HDP ResourceManager启动失败，后台也无进程

原因：手动重启过RM，或者因为其他原因导致RM后，Ambari未监控到

报错信息

stderr: 
Traceback (most recent call last):
  File "/var/lib/ambari-agent/cache/stacks/HDP/3.0/services/YARN/package/scripts/resourcemanager.py", line 261, in 
    Resourcemanager().execute()
  File "/usr/lib/ambari-agent/lib/resource_management/libraries/script/script.py", line 352, in execute
    method(env)
  File "/usr/lib/ambari-agent/lib/resource_management/libraries/script/script.py", line 996, in restart
    raise Fail("Stop command finished but process keep running.")
resource_management.core.exceptions.Fail: Stop command finished but process keep running.
 stdout:

进入ResourceManager记录进程号的目录

/tmp/hsperfdata_yarn

此时目录中肯定是多个进程号文件，导致RM启动时不知道找哪个启动

故障解除：删掉失效的进程号文件，或者删除所有RM进程号文件，重启RM服务