Re: db block size, too big wasting buffer?

In article <w5YR8.20600$Hj3.64863_at_newsfeeds.bigpond.com>, you said (and I quote):
> Hi Nuno,
>
> Can you expand a little on the problems you've had with larger block sizes
> (16K). I'm currently looking at setting up a 9i DB for a client
> (predominantly data warehouse application on NT) and would be interested to
> hear of issues you've had with 16K block sizes.
>

If you can format your datafile partitions using a sector size of 16K, then I guess it wouldn't be too much of a problem. The problem I had was with partitions with sector size of 4K and db block size of 16K.

At first one could say that being a multiple of the sector size (sector size is the unit of I/O NT will use, like it or not), it shouldn't have been a problem. But it was. The read-ahead characteristics for a 16K db block size are completely different from what the NT cache will try to do.

I was getting huge amounts of I/O, way above what it should have been with such a large block size. Changed the partition sector size to 8K and the db block size to 8K and all was sweet.

Moral of the story? Format a partition with the EXACT size of db block you're gonna use in it.

Oh, and of course: do NOT put normal NT folders/files in a partition formatted with a sector size of 8K! Keep it for database files.

-- 
Cheers
Nuno Souto
nsouto_at_optushome.com.au.nospam