Re: Oracle DB cannot restart due to missing REDO-file

From: Antoine BRUNEL <antoinebrunel_at_nospam_yahoo.fr>
Date: Mon, 17 Jan 2005 22:45:19 +0100
Message-ID: <41ec31ef$0$19070$79c14f64@nan-newsreader-06.noos.net>

Hi from Paris

if you did not destroyed the db too much (by editing logs files for example), you can still open it in order to full export it:

- put the _allow_resetlogs... parms
- mount the instance, and begin a recover session with "... until cancel"
- when you have reached the point of no return (last archived applyed, or in 

your case, at the first required), you issue the cancel

- you should normally be able to do an "alter database open resetlogs"
- online redo logs will be recreated, and database be opened
- ... this db is now corrupted (as indicated in the alert.log): pray, and 

export it as soon as possible...

"T. Lehrmann" <Thomas.Lehrmann_at_gmx.de> a écrit dans le message de news: 350gugF4g0c3hU1_at_individual.net...
> Thanks also for your advice, Antoine,
>
> I already tried out the _allow_resetlogs_corruption -flag in the init.ora
> file but it didn't help....
>
> I wasn't very familiar with Oracle, all I knew till now was that log-files
> aren't vital for programs to continue to operate.. now I know, there are
> some exceptions.. at least concerning Oracle and its way of naming files
> (i.e. file extensions). Nevertheless it was clearly my fault, and the DB
> seems lost, so I paid my dues...
>
> "Antoine BRUNEL" <antoinebrunel_at_nospam_yahoo.fr> schrieb im Newsbeitrag
> news:41eb0901$0$17653$79c14f64_at_nan-newsreader-06.noos.net...
>> Hi from Paris
>>
>> if you losed online redo logs, you have two options:
>> - do a full database restore, and apply archived up to when you dropped
>> redo logs... this is an incomplete recovery (recover... until cancel)
>> - open the database without doing the crash recover, which will introduce
>> corruption (as commited transactions may have not been fully applyed to
>> dbfiles)... this is an unsupported way, you have to open an iTar at
>> Oracle Support, and ask them for _allow_resetlogs_corruption. Anyway
>> after that, the current DB will be considered as unsupported, and you
>> will have to export full the DB, and create a new to one to import data.
>>
>> whatever option you take, take a whole backup before doing anything...
>>
>> But: NEVER, NEVER, NEVER modify Oracle files with an hex editor... this
>> would the perfect way to totally broke down you db....
>>
>>
>>
>> "T. Lehrmann" <Thomas.Lehrmann_at_gmx.de> a écrit dans le message de news:
>> 34vkk0F4envhqU1_at_individual.net...
>>> Hello,
>>>
>>> I have a big problem with my Oracle DB (V 9.1): the service cannot start
>>> and when trying to connect Oracle error message "ORA-01033" pops up,
>>> saying "ORACLE initialization or shutdown in progress". In the log file
>>> is written:
>>> "Errors in file c:\oracle\admin\DBNAME\bdump\DBNAME_lgwr_480.trc:
>>> ORA-00313: open failed for members of log group 2 of thread 1
>>> ORA-00312: online log 2 thread 1: 'C:\ORACLE\ORADATA\DBNAME\REDO02.LOG'
>>> ORA-27041: unable to open file
>>> OSD-04002: File cannot be opened
>>> O/S-Error: (OS 2) System cannot find the specified file"
>>>
>>> This was casued because I deleted the 3 REDO-logfiles REDO01.log,
>>> REDO02.log and REDO03.log of my database (I am no real DBA, I thought
>>> Oracle only needs this files if one wants to recover a DB!).
>>>
>>> Apparently, Oracle recreated REDO01.LOG, but not the other two files and
>>> now has problems to even start the DB.
>>>
>>>
>>> Is there any way (e.g. nologging adjustments in init.ora) to get my DB
>>> started again?
>>>
>>>
>>> Any help will be appreciated! Thanks in advance,
>>> Thomas
>>>
>>
>>
>
>
Received on Mon Jan 17 2005 - 15:45:19 CST