Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Mailing Lists -> Oracle-L -> corruption theory link / RE: SAC NORAD .... how to break it?
Not sure if this helps or not, but since there is some general
discussion of how (supposedly) precarious buffering and controller
functions can be in an OS, it might be worth checking out:
http://home.att.net/~tbcox/corrupt2.txt
Problem ID : 1015744.6 Affected Platforms : Generic: not platform specific Affected Products : Oracle7 Server Affected Components : RDBMS Generic Affected Oracle Vsn : Generic
Summary:
BLOCK CORRUPTIONS IN ORACLE AND UNIX
...
It is important to realize that monitoring of hardware is required
for all operating systems. Hardware monitors can sense electrical
signals on the busses and can accurately record them even at high
speed. A hardware monitor keeps observing the system even when it
is malfunctioning, and thus, it can be used to debug the system.
(Jain, 1991 99) These tools can help determine the cause of the
problem and detect problems like controller error and media
faulting which are frequent corruption contributors. In any case,
there are many opportunities for blocks, either on disk or in the
buffer cache, to become corrupt.
[***]Fixing the corruption can
sometimes provide even greater opportunities.
Conclusion
Data block corruption is an ongoing problem on all operating systems, especially UNIX. There are many types and causes of corruptions to consider. Advanced system configurations can increase the chance and hardware problems are a common source of corruptions. When receiving block corruption errors, remember that a couple of them are
[***]not physical corruptions but memory
corruptions that are never written to disk.
Oracle Customer Support provides a number of bulletins on block corruption problems that help recover what is left of the data once corruption has occurred. If block corruption occurs on a machine, be sure to identify the type of corruption and establish a plan for its correction.
On 17 Oct 2000, at 9:27, Peter Barnett wrote:
Date sent: Tue, 17 Oct 2000 09:27:12 -0800 To: Multiple recipients of list ORACLE-L <ORACLE-L_at_fatcity.com> Send reply to: ORACLE-L_at_fatcity.com From: Peter Barnett <pnbarne_at_shep.bcbso.com> Subject: RE: SAC NORAD .... how to break it?
...
> How/why excessive paging can cause a mirror to fail is beyond my
> understanding. I also, do not know how to test it in advance.
Received on Tue Oct 17 2000 - 22:45:58 CDT
![]() |
![]() |