基于ambari的yarn-site.xml的分析与介绍

key	value	描述
hadoop.registry.rm.enabled	true	yarn资源管理器中是否启用了注册表
hadoop.registry.zk.quorum	node1:2181,node2:2181,node3:2181	用于定义注册表的zookeeper仲裁绑定
manage.include.files	false
yarn.acl.enable	false	是否启用了acls
yarn.admin.acl	activity_analyzer,yarn	谁可以成为YARN集群管理员的ACL
yarn.application.classpath	/usr/hdp/2.6.5.0-292/hadoop/conf,/usr/hdp/2.6.5.0-292/hadoop/,/usr/hdp/2.6.5.0-292/hadoop/lib/,/usr/hdp/current/hadoop-hdfs-client/,/usr/hdp/current/hadoop-hdfs-client/lib/,/usr/hdp/current/hadoop-yarn-client/,/usr/hdp/current/hadoop-yarn-client/lib/,/usr/hdp/current/ext/hadoop/*
yarn.client.failover-proxy-provider	org.apache.hadoop.yarn.client.RequestHedgingRMFailoverProxyProvider	在HA模式下配置故障转移的代理程序
yarn.client.nodemanager-connect.max-wait-ms	60000	连接到NM的最大等待时间
yarn.client.nodemanager-connect.retry-interval-ms	10000	每次尝试连接到NM的时间间隔
yarn.http.policy	HTTP_ONLY	设置是支持HTTP还是HTTPS
yarn.log-aggregation-enable	true	是否启用日志聚合功能
yarn.log-aggregation.file-controller.IndexedFormat.class	org.apache.hadoop.yarn.logaggregation.filecontroller.ifile.LogAggregationIndexedFileController
yarn.log-aggregation.file-controller.TFile.class	org.apache.hadoop.yarn.logaggregation.filecontroller.tfile.LogAggregationTFileController	支持TFile的读写操作类
yarn.log-aggregation.file-formats	IndexedFormat,TFile	收集日志的控制器必须要有TFile
yarn.log-aggregation.retain-seconds	2592000	多久删除聚合日志
yarn.log.server.url	http://node1:19888/jobhistory/logs
yarn.log.server.web-service.url	http://node1:8188/ws/v1/applicationhistory	日志聚合服务器Web服务的URL
yarn.node-labels.enabled	false	启用节点标签功能
yarn.node-labels.fs-store.retry-policy-spec	2000, 500
yarn.node-labels.fs-store.root-dir	/system/yarn/node-labels	节点标签的路径，/tmp/hadoop-yarn-${user}/node-labels/
yarn.nodemanager.address	0.0.0.0:45454
yarn.nodemanager.admin-env	MALLOC_ARENA_MAX=$MALLOC_ARENA_MAX	nodemanager的环境变量
yarn.nodemanager.aux-services	mapreduce_shuffle,spark2_shuffle	shuffle服务的列表
yarn.nodemanager.aux-services.mapreduce_shuffle.class	org.apache.hadoop.mapred.ShuffleHandler	MR的shuffle配置
yarn.nodemanager.aux-services.spark2_shuffle.class	org.apache.spark.network.yarn.YarnShuffleService
yarn.nodemanager.aux-services.spark2_shuffle.classpath	/usr/hdp/${hdp.version}/spark2/aux/*
yarn.nodemanager.aux-services.spark_shuffle.class	org.apache.spark.network.yarn.YarnShuffleService
yarn.nodemanager.aux-services.spark_shuffle.classpath	/usr/hdp/${hdp.version}/spark/aux/*
yarn.nodemanager.bind-host	0.0.0.0	绑定的主机
yarn.nodemanager.container-executor.class	org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor	谁将执行（启动）容器
yarn.nodemanager.container-metrics.unregister-delay-ms	60000	完成后取消注册容器指标的延迟时间ms。默认值是10000
yarn.nodemanager.container-monitor.interval-ms	3000	检查的间隔时间默认值3000ms
yarn.nodemanager.delete.debug-delay-sec	0	本配置就是在程序运行完后是否立即删除所缓存的配置文件以及环境信息，默认值0
yarn.nodemanager.disk-health-checker.max-disk-utilization-per-disk-percentage	90	配置磁盘的最大使用比例，默认值90
yarn.nodemanager.disk-health-checker.min-free-space-per-disk-mb	1000	磁盘上必须使用的最小空间，默认值0
yarn.nodemanager.disk-health-checker.min-healthy-disks	0.25	启动新容器的最佳磁盘数量的最小比例，默认值0.25
yarn.nodemanager.health-checker.interval-ms	135000	运行节点运行状况脚本的频率，默认值600000
yarn.nodemanager.health-checker.script.timeout-ms	60000	脚本超时期限，默认值1200000
yarn.nodemanager.kill-escape.launch-command-line	slider-agent,LLAP
yarn.nodemanager.kill-escape.user	hive
yarn.nodemanager.linux-container-executor.cgroups.strict-resource-usage	false	此标志确定应用程序是应以严格的资源限制运行，还是允许在需要时使用备用资源，打开此值会降低吞吐量
yarn.nodemanager.linux-container-executor.group	hadoop
yarn.nodemanager.local-dirs	/data/hadoop/yarn/local	在运行时对资源的一个本地化路径
yarn.nodemanager.log-aggregation.compression-type	gz	日志的压缩方式
yarn.nodemanager.log-aggregation.debug-enabled	false	是否开启dubug的日志
yarn.nodemanager.log-aggregation.num-log-files-per-app	336	定义远程文件系统中每个NM可以拥有的每个应用程序的聚合日志文件数，默认值30
yarn.nodemanager.log-aggregation.roll-monitoring-interval-seconds	3600	上传聚合日志的时间间隔，默认值3600
yarn.nodemanager.log-dirs	/data/hadoop/yarn/log	本地存储日志的目录
yarn.nodemanager.log.retain-seconds	604800	保留用户日志的时间（以秒为单位）仅适用于禁用日志聚合的情况
yarn.nodemanager.recovery.dir	/data/var/log/hadoop-yarn/nodemanager/recovery-state	启用恢复时节点管理器将在其中存储状态的本地文件系统目录
yarn.nodemanager.recovery.enabled	true	启动后启用节点管理器以进行恢复，默认值flase
yarn.nodemanager.remote-app-log-dir	/app-logs	将日志聚合后的位置
yarn.nodemanager.remote-app-log-dir-suffix	logs	后缀名
yarn.nodemanager.resource.cpu-vcores	24	设置可用CPU的数量
yarn.nodemanager.resource.detect-hardware-capabilities	false	是否启用节点的自检功能，当内存与CPU配置为-1时将通过自检的方式获取
yarn.nodemanager.resource.memory-mb	57344
yarn.nodemanager.resource.percentage-physical-cpu-limit	100	可以为容器分配的比例
yarn.nodemanager.vmem-check-enabled	false	是否将对容器强制实施虚拟内存限制，默认值是ture
yarn.nodemanager.vmem-pmem-ratio	2.1	虚拟内存与分配的物理内存的比例
yarn.resourcemanager.address	node:8050
yarn.resourcemanager.admin.address	node:8141	RM管理界面的地址
yarn.resourcemanager.am.max-attempts	2	AM全局重试次数，默认值2
yarn.resourcemanager.bind-host	0.0.0.0
yarn.resourcemanager.connect.max-wait.ms	-1	等待与ResourceManager建立连接的最长时间
yarn.resourcemanager.connect.retry-interval.ms	15000	间隔时间，默认值30000
yarn.resourcemanager.fs.state-store.retry-policy-spec	2000, 500
yarn.resourcemanager.fs.state-store.uri		当使用文件存储时必须设置，对于RM的数据备份机制
yarn.resourcemanager.ha.enabled	false	是否启用HA模式
yarn.resourcemanager.hostname	localhost	RM主机地址
yarn.resourcemanager.monitor.capacity.preemption.natural_termination_factor	1
yarn.resourcemanager.monitor.capacity.preemption.total_preemption_per_round	0.13
yarn.resourcemanager.nodes.exclude-path	/etc/hadoop/conf/yarn.exclude	包含要排除的节点的文件路径
yarn.resourcemanager.recovery.enabled	true	yarn.resourcemanager.store.class必须设置。启用RM恢复机制，默认值false
yarn.resourcemanager.resource-tracker.address	node:8025
yarn.resourcemanager.scheduler.address	node:8030
yarn.resourcemanager.scheduler.class	org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler	用作资源调度程序的类
yarn.resourcemanager.scheduler.monitor.enable	false	启用一组影响调度程序的定期监视器，默认值false
yarn.resourcemanager.state-store.max-completed-applications	${yarn.resourcemanager.max-completed-applications}	保存完成应用的数量，默认值1000
yarn.resourcemanager.store.class	org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
yarn.resourcemanager.system-metrics-publisher.dispatcher.pool-size	10	发送yarn系统指标数据的工作线程数
yarn.resourcemanager.system-metrics-publisher.enabled	true	控制纱线系统指标是否在RM和NM上在时间轴服务上发布的设置，默认值false
yarn.resourcemanager.webapp.address	node1:8088
yarn.resourcemanager.webapp.delegation-token-auth-filter.enabled	false	一种身份验证策略，默认true
yarn.resourcemanager.webapp.https.address	node:8090
yarn.resourcemanager.work-preserving-recovery.enabled	true	启用RM工作以保留恢复。此配置对YARN是专用的，用于试验该功能
yarn.resourcemanager.work-preserving-recovery.scheduling-wait-ms	10000	设置RM在为工作保留 - 恢复分配新容器之前等待的时间。在将新容器分配给应用程序之前，这样的等待期使RM有机会在恢复时解决与集群中的NM重新同步的问题
yarn.resourcemanager.zk-acl	world:anyone:rwcda
yarn.resourcemanager.zk-address	node:2181	zk的连接地址
yarn.resourcemanager.zk-num-retries	1000
yarn.resourcemanager.zk-retry-interval-ms	1000
yarn.resourcemanager.zk-state-store.parent-path	/rmstore
yarn.resourcemanager.zk-timeout-ms	10000
yarn.scheduler.capacity.ordering-policy.priority-utilization.underutilized-preemption.enabled	false
yarn.scheduler.maximum-allocation-mb	57344	RM中每个容器请求的最大分配，以MB为单位
yarn.scheduler.maximum-allocation-vcores	12
yarn.scheduler.minimum-allocation-mb	1024	RM中每个容器请求的最小分配（MB）。低于此值的内存请求将设置为此属性的值。
yarn.scheduler.minimum-allocation-vcores	1
yarn.timeline-service.address	node:10200
yarn.timeline-service.bind-host	0.0.0.0
yarn.timeline-service.client.fd-flush-interval-secs	5
yarn.timeline-service.client.max-retries	30
yarn.timeline-service.client.retry-interval-ms	1000
yarn.timeline-service.enabled	true
yarn.timeline-service.entity-group-fs-store.active-dir	/ats/active/
yarn.timeline-service.entity-group-fs-store.app-cache-size	10
yarn.timeline-service.entity-group-fs-store.cleaner-interval-seconds	3600
yarn.timeline-service.entity-group-fs-store.done-dir	/ats/done/
yarn.timeline-service.entity-group-fs-store.group-id-plugin-classes	org.apache.tez.dag.history.logging.ats.TimelineCachePluginImpl
yarn.timeline-service.entity-group-fs-store.group-id-plugin-classpath
yarn.timeline-service.entity-group-fs-store.retain-seconds	604800
yarn.timeline-service.entity-group-fs-store.scan-interval-seconds	15
yarn.timeline-service.entity-group-fs-store.summary-store	org.apache.hadoop.yarn.server.timeline.RollingLevelDBTimelineStore
yarn.timeline-service.generic-application-history.store-class	org.apache.hadoop.yarn.server.applicationhistoryservice.NullApplicationHistoryStore
yarn.timeline-service.http-authentication.proxyuser.root.groups	*
yarn.timeline-service.http-authentication.proxyuser.root.hosts	node
yarn.timeline-service.http-authentication.simple.anonymous.allowed	true
yarn.timeline-service.http-authentication.type	simple
yarn.timeline-service.leveldb-state-store.path	/hadoop/yarn/timeline
yarn.timeline-service.leveldb-timeline-store.path	/hadoop/yarn/timeline
yarn.timeline-service.leveldb-timeline-store.read-cache-size	104857600
yarn.timeline-service.leveldb-timeline-store.start-time-read-cache-size	10000
yarn.timeline-service.leveldb-timeline-store.start-time-write-cache-size	10000
yarn.timeline-service.leveldb-timeline-store.ttl-interval-ms	300000
yarn.timeline-service.recovery.enabled	true
yarn.timeline-service.state-store-class	org.apache.hadoop.yarn.server.timeline.recovery.LeveldbTimelineStateStore
yarn.timeline-service.store-class	org.apache.hadoop.yarn.server.timeline.EntityGroupFSTimelineStore
yarn.timeline-service.ttl-enable	true
yarn.timeline-service.ttl-ms	2678400000
yarn.timeline-service.version	1.5
yarn.timeline-service.webapp.address	node:8188
yarn.timeline-service.webapp.https.address	node:8190

注：关于时间线的不做过多介绍。