My worst nightmare - ORA-8103

Original post to BOracle list...apologies to those who are seeing this again.

On 5/31, the refresh of a materialized view in our reporting instance failed with an ORA-8103 error, as did a gather stats job for the same table in the production database. We had an application upgrade done on 6/1 and copied our production database to a preprod version on 6/2, so the errors weren't caught immediately.

Users complained that the current data wasn't available in the reporting instance but my attempts to refresh and recreate all failed with the same ORA-8103 error.

I've been working with Oracle Support since early last week and keep hitting brick walls. At the moment, we've managed to copy 'uncorrupted' rows out of the table in our preproduction database and are 480 rows short in a table with 550M rows in it. Oracle is currently working to identify the rowids of the corrupt blocks based on the output in a trace file generated by a failing query.

I'm looking at Document 336133.1 while I wait for more info from Oracle.

Anyone have any comments/suggestions/other info that might help identify and fix the problem? We really can't afford to lose that much data. We're working on recovering our database to a different location to try to get back any data that we lose, but since I don't have any idea what caused the corruption, I'm lost.

• Maureen -- http://www.freelists.org/webpage/oracle-l