambari.apache.org by "Hudson (JIRA)" <ji...@apache.org> on 2015/05/15 14:38:00 UTC

[jira] [Commented] (AMBARI-11162) Ambari should configure NameNode to terminate on OutOfMemoryError.

    [ https://issues.apache.org/jira/browse/AMBARI-11162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14545427#comment-14545427 ] 

Hudson commented on AMBARI-11162:
---------------------------------

FAILURE: Integrated in Ambari-trunk-Commit #2606 (See [https://builds.apache.org/job/Ambari-trunk-Commit/2606/])
AMBARI-11162. Ambari should configure NameNode to terminate on OutOfMemoryError. (aonishuk) (aonishuk: http://git-wip-us.apache.org/repos/asf?p=ambari.git&a=commit&h=c36ec90919918aeeeb5217f8280d40cc36ae0082)
* ambari-server/src/main/resources/stacks/HDP/2.3/services/HDFS/configuration/hadoop-env.xml


> Ambari should configure NameNode to terminate on OutOfMemoryError.
> ------------------------------------------------------------------
>
>                 Key: AMBARI-11162
>                 URL: https://issues.apache.org/jira/browse/AMBARI-11162
>             Project: Ambari
>          Issue Type: Bug
>            Reporter: Andrew Onischuk
>            Assignee: Andrew Onischuk
>             Fix For: 2.1.0
>
>
> It is dangerous for the NameNode to keep running if it encounters
> `OutOfMemoryError`. There is some legacy code in the RPC handling layer that
> catches `OutOfMemoryError`, and we're not going to be able to revert this
> code. Instead, we can kill it externally by setting a JVM option. We would
> like Ambari to configure the NameNode to run with this option:
>     
>     
>     
>     -XX:OnOutOfMemoryError="kill -9 %p"
>     
> This can be set in `HADOOP_NAMENODE_OPTS` in hadoop-env.sh.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)