Re: Recovery: Nightmare Scenario

From: Howard J. Rogers <howardjr_at_www.com>
Date: Mon, 18 Sep 2000 13:19:41 +1000
Message-ID: <39c57b57@news.iprimus.com.au>

I'm a bit confused as to why you were apparently asked to supply the archived redo logs, since the backup controlfile to trace method is simply a script that recereates all copies of your control files in the locations specified by the relevant init.ora. There should be no application of redo required.

So, assuming you have your trace file handy, and assuming you've stripped out the rubbish at the top of the file, added a connect string, and (perhaps) specified the path to the init.ora on the STARTUP NOMOUNT line, you should be able (1) to delete all copies of the existing control file (because obviously the versions are now all over the place), (2) fire up Server Manager, and (3) type @nameoftracefile.

If you have ONLY done exactly what you've described here, that should do it.

Trying to do any recovery at all when you *know* the issue is one of dodgy control files is really piling on the agony. Must have been a very bad Monday.

Regards
HJR
--

-

Opinions expressed are my own, and not those of Oracle Corporation Oracle DBA Resources:
http://www.geocities.com/howardjr2000

  "s.hunter" <sarahxhunterYOURPANTS_at_hotmail.com> wrote in message news:39BE17F5.F2D76824_at_hotmail.com...
  Hello,
  I would like to pretend that this is a hypothetical situation, but unfortunately, it's not...

A colleague came to me with an ORA-00214 in his database when trying to start up an instance:

control file '/u02/oradata/DATA/control01.ctl' version 368714 inconsistant with file '/u03/oradata/DATA/control02.ctl' version 367409

Now, I believe the correct course of action would have been a full cold backup, followed by copying controlfile 01 to the other controlfile. However, I didn't take a backup, and I copied controlfile02 over 01 . Therefore, when I tried to startup the database again I got

  ORA-01122 database file 1 failed verification check 
  ORA-01110 datafile1: '/U02/oradata/DATA/system01.dbf' 
  ORA-01207 file is more recent than controlfile - old controlfile.

So don't ask me why I did this (it was late on a crappy monday, OK?!!!) but I issued a recreate controlfile from a previous backup to trace. This produced the following ORA errors: 00279, 00289, 00280, 00368, 27037. i.e. the logs needed for recovery weren't present (no archive log mode) and so the file needed media recovery when alter db open was issued

I tried running cancel based recovery:

STARTUP MOUNT
RECOVER DATABASE USING BACKUP CONTROL FILE UNTIL CANCEL [CANCEL] but this then produced

ORA 01547 Recover suceeded but open resetlogs would get an error

The error would be ORA-01194: file 1 needs more recovery to be consistent (system01.dbf)

Is there anything anyone can suggest or am I up sh*t creek without a paddle??? (I believe it is the latter and I am going to have to reload all the data in the database)

Thanks in advance for your help anyway!

Sarah

p.s. SYSTEM INFO: SunOS running Oracle 8.0.5.0.0 on an i68pc machine p.p.s. there is no need to point out what a complete moron I have been. I realise this already!

Thanks again...

--
Received on Sun Sep 17 2000 - 22:19:41 CDT