RE: Get data after redologs corruption
Date: Thu, 27 Mar 2008 10:05:40 -0700
My consultancy makes a business of helping clients recover from these types of failures. A number of points this post need clarification, including version and platform. In short, if all you have is a redo corruption that caused instance termination, you should be able to recover almost completely using tried and true techniques. The lack of details such as errors in this post make it impossible to determine the correct course of action in your case.
Ingrid Voigt (giantpanda_at_gmx.net) wrote:
> A virus scanner ran over the database causing it to crash (shutdown abort)
Did it crash or did you shutdown abort? If it crashed, what errors? If you shutdown abort, what prompted you to do so?
> ...and shredded all the redologs. And all recent backups.
What does "shredded" mean, exactly? What was the exact problem with the redologs? How would a virus scan damage backups?
> I was able to open the database with
> _allow_resetlogs_corruption = TRUE in init.ora and
> recover database until cancel;
> cancel immediately
> alter database open resetlogs;
Wait, what happened when you tried to open the database normally? What errors did you receive? Did you consider trying the method from the documentation for overcoming logfile corruption?
> It crashes again when trying to do any work.
With what error? This is important. There are things that you can do to keep from crashing a database opened with _allow_resetlogs_corruption=true.
> After repeated attempts
> I could (once) get an export of the test schema for a fresh database,
> but this was not reliable.
Again, what was the error you were getting? What prevented you from exporting the whole database?