ORA-1092 During Startup Using AUM (Automatic Undo Management)

@Checked for relevance on 30-MAY-2007

Problem Description
-------------------

You start up the database and receive the following error message:

    SQL> startup
    ORACLE instance started.

    Total System Global Area  252478536 bytes
    Fixed Size                   279624 bytes
    Variable Size             167772160 bytes
    Database Buffers           83886080 bytes
    Redo Buffers                 540672 bytes
    Database mounted.

    ORA-01092: ORACLE instance terminated. Disconnection forced

No background processes are started.

ORA-01092: Oracle instance terminated.  Disconnection forced
    Cause: The instance connected to was terminated abnormally, probably
           due to a SHUTDOWN ABORT.  The current process was forced to
           disconnect from the instance.
   Action: Contact the database administrator to determine when the
           instance is restarted.  Attempt to reconnect after the instance
           is running again.

Solution Description
--------------------

In the alert.log file there is additional information on the error context:

    SMON: enabling cache recovery
    Thu Aug  2 14:19:45 2001
    Errors in file /ora/ora901/admin/V901/udump/v901_ora_21863.trc:
    ORA-30012: undo tablespace 'UNDO1' does not exist or of wrong type
    Thu Aug  2 14:19:45 2001
    Error 30012 happened during db open, shutting down database
    USER: terminating instance due to error 30012
    Instance terminated by USER, pid = 21863
    ORA-1092 signalled during: ALTER DATABASE OPEN...

You have three options to resolve this issue:

1.  If an UNDO tablespace currently exists in the database, edit the
    init<SID>.ora file to set the UNDO_TABLESPACE parameter to the name of the
    existing UNDO tablespace and restart the database.

- OR -

2.  If no UNDO tablespace exists, edit the init<SID>.ora file to comment out
    the UNDO_MANAGEMENT parameter: the default value for UNDO_MANAGEMENT is
    MANUAL.  

    Then, after starting the database, do one of the following:

        A.  Create a non UNDO tablespace with at least one online rollback
            segment.
            The instance can work this way.

        -OR-

        B.  Create a new UNDO tablespace.
            If you want the instance to work under Automatic Undo Management,
            a. Set the UNDO_MANAGEMENT back to AUTO
            b. Set the UNDO_TABLESPACE parameter to the name of the new UNDO
               tablespace
            c. Restart the database

Explanation
-----------

The database is using Automatic Undo Management (AUM).  In the init<SID>.ora
file, the parameter UNDO_MANAGEMENT is set to AUTO and UNDO_TABLESPACE to a
tablespace that does not exist or is not an UNDO type tablespace. 

References
----------

Note 135090.1  Managing Rollback/Undo Segments in AUM (Automatic Undo
                 Management)
Note 150216.1  9i: ORA-01552 Though Non-System Rollback Segments can be
                 Onlined
@ Bug 1783809  DEFAULT NAME FOR UNDO TABLESPACE SHOULD BE TAKEN FROM 
@                PARAMETER UNDO_TABLESPACE

Search Words
------------

ORA-1092

© 2008 – 2011, www.oracledatabase12g.com. 版权所有.文章允许转载,但必须以链接方式注明源地址,否则追究法律责任.

相关文章 | Related posts:

1. AUTOMATIC UNDO INTERNALS
2. EVENT:10226 trace CR applications of undo for data
3. Cheat Sheet for UNDO/Rollback problems
4. EVENT:10218 dump uba of applied undo
5. DIAG: Alter System Dump Undo – Quick Reference
6. dump logfile prints minimal info for KTSFUNLK undo record
7. EVENT: 10015 "Write Trace for Undo Segment Recovery"
8. Automatic Checkpoint Tuning
9. Oracle9i:Recovering lost data through Flashback Query