Please help! - ORACLE Open database fails

[Benoît]

SQL> startup
ORACLE instance started.

Total System Global Area 727674832 bytes
Fixed Size 737232 bytes
Variable Size 452984832 bytes
Database Buffers 268435456 bytes
Redo Buffers 5517312 bytes
Database mounted.
ORA-01122: database file 2 failed verification check
ORA-01110: data file 2: '/oracle/oradata/SID/SID_raw_undo'
ORA-01251: Unknown File Header Version read for file number 2

I guess it's possible to drop this datafile in some way ...

Thanks

Ben

[Benoît]

I succeeded to do this:

SQL> ALTER DATABASE DATAFILE '/oracle/oradata/SMCPP/SMCPP_raw_undo' offline;

Database altered.

SQL> alter database open;

Database altered.

SQL> shutdown immediate
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> startup
ORACLE instance started.

Total System Global Area 727674832 bytes
Fixed Size 737232 bytes
Variable Size 452984832 bytes
Database Buffers 268435456 bytes
Redo Buffers 5517312 bytes
Database mounted.
Database opened.


But my undo is no OFFLINE and I can't put it ONLINE.
What to do next?

[Patrick Wallek]

The key is these 3 lines:

ORA-01122: database file 2 failed verification check
ORA-01110: data file 2: '/oracle/oradata/SID/SID_raw_undo'
ORA-01251: Unknown File Header Version read for file number 2

It appears that you have a corrupt file. I would place a call with Oracle, if you have support, and enlist their help.

[Benoît]

I've just created a new undo tablespace:

CREATE UNDO TABLESPACE UNDOTBS2 DATAFILE '/oracle/oradata/SID/undotbs2.dbf' SIZE 50M;
ALTER SYSTEM SET UNDO_TABLESPACE=undotbs2;

Drop the corrupted UNDO tablespace.

DROP TABLESPACE UNDOTBS;

Then stop and restart the database.
Recreate the original UNDO tablespace.

CREATE UNDO TABLESPACE UNDOTBS DATAFILE '/oracle/oradata/SID/SID_raw_undo' SIZE 1014M;
ALTER SYSTEM SET UNDO_TABLESPACE=undotbs;

Drop the temporary new UNDO tablespace.

ALTER DATABASE DATAFILE '/oracle/oradata/SID/undotbs2.dbf' OFFLINE;
DROP TABLESPACE undotbs2;

Sorry disturbing with a problem I could solve myself ...
This was just urgent.

Thanks.

[Joaquin_13]

Yup!! the key there next time is ones or if there is an instance that one of your undo dbf file is corrupted you must never shutdown your DB cause if you did i'm sure the next time you start your instance then you will have to start recovering your DB. Your solution was right . First you must turn the corrupted undo file to offline then create another group for your undo tablespace then that is the time you can perform DB bounce.


regards,

Joaquin

[Benoît]

Solution on the thread.

Thanks for help.