Re: Large table and buffer cache concerns

On Apr 15, 6:50 pm, DA Morgan <damor..._at_psoug.org> wrote:
> deva..._at_gmail.com wrote:
> > A single customer is only ever handled by a single application
> > server, so there should not be situations in which multiple
> > application servers are attempting to access a single block assuming
> > that block only contains data for a single customer.
>
> Why would think this a valid assumption?

I believe what Mr. Lewis was getting at was merely to bring up the potential overhead associated with updating all 1000 records for a single customer routinely. If that was the intent behind his inquiry; then a simple, no the application does not update all 1000 records for each customer at one time would have sufficed.

I may have misread the concern as being over contention by multiple threads for a single data block.

As to the assumption I'm making that a single data block contains only the data for a single customer...
If I am using a single table b*tree cluster and I've configured it to only allow the records for a single customer to be stored within a single data block, that data block will only contain records associated with a single customer. Received on Sun Apr 15 2007 - 23:21:51 CDT