[
https://issues.apache.org/jira/browse/MAPREDUCE-4351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13397795#comment-13397795
]
Robert Joseph Evans commented on MAPREDUCE-4351:
------------------------------------------------
@Arun,
Having cgroups will block the container in, but there are other use cases.
# There will be some OSs we want to support that do not have very good process
group isolation so having this pluggable allows us to put in a NullMonitor for
cgroups and a default/polling monitor for older versions of Linux.
# There are also cases where we may only want to monitor the resource and not
shoot the process. There was some discussion about providing feedback to the
scheduler about what resources each container is actually using so that it can
potentially over commit. Even in the case of cgroups we would need something
to be able to get that information and return it to the RM.
There is some work that needs to go into figuring out exactly how we want to
monitor resources vs. enforce resources, and how we want to support this for
different resources. I did not see much of an issue with putting in a small
amount of refactoring so that others can start experimenting with the
stipulation that everything may change in the future, @Unstable. If you
disagree I am fine with holding off on it and starting that discussion now.
Either here, on MAPREDUCE-4256, or perhaps in a new rollup JIRA for the full
feature.
> Make ContainersMonitor pluggable
> --------------------------------
>
> Key: MAPREDUCE-4351
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4351
> Project: Hadoop Map/Reduce
> Issue Type: Sub-task
> Components: mrv2, nodemanager
> Reporter: Andrew Ferguson
> Assignee: Andrew Ferguson
> Attachments: MAPREDUCE-4351-v1.patch, MAPREDUCE-4351-v2.patch,
> MAPREDUCE-4351-v3.patch, MAPREDUCE-4351-v4.patch, MAPREDUCE-4351-v4.patch
>
>
> Make the existing ContainersManager pluggable, just as the ContainerExecutor
> is currently. This will allow us to add container resource enforcement using
> other techniques (such as cgroups) in an extensible fashion.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
