I have spent a few very confusing hours in desperate attempts to figure out why OMS instance won't start. The command output suggests checking the NodeManager log file. But it has no new entries, in fact, the file wasn't accessed at all.

emctl start oms
(out)Oracle Enterprise Manager Cloud Control 12c Release 5
(out)Copyright (c) 1996, 2015 Oracle Corporation.  All rights reserved.
(out)Starting Oracle Management Server...
(out)Starting WebTier...
(out)WebTier Successfully Started
(out)Node Manager Could Not Be Started
(out)Check Node Manager log file for details: /u01/app/oracle/gc_inst/NodeManager/emnodemanager/nodemanager.log
(out)Oracle Management Server is Down
(out)Starting BI Publisher Server ...
(out)Node Manager Could Not Be Started
(out)Check Node Manager log file for details: /u01/app/oracle/gc_inst/NodeManager/emnodemanager/nodemanager.log
(out)BI Publisher Server is Down

It's always about the network, you know: typos in /etc/hosts, DNS record suddenly turns to a different IP address, wrong name resolution order. I check them all. Everything was neat and clean, plus you would have a hint in the logs, at least a few new lines. From now on, I have another check mark in my validation list - the  LogLevel property.  Make sure that parameter value matches to the one of the list1 :

  • SEVERE (highest value)
  • INFO
  • FINE
  • FINEST (lowest value)

In my case, LogLevel was set to DEBUG. Good enough to do not drag my attention and bad enough to kill NodeManager even before it opens a log file.

  1. https://docs.oracle.com/javase/7/docs/api/java/util/logging/Level.html