Re: Corrupt indexes on Oracle databases

I'm currently upgrading from 7.1.5 to 7.3.2 on Alpha VMS. During our startup on 7.1.5 we encountered many (and easily reproducable) ORA-600/ORA-1578 for currupt data blocks. After 3 weeks of Oracle/DEC name calling we ended up solving OUR problem by breaking our then current RAID-5's down to RAID-0's. The error never occured when we created the datafiles on the RAID-0+1 use for Logs and Rollbacks. We are now running all 80Gigs on RAID-5's again without the previous problems. I don't know yet is these problems will occur again under 7.3 or ever.

Our architecture:

All Drives TZ29
2 Dual-Redundant HSZ40 controllers - 20 Gig Raid-5's Alpha 8400
2 300 CPUs
1 Gig RAM

SGA's

1 - 208 Meg - Warehouse w/ errors
1 - 50 Meg
1 - 30 Meg

My suggestions: Review your storage. Drop the size of your SGA (if above 200 Megs). (These are not based on scientific evidence, just hunches).

Mike Brown

G McLauchlan wrote:
>
> We have been having a problem with corrupt indexes on various Oracle
> databases at our site. We are running RDMS version no. 7.1.32. All
> databases are run under VAX/VMS systems. The problems are cured by droping
> the indexes and recreating them.
>
> In the trace file we are seeing the message :-
>
> Corrupt block=3d000fcf
> .
> .
> .
> ORA-00600: internal error code, arguments [17090], [], [], [] etc.
>
> Has anyone else experienced similar problems or has anyone any suggestions
> ?
>
> Thanks
>
> Gavin McLauchlan
>
> Email
>
> Mclauchlan_g_at_grpr21.dnet.bp.com
Received on Tue Feb 25 1997 - 00:00:00 CST