[
https://issues.apache.org/jira/browse/SAMZA-375?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14149913#comment-14149913
]
Chris Riccomini commented on SAMZA-375:
---------------------------------------
A key point in the YARN flow is that the scheduling is split. The RM decides
who gets which offers based on the ResourceRequest, but once a resource has
been given to an AM, the AM decides what to do with it.
If we run a single "Samza" framework in Mesos, and that framework makes
scheduling decisions for all Samza jobs, it seems akin to having YARN's RM make
all scheduling decisions for all Samza jobs, including which containers process
which partitions, etc. Having this logic decentralized seems kind of nice, at
first glance. Not sure how Mesos handles this kind of pattern.
> Investigate Mesos Job Support
> -----------------------------
>
> Key: SAMZA-375
> URL: https://issues.apache.org/jira/browse/SAMZA-375
> Project: Samza
> Issue Type: Bug
> Components: hello-samza
> Reporter: Jon Bringhurst
> Assignee: Jon Bringhurst
> Labels: mesos, project
> Attachments: Screen Shot 2014-08-23 at 5.51.39 PM.png, Screen Shot
> 2014-09-22 at 8.59.12 AM.png
>
>
> It would be nice if Samza had support for Mesos (https://mesos.apache.org/).
> The current plan is to create a MesosJob and MesosJobFactory, then look into
> what it would take to allow the AM code to act as a Mesos scheduler.
> The feasibility of this landing in trunk will be better understood after a
> rough prototype has been created.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
