[
https://issues.apache.org/jira/browse/YARN-371?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13571763#comment-13571763
]
Arun C Murthy commented on YARN-371:
------------------------------------
bq. 100,000 tasks at 150 bytes each (name of 3 nodes plus some extra stuff) is
going to take about 15MB to transfer
That seems reasonable, but makes assumptions such as 3 nodes per task etc. That
isn't necessarily true... for e.g. several MR apps have input with higher
replication factor - I've personally fixed JobTracker in the past to not fall
over for (rogue?) jobs such as this.
More importantly - this becomes 15MB per application in the RM. Now, multiply
that by 10,000 concurrent apps... or 100,000 apps as we scale.
Also, it's 15MB of PB on the wire, which in java land (i.e. objects), is
probably much more.
----
Anyway, my point is simple: I'd like to see use-cases where the current
protocol is in sufficient. Even then, I value the ability to scale gracefully
over a couple of corner-case features...
----
Given we are very close to stabilizing YARN all of these changes look, at best,
a 3.x change to me.
It's perfectly fine to experiment as Bobby says, just be aware of current
design centre and context, and particularly when the proposed model was
explicitly rejected when we designed the current system:
http://developer.yahoo.com/blogs/hadoop/posts/2011/03/mapreduce-nextgen-scheduler/
It's fine to change it at some point, it just needs a very, very good reason to
do so. Frankly, I'd like to explore the edges of the current system before we
change it.
> Resource-centric compression in AM-RM protocol limits scheduling
> ----------------------------------------------------------------
>
> Key: YARN-371
> URL: https://issues.apache.org/jira/browse/YARN-371
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: api, resourcemanager, scheduler
> Affects Versions: 2.0.2-alpha
> Reporter: Sandy Ryza
> Assignee: Sandy Ryza
>
> Each AMRM heartbeat consists of a list of resource requests. Currently, each
> resource request consists of a container count, a resource vector, and a
> location, which may be a node, a rack, or "*". When an application wishes to
> request a task run in multiple localtions, it must issue a request for each
> location. This means that for a node-local task, it must issue three
> requests, one at the node-level, one at the rack-level, and one with * (any).
> These requests are not linked with each other, so when a container is
> allocated for one of them, the RM has no way of knowing which others to get
> rid of. When a node-local container is allocated, this is handled by
> decrementing the number of requests on that node's rack and in *. But when
> the scheduler allocates a task with a node-local request on its rack, the
> request on the node is left there. This can cause delay-scheduling to try to
> assign a container on a node that nobody cares about anymore.
> Additionally, unless I am missing something, the current model does not allow
> requests for containers only on a specific node or specific rack. While this
> is not a use case for MapReduce currently, it is conceivable that it might be
> something useful to support in the future, for example to schedule
> long-running services that persist state in a particular location, or for
> applications that generally care less about latency than data-locality.
> Lastly, the ability to understand which requests are for the same task will
> possibly allow future schedulers to make more intelligent scheduling
> decisions, as well as permit a more exact understanding of request load.
> I would propose the tweak of allowing a single ResourceRequest to encapsulate
> all the location information for a task. So instead of just a single
> location, a ResourceRequest would contain an array of locations, including
> nodes that it would be happy with, racks that it would be happy with, and
> possibly *. Side effects of this change would be a reduction in the amount
> of data that needs to be transferred in a heartbeat, as well in as the RM's
> memory footprint, becaused what used to be different requests for the same
> task are now able to share some common data.
> While this change breaks compatibility, if it is going to happen, it makes
> sense to do it now, before YARN becomes beta.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
