[ https://issues.apache.org/jira/browse/YARN-4360?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15138152#comment-15138152 ]
Carlo Curino commented on YARN-4360: ------------------------------------ Also big thank you to [~chris.douglas] who very patiently walked through the {{StageAllocatorGreedyRLE}} with me. We spotted a couple of ways to simplify / increase readability and avoid a potential infinite loop (not sure if in practice would materialized, but increased code hygiene nonetheless). Uploading a v6 including all of the above. > Improve GreedyReservationAgent to support "early" allocations, and > performance improvements > -------------------------------------------------------------------------------------------- > > Key: YARN-4360 > URL: https://issues.apache.org/jira/browse/YARN-4360 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacityscheduler, fairscheduler, resourcemanager > Affects Versions: 2.8.0 > Reporter: Carlo Curino > Assignee: Carlo Curino > Attachments: YARN-4360.2.patch, YARN-4360.3.patch, YARN-4360.5.patch, > YARN-4360.patch > > > The GreedyReservationAgent allocates "as late as possible". Per various > conversations, it seems useful to have a mirror behavior that allocates as > early as possible. Also in the process we leverage improvements from > YARN-4358, and implement an RLE-aware StageAllocatorGreedy(RLE), which > significantly speeds up allocation. -- This message was sent by Atlassian JIRA (v6.3.4#6332)