GitHub user vanzin opened a pull request:
https://github.com/apache/spark/pull/19046
[SPARK-18769][yarn] Limit resource requests based on RM's available
resources.
This change limits the number of resource requests Spark will make to the
YARN RM based on the number of available resources the RM advertises. This
lets Spark "play nice" by reducing load on the RM when making requests,
and also slightly speeding up the allocation code when the driver thinks
it needs a lot of containers but the RM doesn't have enough free resources.
Tested with added unit test, and by monitoring the allocator behavior with
dynamic allocation on a spark-shell session.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/vanzin/spark SPARK-18769
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/19046.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #19046
----
commit 35d57a2d2989024884f7e63906e94e826a2c3801
Author: Marcelo Vanzin <van...@cloudera.com>
Date: 2017-08-23T22:56:17Z
[SPARK-18769][yarn] Limit resource requests based on RM's available
resources.
This change limits the number of resource requests Spark will make to the
YARN RM based on the number of available resources the RM advertises. This
lets Spark "play nice" by reducing load on the RM when making requests,
and also slightly speeding up the allocation code when the driver thinks
it needs a lot of containers but the RM doesn't have enough free resources.
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
