This entry was posted on Thursday 19 April 2007 at 2:45 pm and is filed under rman. You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
2 Responses to “Data block corruption cleared with alter system flush buffer_cache”
Hi Andrew,
this is very interesting. I had a similar efect at a customer’s site recently with 10gR2 where I encountered an object that “didi not exist any more”.
It was possible to describe it but a SELECT COUNT(*) returned an ORA-600.
There were no extents in DBA_EXTENTS and it was not possible to DROP it: also ORA-600.
Since my customer did not use RMAN we ended up with a full export except this one object and full import into a new DB.
Obviousely the Data Dictionary was corrupted.
Andrew Fraser 
Wednesday 16 May 2007 at 7:27 am
Hi Anrew,
it is always a good idea to check proactively for corrupt block (both physical & logical corruption) with RMAN.
Pls see my post about it:
http://sysdba.wordpress.com/2006/04/05/how-to-check-for-and-repair-block-corruption-with-rman-in-oracle-9i-and-oracle-10g/
=;-)
Lutz
Monday 10 December 2007 at 11:09 pm
Hi Andrew,
this is very interesting. I had a similar efect at a customer’s site recently with 10gR2 where I encountered an object that “didi not exist any more”.
It was possible to describe it but a SELECT COUNT(*) returned an ORA-600.
There were no extents in DBA_EXTENTS and it was not possible to DROP it: also ORA-600.
Since my customer did not use RMAN we ended up with a full export except this one object and full import into a new DB.
Obviousely the Data Dictionary was corrupted.
I guess it’s a BUG!
Or is it a feature?
=;-)
Lutz