Re: Sniffing redo logs to maintain cache consistency?

"Noons" <nsouto_at_optusnet.com.au.nospam> wrote in message news:Xns932A1828348D0Tokenthis_at_210.49.20.254...
> Following up on Andrej Gabara, 22 Feb 2003:
>
> Crap! This is PRECISELY why it's a bad idea to have
> Java people in charge of design!!!

<snip stuff> I very nearly wrote something very similar, but put it down to a Friday afternoon reaction.

If you must have a cache in the middle tier, then either pass all database transactions through the middle tier i.e. eliminate your stored procs/loads altogether
or
just pass the updates back to the database and let the errors if any bubble back up to your users.

But the strategy of caching data to add speed and then checking if the cache is valid every time you access it is surely going to end in tears.

--
Niall Litchfield
Oracle DBA
Audit Commission UK