Re: 8.1.6 datablock corruption

From: yong huang <yong321_at_yahoo.com>
Date: Thu, 25 Jan 2001 08:43:17 -0800 (PST)
Message-Id: <10752.127502@fatcity.com>

Hi, Stephen,

This is interesting. For our information, could you tell us more about your system? How much memory do you have? Your shared_pool_size, db_block_size and db_block_buffers. Any hidden parameter in your init.ora such as _use_ism?

I ask all these because 8.1.6 Oracle installation guide says setting shmmax to 4G regardless of anything, while in previous versions it is more conservative and says setting to 2G or based on some calculation (I don't remember which).

Yong Huang
yong321_at_yahoo.com

you wrote:

we are getting oracle datablock corruption on the SYSTEM tablespace during perio
ds of intense insert / update activity on other tablespaces.

We are running 8.1.6 under Solaris 2.7.

Typical errors in the alert file are ...

ORA-00602: internalprogramming exception ORA-07445: exception encountered: core dump [kghsrch()+56][SIGSEGV] [Address not mapped to object] [4294967288] [] [] Fri Jan 19 21:11:212001
Errors in file/u01/app/oracle/admin/FS75TST/udump/fs75tst_ora_7899.trc:

ORA-01578: ORACLE datablock corrupted (file # 1, block # 13968)
ORA-01110: data file 1:'/u02/oradata/FS75TST/FS75TST_system01.dbf'
ORA-00607: Internal error occurred while making a change to a data block
---------------------------------------------------------------------

Changing the UNIX shmmax parameter from 4GB down to 3.5GB seems to have helped.

Any thoughts

Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. Received on Thu Jan 25 2001 - 10:43:17 CST