On Wed, Aug 30, 2006 at 10:18:46 +0200,
Filip Hrbek <filip(dot)hrbek(at)plz(dot)comstar(dot)cz> wrote:
>
> It is possible that the corruption was caused by a HW problem at customer's
> server, and then this problem appeared also at our development environment
> because of the data already beeing corrupted. I will recommend the customer
> to make some memory tests.
You might also ask them if they are using ECC memory. A single bit flip
could have easily happened with good memory that didn't have error correction.