Re: general recommendations for setting block size?

From: Nuno Souto <nsouto_at_nsw.bigpond.net.au.nospam>
Date: Fri, 10 Sep 1999 20:19:15 +1000
Message-ID: <7raob8$2um$1@m2.c2.telstra-mm.net.au>

Thomas, the wasted space claim is not mine. Catch back on deja.com the sequence of this thread.

I do not agree entirely with it, although I do agree that having lots of partially filled 8K blocks can cause a space problem like Keneth mentioned, when space is tight. And I stand by that, having felt it in the skin at a couple of sites .Certainly worse than lots of partially filled 4K blocks.

My only objection to 8K nowadays is the I/O bandwidth problems it may generate. Which I hope may be a non-event with newer hardware.

Maybe the analogy wasn't the best one, I admit that. Then again, it's only an analogy.

Did you know that there are still many "experts" out there recommending 2K database blocks on databases that are used for a mix of OLTP and batch/history processing? I'm serious! And they claim they have the numbers to prove this... But are never too willing to show exactly how those numbers have been arrived at.

--
Cheers
Nuno Souto
nsouto_at_nsw.bigpond.net.au.nospam
Is there a nospam domain?
http://www.users.bigpond.net.au/the_Den

Thomas Kyte <tkyte_at_us.oracle.com> wrote in message news:k7=XN6RwLGEChlEtKPfKfzB=XRZ7_at_4ax.com...
>snip...
> .dat files in an OS. The analogy doesn't hold up.
>snip...
Received on Fri Sep 10 1999 - 05:19:15 CDT