YARN-6453. fairscheduler-statedump.log gets generated regardless of service

(Contributed by Yufei Gu via Daniel Templeton)

(cherry picked from commit e74ca7216aecc998d277e5ca1f4d0abae43c7e31)

hadoop-common-project/hadoop-common/src/main/conf/log4j.properties

@@ -322,12 +322,16 @@ log4j.appender.EWMA.cleanupInterval=${yarn.ewma.cleanupInterval}
 log4j.appender.EWMA.messageAgeLimitSeconds=${yarn.ewma.messageAgeLimitSeconds}
 log4j.appender.EWMA.maxUniqueMessages=${yarn.ewma.maxUniqueMessages}
 
-# Fair scheduler requests log on state dump
-log4j.logger.org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler.statedump=DEBUG,FSLOGGER
-log4j.additivity.org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler.statedump=false
-log4j.appender.FSLOGGER=org.apache.log4j.RollingFileAppender
-log4j.appender.FSLOGGER.File=${hadoop.log.dir}/fairscheduler-statedump.log
-log4j.appender.FSLOGGER.layout=org.apache.log4j.PatternLayout
-log4j.appender.FSLOGGER.layout.ConversionPattern=%d{ISO8601} %p %c: %m%n
-log4j.appender.FSLOGGER.MaxFileSize=${hadoop.log.maxfilesize}
-log4j.appender.FSLOGGER.MaxBackupIndex=${hadoop.log.maxbackupindex}
+#
+# Fair scheduler state dump
+#
+# Use following logger to dump the state to a separate file
+
+#log4j.logger.org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler.statedump=DEBUG,FSSTATEDUMP
+#log4j.additivity.org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler.statedump=false
+#log4j.appender.FSSTATEDUMP=org.apache.log4j.RollingFileAppender
+#log4j.appender.FSSTATEDUMP.File=${hadoop.log.dir}/fairscheduler-statedump.log
+#log4j.appender.FSSTATEDUMP.layout=org.apache.log4j.PatternLayout
+#log4j.appender.FSSTATEDUMP.layout.ConversionPattern=%d{ISO8601} %p %c: %m%n
+#log4j.appender.FSSTATEDUMP.MaxFileSize=${hadoop.log.maxfilesize}
+#log4j.appender.FSSTATEDUMP.MaxBackupIndex=${hadoop.log.maxbackupindex}

hadoop-yarn-project/hadoop-yarn/hadoop-yarn-site/src/site/markdown/FairScheduler.md

@@ -228,3 +228,9 @@ The following fields can be seen for each queue on the web interface:
 The Fair Scheduler supports moving a running application to a different queue. This can be useful for moving an important application to a higher priority queue, or for moving an unimportant application to a lower priority queue. Apps can be moved by running `yarn application -movetoqueue appID -queue targetQueueName`.
 
 When an application is moved to a queue, its existing allocations become counted with the new queue's allocations instead of the old for purposes of determining fairness. An attempt to move an application to a queue will fail if the addition of the app's resources to that queue would violate the its maxRunningApps or maxResources constraints.
+
+###Dumping Fair Scheduler state
+
+Fair Scheduler is able to dump its state periodically. It is disabled by default. The administrator can enable it by setting org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler logging level to DEBUG.
+
+Fair Scheduler logs go to the Resource Manager log file by default. Fair Scheduler state dumps can potentially generate large amount of log data. Uncomment the "Fair scheduler state dump" section in log4j.properties to dump the state into a separate file.