[ https://issues.apache.org/jira/browse/YARN-3612?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14537208#comment-14537208 ]
Varun Saxena commented on YARN-3612: ------------------------------------ [~chris.douglas], rebased the patch and moved the JIRA to YARN as code change is primarily there. > Resource calculation in child tasks is CPU-heavy > ------------------------------------------------ > > Key: YARN-3612 > URL: https://issues.apache.org/jira/browse/YARN-3612 > Project: Hadoop YARN > Issue Type: Bug > Affects Versions: 2.7.0 > Reporter: Todd Lipcon > Assignee: Varun Saxena > Labels: BB2015-05-TBR > Attachments: MAPREDUCE-4469.patch, MAPREDUCE-4469_rev2.patch, > MAPREDUCE-4469_rev3.patch, MAPREDUCE-4469_rev4.patch, > MAPREDUCE-4469_rev5.patch > > > In doing some benchmarking on a hadoop-1 derived codebase, I noticed that > each of the child tasks was doing a ton of syscalls. Upon stracing, I noticed > that it's spending a lot of time looping through all the files in /proc to > calculate resource usage. > As a test, I added a flag to disable use of the ResourceCalculatorPlugin > within the tasks. On a CPU-bound 500G-sort workload, this improved total job > runtime by about 10% (map slot-seconds by 14%, reduce slot seconds by 8%) -- This message was sent by Atlassian JIRA (v6.3.4#6332)