0
votes

I have a Weblogic 12c setup running on Windows 2012 RC2. The NodeManager is used to start/stop a managed server.

What works: Starting the managed server through admin console. Also automatic restarting of the managed server through NodeManager after killing the managed server process works. NodeManager, Admin Console and Managed server logs do not show any errors.

What fails: If the Windows machine reboots (hard power cut or simple reboot) the managed server then is in the "FAILED_NOT_RESTARTABLE" state. Also here NodeManager, Admin Console and Managed server logs do not show any errors.

The Oracle documentation clearly states how to set this up, basically it is important to start the managed servers with the "-Xrs" or "-Xnohup" JVM parameters. (https://docs.oracle.com/cd/E24329_01/web.1211/e21050/nodemgr_config.htm#NODEM164)

Nevertheless, the reboot of Windows ends up in the error state all the time. Any ideas?

1
I found that if options for JMX analysis were set in the start-up (Server Start tab -> Arguments:) that this would cause immediate FAILED_NOT_RESTARTABLE state. Removing the JMX arguments resolved the problem I encountered.mikequentel

1 Answers

0
votes

The solution is actually hidden in the documentation linked above: the "-Xrs" is not enough, you also need to change the "nodemanager.properties" file to allow recovery of managed servers after a reboot (by default this is disabled):

CrashRecoveryEnabled=true;

The documentation is just very chaotic at this point mixes "crash" and "reboot": You will need this flag in both cases!

NOTE:

  • Do not forget to check for the little hints hidden in the documentation when running the node manager as a script (you then need additional start parameters to be passed to the node manager shell script)
  • Do not forget to add the "-Xrs" flags to your managed servers

(There is a Oracle support page that describes this scenario also: https://community.oracle.com/thread/726965)