ITPub博客
首页 博客 专栏 学院 社区 IT新媒体评选
发博

首页 > 大数据 > Hadoop > Hadoop2.7实战v1.0之YARN HA

Hadoop2.7实战v1.0之YARN HA

原创 Hadoop 作者：hackeruncle 时间：2016-03-06 18:10:24 0 删除 编辑

YARN HA实战v1.0

当前环境:hadoop+zookeeper(namenode,resourcemanager HA)   

              resourcemanager

serviceId

init status

sht-sgmhadoopnn-01

rm1

active

sht-sgmhadoopnn-02

rm2

standby

参考:
http://blog.csdn.net/u011414200/article/details/50336735

http://blog.csdn.net/u011414200/article/details/50276257

一.查看resourcemanager是active还是standby

1.打开网页

http://172.16.101.55:8088/cluster/cluster

http://172.16.101.56:8088/cluster/cluster

 

2.查看resourcemanager日志

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-01 logs]# more yarn-root-resourcemanager-sht-sgmhadoopnn-01.telenav.cn.log
  2. …………………..
  3. 2016-03-03 18:10:01,289 INFO org.apache.hadoop.yarn.server.resourcemanager.ResourceManager: Transitioned to active state

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-02 logs]# more yarn-root-resourcemanager-sht-sgmhadoopnn-02.telenav.cn.log
  2. …………………..
  3. 2016-03-03 18:10:34,250 INFO org.apache.hadoop.yarn.server.resourcemanager.metrics.SystemMetricsPublisher: YARN system metrics publishing service is not enabled
  4. 2016-03-03 18:10:34,250 INFO org.apache.hadoop.yarn.server.resourcemanager.ResourceManager: Transitioning to standby state

3. yarn rmadmin –getServiceState 


点击(此处)折叠或打开

  1. ###$HADOOP_HOME/etc/hadoop/yarn-site.xml
  2.               <property>
  3.         <name>yarn.resourcemanager.ha.rm-ids</name>
  4.         <value>rm1,rm2</value>
  5.     </property>


  8. [root@sht-sgmhadoopnn-02 logs]# yarn rmadmin -getServiceState rm1
  9. active
  10. [root@sht-sgmhadoopnn-02 logs]# yarn rmadmin -getServiceState rm2
  11. standby
二.命令

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-01 bin]# yarn --help
  2. Usage: yarn [--config confdir] [COMMAND | CLASSNAME]
  3.   CLASSNAME run the class named CLASSNAME
  4.  or
  5.   where COMMAND is one of:
  6.   resourcemanager -format-state-store deletes the RMStateStore
  7.   resourcemanager run the ResourceManager
  8.   nodemanager run a nodemanager on each slave
  9.   timelineserver run the timeline server
  10.   rmadmin admin tools
  11.   sharedcachemanager run the SharedCacheManager daemon
  12.   scmadmin SharedCacheManager admin tools
  13.   version print the version
  14.   jar <jar> run a jar file
  15.   application prints application(s)
  16.                                         report/kill application
  17.   applicationattempt prints applicationattempt(s)
  18.                                         report
  19.   container prints container(s) report
  20.   node prints node report(s)
  21.   queue prints queue information
  22.   logs dump container logs
  23.   classpath prints the class path needed to
  24.                                         get the Hadoop jar and the
  25.                                         required libraries
  26.   cluster prints cluster information
  27.   daemonlog get/set the log level for each
  28.                                         daemon

  30. ###########################################################################
  31. [root@sht-sgmhadoopnn-01 bin]# yarn rmadmin --help
  32. -help: Unknown command
  33. Usage: yarn rmadmin
  34.    -refreshQueues
  35.    -refreshNodes
  36.    -refreshSuperUserGroupsConfiguration
  37.    -refreshUserToGroupsMappings
  38.    -refreshAdminAcls
  39.    -refreshServiceAcl
  40.    -getGroups [username]
  41.    -addToClusterNodeLabels [label1,label2,label3] (label splitted by ",")
  42.    -removeFromClusterNodeLabels [label1,label2,label3] (label splitted by ",")
  43.    -replaceLabelsOnNode [node1[:port]=label1,label2 node2[:port]=label1,label2]
  44.    -directlyAccessNodeLabelStore
  45.    -transitionToActive [--forceactive] <serviceId>
  46.    -transitionToStandby <serviceId>
  47.    -failover [--forcefence] [--forceactive] <serviceId> <serviceId>
  48.    -getServiceState <serviceId>
  49.    -checkHealth <serviceId>
  50.    -help [cmd]
transitionToActive和transitionToStandby是用于在不同状态之间切换的。

failover 初始化一个故障恢复。该命令会从一个失效的resourcemanager切换到另一个上面(不支持在自动切换的环境操作)。

getServiceState 获取当前resourcemanager的状态。

checkHealth 检查resourcemanager的状态。正常就返回0，否则返回非0值。

三.实验

1.测试YARN的手工切换功能(失败)

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-01 ~]# yarn rmadmin -failover --forceactive rm1 rm2
  2. forcefence and forceactive flags not supported with auto-failover enabled.

#yarn-site.xml 中设置yarn.resourcemanager.ha.automatic-failover.enabled为 true,故提示不能手动切换

2.测试YARN的自动切换功能(成功)

 a.在active resoucemanager机器上通过jps命令查找到resoucemanager的进程号，然后通过kill -9的方式杀掉进程，观察另一个resoucemanager节点是否会从状态standby变成active状态

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-01 ~]# yarn rmadmin -getServiceState rm1
  2. active
  3. [root@sht-sgmhadoopnn-01 ~]# yarn rmadmin -getServiceState rm2
  4. standby

  6. [root@sht-sgmhadoopnn-01 bin]# jps
  7. 2583 Jps
  8. 10162 DFSZKFailoverController
  9. 28432 ResourceManager
  10. 21679 NameNode

  12. [root@sht-sgmhadoopnn-01 ~]# kill -9 28432

  14. [root@sht-sgmhadoopnn-02 bin]# jps
  15. 19147 ResourceManager
  16. 17837 NameNode
  17. 17970 DFSZKFailoverController
  18. 27330 Jps


  21. [root@sht-sgmhadoopnn-01 bin]# yarn rmadmin -getServiceState rm1
  22. 16/03/03 19:23:39 INFO ipc.Client: Retrying connect to server: sht-sgmhadoopnn-01/172.16.101.55:8033. Already tried 0 time(s); retry policy is RetryUpToMaximumCountWithFixedSleep(maxRetries=1, sleepTime=1000 MILLISECONDS)
  23. Operation failed: Call From sht-sgmhadoopnn-01.telenav.cn/172.16.101.55 to sht-sgmhadoopnn-01:8033 failed on connection exception: java.net.ConnectException: Connection refused; For more details see: http://wiki.apache.org/hadoop/ConnectionRefused



  27. [root@sht-sgmhadoopnn-01 bin]# yarn rmadmin -getServiceState rm2
  28. active
  29. [root@sht-sgmhadoopnn-01 bin]#

#### sht-sgmhadoopnn-01 机器上resourcemanager进程已经起来,且状态为standby

  c. 再次切换

点击(此处)折叠或打开

  1. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -transitionToStandby rm2
  2. Automatic failover is enabled for org.apache.hadoop.yarn.client.RMHAServiceTarget@11f69937
  3. Refusing to manually manage HA state, since it may cause
  4. a split-brain scenario or other incorrect state.
  5. If you are very sure you know what you are doing, please
  6. specify the --forcemanual flag.

  8. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -transitionToStandby --forcemanual rm2
  9. You have specified the --forcemanual flag. This flag is dangerous, as it can induce a split-brain scenario that WILL CORRUPT your HDFS namespace, possibly irrecoverably.

  11. It is recommended not to use this flag, but instead to shut down the cluster and disable automatic failover if you prefer to manually manage your HA state.

  13. You may abort safely by answering 'n' or hitting ^C now.

  15. Are you sure you want to continue? (Y or N) Y
  16. 16/03/03 19:29:36 WARN ha.HAAdmin: Proceeding with manual HA state management even though
  17. automatic failover is enabled for org.apache.hadoop.yarn.client.RMHAServiceTarget@4e33967b


  20. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -getServiceState rm1
  21. standby
  22. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -getServiceState rm2
  23. standby

  25. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -transitionToActive rm1
  26. Automatic failover is enabled for org.apache.hadoop.yarn.client.RMHAServiceTarget@54c4f317
  27. Refusing to manually manage HA state, since it may cause
  28. a split-brain scenario or other incorrect state.
  29. If you are very sure you know what you are doing, please
  30. specify the --forcemanual flag.

  32. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -transitionToActive --forcemanual rm1
  33. You have specified the --forcemanual flag. This flag is dangerous, as it can induce a split-brain scenario that WILL CORRUPT your HDFS namespace, possibly irrecoverably.

  35. It is recommended not to use this flag, but instead to shut down the cluster and disable automatic failover if you prefer to manually manage your HA state.

  37. You may abort safely by answering 'n' or hitting ^C now.

  39. Are you sure you want to continue? (Y or N) Y
  40. 16/03/03 19:32:46 WARN ha.HAAdmin: Proceeding with manual HA state management even though
  41. automatic failover is enabled for org.apache.hadoop.yarn.client.RMHAServiceTarget@54c4f317

  43.  [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -getServiceState rm1
  44. 16/03/03 19:32:55 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
  45. active
  46. [root@sht-sgmhadoopnn-01 sbin]# yarn rmadmin -getServiceState rm2
  47. 16/03/03 19:33:02 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
  48. standby
  49. [root@sht-sgmhadoopnn-01 sbin]#

#和HDFS HA切换实验不一样, -transitionToStandby,会自动将standby—>active,active-->standby;

而YARN HA就不一样，需要还有手动再执行一下-transitionToActive.

来自 “ ITPUB博客 ” ，链接：https://blog.itpub.net/30089851/viewspace-2047918/ ，如需转载，请注明出处，否则将追究法律责任。

0
0
上一篇： Hadoop2.7实战v1.0之HDFS HA
下一篇： Hadoop2.x运维实战之入门手册v1.0
请登录后发表评论 登录
全部评论
hackeruncle
专家姓名：武基鹏
主要从事大数据平台产品的技术工作； 负责基于Spark、Flink生产开发及性能调优； 负责对公司的集群的部署、监控、调优和运营； 主导我司的数据平台0到1及实时中间件项目开发； 此外,精通Java、Shell、Python编程； 且荣获2016年IBM的Spark大赛全球第四名，中国区域第一名。

注册时间：2015-01-03

  • 博文量
    187
  • 访问量
    395984

最新文章

支持我们 作者招募 用户协议 FAQ Contact Us 编辑准则

北京盛拓优讯信息技术有限公司. 版权所有  京ICP备16024965号-8 北京市公安局海淀分局网监中心备案编号：11010802021510 niuxiaotong@pcpop.com 17352615567 未成年人举报专区

广播电视节目制作经营许可证(京) 字第1234号 中国互联网协会会员

违法和不良信息举报电话:18101376593