Receiving ORA-00600 [4080] or ORA-600 [4886] Refreshing ON COMMIT Materialized View

Last updated on MAY 28, 2013

Applies to:

Symptoms

When doing an ON-COMMIT refresh on a distributed mview (that is, the mview and master table are in different databases), you get an ORA-600 [4080] error.  In rare cases, you may also see an ORA-600 [4886], although the ORA-600 [4886] can also be caused by issues other than on-commit distributed materialized views.

• You will see something like the following in the alert log: 
  
  
  ORA-00600: internal error code, arguments: [4080], [2], [131], [], [], [], [], []
  Tue Jan 13 23:03:22 2009
  DISTRIB TRAN <instance>.78ae9908.25.47.170457
  is local tran 25.47.170457 (hex=19.2f.299d9)
  insert pending prepared tran, scn=164959181879 (hex=26.68555c37)
  Tue Jan 13 23:03:22 2009
  DISTRIB TRAN <instance>.78ae9908.25.47.170457
  is local tran 25.47.170457 (hex=19.2f.299d9)
  change pending prepared tran, scn=164959181879 (hex=26.68555c3a)
  to pending committed tran, scn=164959181882 (hex=26.68555c3a)
  DISTRIB TRAN <instance>.78ae9908.25.47.170457
  is local tran 25.47.170457 (hex=19.2f.299d9))
  delete pending committed tran, scn=164959181882 (hex=26.68555c3a)
  Tue Jan 13 23:03:25 2009
  Following on-commit snapshots not refreshed :
  <owner>.<mview_name>

• The "Current SQL" in the tracefile may look like this:
  
  
  select log, oldest, oldest_pk, oldest_oid, oldest_new, youngest+1/86400, flag, yscn, oldest_seq from sys.mlog$ where mowner = :1 and master = :2 for update

Cause