HDP 2.1, Ambari 1.6.1, CentOS 6.5, multi-node cluster
On a fresh install of HDP 2.1, not all metrics are getting populated in Ambari. Some of the missing metrics include:
- Dashboard: Namenode CPU WIO
- HDFS: All except HDFS I/O
- YARN: All
None of the ideas suggested in the following two threads have helped resolve the issue:
- http://hortonworks.com/community/forums/topic/ambari-1-4-2-no-data-there-was-not-data-available
- http://mail-archives.apache.org/mod_mbox/ambari-user/201406.mbox/%3CCACDnAuZq=+xoUmEGAPOzcgcYJPVdJTDyjznJ8tBpS-h8b=JUmg@mail.gmail.com%3E
ambari-server.log shows the following warnings and errors (/var/log/messages has no Ganglia specific errors):
16:12:52,905 WARN [Thread-1] HeartbeatMonitor:126 - Heartbeat lost from host localhost.localdomain 16:13:28,435 ERROR [pool-1-thread-6163] JMXPropertyProvider:540 - Caught exception getting JMX metrics : Server returned HTTP response code: 500 for URL: http://namenode.localdomain:8745/api/cluster/summary
Any suggestions on how to identify and resolve the cause for this issue ?
Thanks.