Re: Error 1578

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sun, 13 Jun 1999 19:54:33 +0100
Message-ID: <929300295.9676.1.nnrp-11.9e984b29@news.demon.co.uk>

There is an event to set in the 10,000 range (can't remember which one at present, and don't have HP running) which checks for logically corrupt blocks BEFORE writing them.

Obviously there is a CPU overhead involved, but if you set this event and get 1578s without any reports of logically bad writes then you can blame the hardware/OS.

If you get logical writes, then you absolve the hardware/OS

I don't suppose this happens in a database which you've restored from a backup does it ?

Jonathan Lewis
Yet another Oracle-related web site: www.jlcomp.demon.co.uk

Pradhan wrote in message <4FS83.3181$m93.15584_at_newsfeed.slurp.net>...
>Hi all
>
> This specific error does not occur so often. But when it occurs there is
>only way out which is drop the particular segment and recreate the
segment.
>I am trying to resolve one problem why it happens?
>
>So far, to my understanding it happens only when oracle trying to read a
>block of data and it find the discrepancies in its block header and footer.
Received on Sun Jun 13 1999 - 13:54:33 CDT