Recovery- problem

[badrinathn]

Hi,
I was trying to test out few things with the datafiles on the test database.
THe database is not in the archiving mode. As I have one of the tablespaces with only one datafile and only one table in that.

As there are no activites going on that table,

I did the following:

1) When the database was running, I copied the datafile issuing
copy ora_data01.dbf ora_data01_bkp.dbf

2)I shutdown the database
shutdown immediate

3) I deleted the datafile ora_data01.dbf and copied back ora_data01_bkp.dbf
ie del ora_data01.dbf
move ora_data01_bkp.dbf ora_data01.dbf

4) Tried to restart the database

I get the following
ORA-01588: MUST USE RESETLOGS...

SVRMGR> ALTER DATABASE OPEN RESETLOGS;
ora-01152:file 1 was not restrored from a sufficiently old backup
0ra-01110 data file 1: 'c:\oracle\oradata\orcl\system01.dbf'

How can I recover this database now

[tamilselvan]

When the instance is running in NOARCHIVE mode, you can't copy (backup) the data file. The data file will be in inconsistent state.

[badrinathn]

So does it mean, I cannot bring up my database.
badrinath

[pipo]

the problem is that your control files are more recent than your datafile ... if you do not have the archived redo logs, I do not see how you could reopen your database ...
maybe (i am not sure at all of what it will make) you should try to recreate a controle file (here is an example of script) :

CREATE CONTROLFILE REUSE set DATABASE "GNRX" RESETLOGS ARCHIVELOG
MAXLOGFILES 16
MAXLOGMEMBERS 2
MAXDATAFILES 80
MAXINSTANCES 1
MAXLOGHISTORY 100
LOGFILE
GROUP 1 '/data/GNRX/log1GNRX.dbf' SIZE 20M,
...
GROUP 4 '/data/GNRX/log4GNRX.dbf' SIZE 20M
DATAFILE
'/data/GNRX/dbsGNRX.dbf',
'/data/GNRX/datGNRX2.dbf',
...
'/data/GNRX/idxGNRX31.dbf'
;


then you
1) startup nomount
2) play the script above
3) alter database open resetlogs
4) shutdown / startup to make it clean ...

[rsuri]

That's correct. You need to restore from a full backup.
Seems like you are on Windows environment (c:\oracle\...)
How could you delete a datafile when the database was running. NT will not let you do that. Just curious.

By the way the correct way to copy the datafile is to run in Archivelog mode and then bring the tablespace in backupmode and copy the datafiles underneath.

- Rajeev

[pipo]

note to rsuri :
the file was just copied while the DB was running, it was deleted after the DB was stopped

[badrinathn]

Thanks
Are the logfile names and the datafile names given here
should be the name of the existing redo*.log files and datafiles.

I have the following files in my c:\oracle\oradata\orcl

control01.ctl
control02.ctl
control03.ctl
dr01.dbf
indx01.dbf
orcl_data1.dbf ( this was the datafile which was copied back)
orcl_data2.dbf
rbs01.dbf
redo01.log
redo03.log
system01.dbf
temp01.dbf
tools01.dbf and users01.dbf

[rsuri]

duh...Thanks pipo !!!

[padmam]

Hi, 20th April 2001 20:22 hrs chennai

The reason may be due to this.

When you tried to copy the datafile while it was running the
the datafile header would have had CHECKPOINT_CHANGE#
at that particular point in sync with control file header CHECKPOINT_CHANGE# .

while you shutdown the DB CHECKPOINT_CHANGE# would had been updated or chaged in the control file headers as well as in the datafile header.

Now you are trying to do a incomplete recovery by restoring from the back up datafile with different CHECKPOINT_CHANGE# so its asking for a reset logs.

Since your DB is running in NoArchive log mode you had the problem since the incomplete recovery cannot be done with NoArchive log mode.

if you have a full DB back up restore to work.

If i am wrong please correct me.

Cheers

Padmam

[padmam]

Hi PIPO, 20th april 2001 20:36 hrs chennai

Even if recreate the control file it cannot be of any use.

Because the DB is in NOArchive mode.

There is no incomplete recovery at all.

The only way is restore from the last backup.

cheers

padmam