[
https://issues.apache.org/jira/browse/YARN-4599?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15271794#comment-15271794
]
Sidharta Seethana commented on YARN-4599:
-----------------------------------------
{quote}
We are likely better off setting hard limit for all yarn containers so they
don't interfere anything else on the machine. We could disable OOM control on
the cgroup corresponding to all yarn containers (not including NM) and if all
containers are paused, the NM can decide what tasks to kill. This is
particularly useful if we are oversubscribing the node.
{quote}
[~kasha] By tasks, do you mean entire containers? How would the NM know which
containers are paused ? We would probably need to hook into the OOM
notification mechanism for this. I am not sure how well that works in practice.
That being said, it would be good to know that the container got killed
because of an OOM event - otherwise it would be pretty hard to debug
applications that run into this. (I think we might have hit this already with
some internal testing with the memory cgroup)
> Set OOM control for memory cgroups
> ----------------------------------
>
> Key: YARN-4599
> URL: https://issues.apache.org/jira/browse/YARN-4599
> Project: Hadoop YARN
> Issue Type: Bug
> Components: nodemanager
> Affects Versions: 2.9.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Attachments: yarn-4599-not-so-useful.patch
>
>
> YARN-1856 adds memory cgroups enforcing support. We should also explicitly
> set OOM control so that containers are not killed as soon as they go over
> their usage. Today, one could set the swappiness to control this, but
> clusters with swap turned off exist.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
