High Availability with WebLogic Clusters - Oracle WebLogic Server 12c Advanced Administration - page 46

8. Change the Failure Action drop-down menu to Force immediate shutdown of this
cluster and the Panic Action field to Exit the cluster process (as shown in the
following screenshot):
9. Click on the Save button and then the Activate Changes button.
10. Restart all Managed Servers of the PROD_Cluster cluster.
How it works...
Failure Action and Panic Action are the WebLogic overload settings that can shutdown a
Managed Server when it reaches a FAIL state or when it throws an Out of Memory (OOM)
PANIC error. Both errors would leave the affected Managed Server in an inconsistent state
and would probably hang or return application errors to client requests. The Node Manager
monitors the Managed Servers and restarts the failed instance automatically. Unless there
is an analysis of the root cause of these errors, it's recommended to enable both parameters.
CrashRecoveryEnabled is the Node Manager parameter that must be enabled so Node
Manager can restart a crashed/failed WebLogic Managed Server instance automatically.
The Cluster Address, Number of Servers in Cluster Address, and WebLogic Plug-in Enabled
are cluster configurations for the distribution of requests and load balancing.