Hello Currently i'm facing the following problems. 1. i install apache ambari 2.7.3 in 1 node machine (which serve as both master and slave) 2. without gpu enabled, everything work file . you can see the following pdf for detail Ambari - gpu01_good.pdf 3. but after i enable gpu support through yarn config. yarn omit 3 error messages as follow Ambari - gpu01_bad.pdf # NodeManager Health Connection failed to http://gpu01:8042 (<urlopen error [Errno 111] Connection refused>) # NodeManager Web UI Connection failed to http://gpu01:8042/ws/v1/node/info (Traceback (most recent call last):   File "/var/lib/ambari-agent/cache/stacks/HDP/3.0/services/YARN/package/alerts/alert_nodemanager_health.py", line 171, in execute     url_response = urllib2.urlopen(query, timeout=connection_timeout)   File "/usr/lib/python2.7/urllib2.py", line 154, in urlopen     return opener.open(url, data, timeout)   File "/usr/lib/python2.7/urllib2.py", line 429, in open     response = self._open(req, data)   File "/usr/lib/python2.7/urllib2.py", line 447, in _open     '_open', req)   File "/usr/lib/python2.7/urllib2.py", line 407, in _call_chain     result = func(*args)   File "/usr/lib/python2.7/urllib2.py", line 1228, in http_open     return self.do_open(httplib.HTTPConnection, req)   File "/usr/lib/python2.7/urllib2.py", line 1198, in do_open     raise URLError(err) URLError: <urlopen error [Errno 111] Connection refused> ) # Percent NodeManagers Available affected: [1], total: [1] The message say that port 8042 is refused but a quick check show that there isn't any process that occupy port 8042 Please help me. hadoop@gpu01:~$ sudo netstat -tnlpa | grep 8042 [sudo] password for hadoop:  hadoop@gpu01:~$  ... View more