[
https://issues.apache.org/jira/browse/DRILL-142?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13887693#comment-13887693
]
Michael Hausenblas commented on DRILL-142:
------------------------------------------
Did anyone have a closer look at the Low Latency Application MAster (Llama) -
http://cloudera.github.io/llama/ - as it aims to solve the AM/RM ramp-up
overhead for short-running workloads.
> Implement YARN application master
> ---------------------------------
>
> Key: DRILL-142
> URL: https://issues.apache.org/jira/browse/DRILL-142
> Project: Apache Drill
> Issue Type: New Feature
> Reporter: Jacques Nadeau
>
> As YARN adoption grows, Drill needs to support running within YARN. This
> requires implementing a new Application Master. One thing to consider is
> using Continunity's Weave project to help with this.
> http://continuuity.github.io/weave/ As part of this task we need to figure
> out how resource allocation and reallocation would work. since YARN doesn't
> natively support expanding allocations.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
