"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<4064b567$0$3953$afc38c87_at_news.optusnet.com.au>...
> "Joel Garry" <joel-garry_at_home.com> wrote in message
>
> > My mistake. Somehow I thought that Oracle marked blocks in their
> > headers when they were no longer needed,
>
> The function of the segment header block.
>
> Regards
> HJR
I guess I need it explained even slower to me then. If Oracle knows
the block is no longer needed, why does it need to physically read it
before writing a new one in the buffer? Does it need to resolve that
there isn't a still-used part of the block from a different
transaction? Is there some weird consequence of multiblock reads?
jg
--
@home.com is bogus.
http://www.bizarro.com/stuff/puzzles.htm
Received on Mon Mar 29 2004 - 15:05:33 CST
