APPLIES TO:

SYMPTOMS

The following symptoms are known to be attributed to this issue :-

a) Database is 2 or more node RAC

b) The following error or similar is seen in the database alert log(s) :-

ORA-00600: internal error code, arguments: [ORA_NPI_ERROR], [603], [ORA-00603: ORACLE server session terminated by fatal error
ORA-00569: Failed to acquire global enqueue.
Process ID: 
Session ID: 105 Serial number: 1

c) The database is managed by ASM instance(s) and ORA-04031 errors are observed in the ASM alert log(s) for example on the LMD0 process :-

ORA-04031: unable to allocate 3768 bytes of shared memory ("shared pool","unknown object","sga heap(1,0)","ges enqueues") 

If the same ORA-600 error is seen on a non-RAC/ASM managed database this article might still be appropriate, if the same ORA-600 error is

seen on a non-ASM managed database this article does not offer a solution.  The ORA-04031 error might be on a different process and might

have a different failed allocation request.

CAUSE

The issue is correlated to :-

Bug:16502516 - ORA-600 ARGS [ORA_NPI_ERROR], ALONG WITH ORA-00603 & ORA-00569

This BUG is published but is in 'Suspended' status informing that the ORA-04031 errors must be resolved on the ASM side.
 

SOLUTION

Increase the SGA within the ASM instance(s).  The amount of increase is subject to the machine environment so this article cannot give detailled

recommendations on how much of an increase will be needed.