_ALLOW_RESETLOGS_CORRUPTION = TRUE

_ALLOW_RESETLOGS_CORRUPTION = TRUE

Post by Luis Moreno Camp » Fri, 21 Jun 2002 02:40:11



Failure Scenario:
Disk damaged until a point that we can't figure out. All datafiles,
redo logs and control files are there. I'm operating in NOARCHIVELOG
mode.
Steps taken:
1. Startup - can't open database unless with RESETLOGS
2. ALTER DATABASE OPEN RESETLOGS gives a message saying that database
needs recovery using backup controlfile
3. RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE gives the
following messages:
ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get
error below
ORA-01194: file 1 needs more recovery to be consistent
ORA-01110: data file 1: '/ora2/oradata/URS/urs_sys01.dbf'  

4. Somewhere in METALINK I saw that if this happens the following line
should be added to the init.ora:
         _ALLOW_RESETLOGS_CORRUPTION = TRUE

5. I did it all over again now with this hidden parameter set to true
and my alert.log has given me 2 ORA-600 so far, the SVRMGRL process is
taking 99.9% of machine resources, and controlfiles are beeing touched
or so it seems by their date.

6. I've waited more than I think I should wait, and now what do I do?
Kill the process, or what?

Has anyone ever been throught the same situation (or similar)???

I have Linux Red Hat 7.0, with Oracle 8.1.6.1.0 on it, and my 600
datafiles sum up to 265 GB.

NOARCHIVE MODE, no offline backup... just my pens in my hands!

Help me out if you can.

Luis M. Campos

 
 
 

_ALLOW_RESETLOGS_CORRUPTION = TRUE

Post by K Gopalakrishna » Fri, 21 Jun 2002 16:56:43


Hi:

I guess you are getting the ORA-600 [2662]. In this case you need to bump
the SCNs to higher numbers to open the database. This happens because of the
SCN mismatch in the system tablespace and other tablespaces. You need to
trace the instance recovery using the event 10015 and work accordingly.

Best Regards,
K Gopalakrishnan
Bangalore, INDIA



> Failure Scenario:
> Disk damaged until a point that we can't figure out. All datafiles,
> redo logs and control files are there. I'm operating in NOARCHIVELOG
> mode.
> Steps taken:
> 1. Startup - can't open database unless with RESETLOGS
> 2. ALTER DATABASE OPEN RESETLOGS gives a message saying that database
> needs recovery using backup controlfile
> 3. RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE gives the
> following messages:
> ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get
> error below
> ORA-01194: file 1 needs more recovery to be consistent
> ORA-01110: data file 1: '/ora2/oradata/URS/urs_sys01.dbf'

> 4. Somewhere in METALINK I saw that if this happens the following line
> should be added to the init.ora:
>          _ALLOW_RESETLOGS_CORRUPTION = TRUE

> 5. I did it all over again now with this hidden parameter set to true
> and my alert.log has given me 2 ORA-600 so far, the SVRMGRL process is
> taking 99.9% of machine resources, and controlfiles are beeing touched
> or so it seems by their date.

> 6. I've waited more than I think I should wait, and now what do I do?
> Kill the process, or what?

> Has anyone ever been throught the same situation (or similar)???

> I have Linux Red Hat 7.0, with Oracle 8.1.6.1.0 on it, and my 600
> datafiles sum up to 265 GB.

> NOARCHIVE MODE, no offline backup... just my pens in my hands!

> Help me out if you can.

> Luis M. Campos



 
 
 

_ALLOW_RESETLOGS_CORRUPTION = TRUE

Post by Mike Au » Sat, 22 Jun 2002 20:44:29


Make sure your datafiles aren't in backup mode. I realize you have
said that you are in noarchive mode so technically you shouldn't be
able to put them in backup mode, but who knows. I have a paper on
using this parameter if you would like it email me.

Mike Ault


> Hi:

> I guess you are getting the ORA-600 [2662]. In this case you need to bump
> the SCNs to higher numbers to open the database. This happens because of the
> SCN mismatch in the system tablespace and other tablespaces. You need to
> trace the instance recovery using the event 10015 and work accordingly.

> Best Regards,
> K Gopalakrishnan
> Bangalore, INDIA



> > Failure Scenario:
> > Disk damaged until a point that we can't figure out. All datafiles,
> > redo logs and control files are there. I'm operating in NOARCHIVELOG
> > mode.
> > Steps taken:
> > 1. Startup - can't open database unless with RESETLOGS
> > 2. ALTER DATABASE OPEN RESETLOGS gives a message saying that database
> > needs recovery using backup controlfile
> > 3. RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE gives the
> > following messages:
> > ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get
> > error below
> > ORA-01194: file 1 needs more recovery to be consistent
> > ORA-01110: data file 1: '/ora2/oradata/URS/urs_sys01.dbf'

> > 4. Somewhere in METALINK I saw that if this happens the following line
> > should be added to the init.ora:
> >          _ALLOW_RESETLOGS_CORRUPTION = TRUE

> > 5. I did it all over again now with this hidden parameter set to true
> > and my alert.log has given me 2 ORA-600 so far, the SVRMGRL process is
> > taking 99.9% of machine resources, and controlfiles are beeing touched
> > or so it seems by their date.

> > 6. I've waited more than I think I should wait, and now what do I do?
> > Kill the process, or what?

> > Has anyone ever been throught the same situation (or similar)???

> > I have Linux Red Hat 7.0, with Oracle 8.1.6.1.0 on it, and my 600
> > datafiles sum up to 265 GB.

> > NOARCHIVE MODE, no offline backup... just my pens in my hands!

> > Help me out if you can.

> > Luis M. Campos


 
 
 

_ALLOW_RESETLOGS_CORRUPTION = TRUE

Post by Stev » Fri, 28 Jun 2002 06:50:44


try to issue RECOVER DATABASE and when oracle prompts for the logs
provide the full path and full file name to the online redo logs on at
a time. In this manner we can trick oracle to read the on line redo
logs, hopefully all of your changes in are in the redo logs,
otherwise, email Mike for that paper on how to pop open the db.

steve


> Failure Scenario:
> Disk damaged until a point that we can't figure out. All datafiles,
> redo logs and control files are there. I'm operating in NOARCHIVELOG
> mode.
> Steps taken:
> 1. Startup - can't open database unless with RESETLOGS
> 2. ALTER DATABASE OPEN RESETLOGS gives a message saying that database
> needs recovery using backup controlfile
> 3. RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE gives the
> following messages:
> ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get
> error below
> ORA-01194: file 1 needs more recovery to be consistent
> ORA-01110: data file 1: '/ora2/oradata/URS/urs_sys01.dbf'  

> 4. Somewhere in METALINK I saw that if this happens the following line
> should be added to the init.ora:
>          _ALLOW_RESETLOGS_CORRUPTION = TRUE

> 5. I did it all over again now with this hidden parameter set to true
> and my alert.log has given me 2 ORA-600 so far, the SVRMGRL process is
> taking 99.9% of machine resources, and controlfiles are beeing touched
> or so it seems by their date.

> 6. I've waited more than I think I should wait, and now what do I do?
> Kill the process, or what?

> Has anyone ever been throught the same situation (or similar)???

> I have Linux Red Hat 7.0, with Oracle 8.1.6.1.0 on it, and my 600
> datafiles sum up to 265 GB.

> NOARCHIVE MODE, no offline backup... just my pens in my hands!

> Help me out if you can.

> Luis M. Campos


 
 
 

1. SQL Server True vs Access True...

I have a number of SQL strings that use the syntax:

"If variable = True", where variable is a "bit" column.

In Access this syntax is fine - but in SQL Server, it
raises an error that the column "True" is missing. If I
switch it to:

"If variable = 1", it works for SQL Server - but not
Access (no error, just evaluates incorrectly). Is there
any common syntax that works for both engines ? Thanks.
Mark

2. W95 computer can't see sql server

3. SQL*Loader with Direct=true Parallel=true gives error

4. Oracle8 Partitioning & Parallel Query

5. True/False fields tagged both as True and as False (???)

6. sp_setapprole

7. Checkbox on bound form sets value to -1 for True

8. Time synchronisation between two servers

9. Is this true?

10. Attaching / Detaching in TRUE path having some problem.

11. SQL bit - True /False

12. Argh Bit Values 1/True

13. How to Generate True Random IDs?