Printable View
Hi:
A tough situation..
I have a db instance just created and was up and running.
oracle 8i, no archive log and no backup yet..
one of the QA guys accidentally ran a create database script..
and i guess messed up a few things..
after doing a shutdown abort, i cant even bring the databse to mount stage.. Is the db a big trash already?? Any chances..
ORA-00205: error in identifying controlfile, check alert log for more info...
What can be done..Any clues??
Check your init.ora file to make sure the control_files paramter is pointing to one of the correct control files. If the control file was toasted, you will have to do an incomplete recovery and re-create your control files.
It is a toast Jeff..
I check the location of control files and wow..
They dont exist...
I cannot even come to mount stage if I have to do incomplete recovery or even creating a controlfile..
So, what's next??
Thx..
Continuation to the above scenario::
The user ran the database script and when the problem occured he promptly deleted the control files manually and reran the script.. so i guess nothing can be done..
thanks..
1. startup nomount
2. issue create controlfile command (http://technet.oracle.com/docs/products/oracle8i/doc_library/817_doc/server.817/a85397/statem4b.htm#2061202)
3. shutdown immediate
4. startup mount
5. alter database recover until cancel using backup controlfile
6. alter database recover cancel
7. alter database open resetlogs
Create only the control file and you should be able to recover it with above steps
Jeff and sreddy:
Please check the syntax underneath..
tried creating a controlfile and it errors out when i did not mention the datafile name (system.dbf)..
But when i included system.dbf again it errored out..
where am i doing wrong??
The logfile destination, i copied exactly from our create database script so that i will not make a mistake..
*****
create controlfile reuse
database nbs
logfile group 1 '/u02/oradata/nbs/redo01.log' size 5120k,
group 2 '/u02/oradata/nbs/redo02.log' size 5120k,
group 3 '/u02/oradata/nbs/redo03.log' size 5120k
noresetlogs
datafile '/u01/oradata/nbs/system.dbf' size 500M
character set WE8ISO8859P1;
[oracle-dominator]/export/home/oracle>
SVRMGR> @/export/home/oracle/create.ctl
create controlfile reuse
*
ORA-01503: CREATE CONTROLFILE failed
ORA-01565: error in identifying file '/u01/oradata/nbs/system.dbf'
ORA-27037: unable to obtain file status
SVR4 Error: 2: No such file or directory
Additional information: 3
SVRMGR>
thx..
Here you see my production 734 version controlfile backed up to trace. Is system file is the only one you have in the database ? do you have other files too..... you have to include all the files of database.
------------------from here---------------------------
STARTUP NOMOUNT
CREATE CONTROLFILE REUSE DATABASE "HRTS" NORESETLOGS NOARCHIVELOG
MAXLOGFILES 32
MAXLOGMEMBERS 4
MAXDATAFILES 1022
MAXINSTANCES 3
MAXLOGHISTORY 300
LOGFILE
GROUP 1 (
'/oracle/oradata/hrts/redo/log_1a.dbf',
'/oracle2/oradata/hrts/redo/log_1b.dbf'
) SIZE 1M,
GROUP 2 (
'/oracle/oradata/hrts/redo/log_2a.dbf',
'/oracle2/oradata/hrts/redo/log_2b.dbf'
) SIZE 1M,
GROUP 3 (
'/oracle/oradata/hrts/redo/log_3a.dbf',
'/oracle2/oradata/hrts/redo/log_3b.dbf'
) SIZE 1M
DATAFILE
'/oracle/oradata/hrts/data/system_01.dbf',
'/oracle/oradata/hrts/data/rollback_01.dbf',
'/oracle3/oradata/hrts/data/tools_01.dbf',
'/oracle3/oradata/hrts/data/users_01.dbf',
'/oracle2/oradata/hrts/data/hr11_01.dbf',
'/oracle3/oradata/hrts/data/hr11indx_01.dbf',
'/oracle3/oradata/hrts/data/temp_01.dbf'
;
[Edited by sreddy on 01-10-2001 at 07:43 PM]
Thanks for the links sreddy..
It is very clear in metalink(with syntax)..
let me ask you, once i create the control file and run it through as per the one in metalink, i dont have to follow the remaining steps of what jeff mentioned right... such as 4 to 7, coz the end of control file script includes alter db open..
once the db is open, i am in shape, right..
Am i making sense?? If not, please let me know..
Ps: The logfiles and datafiles might be having an SCN and when new control file gets created how will it share the new number.. when i dont go with resetlogs..
Thx..
Usually, we do these steps @ command mode interactively
*creating control file nomount mode
*mounting database
*recovering the database if its production/development and has backup,archivelog,wanted transactions in redolog .
*opening database
If you want to include these, the way trace file gives it can be done thru script.
Assuming,thinking there might be any problem while bringing back database from loss of cotrol file, in general people does stepwise interactively rather driven by a script.
ASA you bring db up do
SVRMGR> ALTER DATABASR BACKUP CONTROLFILE TO TRACE ;
And mirror controlfiles atleast 2 locations.
NORESETLOGS and SCN sync makes difference if you have turned on archivelog mode and you are doing recovery otherwise, no difference, as your logs will be recycled anyway as you are not moveing 'em to archivedlog. So,RESETLOGS resets your logs and starts logging with beginig SCN.
Hope you got what Jeff steps meant now. You said no archivelog, no backup. So, Jeff gave you the steps to match your situation.
[Edited by sreddy on 01-11-2001 at 12:29 AM]
consolidate the concept of recreating controlfile....
http://www.dbasupport.com/forums/sho...?threadid=5619
[Edited by sreddy on 01-11-2001 at 12:21 AM]
Thanks a lot sreddy and jeff..
Now everything is back in form..
creating the control file and recovery got done successfully.
except that i had some db links and they failed.. i am working on it now.. (listener was shut down)..