Re: Bug #922: PANIC: open of /scratch/production_2/pg_clog/0996

Hi,

The server (IBM) is one year old (I think).
Anyway, I found something like a workaround:
If I vacuum before executing my index script (after my loading
script has terminated), or after other huge data movements, the server
doesn't crash. So, is this typical for hardware failures?

Thanks,
Andre.

On Sun, 30 Mar 2003, Tom Lane wrote:

>
> I'd wonder about hardware problems, especially if this is a new server.
> Messages of the above type are one of the most common consequences when
> a data page has been trashed by hardware misfeasance.
>
> (While I wouldn't sit here and swear that it couldn't be a software
> problem, all of the cases I've been able to examine seemed to be
> hardware failures...)
>
> regards, tom lane
>