rman and dbv disagree on datafile quality ?

Hi!

We have gone through some database corrupt blocks in the last months, and we have seen that in some cases, RMAN and DBV disagree. Running a level 0 backup with RMAN will tell me that I have a bad block on one datafile (the backup will stop at that moment). If I run DBV on the datafile, it will return totally clean, no trace of block corruption. In the view V$DATABASE_BLOCK_CORRUPTION there is nothing. There must be something in the datafile, but rman does not say where it is located the bad block.

In Metalink I have found recipes for bad blocks, reported by both rman and dbv (when they agree), that do not appear in DBA_EXTENTS as being allocated to any database object, but I can not apply those recipes to my case.

So apart from recreating the tablespaces, I have no clue how to get rid or reinitialize those bad blocks. Any way to force rman to log more info on the bad blocks ?

Any idea would be very much appreciated.

ciao.

p.s we run 10.2.0.2 on Red Hat 3.x
--

Nilo Segura
Oracle Support - IT/DES
CERN - Geneva
Switzerland

--

http://www.freelists.org/webpage/oracle-l Received on Thu Mar 15 2007 - 20:31:04 CDT