Recovery error....

Question

Hi sir,
I have encountered a very serious problem in the recovery. I have been sent a mail from a friend who is having a DSS system without a backup and in NoARch mode. I agree that its not at all a right thing but that’s what is in there.
3 days prior, one of the DSS db tablespace having 12 files(total of 75g) was overwritten when the dba was trying to move the files from one partition to another. After that whole tablespace went offline and also for the datafile it said, header failed verification. Now the task that was assigned is that lose that datafile and its data but bring the tablespace and rest 11 files back. That’s what I have done on a test box.I am listing the commands here.
1. Create a tablespace with 3 datafiles.
Two datafiles share the same name but on different partition/folder say TEST_TS02.dbf,
one lying on 'c:\oracle\oradata\<DB>\test_ts02.dbf' size 10M and other lying at 'c:\testing\test_ts02.dbf' size 20M
2. Make the Tablespace Offline
“alter tablespace teset_ts offline;”
3. Apply log switches/commits/checkpints
To replicate the working of the real db as its also having some DML’s going on.So I am moving ahead the checkpoint and SCN nos.
4. replace the file at 'c:\testing\test_ts02.dbf' with 'c:\oracle\oradata\<DB>\test_ts02.dbf' by copy paste
5. Try to make the tablespace online
“alter tablespace teset_ts online;”
It gives error --> Header version mismatch for 'c:\testing\test_ts02.dbf'.
6. Now Create a datafile using alter database create datafile option
“alter database create datafile 'c:\testing\test_ts02.dbf' as 'c:\oracle\oradata\<DB>\test_new.dbf';”
This was done to mmake sure that file header is rebuilt by create command.
7. Apply recover command for datafile
“recover datafile 'c:\oracle\oradata\<DB>\test_new.dbf';”
This command failed asking for exact sequence/change number.
Now when this failed, I tried to give recover database until cancel.After that I opened the database with Resetlogs but when I tried to make tablespace online it said to me ORA-1190, either datafile or controlfile are before the resetlog.Why its not accepting the file that I had made? And the file was made before the restlog was issued.I can assume this error if I had used some backup copy or some thing but I didn’t.
I also tried to recreate the controlfile.I got the “missing00n” file message.I tried to rename the file with the existing copy too but it again gave me the same error.I am not able to think anything more about.
Please guide me what can be done just to bring the tablspace and rest 11 files online. The 12th datafile and its data which will be lost, is bearable.
Thanks in advance and best regards
Aman….

Answer 1

Once u offline drop the table u have to use the following mechanism to recover that for those tables which had a data in the datafile which is lost.
1.) Find the "data object id" of the table:
select data_object_id,object_name,owner from dba_objects
where object_name = ' < CORRUPTED TABLE NAME >' and owner = '< OWNER >';
2.) Find the relative_file_id of the table that has got dropped (dba_data_files).
3.) Find all the datafiles involved in the table:
select distinct file_id from dba_extents
where segment_name=< CORRUPTED TABLE NAME > and owner= < Owner >;
4.) Calculate the low rowid of this table for the corrupted datafile:
SELECT dbms_rowid.rowid_create(1,< DATA_OBJECT_ID >,< FILE ID >,0,0) Low_Row_ID from DUAL;
5.) Find the max block_id+blocks-1 for the table.
Select max(block_id+blocks-1) MAX_EXT from dba_extents where SEGMENT_NAME=< Table_name > and owner= < Owner >;
6.) Found the high rowid of this table for the corrupted datafile:
SELECT dbms_rowid.rowid_create(1,< DATA_OBJECT_ID >,< FILE ID >,< MAX_EXT from previous > ,0) High_Row_ID from DUAL;
7.) Create Table As (CTAS) using these 2 limits:
create table salvage_tab_new as select /*+ ROWID(A) */ * from
< Corrupted Table Name > A where rowid > 'High_Row_id>' and rowid < 'Low_Row_id>' ;