Bogusław Brandys wrote:

>
> I feel that all those errors are related each other (the same table for
> example ). Could it be that the reason of broken index is something
> which caused two others problems ?
> Any possibility that my assumption is correct ?
>

My guess is that there's a problem that happens while splitting an
upper level index page. Again, guessing, under some circumstances,
the page isn't being removed from the free list before being written
to the index. That behavior would be invisible in superserver, but
would show up in classic, depending on how the free space page and
the index page were being requested elsewhere. But that's just a
guess and the code there is really complicated... without a repeatable
test case, it will be very hard to find.


Regards,


Ann